Docstoc

Whole-Plan Harvest, Storage and Pretreatment of Corn Grain And Stover for Biochemical Conversion

Document Sample
Whole-Plan  Harvest, Storage and Pretreatment  of  Corn  Grain  And  Stover  for Biochemical  Conversion Powered By Docstoc
					    WHOLE-PLANT HARVEST, STORAGE, AND
PRETREATMENT OF CORN GRAIN AND STOVER                 FOR
          BIOCHEMICAL CONVERSION


                            By

                      DAVID ERIC COOK




         Thesis submitted in partial fulfillment of

           the requirements for the degree of




                    Masters of Science

             (Biological Systems Engineering)




                          at the

         UNIVERSITY OF WISCONSIN – MADISON

                           2011
                 Approved By:

__________________________________________

           Professor Kevin J Shinners

   Biological Systems Engineering Department

       University of Wisconsin – Madison

     Date: ___________________________
                                                                                              i

Acknowledgments


This manuscript is dedicated to the four pillars in my life: my Lord and Savior, Jesus Christ,
my wife, my children, and my parents. Without the support of all of them this work and
thesis would never have been accomplished. I provide the greatest thanks to God for
providing the abilities, opportunities, and constitution to complete this work. I would like to
thank Brenda Cook, my wife, partner, and best friend for all of her love, support, and
patience. I appreciate the shared sacrifice my wife and children, Jacob and Emmalyn, have
endured during this time. Without Cindy Cook, my mother, providing child care during this
time, none of this would have been possible. Thank you all for the support.

I would like to express a deep gratitude and appreciation to my advisor, Dr. Kevin Shinners,
for his mentorship, knowledge, and dedication that provided for the success of this
research. I would also like to thank Dr. Paul Wiemer, Dr. Richard Muck, Dr. Matthew
Digman, Dr. Mary Beth Hall, Dr. Ron Hatfield, Dr. Randy Shaver, and Dr. Peter Vadas for all
of the discussion and expertise provided in the design and understanding of this research.
Many thanks to my fellow graduate students Shane Williams, Jeff Duncan, Kody Habeck,
Tom Hoffman, Steve Spranger, Aaron Wepner, Joe Keene, Phipipp Lehmannm, and Craig
Slattery; and undergraduate students Justin Orrick, Monique Kauke, Jake Langer, Ryan
Blasiak, Bryan Roundtree, Chris Hargot, and Brandon Nigon. The staff at University of
Wisconsin Agricultural Research Station – Arlington deserve thanks for all of their assistance
and cooperation.

I owe a debt of gratitude to my thesis committee: Kevin Shinners, Paul Weimer, Richard
Muck, and Peter Vadas for their guidance, expertise, and time. They were instrumental in
this work and their deep knowledge of the area was an invaluable well from which I could
draw.

This research benefited from funding and support from Archer Daniels Midland, Shell Global
Solutions, and the USDA–ARS Dairy Forage Research Center.
                                                                                                ii

Abstract

       It is expected that agricultural grains, residues, perennial grasses, and manures will

become a significant portion of the US and world energy portfolio. The three primary

causes are the likelihood of reaching or having reached peak oil production, new levels of

commodity prices, and climate change mitigation policy. There are significant logistical

challenges in the harvest, transport, and storage of the proposed one billion plus tons of

agricultural and forestry biomass. Some of the primary logistical issues are, but not limited

to, low bulk density of biomass, short harvest window, labor and truck shortages, and

potentially slower grain harvesting.

       This research investigates the whole-plant harvest, transport, storage, and pre-

treatment of corn. Harvesting the whole corn plant, both grain and most of the above

ground stover, after physiological maturity can reduce the intense logistics challenges and

expanding the harvest window by a factor of two to three. Harvesting was done using a

forage harvester without a kernel processor at long theoretical length of cut, to minimize

the exposure of the starch. The combined wet grain and stover storage system investigated

here uses dilute acid as a pretreatment for the stover fraction and preservative for both

stover and grain fractions.

       Pretreatment with sulfuric acid, ensiling, and earlier harvest maturity were all found

to improve ethanol yield by SSF and reduce the hemicellulose content of the stover.

Hydrodynamic separation was found to be more effective at fractionating starch and fiber
                                                                                            iii

than hand shelling. At the pilot and farm-scale, whole-plant silage harvested from 65% to

19% MC was well conserved, rarely exceeding 5% losses.

       Economic analysis of the whole-plant silage and other fractionated systems showed

the whole-plant silage to be the most cost effective biomass system, requiring the lowest

energy inputs.
                                                                                                                                        iv

Table of Contents

  Acknowledgments .............................................................................................................. i
  Abstract ............................................................................................................................. ii
Chapter 1 – Introduction and Literature Review ................................................................1
  Introduction ......................................................................................................................1
  1.1      Corn as a Biomass Feedstock ...................................................................................1
  1.2      Silage Storage System..............................................................................................7
  1.3      Pretreatment ........................................................................................................11
  1.4      On-Farm Pretreatment ..........................................................................................13
  1.5      Rationale for Whole-Plant Corn Biomass System ...................................................15
  1.6      Need for Research ...............................................................................................17
  1.7      Research Strategy .................................................................................................18
Chapter 2 – Evaluation of Whole-Plant Grain and Stover System .......................................20
  2.1      Materials and Methods .........................................................................................20
     2.1.1        Substrate........................................................................................................20
     2.1.2        Harvest...........................................................................................................20
     2.1.3        Experimental Design ......................................................................................21
     2.1.4        Pilot-Scale Silos ..............................................................................................22
     2.1.5        Farm-Scale Silo Bag ........................................................................................24
     2.1.6        Fraction Separation ........................................................................................25
     2.1.7        Removal Procedure ........................................................................................26
     2.1.8        Sample Preparation ........................................................................................27
     2.1.9        Post-storage Fermentation Products ..............................................................29
     2.1.10        Composition Analysis ....................................................................................29
     2.1.11           SSF Ethanol Yields .......................................................................................30
  2.2      Results ..................................................................................................................32
     2.2.1        Physical Properties .........................................................................................32
     2.2.2        Pre-Storage Substrate Composition ................................................................33
     2.2.3        Storage Losses................................................................................................35
     2.2.4        Post-storage fermentation products ..............................................................38
     2.2.5        Post-Storage Composition ..............................................................................42
                                                                                                                                   v

     2.2.6        Post-Storage Starch ........................................................................................44
     2.2.7        Post-Storage Starch Digestibility .....................................................................46
     2.2.8        Post-Storage Pretreatment ............................................................................49
     2.2.9           Post-Storage Liquid Fraction .......................................................................51
     2.2.10         Post-Storage SSF ..........................................................................................52
     2.2.11          Moisture Content of Fractions Post separation ...........................................54
     2.2.12 Aerobic Stability .............................................................................................55
  2.3      Discussion and Conclusions ...................................................................................57
     2.3.1        Composition ...................................................................................................57
     2.3.2        Losses and Starch Losses ................................................................................58
     2.3.3        Pretreatment .................................................................................................59
     2.3.4        Fractionation ..................................................................................................60
     2.3.5        Scaling to Production System .........................................................................61
     2.3.6        Overview ........................................................................................................61
  References ......................................................................................................................63
Chapter 3 – Economic Assessment of Corn Stover Logistics Systems ...................................68
  3.1      Introduction ..........................................................................................................68
  3.2      Model Background and Assumptions ....................................................................70
     3.2.1 Material Description, Nutrient Removal, and Soil Quality .................................70
     3.2.2        Harvest Methods............................................................................................74
     3.2.3        Handling, Storage and Transport ....................................................................77
     3.2.4        Stover Transport ...........................................................................................87
     3.2.5        Stover Processing ..........................................................................................87
     3.2.6        Profit Margin ..................................................................................................88
  3.3      Model Results and Discussion ...............................................................................89
     3.3.1        Feedstock Costs for 670 Gg Supply .................................................................89
     3.3.2        Whole-Plant Silage System .............................................................................99
     3.3.3        Sensitivity Analysis ......................................................................................102
  3.4      Discussion ...........................................................................................................107
  3.4.1       Comparison to Literature .................................................................................110
  3.5      Summary .............................................................................................................113
  References ....................................................................................................................115
                                                                                                                                  vi

Chapter 4 – Discussion and Conclusions ............................................................................119
   4.1      Discussion ...........................................................................................................119
   4.2 Conclusions..............................................................................................................120
   References ....................................................................................................................121
Chapter 5 – Future Work ...................................................................................................122
   References ....................................................................................................................123
Appendix A – Collection Area and Supply Curve ................................................................124
Appendix B – Additional Model Results .........................................................................133
                                                                                                                                 vii

List of Figures

Figure 1.1. Schematic representation of the typical grain harvest and logistics system used
to supply starch to grain based ethanol biorefinery. ..............................................................4
Figure 1.2. Schematic representation of the typical corn stover harvest and logistics system
used to supply feedstock to cellulosic based ethanol biorefinery. .........................................5
Figure 1.3. Schematic representation of the proposed whole-plant grain and stover harvest
and logistics system used to supply whole-plant corn to combine starch and cellulosic based
ethanol biorefinery. ...............................................................................................................6
Figure 2.1       Temperature history of whole-plant corn (grain and stover) stored in a farm-
scale silo bag. 38
Figure 2.2 Temperature of whole plant silages with biological pretreatments during
aerobic exposure following anaerobic storage and fermentation in a farm-scale silo bag. ...57
Figure 3.1.      Sensitivity of corn stover feedstock costs to fraction of total available stover
removed by eight different harvest scenarios supplying a biorefinery requiring 670 Gg
annually.        103
Figure 3.2.      Sensitivity of corn stover feedstock costs to size of biorefinery for eight
different harvest scenarios. ...............................................................................................105
Figure 3.3.      Sensitivity of corn stover feedstock cost to post storage transport density
for four different bulk harvest scenarios supplying a biorefinery requiring 670 Gg annually.
Density of bales was great enough that the load was always weight limited, so bale
scenarios were not included. .............................................................................................106
Figure 3.4.      Sensitivity of corn stover feedstock cost to harvest moisture content for
eight different harvest scenarios supplying a biorefinery requiring 670 Gg annually..........107
Figure A1. Diagram of collection area with a centralized collection point. .......................125
Figure A2.       Stover supply curve (producer participation ratio) as a function of stover
farm-gate price. Assumes explicit costs of $50(Mg DM)-1 and estimated value of $80(Mg
DM)-1.           127
Figure A3. Participation rate (i.e. supply curve) and land area required for a 227 Gg annual
stover requirement as a function of distance from processor for scenario 4 (table 3.3). ....129
Figure A4.       Mass of material supplied from each tranche for a 227 Gg annual stover
requirement as a function of distance from processor for scenario 4. ...............................130
Figure A5.       Diagram of a supply area and the dimensions used in Figure A7. The
aggregation point is shown as a star. .................................................................................131
Figure A6.       Excel equation that calculates the area of each tranche, where x represents
the distance from the aggregation point. ..........................................................................132
Figure A7.       Diagram of a collection area with a centralized point for the processor,
surrounded by intermediate aggregation points that are off-center from their geographic
centers.         132
Figure B1.       Sensitivity of diesel fuel required to harvest moisture content for eight
different harvest scenarios supplying a biorefinery requiring 670 Gg annually. .................133
                                                                                                                                 viii

Figure B2.        Sensitivity of final transport distance to harvest moisture content for eight
different harvest scenarios supplying a biorefinery requiring 670 Gg annually. .................134
Figure B3. Sensitivity of diesel fuel required to fraction of stover removed for eight
different harvest scenarios supplying a biorefinery requiring 670 Gg annually. .................135
Figure B4. Sensitivity of final transport distance required to fraction of stover removed for
eight different harvest scenarios supplying a biorefinery requiring 670 Gg annually..........136
Figure B5. Sensitivity of feedstock cost on a steam equivalent energy basis to harvest
moisture content for eight different harvest scenarios supplying a biorefinery requiring 670
Gg annually. ......................................................................................................................137
Figure B6. Sensitivity of final transport distance to annual biorefinery mass requirements
for eight different harvest scenarios supplying a biorefinery requiring 670 Gg annually.
Scenarios 6 and 1 reside directly behind scenarios 7 and 5, respectively. ..........................138
                                                                                                                                         ix

List of Tables


Table 2.1. Experimental design and level of treatments for experiments conducted in 2009
and 2010. ............................................................................................................................22
Table 2.2. Assays conducted on samples pre- and post- storage for experiments conducted
in 2009 and 2010. ................................................................................................................28
Table 2.3. Geometric mean particle size and fraction kernels left intact after harvesting
with two types of forage harvesters. Results are averages across all harvest dates for both
years. ..................................................................................................................................32
Table 2.4. Pre-storage composition of whole-plant corn (grain plus stover) and separated
grain and stover fractions harvested in 2009 and 2010 and shell corn harvested in 2010,
averaged across all experiments conducted in each year.....................................................33
Table 2.5. Pre-storage water soluble carbohydrate (WSC) content of whole-plant silage
(grain plus stover) used in 2010 experiments. .....................................................................34
Table 2.6. Storage losses of whole-plant silage (grain plus stover) as percent of initial OM
for 2009 harvests and treatments. ......................................................................................36
Table 2.7. Storage losses of whole-plant silage (grain plus stover) as percent of initial OM
for 2010 harvests and treatments. ......................................................................................37
Table 2.8. Storage losses of whole-plant silage (grain plus stover) harvested and stored in
2010 in a farm-scale silo bag[a]. ............................................................................................37
Table 2.9. Post-storage fermentation products for 2009 harvests of whole-plant silage
(grain plus stover) after 120 days in storage for control with no pretreatment and
pretreatment with sulfuric acid at a rate of 100g (kg DM)-1. ................................................40
Table 2.10. Post-storage fermentation products averaged across three harvest dates in
2010 for whole-plant silage (grain plus stover) after 60 days in storage for control with no
pretreatment and pretreatment with sulfuric acid or lime. .................................................40
Table 2.11 Post-storage fermentation products averaged across three harvest dates in 2010
for grain fraction hydrodynamically separated from the whole-plant silage after 60 days in
storage for control with no pretreatment and pretreatment with sulfuric acid or lime. .......41
Table 2.12. Post-storage fermentation products averaged across three harvest dates in
2010 for stover fraction hydrodynamically separated from the whole-plant silage after 60
days in storage for control with no pretreatment and pretreatment with sulfuric acid or
lime. ....................................................................................................................................41
Table 2.13. Post-storage fermentation products whole-plant silage (grain plus stover) for
2010 harvest after 60 days in storage for control with no pretreatment and pretreatment
with biological amendments. ..............................................................................................42
Table 2.14. Composition of whole-plant silage (grain plus stover) and separated stover and
grain fractions after anaerobic storage for silage with no pretreatment averaged across all
harvests in that year. ...........................................................................................................43
                                                                                                                                          x

Table 2.15. Composition of whole-plant silage (grain plus stover) for control with no
pretreatment and pretreatment with sulfuric acid or lime averaged across all harvests in
that year..............................................................................................................................44
Table 2.16. Composition of separated grain and stover after anaerobic storage in farm-
scale silo bag for silage with no pretreatment[b]. .................................................................44
Table 2.17. Ratio of post-storage to pre-storage starch following pretreatment and
anaerobic storage for sulfuric acid pretreated and control treatments for 2009 harvests. ...46
Table 2.18. Ratio of post-storage to pre-storage starch following pretreatment and
anaerobic storage for sulfuric acid and lime pretreated and control treatments for 2010
harvests...............................................................................................................................46
Table 2.19. Starch digestibility of the 2010 grain fraction following storage and
hydrodynamic separation, as measured by percent of starch disappearance during modified
7 h IVSD assay. ....................................................................................................................48
Table 2.20. Starch digestibility of the 2010 grain fraction following storage in a bag silo and
hydrodynamic separation, as measured by percent of starch disappearance during modified
7hr IVSD assay. ....................................................................................................................48
Table 2.21. Post-storage water soluble carbohydrate (WSC), ash, and sulfur composition .50
of whole-plant silage, stover, and grain fractions ................................................................50
Table 2.22. Ratio (%) of the post- to pre-storage content of hemicellulose or cellulose in
the separated stover fraction for 2009 harvests. .................................................................51
Table 2.23. Ratio (%) of the post- to pre-storage content of hemicellulose or cellulose in
the separated stover fraction for 2010 harvests. .................................................................51
Table 2.24. Components from substrate that were dissociated into the hydrodynamic
separation liquid fraction, as a percent of the component of the whole. .............................52
Table 2.25. Comparison of pretreatments and no pretreatment prior to anaerobic storage
and hydrodynamic fractionation on corn stover ethanol yield and percent of stover cellulose
converted to ethanol by SSF. ...............................................................................................53
Table 2.26. Comparison of effects of harvest date in 2009 on percent of stover cellulose
converted to ethanol by SSF following pretreatment, anaerobic storage, and hydrodynamic
fractionation. ......................................................................................................................54
Table 2.27. Effects of harvest date and moisture on the moisture content of stover and
grain fractions immediately following hydrodynamic separation. ........................................55
Table 2.28. Stability during two durations of aerobic exposure as quantified by heating
degree days (°C) and dry matter losses using whole-plant silage (stover plus grain) stored in
farm-scale silo bag. ..............................................................................................................56
Table 2.29. Mold and yeast counts following anaerobic storage and subsequent aerobic
exposure of two durations for whole-plant silage (stover plus grain) stored in farm-scale silo
bag. .....................................................................................................................................56
Table 3.1. General assumptions and variable ranges for corn stover harvest. ....................72
Table 3.2. Removal ratio, price, and specific cost for replacement of key soil
macronutrients....................................................................................................................74
Table 3.3. Stover harvest scenarios modeled. ....................................................................81
                                                                                                                                        xi

Table 3.4. Machine configurations and assumptions. .......................................................83
Table 3.4. Machine configurations and assumptions (continued). ......................................84
Table 3.5. Time for queuing, loading, and unloading for corn stover transport systems.....85
Table 3.6. Specifications and assumptions for corn stover storage. ..................................85
Table 3.7. Summary of median transport distance, and cost plus profit to harvest, handle,
store, transport and process a 670 Gg annual supply of corn stover. ...................................95
Table 3.8. Cost to bale, deconstruct bale, and size-reduce or cost to size-reduce stover with
SPFH. ...................................................................................................................................96
Table 3.9. Summary of diesel fuel, labor and capital requirements to harvest, handle,
store, transport and process a 670 Gg annual supply of corn stover. ...................................97
Table 3.10. Cost plus profit to load, handle and transport stover in bale or bulk form. .......98
Table 3.11. Farm-gate costs ($(Mg DM)-1) of biomass for use on farm as roughage feed or
animal bedding. ...................................................................................................................99
Table B1. Polynomial and linear models of the final transport distance (km) as a function of
annual biorefinery mass requirements (Gg) for eight different harvest scenarios supplying a
biorefinery requiring 670 Gg annually (fig. A6). .................................................................139
                                                                                              xii

Nomenclature

   DM      dry matter
   LRB     large round bale or baler
    LSB    large square bale or baler
    MC     moisture content – wet basis
  MOG      material other than grain -- all non-grain material passing through the combine.
  PTFH     pull-type forage harvester
   OM      organic matter
   SOC     soil organic carbon
  SPFH     self-propelled forage harvester


Definitions

Biomass Cost      The modeled fixed and variable costs, including soil amendment value,
                  depreciation, insurance, interest, and taxes. This is considered to be a
                  producer’s median cost and uses the geometric mean haul distance to
                  assess hauling costs.

Biomass Value     The median expectation producers would have for stover residue value,
                  process costs, plus profit up to the point indicated. This is calculated for
                  the geometric mean haul distance and the value of storage losses.

Biomass Price     The price offered for biomass, only used in relationship to the supply
                  curve. (Some producers may value biomass below the median
                  expectation (biomass value), thus sell at a lower price. Likewise all
                  producers will experience different hauling costs, so their farm’s stover
                  value is altered from the median.)

Feedstock Cost    The biomass price, plus the cost of preparing the as-received biomass
                  into a form that is useable in the downstream processing or use.

Final Transport   The transportation from the storage site to the processor

Process Cost      The modeled cost (fixed and variable, including depreciation, insurance,
                  and interest) of an operation(s) e.g, the harvest process cost would be all
                  the costs associated with the harvest -- labor, equipment, diesel, etc. --
                  but not the intrinsic value of the stover being harvested.

Process Price     The modeled Process Cost plus the defined profit margin over costs.
                                                                                             1



            Chapter 1 – Introduction and Literature Review
Introduction
       There is considerable interest and work being done to develop and commercialize

renewable bio-based feedstocks to replace fossil fuel feedstocks for purposes of

transportation fuels, electrical generation, and chemical production. It is expected that this

change in feedstocks will reduce our dependence on foreign and/or non-renewable energy

sources, promote rural development, and reduce the amount of greenhouse gases emitted.

Today corn starch is the primary feedstock for ethanol production, because simple sugars

from hydrolysis of corn starch are cheaper than from lignocellulosic sources. This economic

difference is due to high pretreatment and enzyme costs for lignocellulosic feedstocks (Yang

and Wyman, 2008); however feedstock costs that are sufficiently low can aid in overcoming

some of these barriers.




1.1    Corn as a Biomass Feedstock
       In the US alone, the average area planted to corn was 35 million hectares between

2007 and 2011. In 2010, there was 35 million hectares of corn planted, yielding 267 Tg DM

of corn grain (USDA-NASS, 2011). During the 2009-2010 marketing year, US corn grain

production was 13,092 million bushels, and the starch from 4,568 million bushels (34.9%)

was used for ethanol production (USDA, 2011). In calendar year 2010, the US ethanol

production was 50 billion liters (Renewable Fuels Association, 2011). Ethanol yield from

corn grain was reported in 2008 as 401 l(Mg DM)-1 (USDA, 2010). Corn stover has also been
                                                                                                 2

considered as a raw material for energy production, using the traditional 1:1 stover to grain

ratio (Wilhem, et al., 2011), in 2009, the US had corn stover production of 282 Tg DM.

        It is estimated that under current technology cellulosic ethanol yields could be 289 l

(Mg DM)-1 (Kazi, et al., 2010), and theoretical yields of 472 l(Mg DM)-1 could be achieved

(USDOE, 2006). Currently, the expected costs between corn grain and stover are

significantly different, current corn grain prices are over $5(bu)-1 ($0.481(l ethanol)-1),

compared with corn stover estimated at $85.7(Mg DM)-1 ($0.229(l)-1, assuming 375 l(Mg)-1)

(Cook and Shinners, 2011). Conversion costs also vary significantly; for corn grain to ethanol

the total cost of conversion was reported as $0.206(l)-1 (Khanna, 2008). While the

conversion of cellulosic ethanol is of much speculation, Huang et al. (2009) estimates the

conversion cost to be $0.375(l)-1 and a capital investment of $0.771(l)-1, assuming a yield of

375 l(Mg)-1. These estimated total costs, with no return on investment are $0.687(l)-1 for

corn grain ethanol production and $0.72(l)-1 for a cellulosic ethanol production (assuming

the charge for depreciation and amortization was 15% of capital investment). McAloon et

al. (2000) estimated capital costs for cellulosic ethanol processing facility to be 4.9 times

greater than for grain ethanol; therefore when a 15% return on assets to the ethanol

conversion facility is considered this $0.033(l)-1 difference can be expected to increase to

$0.123(l)-1.

        Because of the nearly 300 million ton stover yield that occurs every year as a by-

product of corn production, corn stover has been looked at as the largest immediate

potential source of biomass. (USDOE and USDA, August 2011) Currently the majority of the
                                                                                                3

stover is left on the field. Corn stover that is left on the field has value as a soil amendment,

both to prevent soil erosion and as a soil nutrient source, providing phosphorus and

potassium (Petrolia, 2008). However it can also have negative effects for the following

year’s crop, such as causing tillage and planting problems, insect and disease harboring, and

slowing soil warming and drying, thus delaying the next year’s planting (Blacno-Canqui and

Lal, 2009).

       The current harvest and logistics system for corn grain for ethanol production after

the crop is near maturity is as follows: combine, transport, dry the corn, store, transport,

rehydrate in an acidic environment, milling and separation, saccharification, fermentation,

distillation, and dehydration (fig. 1.1). The generally considered method for corn stover to

ethanol production is as follows: after combining, field dry the stover, windrow, bale,

transport, store, transport, rehydrate, pretreat, saccharification and fermentation,

distillation, and dehydration (fig. 1.2). Both systems rely on drying the feedstock for

satisfactory aerobic storage followed by a rehydration. This drying is costly in terms of

weather risks, energy inputs, costs, and harvest timeliness. Additionally, the rehydration

adds to the water requirements of biorefineries. A new system is proposed here which

intends to lower the cost of both the starch and cellulose factions of the corn plant destined

for ethanol production. This “whole-plant silage” system simplifies these fractionated

systems into the following single-pass system: whole-plant harvest with a forage harvester,

transport, anaerobic storage, transport, fractionate, pretreatment, saccharification followed

by fermentation, distillation, and dehydration (fig 1.3).
                                                                                              4




Figure 1.1. Schematic representation of the typical grain harvest and logistics system used
to supply starch to a grain-based ethanol biorefinery.
                                                                                           5




Figure 1.2. Schematic representation of the typical corn stover harvest and logistics system
used to supply feedstock to a cellulosic-based ethanol biorefinery.
                                                                                              6




Figure 1.3. Schematic representation of the proposed whole-plant grain and stover harvest
and logistics system used to supply whole-plant corn to a combine starch- and cellulosic-
based ethanol biorefinery.



