Docstoc

“Off-the-Shelf” Antifreeze Concrete

Document Sample
“Off-the-Shelf” Antifreeze Concrete Powered By Docstoc
					            CBR (CALIFORNIA BEARING RATIO) DATABASE
                  FOR SOIL STRENGTH PREDICTION
                                Sally A. Shoop (presenting)
                   Research Civil Engineer, ERDC-CRREL, Hanover, NH
                              Sally.A.Shoop@usace.army.mil

                                      Peter M. Seman
                   Research Civil Engineer, ERDC-CRREL, Hanover, NH

                                    Deborah Diemand
                 Research Physical Scientist, ERDC-CRREL, Hanover, NH

                                      George L. Mason
                     Research Civil Engineer, ERDC-GSL, Vicksburg, MS

                                     Larry S. Danyluk
                   Research Civil Engineer, ERDC-CRREL, Hanover, NH

                                      Rosa T. Affleck
                   Research Civil Engineer, ERDC-CRREL, Hanover, NH
Abstract
          The Opportune Landing Site (OLS) program objective is to develop technology for locating suitable
landing sites in natural terrain, and the prediction of adequate soil bearing capacity is a critical component. A high-
quality soil strength database was developed to provide data for the development and testing of the soil strength
algorithms. For the OLS program, only California Bearing Ratio (CBR) strength was considered and only high-
quality, true CBR measurements were used (no derived measurements). The database was carefully designed to
ensure that as many soils data as feasible, along with associated parameters, were incorporated. Later expansion, to
include additional measurements related to soil strength, including geographic, geomorphologic, and geospatial
information, can also be accommodated.

          The database is composed of four separate but related datasets: 1) high quality field CBR measurements
taken as part of the United States Air Force (USAF) airfield evaluation program, 2) measurements from an extensive
laboratory program to evaluate the effects of moisture content and percentage of fines for various compaction
efforts, 3) a specific laboratory study of low-density soils that included wetting and drying experiments, and 4) a
dataset of field measurements of CBR taken in conjunction with vehicle trafficability studies on primarily natural,
un-engineered soils. In all, these datasets include roughly 20,000 entries.
Introduction
          The Opportune Landing Site (OLS) program uses remote sensing, weather, and soil state prediction to find
opportune or expedient landing zones (Ryerson et al., 2008). Surface soil strength is a critical design and selection
criterion for an OLS. The strength of the soil on the surface changes as a function of weather, surface conditions,
and physical properties (i.e., in-situ density, permeability, water content). This paper describes a database that is a
compilation of measured soil properties and strength data from four separate sources: two collected from the field
and two produced under controlled conditions in the lab.
          Decisions supporting rapid emplacement of opportune landing zones are based, in part, on the soil
conditions in the area. Strength, physical properties, and type of surface soil are critical information for design and
decision aids. The soil information is often limited or missing and is inferred from historic databases of similar
and/or nearby areas. This paper describes the four main data sources comprising the overall OLS database. The
source material includes high quality soils data from United States Air Force (USAF) airfield evaluation reports and
from nearly 50 years of trafficability and vehicle mobility studies, as well as data rigorously determined in the lab.
Data drawn from reports includes entries for common geotechnical parameters related to soil strength and
geomorphological features associated with soil formation. Roughly 20,000 records were included in the database
from a wide variety of Unified Soil Classification System (USCS) soil types and from a diversity of natural
environments and geographical locations, in addition to controlled material prepared in the lab. The data support
inferring missing data, correlations between physical properties of soils, and statistical assertions of some soils
models. These datasets were used for generating the OLS soil strength algorithms described in Seman (2006 and
2008), Seman et al. (2006), Shoop et al. (2008a), Ryerson et al. (2008), and Grant and Mason (in prep.).
          The four contributing datasets used in the compilation of this database were
     1. Field CBR dataset using high quality field California Bearing Ratio (CBR) measurements taken as part of
          the USAF airfield evaluation program (Seman and Shoop, 2007);
     2. Lab dataset using measurements from an extensive laboratory program to evaluate the effect of percentage
          of fines, compactive effort, and moisture content on CBR. (Danyluk et al., 2008);
     3. Low-density dataset in which laboratory CBR testing was undertaken specifically for low-density soils
          (Danyluk et al., 2008);
     4. Cone Index (CI) dataset, which includes field measurements of CBR taken in conjunction with vehicle
          mobility and trafficability studies. These soils were primarily natural, un-engineered soils, likely to be
          similar to OLS surface materials, and also include corresponding CI measurements enabling CBR
          calibration against CI for use in evaluating CI strength prediction methods. (Diemand et al., 2008, Shoop et
          al., 2008b).
Each of the datasets is described briefly below.

