Get documents about "A REVIEW OF TOOLS FOR THE MODELLING AND PREDICTION
Document Sample
COOPERATIVE RESEARCH CENTRE FOR COAL IN SUSTAINABLE DEVELOPMENT
Established and supported under the Australian Government’s Cooperative Research Centres Program
A REVIEW OF TOOLS FOR THE MODELLING AND PREDICTION OF
TRACE ELEMENT DEPORTMENT IN COMBUSTION PROCESSES
RESEARCH REPORT 36
Authors:
M. Attalla
C. Chao
P.F. Nelson
CSIRO Energy Technology
March 2003
QCAT Technology Transfer Centre, Technology Court
Pullenvale Qld 4069 AUSTRALIA
Telephone (07) 3871 4400 Facsimile (07) 3871 4444
Email: Administration@ccsd.biz
This page is intentionally left blank
DISTRIBUTION LIST
CCSD
Chairman; Chief Executive Officer; Research Manager, Manager Technology; Files
Industry Participants
Australian Coal Research Limited.....................................................................................Mr Ross McKinnon
BHP Billiton Minerals - Coal .....................................................................Mr Ross Willlims
CNA Resources ..........................................................................................Mr Rod Hall
CS Energy...................................................................................................Dr Chris Spero
Delta Electricity..........................................................................................Mr Steve Saladine
Queensland Natural Resources and Mines .................................................Ms Mary Worthy
Rio Tinto (TRPL) .......................................................................................Mr David Cain
Tarong Energy ............................................................................................Mr Burt Beasley
The Griffin Coal Mining Co Pty Ltd ..........................................................Mr Jim Coleman
Wesfarmers Premier Coal Ltd ....................................................................Mr Peter Ashton
Western Power............................................................................................Mr Keith Kirby
Xstrata Coal Australia Pty Ltd ...................................................................Mr Barry Isherwood
Research Participants
CSIRO …….. .............................................................................................Dr Adrian Williams
Curtin University of Technology................................................................Dr Barney Glover
Macquarie University ................................................................................Prof Peter Bergquist
The University of Newcastle ......................................................................Prof Adrian Page
The University of New South Wales..........................................................Prof David Young
The University of Queensland....................................................................Prof Don McKee
Retiring Participants
ARCO Resources Ltd .................................................................................Mr William Ash
BHP Coal Pty Ltd .......................................................................................Mr Alan Davies
Pacific Power..............................................................................................Dr Allen Lowe
Rio Tinto Pacific Coal ................................................................................Mr Duncan Waters
Stanwell Power ...........................................................................................Mr Des Covey
This page is intentionally left blank
Cooperative Research Centre for
Coal in Sustainable Development
QCAT Technology Transfer Centre
Technology Court
Pullenvale, Qld 4069
Telephone: (07) 3871 4400
Fax: (07) 3871 4444
Feedback Form
To help us improve our service to you may I ask you for five minutes of your time to complete this
questionnaire. Please fax or mail it back to me, or, if you would prefer, give me a call.
ATTENTION MANAGER TECHNOLOGY
FAX NO 07 3871 4444 DATE ………………………………………...
FROM NAME: ………………………………………………….…………………...
COMPANY: .………………………………………………………...……………
REPORT TITLE: A REVIEW OF TOOLS FOR THE MODELLING AND PREDICTION OF TRACE ELEMENT
DEPORTMENT IN COMBUSTION PROCESSES
AUTHORS: M. ATTALLA, C. CHAO, P. F. NELSON
Would you please rate our performance in the following areas by ticking the appropriate box:
The Research Agree Neutral Disagree Don’t
Know
This work has achieved its research objectives
This work has delivered to agreed milestones
This work is relevant to my organisation
Communication of progress has been timely & useful
The Report
The report is well presented
The context of the report is clear
The content of the report is substantial
The report is clearly written and understandable
This product is good value
Any further comments which would assist us in better serving your future requirements?:
…………………………….……………………………………………………………………
…………………………………………………………………………….……………………
Are there any people you would like to add to our distribution list?:
…………………………………………………………………………………………………
…………………………………………………………………………………………………
Thank you for your response
This page is intentionally left blank
TABLE OF CONTENTS
1 INTRODUCTION...................................................................................... 3
2 EXPERIMENTAL EVIDENCE FOR THE NATURE OF TRACE ELEMENT
DEPORTMENT IN PULVERISED FUEL-FIRED POWER STATIONS DURING
COAL COMBUSTION..................................................................................... 4
2.1 INTRODUCTION.................................................................................................................... 4
2.2 FATE OF TRACE ELEMENTS ............................................................................................ 4
2.2.1 General Comments ........................................................................................................... 4
2.2.2 Arsenic.............................................................................................................................. 5
2.2.3 Boron ................................................................................................................................ 6
2.2.4 Beryllium.......................................................................................................................... 6
2.2.5 Cadmium .......................................................................................................................... 6
2.2.6 Chlorine ............................................................................................................................ 6
2.2.7 Cobalt ............................................................................................................................... 6
2.2.8 Chromium......................................................................................................................... 7
2.2.9 Copper .............................................................................................................................. 7
2.2.10 Fluorine............................................................................................................................. 7
2.2.11 Mercury ............................................................................................................................ 7
2.2.12 Manganese........................................................................................................................ 8
2.2.13 Molybdenum..................................................................................................................... 8
2.2.14 Nickel ............................................................................................................................... 8
2.2.15 Lead .................................................................................................................................. 9
2.2.16 Selenium ........................................................................................................................... 9
2.2.17 Zinc................................................................................................................................... 9
2.2.18 Uranium and Thorium ...................................................................................................... 9
2.3 DISCUSSION ......................................................................................................................... 10
3 TOOLS FOR THE MODELLING AND PREDICTION OF TRACE ELEMENT
DEPORTMENT IN COMBUSTION PROCESSES ........................................ 11
3.1 INTRODUCTION.................................................................................................................. 11
3.2 GENERAL MODELLING APPROACHES ....................................................................... 11
3.3 INTEGRATED MODELLING APPROACHES ................................................................ 14
3.3.1 Computational Fluid Dynamics...................................................................................... 24
3.3.2 The Electric Power Research Institute (EPRI), PISCES software.................................. 24
3.3.3 COFERS Co-burning Feed Rate Simulator .................................................................... 25
3.4 GENERIC COAL QUALITY SYSTEMS ........................................................................... 26
3.4.1 The Coal Quality Management System (CQMS) ........................................................... 26
3.5 INFORMATION DATABASES ........................................................................................... 27
3.5.1 CONSOL Inc. Coal Database and Coal Quality/Power Cost Model .............................. 27
3.5.2 British Coal Corporation Emissions Monitoring Database............................................. 28
3.5.3 The United States Geological Survey (USGS) COALQUAL database.......................... 28
3.5.4 Aerometric Information Retrieval System (AIRS) ......................................................... 30
3.5.5 Centre for Air and Toxic Metals (CATM) Database ...................................................... 31
3.6 THERMODYNAMIC DATABASES ................................................................................... 34
3.6.1 GFEDBASE Thermodynamic Database......................................................................... 34
3.6.2 SOLGASMIX, ChemSage and Other Thermochemical Databanks ............................... 35
3.6.3 CSIRO Thermochemistry System (version V) ............................................................... 37
3.6.4 HSC Software and Database........................................................................................... 38
1
3.6.5 MANLABS Thermochemical Database (USA).............................................................. 39
3.6.6 MTDATA Thermochemical Database (UK) .................................................................. 39
3.6.7 THERMO-CALC Thermochemical Database (Sweden)................................................ 40
3.6.8 THERDAS Thermochemical Database (Germany)........................................................ 40
3.6.9 THERMODATA Thermochemical Database (France) .................................................. 40
3.6.10 Lewis Research Centre, NASA (USA)........................................................................... 41
3.7 Overview of Thermodynamic Databases ............................................................................. 41
4 CONCLUSIONS AND RECOMMENDATIONS...................................... 42
5 REFERENCES....................................................................................... 44
APPENDIX 1 ................................................................................................. 47
APPENDIX 2 ................................................................................................. 48
APPENDIX 3 ................................................................................................. 61
2
1 INTRODUCTION
Release of trace elements1 to the environment is an area of increasing concern given
the relatively high toxicity of some of these species. Utilisation of coal in combustion
and gasification processes results in potential releases in a number of ways. The more
volatile may be emitted in the gas phase or enriched on the fine (sub micron)
particulate fraction, and hence escape capture by electrostatic precipitators or bag
filters. Alternatively, trace elements may reside in the fly ash collected by gas
cleaning devices or in the bottom ashes or slags. Their ultimate fate in the latter case
will depend on the utilisation and/or disposal options chosen for the ash or slag, and in
many cases will be determined by the leachability of the trace elements.
In order to predict environmental impacts of trace elements in coal, information on the
contents of trace elements, and how the trace elements partition to the various waste
streams (flue gas, fly ash and bottom ash) is required. The objective of this task is to
develop techniques to estimate the partitioning, based on analyses of the trace element
contents and modes of occurrence in the original coal.
In this report results of a review of existing information in this area is presented. The
review examines the kinetic and thermodynamic approaches to predicting the trace
element speciation and research on ash and slag formation in oxidising and reducing
conditions. The following components were examined and also reviewed:
· the findings of completed Black Coal CRC projects 4.2 (trace element speciation using
selective chemical leaching techniques), 4.4 (chemical kinetic computations of
vaporisation of inorganic species), 4.5 (leaching studies of combustion ashes) and 4.7
(assessment of F*A*C*T for predicting trace element behaviour during PF combustion)
· overseas data and modelling approaches to predicting the fate of trace elements in
combustion systems
1
trace elements of concern include those regulated under the 1990 US Clean Air Act amendments (antimony, arsenic, beryllium,
cadmium, chromium, cobalt, lead, manganese, mercury, nickel and selenium, as well as the radionuclides uranium and thorium
3
2 EXPERIMENTAL EVIDENCE FOR THE NATURE OF TRACE
ELEMENT DEPORTMENT IN PULVERISED FUEL-FIRED
POWER STATIONS DURING COAL COMBUSTION
2.1 INTRODUCTION
This is a review of relevant literature on the behaviour of trace elements during
combustion of coal in pulverised fuel-fired power stations. It follows the extensive
review of published material by Dr E. Jak (University of Queensland) in June, 1999.
The elements targeted are those listed on Australia’s National Pollutant Inventory:
these are arsenic, boron, beryllium, cadmium, chlorine, cobalt, chromium, copper,
fluorine, mercury, manganese, molybdenum, nickel, lead, selenium and zinc. The
radioactive elements thorium and uranium have also be included.
There are a number of key reviews and scientific papers on the fate of trace elements
during, and following, the combustion of coal. IEA Coal Research (Smith, 1987;
Clarke & Sloss, 1992) has produced two comprehensive reviews. There is also a
special issue of Fuel Processing Technology devoted to the transformations of trace
elements in coal fired systems (Benson et al, 1994).
Meij from KEMA (The Netherlands) has published extensively on the transformation
and fate of trace elements in Dutch power stations (Meij, 1999; Meij, 1994; Meij,
1992; Meij et al, 1986; Meij et al, 1984).
The PISCES research program of EPRI has produced a database of trace element
distributions in the coal-fired utilities of the USA (Chow et al, 1994). Part of this
research program is to investigate the applicability of analytical methods.
2.2 FATE OF TRACE ELEMENTS
2.2.1 General Comments
Trace elements have been classified by a number of researchers in classes based on
their apparent volatility during combustion. The volatility is determined empirically
by the comparison of the concentrations of the trace element in the flyash and bottom
ash compared with that in the feed coal. Clarke and Sloss (1992) summarise these as:
· Group 1: low volatility, equally distributed between bottom ash and fly ash
Al, Ca, Ce, Cs, Eu, Fe, Hf, K, La, Mg, Sc, Sm, Si, Sr, Th, Ti
· Group 2: volatile, enriched in the flyash.
Meij (1994) separates the elements of Group 2 into three classes: IIc,
IIb and IIa. The more volatile elements are in IIa, the least in IIc. Meij
bases this separation on the relative concentration of the trace element
in the “flyash” defined by the author as the particulate matter emitted
through the stack (Meij, 1999).
4
IIc: Ba, Cr, Mn, Na, Rb
IIb: Be, Co, Cu, Ni, P, U, V, W
IIa: As, Cd, Ge, Mo, Pb, Sb, Tl, Zn
· Group 3: highly volatile, at least partially present in the gas phase and not
enriched in the flyash
B, Br, Cl, F, Hg, I, S, Se.
Yokoyama et al (1991) used thermodynamic data to similarly classify trace elements
based on the volatility of the element and its compounds, generally oxides (see Table
17, p 36, Clarke and Sloss, 1992).
KEMA (Meij, 1999) has developed a model to estimate the distribution of trace
elements in the various streams of coal-fired power plants. This appears to be based
on factors that have been calculated from the database of studies completed at plants
in the Netherlands. A comparison of the data from Meij and that recently obtained by
CSIRO from studies in Australia is given in Appendix 1.
In the following sections, comments on the transformations and fate of the elements
considered to be of environmental importance (as listed in the NPI, and the
radionuclides) are made. Reference is also made as to the likely form in which each of
the elements are present in coal. The monograph by Swaine (1990) contains a
comprehensive review of the literature pertaining to the occurrence of trace elements
in coal. Finkelman (1994) reports studies into the modes of occurrence and ranks the
confidence (from one to a maximum of ten) in the assignment of the likely mode. The
results of Dale et al (1999), who used selective leach techniques to estimate
mineralogical associations of trace elements in six Australian coals, are also reported;
the associations were grouped into four categories: oxide/carbonate/monosulphide
(HCl soluble), pyrite (HNO3 soluble), silicate (HF soluble) and organic (residual).
2.2.2 Arsenic
Meij (1999) states that in the Netherlands, arsenic is not classified as a Group 3,
highly volatile element and it is generally not found in the flue gases past the
electrostatic precipitator (ESP). It appears that volatile As2O3 is produced during
combustion (see also Clarke and Sloss, 1992) but calcium arsenate is formed if
calcium is present in the mineral matter:
3 CaO + As2O3 + O2 ® Ca3(AsO4)2
Meij cites the work of Gutberlet (1988). Cramer (1986) also reported that both Ca and
Se were less volatile in the presence of CaCO3. He used a tube furnace at
temperatures from 600oC to 1200oC to determine the volatility of a number of
elements. The coal was heated (combusted) in 100% oxygen. How the reported
findings relate to the actual mechanisms occurring in a PF furnace is unclear, but the
effect of the CaCO3 on the behaviour of Ca and Se was marked.