       Significant work has been done to develop corn stover harvest processes and

equipment, reduce transport costs and energy inputs, and develop a suitable storage

system. Single-pass systems have been considered by many as the ideal system, due to

lower soil contamination and a more efficient harvest. However, single-pass harvest

systems are potentially less economical than two-pass systems due to reduced area

productivity in high yield situations and long transport distances in low yield situations.

Utilizing a high-capacity self-propelled forage harvester (SPFH) for harvesting corn and
                                                                                                7

stover together might yield an ideal single-pass system, whereby the productivity remains

high and high yields of stover result in short transport distances.

       A common method of corn harvest for cattle feed is whole-plant harvesting with a

SPFH followed by fermentation in anaerobic storage (i.e. corn silage). The system

eliminates all the drying steps, which have the effect of reducing availability of starch and

fiber. Further, this method of harvest attains all of the much touted goals of cellulosic

biomass harvesting; a single-pass harvest for low soil contamination and high yields. The

harvest results in a non-fractionated harvest, and separation of the grain and stover

fractions would likely need to be done at some point in the process. Due to the harvest of a

wet or moist crop, the grain and stover must be stored anaerobically to prevent spoilage

and excess starch losses in the grain.




1.2    Silage Storage System
       There are two primary methods for storing a crop until it is ready for use, dry

storage or anaerobic storage. In dry storage, a crop is dried until the water activity is

sufficiently low that minimal biological activity occurs and the crop is conserved. In

anaerobic storage, the crop is preserved by two factors, the anaerobic environment that

limits biological activity to anaerobic microbes and anaerobic fermentation, which lowers

the pH and minimizes further biological activity by anaerobic microbes. The bacterial

fermentation of crop sugars to organic acids, primarily acetic and lactic acid, causes some

dry matter (DM) losses, however results in minimal energy losses.
                                                                                                 8

       In corn silage for cattle feed, the preferred harvest window depends on storage

structure, but is typically between 60-70% whole plant moisture content (MC) (Roth, et al.,

1995). This moisture range typically occurs prior to grain physiological maturity. The

reason for this preferred harvest timing is pinnacle yield and overall forage digestibility in

cattle (Wiersma et al., 1993). Also, harvesting in this moisture range insures a strong

fermentation, which helps the aerobic stability at removal and feeding. In a corn silage

harvest intended for biochemical conversion, the primary goal is high starch yield in the

grain fraction, secondarily followed by stover fraction quality. Because of the need for a

high starch yield, the harvest would likely begin after kernel physiological maturity, also

called black layer. The harvest could continue for the duration of the weather constrained

harvest season, enabling a much longer harvest window as compared to a dry grain harvest

system.

       There are several phases of the forage fermentation process, the first of which is an

aerobic phase. In this phase, organisms that reside on the forage surface are mixed with

the freshly cut forage during harvest. These are primarily aerobic microorganisms and until

an anaerobic environment excludes them through respiration, they consume water soluble

carbohydrates (WSC) and produce water, carbon dioxide, and heat. This aerobic activity

results in DM and energy losses, and the heat can build-up causing other problems if the

aerobic phase continues too long.

       The next phase of the fermentation process is anaerobic fermentation. This phase

begins as oxygen is excluded in storage and then consumed by the aerobic microorganisms.
                                                                                              9

Here acetic acid-producing bacteria consume WSC and produce acetic acid. This phase ends

when the acetic acid production has caused the pH to drop below 5, typically in the first few

days of storage. After the pH drops below 5, lactic acid-producing bacteria begin to take

over the fermentation. This phase is a prolonged phase and eventually slows to a halt when

a sufficiently low pH prevents further microbial activity. The silage is then considered

stable. Through this phase, the primary fermentation acids produced are lactic acid, acetic

acid, and propionic acid. A clostridial fermentation can also occur when a crop is ensiled

too wet, or due to poor management. Here clostridia bacteria use water-soluble

carbohydrates (WSC) and lactic acid to produce butyric acid.

       There are three important steps toward achieving low losses and high quality feed

when using silages: pre-storage, storage, and feedout. The time between crop harvest and

storage should be minimized to prevent the aerobic respiration of sugars. To achieve the

maximum quality out of storage, compacting to a porosity of 0.3-0.4 is recommended

(Holmes, 2008). Porosity is essentially a measure of the pore space in the compacted silage.

The greater the porosity the greater the extent that air can infiltrate into the compacted

mass. Thus a low porosity reduces the degree of aerobic conditions during storage. Muck

and Holmes (2006) found that increased particle size causes higher porosity and decreased

DM density. Porosity is difficult, if not impossible, to measure on a farm; therefore the

previously used density recommendation of 240 kg DM(m)-3 for corn silage is still generally

employed (Holmes, 2008).
                                                                                             10

       When a high degree of compaction in the silage is achieved, feed-out losses are

dramatically reduced. Losses during feed-out are caused by oxygen penetrating into the

silage through the surface that is exposed as feed is removed daily. During winter months

feed-out losses are minimized due to low ambient temperatures which limit microbial

activity; however during the summer, the recommended removal rate is 15 cm(day)-1.

Ruppel et al. (1995) reported that at a removal rate of 15 cm(day)-1, feed-out losses were

3%, while removal rates that are only 2.5 cm(day)-1, increased the feed-out losses to 18%. If

whole-plant corn would be used as a biomass feedstock rather than as an animal feed, the

feed-out rate is generally not a consideration as removal rates will dwarf those seen with

typical animal feed silos. Therefore the aerobic stability of the silage time between removal

from storage and use is the primary feedout-loss concern.

       Storing the whole-plant feedstock in a moist form necessitates a modified storage

environment that minimizes microbial growth that would result in respiration losses,

fermentation products, and potentially excess heating. Another consideration in a non-

ruminant feed system is microbes utilizing the readily available soluble carbohydrates and

starch to produce fermentation acids, which may be considered a loss depending upon the

conversion method. In an animal feeding situation, these fermentation acids have high

energy density and can be utilized by the animal. In a biomass feedstock system, these

fermentation acids may have value, or could pose a negative value, depending on the

conversion platform. Further, storage of corn stover or whole-plant silage has the

requirement of preserving the DM or energy content, but it may also be an opportunity to
                                                                                           11

add value in the form of a long-duration pretreatment at ambient temperatures and

pressures. When whole-plant corn is to be used as a biomass feedstock, it is expected that

the primary goal of storage will be the preservation of the starch fraction. A secondary goal

would be the pretreatment of the fiber fraction. Use of some pretreatments can prevent

fermentation, preserving the crop in storage, while pretreating the stover.

        Previous research suggests on-farm pretreatment by dilute acid is effective at

preserving the biomass, as well as improving availability of cellulose for enzymatic

hydrolysis (Digman M. F., et al., 2007). Exogenous acid can create a more rapid drop in pH

than the microbial fermentation process (Muck, 1988). This rapid drop in pH, coupled with

an anaerobic environment should preserve the readily accessible sugars and starches.




1.3     Pretreatment
        Pretreatment refers to chemical, biological, thermal, pressure, and/or physical

treatments of lignocellulosic biomass to overcome the recalcitrance of the complex cell wall

structure, making more of the cellulose and hemicellulose available for enzymatic

breakdown. The breakdown of the cell wall to monosaccharides has been reported to be

one of the primary technical and economic challenges to biochemical conversion (Mosier, et

al., 2005).

        Most saccharides contained in lignocellulosic biomass are contained in a complex

cell wall structure. This complex structure is comprised of three main components:

cellulose, a homopolymer comprised of several hundred to several thousand glucose
                                                                                               12

molecules; lignin, a heteroygeneous polymer lacking primary structure; and hemicellulose, a

class of heteropolymers that are shorter than cellulose. The heteroygeneity of lignin

prevents most enzymatic hydrolysis. This in turn makes enzymatic hydolysis of

hemicelluloses and cellulose difficult because lignin prevents many of the polymer linkages

from coming in contact with enzymes. The breakdown of the sugar polymers into

monomers is necessary for further processing into ethanol, hydrocarbons, or specialty

chemicals.

       Cellulose is a linear polysaccharide, consisting of D-glucose molecules bound

together by beta-1,4 glycosidic bonds. The glucose molecules have three hydroxyl groups,

which form hydrogen bonds with neighboring cellulose chains, resulting in microfibrils.

These microfibrils have a high tensile strength and provide rigidity for the cell wall.

Cellulose is also considered to be crystalline. Cellulases are a class of enzymes that disrupt

the crystalline structure of cellulose and hydrolyze the glycosidic bonds, breaking down

crystalline cellulose to glucose monomers. This process of breaking down cellulose is

required prior to its fermentation to ethanol.

       Hemicellulose provides a physical barrier to cellulose microfibrils, and along with

lignin, prevents cellulases from accessing cellulose by forming the lignocellulosic matrix.

There is a twofold purpose to hydrolysis of the hemicelluloses: one is to open up the matrix

to allow for the hydrolysis of cellulose to glucose, and the other is to provide for another

source of fermentable carbohydrates. Hemicellulose is composed of many C5 and C6
                                                                                              13

sugars, including glucose, xylose, mannose, galactose, rhamnose, and arabinose. For these

non-glucose sugars to be fermented to ethanol, other microbes must be employed.

       Drying of cellulose microfibrils results in the irreversible shrinking of the pore space,

reducing the accessible surface area (Esteghlalian et al., 2001). This would suggest that

biomass that has been dried will require more pretreatment to have similar yield and rate

of reaction.

       There are many types of pretreatment methods including biological, physical,

chemical, and combined physical and chemical pretreatment. Physical pretreatments, such

as milling and irradiation, work by increasing the accessible surface area by physically

degrading the biomass. Chemical pretreatments, such as acid, alkali, and oxidizing agents,

work by decomposing or removing hemicellulose, lignin, or both. Combination treatments,

such as steam explosion, liquid hot water, and AFEX, work by combining the physical and

chemical pretreatment effects. Biological pretreatments, such as rumen fluid and fungi, use

microorganisms to degrade the lignin, hemicellulose, and in some treatments the cellulose.




1.4    On-Farm Pretreatment
        Moist, bulk storage offers the opportunity for “on-farm” pretreatment of biomass.

This ambient condition pretreatment could also be utilized at regional aggregation facilities

as well. Pretreatment on-farm or at a regional aggregation facility will most likely be limited

to ambient temperatures and pressures, but has the advantage of long duration times. This

is the opposite of a biorefinery, where high temperature and pressures are required to
                                                                                                14

increase reaction rate and shorten pretreatment duration. Due to these different

conditions, it is likely pretreatment, both on-farm and in processing, will result in different

and potentially complementary effects.

       Possible pretreatment opportunities on-farm are acid or alkali chemicals, or

biological pretreatments. Biological treatments of microbials and enzymes that promote

the formation of volatile fatty acids (VFAs) have been widely researched for animal feed.

For ethanol fermentation, these methods would likely be avoided because the energy in the

VFAs, while conserved in an animal, could be considered losses in ethanol production.

Other biochemical conversion platforms that produce VFAs as an intermediate, may

consider biological pretreatments an effective method. Acid and alkali pretreatments have

been widely studied, due to their effectiveness and material costs.

       Sulfuric acid has been shown to readily hydrolyze the hemicellulose sugar arabinose.

By degrading this sugar, the cell wall matrix begins to degrade significantly (Digman et al.,

2007). Lime has the effect of swelling cellulose fibers, delignification, and solubization of

xylan, thereby improving cellulose accessibility.

       Storage pretreatment with sulfuric acid has been shown to significantly reduce the

production of total fermentation products by greater than 70%, more specifically reducing

production of lactate, butyrate, and ethanol, while increasing production of acetate

(Digman et al., 2007). This treatment was also shown to reduce DM losses in storage of

reed canary grass (Digman et al., 2010).
                                                                                             15

       Alkali pretreatment with calcium hydroxide during storage was demonstrated to

have slightly higher DM losses than no pretreatment. The effects of this pretreatment on

fermentation products were to increase butyrate and acetate production. The results on

ethanol production were mixed depending on species (Digman et al., 2010).

       Digman (2009) reported acid pretreatment in reed canarygrass and switchgrass

improved cellulose conversion to ethanol by SSF from 21% to 56% and from 13% to 25%,

respectively. Acid pretreatment was effective at both preserving substrate and degrading

the cell wall matrix to enhance cellulose hydrolysis in subsequent steps (Digman et al.,

2007). Further the residual sulfuric acid present in the substrate at removal from storage

could be used to further degrade the cell wall matrix by additional thermal processing

(Digman et al., 2008).




1.5    Rationale for Whole-Plant Corn Biomass System
       Currently dairy cattle feed is dominated by corn silage, harvested as a whole-plant,

due to the lower cost, higher yield, and more digestible feed resulting from the whole-plant

system compared to a feed dominated by dry grain, and supplemented with dry stover. In

an industrial setting where more aggressive pretreatment than the rumen is available,

delaying that harvest to maximize starch yield and facilitate fractionation of the starch from

the cellulosic biomass would be desired.

       The rationale behind this work is that a whole-plant corn silage harvest system,

similar to what is employed for cattle feed, can be employed over a wide range of harvest
                                                                                                16

moistures to significantly decrease the harvest risks and costs as compared to a dry grain,

fractionated biomass harvest. A whole-plant system would also eliminate or reduce many

of the costs associated with the traditional corn systems including significant handling

reductions; elimination of grain drying; and elimination of grain storage in capital intensive

steel bins. The system would also lengthen the harvest window; machinery can be used

more hours annually and over more acres; the onerous labor and truck requirements of

semis hauling biomass can be spread out over the longer harvest window; and harvest is

more reliable because it will be less affected by weather conditions. The reliability of

harvest has been recognized as a primary driving factor for the economic feasibility and

sustainability of a biorefinery (Kou and Zhao, 2011).

       Employing a whole-plant corn silage system changes the method of grain storage

and handling. The grain could be fractionated from the crop prior to storage; however

combined storage will have lower explicit costs and avoid grain drying costs. Grain losses in

storage are a concern in this system, and a major component of this work.

       Transportation of the combined grain and stover fractions is advantageous over a

fractionated crop transport because of the balancing of weight and volume limitations of

each; grain is significantly weight limited, while stover is significantly volume limited. By

balancing these two, the monetary and energy costs of biomass densification can be

avoided, while still achieving the objective of a legal weight load.

       Other potential advantages of the whole-plant corn biomass system include: a

flowable material is produced at harvest; less trips across the field resulting in less soil
                                                                                                17

compaction; early harvest provides a better opportunity for planting fall cover crops; silage

system is very compatible with intercropping; stover fraction will be harvested with low soil

contamination as it will never be placed on and picked up off the soil during harvest; the

machinery legacy already exists; and a longer harvest window allows for reduced systemic

risk and lower capital requirements of the logistics chain.

       It is expected that grain and stover fractionation would occur post-storage and yield

any of or combination of the following: lignocellulosic and starch ethanol feedstocks; if

harvested dry, fuel for solid-fueled boilers; low-quality fiber for incorporation into heifer

and dry cow rations; and high moisture corn for lactating animal rations.




1.6    Need for Research
       While there has been a major effort focused on reducing the cost of the stover

fraction, literature review has shown no research on modifying the harvest, storage, and

logistics systems of both grain and stover to achieve an overall lower cost system for both

fractions. Specifically, the co-harvest, transport, and storage of corn grain and stover post

physiological maturity for industrial and/or agricultural uses have not been previously

evaluated. Current research on whole-plant systems focuses on high quality animal feed

and research on energy uses focuses on split-stream harvest, transport, and storage.

       Much of the work for lignocellulosic feedstocks focuses on four areas: improving

bulk density, reducing handling, creating a flowable feedstock, and reducing the energy

inputs into the system. A whole-plant system would greatly reduce energy inputs by
                                                                                              18

eliminating crop drying inputs, immediately create desired material densities in the field,

create a flowable feedstock from the field, and significantly reduce handling. This suggests

that economic and energy input savings could be realized by this system. These reasons,

along with the robustness of the system that utilizes current equipment legacy,

necessitated this research.




1.7    Research Strategy
       The reasoning for this work on the whole-plant co-harvested silage system is from a

holistic farm system overview; this system has the potential to be the least cost method of

harvest, transport and storage of a biomass crop, utilizing the least energy inputs. The

purpose of this work is to investigate the feasibility of modifying the mature whole-plant

corn silage cattle feed system to be a viable biomass feedstock system.

       The first task was to evaluate the economic potential and bottlenecks of the system.

Initial evaluation suggested that the greater density of the combined crops resulted in lower

transport costs, harvest costs were found to be similar to grain only harvest, the bulk

handling of material reduced the costs of handling and feedstock preparation, and grain

drying was eliminated. However, losses and potentially detrimental changes in storage of

both the grain and stover could be the Achilles heel for the system. To this end, whole-plant

corn was harvested at varying maturity levels and on-farm pretreatment strategies were

employed to preserve and pretreat the corn silage. Corn was harvested at a wide range of

moisture contents to identify the potential harvest window for this corn silage system.
                                                                                               19

       The second portion of this work was to assess the change and losses in storage the

corn crop experiences. In Chapter 2, the methods and results for the storage and

pretreatment of whole-plant corn silage are presented. This work was done at the pilot

scale, as well as one trial on the farm scale. Harvest moisture contents varied between 16%

and 65% MC (w. b.). Pretreatments with sulfuric acid, calcium hydroxide, and biological

agents were evaluated, on the basis of changes to the fiber fraction composition and by

ethanol yield. Changes to the grain fraction were evaluated on the basis of total starch

losses, fractionation efficiency, fermentation acids, and pretreatment effects.

       Chapter 3 presents a holistic economic analysis for corn stover harvest, storage,

transport, through preparation for a biorefinery. Several single, two, and three pass

fractionated harvest systems were evaluated, as well as the whole-plant silage system being

proposed here. Metrics used for evaluation were economics, labor requirements, energy

inputs, and capital requirements. This holistic full-system modeling, combined with the

results from the storage experimentation, were used to validate the initial economic

assessment; i.e. that the whole-plant silage system can provide significant efficiencies and

cost reduction over fractionated harvests, where corn stover harvest is also desired.
                                                                                             20



Chapter 2 – Evaluation of Whole-Plant Grain and Stover System
2.1    Materials and Methods


2.1.1 Substrate
       Whole-plant corn was harvested from plots in 2009 and 2010 at the University of

Wisconsin Arlington Agricultural Research Station located near Arlington, WI. In 2009,

DeKalb 6169, a 111-day comparative relative maturity (CRM) corn, was planted on May 6th

and first harvest occurred on Oct. 8, 2009. In 2010, Dekalb 57-79, a 107-day CRM corn, was

planted on May 27th and first harvest occurred on Sept. 28, 2010. Subsequent harvests

were conducted as the whole-plant matured further and moisture content decreased to the

target levels. Generally speaking, 2009 had record low growing degree days; a large portion

of corn in the state, including the corn harvest for this experimentation had visible mold on

the ears at all harvest dates (Lauer et al., 2009). Throughout the state, corn harvest was

done significantly later and at higher than normal moisture contents. In 2010, growing

conditions were very good with a warmer than normal fall harvest season so the corn crop

matured earlier than normal (Lauer et al., 2010). NASS reported the corn harvest progress

5-year average on Oct. 31 in Wisconsin was 43% (2005-2009). For 2009, it was 12%; and in

2010, it was 76% (USDA-NASS, 2011).


2.1.2 Harvest
       Harvesting was done with a New Holland 900 experimental pull-type forage

harvester (PTFH) which was configured to allow the crop to bypass the kernel processor. In
                                                                                            21

2009, three different theoretical-length-of-cut (TLC) were utilized: 19, 25, and 38 mm. The

cutterhead was configured with only six knives to accomplish these TLC. The results from

2009 indicated no significant differences in storage losses or fermentation acid production

between the different TLC treatments, so in 2010 only the 19 mm TLC was used. No matter

the harvest conditions, it was observed that after processing through the PTFH, the grain

was fully removed from the cob at harvest. The primary reason for harvesting with the

shorter TLC in 2010 was to more easily sub-sample the heterogeneous stover material.

       From each harvest or treatment where TLC was varied, three sub-samples were

taken for particle-size analysis using ASABE Standard S424.1 (ASABE, 2007) and kernel

damage assessment. The sub-samples were approximately 5 kg wet weight. The kernel

damage assessment follows the methods described in Shinners et al. (2000).




2.1.3 Experimental Design
       The 2009 a 4 x 2 x 3 replicated experimental design was used to investigate the

effect of plant moisture, acid pretreatment, and TLC, respectively (table 2.1). In 2010 a 3 x

4 experimental design was used to investigate the effect of plant moisture and

pretreatment, respectively (table 2.1). Crop moisture was altered by harvesting on different

days as the crop senesced and dried. Each treatment was replicated three times per

harvest date.
                                                                                            22

                Table 2.1. Experimental design and level of treatments for
                        experiments conducted in 2009 and 2010.
                                  Pretreatment
               Treatment                                  Units                Level
                                    Material
                                                2009
               Moisture[a]                               % w.b.           65, 56, 40, 34
             Pretreatment         Sulfuric Acid        g(kg DM)-1             0, 100
                   TLC                                     mm               19, 24, 38
                                                2010
              Whole-plant
                                                         % w.b.             45, 34, 16
              Moisture[b]
                Grain
                                                         % w.b.             29, 19, 16
               Moisture[c]
             Pretreatment         Sulfuric Acid        g(kg DM)-1            0, 10, 30
                    ″                Ca(OH)2           g(kg DM)-1               10
          [a] Harvest dates were 09-Oct; 27-Oct; 12-Nov.; and 14-Dec, 2009, respectively.
          [b] Harvest dates were 28-Sept; 07-Oct; and 19-Oct, 2010, respectively.
          [c] Corresponding grain moisture content measured on harvest date.




2.1.4 Pilot-Scale Silos
       After harvest and prior to treatment and storage, the substrate was sub-sampled

and analyzed for dry matter (DM) content using a microwave oven according to ASABE

Standard S358.2 (ASABE, 2008) so that amendments, if any, could be applied on a DM basis.

       From each replicate pilot-scale silo, one sub-sample was taken for separation of the

grain and stover fractions and two sub-samples were then taken for moisture content and

two for constituent analysis. Moisture and constituent sub-samples were dried at 60C for

72 hours in a forced air oven following ASABE Standard S358.2 (ASABE, 2008). All
                                                                                              23

substrates were homogenized with a Hobart commercial kitchen mixer and pretreated, if

applicable. Pretreatment amendments were applied by top dressing over the course of the

two minutes the substrate was in the mixer. After mixing for two minutes, the chopped

whole-plant corn (either with and without pretreatment) was then transferred and packed

into 19 l plastic bucket containers, sealed for storage and stored indoors at approximately

20C for 120 days in 2009 and 60 days in 2010. These pilot-scale silos were filled with 4.3 kg

organic matter (OM) and compacted using a hydraulic cylinder and platen to a target

density of 225 kg OM(m)-3. The pilot-scale silos were sealed by a snap-on lid, with a

neoprene gasket in the rim of the lid. In 2009, plastic was placed under the lid to ensure

anaerobic conditions, however this was deemed unnecessary and not done in 2010. After

the filling and sealing was complete, the silo was weighed to the nearest 0.01 kg.

       During the 2010 harvests, from the same field, several ears of corn were hand

harvested, shelled, and the grain dried in a 60C forced air oven until they reached

approximately 13.5% moisture content (MC). Five sub-samples of hand shelled grain from

each harvest were dried for 72 h in a 60C forced air oven for constituent analysis, and

another five sub-samples were used for quantifying starch digestibility. This dry shell corn

was a dry grain control to compare the control ensiled grain samples against for starch,

starch digestibility, grain fiber, and soluble carbohydrates assays.
                                                                                            24

2.1.5 Farm-Scale Silo Bag
       In 2010, a farm-scale silo bag was used to investigate whole-plant corn biomass on

the farm-scale. A John Deere model 7800 self-propelled forage harvester (SPFH) with a

model 686 corn header was used to harvest whole-plant corn at a TLC of 19 mm and an

average whole-plant MC of 35% (w.b.). The kernel processor was removed from the

harvester prior to harvesting. Two amendments (plus control) were considered:

Lactobacillus buchneri (Lallmand Animal Nutrition Biotal 500 containing L. buchneri 40788

and Pediococcus pentosaceus 12455); and L. buchneri + enzyme (Lallmand Biotal 500 plus

glucose oxidaset; carboxymethylcellulase, xylanase; and polygalacturonase enzymes).

Biotal 500 was used to quickly produce fermentation acids, drop the pH, and improve the

aerobic stability of the removed material. The addition of the enzymes to the Biotal 500

was to produce additional soluble sugars for fermentation substrate through fiber

degradation, as well as to improve aerobic stability by producing antimicrobial compounds.

The bacterial inoculants were applied using an on-harvester Dohrmann model DE-1000

inoculant applicator. The applicator was set to deliver the Biotal 500 so that 100,000 cfu(g)-1

Pediococcus pentosaceus and 400,000 cfu(g)-1 L. buchneri were applied.

       Harvested material was transported to the storage location, and randomly collected

sub-samples of about 5 kg wet matter (WM) were placed into sub-sample parcels consisting

of polypropylene mesh bags measuring 53 cm by 80 cm with 10 mm mesh (MacMaster-Carr

part no. 9883T53). Before placing these replicate sub-sample parcels into the silo bag, two

sub-samples were collected from each mesh bag for moisture determination and one for
                                                                                           25

compositional analyses. Moisture sub-samples were oven dried at 103°C for 24 hours and

composition sub-samples were dried at 60°C for 72 hours following ASABE Standard S358.2

(ASABE, 2008). Wireless temperature data loggers (Onset model UA-001-08) were placed in

every other mesh bag to monitor temperature at a sampling rate of four times per day:

12:00 AM, 6:00 AM, 12:00 P.M., and 6:00 PM. Before placing in the silo bag, the sub-sample

parcels were weighed to the nearest 0.005 kg. Six replicate sub-sample parcels were placed

in each of the three treatments.