Why CBR?
          The CBR test was originally developed by O.J. Porter for the California Highway Department during the
1920s. It is a load-deformation test performed in the laboratory or the field, whose results are then used with an
empirical design chart to determine the thickness of flexible pavement, base, and other layers for a given vehicle
loading. Though the test originated in California, the California Department of Transportation and most other
highway agencies have since abandoned the CBR method of pavement design. In the 1940s, the US Army Corps of
Engineers (USACE) adopted the CBR method of design for flexible airfield pavements. The USACE and USAF
design practice for surfaced and unsurfaced airfields is still based upon CBR today (US Army, 2001; US Army and
USAF, 1994).
          The CBR determination may be performed either in the laboratory, typically with a recompacted sample, or
in the field. Because of typical logistics and time constraints with the laboratory test, the field CBR is more typically
used by the military for design of contingency roads and airfields. The field CBR test procedure is described in
American Society for Testing and Materials (ASTM) standard D 4429-04 (2004) and Army FM 5-530 (1994, 1987).
          The laboratory CBR test method is defined by ASTM D 1883-05 Standard Test Method for CBR of
Laboratory Compacted Soils (2005). The laboratory CBR test is performed by measuring the penetration resistance
of a 1.954-inch- (in.) diameter (3 in.2 end area) cylindrical steel piston advanced into a soil sample at a rate of 0.05
in. (1.27 mm) per minute. The soil sample is contained in a standard mold, and a surcharge ring weighing ten
pounds is placed on the top of the sample to provide containment for the material.
         The reaction force, in pounds per square inch (psi), is measured at increments of 0.025 in. (0.64 mm) until a
total penetration of 0.500 in. (12.70 mm) is reached. To determine the CBR value, the reaction force measured at
0.100-in. penetration is compared to a standardized value of 1,000 psi. This represents the resistance of a high-
quality, well-graded crushed limestone gravel with ¾-in. maximum aggregate-sized particles. The value of the force
measured in the test is divided by the standardized value (1,000 psi), and then multiplied by 100, to yield an index
value. This value is reported as the CBR of the soil, in percent.

Objectives of the Database
From the beginning, it was apparent that the database would need to meet several unique requirements to be suitable
for generating useful relationships among CBR and other fundamental material properties of soil. The constraints
that guided the search for data included the following goals and motivations (from Seman and Shoop, 2007):
     1. Attempt to incorporate as many of the 26 Unified Soil Classification System (USCS) soil types into the
         database as possible. Because they are based on separating different regimes of engineering behavior in
         soil, a diversity of USCS classes should expose machine learning methods to all the mechanisms that drive
         soil strength.
     2. Ensure that the database is representative of the relative prevalence of the USCS soil types worldwide. In
         effect, the data should reflect to some degree how likely we are to encounter each of the different soil types
         in practice and encompass the larger variety that can be present in some of the more common soil types.
     3. Focus specifically on geotechnical parameters, especially those typically used to characterize engineering
         behavior of soils in the civil engineering community.
     4. Concentrate on records that contain actual CBR measurements, not merely another soil strength index or
         parameter that can be correlated only to CBR.
     5. Make sure that the data encompass the range of conditions that we would expect to find in naturally
         deposited soils. In this respect, care must be taken to ensure that the acceptance criteria placed on
         construction materials do not skew the database. For example, standardized laboratory tests limited to high
         quality material for airfields and pavement applications could reflect higher densities, lower fines contents,
         and lower moisture contents.
     6. Incorporate as much geographic, geologic, environmental, and depositional diversity as possible. In this
         manner, there is some attempt at trying to reflect the wide variety of unique conditions under which natural
         soils can form.
     7. Bring together a consistent and well-documented dataset. The use of standardized test methods is critical
         for high confidence. Ensuring that individual data records are tied to their original sources can be useful in
         many respects. A peculiar soil can be isolated and dealt with separately if necessary. Further information
         may be collected from documented sources to support future efforts. Inferences due to test locations or
         seasonal variation might also be possible.
These principles formed the basis for evaluating prospective sources of data for the OLS soil strength prediction
study and the design of the database.

Selection of Data Fields
          For each data entry more than 50 fields store information about data identification, reference source
documentation, sample site description, soil classification, physical property data, strength index testing (both
laboratory and field), particle sizes and shapes, and remarks (see Table 1). The definition and contents for each of
these fields is described in further detail in Seman and Shoop (2007). Features were chosen by consulting with a
group of subject matter experts to determine a broad range of data types that may either have a qualitative
relationship to soil strength or allow inferences to be made about soil conditions.
          Even though many were not filled either at all or to a significant degree, this large number of fields was
useful in providing a comprehensive scheme for all data types that might be encountered in any of the literature
sources, flexibility for further data collection in the future, crossover with other databases for possible merging at a
later date, and expansion for future inclusion of geospatial, geomorphological, and soil formation process
contributions to soil strength. All contributing datasets to the database are given in the same format.