Finkelman (1994) reports that arsenic is most likely (confidence level 8) present in the
pyrite in coal, with a possibly minor amount organically bound. Swaine (1990)
5
suggests that some arsenic may also be present as arsenate ions in clays or phosphate
minerals. Dale et al (1999) report that As is associated with pyrite.
2.2.3 Boron
Boron is regarded as being highly volatile (Meij, 1999). Approximately 50% of the
element is found in the flyash. It is depleted in the bottom ash. Clarke and Sloss
(1992) state that B2O3 is the likely volatile species.
Swaine (1990) indicates that most of the boron is organically associated in coal.
2.2.4 Beryllium
Beryllium appears to be partly volatilised and, according to the KEMA model (Meij,
1999), the element is slightly enriched in the flyash.
Both Swaine (1990) and Finkelman (1994) suggest (level of confidence 4) that Be is
associated with the organic matter, with some present in clays. Dale et al (1999)
report that Be is present in the silicates in the Australian coals studied.
2.2.5 Cadmium
Cadmium is volatile and is enriched in the flyash (Meij, 1999). Finkelman (1994)
reports (level of confidence 8) that Cd is predominantly associated with sphalerite,
ZnS, although it may be found in other sulphides. This is in agreement with the
summation of Swaine (1990). Dale et al (1999) report Cd as being present in the
monosulphides, pyrite and silicates.
2.2.6 Chlorine
Chlorine is regarded as being totally volatilised. There is an IEA publication on
halogen emissions from coal combustion (Sloss, 1992); this is a comprehensive
review of the research on the behaviour of the halogens during coal combustion.
Most researchers are of the opinion that most (>90%) of the chlorine is emitted as
HCl (Sloss, 1992, p.24).
CSIRO’s data on the emissions from Queensland power stations indicates that most of
the chlorine is emitted in the gas phase.
The amount of chlorine present in the coal as fired is an important factor in the nature
of the mercury compounds formed during combustion (see 2.11).
2.2.7 Cobalt
Cobalt is not regarded as volatile and is evenly distributed between the bottom ash
and flyash.
Cobalt is most likely associated with sulphide minerals but also in clays and in the
organic matter (level of confidence 4, Finkelman, 1994). Dale et al (1999) found that
Co was associated with the silicates in the Australian coals studied.
6
2.2.8 Chromium
Chromium is in Group II and is regarded as an element that can be vaporised and then
condensed on the flyash. However the KEMA model (Meij, 1999) shows Cr to be
evenly distributed between the bottom ash and flyash.
The final valency of chromium is of interest. Huggins et al (1993) used XAFS
spectroscopy to determine the valency of the Cr specks in both coal and coal ash. The
researchers concluded that Cr was present as Cr3+ in the coal and “is predominantly
(>95%), if not exclusively present as Cr3+ in ash samples”, even in the first flyash
fractions.
Finkelman (1994) states that there is “insufficient data to specify the modes of
occurrence of Cr in coal”. Some organic association is suspected (see also Swaine,
1990). Dale et al (1999) report that Cr is present in the oxide/carbonate/monosulphide
group but that it may be present in the silicates and also associated with the organic
matter.
2.2.9 Copper
Copper is regarded as slightly volatile. The element is generally depleted in the
bottom ash and somewhat enriched in the flyash.
Copper is likely to be present in coal as chalcopyrite or other sulphides and possibly
as organically bound species (Swaine, 1990). Dale et al (1999) report that Cu is
present in the oxide/carbonate/monosulphide group, in the pyrite and in the silicates.
2.2.10 Fluorine
Fluorine is a Group III element. During combustion, fluorine is released as HF. As
stated in section 2.6, the behaviour of fluorine and the other halogens is reviewed by
Sloss (1992).
According to Swaine (1990), the mode of occurrence is “by no means settled”.
Fluorapatite has been reported. It is also thought that F is present in the clays as a
-
replacement for the OH .
2.2.11 Mercury
Mercury is perhaps the most studied of the elements released by coal combustion. The
KEMA model (Meij, 1999) predicts that 49% of Hg is retained in the flyash.
Curtis (1999) reports on the development of a method to iso-kinetically sample stack
gas and determine the speciation of Hg present. Two volatile forms of Hg are present
in the stack gas of power stations: a water soluble compound, most probably HgCl2,
and elemental Hg. Curtis reports that in the Canadian power stations studied, the
majority of Hg is present in the stack gas as elemental Hg.
7
Devito and Rosenhoover (1999) studied the fate of Hg in six power stations in the
USA. These were burning coals containing Cl at 1000 mg/kg or more. Greater than
50% of the total Hg (80 to 95% of the oxidised Hg) in the stack gas was removed by
the scrubbers.
Meij (1999a) reports similar findings in Dutch power stations where mercury
emissions are low. Flue gas desulphurisation systems remove at least 50% of the
mercury; total mercury removal overall is an average 72%.
It is apparent that the speciation of Hg is to some extent dependent upon the levels of
chloride present. Meij (1999) states that: “Regardless of the form in which mercury is
present in the coal, elemental mercury is released during combustion...... However, in
the presence of HCl, Hg (0) can be (partly) converted into HgCl2 at temperatures
o o
between 300 C and 400 C......” HgCl2, being less volatile than Hg (0), will more
readily condense on the flyash in the ESP or baghouse.
In coal, it seems that much of the mercury is associated with pyrite (Finkelman, 1994,
level of confidence 6). Swaine (1990) states that Hg is “probably associated with
pyrite and sometimes sphalerite, with organically bound Hg still an uncertainty”. Dale
et al (1999) report that Hg is associated with the pyrite and that there is residual Hg
present after acid extractions; this could be organically bound, or more likely Hg
associated with framboidal pyrite, which is protected by coaly matter during acid
leaching.
2.2.12 Manganese
Clarke and Sloss (1992) have manganese listed in Group I, although Meij (1999)
places it in Group IIc (the least volatile of the volatile elements).
Swaine (1990) states that Mn is associated with carbonate minerals and clays. Dale et
al (1999) found Mn present in the oxide/carbonate/monosulphide group.
2.2.13 Molybdenum
Molybdenum is a Group II element and is slightly enriched in the flyash (Clarke &
Sloss, 1992; Meij, 1999).
Swaine (1990) states that “the mode of occurrence of Mo in coals ranges from mostly
inorganic to mostly organic”. Dale et al (1999) report that Mo is present in the
monosulphides, pyrite and is possibly associated with organic matter (this may in fact
be Mo present in framboidal pyrite).
2.2.14 Nickel
Nickel is in Group II and is generally enriched to a slight extent in the flyash (Clarke
& Sloss, 1992; Meij, 1999).
According to Finkelman (1994), there is a lack of any direct evidence for the modes
of occurrence of Ni in coal; Ni may be either organically bound or associated with
8
sulphides (level of confidence 2). The results of Dale et al (1999) indicate that Ni is
present in both the monosulphides and the organic matter.
2.2.15 Lead
Lead is vaporised during combustion and condenses on the flyash particles; it is
depleted in the bottom ash (Clarke & Sloss, 1992; Meij, 1999).
Lead occurs as sulphides (galena) or associated with sulphide minerals (Finkelman,
1994; level of confidence 8). Dale et al (1999) report that some of the Pb is present in
the silicates.
2.2.16 Selenium
Selenium is a Group III element and is volatilised during coal combustion. Finkelman
o
et al (1992) report that Se is largely volatilised at 550 C. Cramer (1986) reports that
Se is less volatile in the presence of CaCO3. He postulates that calcium selenate is
formed. As stated in section 2.2, a tube furnace at temperature was used to combust
the coal in 100% oxygen. Clemens et al (1999) report the likely formation of calcium
selenate during the stoker combustion of alkaline sub-bituminous coal in NZ.
However, CSIRO studies into the leachability of flyash indicates that the selenium is
present as selenite, Se (IV), in the flyash from NSW power stations.
The work of Andren et al (1977) is interesting. The researchers report that all the Se
present in the slag, flyash and vapour phases of a steam plant in the USA was
elemental Se. They speculate that Se (IV) is reduced to Se (0) by SO2. No Se (IV) was
found in the flyash. This early work is contradictory to the observations of other
researchers, although the mass balances obtained by CSIRO on power stations are
often low, with little or no selenium being retained in the nitric acid/peroxide trap
used in the iso-kinetic sampling of the stack gas.
Finkelman (1994) states that the majority of Se in coal is associated with the organic
matter (level of confidence 8). Lesser amounts are present in sulphides or as selenides
or bonded to iron oxides (see also Swaine, 1990). Dale et al (1999) report selenium as
being associated with pyrite and organic matter.
2.2.17 Zinc
Zinc is listed as a Group II element. It is enriched in the flyash and thus depleted in
the bottom ash (Meij, 1999).
Zinc is most likely to occur as sphalerite in coal (Swaine, 1990). The results of Dale et
al (1999) are consistent with this.
2.2.18 Uranium and Thorium
Uranium exhibits behaviour intermediate between that of Group II and Group III
elements. Thorium is listed as an involatile Group I element (Clarke & Sloss, 1992).
9
Clarke and Sloss (1992) report the work of Chadwick et al (1987), who reported that
the behaviour of U may be affected by its oxidation state in the mineral matter in the
coal, eg. if present as the refractory silicate mineral coffinite, (USiO4)1-x.(OH)4x, the U
is unlikely to be vaporised but if present as uraninite, UO2, will be vaporised and
condense on the flyash. Swaine (1990, pp 170-171) concludes that U may be in the
mineral matter but also organically bound to the coal. In the latter form, it would
presumably be volatile.
Thorium is unlikely to be organically bound. It is found in minerals such as monazite
and zircon (Swaine, 1990).
2.3 DISCUSSION
There is no direct identification of the speciation of trace element compounds formed
during the combustion of coal in PF-fired power stations. Conclusions have been
drawn from the distribution of the trace elements in the ash fractions (bottom ash,
flyash and fine particulate matter) and stack gas, as well as the observed behaviour in
the scrubbers (FGD).
Based on these observations, the trace elements have been listed according to their
relative volatilities. This has been combined with thermodynamic data and the likely
forms of the trace elements in the combustion zone, which have been suggested by
researchers worldwide.
For some trace elements there is doubt as to the original mode of occurrence of the
trace element in the coal. It is unclear whether this has a bearing on the ultimate fate
of a particular trace element; in some cases it will have an effect, eg. the same trace
element incorporated into silicates would behave differently if associated with
sulphides or organic matter.
Most of this data has come from researchers overseas. There is an opportunity to
collate similar data obtained from Australian power stations. The recent tests
completed by CSIRO for the electricity generators in Queensland contain extremely
good data (as indicated by the mass balances obtained). Valuable knowledge of the
behaviour of trace elements in these power stations could be gained by determining
the modes of occurrence of the trace elements in the feed coals, as well as obtaining
more information on the station operating conditions (eg. temperature of ducts,
temperature at the inlet and outlets of ESPs, etc). The results would be extremely
useful in assessing any computer models based on thermodynamic or kinetic data.
10
3 TOOLS FOR THE MODELLING AND PREDICTION
OF TRACE ELEMENT DEPORTMENT IN
COMBUSTION PROCESSES
3.1 INTRODUCTION
Coal is the most abundant organic-rich sedimentary rock on the face of the planet.
Some constituents of coal however are potentially toxic. That potential toxicity arises
from the trace metals and metal compounds bound with the coal’s mineral matter and
organic component.
These potentially toxic trace element species are released during the combustion of
coal in power generation processes and pose an environmental and human health risk,
depending on their concentration, physical and chemical forms and toxicity. Other
factors that govern the release of these toxic species are the partitioning behaviour of
the trace element in the combustion and environmental control process systems.
In order to minimise the potential risk of these toxic trace elements, it is essential to
understand and be able to predict the behaviour of the elements in the combustion
process. In order to get a full understanding of the processes involved, it is necessary
to characterise, quantify and model the physical and chemical mechanisms that
govern the fate of the trace elements both during and after the coal combustion
process.
It is recognised that to fully understand and predict the fate of trace elements in
combustion processes, several factors need to be addressed. These factors are:
· Combustion system design – for this dictates the fraction split between
bottom ash/slag and fly ash
· Coal rank and composition – mode of occurrence of the trace element
species and their interaction with mineral matter transformations and
partitioning mechanisms (Ratafia-Brown, 1994)
· Firing conditions and level of equipment maintenance
The aim of this project, “Trace element deportment”, is to provide well-characterised
information on emissions from New South Wales and Queensland power plants and to
develop accurate predictive tools and techniques for the relevant emissions and
hazardous air pollutants.
The intention of this review is to examine the predictive models, tools and techniques
available to the power generation industry for the prediction of transformation and
partitioning behaviour of a coal’s trace elements.
3.2 GENERAL MODELLING APPROACHES
One of the earliest and simplest approaches to modelling and predicting trace element
deportment or hazardous air pollutants (HAPs) is the use of generalised emission
11
factors. An emission factor, as defined in Compilation Of Air Pollutant Emission
Factors (AP-42) documentation “is a representative value that attempts to relate the
quantity of a pollutant released to the atmosphere with an activity associated with the
release of that pollutant. These factors are usually expressed as the weight of pollutant
divided by a unit of weight, volume, distance, or duration of the activity emitting the
pollutant (for example, kilograms of particulate emitted per megagram of coal
burned). Such factors facilitate estimation of emissions from various sources of air
pollution. In most cases, these factors are simply averages of all available data of
acceptable quality, and are generally assumed to be representative of long-term
averages for all facilities in the source category (that is, a population average).
The general equation for emission estimation is:
E = A x EF x (1-ER/100)
Where:
E = emission,
A = activity rate,
EF = emission factor, and
ER = overall emission reduction efficiency, %.
ER is further defined as the product of the control device destruction or removal
efficiency and the capture efficiency of the control system. When estimating
emissions for a long time period (eg. one year), both the device and the capture
efficiency terms should account for upset periods as well as routine operations” (AP-
42, 1995).
To facilitate the implementation of the information contained in the Compilation Of
Air Pollutant Emission Factors (AP-42), the US Environment Protection Agency
(EPA) have released a series of software application models. These include:
· The Factor Information Retrieval (FIRE) Data System
· PM Calculator
· The Integrated Air Pollution Control System (IAPCS) for coal-fired utility
boilers
· SPECIATE
The Factor Information Retrieval (FIRE) Data System is a database management
system containing the US EPA's recommended emission estimation factors for criteria
and hazardous air pollutants. FIRE includes information about industries and their
emitting processes, the chemicals emitted, and the emission factors themselves. FIRE
allows easy access to criteria and hazardous air pollutant emission factors obtained
from the Compilation Of Air Pollutant Emission Factors (AP-42), Locating and
Estimating (L&E) series documents, and the retired AFSEF and XATEF documents.