2.1.6 Fraction Separation
       In 2009, the fractionation sub-samples taken prior to pretreatment and storage

were dried in a forced air oven at 60°C for 72 hours following ASABE Standard S358.2

(ASABE, 2008). The sub-samples were then separated by hand into two grain fractions

(damaged and undamaged) and stover. Constituent analysis was conducted on the

separated stover and undamaged grain fractions.

       In 2010, the fractionation sub-samples (~200g DM) taken prior to pretreatment and

storage, were fractionated at harvested moisture into a "grain fraction" and a "stover

fraction" on the basis of specific gravity, using a previously developed hydrodynamic

fractionation technique (Savoie et al., 2004). The sample was placed into water and

underwent a single floatation step. All material that floated was considered the "stover

fraction" and all material that sank was considered the "grain fraction". The water was

decanted off the grain through wire cloth with an opening size of 0.086 mm; all material left
                                                                                               26

on the screen was included as part of the stover fraction. A sample of the water was taken

to evaluate any solids and solubles that dissociated into the liquid fraction.




2.1.7 Removal Procedure
       Silage was stored in the pilot-scale silos after 120 or 60 days storage in 2009 and

2010, respectively. The mass of each silo and its contents were weighed to the nearest

0.01 kg at the time of removal from storage. The contents of each silo were then removed

and homogenized by hand prior to sub-sampling. Two sub-samples from each silo were

taken for moisture content determination in a forced-air oven at 60C for 72 hours (ASABE,

2008). In 2009, a subsample of ~300g DM was taken and frozen for constituent analysis of

the grain and stover fractions. The remainder of the silage was size-reduced in a

hammermill with a 32 mm screen, sub-sampled into plastic bags, and frozen for constituent

analysis of the whole-plant silage. Another sub-sample of ~300g DM was taken for

hydrodynamic separation using the method described above.

       In 2010 the removal technique was modified to reduce sampling error in a

heterogeneous material that was too easily fractionated by mechanical handling. Upon

removal from storage, a sub-sample was taken to evaluate moisture content, and the

remainder was hydrodynamically separated, using the method described above. Total

approximate mass of each of the three fractions; grain, stover, and liquid (typically 2-3 kg

DM; 1-2 kg DM; and 40-50 kg WM, respectively) was measured to the nearest 0.01 kg and

each fraction was analyzed separately. In this way more accurate sub-sampling was
                                                                                           27

enabled, and total analysis could be inferred by summation of the three fractions. From

each of these fractions, two sub-samples were taken for moisture content and a sub-sample

was taken and frozen for further preparation and analysis. From the grain fraction, two

additional sub-samples (~1 kg DM) were taken and stored frozen for starch digestibility

assay.

          The total water used for the laboratory separation was a function of depth, more so

than a solids loading. A physical separation of greater than 10 cm was needed for the grain

and stover to separate. At the commercial scale, where water efficiency is required, a

higher solids loading could be employed, and water used for more than one fractionation in

a continuous process.




2.1.8 Sample Preparation
          Post-storage sub-samples that were frozen prior to analysis had to be further

analyzed and prepared prior to constituent analysis (tables 2.2). All of these samples were

assayed for fermentation products and pH (see below). After this, the samples were

titrated to neutral pH using 10M sodium hydroxide or 24N sulfuric acid, and the suspension

was subsequently frozen and freeze-dried. The dry pre- and post-storage samples were

then ground in a Wiley mill through a 1 mm screen. Finally, the grain samples had a portion

further sub-sampled that would be evaluated for starch content. This starch assay sub-

sample was ground in a vortex mill (Udy Corporation, Fort Collins, CO) through a 1 mm

screen.
                                                                                                 28



               Table 2.2. Assays conducted on samples pre- and post- storage for
                           experiments conducted in 2009 and 2010.
                                     Fractionated by Hydrodynamic
                                               Separation              Whole-             Shelled
                 Assay                         Grain         Stover      Liquid   Plant    Corn
                                   [b]       [c]       [c]
Pre-Storage         Total Starch          09       10        09 10                09 10
                    NDF/ADF/ADL                    10        09 10                09 10
                         WSC                09 10            09 10           10   09 10
                          CP
                          Ash               09 10            09 10           10   09 10
                         Sulfur                                              10
    Post-           Total Starch[a]         09 10            09 10                09        10
  Storage[a]        NDF/ADF/ADL                    10        09 10                09        10
                         WSC                09 10            09 10       09 10    09        10
                          CP                09 10            09 10                09
                          Ash               09 10            09 10       09 10    09        10
                          pH                       10            10          10   09
                          VFA                      10            10      09 10    09
                         Sulfur             09 10            09 10       09 10    09
                        Starch
                     Digestibility                 10                                       10
                          SSF                                09 10                          10
[a] Neutralized, freeze-dried and ground through 1 mm mill (Wiley).
[b] Sub-sample ground through 1 mm mill (Udy).
[c] Indicates that assay was conducted in 2009 and 2010, respectively.
                                                                                             29

2.1.9 Post-storage Fermentation Products
         Post-storage samples to be assayed for fermentation products and titrated were first

mixed in a stomacher (Tekmar, Stomacher Lab Blender 400). From the suspension a 1.5 mL

sub-sample was drawn and centrifuged to have a cell-free supernatant to analyze for

fermentation products . The prepared supernatant was measured by high performance

liquid chromatography (HPLC, Varian Model 410, Varian Inc., Palo Alto, CA) with a refractive

index detector, for the following fermentation products: lactate, acetate, butyrate,

proprionate, ethanol, succinate, 1,2 -propanediol, and 2,3 -butanediol. Samples were

injected (100uL) onto an organic acid column (HPX-87H, Bio Rad Labratories Inc., Hercules,

CA) and eluted with 0.015 N H2SO4 in 0.0034 M EDTA free acid at 0.7 ml(min)-1 at 45˚C.




2.1.10          Composition Analysis
         Wet chemistry constituent analysis was done by Dairyland Labs (217 E Main Street

Arcadia, WI) of prepared samples for ash-corrected neutral detergent fiber (NDF), ash-

corrected acid-detergent fiber (ADF), ash-corrected acid detergent lignin (ADL), water-

soluble carbohydrates (WSC), ash, and crude protein (CP). Fiber analysis was done using the

crucible method (Goering and Van Soest, 1970), and WSC assay was done using the phenol-

sulfuric method (Dubois et al., 1956). Dairyland Labs also conducted sulfur analysis using

inductively coupled plasma mass spectrometry. A slightly modified starch digestibility assay

was also conducted. The typical method for in vitro starch digestibility (IVSD) used by

Dairyland Labs involves grinding the dry sample through a 4 mm Wiley mill and conducting a
                                                                                             30

rumen fluid digestion assay. Due to the acid pretreatment of the samples and acid

hydrolysis of the starch that would occur during drying, the frozen samples were ground

through a 4 mm Wiley. This was followed by the rumen fluid digestion assay. Crude protein

was assayed in some water samples, but no nitrogen was found in any of the samples

assayed, so further testing was discontinued.

       Cellulose content was estimated in the samples by the difference between ash-

corrected acid-detergent fiber and ash-corrected acid detergent lignin. Hemicellulose was

estimated in the samples by the difference between ash-corrected neutral detergent fiber

and ash-corrected acid detergent fiber. All fiber assays were ash-corrected; therefore they

were non-sequential assays.




2.1.11 SSF Ethanol Yields
       Ethanol and pentose sugar yields were determined using a modified simultaneous

saccharification and fermentation (SSF) method (Dowe and McMillan, 2001). Yeast,

Saccharomyces cerevisiae strain D5a, was grown overnight in a sterile culture of 10 g(l)-1

yeast extract, 20 g(l)-1 peptone, and 50 g(l)-1 glucose at 35ᵒC while shaking at 250 rpm. The

yeast cells were harvested by centrifugation (10,000 x g for 20 min) and were resuspended

in a sterile solution of 200 mL water and 0.2 g of peptone.

       A sterile inoculation solution was made from 1.842 L H2O, 30 g peptone, 40 g yeast

extract, and 100 mL of a citrate buffer (1 M sodium citrate added to 1 M citric acid until pH

reaches 4.8). To this solution, the yeast suspension, Celluclast 1.5L (Novozymes, Bagsvared,
                                                                                                 31

Denmark) at 5 FPU(g)-1 dry stover, Novozyme 188 beta-glucosidase (Novozymes, Bagsvared,

Denmark) at 40 IU(g)-1 cellulose, and 8 mL of a sterile tetracycline solution (10 g(L)-1) were

added.

         A 1 g DM aliquot of the freeze-dried stover was added to a 30 ml media bottle and

then autoclaved. To the substrate 10 mL of the inoculation solution was added, resulting in

a solids loading of 10%. Fermentation was conducted for 72 h at 35C with gentle shaking

(150 rpm). After 72 h had elapsed, the bottles were put in ice and sampled for analysis by

the HPLC described above. The samples were prepared and assayed in the sample method

as described in the post-fermentation products section. To account for assay inefficiency

due to contamination, any lactate or acetate that was produced during the SSF experiment

was converted to potential ethanol yield on a one mole lactate or acetate to one mole

ethanol basis.

         Cellulose conversion efficiency ( ) was estimated on the basis of percent of

theoretical ethanol production from cellulose. The ethanol produced was divided by 0.569

times the sample cellulose content, 0.569 being the theoretical yield of ethanol production

from glucose on a mass basis.


                                            (           )
                                                ( )

where:           Ep    ethanol produced (g)
                 C     cellulose in the sample (g)
                                                                                             32

2.2    Results


2.2.1 Physical Properties
       The physical properties of the whole-plant corn silage harvested in 2009 and 2010

are summarized in table 2.3. The increase in theoretical length of cut (TLC) from 19 to 25

mm did result in more intact kernels; however there was no further improvement in intact

kernels when TLC was increased to 38 mm. The SPFH resulted in a shorter mean particle-

size, fewer intact kernels, and a visibly greater degree of processing than experienced with

the PTFH.

                Table 2.3. Geometric mean particle size and fraction kernels
                    left intact after harvesting with two types of forage
                harvesters. Results are averages across all harvest dates for
                                          both years.
                 Harvester Type      Intact Kernels ..          Geometric Mean
                     and TLC          % of the total           Particle Size .. mm

                                                2010

               PTFH[a] – 19 mm               40                        18.8

               SPFH[a] – 19 mm               29                        12.6

                                                2009

               PTFH – 19 mm                44 a[b]                   16.7 a[b]

               PTFH – 25 mm                 53 b                      21.5 b

               PTFH – 38 mm                 53 b                      28.5 c
               [a]    Pull-type and self-propelled forage harvesters, respectively.
               [b]    Means within columns with the different letters were statistically
                      different based on two-way analysis of variance at P = 0.05.
                                                                                                                                33

           In 2009, all three TLC were used for harvest as a substrate treatment. Results

showed that there were no statistically significant differences in the constituent and

conversion parameters and many times no numerical differences across the different TLCs,

therefore TLC treatments were pooled. The primary effect of a longer TLC was increased

variability in sub-sampling due to the substrate’s highly fractionable nature.



2.2.2 Pre-Storage Substrate Composition
           The composition and moisture content of the pre-storage substrate is summarized in

table 2.4. In 2009, approximately 52% of the whole-plant corn organic matter (OM)

harvested was grain and 48% was stover, while in 2010, approximately 60% of the whole-

plant corn OM harvested was grain and 40% was stover.


 Table 2.4. Pre-storage composition of whole-plant corn (grain plus stover) and separated
  grain and stover fractions harvested in 2009 and 2010 and shell corn harvested in 2010,
                  averaged across all experiments conducted in each year.
 Year       Fraction      Starch    WSC      Cellulose Hemicellulose        ADL       Ash
                                                                           g (kg OM)-1
 2009            Grain[a]            721            26              -[e]                -[e]                -[e]           17
   “             Stover[a]           37             53              430                 276                 67             58
      “       Whole-plant            367            38              223                 171                 35             38
                              [b]
 2010         Shell Corn             709b           33a             8a                  43a                12a             12a
   “           Grain [b, c]          722a           24b             8a                  35a                13a             11a
                  LSD[b, d]           9[d]          4[d]            5[d]               11[d]                4[d]           2[d]
      “        Stover[c]             17             33              444                 297                 77             53
      “       Whole-plant            414            39              179                 155                 45             30
[a]   Grain and stover fraction hand separated from the whole-plant mass.
[b]   Statistical analysis was only conducted on averages between shell corn and fractionated grain
[c]   Grain and stover separated hydrodynamically from whole-plant mass.
[d]   Least significant difference. Means within columns with different letters are statistically different at P = 0.05.
[e]   Fiber analysis was not quantified for the grain fraction in 2009.
                                                                                                 34

       In 2010, the grain was fractionated in two ways, by hand shelling and hydrodynamic

separation of the grain from the chopped substrate. The component analysis showed that

hydrodynamic separation of the chopped corn was more effective at concentrating the

starch in the grain fraction than hand shelling corn (table 2.4). The reason for this is likely

that during the hydrodynamic separation, many corn seed coats were observed to be

removed in the water and remained with the stover fraction, while a small amount of stover

resided with the grain fraction. The hydrodynamic fractionation process resulted in some

soluble carbohydrates going into the liquid fractionation so the WSC content of the

fractioned grain was less than that of the shell corn. Early harvest maturity resulted in

greater WSC content in the whole-plant silage at the time of harvest (table 2.5).


                      Table 2.5. Pre-storage water soluble carbohydrate
                    (WSC) content of whole-plant silage (grain plus stover)
                                  used in 2010 experiments.
                                        Whole-plant              WSC
                     Harvest Date         moisture           g (kg OM)-1
                                           % w.b.
                       28-Sept.              45                  44a
                       07-Oct.               34                  39b
                        16-Oct.                     16                         35c
                   LSD[a] (P = 0.05)                                            2
                 [a] Least significant difference. Means within columns with different letters
                     are statistically different at P = 0.05.



       The post-storage moisture content ranges, within harvest dates, were very tight and

had low variance. This has been consistently found in anaerobic storage (Williams, 2011),

as compared to widely ranging in moisture contents for bales aerobically stored outdoors

(Shinners et al., 2007).
                                                                                             35



2.2.3 Storage Losses
       Following storage, losses of DM were evaluated by the difference in the DM content

post- and pre- storage, expressed as the percent reduction of initial mass. Total dry mass

was estimated by weighing the WM of the silo contents and sub-sampling for DM content.

The DM losses were converted to OM losses by adjusting for the ash content found post-

storage.

       In 2009, there were significant differences in losses between treatments only for

material harvested at 65% (w.b.) moisture (table 2.6). On the remaining three harvest dates

there were no significant differences and losses were less than 3% of OM. However losses

were numerically lower at each subsequent harvest date. Storage losses for the economic

modeling (see chapter 3) were assumed to be 5%. The maximum allowable losses to

achieve an economic advantage over fractionated grain and stover harvest systems was

found to be 16%. In all treatments, except the control at 65% (w.b.) moisture, OM losses

were less than 3%, indicating the moist anaerobic conditions conserved the substrate well

and met the desired economic threshold.

       Storage losses were statistically less at each subsequent harvest date for all

treatments in 2010 (table 2.7). This was consistent with other anaerobic storage studies

(Shinners et al., 2010; Williams and Shinners, 2011). The acid pretreatment resulted in

numerically lower storage losses by immediately lowering the substrate pH and inhibiting

microbial growth. The lime pretreatments in the first two harvests of 2010 had high OM

losses that were a result of poor storage conditions due to a low loading of lime that was
                                                                                                                          36

insufficient to cause a pH greater than 8. By keeping the pH near neutral, the silage pH

never inhibited anaerobic microbial growth. Here, only the lime treatment from the first

harvest of the year exceeded the 5% storage loss benchmark. On the farm-scale, neither

the control, nor biological treatments exceeded the 5% benchmark (table 2.8). The

inoculants used in the farm-scale experiment had a trend toward lower losses, (p<0.25)

(table 2.8).


Table 2.6. Storage losses of whole-plant silage (grain plus stover) as percent of initial OM for
                               2009 harvests and treatments.
           Harvest Date      9-Oct        27-Oct.        12-Nov.       14-Dec.       Average by
      Moisture (% w.b.)       65             56             40           34         Treatment[b]
                       Control           5.7b               2.3a              2.4a              1.0a               2.8a
                   Acid – 10%            0.5a               2.8a              1.3a              2.7a               1.8a
                          LSD[a]          3.7               2.1               1.8               1.9                 1.2

      Average by Date[c]                 3.1a               2.5a              1.9a              1.9a          LSD[d] = 2.6

[a]    Least significant difference. Means within columns with different letters are significantly different at P < 0.05.
[b]    Data was pooled by treatment and analyzed using two‐way analysis of variance.
[c]    Data was pooled by harvest date and analyzed using two‐way analysis of variance.
[d]    Least significant difference for pooled data averaged across harvest date. Means within row with different letters are
       significantly different at P = 0.05.
                                                                                                                         37

          Table 2.7. Storage losses of whole-plant silage (grain plus stover) as percent of
                           initial OM for 2010 harvests and treatments.
               Harvest Date 28-Sept.           7-Oct.        19-Oct.         Average by
           Moisture (% w.b.)        45           34             16          Treatment[b]
                        Control           4.3a             3.2ab              1.7b                   3.1ab
                      Acid – 1%           3.4a              2.8a              0.9ab                   2.4a
                      Acid – 3%           2.7a              1.8a              1.1ab                  1.9a
                      Lime– 1%            7.5b              4.6b               0.8a                  4.3b
                           LSD[a]          3.1               1.5                0.8                   1.2

         Average by Date[c]               4.5c              3.1b               1.1a              LSD[d] = 1.2
[a]   Least significant difference. Means within columns with different letters are significantly different at P = 0.05.
[b]   Data was pooled by treatment and analyzed using two‐way analysis of variance.
[c]   Data was pooled by harvest date and analyzed using two‐way analysis of variance.
[d]   Least significant difference for pooled data averaged across harvest date. Means within row with different letters are
      significantly different at P = 0.05.




                           Table 2.8. Storage losses of whole-plant silage (grain
                            plus stover) harvested and stored in 2010 in a farm-
                                              scale silo bag[a].
                                                                    Moisture           Losses
                                                                      % w.b.           % OM
                                                      Control            32             4.1a
                                                  L. buchneri            38             3.1a
                                    L. buchneri + Enzymes                36             1.8a
                                            LSD[b] (P = 0.05)                            4.3
                         [a] Material stored on 7-Oct, 2010 and removed 7-May 2011 after 212
                              days in storage.
                         [b] Least square difference. Means within columns with different letters
                              are significantly different at P = 0.05




In the farm-scale bag silo, thermocouples were used to measure the temperature of the

silage. Figure 2.1 shows that none of the treatments reached 30C. The pretreatment with
                                                                                                          38

L. buchneri and enzymes had a shorter time to maximum temperature during the aerobic

phase, and then had a more consistent temperature.

                       30


                       20


                       10
    Temperature (°C)




                        0


                       -10


                       -20


                       -30
                        9/29/10        11/13/10    12/28/10      2/11/11       3/28/11       5/12/11


                             Control        L. buchneri       L. buchneri + Enzyme         Ambient

Figure 2.1                   Temperature history of whole-plant corn (grain and stover) stored in a farm-
                             scale silo bag.




2.2.4 Post-storage fermentation products
                       In 2009, the control treatment produced primarily lactate, acetate, and ethanol,

with greater quantities at higher moisture levels (table 2.9). In untreated silages, lactic

acid was the primary fermentation product, while in acid treated silages, acetic acid was the

primary fermentation product. This occurs due to the exogenous acid immediately

dropping the pH, reducing the time available for lactic acid producing microorganisms to
                                                                                              39

flourish. Acetic acid can be created by cleaving the ester linkage of the acetyl groups from

acylated xylans, and is likely the primary source of acetic acid production in the silages with

high acid loading, rather than anaerobic fermentation. The pH for the control treatment,

where sufficient moisture existed for fermentation to take place, generally stabilized near 4.

This is similar to results with ensiled corn stover reported in Shinners et al. (2010). Butyrate

was observed in very low levels.

       In 2010, it was found that low levels of acid pretreatment (10 g sulfuric acid (kg

substrate DM)-1) resulted in greater levels of ethanol production and suppressed the

production of lactic and acetic acid compared to the control (tables 2.10 – 2.12). The likely

cause of this was the acid pretreatment effectively minimized bacterial growth, but not

yeast growth. The fermentation products were greater in the stover fraction than the grain

fraction. A likely cause of this is the lower buffering capacity of the grain. As fermentation

products accumulate, the pH is brought down more quickly and further in the grain, ending

further fermentation (tables 2.10 - 2.12)

       The on-harvester application of L. buchneri and L. buchneri plus enzymes increased

production of all fermentation products in the farm-scale silo bag and brought the pH of the

silage down to 4.9, while the control was relatively high at 6.3 (table 2.13). The addition

enzymes to the L. buchneri significantly increased total fermentation products.
                                                                                                                         40


Table 2.9. Post-storage fermentation products for 2009 harvests of whole-plant silage
  (grain plus stover) after 120 days in storage for control with no pretreatment and
              pretreatment with sulfuric acid at a rate of 100g (kg DM)-1.
 Harvest           Moisture            pH          Lactate      Acetate        Ethanol       Butyrate          Total
   Date             (% w.b.)                                                g (kg OM)-1
                                              Control Treatment
 9-Oct.                 65            3.9a          40.4c        27.7c             8.1bc           1.1a          83.6d
27-Oct.                 56            3.9a          33.2b        11.3b             6.6b            1.1a          58.1c
12-Nov.                 40           4.1ab          21.9a        7.7ab             4.3ab           0.6a          41.2b
14-Dec.                 34           4.3b           16.8a         4.9a              2.0a           0.8a          29.2a
  LSD[a]                               0.1           6.6            3.9              3             1.2            7.1
     Average by Date                                28.1            12.9            5.2            0.9           53.0
                                                Acid Treatment
                          [b]
   Average by Date                     1.3           1.4            17.2            0.3              -           21.9

[a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.
[b] Harvest date had no significant effect on the fermentation products produced in the acid treated material.



    Table 2.10. Post-storage fermentation products averaged across three harvest
   dates in 2010 for whole-plant silage (grain plus stover) after 60 days in storage for
       control with no pretreatment and pretreatment with sulfuric acid or lime.
                           pH             Lactate           Acetate        Ethanol        Butyrate        Total
                                                                      g (kg OM)-1
      Control             6.6c               8.1d            2.1b           7.7b           1.8b           21.4c
    Acid – 1%             5.1b               1.7b            1.5a           9.9c            0.0a          13.3b
    Acid – 3%             2.5a               0.0a            5.1c           0.9a            0.0a          6.2a
    Lime – 1%             7.8d               4.6c            2.1b           7.6b            4.2c          20.7c
       LSD[a]              0.3               0.9              0.5            1.4            0.6            2.7
[a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.
                                                                                                                         41


          Table 2.11 Post-storage fermentation products averaged across three harvest
       dates in 2010 for grain fraction hydrodynamically separated from the whole-plant
       silage after 60 days in storage for control with no pretreatment and pretreatment
                                     with sulfuric acid or lime.
                              pH             Lactate          Acetate         Ethanol       Butyrate          Total
                                                                         g (kg OM)-1
         Control              4.8c             2.7d              0.7a           2.5c           0.5b            7.0c
        Acid – 1%            4.1b              0.6b              0.5a           3.0d           0.0a           4.2b
        Acid – 3%            2.2a              0.0a              2.0b           0.2a           0.0a           2.3a
       Lime – 1%             5.8d               1.1c             0.3a           1.7b           0.8b           4.5b
                 [a]
           LSD                0.1               0.3              0.5             0.4            0.1            0.9
[a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.



         Table 2.12. Post-storage fermentation products averaged across three harvest
         dates in 2010 for stover fraction hydrodynamically separated from the whole-
           plant silage after 60 days in storage for control with no pretreatment and
                             pretreatment with sulfuric acid or lime.
                                pH            Lactate         Acetate        Ethanol       Butyrate         Total
                                                                         g (kg OM)-1
           Control            5.6c              7.8d            1.9a           6.0c           1.4b          18.9c
         Acid – 1%            3.9b              1.6b            1.5a           7.3d           0.0a          10.4b
         Acid – 3%            2.4a              0.0a            4.2b           0.6a           0.0a           5.2a
         Lime – 1%            6.8d              4.3c            2.2a           4.5b           4.0c          17.5c
            LSD[a]             0.2               0.8             0.9            1.1            0.3           1.9
[a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.
                                                                                                                         42


    Table 2.13. Post-storage fermentation products whole-plant silage (grain plus stover)
       for 2010 harvest after 60 days in storage for control with no pretreatment and
                         pretreatment with biological amendments.
                           pH             Lactate          Acetate        Ethanol        Butyrate         Total
                                                                      g (kg OM)-1
      Control             6.3b              2.0a             2.0a           1.1a              -            5.4a
    L. Buchneri           4.9a              6.7b             6.4b           2.2a              -           16.2b
    L. Buchneri
                          4.9a              9.5b             13.6c          5.2b              -           30.5c
    + Enzymes
       LSD[a]              0.7               4.7              3.9            2.8                            8.6
[a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.




          The VFAs that were dissociated into solution during hydrodynamic separation were

variable and dependent on duration and agitation in the liquid. These variables were not

strictly controlled, but were a function of the processing time for each silo, all of which were

at least 15 minutes. An analysis of the fermentation products dissociated into the liquid

fraction suggests that 40% of total fermentation products dissociated into the liquid fraction

during the fractionation procedure. The primary fermentation products lactate, acetate,

and ethanol had average dissociations of 28%, 40%, and 50%, respectively. No statistical

analysis was done due to its dependency on processing time, which was not a strictly

controlled part of the experiment.