Field CBR Dataset
        The initial field CBR dataset was completed using ERDC Joint Rapid Airfield Construction (JRAC) and
C17 unsurfaced airfield flight testing data. This was evaluated for distribution of soil types and properties and used
to guide additional development of the database. To fill in gaps in the JRAC and C17 test soil types we mined USAF
evaluation reports for high quality CBR field data. The evaluation reports were prioritized for dataset entry to fill in
gaps in the dataset, particularly with regard to soils that are
     1. Important for potential OLS sites (based on experience on expedient and other runways, and
          geomorphology of soils typically found in flat areas).
     2. Soils that are most common globally.
     3. Soils that are difficult to test in the laboratory.
          Statistical analysis techniques (for predicting CBR) that were evaluated using the initial version of the
dataset included Neural Networks, k-Nearest Neighbor, recursive partitioning, and various regressions. These
techniques were further pursued and validated with the final CBR database. The field CBR dataset is fully described
and statistically summarized in Seman and Shoop (2007).
          A total of 4,608 records of separate field test conditions was collected from all sources. Before proceeding
with further analysis, 16 of the records were not used because they were either stabilized with cement (10 records)
or had compaction energy (CE 26, six records) that differed from all other records (CE 55). Consequently, all
descriptions of the dataset that follow are for the remaining 4,592 records.
          Approximately one-third of all records (1,580) contained information regarding CBR. The remaining two-
thirds were collected because these records provided useful soil condition information for determining relationships
among the non-CBR features and could be valuable in further data mining efforts not focused on CBR. For 47
records, non-numeric CBR data were recorded (e.g., CBR ≥ 100) in order to retain full information. Also, these
could be used for models involving classification or probability distribution. However, most of the records (1,533)
containing CBR information had an integer value for the strength index.
          The data collected for the field CBR dataset came from 46 separate test sites. These sites include 34 from
within the continental United States, seven located around the Pacific Ocean, and five from in, or near, Europe. The
geographical distribution of these sites, shown in Figure 1, represents a variety of locations around the world. They
encompass a broad range of geologic and environmental conditions, such as arid deserts, humid tropics, glacial till,
coral islands, alluvial plains, volcanic deposits, dry lakebeds, and frost-active areas. Therefore, they should cover
many of the different combinations of conditions and processes that lead to soil formation.
          Figure 2 shows the percentage distribution of each soil type relative to the total number of records in each
dataset, while the associated values from the literature are an estimated percentage based on overall global land area.
The chart shows that the distribution in the numerical CBR subset tracked the overall dataset quite closely. Some
exceptions to this include a slight increase in the number of gravel soils (GW, GP, GM, GC) and a significant
decline in low plasticity clays (CL) and high plasticity silts (MH) for the CBR records.

Cone Index Dataset
          In addition to the airfield data, off-road vehicle trafficability studies were used in generating a soil strength
data set for the softer soils. Cone Index (CI) is a simple measure of soil bearing strength used for cross-country
vehicle trafficability studies. A large amount of CI data and prediction routines exist for unprepared soils, such as
may be found at OLS sites. Historic CI prediction algorithms have been incorporated into FASST (the Fast All-
season Soil STrength model, Frankenstein and Koenig, 2004), a module of the OLS program, although the basis and
data used for these strength algorithms are not well documented. This effort served to build a high quality database
of CI measurements and document the performance quality of the relationships used to predict CI from soil
properties, specifically soil moisture content, and also to relate CI to CBR. This dataset is particularly useful for
strengths of in-situ surface soils, which are uncommon in the civil engineering literature.
          The CI dataset effort was jointly performed by the Engineering Research and Development Center (ERDC)
Geotechnical and Structures Lab (GSL) and ERDC-CRREL (Cold Regions Research and Engineering Laboratory),
with leveraging and funding from the Army JRAC and Battlefield Terrain Reasoning Awareness (BTRA) programs.
The CI dataset has over 14,000 entries, the contents of which are summarized in Diemand et al. (2008). The CI
dataset includes over 500 entries that have both a CI and a CBR value enabling the generation of soil-specific
correlations between CI and CBR. These are documented in Shoop et al. (2008b).
          Most of the entries in the dataset come from testing locations in either the United States or Costa Rica.
Testing conducted at locations all across the United States accounted for 5,682, or 39%, of the entries, and testing
conducted in Costa Rica accounted for 8,861, or 61% of all dataset entries. Of the few remaining entries, 13 came
from Panama (0.1%), 55 from Puerto Rico (0.4%), 33 from Thailand (0.2%), and 24 from Hawaii (0.1%). The
geographic distribution of these locations is shown in Figure 3.
          One of the unique contributions of this dataset is its representation of a wide variety of in-situ soils. The
ranges of values for the CI dataset by soil type, for the moisture content, dry density, CBR, Trafficability CI,
Remolding Index, and Atterberg limits are given in Diemand et al. (2008), along with histograms showing the
distribution of other properties given in the dataset.
          Of the 14,645 total entries, there were 4,068 noted as plastic soils. The soil types having plasticity
characteristics include CH, CL, CL-ML, MH, ML, OH, OL, Pt, SC, SM, and SM-SC.
          Of particular interest for the OLS program is the comparison of the physical properties of the CI dataset,
which represents natural sites, to those of the field CBR dataset using data from airfield evaluations, which includes
base, subbase, and subgrade soils data. Table 2 lists the CBR values for the two datasets, as well as the expected
CBR values published by Fang (1991) for USCS categories. Notably, the CI dataset values occur near the lower end
of the expected CBR range, with field CBR data comprising the higher end. Thus, these two datasets form
complementary data sources for the wide range of site conditions that could be expected for an OLS.

Lab CBR Dataset and Low-Density CBR Dataset
          In order to complement the CBR data collected in the field and compiled in the datasets discussed in
previous sections, this lab study focuses on the determination of controlled laboratory CBR measurements.
Laboratory CBR tests were performed in the CRREL Soil Laboratory on soil samples obtained at the OLS
assessment sites in California, Texas, and Indiana, and on specially designed combination soil types. In addition to
increasing the amount of high quality data in the database, this study also considers the effect of specific factors on
soil strength: compaction effort, moisture content, and percentage of fines. A brief summary of each of these studies
is given below; a full discussion of these methods and the associated CBR results are discussed in Danyluk et al.
(2008).