The information can be browsed through records in the database or obtained via
specific emission factors by source category name or source classification code
(SCC), by pollutant name or CAS number or by control device type or code. FIRE
6.22 contains emission factors from AP-42 through Supplement D and part of
Supplement E (through 3/15/99) of the Fifth Edition. FIRE 6.22 also contains all EPA
12
point and area Source Classification Codes (SCC) through March 1999, and provides
a convenient interface to these codes (EPA web site, 1999).
The PM calculator was developed by the Office of Air Quality Planning and
Standards (OAQPS) to help the US States PM-10 (particulate matter less than 10
micrometres in diameter) and PM-2.5 (particulate matter less than 2.5 micrometres in
diameter) inventories. The Calculator requests the SCC (source classification code),
the primary particulate control device code, the secondary control device code, and
the uncontrolled emissions. The uncontrolled emissions may be either TPM (total
particulate measured) or PM-10.
The PM Calculator uses information from AP-42 to calculate controlled particulate
emissions. It calculates controlled PM-10 and PM-2.5 from uncontrolled PM-10 or
total particulate emissions for point sources with up to two control devices (EPA web
site, 1999).
Version 5 of the IAPCS program is a product of The Air Pollution Prevention and
Control Division of EPA's National Risk Management Research Laboratory. IAPCS
is a Windows-based program that estimates the capital and annual costs for sulfur
dioxide (SO2), nitrogen oxides (NOx) and particulate matter controls for coal-fired
utility boilers. It addresses boiler characteristics, coal analyses, pollution controls,
and economics. Output from the program includes material balances, emission
summaries, capital costs and annual costs. The program addresses 16 control
technologies applicable to power plants ranging in size from 100 to 1300 MWe of
generating capacity.
IAPCS can estimate costs for each of these control technologies, or logical
combinations of them:
· SO2 control technologies: wet flue gas desulphurisation, lime spray drying,
advanced silicate process (ADVACATE), coal supply option, lime
injection (LIMB) and dry sorbent injection
· NOx control technologies: low NOx combustion, natural gas reburning,
selective catalytic reduction, and selective non-catalytic reduction
· Particulate matter control technologies: electrostatic precipitators, fabric
filters, and gas conditioning
· Integrated combustion technologies: fluidised bed combustion and
integrated gasification combined cycle
The software is readily available from the US EPA free of charge.
The SPECIATE 3.0 for Windows database contains organic compound and particulate
matter speciation profiles for more than 300 source types. SPECIATE 3.0 contains
262 new TOC profiles and 13 new PM profiles. The assignment of profiles to SCCs
has been updated to include these new profiles. Currently, every SCC has at least one
profile assignment. Some SCCs have more than one profile assigned to it. The profiles
attempt to break the total volatile organic compound (VOC) or particulate matter
(PM) emissions from a particular source into the individual compounds (in the case of
VOC) or elements (for PM) (EPA web site, 1999).
13
Generalised emission factors have however been uniformly dismissed as a predictive
tool for trace element emissions from coal-fired power plants. This is due to the fact
that variations in fuel composition are not taken into account during model
development or the modelling process itself. However, test data from individual
sources are not always available, and may not reflect the variability of actual
emissions over time.
Another generalised method that had minor implementation is the use of enrichment
factors. Enrichment factors are the ratio of particulates emitted to particulates
captured. The assumption is that trace elements that volatilise during the combustion
process, such as As, Se and Cd, have an enrichment factor greater than 100. Metals
with low volatility, such as Be, Cr and Ni, have lower enrichment factors. Many of
these enrichment factors were estimated on measurements from what is now
considered antiquated particulate control equipment and analytical methods which
would render a vast majority of the data useless when considering new control
technologies. It is now only Hg and Se that will escape capture by an electrostatic
precipitator, or ESP (Davidson & Clarke, 1996).
As with the “generalised emission factors” approach, the enrichment factor method of
modelling and prediction does not take into account mineralogy or mode of
occurrence of the trace elements in the fossil fuel used. Davidson and Clarke (1996)
quote the example of calcium absorption as well as the low volatile nature of elements
associated with aluminosilicates to further illustrate the point. They also emphasise
that this must cast substantial doubt on modelling approaches that are based solely on
trace element volatility alone. However, emission factors are frequently the best or
only method available for estimating emissions, in spite of their limitations (AP-42,
1995).
3.3 INTEGRATED MODELLING APPROACHES
Deluliis and co-workers (1993) developed a more sophisticated approach to modelling
and prediction of trace element partitioning and emission from coal-fired power
stations. The modelling approach combines the use of thermodynamic data, an
electrostatic precipitator (ESP) model, as well as mass balance calculations. The
authors found that, under normal plant operating conditions, almost all of the Hg and
a significant fraction of Se and F exit the stack in vapour form if no flue gas
desulphurisation system is present. Hence, only small fractions of these trace elements
are retained in cold-side electrostatic precipitators. The elements As, Be, Cd, Cr, Mn,
N, Pb and Sb were found to be partitioned between bottom ash and fly ash. The
retention of these elements in the cold-side electrostatic precipitators was estimated
using elemental mass distribution in the electrostatic precipitator inlet fly ash and ESP
performance model. The calculated retention efficiencies produced by the model
compared well with measured efficiencies from a pilot-scale electrostatic precipitator.
Hypothetical investigation carried out by the authors revealed that almost 90 % of the
non-halogen element emissions can be attributed to vapour-phase selenium. Tumati
and Bilonick (1996) carried out similar modelling work and obtained similar results.
Rizeq et al (1994), and Seeker and co-workers (1995), have developed a computer
modelling approach to predict metals behaviour in coal-fired combustion systems.
The model was based on earlier work carried out by Barton et al (1990) on the fate of
14
metals in waste combustion systems. To test the predictive capability of their
modelling approach, the authors compared the predictions of the overall partitioning
of metals in a pulverised bituminous coal-fired combustor with measured data from
full-scale facilities in the Netherlands. It was generally found that the comparison
between predicted and measured data was favourable.
The model consists of a group of sub-models designed to simulate the chemical and
physical influences on the behaviour of the trace elements in combustion systems. A
schematic representation of the sub-models in trace element partitioning analysis is
shown in Figure 1.
Combustion System
Thermal Analysis
Reaction/Vaporisation
Submodel
Entrainment
Submodel
Aerosol Dynamics
Submodels
Air Pollution Control
Submodel Emissions
Figure 1 Metal partitioning model (Rizeq et al, 1994)
As can be seen from Figure 1, the Combustion System Thermal Analysis is carried
out first in order to define the environment for the burning fuel and metal
vaporisation, as well as the condensing metal vapour.
The metal reaction and vaporisation sub-model is based on thermodynamic data for
trace elements and their compounds. The chemical reactions and phase behaviour
were determined using the program developed at NASA’s Lewis Research Centre.
The program, outlined later, utilises free energy minimisation as well as the JANAF
tables, and is based on two assumptions:
15
· All reactions achieve equilibrium at local conditions
· All component elements being coal and combustion air are intimately
mixed
Rizeq et al (1994) highlight that although it is recognised that equilibrium may not be
maintained throughout a combustor, the assumption holds true when considering the
high temperatures typical of combustion chambers. Kinetic models cannot be
developed due to the lack of sufficient kinetic data.
Uncontrolled partitioning of metals due to vaporisation was calculated from the
saturated vapour pressure of metals under predicted operating conditions and fuel
composition. The saturated concentration of each metal was then estimated and
compared to the available concentration of the same metal.
The aerosol dynamics sub-model simulates the nucleation and condensation behaviour
of the metals with nearby particles. The model can perform calculations to determine
the fraction of metals condensed onto or into each particle size range. A separate
model can simulate particle coagulation. Sub-models simulating the behaviour of the
following types of equipment can, and have been, incorporated into the model:
· Electrostatic precipitators (Faulkner & DuBard, 1984)
· Venturi scrubbers (Calvert, 1972; Rudnick et al 1986)
· Cyclones (Leith & Licht, 1972)
· Baghouses (Dennis, 1977)
· Baffle plate scrubbers (Calvert, 1972)
Rizeq et al (1994) were able to draw the following conclusions from the modelling
studies:
· Changes and variations in combustion temperature and coal chlorine
content influence the vaporisation of non-metals such as Ba, Ni, Be and
Cr, however, there is no influence on medium volatile elements such as
Cd, Pd, As, Sb and Ti and volatile metals such as Se and Hg. The latter
two types were predicted to partition 100% to the gas phase at between
800°C and 1800°C. This conclusion depends on coal type and operating
conditions.
· The main form of Hg at the combustor exit was predicted to be elemental
Hg and the main form at the APCD inlet was predicted to be HgCl2. Wet
scrubbers are sufficient for the chloride due to its partial solubility.
· The proportion of metals condensing into or onto small nucleating
particles in the flue gas increases with increasing volatility of the metal
and as the particle loading in the gas decreases.
· The results of the study highlight the importance of knowing and using
operating conditions and fuel composition during a test to determine
representative emissions.
· Good comparisons between predicted and measured data provide
confidence in the modelling approach developed. This approach can be
effective in planning tests and optimising operating conditions. The model
16
can also be implemented to determine the impact of different combustor
and APCD types, operating conditions and coal types on metal emissions.
One criticism of the model is that, being based mainly on thermodynamic volatility of
the elements, any significant chemical interactions, such as absorption by host
minerals, will affect the accuracy of the predictions (Davidson & Clarke, 1996).
Work by Bool and Helble (1995) further examined this problem to enhance the
predictive capabilities of their model, recognising that vaporisation and the
subsequent recondensation of many trace elements during coal combustion may
depend upon the form of occurrence of the trace elements in the fuel coal. They tested
their hypothesis that the forms of occurrence of the elements As, Hg, Se, Zn, Cr and
Sb would dictate their partitioning behaviour. By undertaking well-controlled
laboratory combustion tests and measuring the concentration of each element as a
function of fly ash particle size, they were able to show that those elements associated
with the coal organic matrix either ion-exchanged or were covalently bound and those
associated with pyrite were highly volatile. Examples of these elements are As, Hg,
Se and Sb.
In contrast, elements associated with silicate or oxide minerals are relatively non-
volatile. Bool and Helble (1995) also found that for all elements, with emphasis on
Zn, differences in the form of occurrence between coals led to differences in the
fraction volatilised. Also, for As and Se, condensation was dependent on fly ash
chemistry.
By using thermochemical equilibrium modelling to interpret the experimental data,
they concluded that As vapour reacted to form calcium arsenate during combustion of
the high Ca sub-bituminous coal and arsenic oxide in the low Ca bituminous coal.
Linak and Wendt (1994) also modelled trace metal transformation mechanisms during
coal combustion. Emphasis was given to predicting the size-segregation of trace
metals in pulverised coal-fired power plant effluent and the impact of fuel
composition and combustion conditions. The metals examined were Be, Cr, Ni, As,
Se, Cd, Sb, Hg and Pb.
Figure 2 illustrates a controlling mechanism for particle formation in combustion
systems put forward by Linak and Wendt (1994). As can be seen from the schematic,
the formation of small particles enriched in trace metals is through high temperature
vaporisation. Trace elements may be vaporised as they are introduced or after
transformation within the combustor. Supersaturated vapour then condenses
heterogeneously on the surfaces of existing particles or nucleates homogenously to
form new particles (Linak & Wendt, 1994).
Linak and Wendt (1994) note that vapour pressures of pure compounds do not show
which species are favoured under equilibrium considerations and cannot alone be used
to predict when and under what conditions condensation will occur. Utilising NASA
CET89 computer code to calculate multi-component equilibria, predictions were
made for a list of the given species. Linak stipulates however that the code allows
only for ideal gas mixtures and pure condensed species, neglecting the possible
formation of complex slags and glasses. Solutions obtained from the code depend on
17
the species for which thermodynamic data are available. Any omitted species will lead
to erroneous predictions.
Figure 2 Controlling mechanism for particle formation in combustion systems (Linak
& Wendt, 1994)
Another limitation of the modelling approach noted by the authors is the metal-metal
interactions that may be important, as well as any likely dependence on the mix of
metal elements that are specified in the initial mix. Thermodynamic equilibrium
predictions do not account for any kinetic or mixing limitations that may control
species formation in practice. However, irrespective of these limitations, equilibrium
predictions provide a reasonable starting point to describe the behaviour of metals
(Linak & Wendt, 1994).
Predictions using the model indicated that Hg was unlikely to form any condensed
species, even at temperatures as low as 200 K. It was also predicted by Linak and
Wendt’s model that when sulphur was present, Ni, Be and Cr were the least volatile
metals, while Hg, Se and As were the most volatile. In the absence of sulphur, the
formation of sulphates or sulphides was non-existent, leading to the formation of the
more volatile chlorides. In effect, the result of chlorine in the system was to lower the
temperature at which the transition between vapour and condensed trace metals
occurred.
The authors point out that sulphur removal from coal may have undesirable effects
elsewhere in the combustion system.
Overall, the authors concluded from their study that experimental data from practical
combustors, utilising individual metals in their ideal form, supported the calculated
18
equilibrium predictions, in particular the role of chlorine. Coagulation processes can
also be modelled using existing software and these models agree well with
experimental data.
A similar study was undertaken by Lyyranen and co-workers (1995), where the
behaviour of 17 elements (As, Be, Cd, Co, Cr, Cu, Hg, Mn, Mo, Ni, Pb, Sb, Se, Sn,
Ti, V and Zn), and their compounds, during pressurised fluidised bed combustion and
gasification (p=12 or p=20 bar, T= 380 to 1380°C) of coal were modelled. The
process was modelled using the ChemSage 3.0 equilibrium model. No inter-metallic
compounds were taken into account. The authors found that the effect of
silicates/minerals on the behaviour of trace elements could be remarkable and it was
possible that the reactions with silicates would bind them so strongly that they were
unable to vaporise/react further. To overcome this, the reactions of the trace elements
with mineral species were taken into account via the input amount of the trace
elements given for the equilibrium calculation. Thus only a small fraction of the coal
containing trace elements are allowed to volatilise.