2.2.5 Post-Storage Composition
          Post storage analysis of the control pretreatments across harvest dates are reported

in tables 2.14 – 2.16. Both the grain and stover fractions retained compositional

characteristics that make them attractive for use as feed or fuel: low ash, high starch
                                                                                                43

content in the grain, and high cellulose and hemicellulose components in the stover. Acid

pretreatment had the effect of increasing WSC and decreasing hemicellulose, likely through

acid hydrolysis of cell wall components (table 2.15). Biological pretreatments used in the

farm-scale silo bag resulted in WSC contents of 2.0%, with no statistically significant

difference in measured compositional characteristics (ash, crude protein, ADF, NDF, ADL,

starch, and WSC) between treatments (data not shown).

         The grain fraction in the control sample resulted in numerically greater starch and

WSC values post-storage as compared to pre-storage, which was shown in table 2.1 to have

a significantly greater starch content than the shell corn (709 g starch (kg shell corn)-1).


Table 2.14. Composition of whole-plant silage (grain plus stover) and separated stover and
 grain fractions after anaerobic storage for silage with no pretreatment averaged across all
                                     harvests in that year.
  Year                    Starch    WSC    Cellulose Hemicellulose        ADL      Ash         CP
                                                      g (kg OM)-1
 2009        Grain         763      16        -[a]           -[a]          -[a]     16         74
             Stover        39       35        419            243           66       53         69
             Whole         433      33        187            109           37       41         86
                                                [a]           [a]
 2010        Grain         731      19         -             -             13       14         64
             Stover            15   26        461            283           74       46         39
             Whole         519      22         -[a]          -[a]          34       30         55
[a]      Data not available.
                                                                                                                           44


    Table 2.15. Composition of whole-plant silage (grain plus stover) for control with no
  pretreatment and pretreatment with sulfuric acid or lime averaged across all harvests in
                                       that year.
 Year        Treatment         Starch       WSC        Cellulose Hemicellulose                   ADL      Sulfur         CP
                                                                        g (kg OM)-1
 2009          Control          433a         33a          187a               109b                37a         1a          86b
                Acid –
                                431a        125b          168a                  37a              32a         31b         80a
                 10%
                LSD[a]           39            5            31                  20                6           2           4
                                                             [b]                 [b]
 2010          Control          519c         22a            -                   -                34a         1a          55ab
                                                             [b]                 [b]
             Acid – 1%         500ab         24a            -                   -                33a         5b          54a
             Acid – 3%          493a         46b            -[b]                -[b]             33a         11c         57c
             Lime – 1%         514bc         21a            -[b]                -[b]             34a         1a          55b
                LSD[a]           17            4                                                  2           1           1
[a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.
[b] Data not available.




   Table 2.16. Composition of separated grain and stover after anaerobic storage in farm-
                     scale silo bag for silage with no pretreatment[b].
                     Starch           WSC          Cellulose Hemicellulose               ADL           Ash          CP
                                                                   g (kg OM)-1
     Grain            743              7             -[a]                -[a]             -[a]         14           68
     Stover           -[a]            15             452                 341              69           40           37
   [a] Data not available.
   [b] No significant difference in compositional characteristics between treatments, so only control values shown.




2.2.6 Post-Storage Starch

          One of the most important questions concerning the viability of the whole-plant

silage system for both starch and cellulosic ethanol conversion is how well the grain starch
                                                                                                45

would be preserved when ensiling at relatively low moistures. The results indicate that

starch and starch derived sugars were well conserved during storage (table 2.17 and 2.18).

       In 2009 the post-ensiled samples were hammermilled to homogenize the grain and

stover fractions. In 2010, the majority of the post-ensiled sample was hydrodynamically

separated and the more homogeneous fractions were sampled. However, pre-ensiled

samples were collected by grab sampling from the mixture. It was difficult to achieve a

uniform sample from this heterogeneous mixture, especially in 2009 when very long TLC

were used. The starch ratio was considerably greater than one in 2009 which suggests that

the grab sampling technique used biased the samples against the grain fraction (table 2.17).

The shorter TLC used in 2010 made it possible to collect more uniform pre-storage samples,

so the starch ratio was closer to one (table 2.18). The LSD was less in 2010 than 2009,

which indicates the variance was less in 2010.

       Although the 2009 results indicated significant starch gains, the significantly lesser

post to pre storage starch ratio for the acid pretreated material indicates that some acid

hydrolysis of the grain starch did occur (table 2.17). In 2010, starch losses averaged across

all moisture contents and treatments were 1%. Acid hydrolysis of the starch was likely one

source of the increased WSC levels in the acid pretreated silages (table 2.16). Further

evidence of starch conservation is that the post-storage corn grain fraction had numerically

greater starch content than pre-storage grain fraction (tables 2.4 and 2.14). When this

evidence is coupled with the low overall dry matter losses, generally 3% or less (table 2.6 –

2.8), it may be inferred that starch losses were less than DM losses from the silage.
                                                                                                                          46



Table 2.17. Ratio of post-storage to pre-storage starch following pretreatment and anaerobic
        storage for sulfuric acid pretreated and control treatments for 2009 harvests.
           Harvest Date        9-Oct        27-Oct.      12-Nov.      14-Dec.     Average by
      Moisture (% w.b.)          65           56            40           34      Treatment[b]
                       Control          1.08a              1.15a             1.41a             1.37a              1.25a
                   Acid – 10%           0.98a              1.09a             1.09a             1.10a               1.06
                                [a]
                          LSD             .41               .14               .33               .55                0.15

      Average by Date[c]                1.03a             1.12ab             1.25b             1.23b          LSD[d] = 0.18

[a]    Least significant difference. Means within columns with different letters are significantly different at P < 0.05.
[b]    Data was pooled by treatment and analyzed using two‐way analysis of variance.
[c]    Data was pooled by harvest date and analyzed using two‐way analysis of variance.
[d]    Least significant difference for pooled data averaged across harvest date. Means within row with different letters are
       significantly different at P = 0.05.



           Table 2.18. Ratio of post-storage to pre-storage starch following pretreatment
              and anaerobic storage for sulfuric acid and lime pretreated and control
                                   treatments for 2010 harvests.
                Harvest Date 28-Sept.           7-Oct.        19-Oct.        Average by
            Moisture (% w.b.)       45            34            16          Treatment[b]
                         Control          1.06a             0.93a              1.17b                  1.06b
                       Acid – 1%          1.12a             0.81b              0.97a                  0.97ab
                       Acid – 3%          0.95a             0.92a              0.97a                   0.95a
                       Lime– 1%           1.01a             0.96a              1.01a                  0.99ab
                            LSD[a]         0.28              0.08               0.11                   0.09

          Average by Date[c]              1.04a              .91b              1.03a             LSD[d] = 0.08
[a]    Least significant difference. Means within columns with different letters are significantly different at P = 0.05.
[b]    Data was pooled by treatment and analyzed using two‐way analysis of variance.
[c]    Data was pooled by harvest date and analyzed using two‐way analysis of variance.
[d]    Least significant difference for pooled data averaged across harvest date. Means within row with different letters are
       significantly different at P = 0.05.



2.2.7 Post-Storage Starch Digestibility
           Ensiling tends to minimizes the effect of vitreousness (horneiness) on starch

digestion. This is done by degradation of the hydrophobic zein protein-starch matrix,
                                                                                                  47

resulting in higher starch digestibility (Philippeau and Michalet-Doreau, 1998) (Jurjanz and

Monteils, 2005) (Hoffman, et al., 2011). This effect was observed for all the ensiled samples

where the higher moisture at earlier harvests increased the fermentation products

produced and lowered pH which then resulted in increased starch digestibility (table 2.19).

The shell corn had higher starch digestibility at lower harvest moisture contents. Overall

shell corn had the highest IVSD result, a surprising result because ensiling is known to

improve starch digestibility (Allen et al., 2003) (Philippeau and Michalet-Doreau, 1998)

(Firkins et al., 2001). Lime pretreatment kept the starch digestibility at a similar, low level

for all harvest maturities. Alkaline steeping has been found to be an effective method of

solubilizing the starch and improving its enzymatic accessibility (Mistry et al., 1992). This

incongruence may have been due to the lime negatively affecting the rumen fluid used in

the digestibility assay.

        Biological inoculation of the silage was shown to improve starch digestibility (table

2.20). The stronger fermentation that occurs with the inoculated silages likely shifts some

of the slowly degradable starch to the rapidly degradable starch fraction (Jurjanz and

Monteils, 2005).
                                                                                                                   48

 Table 2.19. Starch digestibility of the 2010 grain fraction following storage and
   hydrodynamic separation, as measured by percent of starch disappearance
                        during modified 7 h IVSD assay.
        Harvest Date      28-Sept.         7-Oct.       19-Oct.       Average by
    Moisture (% w.b.)           45           34           16         Treatment[b]
                   Control             66.6a              64.8a           51.2b                60.9 b
               Acid – 1%              64.9ab             59.4b            52.5b                58.9 b
               Acid – 3%               69.8a             56.0bc           53.9b                59.9 b
               Lime – 1%               54.8c              51.6c           51.5b                52.6 c
                    Shell             59.7bc             64.9a            67.7a                64.1 a
                      LSD[a]             6.0               4.5             6.3                   2.8
                           [c]
    Average by Date                    64.0c              58.0b           52.3a             LSD[d]=2.4
[a] Least significant difference. Means within columns with different letters are significantly different at P =
    0.05.
[b] Data was pooled by treatment and analyzed using two‐way analysis of variance.
[c] Data was pooled by harvest date and analyzed using two‐way analysis of variance. Includes only ensiled
    treatments.
[d] Least significant difference for pooled data averaged across harvest date. Means within row with
    different letters are significantly different at P = 0.05.




                Table 2.20. Starch digestibility of the 2010 grain fraction
                   following storage in a bag silo and hydrodynamic
                     separation, as measured by percent of starch
                    disappearance during modified 7hr IVSD assay.
                                       Shell          64.9a
                                     Control          59.8a
                                 L. buchneri          66.1a
                         L. buchneri + Enzymes                   75.1b
                                               [a]
                                         LSD                       6.9
               [a] Least significant difference. Means within columns with different letters
                   are significantly different at P = 0.05.
                                                                                               49

2.2.8 Post-Storage Pretreatment

       Chemical pretreatments of the substrate increased the ash and sulfur content by

introducing them to the substrate. While applied on a dry matter basis, the application

appears to be biased toward more application of amendments to the stover fraction, as

seen in the ash and sulfur contents of the fractions (table 2.21). This is likely due to the

large difference in surface area of the two fractions.

       Degradation of hemicellulose in the cell wall by acid hydrolysis in the sulfuric acid

pretreatment was very effective. At 10% acid loading hemicellulose was reduced by greater

than 85% for the last three harvest dates in 2009 (table 2.22). In 2010 the 3% acid loading

yielded hemicellulose degradation of greater than 35% on the last two harvest dates (table

2.23). The 1% acid loading pretreatment produced greater degradation of hemicellulose

than the control only at the lowest crop moisture.

       Sulfuric acid pretreatment seems to have a distinct breakpoint in loading level and

moisture content, where beyond that point, a majority of the hemicellulose in the cell wall

is rapidly degraded with small changes in acid loading or moisture content. This breakpoint

seems to be a function of moisture content, where high harvest moisture content silage

also needs a high sulfuric acid loading (tables 2.22 and 2.23).

       The combination of early maturity and ensiling also had a pretreatment effect.

Higher moisture content, at earlier harvest maturities, resulted in higher levels of

fermentation products (table 2.9). This coincided with hemicellulose degradation, as clearly

seen in the non-pretreated control substrates (tables 2.22 and 2.23). Lime pretreatments
                                                                                                                            50

experienced a similar trend due to levels of fermentation products similar to that of the

control treatments.

          Water soluble carbohydrates increased with acid pretreatment (table 2.15 and 2.23),

as a result of acid hydrolysis of hemicellulose sugars and inhibition of fermentation.

          There was an apparent cellulose gain for all treatments. All of the treatments had

higher cellulose content, as a percent of mass, following storage (table 2.22 and 2.23). This

suggests that acid hydrolysis of the cellulose did not occur, and that it was conserved as

other components of the stover were consumed during storage.


  Table 2.21. Post-storage water soluble carbohydrate (WSC), ash, and sulfur composition
                      of whole-plant silage, stover, and grain fractions

                                 WSC                                   Ash                                 Sulfur

                    Whole                                Whole                                Whole
                                 Grain       Stover                   Grain       Stover                   Grain      Stover
                    -Plant                               -Plant                               -Plant
                                                                  g (kg OM)-1
                                                            2009
  Control            33b          16a          35a         41a         16a          53a          1a          1a         1a
 Acid – 10%          125a         67b         143b        109b         53b         132b         31b         15b        31b
    LSD[a]              5           4           9            4           4          10            2             1       2
                                                            2010
  Control             22a        19b          26a          30a         14a          46a        1.1a         0.9a      0.7a
 Acid – 1%            24a        17ab         31b          39b         16b          51b        5.1b         2.0b      3.8b
 Acid – 3%            46b         30c          60c         53c         27c          67c        11.2c        4.4c       8.5c
 Lime – 1%            21a         15a          23a         36b         15a          53b         1.1a        0.9a       0.6a
          [a]
    LSD                 4           3           4            3           1           4          0.8             0.3    0.4
[a] Least significant difference, different letters in the same columns are statistically different (P=0.05).
                                                                                                                       51

  Table 2.22. Ratio (%) of the post- to pre-storage content of hemicellulose or cellulose
                    in the separated stover fraction for 2009 harvests.
   Harvest Date        9-Oct 27-Oct. 12-Nov. 14-Dec.
                                                              Average[b]      Average[c]
  Moisture (% w.b.)      65       56         40        34
                                        Hemicellulose                         Cellulose
      Control           81 a     86 b      94 b       92 b       88 b            107 a
    Acid – 10%          56 a     14 a       7a         9a        21 a            106 a
           [a]
       LSD               33       12          8        13         7                3
[a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.
[b] Hemicellulose data was pooled by date and analyzed using two‐way analysis of variance.
[c] Cellulose ratio was not affected by harvest date, so data was pooled by date and analyzed using two‐way analysis of
    variance.



  Table 2.23. Ratio (%) of the post- to pre-storage content of hemicellulose or cellulose in
                      the separated stover fraction for 2010 harvests.
    Harvest Date          28-Sept.       7-Oct.    19-Oct.
                                                              Average[b]        Average[c]
  Moisture (% w.b.)           45           34         16
                                          Hemicellulose                         Cellulose
        Control              88a          93b       104c         96b              104a
      Acid – 1%              97b          95b        85b         92b              101b
      Acid – 3%             94ab          64a        48a         69a              105a
      Lime – 1%              87a          94b       101c         94b              101b
             [a]
         LSD                   8           14         12          6                  2
  [a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.
  [b] Hemicellulose data was pooled by date and analyzed using two‐way analysis of variance.
  [c] Cellulose ratio was not affected by harvest date, so data was pooled by date and analyzed using two‐way analysis
      of variance.




2.2.9        Post-Storage Liquid Fraction
         Analysis of the liquid fraction after hydrodynamic separation was done to assess the

components leaving the substrate and going into solution or suspension. Overall, a small

proportion of the whole (1 - 3% of DM) was dissociated into the separation liquid. The

liquid fraction was assayed for WSC, ash, sulfur, and fermentation products (table 2.24).

The constituents analyzed accounted for most of the dry matter in the liquid fraction.
                                                                                                                         52

Further analysis for starch or other potential constituents was not attempted as the

remaining components were diluted to the point that they were near the detectable limits

of the assays. Almost half the fermentation products left the substrate during

hydrodynamic separation, which may be an advantage if some fermentation products prove

inhibitory to downstream processes.


    Table 2.24. Components from substrate that were dissociated into the hydrodynamic
           separation liquid fraction, as a percent of the component of the whole.
                                                                            Fermentation
       Year       Treatment        WSC            Ash          Sulfur
                                                                              Products
      2009         Control          7a            14a           18 a            41a
                 Acid – 10%        14 b            8a           33 b            39a
                               LSD[a]           5                  8                    8                      9
        2010             Control               4a               13 a                  23 a                   30 a
                        Acid – 1%              6a               23 b                  48 b                  46 b
                        Acid – 3%             12 b              22 b                  48 b                  44 ab
                        Lime – 1%             13 b              18 b                  22 a                   49 b
                                     [a]
                               LSD              3                  5                    7                     15
[a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.




2.2.10 Post-Storage SSF
          SSF results are yields of ethanol from S. cerevesiae D5A, a yeast that converts glucose

to ethanol, but does not utilize pentose sugars. Acid pretreatments were effective at

increasing ethanol yields and cellulose conversion by hydrolyzing hemicellulose sugars and

opening the cellulose up for enzymatic degradation (table 2.25). Increased acid loading

improved the ethanol yield. Early harvest maturities had a significant impact on cellulose

conversion efficiency. In 2009, the first harvest done at a whole-plant moisture content of
                                                                                                            53

65% (w.b.) had a cellulose conversion efficiency 18 percentage units higher than the harvest

conducted at 56% moisture content (table 2.26). L. buchneri improved ethanol yields, but

the L. buchneri + enzyme treatment was similar to that of the control.




          Table 2.25. Comparison of pretreatments and no pretreatment prior to
         anaerobic storage and hydrodynamic fractionation on corn stover ethanol
             yield and percent of stover cellulose converted to ethanol by SSF.
            Year                Treatment               Ethanol        Cellulose
                                                      Yield (l(Mg     conversion
                                                              -1
                                                        OM) )        efficiency (%)
                [a]
           2009                   Control                123b             41 b
                                Acid – 10%               179a              60 a
                                                         LSD[c]           7                     2
                  [a]
           2010                          Control                       106c                   32 c
                                        Acid – 1%                      117b                   36 b
                                        Acid – 3%                      143a                   43 a
                                        Lime – 1%                      112bc                  35 b
                                                         LSD[c]           7                     2
                  [b]
           2010                          Control                       96b                    30 b
                                       L. buchneri                     119a                   38 a
                                L. buchneri +Enzymes                    87b                   27 b
                                                               [c]
                                                         LSD             10                     4
       [a] Stored in pilot scale silos. Data averaged across all harvest dates and analyzed using two-way
           analysis of variance.
       [b] Stored in farm-scale silo bag.
       [c] Least significant difference. Means within columns with different letters are significantly
           different at P = 0.05.
                                                                                                                         54


  Table 2.26. Comparison of effects of harvest date in 2009 on percent of stover cellulose
       converted to ethanol by SSF following pretreatment, anaerobic storage, and
                               hydrodynamic fractionation.
       Harvest Date               9-Oct            27-Oct.            12-Nov.            14-Dec.
                                                                                                           Average by
  Moisture (% w.b.)                  65               56                  40                 34
                                                                                                          Treatment[b]
                                                % of theoretical ethanol yield
          Control                  54 b              38 b                37b                36b                 41b
        Acid – 10%                 76 a              58 a               55 a                52 a               60 a
                  [a]
            LSD                      7                 4                  6                   3                  2
  Average by date[c]               65 a              48 b               46 bc               44 c             LSD[d]=3
[a]   Least significant difference. Means within columns with different letters are significantly different at P = 0.05.
[b]   Data was pooled by treatment and analyzed using two‐way analysis of variance.
[c]   Data was pooled by harvest date and analyzed using two‐way analysis of variance.
[d]   Least significant difference for pooled data averaged across harvest date. Means within row with different letters are
      significantly different at P = 0.05.




2.2.11 Moisture Content of Fractions Post separation
          Following the hydrodynamic fractionation procedure, the moisture content of each

fraction was evaluated. Analysis of these post separation fraction’s moisture contents

revealed that following the approximately fifteen minute fractionation procedure, a very

clear trend in moisture contents was found, most acutely in the grain (table 2.27). Stover

and grain that has reached low moisture contents at harvest resisted rehydration. This was

likely because as the cell wall and starch dehydrate, the hydrogen bonding that had

previously been with the water changes to hydrogen bonding among adjacent molecules.

When this occurs there is a shrinking of the capillary space that inhibits rehydration, a

phenomenon that is known as hornification (Laivins and Scallan, 1996) (Lou et al., 2011).
                                                                                                                            55

        Table 2.27. Effects of harvest date and moisture on the moisture content of
        stover and grain fractions immediately following hydrodynamic separation.
                                                      Moisture Content .. % w.b.
                            Whole-Plant at                  Stover After                       Grain After
           Date
                              Harvest                        Separation                        Separation
         28-Sept.                45                             77 a                              46 a
          07-Oct.                    34                          73 b                              39 b
          16-Oct.                    16                          72 c                               35 c
           LSD[a]                                                  1                                 1
   [a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.

                                                            .




2.2.12 Aerobic Stability
        For whole-plant silage stored in a farm-scale silo bag at 34% MC, aerobic stability

was good for the first few days, but after prolonged exposure, began to heat (table 2.28).

The exception to this was the silage treated with L. buchneri and enzyme, which took 84

hours to heat by 2ᵒC. Pretreatments that effectively alter the pH of the silage would be

expected to provide good aerobic stability. Figure 2.2 shows the slower heating that occurs

with the biological pretreatment and the efficacy of the L. buchneri and enzyme

pretreatment at minimizing biological activity, as seen by the low levels of heat production.

Treatment with L. buchneri and enzyme reduced mold growth during aerobic exposure, but

there was no significant difference between treatments on yeast growth except at removal

where the untreated silage had the highest counts (table 2.29).
                                                                                                                             56


                  Table 2.28. Stability during two durations of aerobic exposure as
                 quantified by heating degree days (°C) and dry matter losses using
                 whole-plant silage (stover plus grain) stored in farm-scale silo bag.
                                                  Aerobic Exposure Duration
                                                 2 days                   7 days
                                  Control                       12                              132
                             L. buchneri                        2                               108
               L. buchneri + enzyme                             0                                16
                                                                     Losses .. % of DM
                                  Control                    1.8a                              -1.7a
                             L. buchneri                     1.0a                              -0.4a
               L. buchneri + enzyme                          3.8a                              1.0a
                        LSD[b]                                3.6                               3.9
[a] Heating degree days are based on the difference of the silage average temperature (mean of maximum and minimum
temperature) from the ambient average temperature. (Williams, 2011)
[b] Least significant difference. No significant differences found at P = 0.05.




      Table 2.29. Mold and yeast counts following anaerobic storage and subsequent
    aerobic exposure of two durations for whole-plant silage (stover plus grain) stored in
                                     farm-scale silo bag.
                        Mold .. log CFU(g)-1                    Yeast .. log CFU(g)-1
                     At
                              2 days     7 days        At Removal 2 days            7 days
                  Removal
     Control        5.1a        5.8b       6.6b            8.2b          8.1a         8.0a
    L. buchneri            4.8a           3.0a           7.1b                 6.7a             7.9a            7.8a
   L. buchneri +
                           4.3a           3.0a           3.0a                 5.5a             8.2a            7.6a
      enzyme
       LSD[a]               3.0            0.2            1.7                  1.4              1.8             1.6
    [a] Least significant difference. Means within columns with different letters are significantly different at P = 0.05.
                                                                                                        57

                      45

                      40

                      35

                      30
  Temperature .. °C




                      25

                      20

                      15

                      10

                       5

                       0
                        5/11         5/12   5/13      5/14    5/15      5/16     5/17      5/18      5/19
                           Control          L. buchneri       L. buchneri + Enzyme           Ambient


 Figure 2.2 Temperature of whole plant silages with biological pretreatments during
            aerobic exposure following anaerobic storage and fermentation in a farm-
            scale silo bag.




2.3                   Discussion and Conclusions


2.3.1 Composition
                      The composition of the silage following storage has many characteristics desired for

biochemical conversion, specifically low ash, high cellulose and hemicellulose contents.

(tables 2.14 - 2.16). However, fermentation products are also present which may be

inhibitory to some of the proposed biochemical processes. This may be partially overcome

by their observed volatility and the propensity to dissociate in water (table 2.24). During a
                                                                                             58

prolonged hydration process, it is expected most of the fermentation products would

dissociate into the water. From this water they could be recovered for use.

       The moisture content of the whole-plant silage has a twofold advantage over dry

feedstock. Drying the feedstock causes hornification of the biomass, resulting in a

feedstock that is more resistant to enzymatic degradation. Secondly, moist feedstock brings

along some of the water required for some biochemical conversions, resulting in a lower

water requirement for the biorefinery. Although industrial drying of feedstocks has been

proposed (Hess et al., 2009), this causes a much greater energy input to the system that

moist feedstock systems are not saddled with.




2.3.2 Losses and Starch Losses
       The whole-plant silage of grain plus stover was well conserved during storage. The

results of this work suggest that if this system was employed and harvest took place over 10

weeks, with whole-plant harvest moisture contents ranging from 60-15% and no

pretreatments used to further conserve the crop, losses greater than 5% would be

uncommon as long as anaerobic conditions are maintained (tables 2.6 – 2.8).

        Grain starch losses are an important concern with the proposed harvest system

because the grain carries the majority of the value. Also of concern is the grain’s ability to

be converted to energy and the possible requirement of significant changes to an ethanol

process that could be required to utilize corn grain originating from the silage system. Any

impairment to the value of grain harvested, by storage losses or a less advantageous

ethanol process, would have to be assigned to the stover. Due to its relatively low value
                                                                                               59

compared to grain, only modest grain value impairment can be experienced in the system.

Conversely, any net benefit this system provides to the grain would have a marked effect on

further cost reductions to the stover fraction.