          Percentage of Fines
          The first aspect of the laboratory effort was to look at the effect of the percentage of fines on soil strength
and density. As prediction and modeling efforts for soil strength parameters have continued under the OLS program,
consistent values of CBR, and consistent correlations of CBR and CI, or CBR with moisture content, for sandy soils
have been difficult to obtain. The focus of this work was a detailed analysis of the effect of the percent of high- and
low-plasticity fine-grained soil particles on CBR values. A well-graded (SW) and poorly graded (SP) sand were
used as the base material. A clay (CL) and silt (CL-CH) were added to the sand in varying proportions.
          This study suggests that the proportion of fines in a soil does in fact influence its strength, but that the
relationship is complex. For example, in one test series, when a CL soil was added to a SW soil, the additional fines
increased the CBR if the moisture content was fairly high. This was also true for the combination of SW and CH
soils. There was also a clear progression of optimum moisture content increasing and density decreasing with
additional fines once a threshold of 30% fines was exceeded.
          For an SP soil, adding additional percentages of CH soil resulted in higher optimum moisture contents and
lowered densities as fines content increased, except for the 5% by weight of CH soil, which had an increase in
density. The 5% CH mixture had the highest CBR values, but at lower moisture content. Mixtures with a higher
percentage of fines resulted in reduced CBR values at higher moisture contents.
          The combination of SP with CL soils had results similar to the SP and CH soil combinations in that initially
low contents (less than 10%) of CL additive increase density and CBR, followed by a decrease in these values as the
fines content is increased. However, the CBR difference is less pronounced using the CH additive rather than the
CL.
          It appears from this limited study that when the fines content is increased relative to a base granular soil,
i.e., going from an SP or SM to an SC or CL, there is an initial gain in density and CBR values (with 10% fines)
followed by a decrease in CBR and densities as fines content increases. Generally, at higher moisture contents, CBR
values can be expected to increase as fine content increases.

          Compaction Effort and Density
          Since the OLS program is focusing on using natural unprepared sites for landing zones, a repeatable
method of reproducing these in-situ conditions and CBR values in the laboratory was highly desirable. In-situ soils
typically have much lower density than soils prepared in the laboratory for Civil Engineering methods, or
mechanically compacted to provide a traffic surface. In order for the OLS program to select usable sites for aircraft
operations, it must also be able to eliminate sites with in-situ soils of insufficient density and strength to support
aircraft operations. Therefore, the OLS database must also be populated with soils at a range of densities typical of
undisturbed natural conditions. This aspect of the effort included applying new methods to reliably and repeatedly
prepare low-density samples.
         Standard compaction methods were modified by varying hammer weight, drop height, number of blows per
layer, and number of layers. The reduced compaction effort (ft-lb/ft3) resulted in lower densities while still retaining
a uniformly compacted soil specimen. Note that the low density compactive energies we used are less than 1/4 of
those used in even the lowest standard compactive effort within the ASTM test methods. CBR samples were
prepared using both standard and low-density compaction methods. Very low-density samples were prepared by
pouring relatively dry soils, moisture contents of 2 to 3% of dry soil weight, into the compaction mold from a height
of seven inches. The poured samples were not compacted in any way.
         An illustration of the influence of compactive effort and the impact of density on CBR for an SW sand with
30% fines added, at different compaction efforts, is given in Figures 4 and 5. By reducing the compaction effort in
the laboratory samples we could more closely match the density and moisture regimes found in the field conditions.
         Results from this study showed that for these relatively low-density samples that were used to simulate
naturally occurring soils, CBR increased with increased compaction effort. This would be expected, as it is logical
that a denser sample of a given soil would have a higher bearing capacity. The data also show that the optimum
moisture contents for the samples tend to decrease slightly with increased compaction effort, particularly for the
higher compaction efforts. The lower compaction effort soils have flatter, less defined curves that might be expected
with soils of low densities. When the soil particles are so loosely packed, void ratios and associated moisture
contents can vary widely for the same density.

          Soil Moisture
          Another aspect of the work was to simulate variation in moisture content, as it is known that in-situ strength
varies significantly with moisture content. A potential landing site that may be acceptable for use after long periods
of dry weather can quickly become unsuitable after precipitation. The presence of groundwater can also contribute
to decreased soil strength, and was also examined in this effort.
          The effect of soil wetting and drying was investigated using the very low-density samples. Strong effects
from soil wetting and drying cycles on soil strength were observed during the OLS field testing, and laboratory tests
were designed for a controlled study on the impacts of wetting and drying on CBR. This included simulated rain
events and controlled drying, soaking, and settling as described in Danyluk et al. (2008). Samples were divided into
four groups: 1) dry, 2) rain, 3) rain and dry, and 4) water table. CBR was then determined for each of the groupings.
          For a given soil, it was evident that the CBR may vary considerably for a specific moisture content
depending on the density of the soil. However, in general, it appears that no matter what compaction effort is used,
once the moisture content reaches a range between 12 and 15%, CBR value is greatly reduced or nonexistent. The
value of the moisture content for this decrease varies with soil, but in general, this reaction can be seen for all the
soils tested.