Pires and co-workers (1997) hold the belief that knowledge of the distribution and
concentration of trace elements in coal is very important not only to the geochemistry
but also for studies related to the mobilisation of these elements into the environment.
The volatility of trace elements was calculated using their partial concentrations in
organic and sulphide coal fractions. The applicability of trace element volatility to
estimate the atmospheric emissions was determined against experimental and
theoretical data. They also classified trace elements into different groups as a function
of volatility on elemental enrichment in fly ash and on Klein’s classification for a
Spanish coal.
The authors also modelled emissions from Leao coal combustion and showed that it
was possible to use the volatility of trace elements to estimate the mobilisation of
these species into the atmosphere. However, they found that complex coal combustion
models have not been readily incorporated into the mechanism that could explain the
noted trace element emissions in the vapour phase.
In order to analyse the behaviour of trace elements in coal-fired steam generators and
flue gas cleaning systems, Fahlke (1993) implemented a purely thermodynamic
approach to the problem. He found that the behaviour of elements and their species,
during combustion and in the course of the cooling of the flue gases, might be
explained by thermodynamic equilibria. For the relevant temperatures noted in
combustion, the DEXON, ESP and FGD systems, and calculations using
minimisations of the Gibbs energy by way of the program equiTherm, provided good
predictions for the behaviour of element species. Fahlke’s (1993) calculations
included simultaneous interactions of 50 species of 14 elements, each of which may
occur in either gaseous, liquid or solid phase. As soon as one species exits the gaseous
phase because of condensation of formation of a solid phase, it is no longer
considered in the next cooling stage.
Mathews (1988) took a similar chemical modelling approach in studying trace metal
speciation in hazardous waste incineration. Adopting the model illustrated in Figure 3,
Mathews implemented a combination of heat balance calculations and chemical
equilibrium calculations to determine the speciation of trace elements in combustion.
19
The heat balance computations were made to determine the excess air and fuel
requirement for a specified waste fee composition and incinerator operating
temperature. The heat balance model uses the enthalpies of the input stream, heats of
combustion, enthalpies of combustion products and a percentage heat loss from the
incinerator. The chemical equilibrium analysis is based on the assumption that at
equilibrium the free energy of the system must be at a minimum.
Mathews (1988) concluded from his modelling work that the fate of trace elements in
hazardous waste incineration systems depends on the following factors:
· Waste composition
· Operating conditions
· Air pollution control efficiencies due to different condensation
temperatures and mechanisms for the different species formed
The author also highlighted an important point, in that the goal of partitioning analysis
is to determine how much of the volatilised species will condense and be removed in
the air pollution control devices such as cyclone, ESPs, etc. Condensation
mechanisms for the various species must be examined further to predict enrichment
factors as a function particle size distribution (Mathews, 1988).
Operating Temperature
Heat Balance Ex Air Elemental
Waste Feed Calculation Composition of
Inputs
Solid
Phase
Free Energy
Data Chemical
Equilibrium Gas
Calculations Phase
Figure 3 Partitioning analysis flow chart (adapted from Mathews, 1988)
In 1995, Benson and co-workers developed a model to predict initial partitioning of
metals in the gasifier and the gas clean-up system. The model was developed as part
of the Energy and Environment Research Centre’s (EERC) investigation into methods
of predicting the fate of selected elements in integrated gasification combined cycle
(IGCC) and integrated fuel cell (IGFC) systems. The modelling work has been
focused on modification of two existing codes at the EERC: ash transformation
(ATRAN) and thermochemical equilibrium analysis of coals and ashes (TEACH).
The codes are being modified to include specific algorithms associated with trace
element behaviour. The resultant modelling package, TraceTran, is a consolidation of
the codes into a single program capable of predicting the size, composition and phase
of the inorganic species at a given temperature and pressure. TraceTran can be used to
predict the evolution of major, minor and trace elements during coal combustion and
gasification. The TraceTran model is based on the algorithm shown in Figure 4.
20
Analytical Characterisation
ASTM Analysis Trace Element Trace Element Associations
· Proximate/Ultimate Associations · CCESM/AIA/WDS
· Bulk & Trace Comp. · Chemical Fractionation · Size, Comp., Assoc.
Mass Balance
Inorganic Mass Balance
and Associations
PSCD Calculations
TEACH TRACETRAN
Gasifier · Vapours, Liquids, Solids · Locked/Liberated
System for release from coal · Mineral Fragmentation
Conditions · Mineral Coalescence
· Homogeneous Conden.
· Heterogeneous Conden.
System TEACH · Particle Size and
Conditions Composition Evolution
· Vapours, Liquids, Solids
Downstream of for release from coal
Gasifier
Outputs
Gas species Liquid/Solid Species
· Composition · PSCD and Phases
· Species
Inputs Data Processing Outputs
Figure 4 Model structure (Benson et al, 1995)
As can be seen from Figure 4, the first task of the model is to determine the
association of the major, minor and trace elements in the coal prior to utilisation, as
this will affect their phase, size and composition in the waste streams. To predict the
21
transformation of the trace and minor elements effectively, their distribution among
the minerals needs to be determined. Once the mass balance is performed for the coal-
input data, it is necessary to determine which of the inorganic components will be
vaporised during the initial stages of the conversion process. The calculations on
vaporisation of the species are performed with the use of the EERC thermodynamic
program. The authors state that the EERC thermodynamic program has been
appropriately upgraded to include trace element phases as well as enhancements made
with the ability to include more phases through the use of additional thermochemical
programs, such as FACT.
Once the vaporised species have been removed from the system, the remaining
constituents are processed through algorithms for mineral fragmentation and
coalescence for both mineral and organically associated species. The fragmentation,
coalescence and shedding algorithms were developed with the aid of data from full-
scale systems and data generated in pilot-scale facilities.
The state of the volatile species at the appropriate downstream conditions are
determined using the TEACH code. The TEACH data is used to calculate the species
that will condense prior to a given process condition. The resultant species are
compiled and manipulated into various compositions and sizes. These distributions
can be used to determine effective control technologies for a specific coal or to locate
a coal compatible to a specific control technology (Benson et al, 1995).
The authors emphasise that although the model was intended to aid in the control of
trace element emissions, only minor attention was given to the engineering
mechanisms of the process. By implementing a series of low level engineering
algorithms for the bed fluidisation, cyclone efficiency and filter performance, the
model was found to predict reasonable results for Hg and Na and only marginal
results for Pb. This result was attributed to the lack of robust engineering models.
One of the largest modelling projects reported to date has been that of the
development of the new Toxic Partitioning Engineering Model (ToPEM). The model
was developed with support from the Federal Energy Technology Center (FETC), the
Electric Power Research Institute (EPRI) and VTT (Finland). The overall purpose of
the project was to bring together a team of researchers of specific expertise to develop
sub-models for the existing Engineering Model for Ash formation (EMAF). The team
of researchers was made up from Physical Sciences Inc. (PSI), USGS, MIT, the
University of Arizona, the University of Kentucky, University of Connecticut and
Princeton University (Bool et al, 1996). The Toxic Partitioning Engineering Model
was developed to be broadly applicable to regulators and utility planners. The sub-
model processes and the associated research team are illustrated in Figure 5.
22
Stack
Combustion Organics Model
MIT Princeton
PSI Validation
University U.Conneticut
U. Kentucky
Boiler Air heater
Burners
Particulate SO2 Induced
Collector Scrubber Draft Fan
Fuel
Post-Combustion
U. Arizona
MIT
PSI
Forms of
Occurrence
U. Kentucky
USGS
MIT
Coal Pulverisation Coal Supply
Figure 5 Project organisation (Bool et al, 1996)
One of the most important questions addressed by the research team is whether the
form of occurrence of an element in the coal affects its form of emission at the end of
the process. To address this issue, the specific mineral associations of elements and
the relationship between trace metal form and standard analyses were examined. This
issue is being addressed with the aid of XAFS analysis as well as low temperature
(200°C) ashing, chemical analysis, X-ray diffraction, coal segregation via flotation,
synchrotron radiation X-ray fluorescence microscopy (SRXFM), ammonium acetate
and acid leaching, electron micro-beam measurements and low and moderate
temperature heating tests to determine the forms of elements in coal (Bool et al,
1996). The model is still under development.
The fate of trace elements in fuel combustions system has also attracted considerable
attention from the UK Department of Trade and Industry Coal R&D Program. The
program highlights that an essential prerequisite for determining the occurrence and
distribution of trace metals in both raw and cleaned coal is a rapid method of
determining their concentration. The program aims to assess the effectiveness of
conventional coal cleaning as a means of removing trace metals (Department of Trade
and Industry UK, Coal R&D Program, PP 203).
23
The specific objectives of this project are:
· Determine mode of occurrence and distribution of trace elements in coal as
a function of coal preparation and cleaning.
· Develop on-line measurement techniques for trace metals.
· Determine the fate of trace metals as a function of coal and ash
composition during combustion.
· To correlate trace metal distribution and deportment data for cleaned and
raw coals and for in-furnace and post-boiler pollutant concentrations.
· To ascertain the effects of cleaning on gaseous emissions, combustion
efficiency, flame stability, slagging, fouling and corrosion.
No further information is available on this project.
Another study currently being undertaken by the UK Department of Trade and
Industry Coal R&D Program is the Prediction of the Behaviour of Trace Elements in
Industrial Combustion and Gasification Systems - A Jointly Funded Project with
BCURA (Department of Trade and Industry U.K., Coal R&D Program, PP 216,
1999). The project was due to be completed in September 1999. As of the completion
of this review, the author has not sighted any formal results or report from this work.
3.3.1 Computational Fluid Dynamics
The current state of development of models based on heat balance methods and
computational fluid dynamics (CFD) is the subject of a detailed and excellent review
of modelling of combustion processes in stationary plant appliances. The review,
produced by Pourkashanian and co-workers (1998) was undertaken on behalf of the
Department of Trade and Industry, UK. The authors cover such topics as:
· Combustion modelling methods
· Heat balance modelling
· CFD modelling
· Physical modelling
· Commercial code
· Modelling of combustion plants
· State of the art in CFD
The authors concluded from their review “very considerable advances to CFD models
are taking place, ranging from power station applications to gasification. Coal blends
present many problems, however CFD codes have the potential capability of handling
many environment aspects, SOx, NOx, unburned carbon and toxic metals”.
No example of CFD application to trace element deportment was given in the review.
3.3.2 The Electric Power Research Institute (EPRI), PISCES software
The Electric Power Research Institute (EPRI) was formed over 25 years ago as a non-
profit organisation committed to providing science and technology-based solutions of
24
value to the energy industry. The institute manages a program of scientific research,
technology development and product implementation. One of the institute’s projects
relevant to the current study is the project to evaluate trace element emissions from
power stations that use fossil fuels. The PISCES model was intended to provide
power station personnel with a greater understanding of the operational factors that
contribute to the emission of trace elements. The major objective of the project is to
provide a structural framework by which economic studies can be undertaken to
reduce trace element emissions (Clarke & Sloss, 1992).
The EPRI program is known as the Power Plant Integrated System: Chemical
Emission Study (PISCES). The PISCES program consists of several activities that
include:
· An extensive database of coal trace element species in power stations
· A model to track and estimate power station emissions
· An ongoing field program for the collection and collation of new power
station data
· Guidelines for the selection of emission control technologies
· A guide for trace element sampling and analytical methods
Utilising either a deterministic or probabilistic estimate of the gaseous, liquid and
solid emissions for a given plant configuration, the PISCES model can simulate the
environmental performance of a power plant based on its sub-processes. By
characterising the input and output streams of such sub-processes as the electrostatic
precipitators, waste handling, scrubbers and water treatment, the flow rates of the
process are quantified based on internal mass and energy balance calculations for the
plant size, design and fuel composition. The model can also accommodate variations
in plant performance, emission constraints and coal properties.
The probabilistic component of the PISCES model can be used to determine
uncertainty in mass balance calculations caused by variations in trace element input
and efficiency of the control technologies implemented (Rubin et al, 1991). The
probabilistic mode of the application yields rates of emission occurrence.
The most important implementation of the PISCES model is the determination of
trace element partitioning between the output streams (Clarke & Sloss, 1992).
3.3.3 COFERS Co-burning Feed Rate Simulator
Another software package from the Electric Power Research Institute (EPRI),
COFERS is a computer model developed for evaluating the implementation of co-
burning coal and waste material in a power plant (Stadler et al, 1994). The program
calculates the distribution of furnace feed constituents arising from feed coal and co-
burned waste to all major power plant solid waste streams. The program also allows
for determination of the rates and quantities at which waste can be burned with coal
under user-defined mass and waste composition limits. COFERS can also estimate
both the composition and rates of bottom and fly ash as well as FGD waste produced
during the co-burning process. Sensitivity analyses can be performed on a variety of
parameters.
25
Coal properties are provided in the software model by seam type. The application
contains coal trace element data for 74 different coal seams located in the USA. The
properties of co-burned wastes are modeled the same way as coal, hence it is highly
recommended that characterisation is obtained for all wastes before inclusion in the
COFERS model.
COFERS version 1 was proposed for release by EPRI in 1994.
3.4 GENERIC COAL QUALITY SYSTEMS
3.4.1 The Coal Quality Management System (CQMS)
In a collaborative project with BNI Coal Ltd., and the North Dakota State Industrial
Commission, MTI has developed a Windows-based software package, the Coal
Quality Management System (CQMS).
CQMS was developed with the purpose that coal mining and power plant operations
personnel use fuel quality information to make operational adjustments that increase
plant efficiency and minimise maintenance costs.
Among other applications, using CCSEM fuel quality data CQMS can provide
relative comparisons of fuels by calculating the following power system performance
indices:
· sulfate index (low-temperature convective pass fouling)
· silicate index (high-temperature convective pass fouling)
· wall slagging index
· abrasion index
· erosion index
· cyclone slagging index
· strength index
· ash resistivity
These advanced power system performance indices relate coal characteristics, as
determined by CCSEM and chemical fractionation, to ash behaviour in a coal-fired
utility boiler. The indices are in the process of being validated. The ability of the
indices to provide quantitative information related to specific boiler operation is
currently being assessed. The results of the assessment to date appear promising,
however, the database of coals and boiler types is still limited. Fuel performance is
estimated in terms of slag flow behaviour, abrasion and erosion wear, wall slagging,
high-temperature silicate-based convective pass fouling, and low-temperature sulfate-
based convective pass fouling. In addition, the resistivity of the fly ash related to
collection in an electrostatic precipitator can be calculated (MTI public relations
information, 1999).