       To this end, starch losses were assessed following anaerobic storage. Given the low

starch losses or apparent starch gains (tables 2.17 and 2.18), higher starch concentration in

the grain (tables 2.4 and 2.14), and low overall silage losses (tables 2.6 – 2.8), these results

gives strong evidence the starch fraction was very well conserved.




2.3.3 Pretreatment
      The effects of pretreatment on the silage were very pronounced for the 30 g(kg DM)-1

and 100 g(kg DM) -1 sulfuric acid pretreatments, particularly effective at lower moisture. At

the high sulfuric acid loading, less than 10% of the hemicellulose remained bound in the cell

wall at less than 50% moisture (table 2.22). And the cellulose conversion to ethanol was

improved dramatically, 60% as compared to the control at 41%. The pretreatment by lime

and sulfuric acid at 10g(kg DM)-1 produced only a minor increase in cellulose conversion to

42% and 43%, respectively, from the control conversion efficiency of 38% (table 2.23).

       All acid pretreatments were effective at inhibiting anaerobic fermentation during

storage as seen by the low levels of fermentation products produced (tables 2.9 and 2.10).

The lime pretreatment likely had an insufficient loading level to inhibit anaerobic

fermentation.

       The pretreatment was also found to affect the grain fraction. The sulfuric acid

would be expected to promote starch availability by a step similar to the steeping of corn in
                                                                                               60

lactic acid and sulfuric acid prior to some milling processes; however, this was not found in

this starch digestibility assay.

        Sulfuric acid pretreatment increased the sulfur content of the grain (table 2.21),

although at a rate lower than the pretreatment. This excess sulfur in the grain fraction may

require additional amendments (e.g. copper) to offset the sulfur causing a copper deficiency

if grain or byproduct is fed to animals.




2.3.4 Fractionation
        The hydrodynamic separation method was chosen over a mechanical system in an

effort to replicate what was felt was the most likely industrial scale fractionation technique

to be employed. The system seems to be the most logical choice as it results in a very high

degree of starch and fiber fractionation. In fact, it was found to be more effective than a

combine in the traditional dry grain system at separating the starch and fiber (72.2% starch

of the hydrodynamic separated grain as compared to 70.9% for the shelled corn – table

2.4). The material needs to be rehydrated; therefore the water separation is plausible. If

this system is employed on-farm for animal feed production, mechanical separation

techniques are likely to be used. This area needs further evaluation as to efficacy across a

range of moisture contents.
                                                                                              61

2.3.5 Scaling to Production System
       In an effort to evaluate if this system has the potential to scale up and be done

commercially, a farm-scale bag silo whole-plant corn pretreated with biological

amendments was evaluated. The results were similar to those seen in the pilot-scale silos,

with the exception of the pretreatment differences resulting from biological rather than

chemical pretreatments. It was found that storage losses at the farm-scale can also be

expected to remain low and result in a feedstock that is a dense, flowable, size-reduced,

consistent product (table 2.8).




2.3.6 Overview
       The desire of biorefineries to have a dense, size-reduced, flowable, consistent

product at a low price, that has a very low risk of a failure to harvest, makes this system

appear to be a promising feedstock system for biorefineries that have a desire to convert

both grain and stover to ethanol or other products in a biochemical platform.

       Prior to this research, this system had not yet been fully investigated, but this work

strongly suggests that the silage system needs to be seriously considered. Cultural changes

will be a significant concern to growers, as it will necessitate changes in the way they

harvest, haul, and market their grain. Grain producers will see new risks in storage, as

ensiling requires different management approaches than dry grain in bins. But at the same

time they will be able to eliminate grain drying and increase the value of their crops.

       This research did not focus on the use of this system in animal agriculture, but there

may be benefits to its utilization on-farm. This system can result in two product streams
                                                                                              62

out of storage: ensiled corn grain and stover. The grain can be used in feed after

processing or potentially sold to an ethanol facility accepting moist corn. Stover can have a

value as a both a feed amendment and as bedding. Recently wheat straw in Wisconsin has

been selling for over $100(ton)-1 for use on dairy farms as bedding. This suggests that the

value of the silage as a bedding and feedstuff could add value to a producer’s operation

over the traditional grain only harvest systems.
                                                                                                   63

References

Allen, M. S., Coors, J. G., and Roth, G. W. (2003). Corn silage. In: silage science and
technology, D. Buxton, R. Muck, and J. Harrison, eds. American Society of Agronomy, Crop
Science Society of America, and Soil Science of America, Madison, WI.

ASABE. (2007). Standard S424.1: Method of determining and expressing particle size of
chopped forage materials by screening. ASABE.

ASABE. (2008). Standard S358.2: Moisture measurement - forages. ASABE.

Blacno-Canqui, H., and Lal, R. (2009). Crop residue removal impacts on soil productivity and
environmental quality. Critical Reviews in Plant Sciences, 28(3):139-163.

Cook, D., and Shinners, K. (2011). Economics of alternative corn stover logistics systems,
Paper number: 1111130. 2011 ASABE Annual International Meeting, (p. 38). Lousiville, KY.

Digman, M. F., Shinners, K. J., Dien, B. S., Hatfield, R. D., Li, X., Muck, R. E., et al. (2007). On-
farm pretreatment technologies for improving enzymatic degredability of cellulose and
hemicellulose present in perenial grass. ASABE Annual International Meeting, Paper
Number: 071021.

Digman, M., Shinners, K., Dien, B., Hatfield, R., Muck, R., and Weimer, P. (2008).
Investigation of a two-stage process for increasing yields of ethanol from switchgrass. 2008
ASABE Annual International Meeting, no. 084954.

Digman, M. F. (2009). Simultaneous preservation and pretreatment of perennial grasses for
fuel ethanol production. PhD Thesis, University of Wisconsin - Madison.

Digman, M. F., Shinners, K. L., Muck, R. E., and Dien, B. S. (2010). Pilot-scale on-farm
pretreatment of perennial grasses with dilute acid and alkali for fuel ethanol production.
Transactions of the ASABE, 53(3): 1007-1014.

Dowe, N., McMillan, J., 2001. SSF experimental protocols: lignocellulosic biomass
hydrolysis and fermentation. National Renewable Energy Laboratory, Golden,
CO

Dubois, M., Gilles, K. A., and Hamilton, P. A. (1956). Colormetric method for determination
of sugars and related substances. Anal. Chem., 28(3): 350-356.
                                                                                                    64

Esteghlalian, A. R., Bilogesu, M., Mansfield, S. D., and Saddler, J. N. (2001). Do enzymatic
hydrolyzability and simons’ stain reflect the changes in the accessibility of lignocellulosic
substrates to cellulase enzymes? Biotechnology Progress, 17(6): 1049-1054.

Firkins, J. L., Eastridge, M. L., St-Pierre, N. R., and Noftsger, S. M. (2001). Effects of grain
variability and processing on starch utilization by lactation dairy cattle. J. Animal Science,
79(E. Supl.), E218-E238.

Goering, H. K., and Van Soest, P. J. (1970). Forage fiber analysis (Vol. Agriculture Handbook
379). ARS, USDA.

Hess, J. R., Wright, C. T., Kenney, K. L., and Searcy, E. M. (2009). Uniform-format solid
feedstock supply system. Idaho Falls, Idaho: Idaho National Laboratory.

Hoffman, P. C., Esser, N. M., Shaver, R. D., Coblentz, W. K., Scott, M. P., Bodnar, A. L., et al.
(2011). Influence of ensiling time and inoculation on alteration of the starch-protein matrix
in high-moisture corn. J. Dairy Sci., 94:2465-2474.

Holmes, B. (2008). Density and porosity in bunker and pile storage. Retrieved 8 5, 2011,
from http://www.uwex.edu/ces/crops/uwforage/Density-Porosity3.pdf

Huang, H.-J., Ramaswamy, S., Al-Dajani, W., Tschirner, U., and Cairncross, R. (2009). Effect of
biomass species and plant size on cellulosic ethanol: A comparative process and economic
analysis. Biomass and Bioenergy, 33(2):234-246.

Jurjanz, S., and Monteils, V. (2005). Ruminal degradability of corn forages depending on the
processing method employed. Animal Research, 54:3-15.

Kazi, F. K., Fortman, J. A., Anex, R. P., Hsu, D. D., Aden, A., Dutta, A., et al. (2010). Techno-
economic comparison of process technologies for biochemical ethanol production from
corn stover. Fuel, 89:20-28.

Khanna, M. (2008). Cellulosic Biofuels: Are they economically viable and environementally
sustainable? Choices, 23(3): 16-21.

Kou, N., and Zhao, F. (2011). Effect of multiple-feedstock strategy on the economic and
environmental performance of thermochemical ethanol production under extreme weather
conditions. Biomass and Bioenergy, 35(1): 608-616.

Laivins, G. V., and Scallan, A. M. (1996). The influence of drying and beating on the swelling
of fines. J. Pulp Paper Science, 22(5): 178-183.
                                                                                                 65

Lauer, J., Kohn, K., and Diallo, T. (2009) 2009 Wisconsin corn hydbrid performance trials
grain and silage. UW-Extension Publication A3653.

Lauer, J., Kohn, K., and Diallo, T. (2010) 2010 Wisconsin corn hydbrid performance trials.
UW-Extension Publication A3653.

Lou, X. L., Zhu, J. Y., Gleisner, R., and Zhan, H. Y. (2011). Effects of wet-pressing induced fiber
hornification on enzymatic saccharification of lignocelluloses. Cellulose, 18:1055-1062.

McAloon, A., Taylor, F., Yee, W., Ibsen, K., and Wooley, R. (2000). Determining the cost of
producing ethanol from corn starch and lignocellulosic feedstocks. U. S. Department of
Agriculture and U. S. Department of Energy. Golden, CO: NREL / TP-580-28893.

Mistry, A.H., Schmidt, S. J., Eckhoff, S. R., and Sutherland, J.W. (1992) Alkali extraction of
starch from corn flour. Starch, 44:284-288.

Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y., Holtzapple, M., et al. (2005). Features of
promising technologies for pretreatment of lignocellulosic biomass. Bioresource
Technology, 96:673-686.

Muck, R. E. (1988). Factors influencing silage quality and their implications for management.
Journal of Dairy Science, 71:2992-3002.

Muck, R. E., and Holmes, B. J. (2006). Bag silo densities and losses. Transactions of the
ASABE, 49(5):1277-1284.

Petrolia, D. R. (2008). The economics of harvesting and transporting corn stover for
conversion to fuel ethanol. Biomass and Bioenergy, 32(7):603-612.

Philippeau, C., and Michalet-Doreau, B. (1998). Influence of genotype and ensiling of corn
grain on the in situ degradation of starch in the rumen. J. Dairy Science, 81:2178-2184.

Renewable Fuels Association. (2011). Ethanol Statistics. Retrieved June 22, 2001, from
http://www.ethanolrfa.org/pages/statistics

Roth, G., Undersander, D., Allen, M., Ford, S., Harrison, J., Hunt, C., et al. (1995). Corn silage
production, management, and feeding. Madison, WI: American Society of Agronomy.

Ruppel, K. A., Pitt, R. E., Chase, L. E., and Galton, D. M. (1995). Bunker silo management and
its relationship to forage preservation on dairy farms. Journal of Dairy Science, 78(1):141-
153.
                                                                                                 66

Savoie, P., Shinners, K. J., and Binversie, B. N. (2004). Hydrodynamic separation of grain and
stover components in corn silage. Applied Biochemistry and Biotechnology, 113:41-54.

Shinners, K.J., A.G. Jirovec, R.D. Shaver and M. Bal. (2000). Processing whole-plant corn
silage with crop processing rolls on a pull-type forage harvester. Applied Engineering in
Agriculture. 16(4):323-331.

Shinners, K. J., Binversie, B. N., Muck, R. E., and Weimer, P. J. (2007). Comparison of wet and
dry corn stover harvest and storage. Biomass and Bioenergy, 31(4):211-221.

Shinners, K. J., Wepner, A. D., Muck, R. E., and Weimer, P. J. (2010). Aerobic and anaerobic
storage of single-pass, chopped corn stover. Bioenergy Research, 4(1):61-75.

USDA. (2010). 2008 Energy Balance for the Corn-Ethanol Industry. Retrieved July 2011,
from: http://www.usda.gov/oce/reports/energy/2008Ethanol_June_final.pdf.

USDA. (June 9, 2011). World Agricultural Supply and Demand Estimates. Retrieved June
2011, from: http://usda01.library.cornell.edu/usda/waob/wasde//2010s/2011/wasde-06-09-
2011.pdf

USDA-NASS. (2011). Agricultural statistics database; National crop statistics. United States
Department of Agriculture National Agricultural Statistics Service. Retrieved July 2011, from
http://www.nass.usda.gov/index.asp

USDOE. (2006). Theoretical ethanol yield calculator. Retrieved 8 5, 2011, from
http://www1.eere.energy.gov/biomass/ethanol_yield_calculator.html

US DOE and USDA, A joint study sponsored by. (August 2011). U.S. Billion-Ton Update:
Biomass Supply for a Bioenergy and bioproducts industry. Available from:
http://www1.eere.energy.gov/biomass/pdfs/billion_ton_update.pdf: DOE/EE-0363.

Wiersma, D. W., Carter, P. R., Albrecht, K. A., and Coors, J. G. (1993). Kernel milkline stage
and corn forage yield, quality, and dry matter content. J Prod. Agric., 6:94-99.

Wilhem, W. W., Johnson, J., Lightle, D., Barbour, N. W., Karlen, D. L., Laird, D. A., et al.
(2011). Vertical distribution of corn stover dry mass grown at several U.S. locations.
Bioenergy Research, 4(1):11-21.

Williams, S. D. (2011). Anaerobic storage and aerobic stability of moist biomass feedstocks.
MS Thesis, University of Wisconsin - Madison.
                                                                                                67

Williams, S. D., and Shinners, K. J. (2011). Anaerobic storage and aerobic stability of moist
biomass feedstocks. ASABE Annual International Meeting, Paper Number:1111082.
Louisville, KY.

Yang, B., and Wyman, C. E. (2008). Pretreatment: the key to unlocking low-cost cellulosic
ethanol. Biofuels, Bioproducts, Biorefining, 2(1):26-40.
                                                                                               68

Chapter 3 – Economic Assessment of Corn Stover Logistics Systems
3.1    Introduction
       Corn stover is widely recognized as the most promising high volume lignocellulosic

feedstock to produce a sustainable, renewable energy supply. However, the costs

associated with harvest, storage, and transport of this relatively low-value commodity

challenges its economic viability for both producers and end-users. In 2009, US corn

producers harvested approximately 298 Tg of corn grain dry matter, of which 18 Tg was

harvested as silage (USDA-NASS, 2010). Using a stover to grain ratio of 1:1 (Shinners et al.,

2007a), the US has a corn stover production potential of 280 Tg. If half of this corn stover

were utilized that year, based on an estimated yield of 300 L(Mg stover)-1 (McAloon et al.,

2000), it could have produced 42 GL of ethanol. On an energy basis, this equates to 8% of

the US motor gasoline consumption (DOE – EIA, 2011).

       Stover harvest and storage schemes can be grouped broadly into bale or bulk

systems. Bale systems use large round or square balers to package stover harvested

directly from the combine (single-pass systems) or baled from windrows (two-, three- or

four-pass systems). Some aspects of the economics of baled systems have been presented

by Shinners et al. (2003); Sokhansanj et al. (2006); Brechbill and Tyner (2008); Wright et al.

(2006); Cundiff and Grisso (2008); and Petrolia (2008). They have suggested the baling and

storage be done on the farm, followed by year round transport to the processor, by varying

means, including bales, ground stover, and containerized systems. The costs were

estimated between $40 (Mg)-1 and $93 (Mg)-1 for transport distances between 32 and 163

km.
                                                                                             69

       Bulk systems involve size-reducing and collecting the stover at the grain combine

using additional machine components at the rear of the combine (single-pass systems) or

picking-up windrowed stover using a forage harvester with windrow pick-up (two-, three-,

or four-pass systems). The size reduced stover would then be anaerobically stored and

preserved. Shinners et al. (2003) estimated that bulk stover systems reduced total costs of

stover logistics by 20 to 31% compared to baled systems practiced today.

       The manner in which stover is harvested is a key component to the overall stover

logistics system. It influences the entire storage and logistics system around the harvested

biomass physical form and moisture content. If stover moisture must be less than 25%,

then the time available for harvest may be shortened due to weather, dramatically

increasing the cost, labor (mass per harvest time laborer), and capital requirements of

biomass harvesting. To limit the negative consequences to soil quality, stover yields should

be limited to 10-65%, depending on field characteristics. However, the consequences to

limiting stover yield are increased transport distances, more land suffering soil compaction,

increased use of fossil fuels, more expensive harvest operations, and increased soil

contamination of the crop.

       The objective of this research was to evaluate the effects of corn stover yield,

moisture content and bulk density on the overall system economics, as quantified by costs,

labor and capital requirements, and energy use. Specifically, this research evaluated several

corn stover harvest, storage, and logistics schemes for the sensitivity to primary harvest
                                                                                             70

parameters, cost and energy balance comparisons, the system’s labor and capital intensity,

and estimates for corn stover costs on different supply scales.




3.2    Model Background and Assumptions
       A model was developed for corn stover logistics systems for harvest through

preparation for use in a biorefinery. The model was built as a set of modules to evaluate

standing crop value, harvest, storage, transport, processing, and the supply availability. The

modules can be operated either individually or as a network to evaluate the entire system.

Overall system assumptions used for all systems and variables are given in table 3.1. All

monetary values are set in 2011 US dollars. The outputs of the model are the additional

cost, labor, energy, and capital above and beyond the traditional corn for grain system.




3.2.1 Material Description, Nutrient Removal, and Soil Quality
       The impact of corn stover harvest on soil quality is a concern with respect to system

sustainability. However, the value of stover as a soil amendment is not agreed upon in the

literature. Soil type, topography, residue removal ratio, nitrogen fertilization rate, and

tillage are major factors that affect both soil erosion and soil organic carbon (SOC).

Unharvestable biomass, made up of structural root biomass and rhizodeposition, makes up

a greater portion of the SOC pool than does the aboveground portion (Allmaras et al.,

204). In the model, the potential SOC contribution from corn stover was assumed to be

13% of stover OM and its value was estimated at $7(Mg)-1 (Lang, 2002). It was also
                                                                                             71

assumed that $15(Mg DM)-1 was sufficient producer payment to cover the cost of soil

erosion in most circumstances.

       The nutrient replacement ratio of stover was reported as 0.0; 3.1; and 16.5 kg(Mg

DM)-1 for N, P, and K, respectively (Petrolia, 2008). However, Sheehan et al. (2004)

suggested the replacement ratio to be 8.8; 0.6; and 7.2 kg(Mg DM)-1 for N, P, and K,

respectively. It has also been suggested that the following year’s corn crop needs less

nitrogen fertilization when stover is removed, likely due to reduced immobilization of

nitrogen caused by the high C:N ratio of corn stover (Coulter, 2008). In the model, it was

assumed no replacement requirement for N (Table 3.2).
                                                                                   72

 Table 3.1. General assumptions and variable ranges for corn stover harvest.
       Assumptions or variables                     Values         Units
                             Corn grain yield        8.54          kg DM(ha)-1
                   Stover : grain mass ratio          1:1
    Localized land area available for stover          27           %
                                    harvest
             Average in-field haul distance           0.3          km
Average field-edge to storage haul distance           3.2          km
  Median payment above soil amendment                 15           $(Mg DM)-1
                        value of stover
                              Fuel cost              1.06          $(L)-1
                                Labor Rate [1]        12           $(hr)-1
                          Grain Drying cost          0.049         $-point(bu)-1
               Grain Drying, MC reduction              6           points/bu
                                                                   $(bu-6
                               Grain Storage         0.138         months)-1

                             Cash Corn Price           5           $(bu)-1
               Profit for harvest processes           20           %
             Profit for transport processes           15           %
               Profit for storage processes           10           %
              Profit over storage loss costs          35           %
            Inflation adjusted interest rate           3           %
              Round bales
                                         Size     1.57 x 1.83      m
                                                             [2]
                                      Density     110, 152         kg DM(m)-3
              Square bales
                                         Size    0.89 x 1.19 x     M
                                                      2.59
                                      Density    140, 152 [2]      kg DM(m)-3
           Forage transporter
                                         Size     2.5 x 2.9 x      M
                                                     15.8
         Transport density – pre-storage[3]       70, 80, [5]      kg DM(m)-3
                                                                                                           73

                                – post-storage[4]              160, 140, [5]
[1]   On top of the labor rate is 20% benefits and 25% employer costs (7.65% FICA taxes, 4%
      unemployment, 4% worker’s compensation, 9.35% overhead and administration) - $18(hr)-1 total
      base labor cost. A premium of 20% and 10 % are applied to harvester and semi operators,
      respectively.
[2]   Density for bales of stover and material other than grain (MOG), respectively.
[3]   Density for bulk stover, MOG and whole-plant corn, respectively, during transport from field to
      storage.
[4]   Density for bulk stover, MOG and whole-plant corn, respectively, during transport from storage
      to biorefinery.
[5]   Whole-plant corn silage density was calculated as the sum of the partial volumes of its respective
      grain and bulk stover fractions.
                                                                                                                 74




             Table 3.2. Removal ratio, price, and specific cost for replacement of key
                                       soil macronutrients.
                                           Removal          Price[2]        Specific
                                                   Ratio[1]                                 Cost[1]
                                               kg(Mg DM)-1             $(kg)-1           $(Mg DM)-1
                         N                             -                 1.26                   -
                       P2O5                          2.35                1.96                 4.6
                        K2O                           14                 1.16                 16.2
                                                                              [3]
              Soil organic matter                    130               0.007                  0.9
         Total economic value                                                                 21.8
       [1]    Unit mass or cost of nutrient removed per mass of stover DM removed.
       [2]    Price per unit mass nutrient from urea (0-0-46) $525/ton, DAP (12-40-0) $765/ton, and Potash (0-
              0-62) $653/ton. (Prices quoted by Agriland Co-op, Fond du Lac, WI on May 11, 2011)
       [3]    (Iowa State Extension, 2010)




3.2.2 Harvest Methods
       Harvest methods were defined by the number of field operations, the physical form

of the stover as it exits the field, and the quantity of stover harvested. Single-pass harvest

refers to the combine doing all the stover harvesting in conjunction with the grain harvest

(Shinners et al., 2007b; 2009). The type of head used on the combine dictates the fraction

and quantity of stover harvested. Two-pass harvest refers to the combine harvesting grain,

and simultaneously windrowing the stover that will be harvested in a second field operation

(Shinners et al., 2011a). Three-pass harvest refers to the combine harvesting the grain,

followed by a windrowing stalk chopper being used in the second field operation, and finally

a third operation of baling. Two- and three-pass harvest systems utilized either a large

round baler (LRB) or large square baler (LSB) to harvest and package the stover; or a self-
                                                                                              75

propelled forage harvester (SPFH) to harvest size-reduced stover in bulk form. Several of

the most promising and/or the most often considered scenarios were modeled (table 3.3).

       Two, single-pass scenarios were modeled where the stover was chopped at the

combine and handled thereafter as a bulk material. The combine was configured with

either a whole-plant or ear-snapper head where 55% or 15% of the available stover was

collected (scenarios 1 and 2, respectively). Another single-pass scenario was modeled

where a LRB was towed by the combine that had an ear-snapper head to capture 17% of

the available stover (scenario 3). The LRB was configured so that material accumulated

during the bale wrapping and unloading process so that the combine was not required to

stop. Two, two-pass scenarios were modeled, harvesting the windrows formed at grain

harvest either by chopping with a SPFH and handling in bulk form or by baling with a LSB

(scenarios 4 and 5, respectively). Conventional stover harvest practices were also modeled:

one using a LSB and storing bales indoors, the second using a LRB and storing bales

outdoors (scenarios 6 and 7, respectively).

       An additional harvest scenario, not previously considered, was also modeled –

single-pass harvest of the whole-plant. Corn grain and stover at black-layer or later

maturity would be harvested simultaneously using a SPFH, and the combined fractions

transported and placed into anaerobic storage (Cook and Shinners, 2011). Amendments

may be used to help preserve the feedstock and begin pretreatment of the stover fraction;

however, this was not modeled. After storage, either the grain or stover fractions could

also be utilized as animal feed, but both fractions were targeted as a biomass feedstock in
                                                                                            76

this model. Therefore the combined grain and stover would be shipped to the processor,

where the grain and stover fractions will be separated and both fractions further processed.

        Harvest equipment specifications and assumptions are given in table 3.4.

Parameters such as grain loss, fraction of total stover harvested, and mass-flow-rates were

determined from previous work on stover harvest (Shinners et al., 2007b; 2009; 2011a).

The equipment productivity was modeled by the lesser of the results from equations 1 and

2, with the maximum speed and mass-flow-rate plus assumed field efficiency provided in

table 3.4:



                                                    [1]


                                          (   )     [2]


       where:                 Pa1    harvester area productivity limited by ground speeds
                              Pa2    harvester area productivity limited by mass flow
                               w     width of harvested strip
                                v    maximum harvester speed
                               ηf    field efficiency
                               M     maximum mass-flow-rate
                              Ym     dry mass yield per unit area




       The length of the typical annual stover harvest was estimated at 200 h (5 week

season x 40 machine-hours per week) for any harvest scenario where storage by ensiling

occurred. Dry bale harvest was estimated as 75% of this duration because of the limitation

to produce low crop moisture by field drying. The time for stover to dry to baling moisture
                                                                                               77

has been reported to take at least several days because of low ambient temperatures and

frequent precipitation (Glassner et al., 1998; Shinners et al., 2007a). Labor was calculated

as 25% more time than machine hours to account for machine downtime and non-

productive time related to harvest.