Comparison of Dataset Values
          The four datasets allow a comparison of the values from the different collection sources. Of particular
interest is the comparison of soil properties in the dataset consisting of airfield evaluations (field CBR dataset) with
the data from the trafficability evaluations (CI dataset). These provide some insight into the differences in a soil used
in airfields base, subbase or subgrade, which are likely to be prepared and compacted for structural integrity of the
airfield, with soils occurring naturally on the surface that have not been prepared or engineered. The latter is likely
to be a major contributor to OLS soil conditions. A comparison of the dry density, moisture content, and CBR for
each of the datasets, for major soil types, is shown in Figures 6, 7, and 8. Further clarification of these unique
differences between engineering properties of natural soils compared to engineered soil material should be included
in future work.
          Based on the large amount of data in the CBR datasets, default values of characteristic soil properties could
be generated based on USCS soil class (or other variables). Default values of soil properties needed for the soil
strength prediction algorithms (Ryerson et al., 2008) were developed for occasions when these values were not
available and these are given in Table 3.

Summary
         To generate new methods and algorithms for predicting soil strength, a high quality database of soil
strength and associated parameters was developed. Only true CBR measurements were considered (no derived
measurements). CBR strength data were targeted in order to be applicable to current airfield design methodology,
experience, and historical data. The database was carefully designed to ensure inclusion and expansion to additional
measurements related to soil strength, including geographic and geomorphologic, and geospatial information.
          The database is composed of four separate but complementary datasets: 1) high quality field CBR
measurements taken as part of the USAF airfield evaluation program, 2) measurements from an extensive laboratory
program to evaluate the effects of compaction effort, moisture content, and percentage of fines on CBR, 3) a specific
effort for low-density soils, including wetting and drying, and 4) a dataset of field measurements of CBR taken in
conjunction with vehicle mobility and trafficability studies. This last section of the database includes primarily
natural, un-engineered soils, likely to be similar to OLS surface materials, and also included corresponding CI
measurements enabling calibrating CBR against CI for use in evaluating CI strength prediction methods.
          In all, these datasets include roughly 20,000 entries. The database has been used to generate algorithms to
predict CBR using standard statistical techniques (Ryerson et al., 2008) as well as machine learning techniques
(Seman, 2006; Seman et al., 2007; and Seman, 2008) and physics-based methods (Grant and Mason, in prep). The
database has also been used to generate relationships between CI and CBR (Shoop et al., 2008b).