26
3.5 INFORMATION DATABASES
3.5.1 CONSOL Inc. Coal Database and Coal Quality/Power Cost Model
In anticipation of the 1990 Title III of the Clean Air Act Amendment (CAAA), which
empowered the US Environment Protection Agency to set emission standards for a
variety of air emissions from combustion sources, CONSOL Inc. began a systematic
comprehensive analysis for all its product coals in 1981 (Devito et al, 1994).
CONSOL Inc. operates over 30 mines in the USA with an annual coal production of
about 70 million tons. The CONSOL Inc. coal database includes annual trace element
analyses for most CONSOL mines and coals since 1981. Since 1990, the number of
trace elements determined has been increased and includes all of the eleven CAAA
trace elements. The eleven elements or compounds of these elements found in coal
are among the 189 substances identified in the legislation as air toxins. In addition to
these trace elements, the CAAA also addresses HCl and HF emissions, derived from
the chlorine and fluorine content of the fuel.
The CONSOL database consists of the traditional coal quality parameters and
generally includes the trace elements listed in the CAAA for eastern and mid-
continent coals ranging from high volatile to mid-volatile bituminous. In 1992, the
database was expanded with trace element analyses of paired raw and clean samples
from eight coal preparation plants.
The database has been reported to be particularly useful for addressing utility
environment concerns because it represents commercial coal shipments and provides
an excellent basis for the variability of commercial coal quality over time for a given
coal.
Devito et al (1994) utilised the CONSOL database to determine trace elements in
coals and their emissions. They found that trace element variability, expressed as
percent standard deviation, ranged between 10 to 30 % for coals from a single mine
processed through a well-controlled preparation plant. Trace element variation
throughout a seam is similar to the overall inherent ash and major element variability.
Correlation analysis was conducted to determine if the ash content had a direct effect
on trace element levels. It was found that individual trace element concentrations have
a weak to moderate correlation to the ash content of the coals examined. However, the
combined total concentration of the 11 CAAA trace elements is strongly correlated to
the ash concentration for cleaned Pittsburgh seam coals. The authors highlight,
however, that while general correlations were observed for specific trace elements,
they were of little practical value in predicting trace element concentrations.
In an effort to assess the usefulness of fuel options for power generation units to meet
the CAAA 1990, CONSOL Inc. developed the Coal Quality/Power Cost Model
(CQ/PCM). The objective of the CONSOL CQ/PCM is to predict power station
equipment performance from coal quality data, and determine the effect on
busbarpower production costs. Abbott et al (1994) utilised the CQ/PCM for boiler
performance prediction and enhanced the model by the addition of slagging and
fouling factors. The slagging and fouling factors were developed from the CONSOL
0.4 MW pilot-scale combustion test facility, and a theoretical model for predicting
27
elemental ash deposition composition using coal ash chemistry and combustion
environment data. The paper describes the application of the model for predicting ash
deposit composition in the pilot-scale combustor and boiler performance. Model
predictions for mill, boiler and ESP performance were compared to pilot and
commercial test burn data. The model was found to be useful for the assessment of
fuel options for existing units, in particular for blending or switching coals (Tumati,
1994).
DeIuliis and Tumati (1995) have adapted and implemented the CQ/PCM to the 11
CAAA trace elements. Trace element partitioning, and its effects on stack emissions,
in a 500 MW power utility was investigated for different coals. The trace element
partitioning and cost analysis were also performed for various scenarios (Jak, 1999).
3.5.2 British Coal Corporation Emissions Monitoring Database
Over the years, the British Coal Corporation has collected a considerable quantity of
emissions data from coal-fired plants and more recently, from appliances such as
incinerators and oil and gas boilers. In an effort to make use of that information, a
database was developed on air pollutant emissions from combustion plants in the UK.
The database was designed to be interrogated in a flexible way and to select the
information that is required. The database is made up of data sheets that were
prepared for each type of combustion appliance, giving typical emissions of the major
pollutants of concern. These values are given for operation at maximum, or nearly
maximum, continuous rating (Hughes & Beer, 1994).
The database has been used to derive typical emission profiles for a range of power
plants and to draw conclusions on the relationship between pollutant emissions and
fuel and appliance characteristics. The specific findings, as reported by Hughes and
Beer, (1994) are:
· NOx emissions were found to increase with increasing O2 content in the
flue gas
· Hydrocarbon and CO emissions exhibit a correlation in terms of efficiency
of combustion
· SO2 emissions are dependent on fuel sulphur content, however
· The database does not allow for the comparison of the effectiveness of the
various particulate control systems due to the site-specific nature of the
results.
As of 1994, commercial prospects for the database were under investigation
3.5.3 The United States Geological Survey (USGS) COALQUAL
database
The USGS coal trace element database is the largest compilation of trace element data
that provides information on virtually all coal seams in the United States of America
(Tumati & Bilonick, 1996). The database contains information on more than 7,300
full-bed channel samples representing the in-ground composition of the coals. There
are approximately 136 properties identified in the database for each coal sample,
which include:
28
· Geological, stratigraphic, and geographic data
· Ultimate and proximate analysis
· Forms of sulphur
· Grindability and free swelling index
· Ash fusion temperature
· Calorific value
· Concentrations of up to 60 elements
Many researchers have used this database to estimate the quantity potential of trace
element emissions from coal-fired power stations. Typical of the use of the database is
the work of Oman and Finkelman (1994) to determine if coal switching produces a
benefit or penalty with respect to hazardous air pollutants. They concluded that
comparisons will have to be made on a coal bed by coal bed basis and that differences
in the calorific value of coals from different basins, as well as trace element loading
within a basin, can have a significant impact on input load. The impact of coal
cleaning on the input load to the boiler was also identified as a crucial factor in
assessing the merits of coal switching.
Tumati and Bilonick (1996), however, highlight several inherent problems when
considering application of the database to estimate commercial steam coal product
trace element concentrations and the resulting power station emissions. They note that
the USGS database represents trace element concentrations of various channel and
core samples collected without an appropriate comprehensive sampling strategy for
estimating reserves. In a sense, for ordinary sampling, a probabilistic scheme should
have been used to get a true representation of the population and, for spatial sampling,
an exhaustive systematic grid should have been utilised to ensure an unbiased
estimate of the deposit. Neither is observed.
Another shortcoming noted by Tumati and Bilonick (1996) is that the database only
represents coal deposits prior to any preparation processes such as physical or
chemical cleaning for mineral matter removal. The USGS coal data does not
represent commercial coal products used by power stations, since 75 to 80 % of the
coal east of the Mississippi River is processed through a coal preparation plant. Other
factors of concern are that some of the mines in the database have subsequently
closed, as well as the unknown precision or accuracy of the analytical methods
employed in determining the trace element concentrations of the samples.
To fully utilise the database, Tumati and Bilonick (1996) express the need to analyse
the data carefully and develop factors that would relate the USGS trace element
concentrations to currently used commercial coal supplies.
By implementing complex statistical techniques, Tumati and Bilonick (1996)
demonstrated that spatially weighted average trace element concentrations for the
entire coalfield differed substantially from simple equally weighted average
concentrations. The percentage difference ranged from –36 % to +48 %. Regression
models were also constructed to predict the effect of coal washing on trace element
concentration. They also noted that the percent of trace element variance accounted
for by ash and sulphur ranged as high as 89 %. Applying these models to two western
29
Pennsylvania CONSOL mines tended to be significantly closer to the actual values
than those produced by the US EPA using equally weighted averages.
3.5.4 Aerometric Information Retrieval System (AIRS)
AIRS is a database of information about airborne pollution in the United States and
various World Health Organisation (WHO) member countries. The system is
administered by the US EPA Office of Air Quality Planning and Standards (OAQPS),
Information Transfer and Program Integration Division (ITPID), located in Research
Triangle Park, North Carolina.
AIRS contains all the air quality, emissions, compliance and enforcement information
that OAQPS and state agencies need to carry out their respective programs for
improving and maintaining air quality. Reporting capabilities allow states to access
and use their data. It eliminates the need for individual states to maintain their own
databases of air pollution information and to reformat or reorganise data for
submission to the EPA's database. AIRS provide standard information requirements
and information handling procedures, which enables OAQPS to compare and to use
data from different sources.
The AIRS database contains a series of sub-systems and applications:
Air Quality Sub-system (AQS) contains measurements of ambient concentrations of
air pollutants and meteorological data from thousands of monitoring stations operated
by EPA, state and local agencies. The Air Quality Sub-system contains descriptive
information about each monitoring station, including its geographic location and who
operates it.
AIRS Facility Sub-system (AFS) contains both emissions and compliance data on air
pollution point sources regulated by the US EPA and/or state and local air regulatory
agencies. AFS contains data on industrial plants and their components: stacks, the
points at which emissions are introduced into the atmosphere; points, the emission
point or process within a plant that produces the pollutant emissions; and segments,
which are components of the processes that produce emissions. Compliance data is
maintained at the plant and point levels, tracking classification status, inspections, and
compliance actions. AFS also includes data for management of operating permits
applications and renewals.
Geographic, Common, and Maintenance Sub-system (GCS) is a repository of
reference data shared by the AQS, AFS and AMS sub-systems. The data includes
codes and code descriptions used to identify places, pollutants, processes, geographic
information, and values such as air quality standards and emission factors.
AIRS Graphics (AG) integrates data from multiple AIRS sub-systems into maps and
charts that enable users to identify patterns, trends and anomalies in air pollution data.
AIRS Graphics has interactive menus that make it easy to choose “graphical reports”,
and to select options that control their contents. AG software takes care of all the
details and displays colour graphics.
30
AIRS Executive (AE) is an IBM PC program that contains a select subset of data
extracted from the AIRS database. Its interface guides the user to air pollution
information on ambient air and plant emissions sources.
AIRSData provides access from the Web to the same subset of AIRS data. AIRSData
has summaries of air monitoring data for the past several years, the latest available
estimates of air pollutant emissions from major point sources, the overall regulatory
compliance status of those sources, and names of contacts in EPA and state/local air
pollution agencies. All this data pertains to the criteria pollutants (carbon monoxide,
nitrogen dioxide, sulfur dioxide, ozone, particulate matter and lead). AIRSData
includes data for the 50 states plus the District of Columbia, Puerto Rico and Virgin
Islands.
3.5.5 Centre for Air and Toxic Metals (CATM) Database
The CATM database was established, as per Folkedahl et al (1999), to facilitate
research on the prevention, transformation, behaviour and control of toxic metal
emissions from energy producing and incineration systems. The primary applications
of the database are to aid in the development of methods to predict the fate of metals
in fossil fuel systems, determine the effectiveness of control devices, assist in
identifying new control technologies and provide source-emission data that can be
used to assess health risks. The database provides an interactive user application from
which users can quickly and efficiently access and display information relevant to
their particular need or area of concern. The data is maintained in a relational database
engine at the Energy and Environmental Research Centre (EERC); however, the
application has been designed such that the data is accessible through the Internet.
The database has many practical applications, including investigating collection and
removal efficiencies of control devices, comparing source emissions and comparing
analytical results from bench and full-scale operations. The database can also be used
for investigating the usefulness of various sorbents, comparing the size distribution of
ash produced from a coal at varying run configurations, or comparing plant
performance as a function of fuel type and quality.
The information in the database was derived from reports and test results from a
variety of commercial, federal and academic sources, including data derived from the
ongoing CATM technical projects pertaining to trace element transformation, analysis
and control (Folkedahl et al, 1999).
Folkedahl et al (1999) state that the database is organised into three categories. These
categories are analytical, engineering and materials. To maintain the focus of the
database on trace element emissions, the database uses a sample-based approach, that
is the sample must have the associated analytical data to be included as well as the
engineering data. The engineering data is not entered solely. The database links the
information together so queries can be related to the analytical and engineering
information along with geographical information in a GIS system.
An example of the application of the database is provided by Folkedahl et al (1999):
Question: Since coal cleaning processes commonly remove a substantial amount of
pyrite, could we expect also a removal of As and Hg?
31
Approach: Correlate the Fe concentration of a number of coals with As and Hg
contents. In addition, compare the As and Hg concentration of the pulveriser rejects
produced during coal processing to the resulting feed coal.
The authors highlight that the continuous and timely collection and input of relevant
data is extremely important to maintaining the relevance of the CATM database. Also
they point out that industrial participation as well as participation from other research
organisations and government agencies will greatly benefit the implementation of the
database. The authors also put forward that there are many practical applications for
the database, which include:
· Investigating the collection and removal efficiencies of control devices
· The comparison of source emissions, analytical results from bench and
full-scale operations
· Investigating the usefulness of various sorbents
· The analysis of the size distribution of ash produced from a coal at varying
run configurations
· The effect of plant performance as a function of fuel type and quality
However, no references are cited for these applications.
Aside from the CATM database, other databases maintained by the Centre for Air and
Toxic Metals are:
Coal Ash Properties Database (CAPD) – Contains chemical, physical, and
mineralogical characterisation on more than 800 coal fly ash samples for use by the
ash utilisation industry.
Filter Ash Systems Database – Contains engineering data and analytical
characterisation of samples collected from pressurised fluidised bed combustor and
integrated gasification combined cycle systems.
The EERC also maintains a range of information and computing technologies. Such
technologies are the predictive models that have been developed to enhance
fundamental research and aid clients in making operation decisions without the high
cost of test work.
The models maintained by the EERC are:
ATRAN – Predicts entrained ash particle-size and composition distribution for a
conventional pulverised coal combustion system
TraceTran – Utilised to predict the physical state of trace elements during coal
gasification. Jak (1999) has covered the properties and implementation of TraceTran.
LEADER – Predicts the potential formation of problematic deposits in the lower-
temperature regions of a commercial boiler
Fouler – Dynamically predicts the formation of deposits within the convective pass of
a commercial boiler.
32
PCQUEST – Ranks various operational effects for a set of potential coals.
FACT – The Facility for Analysis of Chemical Thermodynamics model is a
comprehensive thermodynamic model that performs chemical equilibrium
calculations for systems consisting of solid, liquid, gaseous and/or aqueous phases.
The development of a comprehensive trace metal thermochemical database is a main
priority at the EERC. The database is the result of collaboration among the EERC, the
University of Connecticut and the Institute for Kemiteknik (Denmark) and the EPA.
The EERC is currently using the FACT model to quantify chemical equilibrium in
combustion and gasification processes. Although it is recognised that a stable
equilibrium is seldom achieved in such dynamic processes, an equilibrium analysis
can provide useful information on speciation under given reaction conditions of
temperature, pressure and composition. The FACT model began as a metallurgical
tool, thus it contains a considerable amount of data on trace metals (Vit Kuhnel,
1998).