       All machinery had lubrication costs estimated as 15% of the total fuel cost, machine

housing costs were estimated to be $32(yr-m2)-1, and annual insurance cost was estimated

to be 0.5% of the machine purchase price. Purchase price was estimated as a 15% discount

to retail list price. The cost of net wrap for LRBs was estimated as $1.50(bale)-1 and twine

for LSBs was estimated as $0.50(bale)-1. Amortization and depreciation of all equipment

utilized the methods described in ASABE Standards EP496.2 and D497.4 (ASABE, 2006a, b)

and used 0% inflation and an estimated real interest rate of 3% to return real dollar values.

Repairs were estimated for the lifetime of the machine (table 3.4) as a fraction of the

purchase price.




3.2.3 Handling, Storage and Transport
       We assumed all of the baled systems employed a self-propelled bale transporter

which collected the bales and placed them in a stack at the field edge. The staged bales

would later be moved with a truck-trailer combination with a capacity of 27 large round

bales or 36 large square bales (0.9 x 1.2 x 2.6 m). Bales were assumed to be handled at the

field or storage site with a telehandler, capable of handling multiple bales at a time with

weight capacity of 44 kN. The average in-field haul distance was assumed to be 0.3 km, and

the distance from field side to storage was 3.2 km. For final transport from storage to the
                                                                                             78

processor, bales were handled with the same telehandler at both loading and unloading.

Moist, bulk material was harvested with a SPFH and blown directly into a self-unloading

semi-trailer with a capacity of 114 m3. The transport distance from field to storage was 3.2

km.

       The whole-plant silage system yielded a negative harvest cost for the stover fraction

because the harvest costs for the whole-plant were similar to the harvest costs of

harvesting the grain fraction alone with the combine harvester. When coupled with slightly

decreased grain losses with the SPFH compared to the combine, the whole-plant harvest

with the SPFH was slightly less costly than grain harvest alone with a combine, yielding a

negative stover harvest cost for the whole-plant system.

       Costs for truck and trailer transport of bales or bulk stover was assumed to be

$0.79(km)-1 plus labor, fuel and lubrication, and trailer costs. Transport labor rate is

estimated as the system labor cost (table 3.1) plus 10% to account for annual non-

productive hours. Fuel use was estimated by the vehicle gross weight on a linear regression

of 3.4 km(L)-1 unloaded and 2.3 km(L)-1 at maximum regulated weight of 356 kN (80,000

lbs). Lubrication was assumed to be 15% of fuel cost. Transport speeds for trucks were

assumed to average 70 km(h)-1. All other transport vehicles and in-field transport speed

was estimated by traction modeling, where the in-field rolling resistance coefficient was

0.12 and the road rolling resistance coefficient was 0.03 (Goering et al., 2003).

       Bales that are less than 30% moisture content (w.b.) were assumed to be stored

under tarp on a well-drained base of stone, while bales over 30% were assumed to be line
                                                                                           79

wrapped with plastic film and preserved by anaerobic fermentation. Storing biomass

uncovered increases dry matter loss, moisture content and variability of moisture content,

the risk of self-combustion, and material inconsistency, making it less desirable to

processors (Shinners et al., 2007a; 2010). However, because storing uncovered round bales

of stover outdoors is common practice, one scenario in which dry LRB would be stored

outdoors in direct contact with the ground and without cover was modeled.

       Moist, bulk stover was assumed to be stored in 2.87 m diameter plastic silo bags.

Shinners et al. (2011b) reported that stover density in silo bags averaged 265 kg WM(m)-3.

Typical road transport weight and volume regulations limit cargo density to about 240

kg(m)-3. Therefore, it was assumed that when the moist, bulk stover was removed from

storage, its in-storage density would be preserved to insure transport efficiency. It is

envisioned that sections of bags would be loaded onto transport trucks in a similar fashion

that cotton modules are loaded and transported. Whole-plant stover and grain was

assumed to be stored in bunk silos and was removed from storage with large wheel loaders.

The density of whole-plant material with stover plus grain was greater than the legal limits

so no efforts or costs were expended to maintain the in-silo density. Improved storage sites

for bales or bag silos were assumed to be located an average of 3.2 km from any field on

the farm. Any site would have full access throughout the year.

       The major factors that affect loss of DM during aerobic storage are initial moisture,

ambient temperature, exposures to precipitation, precipitation amount, type of bale wrap

(twine or net mesh), drainage, and length of storage period. Most of these factors also
                                                                                           80

affect the aggregate bale moisture at removal from storage. Stover DM losses in net

wrapped bales stored outdoors directly on the ground ranged from 10.7 to 14.2% (Shinners

et al., 2007a) and 10% to 23% (Richey et al., 1982). Storage losses of 12% of DM were

assumed.

       Net wrapped stover bales stored outdoors gained 8.1 to 36.9 percentage units of

moisture during storage (Shinners et al., 2007a) and had highly variable moisture on a daily

basis and within the bales themselves. A final moisture content of 30% for bales stored

outdoors and 20% for bales covered during storage was assumed, based on the observation

from literature that aerobically stored bales tended to move toward those values. Stover

stored in a bag-type silo and preserved by anaerobic fermentation averaged DM losses of

3.3% (Shinners et al., 2011a) so for the purposes of the model it was assumed storage losses

of 3-5% for the anaerobic storage schemes considered (table 3.5).
                                                                                          81

                           Table 3.3. Stover harvest scenarios modeled.
                                                                                   Whole-
                                                                                    Plant
                                                                                   Single-
                                     Single-Pass         Two-Pass    Three-Pass     Pass
  Harvest Scenarios             1         2       3      4     5      6      7        8
                               SP-       SP-     SP-    TP-   TP-    ThP- ThP-
    Nomenclature                                                                    WPS
                               WP       MOG LRB        SPFH LSB      LSB    LRB
   Primary Stover
 Harvesting Machine
 Conventional combine                                   X      X      X      X
    Combine w/ stover
              chopper           X        X
  Combine w/ LSB, LRB                            X
                 SPFH                                                                X
Combine or SPFH Head
          Whole-plant           X                                                    X
         Ear-snapper -
         conventional                    X       X                    X      X
         Ear-snapper -
          windrowing                                    X      X
  Stover Properties
          Stalks/leaves         X                       X      X      X      X       X
       Cob/husk/MOG             X        X       X      X      X                     X
   Quantity - % of total
                 stover         55       15     17      55    55      45    45       75
Aggregate moisture - %
                   w.b.         40       35     35      40    25      20    25       35

Harvesting Machine –
   Second or Third
     Operation
             Shredder                                                 X      X
       Rake or merger                                  X[1]   X[1]   X[1]   X[1]
                 SPFH                                   X
              LSB, LRB                                         X      X      X
    Physical Form
    Size-reduced, bulk          X        X              X                            X
       Baled - LSB, LRB                         LRB           LSB    LSB    LRB
   Storage System
   Anaerobic – bunk or
               bag silo        Bag      Bag            Bag                          Bunk
       Anaerobic – film
         wrapped tube                           X[2]          X[2]   X[2]
                                                                                                         82

      Aerobic – covered                                X                    X            X
    Aerobic – uncovered                                                                        X
Additional Assumptions
 Days available for field
                   work             25       25       25           25       23          20     21   40
Hours available per day             10       10       10           10        8           8      8   10
[1] Merging was done as deemed economically advantageous
[2] Bales wrapped only if moisture above 30% (w.b.) otherwise bales were stored under cover.
                                                                                                             83

                             Table 3.4. Machine configurations and assumptions.
                                                                                                    Three-
                                                                Single-pass            Two-pass      pass
                                                                                       Combine
                                                                     Combine w/            w/       Combine
                                                       Combine       conventional     windrowing     w/ ear
                                                      w/ whole-       ear snapper     ear snapper   snapper
                                                      plant head         head            head         head
Stover Harvest Scenario                                   1           2        3         4, 5         6, 7


       Productivity
       Assumptions
           Working width               m                 6.1         9.1       9.1        9.1         9.1

       Assumed grain loss          % of total            2.5          2        2          2            2

       Engine rated power             kW                 350         375      375        350          350
           Field efficiency            %                 73           77       75         78          80
                                            -1
  Maximum field speed               km(h)                7.2         8.0       8.0        8.5        10.0


         Maximum
       Instantaneous
         Mass-Flow
                                    Mg DM
           Grain and MOG                -1               54           69       80         78          83
                                     (h)


Economic Assumptions
Additional annual usage                h                 150         150      150        150          100
              Machine life             h                3,000       3,000     3,000     3,000        3,000
                             [1]
         Retail list price             $                540k         570k     560k       560k        500k
                                           -1
           Typical fuel use          L(h)                63           69       68         68          66
       Accumulated repair
                                   % of Pu[1]            75           65       65         66          65
                      costs
 [1]        Pu – inflation adjusted purchase price.
                                                                                                                              84



                    Table 3.4. Machine configurations and assumptions (continued).
                                                                                  SPFH w/          SPFH w/
                                                                                  windrow        whole-plant           Stalk
                                                           LSB         LRB        pick-up         corn head          shredder
 Stover Harvest Scenario                                   5, 6        3, 7            4                8                6, 7

Productivity Assumptions
            Working width                m                                                             7.5               6.1
      Engine rated power                kW                112          67            460              600                112
                                                                         [4]
           Field efficiency              %                 70         70              75               75                 80
                                            -1
    Maximum field speed                km(h)              11.3        11.3           12.9             12.9               9.7

  Machine Parameters
                                       Mg DM
Capacity mass flow rate[1]                 -1
                                        (h)                 55         35             75              100

          Assumed grain loss          % of total                                                     0.5%
                                                          Bale
      Stover physical form
                                                           d         Baled           Bulk             Bulk          Windrowed

 Economic Assumptions
  Additional annual usage                 h                100         100           300              300                 0
               Machine life[2]            h                [2]         [2]          4000             4000               2000
                                [3]
            Retail list price             $               121k         45k          400k             550k                37k
             Typical fuel use           L(h)-1              26         16             84              109                 28
 Accumulated repair costs             % of Pu               75         90             87               65                 75
    [1]    Baler instantaneous maximum mass-flow was assumed for dry stover. If moist stover is baled, mass-flow is
           typically reduced, but this was not accounted for in the model.
    [2]    Typically baler life is quantified by number of bales, other machines in h. Baler life when baling primarily hay was
           assumed to be 50,000k bales, however stover bales were counted as the equivalent to 1.4 hay bales because
           stover is known to reduce baler life and have greater repair costs per bale. For example a baler could experience
           25,000 bales of hay and 15,000 bales of stover before it is sold for salvage value.
    [3]    Machine purchase price by producer (Pu) was assumed to be 85% of the retail list price.
    [4]    LRB field efficiency does not include time to stop, wrap, and unload bales. This is directly calculated and
           accounted for in the model.
                                                                                                                       85

  Table 3.5. Time for queuing, loading, and unloading for corn stover transport systems.
                                                  LRB             LSB            SPFH
                                                 Bales           Bales            Bulk
       Field to Storage
              Loading queue min(load)-1             5               5               2
                Loading time       "               15              10              [1]
           Unloading queue         "                5               5               2
             Unloading time        "               15              10               5

     Storage to Processor
              Loading queue min(load)-1                              5                    5                     5
                Loading time     "                                  10                   10                     5
           Unloading queue       "                                  10                   10                     2
             Unloading time      "                                  10                   10                     3
           [1]       Loading time was calculated by the harvest rate of the SPFH (Mg/hr) and the load limit
                    (Mg or cubic meters) of the transport container.




                 Table 3.6. Specifications and assumptions for corn stover storage.

                                            Moist, Bulk                                     Bales
                                                                           Film           Covered[1]          Uncovered
                                         Bag          Bunker
                                                                         Wrapped            - LSB               - LRB
        Ground                                                                             Gravel
 improvement                           None          Concrete               None             pad                None
       Storage    Mg                    0.18,          0.78,
      footprint DM(m)-2                0.33[2]        1.43[2]                0.16             0.65              0.08
        Capital
   expenditure $(m)-2                    1.1            26.9                  1.1             10.8               1.1
  Annual fixed   $(Mg
          costs  DM)-1                   0.3            3.6                   0.3              0.9               0.5
Annual variable
          costs    "                    13.9            5.1                   5.2              0.3               0.0
  Total annual
          costs    "                    14.2            8.7                   5.5              1.2               0.5
 Storage losses % of DM                   3               5                    3                3                12
 Final moisture         % w.b.                                                                 20                30
      Change in
       moisture        % units            0               2                    0
   [1]   Covered storage is stover bales on a rock pad and covered with a tarp. Indoor storage was evaluated and was
         considerably more expensive than covering with tarp.
   [2]   First number is for corn stover, second number is for whole-plant corn silage
                                                                                                 86



         Storage losses in the whole-plant silage harvest system were considered as starch

losses and fiber losses, as these fractions have very different economic values. The losses,

2-3% of the whole and 1% starch losses, were increased from the values reported in Cook

and Shinners (2011) to be conservative and with the assumption of greater storage losses in

a bunker silo as compared to well-sealed bag silos. The assumed 5% loss of grain in storage

is the primary reason for the high costs of storage losses associated with this system.

However, the savings from the avoidance of grain drying and grain bin storage was

incorporated into the storage costs of the whole-plant system, which frequently provided a

savings greater than the cost of ensiling the whole-plant material, so the net storage costs

of the whole-plant material was often negative.

         Storage costs were broken into two parts, storage costs and costs associated with

DM losses. Storage costs were calculated as the annualized storage structure fixed costs,

plus the annual operating costs. Examples of operating costs included covering bales with

tarps, filling silo bags, packing bunker silos, or stacking bales. Storage loss costs were

calculated by:

                                            1       
                    C         V         1  L   1
                                                                         [3]
                        st         sv           
                                                       
                                                      




where:                                    Cst   costs of DM loss, $(Mg DM)-1
                                          Vsv   modeled value of stover in storage, $(Mg DM)-1
                                            L   loss of DM during storage
                                                                                             87

3.2.4 Stover Transport
       The biomass supply area for the scenarios modeled were described as a set of

concentric diamonds nested around a central processing site (Appendix A). The variable

costs of hauling for each concentric area as a function of its distance from the processor

were then calculated. It was assumed that the primary driver for producers participating in

stover collection was the difference between their costs and the price they receive (i.e.

profit). It was further assumed that stover harvest participation would follow a gamma

distribution which described participation as a function of this difference (Appendix A).

Assuming cost of production was similar across all farms, the difference in profit was

primarily a function of hauling distance and cost. The total yield and participation ratio for

each concentric area along with average haul distance for that area was used to create a

supply curve. The supply curve was integrated over an infinite land area, limited only by

distances that the transportation costs cause the total costs to exceed the price paid at the

plant. This integration defined the annual supply for a given price and the average,

geometric mean transport distance (Appendix A). Using the integration of the supply curve

over the supply area and the Excel spreadsheet solver function, the biomass price could be

calculated to achieve a specified annual supply.




3.2.5 Stover Processing
       We assumed stover would be biochemically converted to ethanol. Before the

conversion process can begin, delivered stover must be size-reduced and made into a slurry.

For this analysis generic feedstock assumptions to evaluate each system were made. For all
                                                                                                88

systems the feedstock was taken to a 20 mm particle size and 55% moisture content. The

moisture content is based on an estimated water requirement of a net 1.25 Mg of water

required for processing 1 Mg DM of stover into ethanol. The added water was estimated to

cost $0.95(Mg)-1 water and require 1 MJ(Mg)-1 water for well pumping. For this analysis we

assumed a coarse grind that would result in the bale being turned into a bulk material

similar to the bulk systems analyzed herein. Kaliyan et al. (2010) found that for bales at 15%

moisture content (w.b.) ground to a mean particle size of 17.4 mm, the specific energy

consumption was 241.3 MJ(Mg)-1. Using the same grinder and the ASABE standards for

capital recovery and repairs and maintenance; the grinding cost was estimated at

$6.1 (Mg)-1 (electric power), plus $2.1(Mg)-1 and 21 MJ(Mg)-1 for handling and loading

(Kaliyan et al., 2010). It is known that moisture content of biomass drastically reduces the

efficiency of bale grinders; however these decreases in performance were not modeled.

Handling costs at the processor, beyond the unloading of the stover, of all bulk material was

estimated as $2 (Mg)-1 and 20MJ (Mg)-1. For the above mentioned processing costs, a 15%

margin was assumed over the feedstock preparation and processing costs to estimate the

feedstock cost to the conversion facility.




3.2.6 Profit Margin
       To produce a more realistic estimate of feedstock cost, a profit margin was assumed

for the harvest, storage, and transport operations (table 3.1). The profit margins were

assumed to provide compensation to the service provider in exchange for their acceptance

of risk. In the following results and discussion sections, the term cost is related to one or
                                                                                               89

several logistics steps and refers to the actual cost plus the assumed profit margin (see

nomenclature section). Biomass value is the modeled cost plus the profit margin (table

3.1), while biomass price is the modeled price that needs to be paid by the processor, to

achieve the given annual supply. The term feedstock cost is the sum of the biomass price

and the cost, plus margin, for handling and processing after being unloaded at the facility.




3.3    Model Results and Discussion


3.3.1 Feedstock Costs for 670 Gg Supply
       During the 1990s, grain ethanol plants were frequently of the 100 million liter per

year size, and climbed to 500+ million liter per year facilities through the present time. For

this analysis a supply of 670 Gg was used, which if utilized for ethanol would likely be in the

200 to 225 million liter per year range. It was assumed no other nearby facilities that would

be competing for corn stover, which would have the effect of constricting the supply area

and increasing the price paid for biomass. Table 3.7 summarizes the model results for each

of the eight harvest scenarios (table 3.3) for a 670 Gg supply. The sum of handling and

transportation for the entire annual biomass supply were the greatest cost associated with

any biomass harvest scenario, ranging from 36% to 64% of the total costs. A single-pass

combine that harvests 55% of the available stover as a chopped, bulk material (scenario 1)

slows the combine and increases grain harvest losses, resulting in high harvest costs.

However, the high yield reduces the transport distance and cost. A similar system which
                                                                                                90

only harvests 15% of the stover (scenario 2) had lower harvest costs but much greater

transport costs due to the long transport distance, caused by a low yield. This resulted in

the costs of these two systems being similar. Re-configuring the single-pass combine to

produce LRBs (scenario 3) instead of chopped, bulk material increased harvest costs due to

reduced combine productivity and greater capital costs. The costs of handling and

processing bales were substantial, which could not be offset by the lower costs of storage

and transport of the greater density, dry bales, so this scenario had the greatest cost of the

single-pass systems. An unpublished scenario identical to scenario 3 using a LSB, rather

than a LRB, resulted in greater costs. This was due to the greater capital cost of the baler,

which is underutilized in a low yield situation, such as MOG baling.

       In the two-pass scenarios (4 and 5), the cost of forming windrows with the combine

was $2.7(Mg)-1. It was assumed windrows could either be chopped or baled (scenarios 4

and 5, respectively). The combination of high yields (55%) and high capacity harvesting with

the SPFH (scenario 4) resulted in harvest costs of $9.7(Mg)-1, substantially less than baling

costs of $28.8(Mg)-1 (scenario 5). Transport costs were similar between the two systems

because it was assumed that the density of the chopped material in the silo bag (which is

similar to that of a LSB) would be maintained during transport. Although storage costs

were high for the chopped, bulk material, these costs were more than offset by lower

handling, harvest, and processing costs compared to the baled system (table 3.8).

Therefore, the two-pass, chopped scenario had the lowest cost of the harvest scenarios

considered where grain and stover were separated at harvest.
                                                                                                 91

       The conventional three-pass scenarios generally had the greatest costs of those

considered. Harvest costs were high because of the additional cost of forming windrows

with a separate operation in addition to the high costs of baling. LRBs have a lower harvest

cost than the LSBs; however the inefficiency of transporting of LRBs, caused by their shape

and low density, offset the harvest savings. The major divergence point of the two systems

is the storage costs between the LSB/covered storage and LRB/outdoor storage scenarios (6

and 7, respectively). High storage losses, moisture content gains, and material

deterioration associated with outdoor storage of stover (Shinners et al., 2007a) increases

storage costs by $10(Mg)-1 and results in a lower quality feedstock. This quality impairment

was not monetized or captured by our modeling, but could have an important impact on

final product costs.

       Baling corn stover is rigorous and increases wear and the depreciation rate of balers.

To account for this the baler repair costs (table 3.4) were held constant while reducing the

expected baler life in the model by 40%. With these assumptions, the baling costs of $6.7

to $10.8(bale)-1 returned by the model were within the range reported in custom rate

surveys of $6-13(bale)-1 (USDA-NASS, 2011; Edwards et al., 2011), suggesting that our bale

costs are reasonable.

       The biomass price (table 3.7) is the price that a processor must offer to achieve a

desired annual supply (in this case 670 Gg). The biomass price is calculated from the sum of

explicit process costs, the stover fertilizer value in the field of the stover ($21.8(Mg) -1), and

the profit necessary for a given annual supply, as modeled by the supply module (Appendix
                                                                                               92

A). For larger annual supply requirements a greater price must be offered to induce more

market participants. Systems that have high densities and low moisture contents are more

efficiently transported. As such, we found the average transport distance increased for the

dry bale systems, increasing the collection area and potential market participants that

would accept smaller profit margins. This had the effect of lowering the price and producer

profit potential that was needed to be offered to achieve the given supply size. In other

words, for material with lower transportation efficiencies (i.e. any of the low density or

higher moisture systems), a smaller geographic area is available for economic collection,

and therefore a greater price must be paid to source the same amount of biomass. The

feedstock cost to the biomass processor (table 3.7) is the biomass price plus the biorefinery

processing costs to produce ready-to-be-converted, size-reduced slurries plus the profit

margin of this operation. Of the systems considered where the stover and grain are

separated at harvest (scenarios 1-7), the two-pass chopped, bulk system produced the least

feedstock cost at $86(Mg)-1, a difference of $27.2(Mg)-1 or 24% less than the conventional

three-pass LSB, covered storage scenario.

       The single and two pass systems reduced the harvest labor requirements (h(Mg)-1)

compared to the three-pass systems (table 3.9). For these systems, each laborer during the

harvest season could produce approximately 800-1000 Mg with the exception of the two-

pass chopped system where a single laborer could produce 1300 Mg. A comparison of fuel

use to the farm-gate revealed bale handling versus bulk handling had greater fuel use by 1.5

L(Mg)-1, but less storage fuel use by 1 L(Mg)-1. Fuel use for harvest varied across systems,
                                                                                                93

but depended most on the number of passes, the yield, the field efficiency of the harvester,

and the fuel expenditures required for handling and size-reduction. Transportation from

storage to the processor accounted for 57% to 78% of the total fuel use. The most

significant differentiating factor found in the analysis of the capital requirements ($(Mg)-1)

was that bale systems require a large investment in handling equipment due to the unit

handling that is required.

       Bale systems have greater harvest, handling, and processing costs, whereas the bulk

systems have a greater cost of storage. The bulk harvested systems (scenarios 1, 2 and 4)

incur storage costs four-fold that of well-managed baled systems (scenarios 5 and 6). Baling

corn stover and then deconstruction of the bales to get a size reduced material had a

significant effect on the overall costs, making up 15% to 32% of the total costs (table 3.8).

Size-reduction in the field at the time of harvest with the SPFH is much more cost effective,

making up only 8% of the total feedstock cost.

       When yield was similar, the transport economics slightly favored LSBs and the MOG

LRBs made with the combine over chopped, bulk stover (i.e. distance variable costs for

scenario 5 and 3 versus 4) (table 3.10). Even when mechanized with focused bale handling

equipment, the handling of bales makes up a significant portion of the total feedstock cost

because it was very labor and capital intensive, while adding no value to the material. The

major inefficiencies of the bulk systems, particularly in low yield situations, is the time spent

filling a semi-trailer and transporting small quantities at low density to storage (e.g. scenario

2) (table 3.10). LRBs of stover have a bulk density that is too low and the bale shape does
                                                                                               94

not optimize the load volume capacity. LSBs on the other hand have good density and fully

take advantage of the load volume capacity. Bulk systems have good density, make

reasonable use of the load volume capacity, and have low handling costs associated with

them (table 3.10). To compare bulk and bales more generally, baling is the compaction of

biomass in the field followed by an extremely energy intensive process of bale

deconstruction followed by size-reduction to make the bales into a bulk format. With the

bulk system, size-reduction takes place at harvest, but has greater storage costs associated

with it.

           Stover has value to ruminant animal producers as a high-fiber roughage feed or as

animal bedding. The farm-gate cost of stover(table 3.11) was determined by summing the

non-nutrient value of stover with the costs of harvest through storage, plus processing

costs. Here the stover value does not include the value of P and K because it was assumed

these elements would be returned to the land in the form of manure or a bedding pack.