Acknowledgements
          The authors thank the following individuals and organizations for their assistance through the execution of
the OLS program: Mr. D. Biehle and associates at SEPAC; R. Almassy, P. Blake and C. Hines of Boeing; J.
McDowell, K. Eizenga, Captain J. Rufa, J. Johnson, and R. Haren of AFRL; C. Ventresca and R. McCarty of
Syngenics; D. and D. Ford, North Vernon, IN; R. Curry of N. Vernon Municipal Airport; C. Harig of Fort Bliss; Mr.
C. David at El Centro NAF; D. Walker, G. Machovina, M. Huffman, Maj Dumon, S. King, and M. Elrod at AMC;
F. Scott, C. Scott, and W. Wieder of Science and Technology Corp; Captain Kost and Captain Roope of the Air
Force Academy; R. and M. Rollings of Rollings Consulting; G. Mason, R. Peterson, C. Carter, L. Dunbar, and G.
Brandon of ERDC-GSL; I. Maldonado and B. Cancel Calderon of the University of Puerto Rico Mayagüez; and K.
Laskey of George Mason University. We also thank the following ERDC-CRREL employees who participated in
the program: C. Ryerson, S. Shoop, G. Koenig, P. Seman, L. Barna, R. Affleck, L. Danyluck, J. Quimby, C. Smith,
C. Berini, G. Durrell, J. Buska, S. Frankenstein, V. Keating, M. Beck, A. Maynard, C. Grant, R. Melendy, J.
Berman, J. Richter-Menge, J. Hardy, S. Barrett, K. Bjella, K. Claffey, R. Davis, D. Diemand, G. Gooch, E. Ochs, S.
Orchino, and B. Tracy. Funding was provided by the U.S. Transportation Command through Air Mobility
Command, and managed by the Air Force Research Laboratory Air Vehicles Directorate at Wright-Patterson Air
Force Base.
          Special thanks are due to the Air Force Civil Engineer Support Agency’s Airfield Pavement Evaluation
Team for its support of Air Force Research Laboratory’s Opportune Landing Site System. Captain Michelle
Harwood, Technical Sergeant Jason Rusticelli, Major John Lantz, Staff Sergeant Heidi Hunter, Technical Sergeant
Jacob Sanabia, Major Erik Sell, Mr. Richard Smith, and Mr. Jon Reed all were instrumental in the successful
demonstration of the OLS system.
References
American Society for Testing and Materials, 2004, D4429-04, Test Method CBR (California Bearing Ratio) of Soils
   in Place. In Annual Book of ASTM Standards, Volume 04.08. West Conshohocken, Pennsylvania: ASTM.
American Society for Testing and Materials, 2005, D1883-05, Standard Test Method for CBR (California Bearing
   Ratio) of Laboratory-Compacted Soils. In Annual Book of ASTM Standards, Volume 05.08. West Conshohocken,
   Pennsylvania: ASTM.
Danyluk, L., S.A. Shoop, and W.L. Wieder, 2008, Opportune Landing Site CBR and Low Density CBR Laboratory
   Database. ERDC/CRREL Technical Report 07-xxx. U.S. Army Cold Regions Research and Engineering
   Laboratory, Engineer Research and Development Center, Hanover, NH, USA.
Diemand, D., S.A. Shoop, G. Mason, and G. Brandon, 2008, Trafficability Cone Index Database. ERDC Technical
   Report TR 08-2, U.S. Army Cold Regions Research and Engineering Laboratory, Engineer Research and
   Development Center. Hanover, NH, USA.
Fang, H-Y. 1991. Foundation Engineering Handbook. Second Edition. Van Nostrand Rheinhold, New York.
Frankenstein, S., and G.G. Koenig, 2004, Fast All-season Soil Strength (FASST). ERDC/CRREL SR-04-1, U.S.
   Army Engineer Research and Development Center, Hanover, NH, USA.
Grant, S., and G. Mason, in prep. A simple cavity-expansion expression to describe cone indices and rating cone
   indices for unsaturated soils.
Robinson, J.H., and C.P. Rabalais, 1993, Performance of the combat engineer vehicle with mineplow operating
   worldwide and in theaters of operation. Technical Report GL-93-23. Waterways Experiment Station, Vicksburg,
   MS, USA.
Ryerson, C., S. Shoop, and G. Koenig, 2008, Opportune Landing Site Program: Final Report. ERDC-CRREL
   Report, U.S. Army Cold Regions Research and Engineering Laboratory, Engineer Research and Development
   Center. Hanover, NH, USA.
Semen, P.M., 2006, A generalized approach to soil strength prediction with machine learning methods.
   ERDC/CRREL Technical Report 06-15, U.S. Army Cold Regions Research and Engineering Laboratory,
   Engineer Research and Development Center, Hanover, NH, USA.
Seman, P., 2008, Machine Learning Approaches to CBR Prediction for Unsurfaced Airfields, Transportation
   Systems Workshop, April 2008, Phoenix, AZ, USA.
Seman, P.M., and S.A. Shoop, 2007, Worldwide Database of In-Situ California Bearing Ratio Measurements.
   ERDC/CRREL Technical Report 07-21, Engineer Research and Development Center, Cold Regions Research
   and Engineering Laboratory, Hanover, NH, USA.
Seman, P.M., S.A. Shoop, S.P. McGrath, and R.S. Rolling,. 2006, Soil Strength Prediction with k-Nearest Neighbor
   Method. In Proceedings of the 59th Canadian Geotechnical Conference, Vancouver. 1–4 October 2006.
   Alliston, Ontario: Canadian Geotechnical Society.
Shoop, S.A., C. Ryerson, R. Affleck, J. Buska, J. Kost, 2008a, Predicting Soil Strength for Opportune Landing Sites,
   Transportation Systems Workshop, April 2008, Phoenix, AZ, USA.
Shoop, S., D. Diemand, and W. Wieder, 2008b, Predicting California Bearing Ratio from Trafficability Cone Index
   Values. U.S. Army ERDC/CRREL Technical Report (in press).
US Army Corps of Engineers, 2001, Pavement Design for Airfields, Unified Facilities Criteria Handbook, UFC 3-
   260-02, U.S. Army Corps of Engineers.
US Army and Air Force, 1994, Planning and Design of Roads, Airfields, and Heliports in the Theater of Operations:
   Airfield and Heliport Design, Vol. I and Vol. II. US Army FM 5-530-00-1 and US Army FM 5-530-00- 2 or Air
   Force AFJP 32-8013, Vol I and Vol. II, 1994.
US Army, Air Force and Navy, 1987, Materials Testing. US Army Field Manual 5-530/Air Force AFM 89-3/Navy
   NAVFAC MO-330. Washington, DC: DoD, 17 August 1987.
List of Tables and Figures