The EERC chose the FACT model because it included the most comprehensive
thermodynamic database. The FACT database contains more than 6500 species that
can be modified and expanded to accommodate new thermodynamic data. Optimised
thermodynamic solutions have also been added by the EERC for a number of
nonstoichiometric ash minerals. The FACT code uses this data to determine the
multiphase, multi-component composition of up to 24 reactants by minimising the
total content of Gibbs free energy in the system, while also maintaining a mass
balance of the elements in the system. The resulting equilibrium composition can
contain up to 550 product species distributed either in the gas phase, the pure solid or
liquid phase, the aqueous phase, or a mixed phase (Folkedahl et al, 1999).
In 1998, at the EERC, Vit Kuhnel undertook an examination of the potential and
limitations of the FACT model to trace metal speciation. Vit Kuhnel states that a
simulation begins with definition of the reaction system, that is, all reactants in the
reaction system, whether flue gas, fly ash or sorbent as well as the physical conditions
of the theoretical reactor, are identified. Reactants are entered into the model in their
elemental form, such as an ultimate coal analysis, or in their actual state, such as a gas
composition in terms of partial pressures. Physical conditions are also included, such
as pressure and temperature, as well as any removal of discrete phases from an open
system.
Vit Kuhnel highlights that special attention must be given when using data beyond
existing temperature ranges. The extrapolation of data beyond a temperature range
may be plausible in some cases, however, species that do not disappear outside their
data range may disrupt stable solutions and give erroneous results. The FACT code
contains a facility that warns the user of any type of extrapolation and allows for the
deletion of product species that may lead to erroneous results.
Any species without thermodynamic data in the database will also affect the
modelling results. Vit Kuhnel cites examples of such omissions, in particular
thermodynamic data on chromium (oxy)hydroxides has been reviewed and data
33
entered for these gas phases. The resulting speciation illustrated that CrO2 (OH)2 is
the last gas species to appear upon cooling from 1800 K to 1400 K. At temperatures
below 1400 K, equilibrium calculations indicate that the majority of chromium is in
the solid phase. However, lack of information on chromium (oxy)hydroxides from the
database will place the temperature for solid-phase precipitation at 1800 K.
Similarly, for lead speciation the dominant gas species below 1200 K are lead
chlorides. Most coal combustion simulations are not concerned with chlorine, which
is often left out of coal analyses. However, if chlorine is left out of trace metal
speciation calculations, the equilibrium partitioning will be greatly affected (Vit
Kuhnel, 1998).
3.6 THERMODYNAMIC DATABASES
Jak (1999) has extensively reviewed the application of a thermodynamic equilibrium
approach to trace element speciation in combustion processes for the Black Coal
CRC. The review provides a comprehensive insight into the basis and use of the
thermodynamic equilibrium predictions. The review covers such topics as the
importance of completeness and self-consistency of phases in the prediction process,
as well as an excellent literature review of the FACT model and other thermodynamic
databases dealing with trace elements. Thus, only a summation will be given here as
to the other thermodynamic databases used for trace element predictions.
3.6.1 GFEDBASE Thermodynamic Database
Maintained by the Technical University of Denmark, the GFEDBASE
thermodynamic database is used in conjunction with the MINGTSYS Gibbs Free
Energy minimisation software. The GFEDBASE thermodynamic database contains
reduced data on the Gibbs Free Energy for approximately 800 chemical species of the
elements Al, As, B, Be, Br, C, Ca, Cd, Cl, Co, Cr, F, Fe, Ga, Ge, H, Hg, K, Mg, N,
Na, Ni, O, P, Pb, S, Sb, Se, Si, Sn, Ti, V and Zn. This predictive combination has
been developed and used by Frandsen et al (1994) to clarify the equilibrium chemistry
of trace elements in oxidative as well as reductive flue gases.
The system uses a global equilibrium analysis approach (GEA). The reason for this is
that many of the reactions in coal combustion are believed to be kinetically controlled,
this is the reaction rates are dependent on such factors as temperature and pressure.
However, with the exception of mercury, no kinetic data is available for trace
elements in hot flue gases. Thus the GEA is used to pinpoint, within a combustion
system, the trace element’s thermodynamic stable chemical and physical species as a
function of temperature, pressure and total composition. When the total Gibbs Free
Energy of the system is at a minimum, the system is considered in thermodynamic
equilibrium. No local conditions of pressure and temperature are considered. The
system is at its final state when all homogeneous and heterogeneous reactions have
reached equilibrium (Frandsen et al, 1994).
Frandsen et al (1994) point out that the global equilibrium analysis approach (GEA)
has several limitations when applied to combustion or gasification systems. Those
limitations are:
34
1. In the furnace zone, where the temperature is high (T>1500 K in a pulverised-
coal system), the reaction rates may be high enough to reach equilibrium,
provided that there is sufficient residence time.
2. In low temperature (T<800 K) zones of the combustion system around an
electrostatic precipitator, the reaction rate may be too low to reach
equilibrium, even at long residence times.
3. In the furnace zone, mixing phenomena may introduce local conditions that
are not taken into account in the GEA.
4. All relevant chemical species occurring in the real combustion system must be
taken into account or the output will be erroneous.
5. The presence of particulate material such as fly ash in the system is not
considered. Some of the gaseous components may adsorb onto the surface of
fly ash particles, eg mercury and selenium.
6. The curvature of small particles, which increase the partial pressure of the
reactant species, hence affecting the condensational behaviour.
Despite the limitations listed above, the authors state that the GEA approach of
pinpointing stable phases and assuming global equilibrium is the only computational
possibility for generating information on the chemistry of trace elements in a
combustion and gasification system. The kinetic limitation of the system, especially at
low temperatures, can be overcome if chemical kinetic data of trace elements in flue
gases is made available to the predictive system.
3.6.2 SOLGASMIX, ChemSage and Other Thermochemical Databanks
SOLGASMIX is one of the most widespread computer programs for chemical
equilibrium calculations (Eriksson & Hack, 1990). Covered in Jak’s review (1999)
with respect to its implementation for the modelling of trace elements, SOLGASMIX
has one main deficiency, and that is the user has to implement the proper
parameterised activity-composition relationships separately for any non-ideal mixture
phases. To overcome this limitation, Gibbs energy minimisation routines of
SOLGASMIX were augmented with an extensive library of subroutines of the most
frequently utilised models for non-ideal integral molar Gibbs energies and a refined
user interface. The modified code was ultimately renamed to ChemSage.
ChemSage is an extensive computer program designed to perform three types of
thermochemical calculations in complex systems involving phases exhibiting non-
ideal mixing properties. These calculations are:
· Thermodynamic functions
· Heterogeneous phase equilibria
· Steady-state conditions for the simulations of simple multistage reactors
The thermodynamic function module calculates specific heat, enthalpy, entropy and
Gibbs energy for a given reference state and phase. Chemical equilibrium calculations
can be made for a system that has been uniquely defined with respect to temperature,
pressure or volume and composition. The extensive library for the calculations of
integral and partial excess molar Gibbs energies enables a user to properly treat
systems containing dilute or concentrated, solid or liquid alloys, molten salts, liquid
oxides, non-dilute aqueous phases and non-ideal gases (Eriksson & Hack, 1990).
35
Figure 6 shows the structure sheet with the major blocks of ChemSage.
MAIN
INPUT OF THERMODYNAMIC DATA
Thermodynamic Phase Equilibrium STAGED REACTOR
Functions Calculations MODEL
Reads interactive
Reads interactive input Reads reactor flow
input file
Prints out and stores
Prints out and stores tabular results Prints out and stores
tabular results tabular results
MODELS
1. Polynomial models for simple substitutional and associated solutions
and several sub-lattice formations
2. Metallic dilute and aqueous solutions based on the unified interaction
parameter and Pitzer formalisms, and non-ideal gas models based on
the virial equation
3. Blander-Pelton quasi-chemical formalism and Gaye-Kapoor-Frohberg
cell formalism for ionic liquid mixtures
FIGURE 6 Structure sheet with the major blocks of ChemSage (Eriksson & Hack,
1990)
ChemSage modules permit:
· calculation of thermodynamic properties of single solution and
stoichiometric condensed phases with respect to a chosen reference state
36
· calculation of the chemical equilibrium state of a system that is defined
with regard to temperature, pressure or volume, and total amounts and/or
equilibrium activities of any phase constituent in the system
· calculation of temperatures when precipitates are formed from the liquid,
of adiabatic temperatures, etc.
· simulation of a multi-stage reactor by defining energy and material flows
between stages
· optimisation of thermochemical data based on experimental information
· results to be saved as plot-files, displayed and printed graphically, or
exported to other applications.
The new version of ChemSage (V4) will add more major features to this list,
including the automatic calculation of phase diagrams, new models and data from the
field of geochemistry and geophysics, and the calculation of stoichiometric reactions.
ChemSage is an application program that does not contain any thermodynamic data,
however it can be integrated into any integrated thermochemical database (ITD) such
as THERDAS (Eriksson & Hack, 1990). The software can also be integrated with:
· The SGTE Pure Substance and Solution databases
· A range of other thermochemical databases
· A large number of ready-to-use data-files for specific areas of application
· ‘Customised’ data-files prepared by data specialists to meet your specific
data requirement.
Bale and Eriksson (1990) have extensively reviewed other integrated thermochemical
databases (ITD). A list of the principal metallurgical ITD accessible to the public are:
3.6.3 CSIRO Thermochemistry System (version V)
Initiated in 1975 in collaboration with the National Physical Laboratory NPL (UK),
the CSIRO thermochemistry system is divided into several application programs.
· FILER: used for the fitting of data, creation and management of the
databases as well as index generation.
· CHEMIX: which is a SOLGASMIX-based minimisation program with
built-in activity coefficient correlations for multiple non-ideal phases and
process simulations.
· SYSTEM: diagram generation
· REACT: third law analysis of reactions mass and heat balances
· MODEL: staged modelling of reactors and processes
· EXERGY: second law analysis of process efficiency
· VAPOUR: vapour pressure and critical data
· ESTIMA: estimation of thermodynamic data
The databases accessible to the system include the CSIRO database on over 2000
compounds, the Scientific Group Thermodata Europe (SGTE) database on over 2300
compounds, JANAF with 1500 compounds, NBS with 1400 compounds, NPL with
1700 and vapour pressure databases with 1400 compounds.
37
3.6.4 HSC Software and Database
The HSC software version 3.2 and database was initiated in 1981 and developed by
Outokumpu Research Centre in Finland. The software is menu driven and integrated
with only a pure substance database with information on thermodynamic data such as
enthalpy, entropy and heat capacity. In equilibrium calculations, the activity
coefficients can be assigned values as needed.
HSC Chemistry is designed for various kinds of chemical reactions and equilibria
calculations. The current version contains twelve calculation modules displayed as
twelve options in the HSC main menu:
1. Reaction Equations
2. Heat and Material Balances
3. Equilibrium Compositions
4. Electrochemical Equilibria
5. Formula Weights
6. Eh - pH - Diagrams
7. H, S, C and G Diagrams
8. Phase Stability Diagrams
9. Mineralogy Iterations
10. Composition Conversions
11. Elements
12. Units
The name of the program is based on the feature that all twelve calculation options
automatically utilise the same extensive thermochemical database which contains
enthalpy (H), entropy (S) and heat capacity (C) data for more than 15000 chemical
compounds.
HSC Chemistry offers powerful calculation methods for studying the effects of
different variables on the chemical system at equilibrium.
For example, if the user gives the raw materials, amounts and other conditions of
almost any chemical process, the program will give the amounts of the product as a
result. HSC also makes heat and material balance calculations of different processes
much more easily than any manual method. The Eh-pH-diagrams option of HSC also
offers a very fast way of studying the dissolution and corrosion behaviour of different
materials.
HSC does not solve all chemical problems, because it does not take into account the
kinetics (rates) of the chemical reactions and non-ideality of solutions. However, in
many cases it is a very inexpensive and useful tool that helps to find the optimum
reaction conditions and yields for experimental investigations without expensive trial-
and-error chemistry.
The main thermodynamic calculations available are:
· Reaction Equations: enthalpy, entropy, Gibbs energy
38
· Heat Balances: heat and material balances for industrial processes
· Equilibrium Compositions: theoretical phase compositions and graphical
display
3.6.5 MANLABS Thermochemical Database (USA)
Developed in collaboration with CALPHAD members, the MANLABS
thermochemical database offers extensive binary and multiphase diagram programs,
access to the Thermo-Calc programs as well as the CALPHAD alloy database, oxides
database and the Thermo-Calc solutions database.
3.6.6 MTDATA Thermochemical Database (UK)
The Metallurgical and Thermochemical Data Services is a member of the SGTE and
was developed in 1971 at the National Physical Laboratory in collaboration with
AKAEA, Harwell laboratories.
MTDATA is an integrated package with several levels of operation depending on user
needs. The interface gives access to the in-built modules as follows:
· ACCESS: for recovering data from databases
· THERMOTAB: for the tabulation of substances and equations
· MULTIPHASE: energy minimisation for multi-component multiphase
equilibrium
· UNARY: functions for unary systems
· G-PLOT: functions for solutions
· BINARY: binary phase diagrams
· TRANSITION: phase boundaries and crystallisation paths in multi-
component systems
· TERNARY: ternary phase diagrams
The databases utilised by the MTDATA package include the SGTE pure substance
database of stoichiometric compounds, solution database with 100 binary, 50 ternary
systems as well as AQDATA dilute aqueous systems database.
The Department of Trade and Industry in the UK undertook a study on the application
of the MTDATA thermodynamic database to various coal conversion processes
(Department of Trade and Industry UK, Coal R&D program, PS040, 1995). The
researchers concluded from the study that MTDATA is of considerable value in
predicting equilibrium concentrations for various components in the coal conversion
process. In the case of coal combustion, the project focused on three stages:
· The high temperature capture of SO2 by CaO and MgO
· The release of Na and K during pressurised fluidised bed combustion
(PFBC)
· The release of trace elements during circulating PFBC
39
Certain important considerations however needed to be taken into account when
applying the database in modelling; these are:
· Limitations of the catalytic method for reducing NH3 to levels below those
directed by chemical equilibrium
· A direct relationship between Ca:S ratios and equilibrium concentration of
SO2 and SO3
· The limited effect of MgO as a sorbent under gasification and pressurised
combustion conditions
· The limited concentration of the vapour phase NaCl and KCl
· The pre-eminence of atmospheric N2 in NOx formation during fuel gas
combustion
3.6.7 THERMO-CALC Thermochemical Database (Sweden)
A THERMOchemical databank for equilibria and phase diagram CALCulations
(Thermo-Calc) is a member of the SGTE and was initiated in 1974. Developed at the
Royal Institute of Technology in Stockholm, the system consists of over 600
subroutines divided into modules. The central module is called POLY-3 and is used
for all equilibrium calculations and phase diagrams. It has the same potential for
binary, ternary and multi-component phases as other packages listed here. The other
modules in the system add support to the POLY-3 module.