Similar to when stover was considered a biomass feedstock, the two-pass bulk stover

systems (scenario 4) provided the lowest cost material. Bulk systems have the further

advantage in that they lend themselves to pre-storage application of amendments to

enhance stover fiber digestibility by ruminant animals (Hadjipanayiotou et al., 1984).
                                                                                                                         95

  Table 3.7. Summary of median transport distance, and cost plus profit to harvest, handle,
             store, transport and process a 670 Gg annual supply of corn stover.
                                                                 Scenario No.
                        1             2             3             4             5              6             7            8
Nomenclature         SP-WP       SP-MOG          SP-LRB       TP_SPFH        TP-LSB        ThP-LSB       ThP-LRB        WPS


                                                    Median transport distance .. km
                       73           101           104            65            74             80            81           36


                                                              Baling Cost + Profit
       $(bale)
                -1                                 6.7                     10.0              10.8           9.2
   $(Mg DM)
               -1                                 10.7                        26.1           28.1          20.1


                                                         Cost + Profit[1] .. $(Mg DM)-1
        Harvest       24.5          9.5           26.0           9.7          28.8           36.2          28.1         -2.9
        Storage       17.3          17.0          10.5          16.4           4.0            4.2          14.0         -1.8
      Handling         1.7          1.7            9.9           1.7           8.8            8.8           5.0          0.3
     Transport        34.4          50.3          33.9          27.7          27.0           28.8          27.0          15
                                                                                                   [2]
    Processing         3.5          3.7           14.0           3.5          14.3          14.4           14.3          6.5
                                                                                                   [2]
           Total      82.2          83.4          95.3          59.6          83.6          93.2           89.0         17.1


                                                          Cost + Profit .. % of total
        Harvest        30            11            27            16            35             39            32           -17
        Storage        21            21            11            28             5              5            16           -11
      Handling          2             2            11             3            11             10             6            2
     Transport         43            62            36            47            32             31            30           88
    Processing          4             4            15             6            17             15            16           38


                                                                  $(Mg DM)-1
 Biomass price        99.6         109.8         107.7          82.4          90.4           98.8         100.1         49.7
Feedstock cost        103.1        113.5         121.7          85.9          104.7         113.1         114.4         56.2
 [1] Feedstock quality differences were not taken into account. Single and two-pass systems harvest stover with much less
    soil contamination than conventional three-pass systems (Shinners et al., 2011a). Anaerobic and fermentation of
    stover produces volatile fatty acids (Shinners et al., 2011b) to which some biochemical conversion systems are sensitive.
 [2] LRBs stored outdoors gain moisture during storage due to exposure to precipitation, so Scenario 7 yields bales of higher
    and variable moisture content. The greater cost of grinding moist bales was not accounted for, but should be
    considered.
 [3] Bales resulted in greater handling costs than bulk material, due to the significant unit handling operations of bales as
    compared to bulk material handling.
                                                                                        96



Table 3.8. Cost to bale, deconstruct bale, and size-reduce or cost to
                   size-reduce stover with SPFH.
                                                   Scenario No.

                                  3           5            6           7          4

                                        Price or cost .. $(Mg DM)-1

          Biomass price 107.7               90.4        98.8        100.2        82.4

          Feedstock cost 121.7             104.7        113.1       114.4        85.9

     Baling cost + profit 10.7              26.1        28.0        20.1

  Size reduction cost[1]         7.0         7.0         7.0         7.0         6.9

                                Process cost .. % to total feedstock cost

                     Baling       9          25           25          18

          Size-reduction          6           7            6           6          8

                Bale total       15          32           31          24

                Chopping                                                          8
    [1]   For bales (scenarios 3, 5-7), the cost to unwrap net wrap from bale,
          deconstruct bale and size-reduce in tub grinder or similar. For bulk
          material (scenario 4), the cost to chop in the field with the SPFH.
          Grinding at the biorefinery is not needed.
                                                                                                                    97




 Table 3.9.    Summary of diesel fuel, labor and capital requirements to harvest, handle, store,
                  transport and process a 670 Gg annual supply of corn stover.
                                                              Scenario No.
                        1            2             3           4        5                 6             7           8

                                                   Diesel fuel use[1] .. l(Mg DM)-1
       Harvest        1.4          0.5            2.3       1.7        0.8       2.2                   2.1         1.2
       Storage        1.1          1.0            0.8       1.1        0.1       0.1                   0.4         1.2
     Transport        5.9          7.6            9.7       5.4        7.1       7.5                   7.1         4.2
          Total       8.4          9.2           12.7       8.1        8.0       9.8                   9.6         6.7

                                                Labor requirements .. h(Mg DM)-1
       Harvest       0.036        0.017         0.013    0.033   0.052      0.127                    0.159        0.011
       Storage       0.084        0.086         0.076    0.083   0.018      0.019                    0.025        -.004
     Transport       0.244        0.360         0.373    0.192   0.310      0.318                    0.246        0.109
          Total      0.364        0.463         0.462    0.308   0.380      0.464                    0.430        0.116
Mg / Laborer[2]      1040          820           970     1290     870        550                      730         9900

                                                  Capital intensity .. $(Mg DM)-1
       Harvest         37           25           58         22        25       32                     21           -10
       Storage         16           15           13         15        32       32                     21           -26
     Transport         39           47           82         36        98      114                     78            16
          Total        92           87           153        73       155      178                     120          -20
     [1] Energy requirements for bale deconstruction and size-reduction at the processor are not included because these
         processes are often powered by electric motors.
     [2] Represents the mass removed from storage per laborer, not including labor for transport from storage to
         processor.
                                                                                                                      98

      Table 3.10. Cost plus profit to load, handle and transport stover in bale or bulk form.
                                                                    Scenario No.

                                    3          5            6            7                1           2           4

                                               Bale Systems                                    Bulk Systems

                 Bale type        LRB         LSB          LSB          LRB

                     Yield[1]      17         55           45            45              55          15          55

                                                          Cost + Profit .. $(Mg DM)-1

      Field to Storage

                 First km[2]     14.4        12.5         12.5          4.7             12.8        22.0         7.8

        Distance variable
                   cost[3]       0.22        0.23         0.23            -             0.51        0.51         0.51

                 Sub-total       14.9        12.9          13           4.7             13.8         23          8.9

Storage to Biorefinery

                 First km[2]      6.7         7.0          7.1          7.8              4.7         4.7         4.7

Distance variable cost
                           [3]
                                 0.22        0.23         0.23         0.25             0.25        0.25         0.25

                 Sub-total       30.2        23.8         25.6         29.6             23.2        30.2         21.2

                       Total     45.1        36.7         38.5         34.3             37.0        53.3         30.1
[1]    Assumed yield of stover as percent of total stover available.
[2]    The cost to handle bale (if applicable) and/or load the transport vehicle including any idle time.
                                     -1
[3]    Marginal cost ($(Mg DM-km) ) of transporting the remaining distance from field to storage or storage to
       biorefinery.
                                                                                                                    99

  Table 3.11. Farm-gate costs ($(Mg DM)-1) of biomass for use on farm as roughage feed
                                  or animal bedding.
                                                          Scenario No.
                           3             5            6           7              1          2          4     8
                                     Bale Systems                                    Bulk Systems
      Bale type          LRB           LSB          LSB         LRB
   Process cost,
    less profit[1]       54.2          49.1        55.1        47.7            49.0       44.5       32.3   (0.4)
          Total[2]       70.7          65.6        71.6        64.2            65.5       61.0       48.8   16.1
        [1]   Process costs include; harvest, transport to storage, storage, and post-storage processing.
        [2]   Total is the process cost, plus the non-soil nutrient value of stover




3.3.2 Whole-Plant Silage System
        We have suggested a paradigm shift in which stover and grain, destined for

biochemical conversion or processed animal feed, would no longer be harvested, stored

and transported separately. Alternatively, corn grain and stover at black-layer or later

maturity would be harvested simultaneously using a SPFH and the combined fractions

transported and placed into anaerobic storage (Cook and Shinners, 2011). After storage,

the combined grain and stover would be shipped to a processor, where the grain and stover

fractions would be separated and both fractions biochemically converted or further

processed to value-added animal feeds. The whole-plant system has many advantages.

Only one harvest operation is required and machinery investment is dramatically reduced

(table 3.8). The harvest capacity of a SPFH is much greater than that of a combine. The

whole-plant silage system is more robust than separate grain/stover harvest systems,

resulting in a longer harvest window, thereby significantly reducing the capital and labor
                                                                                               100

intensity of the harvest to storage logistics (table 3.8). For all these reasons, harvesting

grain in this fashion was less expensive than conventional grain harvest, essentially resulting

in negative harvest costs for the stover fraction (table 3.7).

       With this scenario, grain drying and bin storage are eliminated. Therefore, storing

grain by co-ensiling with the stover fraction was less expensive than conventional drying

and bin storage, essentially resulting in negative stover storage costs (table 3.7). In 2011

the estimated cost for grain drying was $138 (ha)-1 (Duffy, 2011) and grain storage for 6

months was $55 (ha)-1 (Edwards et al., 2011). A more conservative assumption for grain

drying of a 6 percentage unit reduction in grain moisture content was made. This results in

a grain drying cost savings of $116(ha)-1 and a heat energy input savings of

307 MJ(Mg stover)-1 that could be realized. Considering the whole plant scenario has

energy inputs of 311 MJ(Mg stover)-1, this system would result in an overall system where

stover could be received and processed at the biorefinery and yield an overall lower energy

input than traditional grain to a refinery gate The transportation costs were also reduced

by combining the low-density stover with the high-density grain where the load is weight

limited, regardless of compaction or moisture content. Processing costs were the highest

among the bulk scenarios considered because of the added costs to separate the grain and

stover fractions. We envisioned a hybrid mechanical-hydrodynamic separation process

which has proven successful in the laboratory (Savoie et al., 2004; Cook and Shinners, 2011)

and assumed a cost to stover of $3.0 (Mg DM)-1 because the separation process can be

diluted over the entire year.
                                                                                              101

       The whole-plant silage system produced feedstock costs that were 51% less than

conventional three-pass LRB system (scenario 7) and 35% less than the least cost two-pass

system (scenario 4) (table 3.7). This system offers the best opportunity to significantly

reduce corn stover feedstock costs. If the stover removal ratio is too great with this

scenario, then cover crops would be required to reduce the erosion of this harvest

approach. A survey of Corn Belt producers found that 56% of producers would use cover

crops if an incentive payment was offered, and would require, on average, a minimum

payment of about $57 (ha)-1 (Singer et al., 2007). In scenario 8, a stover yield of 8.3 Mg(ha)-1

would suggest cover crops would raise the cost of the stover $6.8 (Mg)-1. Even with this

added cost, the overall feedstock cost is still significantly less than any other system

considered in this work.

       Although this system has many positive attributes, there are hurdles to adoption.

First, there is a significant cultural change away from the engrained manner that corn grain

is harvested. The approach to marketing of corn will also change. If there is limited

competition for stover in the area, then we envision that both the producer and processor

will require that the ensiled grain and stover be contracted prior to harvest. This eliminates

the opportunity for the producer to exploit multiple potential buyers for the lowest possible

grain basis (i.e. future minus cash price). However, this contract merely needs to provide

for a basis level, it would not need to be a complete forward contract. A basis only contract

allows the producer to continue all other mechanisms currently employed for hedging and

the ability to market their grain any day of the year, as it fits their marketing plan. For this
                                                                                                 102

system, it is envisioned that forward or basis contracts will be the most likely grain

marketing tools. Another limitation of the system is that it is likely not compatible with

systems that require dry milling or the export markets. However, this new harvest system

was intended primarily for harvesting corn destined for ethanol production in the US.


3.3.3 Sensitivity Analysis
    Sensitivity of feedstock costs, specific energy, labor requirements, and capital intensity

to changes in stover yield, moisture, transport density, annual machine usage, and scale of

biorefinery were conducted. Of the sensitivity analyses conducted, the biorefinery size and

stover yield had the most profound effects on feedstock costs, while the remaining

variables had only minor impact. Transport costs accounted for a significant fraction of a

system’s total costs (table 3.10). As such, yield and facility size have direct effects on the

transport distance, and the most direct effect on total feedstock costs (fig. 3.1 and 3.2).

Additionally, yield is the primary driver of harvest costs, and low yields exacerbate harvest

costs. Figure 3.1 shows a downward trend for the price of whole plant silage system

(scenario 8) as the fraction of stover harvested decreases. The reason for this is the savings

from the grain fraction is spread over a smaller mass of stover.
                                                                                                                        103


                               $200

                               $180

                               $160
   Feedstock Cost .. $/Mg DM




                               $140

                               $120

                               $100

                               $80

                               $60

                               $40
                                      10%       20%      30%       40%      50%      60%      70%       80%      90%
                                                                  Fraction of stover removed .. %
                                            Scenario 1            Scenario 2           Scenario 3             Scenario 4
                                            Scenario 5            Scenario 6           Scenario 7             Scenario 8




             Figure 3.1.                    Sensitivity of corn stover feedstock costs to fraction of total available
                                            stover removed by eight different harvest scenarios supplying a
                                            biorefinery requiring 670 Gg annually.


         Considerable attention has been paid to stover bulk density and moisture, but the

model results show that bulk density had a negligible effect on feedstock costs, unless the

load was volume limited (fig. 3.3). The bale based scenarios all have weight limited

transport because of bale density and moisture content. The bulk systems show

considerable sensitivity to density until about 150 kg(m)-3 is reached (inflection points vary

by moisture content and load volume limitations). An interesting relationship that was
                                                                                              104

found in general for density and moisture content is that, low moisture, high density

biomass is transported more efficiently than high moisture, low density biomass. Therefore

there is economic incentive to transport dense biomass a longer distance. This still results

in a more economical system as the processor can reach more low cost biomass.

Development efforts should focus on assuring that the density achieved during storage of

bulk materials is maintained during transport.

    Moisture content had negligible effect on feedstock cost, unless the load was weight

limited (fig. 3.4). The discontinuity of the bale systems’ cost is related to the shift from dry

storage to wrapped ensiled storage above 30% (w.b.) moisture. Two caveats related to

moisture content should be considered, neither of which is modeled here. A more robust

system that can handle a wider range of harvest moisture contents will result in a more

secure supply, likely adding value for a processor. Also, baler and bale grinder capacity

decreases and cost of deconstruction and size reduction increases with greater moisture

content.
                                                                                                                 105




                            $160


                            $140
Feedstock cost .. $/Mg DM




                            $120


                            $100


                            $80


                            $60


                            $40
                                   91     544        998    1451     1905     2359     2812     3266      3719
                                                           Facility size .. Gg DM annually

                                        Scenario 1          Scenario 2            Scenario 3           Scenario 4
                                        Scenario 5          Scenario 6            Scenario 7           Scenario 8




                            Figure 3.2. Sensitivity of corn stover feedstock costs to size of biorefinery for
                                        eight different harvest scenarios.
                                                                                                              106


                             $220

                             $200
Feedstock cost .. $/Mg DM



                             $180

                             $160

                             $140

                             $120

                             $100

                              $80

                              $60

                              $40
                                    32         72       112         152       192        232         272       312
                                                        Post storage transport density .. kg DM/m3

                                         Scenario 1           Scenario 2         Scenario 3           Scenario 4
                                         Scenario 5           Scenario 6         Scenario 7           Scenario 8




                            Figure 3.3. Sensitivity of corn stover feedstock cost to post storage transport
                                        density for four different bulk harvest scenarios supplying a biorefinery
                                        requiring 670 Gg annually. Density of bales was great enough that the
                                        load was always weight limited, so bale scenarios were not included.
                                                                                                                        107

                                $160


                                $140
   Feedstock cost .. $/Mg DM




                                $120


                                $100


                                 $80


                                 $60


                                 $40
                                       10%      15%       20%     25%      30%      35%        40%   45%      50%

                                                                Moisture content .. % (w.b.)
                                             Scenario 1          Scenario 2           Scenario 3           Scenario 4
                                             Scenario 5          Scenario 6           Scenario 7           Scenario 8




                               Figure 3.4. Sensitivity of corn stover feedstock cost to harvest moisture content
                                           for eight different harvest scenarios supplying a biorefinery requiring
                                           670 Gg annually.



3.4                            Discussion
                               When other factors are relatively equal, the bulk systems generally produced

feedstock at lower cost than baled systems. Although bulk systems had slightly greater

storage and transport costs, these costs were offset by less harvest and processing costs

compared to baled systems. Harvesting chopped material that is stored anaerobically and

preserved by fermentation is a more robust system than dry bales, reducing the weather
                                                                                             108

related harvest risks. Chopped, bulk systems also offer the possibility of on-farm pre-

treatment to enhance stover value for biochemical conversion or animal feed.

       The model results suggest that high yields are encouraged by the producer’s desire

for a profit and the processor’s desire for the lowest cost feedstock, when the transaction is

on a dollars per mass or sugar basis. High yields also result in the least energy, labor, and

capital inputs. However, biomass that results in depletion of SOC or unsustainable soil

erosion renders the system non-sustainable, inconsistent with the goals of a sustainable

renewable energy resource. Using Scenario 4 as an example, the difference in feedstock

cost for harvest removal ratios of 35% and 70% of available stover yield is $15 (Mg)-1. Using

an 8.54 Mg DM(ha)-1 grain yield and the 1:1 corn to stover assumption, this difference

would result in a $90 (ha)-1 difference. If this $90 (ha)-1 of improved profitability could be

used in part for agronomic (e.g. cover crops) or harvest practices that offset or improve the

soil sustainability, a much greater yield could be accomplished, resulting in overall greater

profitability, less inputs, and a more sustainable system.

       The economic modeling did not take into account the risk premium of a failure to

harvest the feedstock. It can be expected that more robust harvest and storage systems

will receive a premium in price, because when corn stover harvest systems are significantly

affected by weather extremes, the conversion facility has a higher probability of economic

failure and lower total returns (Kou and Zhao, 2011). A large portion of the feedstock cost is

the P and K value of the biomass. These inorganic elements are not consumed in
                                                                                              109

conversion processes and if recovered and returned to the field could offset fertilization

costs and bring down net feedstock costs.

       Feedstock quality is not addressed in this model, as it depends highly on the

downstream process specifications. For the purpose of using this model to assess economic

feasibility of downstream processes, all cited stover DM values are assumed to contain 7%

ash. However, ash content increases with each field operation. Ash content was reported

to be 3.9-5.5; 5.8-6.0; and 9.8-10.3% of DM for single-pass, two-pass and conventional

three-pass stover harvest systems (Shinners et al., 2009; 2011a). Reduction of feedstock

quality by soil contamination should be considered when comparing feedstock costs of

different harvest scenarios.

       Both harvest and storage influence biomass composition and quality. Harvest

systems change the compositional characteristics of each system. Scenarios 1 and 8 use a

whole-plant corn head which removes the highest portion of the stalk, most of the leaves,

and all the cobs. The stover collected in scenarios 2 and 3 is dominated by cobs, which have

greater sugar content. Scenarios 4 and 5 can capture most of the cobs and are partially

selective for the leaves and higher stalk. The three-pass conventional scenarios (6 and 7)

recover little cobs (Shinners et al., 2007c) and are not selective for stover fractions. In

storage, the losses generally come from the most digestible, most valuable components. So

systems which have high storage losses should incur a cost associated with both quality and

quantity loss. Dilution of feedstock value by different composition should be considered

when comparing feedstock costs of different harvest scenarios.
                                                                                               110

       Corn stover digestibility decreases with maturity, as well as the total mass remaining

in the field (Bal et al., 1997). Earlier harvests at greater moisture contents may result in

greater yields of more digestible stover. For this greater moisture harvest, anaerobic

storage becomes necessary. This anaerobic storage of biomass can make systems more

robust by widening the harvest window and reducing the risk to processors of not receiving

a crop. It also provides an environment where fermentation occurs and results in

production of volatile fatty acids, and these have positive, negative, or indifferent values,

depending on the downstream processing.




3.4.1 Comparison to Literature
       A review of literature reveals three primary differences of this model from others.

One variation in all work is the nature of changing prices; every paper presented does so at

a different timepoint. During 2006-2008, a new level of commodity prices have emerged,

and through 2011 there has been tremendous variation of these prices, resulting in a wide

variation in reported fertilizer prices and fuel prices. Another difference in this model from

that of the literature is that profit was taken into account to better reflect the cost to a user

of biomass, rather than only the cost of production experienced by the producer. Also

considered was the entire process from the field through the processor gate to the

initiation of the conversion process. Doing this creates a much more complete comparison

of systems and a better understanding of the true feedstock costs by taking it from the field

to a size-reduced, flowable material, beyond the plant gate.
                                                                                              111

         In the literature there exists no standardization of hauling distance, either arbitrary

or for annual tonnage requirements. As transportation costs make up a majority of total

feedstock costs, this lack of standardization creates a wide variation in estimated costs of

delivered biomass. To overcome this lack of standardization and differences in fertilizer

values of stover, several papers were reviewed and compared to the model presented here

using some of their assumptions, to allow for valid comparisons.

         Sokhansanj and Turhollow (2002) found that the costs of stalk chopping, raking,

baling, and covered storage were $21.6 (Mg)-1 and $23.6 (Mg)-1, for large round bales and

large square bales, respectively. The modeling of these steps in a three-pass LSB baling

system indicates the current costs of $44.2 (Mg)-1. Some of the differences between these

values are changes in prices and the assumptions on the cost of baling. Diesel fuel costs

were estimated by Sokhansanj and Turhollow to be $0.29 (L)-1, in 2011 these prices have

increased to approximately $1 (L)-1. Baling costs were the other primary point of

divergence. Sokhansanj and Turhollow’s estimated cost of baling was $9.80 (Mg)—on the

low end of what they reported for estimated baling costs available in literature at the time,

ranging from $6.90 to $29.50. Sokhansanj and Turhollow used costs associated with baling

hay to estimate the cost of baling stover, which is likely an underestimate of stover baling

costs.

         The model estimates current baling costs of stover to be $26.1 (Mg)-1, close to, or

slightly higher than, that reported in custom harvest surveys (USDA-NASS 2011; Edwards et
                                                                                          112

al., 2011). This can be expected due to the higher cost of repairs and maintenance

associated with baling stover as compared to baling hay.

       Morey et al. (2010), reported stover delivered to the gate of a combustion facility to

have a cost of $81.29 (Mg)-1, including a 48 km haul and a stover payment to the producer

of $26.76(Mg)-1. Using the same assumptions for haul distance and stover value in the field,

this model was very similar at $85 (Mg)-1. However, there are some differences in the

logistics systems. The model presented here assumed delivery of bales, while the Morey et

al. model grinds the bales out of storage and compacts the stover with a roll-press

compaction machine, thereby adding value for a combustion facility. A look at the

individual costs shows that Morey et al. have an estimated value of stover in the field of $24

compared to the $34.7 (Mg)-1 estimated in this work. The costs of storage were similar, but

baling and handling to storage were estimated by Morey et al. to be 20% less than the

model presented here.

       Perlack and Turhollow (2003) reported costs of delivered LRBs of stover to be $47.51

(Mg)-1 to $56.88 (Mg)-1 for haul distances of 35 to 100 km, assuming an in-field stover value

$11/Mg. This was much lower than results presented here or others in literature, but a lack

of published assumptions inhibits comparison of the system.

       Petrolia (2008) reported the costs of delivered LRBs of stover to be $62 (Mg)-1 to $93

(Mg)-1 for haul distances of 53 to 183 km, assuming an in-field stover value of $4.64 (Mg)-1.

Using the stover value and haul distance assumptions from Petrolia, the model returns costs
                                                                                             113

of $63 (Mg)-1 to $91 (Mg)-1, essentially the same as the total delivered costs reported by

Petrolia.

       The IBSAL model (Sokhansanj, 2006) found that for a 450.9 Gg annual supply of LSBs,

all the process costs, from field to biorefinery except for storage, would be $53.57 (Mg)-1.

Our analysis suggests those processes have costs of $68 (Mg)-1. The discrepancy here is

primary due to the harvest costs assumed in the IBSAL model, $11.71 (Mg)-1, which is less

than the $29 (Mg)-1 estimated in this model and similar to custom rate surveys. (USDA-

NASS, 2011 and Edwards et al., 2011).




3.5    Summary
       A comprehensive economic model was developed to evaluate the effects of yield,

moisture content, bulk density, and biorefinery size on the economics of corn stover

feedstock, as quantified by costs, labor and capital requirements, and energy use. The

model considered eight different harvest and storage scenarios. Of the systems considered

where the stover and grain are separated at harvest, the two-pass chopped, bulk system

where moist stover is conserved by anaerobic storage produced the least cost feedstock at

$85.9 (Mg)-1, a difference of $27.2 (Mg)-1 or 24% less than the conventional three-pass LSB,

covered storage scenario. Storage and transport costs made up a significant fraction of

total costs with this system, but these greater costs were offset by much less harvest and

processing costs compared to other systems.
                                                                                         114

       A new approach was considered in which stover and grain would no longer be

harvested, stored, and transported separately. Alternatively, mature corn grain and stover

would be harvested simultaneously using a forage harvester and the combined fractions

transported and placed into anaerobic storage. After storage, the combined grain and

stover would be shipped to a processor, where the grain and stover fractions would be

separated and both fractions biochemically converted or further processed to value-added

animal feeds. The advantages of this approach include robustness, elimination of grain

drying, and low harvest and transport costs. The whole-plant silage system produced stover

feedstock costs that were 51% less than conventional stover harvest systems and 35% less

than the lowest cost two-pass system. While not without significant system-wide changes

and adoption hurdles, the whole-plant system offers the best opportunity to significantly

reduce corn stover feedstock costs.

       The sensitivity analysis showed that yield and the size of the processor had the most

impact on feedstock cost. They do so, primarily by the impact on transportation distance,

which was found to be the greatest cost. Achieving a weight-limited load is another

important objective for all systems. Fortunately, many of the currently employed systems

can do this with corn stover.
                                                                                             115

References

Allmaras, R. R., D. R. Linden, and C. E. Clapp. 2004. Corn-residue transformations into root
and soil carbon as related to nitrogen, tillage, and stover management. Soil Sci. Soc. Am. J.,
68: 1366-1375.

ASABE Standards. 2006a. D497.5. Agricultural machinery management data. St. Joseph,
Mich. ASABE

ASABE. 2006b. EP496.3. Agricultural machinery management data. St. Joseph, Mich. ASABE

Bal, M. A., J. G. Coors, and R. D. Shaver. 1997. Impact of maturity of corn for use as silage in
the diets of dairy cows on intake, digestion, and milk production. Journal of Dairy Science
80(10): 2497-2503.

Brechbill, S., and W. Tyner. 2008. The economics of biomass collection, transportation, and
supply to Indiana cellulosic and electric utility facilities. Working Paper 08-03. Purdue Univ.,
West Lafayette, IN.