Table 1. Fields in Opportune Landing System California Bearing Ratio Database.
Table 2. Average CBR values for the four OLS CBR datasets by soil type, compared to published values from Fang
    (1991).
Table 3. Average values of some physical soil properties from the OLS CBR databases.
Figure 1. Geographic distribution of CBR data test sites in the field CBR database for the continental United States
    (top), the Pacific (center), and Europe (bottom). (Image courtesy of Google Earth mapping service.)
Figure 2. Distribution of records in Field CBR dataset by USCS soil type compared to worldwide occurrence of soil
    type according to Robinson and Rabalais (1993).
Figure 3. CI database sample locations for North America and Hawaii (top) and Costa Rica and Thailand (bottom).
    (Image courtesy of Google Earth mapping service.)
Figure 4. The effect of compaction effort on CBR and soil dry density for the same moisture content.
Figure 5. Exponential trend line fit for effect of density on CBR for 9% moisture content (by dry soil weight).
Figure 6. Comparison of average dry density among the four datasets by soil type. LAB = Laboratory fines testing,
    LD = Low Density laboratory tests, CI = Cone Index dataset, Field = USAF airfield dataset.
Figure 7. Comparison of in-situ or as-tested moisture content among the four datasets by soil type. LAB =
    Laboratory fines testing, LD = Low Density laboratory tests, CI = Cone Index dataset, Field = USAF airfield
    dataset.
Figure 8. Average CBR values from non-engineered soils from the trafficability studies compared to airfield soils,
    by USCS soil type. LAB = Laboratory fines testing, LD = Low Density laboratory tests, CI = Cone Index
    dataset, Field = USAF airfield dataset.
     Table 1. Fields in Opportune Landing System California Bearing Ratio Database.
OLS Data Point #                            Soaked CBR (laboratory)
JRAC Soil #                                 Moisture Content as Tested (weight %)
Test or Sample Date                         Moisture Content as Tested (volumetric %)
Report #                                    Trafficability Cone Index (CI)
Report Date                                 Remolding Index
Report Title                                DCP Index (dynamic cone penetrometer)
Country Code (ISO-3166§)                    Field CBR
Location                                    Field Dry Density
Test Station                                Field Wet Density
Latitude                                    ¾-inch Sieve, Maximum Percent Passing
Longitude                                   ¾-inch Sieve, Minimum Percent Passing
Layer                                       ⅜-inch Sieve, Maximum Percent Passing
Landform                                    ⅜-inch Sieve, Minimum Percent Passing
Lithology of Parent Material                #4 Sieve, Maximum Percent Passing
Deposition Type                             #4 Sieve, Minimum Percent Passing
Depth to Water Table                        #10 Sieve, Maximum Percent Passing
Soil Type, USCS                             #10 Sieve, Minimum Percent Passing
Alternate Soil Type                         #40 Sieve, Maximum Percent Passing
Alternate Soil System                       #40 Sieve, Minimum Percent Passing
Soil Description                            #100 Sieve, Maximum Percent Passing
Clay Mineralogy                             #100 Sieve, Minimum Percent Passing
Specific Gravity                            #200 Sieve, Maximum Percent Passing
Sample Depth Below Grade                    #200 Sieve, Minimum Percent Passing
Plastic or Non-Plastic                      0.005 mm, Maximum Percent Passing
LL (liquid limit)                           0.005 mm, Minimum Percent Passing
PL (plastic limit)                          0.001 mm, Maximum Percent Passing
PI (plasticity index)                       0.001 mm, Minimum Percent Passing
Compactive Effort                           Roundness, Gravel
Molding Moisture Content                    Roundness, Sand
Dry Density (laboratory)                    Sphericity, Gravel
Optimum Moisture Content and Max. Density   Sphericity, Sand
Unsoaked CBR (laboratory)                   Remarks
  Table 2. Average CBR values for the four OLS CBR datasets by soil type, compared to published values from
            Fang (1991).

                  CI dataset         Field CBR       Lab dataset*      Low-density     Fang (1991)
    USCS
     Soil      No. of    CBR      No. of    CBR    No. of    CBR     No. of   CBR
Classification data     range      data    range    data    range     data   range     CBR range
CH                170 0.09–11.7       79      4–25                                              3–5
CL                174    0.1–16.7    149      2–48     65   0.3–94.9     42    0.1–9.2         5–15
CL-ML                                                  30   0.5–98.7     16    0.0–8.1          N/A
GC                                    97     5–109                                            20–40
GM                                    57    31–146                                          20–80+
GP                 25    0.1–11.3                                                             25–60
GW                                    17     19–68                                          60–80+
MH                 95    0.04–9.2                                                               4–8
ML                 44    0.1–10.4     42      6–37                                             5–15
SC                                   156     3–158    118   0.5–98.6     17  0.2–12.2         10–20
SC-SM                                                  60   0.4–43.3     22  0.0–34.0           N/A
SM                 48       0.1–8     70      7–73     77   0.3–71.3     32  0.1–20.1         10–40
SP-SM               4        4–25                      24 2.4–105.3                             N/A
SP                                                     20   9.2–59.8                          10–25
SW                                    24      6–60     20   3.2–20.1                          20–40
SW-SM                                                  42   3.5–77.9     16  0.0–11.0           N/A
          * Both soaked and unsoaked CBR were determined in this lab study. Both were included in the range of
          CBR values indicated.
        Table 3. Average values of some physical soil properties from the CBR database.
                                                            %           % Passing % Passing
 USCS Soil         Liquid       Plastic     Plasticity
                                                         Retained        #4 Sieve   #200
Classification     Limit         Limit        Index
                                                         #4 Sieve                   Sieve
GW                  N/A          N/A           N/A          61a            39a            4a
GP                  N/A          N/A           N/A          67a            33a            3a
GM                   36a          27a           9a          50a            50a        20a
                        a            a              a            a              a
GC                   27           14           13           56             44         17a
SW                  N/A          N/A           N/A          15a            85a            3a
SP                  N/A          N/A           N/A          27a            73a            3a
SM                  32.5c        27.5c          5c          14a            86a        22a
SC                  27.5c         16c         11.5c         20a            80a        30a
ML                   36c          27c           9c           3a            97a        70a
CL                   35c         19.5c        15.5c          7a            93a        65a
OL                   45b          42b           3b           5d            95d       67.5d
MH                   71c         42.5c        28.5c          5a            95a        80a
CH                  61.5c         24c         37.5c          2a            98a        73a
OH                  105b          71b          34b          3.5e          96.5e      76.5e
Pt                  207b         166b          41b        No data        No data    No data
                        c            c           c            a                 a
CL-ML                23           17            6            4             96         65a
GW-GM                23a          17 a         6a            4a            96 a       65 a
GW-GC                NP           NP           NP           57 a           43 a           6a
GP-GM                28 f         16 f         12 f         54 f           46 f           7f
GP-GC                NP           NP           NP           60 a           40 a           6a
GC-GM                35 f         22 f         13 f         72 f           28 f           8f
SW-SM                21 f         17 f          4f          33 f           67 f       30 f
SW-SC                NP           NP           NP           25 a           75 a           9a
SP-SM              No data      No data      No data      No data        No data    No data
                                                                  a             a
SP-SC                NP           NP           NP           25             75             9a
SC-SM                24c         18.5c         5.5c         14a            86a        34a
a Values from Field CBR dataset.
b Values from CI dataset.
c Average of values from Field CBR and CI datasets.
d OL grain size is average of ML and CL soils from Field CBR dataset.
e OH grain size is average of MH and CH soils from Field CBR dataset.
f Based on two cases or less in the Field CBR dataset.
g This category is not used in the FASST model.
NP Non-Plastic soils.
Figure 1. Geographic distribution of CBR data test sites in the Field CBR dataset for the continental United
           States (top), the Pacific (center), and Europe (bottom). (Image courtesy of Google Earth mapping
           service.)
                             50%
                                                                 All Records         CBR Records          Worldwide % Area Distribution