The databases utilised by Thermo-Calc are SGTE pure substance and solution alloy
databases. Also available is FEBASE database of Fe-alloys in the range of 700 to
1200°C and the KAUFMAN database with 100 binary alloys. SLAG and ISHIDA
database for semiconductors systems is readily accessible.
3.6.8 THERDAS Thermochemical Database (Germany)
The Thermochemical databank for inorganic substances, THERDAS, is a member of
SGTE. It was developed in collaboration with NPL and Thermodata. THERDAS can
be implemented through four key programs:
· THERDYN: utilised for the calculations of substance properties
· REACTION: utilised for the calculation of standard reaction properties
activity correlation and electrochemical equilibria.
· PHASEDIA: prodimance-area diagrams
· ChemSage, as mentioned earlier, is a more advanced implementation of
SOLGASMIX for complex systems, co- and counter current reaction
calculations and property calculations.
The databases utilised by THERDAS are SGTE pure substance and solution alloy
databases, as well as an extensive dilute solution database.
3.6.9 THERMODATA Thermochemical Database (France)
Initiated and developed in 1974 by the Association Thermodata, Domaine
Universitaire de Grenoble, the THERMOchemical DATAbase on inorganic
thermochemistry, THERMODATA, is also a member of the SGTE group.
40
The four fully integrated services and databases that the system provides are
accessible by the multilingual user interface. The programs and databases comprise of
the following:
· THERMDOC: bibliographic database with access to 30,000 selected
references
· THERMOCOMP: bank of data covering inorganic and metallurgical
thermodynamics
· THERMALLOY: bank of data covering thermochemical properties of
alloys and multi-component systems
· THERMOSALT: bank of data covering thermochemical properties of
molten salts
3.6.10 Lewis Research Centre, NASA (USA)
This database has been covered by Jak (1999). Briefly, this database was developed at
NASA Lewis Research Centre as part of the jet propulsion program. The database and
package are readily available to the public and contain information useful to the
deportment of trace elements.
3.7 OVERVIEW OF THERMODYNAMIC DATABASES
Blander and coworkers (1997) conducted a round robin comparison to evaluate the
major computer thermodynamic packages. The authors concluded from their
investigation that “FACT and ChemSage appear to have largely produced credible
predictions for condensed phases … for the purpose of identifying gas-phase
speciation of ash forming species. All programs appear adequate.”
Research investigating the applicability of thermodynamic packages for trace element
partitioning can be found in papers by Frandsen and co-workers (1996, 1997 and
1999). The comparisons highlighted that similarities and differences in results for
some trace elements can be attributed to the differences in the thermodynamic
databases utilised by the packages. The authors put forward that for a thermodynamic
package to be of any use, the following must be observed:
· All chemical species occurring in the thermal fuel conversion system must
be taken into account, otherwise the output is meaningless
· Consistent thermodynamic data must be used
· Appropriate mixing models should be applied in the condensed phase
· Local equilibria due to boiler design and mixing phenomena (Jak, 1999)
Based on the essential criteria for correct thermodynamic predictions of completeness
and self-consistency, Jak (1999) concluded “only the FACT package from the
databases examined for comparison at the present stage contains non-ideal solution
models necessary to describe complex condensed solution phases. Moreover, Fact
oxide, matte and salt databases are very advanced. It is believed that GFEDBASE,
ABO and NASA Lewis packages do not have condensed solutions thermodynamic
models of major phases of interest to the coal ash incorporating trace elements.”
41
4 CONCLUSIONS AND RECOMMENDATIONS
Based on the information supplied in this review, it is clear that there are many factors
to be considered in understanding, modelling, predicting and controlling the emission
of potentially hazardous airborne trace elements from coal-fired power plants. The
physical and chemical processes that govern the transformation and partitioning of the
trace elements originating from coal are highly complex. This complexity stems from
the heterogenous nature of coal as well as the variability in composition and
concentration of the associated trace elements.
As can be seen from the emission-related modelling research presented here, only a
general understanding of the transformations and partitioning of trace element
processes as a function of the fuel composition and combustion system design
parameters have been gleaned. It is a strongly held belief that only through a
coordinated theoretical and experimental (laboratory, pilot and full-scale) effort to
supply the mechanistic, thermodynamic and chemical kinetic data for the model that a
reliable method of predicting trace element behaviour for a specified coal and boiler
configuration will be achieved. This in particular applies to modes of occurrence of
the trace elements in the coal. However, one of the main obstacles facing this type of
integrated model is the combining of databases and models as those listed above with
standard engineering models to produce a strong predictive tool (Davidson & Clarke,
1996 and Ratafia-Brown, 1993).
Keeping in mind that a high level of integration of models is needed to produce an
accurate predictive tool, this author shares the opinion with Eugene Jak (1999) that
the thermodynamic approach, with appropriate kinetic constraints, should be central
to the development of a model to predict the fate of trace elements during thermal fuel
power generation.
The recommendations put forward by Eugene Jak (1999) provide a very strong
foundation for the development of a predictive model for trace element deportment,
however, only with the following modifications:
1. Focus on two or three trace elements or species of greatest concern to the
Australian power generation industry, for example As, Cd, Se and Hg, for
a given power plant and assess the capabilities and accuracy for a couple
of modelling packages, for example FACT and ChemSage.
2. For the selected elements, review the thermodynamic data and extend for
optimisation as required.
3. Examine the completeness and self-consistency of the chosen database
with up-to-date thermodynamic information and modify as necessary.
4. Enhance the capabilities of the predictive modelling package through a
greater understanding of the thermodynamic and kinetics based on
laboratory scale experiments.
42
5. Enhance the capabilities of the predictive modelling package through the
utilisation and integration of standard engineering models.
From the tour of US research facilities, as outlined in Appendix 3, the following key
conclusions can be drawn in relation to studies on trace element release during
combustion:
· The elemental form of trace species has an influence on the final
partitioning
· For the more volatile elements seem to vaporise independently of mode of
occurrence and the char particle temperature determines the degree of
vaporisation. Under reducing conditions however, some volatile elements
are significantly less volatile.
· More refractory elements show a greater dependence of mode of
occurrence and char particle temperature on vaporisation.
· Some post combustion zone transformations have been identified which
have an impact on model development. The interaction of Ca with As, SE
and Cd dictates that for coals which contain significant CaO the
partitioning behaviour of these elements is controlled by surface reactions
rather than condensation.
· Thermodynamic calculations indicate that the reactions of mercury in the
flue gas are frozen at 800K, the chlorine content of the flue gas is predicted
to have the largest impact on mercury gas phase speciation. Furthermore
there appears to be adsorption of gaseous mercury on char and ash at low
temperature.
The following recommendations are made:
Coal Selection
Because of the importance of the elemental form on partitioning a wide range of
elemental forms of the trace elements of interest be selected for the study.
Analytical Techniques
CCSEM, Mössbauer, XAFS, Microprobe
Bench Scale Studies
The effect of coal type and combustion conditions should be carried out in a laminar
flow drop tube furnace. A range of oxygen stoichiometries and temperatures could be
utilised. Samples could be collected using a number of captive sample and
continuous flow techniques. Captive samples would utilise cyclones and aerosol
filters, whist continuous flow would include low pressure cascade impactor, aerosol
particle sizers and scanning mobility particle sizers. Resulting ash samples would be
43
analysed for chemical composition. These studies would be used to develop a model
for vaporisation.
5 REFERENCES
· Abbott, M., Douglas, R., Fink, C., Deluliis, N. and Baxter, L. (1994) Engineering
Foundation Conference on the Impact of Ash Deposition on Coal-fired Plants,
Solihull UK, pp165-176.
· Bale, C. W. and Eriksson, G. (1990) Canadian Metallurgical Quarterly; Vol. 29
(2), 105-132.
· Barton, R. G., Clark, W. D. and Seeker, W. R. (1990) Combustion Science and
Technology, Vol. 74, 327-342.
· Benson, S., Erickson, T. A., Zygarlicke, C. J., O’Keefe, C., Katrinak, K. A., Allen,
S. E., Hassett, D. J. and Hauserman, Wm. B. (1995) In: Fuel Cells ’95 Review
Meeting, Morgantown, WV, USA.
· Benson, S.A., Steadman, E.N., Mehta, A.K. and Schmidt, C.E. (eds) (1994) 19-22
April 1993, Special Issue, Fuel Process. Technol., Vol. 39.
· Blander, M., Milne, T., Dayton, D., Backman, R., Blake, D., Kuhnel, V., Linal,
W., Nordin, A. and Ljung, A. (1997) In: Impact of Mineral Impurities in Solid
Fuel Combustion, Kona, Hawaii, USA.
· Bool III, L. E. and Helble, J. J. (1995) Energy and Fuels; Vol. 9, 880-887.
· Bool III, L. E., Senior, C. L., Huggins, F., Wendt, J. O. L., Sarofim, A. F. and
Finkelman, R. (1996) Quarterly report DOE/PC/95101—T5.
· Calvert, (1972) In: Rizeq, R. G., Hansell, D. W. and Seeker, W. R. (1994) Fuel
Processing Technology; Vol. 39, 1/3, 219-236.
· Chadwick, M.J., Highton, N.H. and Lindman, N. (1987) In: Environmental
impacts of coal mining and utilisation. National Energy Administration Sweden.
Pergammon Press, Oxford, UK, pp 171-217. Cited in Clarke and Sloss (1992).
· Chow, W., Miller, M.J. and Torrens, I.M. (1994) Fuel Processing Technology,
Vol. 39, 5-20.
· Clarke, L. B. and Sloss, L. L (1992) IEACR/49. Coal Research, London.
· Clemens, A.H., Damiano, L.F., Gong, D. and Matheson, T.W. (1999) Fuel, Vol.
78, 1379-1385.
· Cramer, H. (1986) VGB Kraftwerkstechnik, Vol. 66 (8), 648-651.
· Curtis (1999) Proceedings of TraceElements99, University of Warwick, UK, 9th
September, 1999, IEA, London.
· Dale, L.S., Chapman, J.F., Buchanan, S.J., Lavrencic, S.A. (1999) Project 4.2,
Final Report to Co-operative Research Centre for Black Coal Utilisation, CSIRO
Energy Technology.
· Davidson, R. M. and Clarke, B. L. (1996) IEAPER/21
· Deluliis, N. J. and Tumati, P. R. (1995) In: 3rd International Conference on
Managing Hazardous and Particulate Air Pollutants, Toronto, ON, Canada.
· Deluliis, N. J., Tumati, P. R. and Fink, C. E. (1993) In: Proceedings: 10th Particle
Control Symposium and 5th International Conference on Electrostatic
Precipitation. Washington, DC USA, pp. 13.1-13.14.
44
· Dennis, (1977) In: Rizeq, R. G., Hansell, D. W. and Seeker, W. R. (1994) Fuel
Processing Technology; Vol. 39, 1/3, 219-236.
· Department of Trade and Industry U.K., Coal R&D Program, PP203, 1999
· Department of Trade and Industry U.K., Coal R&D Program, PP216, 1999
· Department of Trade and Industry U.K., Coal R&D Program, PS040, 1995
· Devito, M., Rosendale, L. and Jackson, B. (1994) In: 56th Annual American
Power Conference. Chicago, Il, USA, pp. 438-443.
· DeVito, M.S. and Rosenhoover, W.A. (1999) Proceedings of TraceElements99,
University of Warwick, UK, 9th September, 1999, IEA, London.
· EPA web site (1999) www.epa.gov
· Eriksson, G. and Hack, K. (1990) Metallurgical Transactions B, Vol. 21B, 1013.
· Fahlke, J. (1993) VGB Kraftwerkstechnik; Vol. 73, 216-218.
· Faulkner and DuBard, (1984) In: Rizeq, R. G., Hansell, D. W. and Seeker, W. R.
(1994) Fuel Processing Technology, Vol. 39, 1/3, 219-236.
· Finkelman, R.B. (1994) Fuel Processing Technology, Vol. 39, 21-34.
· Folkedahl et al (1999) www.eerc.und.nodak.edu/catm/ and related documents
· Frandsen and co-workers (1999) In: Jak, E. (1999) Collaborative Research Centre
Black Coal Utilisation: Environmental Issues Program Area 4: Project 4.7.
· Frandsen, F., Dam-Johansen, K. and Rasmussen, P. (1994) Progress in Energy and
Combustion Science, Vol. 20 (2), 115-138.
· Frandsen, F., Erickson, T. A., Kuhnel, V., Helble, J. J. and Linak, W. P. (1996) In;
Proceedings of the 3rd International Symposium on Gas Cleaning at High
Temperatures, pp 462-471.
· Frandsen, F., Helble, J. J., Erickson, T. A. and Kuhnel, V., (1997) In Proceedings
of EPRI-DOE-EPA combined Utility Air Pollutant Control Symposium.
Washington, DC.
· Gutberlet, H., Spiesberger, A., Kastner, Tembrink, J. (1992) VGB
Kraftwerktechnik, Vol. 72 (7), 636-640 (in German) cited in Meij (1994).
· Huffman, G.P., Huggins, F.E., Shah, N., Zhao, J. (1994) Fuel Processing
Technology, Vol. 39, 45-62.
· Hughes, I. S. C. and Beer, A. D. (1994) British Coal Corporation, Coal Research
Establishment, Report No. COAL R031, 1994.
· Jak, E. (1999) Collaborative Research Centre Black Coal Utilisation:
Environmental Issues Program Area 4: Project 4.7.
· Kuhnel, V. (1998) www.eerc.und.nodak.edu/catm/catm3/tech1.html
· Leith and Licht, (1972) In: Rizeq, R. G., Hansell, D. W. and Seeker, W. R. (1994)
Fuel Processing Technology, Vol. 39, 1/3, 219-236.
· Linak, W. P. and Wendt, J. O. L. (1994) Fuel Processing Technology, Vol. 39,
1/3, 173-189.
· Lyyranen, J., Jokiniemi, J., Mojtahedi, W. and Koskinen, J. (1995) Journal of
Aerosol Science, Vol. 26, S687-S688.