Cook, D.E., K.J. Shinners, S.D. Williams, P.J. Weimer and R.E. Muck. 2011. Whole-plant corn
as a biomass feedstock: harvest, storage and pretreatment. ASABE Paper No. 110797, St.
Joseph, MI

Coulter, J. A., A. Nafziger, and D. Emerson. 2008. Continuous corn response to residue
management and nitrogen fertilization. Agronomy Journal. 100: 1774-1780.

Cundiff, J. S., and R. D. Grisso. 2008. Containerized handling to minimize hauling cost of
herbaceous biomass. Biomass and Bioenergy. 32(4): 308-313.

Duffy, M. 2011. Estimate Costs of Crop Production in Iowa – 2011. Available at:
http://www.extension.iastate.edu/agdm/crops/pdf/a1-20.pdf Acessed April 2011

Edwards, W., Johnson, A., and Chamra, A. (2011, March). Iowa custom rate survey.
Available at: http://www.extension.iastate.edu/publications/fm1698.pdf. Accessed May
2011.

DOE – EIA. 2011. Supply and disposition summary. United States Department of Energy
Energy Information Administration. Available at: http://www.eia.gov/petroleum/
Accessed: June 2011

Glassner, D. A., J. R. Hettenhaus, and T. M. Schechinger. 1998. Corn stover collection
project. Proceedings of the 1998 Bioenergy Conference. p. 1100-9.
                                                                                                116


Goering, C.E., M.L. Stone, D.W. Smith, and P.K. Turnquist. 2003. Off-road vehicle
engineering. ASABE, St. Joseph, MI.

Hadjipanayiotou, M. 1984. Effect of level and type of alkali on the digestibility in vitro of
ensiled, chopped barley straw. Agricultural Wastes. 10(3): 187-194.

Iowa State Extension. 2010. Economics of Residue. Available at:
http://www.soc.iastate.edu/extension/watershed/modules/pdf/residue_sheet.pdf.
Accessed May 2011.

Kaliyan, N., D. Schmidt, and V. R. Morey. 2010. Commercial scale grinding of corn stover and
perennial grasses. ASABE Paper No. 1009062. St. Joseph, Mich.: ASABE.

Kou, N., and F. Zhao. 2011. Effect of multiple-feedstock strategy on the economic and
environmental performance of thermochemical ethanol production under extreme weather
conditions. Biomass and Bioenergy. 35(1): 608-616.

Lang, B. (2002, Dec). Estimating the nutrient value in corn and soybean stover. Iowa State
University Extension Fact Sheet BL-112. Retrieved May 2011, from
http://www.extension.iastate.edu/NR/rdonlyres/5D3BD06C-F585-413C-826B-
8452EF7A22DB/4744/stovervalue.pdf

McAloon, A., F. Taylor, W. Yee, K. Ibsen, and R. Wooley. 2000. Determining the cost of
producing ethanol from corn starch and lignocellulosic feedstocks. U. S. Department of
Agriculture and U. S. Department of Energy. Golden, CO: NREL / TP-580-28893.

Morey, V. R., N. Kaliyan, D. G. Tiffany, and D. R. Schmidt. 2010. A corn stover supply logistics
system. Applied Eng. In Agric. 26(3): 455-461.

Perlack, R. D. and A. F. Turhollow. 2003. Feedstock cost analysis of corn stover residues for
further processing. Energy. 28(4):1395-1403.

Petrolia, D. R. 2008. The economics of harvesting and transporting corn stover for
conversion to fuel ethanol. Biomass and Bioenergy. 32(7): 603-612.

Richey, C. B., J. B. Liljedhal, and V. L. Lechtenberg. 1982. Corn stover harvest for energy
production. Transactions of ASAE. 25(4): 834-839.

Savoie, P., K. J. Shinners, and B. N. Binversie. 2004. Hydrodynamic separation of grain and
stover components in corn silage. Applied Biochemistry and Biotechnology. 113: 41-54.
                                                                                             117

Sheehan, J., A. Aden, K. Paustian, K. Killian, J. Brenner, M. Walsh, and R. Nelson. 2004.
Energy and environmental aspects of using corn stover for fuel ethanol. Journal of Industrial
Ecology. 7(3-4): 117-146.

Shinners, K. J., B. N. Binversie, and P. Savoie. 2003. Whole-plant corn harvesting for
biomass: comparison of single-pass and multi-pass harvest systems. ASAE Paper No.
036089. St. Joseph, Mich.: ASAE

Shinners, K. J.,B. N. Binversie, R. E. Muck, and P. J. Weimer. 2007a. Comparison of wet and
dry corn stover harvest and storage. Biomass and Bioenergy. 31(4): 211-221.

Shinners, K. J., G. S. Adsit, B. N. Binversie, M. F. Digman, R. E. Muck, and P. J. Weimer.
2007b. Single-pass split-stream harvest of corn grain and stover. Transactions of the ASABE.
50(2): 355-363.

Shinners, K. J., B. N. Binversie. 2007c. Fractional yield and moisture of corn stover biomass
produced in the northern US Corn Belt. Biomass and Bioenergy. 31(8): 576-584.

Shinners, K. J., G. C. Boettcher, D. S. Hoffman, J. T. Munk, R. E. Muck, and P. J. Weimer.
2009. Single-Pass Harvest of Corn Grain and Stover: Performance of Three Harvester
Configurations. Transaction of the ASABE. 52(1): 51-60.

Shinners, K. J., R. G. Bennet, and D. S. Hoffman. 2011a. Single- and two-pass harvest of corn
stover. Transactions of the ASABE, In Review.

Shinners, K. J., A. D. Wepner, R. E. Muck, and P. J. Weimer. 2011b. Aerobic and anaerobic
storage of single-pass chopped corn stover. BioEnergy Research. 4(1): 61-68.

Singer, J. W., S. M. Nusser, and C. J. Alf. 2007. Are cover crops being used in U.S. Corn Belt?
Journal of Soil Water Conservation. 62: 353-358.

Sokhansanj, S. and A. F. Turhollow. 2002. Baseline cost for corn stover collection. Applied
Engineering in Agriculture. 18(5): 525-530.

Sokhansanj, S., A. Kumar, and A. F. Turhollow. 2006. Development and implementation of
integrated biomass supply analysis and logistics model (IBSAL). Biomass and Bioenergy.
30(10): 838-847.

USDA-NASS. (2010). Agricultural statistics database; National crop statistics. United States
Department of Agriculture National Agricultual Statistics Service. Available at:
http://www.nass.usda.gov/index.asp. Accessed July 2010.
                                                                                         118

USDA-NASS. (2011). Wisconsin Custom Rate Guide 2010. Available at:
http://www.nass.usda.gov/Statistics_by_State/Wisconsin/Publications/custom_rates_2010.
pdf. Accessed May 2011.

Wright, C. T., P. A. Pryfogle, N. A. Stevens, R. J. Hess, and C. W. Radtke. 2006. Value of
Distributed Preprocessing of Biomass Feedstocks to a Bioenergy Industry. ASAE Paper No.
066151. St. Joseph, Mich.: ASAE.
                                                                                               119

                     Chapter 4 – Discussion and Conclusions
4.1    Discussion
       The economic analysis done on the whole-plant system suggests significant cost

savings can be achieved at storage losses of 5%, through elimination of grain drying,

reduced harvest costs, reduced transportation and handling costs through the use of a

flowable high density material, and reduced storage costs. The cost savings overshadow

the grain and stover losses associated with storage that are modest, however higher than

dry grain storage.

       An analysis of the whole-plant silage system to evaluate storage loss limitations,

found that 16% grain and stover losses could be incurred before it is the cost equivalent to

the best fractionated system modeled. Another analysis of the whole-plant silage system

using 5% losses for the stover and 1% losses for the grain, as is likely to be achieved on a

farm scale, the stover has a farm gate cost of $8.3 (Mg)-1 and a cost $46.7 (Mg)-1 lower than

the next lowest cost system for a biorefinery.

       The system is a very robust system, in that it can begin harvest earlier than is

possible with a combine and continue until the end of dry grain harvest, with no limitation

on field drying of stover. The harvest season of 2009 was an example of this need for

robustness. In the last week of November 2009, NASS reported corn combining progress

for Wisconsin as 67% complete, Iowa at 87%, and Illinois at 72% (NASS, 2010). In 2009

harvest first began in earnest in the first week of November when corn harvest progress

across Iowa, Illinois, and Wisconsin was only 20%. It is doubtful that any quantity of stover

could have been harvested in a fractionated harvest system in that year, and the only
                                                                                              120

opportunity for dry stover harvest would have been a spring harvest. In fact, much stover

harvest occurred in the spring and it was noted for its poor quality and high soil

contamination.

       Harvesting and storing the moist grain may improve its value as feed or feedstock

when using moist grain is acceptable. Single-pass harvesting of the whole-plant results in a

stover that has low ash content, and is a flowable, size-reduced, dense material directly

from the field, with no extraordinary inputs required for densification.




4.2 Conclusions
       Whole-plant silages, both pretreated and control were well preserved during

anaerobic storage. Following storage the stover and grain fractionated at least as well as

dry grain harvest, and both fractions had desirable composition. Pretreatment was

effective at degrading the hemicellulose in the cell wall, with up to 93% removal. It also

significantly enhanced enzymatic degradability and subsequent fermentation to ethanol,

increasing the cellulose conversion efficiency by 19, 11, and 4 percentage units for sulfuric

acid pretreatments of 100, 30, and 10 g(kg DM)-1, respectively.

       Scaling up to the farm level yielded similar results to those found at the pilot-scale

silos. Biological pretreatment was found to be effective at improving the anaerobic

fermentation during storage, reducing the pH and improving its aerobic stability.

       A system analysis of the whole-plant silage system demonstrated it to be a cost-

effective system. Its cost-effectiveness is reliant on the quantity and quality of material

removed from anaerobic storage at its time of use. The storage research conducted
                                                                                          121

demonstrated this to be feasible, even in poorly managed silage conditions. The

combination of these two elements provides a very strong preliminary evaluation of this

system to be a cost-effective and the most labor, capital, and energy efficient corn based

biomass system.




References

USDA-NASS. (2010). Agricultural statistics database; National crop statistics. United States
Department of Agriculture National Agricultural Statistics Service. Available at:
http://www.nass.usda.gov/index.asp. Accessed July 2010.
                                                                                              122

                              Chapter 5 – Future Work

       This work was to be a complete feasibility assessment of the whole-plant silage

system. To this end, the system appears to provide a low cost, consistent material that is

dense, low ash, and size-reduced from the field all the way through the logistics chain. The

grain is the most valuable fraction, and more research is needed to understand the effects

of harvest and storage at various moisture contents, as well as the effect of pretreatments.

       From the grain standpoint, it is expected the acid pretreatment would be the most

acceptable pretreatment. However, much work is being done using alkali pretreatments to

improve the digestibility of the stover. In this research, this appeared to have a negative

impact on the grain, therefore this needs to be further quantified and methods to overcome

obstacles need to be considered.

       A water separation technique was employed, but for on-farm use, a mechanical

system would likely be desired to ensure quality bedding or feed products. The

fractionation method will affect the composition of each fraction. Fractionation equipment

would need to be developed and its efficacy, as quantified by compositional analysis of the

fractions, should be evaluated. Where hydrodynamic fractionation is employed, large

amounts of water containing fermentation products, pretreatment amendments, sugars,

ash, and other material will need to be utilized or cleaned for water recycling through the

plant. The potential benefits or consequences of the “wash water” needs to be further

understood.
                                                                                          123

       The farm-scale experiment was conducted utilized a silo bagger, which is very

effective at packing both wet and dry crops. Demonstration of packing dry silage in a pile or

bunker, and its subsequent storage losses need to be verified.

       Kumar (2005) has suggested pipeline transport and simultaneous saccharification of

corn stover, as opposed to the traditionally considered truck transport. If this scheme ever

were to gain widespread acceptance, the acid pretreated whole-plant silage would fit this

system better than a stover only system. The reasons for this would include: the grain could

also be simultaneously saccharified; the return water pipeline could enjoy larger economies

of scale; and the grain could be much more efficiently transported, thereby improving the

economy of scale for stover transport as well. Additionally the inefficient transport of wet

biomass can be partially overcome, allowing earlier harvest of more digestible and valuable

biomass. Kumar also proposed pumping station spacing that coincidently is similar to that

of the supply hauling distances of the whole-plant system.




References

Kumar, A., J.B. Cameron, and P. C. Flynn. 2005. Pipeline transport and simultaneous
saccharification of corn stover. Bioresource Technology. 96:819-829.
                                                                                             124

Appendix A – Collection Area and Supply Curve


        An alternative method was developed to model the collection area for the corn

stover supply. In general roads are laid out in a grid, particularly at the scale of a biomass

collection point. Therefore biomass supply can best be described as a large set of

concentric diamonds, rather than a circle as is more typically assumed. Each point on the

diamonds are equidistant from the center if you can only travel on the grid. Each of these

concentric diamonds has a diagonal 3.2 km (2 mi.) wider than the previous, radiating out

from the aggregation point. In this way the diagonals are laid out on a one mile by one mile

grid that is typically seen of roadways on a large scale.

        A centrally located processor, indicative of an end-user or final processor is shown in

figure A1. For the case of regional collection points, the aggregation point will likely be off-

center, biased towards the final processor. The economically ideal location between the

center and the extreme point closest to the final processor will be primarily determined by

trucking cost differential of the potentially seasonal incoming versus outgoing products. For

all analysis done in this paper, the processor was assumed to be centrally located within the

area.

        Each area in between the concentric diamonds, henceforth referred to as a tranche,

has the appearance of a squared off annular ring, with the tranches radiating out from the

aggregation point (fig. A1). Rather than using the term collection radius, the more accurate

term collection diagonal is used.
                                                                                               125




         Figure A1. Diagram of collection area with a centralized collection point.


       The area and distance from the processor for each tranche was calculated (Cook,

2011). A table of areas and their respective distances from the processor was developed,

enabling a simulation of the variable costs of hauling for each tranche as a function of its

distance from the processor. Utilizing this variable total cost model for all distances, a

supply curve was applied to each tranche with its respective area and cost to deliver

biomass to the processor.



       Producers will choose to harvest stover residue for many reasons, but it was

assumed the primary driver for producers participating in stover collection was the delta

between their costs and price paid. Therefore it was assumed stover harvest participation

ratio as a gamma distribution approximated by:
                                                                                           126




                                                (   )
                                            (        )
                                        (                )
                                                             [A1]

where:               Pr    stover harvest participation ratio
                      a    explicit cost, $(Mg DM)-1
                     b     estimated stover value out of storage, $(Mg DM)-1
                      x    farm-gate price producer receives, $(Mg DM)-1




         The explicit cost was assumed to be 80% of the estimated process costs through

storage plus the crop cost (for stover the fertilizer value and for grasses the model standing

crop cost). Under the assumption that there is not competing processors requiring stover in

the area, a supply curve based on the price offered for an individual tranche was developed

(fig. A2).
                                                                                                                   127


   Fraction of corn acres that have stover


                                             100%


                                             75%
                harvested .. %




                                             50%


                                             25%


                                              0%
                                                    40   50   60   70   80   90 100 110 120 130 140 150
                                                              Farm-gate price offered .. $/Mg DM




                              Figure A2. Stover supply curve (producer participation ratio) as a
                                         function of stover farm-gate price. Assumes explicit costs
                                         of $50(Mg DM)-1 and estimated value of $80(Mg DM)-1.


                                     The model used an Iowa agricultural area, without major geographic features

altering travel routes and 27% of the total land area was used for corn grain production.

Using the supply curve function with roads running along section lines, an estimate the

amount of stover available as a function of distance from the processor was made. In this

way, the transportation costs would affect stover farm-gate prices. Assuming a fixed price is

paid at the facility gate, the closest fields would receive the highest farm-gate price and the

furthest fields would receive the lowest farm-gate price. The producer participation rate

was applied for each tranche to determine total area and stover mass at all distances.
                                                                                            128

        Currently there is no major market for stover, as such competition for stover was

assumed to not alter the supply curve or limit the collection area. Therefore, the maximum

haul distance was not arbitrarily limited, but determined using the supply curve. The

average, geometric mean, haul distance was determined by dividing the total truck-ton

distance by the total mass of stover delivered. Likewise the average diesel consumption

and labor requirements were determined by dividing the total diesel fuel usage by the total

mass of stover delivered.

        A good case study of this supply model, is the supply curves for scenario 4 (figs. A3

and A4), for a 227 Gg annual supply. The longest haul distance was determined to be 155

km and the mean transport distance of 49.7 km. The midpoint of the supply occurs at 46

km, at this distance half the supply comes from a closer and further distance. Figure A4 is

the result of the supply curve and area lines in figure A3 being multiplied together. The

results here show that at close distance the small amount of area limits the stover supply

and at far distances the economic limitations of transport as seen in the supply curve is the

limiting factor.
                                                                                                                                                         129

                                                         40%                                                              900
Fraction of corn acres that have stover harvested .. %
                                                         35%                                                              800

                                                                                                                          700
                                                         30%




                                                                                                                                 Tranche Area .. sq km
                                                                                                                          600
                                                         25%
                                                                                                                          500
                                                         20%
                                                                                                                          400
                                                         15%
                                                                                                                          300
                                                         10%
                                                                                                                          200

                                                         5%                                                               100

                                                         0%                                                               0
                                                               0   8   16 24 32 40 48 56 64 72 80 88 96 104 112

                                                                            km from Aggregation point

                                                                             Supply Curve    Area (sq km)




                                                  Figure A3. Participation rate (i.e. supply curve) and land area required for a 227
                                                             Gg annual stover requirement as a function of distance from processor
                                                             for scenario 4 (table 3.3).
                                                                                                                        130

  Mass available .. Mg per tranche    6000


                                      5000


                                      4000


                                      3000


                                      2000


                                      1000
                                                      Mid-point of supply        Mean transport distance
                                         0
                                             0    8   16    24   32   40    48    56   64    72   80   88    96   104 112

                                                             Tranche distance from aggregation point .. km




                                     Figure A4.   Mass of material supplied from each tranche for a 227 Gg annual stover
                                                  requirement as a function of distance from processor for scenario 4.




                                      The calculations for each tranche are described by the supply area diagram (fig. A5)

and the excel formula (fig. A6). This figure gives a representation of the aggregation point

in relationship to the collection area surrounding it. The dimension “d” is the total diagonal

distance of the supply area. Dimensions “a” and “b” are the distances to the processor

from the nearest corner; therefore they cannot exceed one-half of dimension “d”. In the

case where no intermediate aggregation centers exist, it is recommended to consider the
                                                                                            131

dimension “d” as without limit (any number that doesn’t artificially limit the collection

area). The dimensions “a” and “b” are equal to half of the value of “d”.

       When intermediate aggregation centers are employed, then the dimension “d” must

be set in accordance to the layout of processors and intermediate aggregation facilities.

Dimensions “a” and “b” would likely be values between zero and one half of value “d”, set

to the individual location’s ideal economic situation. An example of what the supply areas

and aggregation points might look like is shown in figure A7.

       Transportation was found to generally be the single largest cost in biomass. The

novel model described herein was found to be a valuable tool in estimating this distance,

and thereby transportation costs, for any distributed supply systems. Utilizing this model

allowed for accurate assessments of varying biomass supply systems.




Figure A5. Diagram of a supply area and the dimensions used in Figure A7. The
           aggregation point is shown as a star.
                                                                                               132



Tranche Area = 4x – 2 – x*(if(x>a,1,0)) – (if(x-1)>a,x-b-1,0) – ((x + b)*(if(x>b,1,0))) – (if(x-
1)>b, x – b – 1,0) – (x + ((a + b – d – 1)/2))*(if(x>(d - b),1,0) – d*(if(x>(d - a),1,0)

Figure A6. Excel equation that calculates the area of each tranche, where x represents the
           distance from the aggregation point.




Figure A7. Diagram of a collection area with a centralized point for the processor,
           surrounded by intermediate aggregation points that are off-center from their
           geographic centers.
                                                                                                                    133

Appendix B – Additional Model Results
                                  18
                                  16
    Diesel fuel use .. l/ Mg DM



                                  14
                                  12
                                  10
                                   8
                                   6
                                   4
                                   2
                                   0
                                       10%    15%     20%      25%      30%      35%       40%   45%      50%
                                                            Moisture content .. % (w.b.)
                                             Scenario 1         Scenario 2          Scenario 3         Scenario 4
                                             Scenario 5         Scenario 6          Scenario 7         Scenario 8


Figure B1. Sensitivity of diesel fuel required to harvest moisture content for eight different
           harvest scenarios supplying a biorefinery requiring 670 Gg annually.
                                                                                                                         134

                                    120
   Final transport distance .. km

                                    100

                                     80

                                     60

                                     40

                                     20

                                      0
                                          10%     15%        20%     25%      30%      35%     40%    45%       50%
                                                                   Moisture content .. % (w.b.)
                                                Scenario 1           Scenario 2          Scenario 3         Scenario 4
                                                Scenario 5           Scenario 6          Scenario 7         Scenario 8

Figure B2. Sensitivity of final transport distance to harvest moisture content for eight
           different harvest scenarios supplying a biorefinery requiring 670 Gg annually.
                                                                                                                  135

                                  25
    Diesel fuel use .. l/ Mg DM

                                  20


                                  15


                                  10


                                   5


                                   0
                                       10%    20%     30%     40%        50%    60%      70%   80%      90%
                                                            Fraction of stover removed .. %
                                         Scenario 1         Scenario 2            Scenario 3         Scenario 4
                                         Scenario 5         Scenario 6            Scenario 7         Scenario 8




Figure B3. Sensitivity of diesel fuel required to fraction of stover removed for eight
           different harvest scenarios supplying a biorefinery requiring 670 Gg annually.
                                                                                                                         136

                                    160
   Final transport distance .. km
                                    140

                                    120

                                    100

                                     80

                                     60

                                     40

                                     20

                                      0
                                          10%    20%         30%    40%      50%      60%      70%    80%      90%

                                                                   Fraction of stover removed .. %
                                                Scenario 1           Scenario 2          Scenario 3         Scenario 4
                                                Scenario 5           Scenario 6          Scenario 7         Scenario 8




Figure B4. Sensitivity of final transport distance required to fraction of stover removed for
           eight different harvest scenarios supplying a biorefinery requiring 670 Gg
           annually.
                                                                                                                    137


                                 $12

                                 $11

                                 $10
   Feedstock cost .. $/GJ (SE)




                                  $9

                                  $8

                                  $7

                                  $6

                                  $5

                                  $4

                                  $3

                                  $2
                                       10%     15%   20%       25%      30%     35%        40%   45%       50%

                                                            Moisture content .. % (w.b.)

                                       Scenario 1          Scenario 2           Scenario 3             Scenario 4
                                       Scenario 5          Scenario 6           Scenario 7             Scenario 8




Figure B5. Sensitivity of feedstock cost on a steam equivalent energy basis to harvest
           moisture content for eight different harvest scenarios supplying a biorefinery
           requiring 670 Gg annually.


                                 Steam equivalent energy and its formula were developed as a method to

estimate the obtainable thermal energy in biomass. The equation estimates the steam

generation capacity of biomass, by taking into account the intrinsic inefficiency of a

biomass due to its composition and moisture content. Using the steam equivalent

energy basis allows for a more accurate comparison of different biomass species and

harvest parameters, in thermal conversion processes.
                                                                                                                                  138



                                      (    )    ((      ) (     ))     ((           )   (           )     (          )) [B1]

where:                                                        MJ(SE)        steam equivalent heat energy, MJ(kg DM)-1
                                                               HHV          higher heat value, MJ(kg DM)-1
                                                                  H         hydrogen content, g(g)-1
                                                                  M         Moisture content, g H2O(g biomass DM)-1
                                                                  O         oxygen content, g(g )-1




                                    200
                                    180
                                    160
   Final transport distance .. km




                                    140
                                    120
                                    100
                                     80
                                     60
                                     40
                                     20
                                      0
                                          91      544     998        1451       1905    2359       2812       3266      3719
                                                                 Facility mass requirements .. Gg DM annually
                                           Scenario 1            Scenario 2                 Scenario 3               Scenario 4
                                           Scenario 5            Scenario 6                 Scenario 7               Scenario 8


Figure B6. Sensitivity of final transport distance to annual biorefinery mass requirements
           for eight different harvest scenarios supplying a biorefinery requiring 670 Gg
           annually. Scenarios 6 and 1 reside directly behind scenarios 7 and 5,
           respectively.
                                                                                        139




Table B1. Polynomial and linear models of the final transport distance (km) as a function of
annual biorefinery mass requirements (Gg) for eight different harvest scenarios supplying a
                     biorefinery requiring 670 Gg annually (fig. A6).
                                               Scenario No.
                 1         2         3          4         5           6         7         8

                                  Polynomial   model: I + aX + (b/1,000)X2
   Intercept   50.31      62.29      66.77     43.61   50.76       55.12      55.56    18.98
           a   0.219      0.376      0.363     0.209   0.220       0.242      0.243    0.165
           b   -0.178    -0.285     -0.285     -0.171 -0.178      -0.199     -0.199    -0.121
          R2     99        99         99         99      99          99        99        99

                                            Linear model: I + aX
   Intercept 60.51         78.61     82.95     53.14     61.01      66.37      66.85     25.77
            a 0.1215 0.2202         0.2075 0.1181 0.1215           0.1341     0.1341 0.1004
             2
           R      95        96        96        96        95          95        95        97
Note: X is the processor’s annual supply requirement in Tg of corn stover organic matter

				
DOCUMENT INFO
Description: Explores economic and biochemical changes in storage of the new method of whole-plant corn harvest and storage. (combined grain and stover harvest, transport, and storage)