                             40%
 Percent of Entire Dataset




                             30%




                             20%




                             10%




                             0%




                                                                                                                                                                                                         & Other*
                                                                                                                                                                                                          Missing
                                                                                                                                                 GP-GC
                                                                                                                                                         GC-GM




                                                                                                                                                                                 SP-SM
                                                                                                                                                                                         SP-SC
                                                                                                                                                                                                 SC-SM
                                                                                                                                        GP-GM
                                             GM
                                                  GC




                                                                 SM
                                                                      SC




                                                                                           MH
                                                                                                CH
                                                                                                     OH
                                                                                                           Pt
                                                                                      OL




                                                                                                                CL-ML
                                        GP




                                                            SP




                                                                           ML
                                                                                CL




                                                                                                                        GW-GM
                                                                                                                                GW-GC




                                                                                                                                                                 SW-SM
                                                                                                                                                                         SW-SC
                                   GW




                                                       SW




                                                                                           USCS Soil Type                                       * Other Worldwide = 1% Tundra


Figure 2. Distribution of records in Field CBR dataset by USCS soil type compared to worldwide occurrence
           of soil type according to Robinson and Rabalais (1993).
Figure 3. CI database sample locations for North America and Hawaii (top) and Costa Rica and Thailand
            (bottom). (Images courtesy of Google Earth mapping service.)
                                        SW+30%CL at 9% Moisture Content

                        40                                                             135.0

                        35                                                             130.0
                        30
                                                                                       125.0




                                                                                               Density (pcf)
                        25
                                                                                       120.0
                  CBR


                        20
                                                                                       115.0
                        15
                                                                                       110.0
                        10                                                   CBR
                        5                                                    Density   105.0

                        0                                                            100.0
                             0   2000   4000   6000    8000    10000 12000 14000 16000
                                          Com paction Effort (ft-lbs/ft 3)



Figure 4. The effect of compaction effort on CBR and soil dry density for the same moisture content.
                                         SW+30%CL, Constant Moisture %


                         40

                         35

                         30          9% Moisture
                                     CBR Trend
                         25
                   CBR




                         20

                         15

                         10

                         5

                         0
                          80.0    90.0      100.0      110.0        120.0   130.0     140.0
                                                    Density (pcf)



Figure 5. Exponential trend line fit for effect of density on CBR for 9% moisture content (by dry soil weight).
                                                             Dry Density VS. Soil Type

                      140




                      120




                      100
Dry Density Average




                      80                                                                                                  CI
                                                                                                                          LAB
                                                                                                                          LD
                      60                                                                                                  Field




                      40




                      20




                       0
                            CL     CL-ML   SC   SC-SM   SM    CH     GC      GM    GP    GW    MH     ML    SP       SW
                                                                      Soil Type


                                 Figure 6. Comparison of average dry density among the four datasets by soil type.
                                                         LAB = Laboratory fines testing
                                                       LD = Low Density laboratory tests
                                                            CI = Cone Index dataset
                                                          Field = USAF airfield dataset
                                                          Gravimetric Moisture Content VS. Soil Type

                           70




                           60




                           50
MC as tested wt% average




                           40                                                                                                    CI
                                                                                                                                 LAB
                                                                                                                                 LD
                           30                                                                                                    Field




                           20




                           10




                           0
                                CL   CL-ML   SC   SC-SM    SM      CH    GC       GM    GP    GW       MH   ML   SP   SW
                                                                           Soil Type


                           Figure 7. Comparison of in-situ or as-tested moisture content among the four datasets by soil type.
                                                            LAB = Laboratory fines testing
                                                           LD = Low Density laboratory tests
                                                                CI = Cone Index dataset
                                                             Field = USAF airfield dataset
                                                        CBR VS. Soil Type

              90


              80


              70


              60
CBR average




              50                                                                                             CI
                                                                                                             LAB
                                                                                                             LD
              40                                                                                             Field


              30


              20


              10


              0
                   CL   CL-ML   SC   SC-SM   SM    CH      GC      GM       GP   GW   MH   ML     SP    SW
                                                            Soil Type




                    Figure 8. Average CBR values from non-engineered soils from the trafficability studies
                                       compared to airfield soils, by USCS soil type.
                                             LAB = Laboratory fines testing
                                            LD = Low Density laboratory tests
                                                 CI = Cone Index dataset
                                              Field = USAF airfield dataset

				
DOCUMENT INFO