· Mathews, A. P. (1988) In: Hazardous Waste: Detection, Control, Treatment edited
by Abbou, R. Elsevier Science Publications, Amsterdam, The Netherlands.
· Meij, R (1992), The fate of trace elements at coal-fired power plants, KEMA
Report 32561-MOC 92-3641, KEMA, Arnhem, The Netherlands.
· Meij, R (1994), Trace element behaviour in coal-fired power plants, Fuel
Processing Technology, Vol. 39, 199-217.
45
· Meij, R (1999) Proceedings of TraceElements99, University of Warwick, UK, 9th
September, 1999, IEA, London.
· Meij, R (1999a) Proceedings of CEM99, International Conference on Emissions
Monitoring, University of Warwick, UK, 6th-8th September, 1999, IEA, London.
· Meij, R., Janssen, L.H.J.M. and van der Kooij, J., (1986) Kema Scientific &
Technical Reports, Vol. 4 (6), 51-69.
· Meij, R., van der Kooij, J., van der Sluys, J.L.G., Siepman, F.G.C. and van der
Sloot, H.A., (1984) Kema Scientific & Technical Reports Vol. 2 (1), 1-8.
· MTI public relations information 1999: http://www.rrv.net/microbeam/cqms.htm
· Oman, C. L. and Finkelman, R. B. (1994) In: Proceedings of the Eleventh Annual
International Pittsburgh Coal Conference, University of Pittsburgh, pp1096-1099.
· Pires, M., Fiedler, H. and Teixeira, E. C. (1997) Fuel, Vol. 76 (14/15), 1425-1437.
· Pourkashanian, M., Williams, A., Jones, J. M. and Sykes, J. (1998) ETSU for the
Department of Trade and Industry, Report No. COAL R146, 1998.
· Ratafia-Brown, J. A. (1994) Fuel Processing Technology, Vol. 39, 1/3, 139-157.
· Rizeq, R. G., Hansell, D. W. and Seeker, W. R. (1994) Fuel Processing
Technology, Vol. 39, 1/3, 219-236.
· Rubin et al, (1991) In: Rizeq, R. G., Hansell, D. W. and Seeker, W. R. (1994) Fuel
Processing Technology, Vol. 39, 1/3, 219-236.
· Rudnick etal. (1986) In: Rizeq, R. G., Hansell, D. W. and Seeker, W. R. (1994)
Fuel Processing Technology, Vol. 39, 1/3, 219-236.
· Seeker, W. R., Rizeq, R. G., England, G. (1995) In: Effects of Coal Quality on
Power Plants: Fourth International Conference Proceedings, Palo Alto, CA, USA,
Electric Power Research Institute.
· Sloss, L. (1992), Halogen emissions from coal combustion, IEACR/45, IEA Coal
Research, London.
· Smith, I. (1987), Trace elements from coal combustion, IEACR/01, IEA Coal
Research, London.
· Stadler, S. P., Shea, S. C., Quinn, A. and Murarka, I. (1994) In: 11th Annual
International Pittsburgh Coal Conference, pp 310-315.
· Swaine, D.J. (1990), Trace Elements in Coal, Butterworth & Co. Ltd, London,
278p.
· Tumati, P. R. and Bilonick, R. A. (1996) Journal of Air and Waste Management,
Vol. 46, 59-65.
· Tumati, P. R., Douglas, R. E. and Abbot, M. (1994) Engineering Foundation
Conference on Coal-blending and Switching of Low Sulfur Western Coals:
Snowtown, UT USA, pp 149-167.
· Yokoyama, T., Asakura, K. and Seki, T. (1991) Komae Research Laboratory
Report No. ET 91002, Komae, Japan, CRIEPI, cited in Clarke and Sloss (1992).
46
APPENDIX 1
Comparison of KEMA Model with CSIRO Data on Distribution
of Trace Elements in Power Station Streams
Element Meij 1999 CSIRO a CSIRO b CSIRO c
flyash b'ash vapour flyash b'ash vapour flyash b'ash vapour flyash b'ash vapour
% distribution % distribution % distribution % distribution
As 95 1 82.7 1.8 <0.1 94.9 3.6 0.1 94.0 1.6 0.4
B 47 5 10 57.4 2.6 34.7 17.6 1.7 62.4 55.1 1.5 34.5
Be 92 8 92.5 10.0 <0.1 101.7 8.6 <0.1 109.5 11.0 <0.1
Cd 95 1 79.2 2.5 1.8 57.3 2.2 1.0 105.0 4.7 <0.1
Cl <0.1 <0.1 71.0 <0.1 <0.1 88.5 <0.1 <0.1 90.8
Co 91 9 87.7 8.1 <0.1 89.7 12.1 <0.1 96.4 8.9 <0.1
Cr 88 11 99.4 9.9 0.4 111.2 12.9 <0.1 116.6 10.9 <0.1
Cu 93 6 92.6 7.6 <0.1 89.7 10.5 <0.1 104.2 10.3 <0.1
F 15 1 12.7 0.3 31.7 12.8 0.6 39.0 23.6 1.3 50.1
Hg 49 1 25 54.3 <0.1 20.1 45.6 <0.1 59.9 61.3 <0.1 nd
Mn 82 11 87.1 14.8 <0.1 ? 32.2 <0.1 78.5 16.9 <0.1
Mo 91 6 111.9 6.1 <0.1 79.6 2.1 <0.1 102.3 3.7 0.4
Ni 91 9 88.2 11.5 0.2 83.0 12.7 <0.1 95.2 9.6 <0.1
Pb 91 5 92.0 4.9 0.1 83.1 3.4 <0.1 115.8 6.3 <0.1
Se 76 2 5 37.9 0.2 3.8 35.8 0.3 0.7 91.6 0.4 0.4
Zn 92 6 93.2 6.1 0.2 80.2 2.4 <0.1 68.8 4.1 <0.1
U 92 7 105.4 11.0 <0.1 85.3 11.0 <0.1 95.8 12.3 <0.1
Th 88 12 86.0 5.6 <0.1 86.5 13.0 <0.1 100.0 12.7 <0.1
“Meij (1999)” are the percentage mass balances according to the KEMA model in a
hypothetical power station with flyash:bottom ash ratio of 88:12. The power station is
equipped with FGD thus the percentages of trace elements quoted in the emitted
vapour (stack gas) are lower than in power station not equipped with scrubbers. Note
especially the value of 25% for Hg; this is equivalent to 50% Hg emissions without
the scrubber.
“CSIRO a” are actual percentage mass balances obtained on power station A. The
flyash:bottom ash ratio is 85:15. The station is equipped with ESP for dust collection.
“CSIRO b” and “CSIRO c” show the actual data obtained on power stations B and C.
The flyash:bottom ash ratio is 90:10 in both stations. The stations are equipped with
ESP for dust collection.
47
APPENDIX 2
Report on the CEM99 and TraceElements99 Conferences
University of Warwick, 6-9 September, 1999
Much of the proceedings of the two conferences directly relate to the CRC project on
trace element deportment in power stations.
CEM99
The conference was held over 3 days (6-8 September). There were five sessions:
· legislation,
· measurement techniques,
· the monitoring of suspended particulate matter,
· calibration/data handling and
· case studies.
Much of the session on measurement techniques consisted of instrument
manufacturers describing advances in monitors used in industry for the continuous
monitoring of emissions.
The author was surprised by the activity within Europe on the efforts being made to
monitor emissions from industry. It appears that there is a requirement by the EU of
individual nations to adopt legislation that makes it mandatory to report such
emissions.
The CD-ROM from this conference is available and if copyright restrictions do allow
the printing of hard copies of the proceedings, Ken Riley can arrange to have copies
prepared for those who are interested.
Conference Outline
Welcoming Address
David King, Regional General Manager, Environment Agency
King stated that the Environment Agency is one of the largest and most powerful
environmental protection agencies in Europe. The primary aim of the Agency is to
protect and improve the environment throughout England and Wales and to contribute
to sustainable development through the integrated management of air, land and water.
In his opinion, the key issues were: addressing climate change, regulating major
industries, improving air quality, managing waste, managing water resources,
delivering integrated river-basin management, conserving the land, managing
freshwater fisheries and enhancing bio-diversity.
King made the point that the EU drove 85% of UK legislation. A system of integrated
permits with ELVs (emission limit values) to be based on best available techniques
48
was being introduced. The jargon statement that there were “opportunities for the
environment, industry and public” was made a number of times. It is evident that the
UK is serious about addressing environmental problems. Pollution prevention and
control bills will be implemented in October 1999. The Agency is also establishing a
certification scheme for continuous emissions monitoring (MCERTS).
It is obvious that much activity in emissions testing or monitoring is directly driven by
legislation. It is interesting that in the UK, the Agency is both enforcing the
regulations and licensing the monitoring systems.
Our past and our future
Walter Smith, Walter Smith Associates, USA.
This was an entertaining and deliberately controversial presentation. Smith discussed
the history of measuring stack emissions. He made the comment that, “ We are hung
up on our electronic gadgets and pretend they will solve our sampling problems.”
A number of examples were given:
· The EPA Method 7E, which uses chemiluminescent analysers and
cold traps, was described as a good compromise in the 1980s when
the majority of the NOx was NO. But with gas turbines with permit
levels less then 10 ppmv @ 7% O2, the problem is that most of the
NOx is now NO2 and there is also NH3 in the gas stream. He
observed that the analysers require the reduction of the NO2 to NO;
the reduction is slow and NH3 interferes with this process.
· The need for continuous emission monitoring systems and
instrumental methods for mercury and dioxins/furans. But the
aerosol part of the emission and its spatial distribution cannot be
ignored. It is also necessary to sample iso-kinetically and traverse
the stack.
He stated that accreditation programs should certify both the organisation and
individual and are not a substitute for audits, co-location and quality checks. The need
for a demonstration of precision (by co-location of instrumental techniques with
manual procedures) was stressed.
The author is not familiar with the operation of chemiluminescent analysers and does
not know whether the limitation referred to by Smith is a significant problem.
Extension of the Agency’s Monitoring Certification Scheme MCERTS to
include Continuous Ambient Air Quality Monitoring Instruments
M. Lewandowski, S. Newstead and J. Tipping, National Compliance Assessment
Service, Environment Agency
The Environment Agency launched its Monitoring Certification Scheme MCERTS for
continuous emission monitoring systems (CEMs) for industrial stacks in 1988. The
agency plans to broaden the scheme later to include other monitoring activities. In this
49
paper presented by Newstead, the proposed extension of MCERTS to cover
continuous ambient air quality monitoring instruments was described.
Within MCERTS, performance standards for instruments have been specified for a
common range of pollutants. It was claimed that the provision of MCERTS
certification for ambient monitoring instruments would have significant benefits for
instrument users, manufacturers and regulatory authorities. The proposed extension to
MCERTS covers continuous ambient monitoring systems (CAMs). The proposed
extension includes performance standards for the measurement of the following
substances: nitrogen monoxide, nitrogen dioxide, sulphur dioxide, carbon monoxide,
ozone, particulate matter (PM10 and PM2.5), lead (Pb) and other metals (cadmium,
arsenic, nickel and mercury), benzene, and poly-aromatic-hydrocarbons (PAHs).
CAMs would be assessed against performance standards appropriate for their
intended applications (rural and remote sites, urban background/centre, suburban,
curbside, roadside and industrial sites).
The proposed standards have been defined so that an instrument which meets the
MCERTS requirements would be capable of meeting the requirements both of the EC
Framework Directive 96/62 EC “Ambient Air Quality Assessment and Management”.
Extension of the Agency's monitoring certification scheme MCERTS to
include continuous ambient air quality monitoring instruments
John Tipping, Environment Agency, UK.
This paper was a description of the impact of the Agency’s regulations on the
installation and operation of real time monitors. The emphasis was on the necessity of
monitoring equipment being “fit for purpose”. Performance standards were being
prepared for NOx, SOx, PAHs and metals. Factors being considered are response
time, repeatability, detection limits and comparison with reference instruments.
It is apparent that contractors not holding certification or manufacturers of
instruments not certified under the MCERTS scheme will be seriously disadvantaged.
It is similar to the NATA system for analytical laboratories but the author assumes the
Agency will require all data generated for their use to come from certified operators
using certified instrumentation.
Sample conditioning system issues relating to emissions monitoring
from combustion processes
David Graham, Alstom Energy Technology Centre, UK
Graham responded to Smith (see comment above). Graham also highlighted some of
the problems involved in the sampling of stack gases, eg. NO is not unreactive and
NO2 is soluble, SO2 is only sparingly soluble and SO3 is soluble; the sampling of the
soluble gases required an awareness of the dew point. Graham also commented that it
was obvious that sample composition must be unaltered during sampling. Peltier
coolers and heated lines may be necessary. He emphasised the necessity to correctly
calculate water and acid dew points.
50
The Big Three - on-line monitoring of Greenhouse gases
Simon Bruce, Servomex Group Ltd, UK
Bruce discussed the problems in the continuous monitoring of some of the greenhouse
gases using IR-based instruments.
Continuous monitoring of the performances of a NOx reduction system
using electrochemical sensors
Derek Stuart, Land Combustion, UK
This was a study of the difficulties to be overcome in measuring NOx required to
monitor the performance of the SCR (selective catalytic reduction) process. Major
problems being the high temperature, high SO2, high dust, irregular flow patterns and
NH4HSO4 formation in the sample probe. Problems were apparently overcome by the
use of dual systems to allow recovery; separate NO and NO2 sensors (ie. no
converter); chemical filter to remove SO2. The instrumentation should be ideal for
monitoring gas turbines, as little or no SO2 is present.
New low gas absorption, electronic water condenser for use in gas
analyser sample conditioning systems featuring the "m class" electronic
water condenser
Tom Baldwin, Baldwin Environmental, USA
Baldwin described his development of electronic water condensers. The aim of this
apparatus is to remove water from stack gases as quickly as possible so as to prevent
the loss of soluble analyte species from the sampled gas. Baldwin stated that
thermoelectric chillers were mandated in Californian legislation relating to gas
sampling. A comparison of the adsorption characteristics of different materials was
made, ie. adsorption of gases such as NO2 and SO2 to apparatus walls. Baldwin stated
that stainless steel was often the worst material and that glass was often the best but it
is of course too fragile to be used in many applications. A Teflon-like polymer,
Kynar, was recommended.
Hydrocarbon monitoring and instrumentation methods
Tibor Bernath, Bernath Atomic GmbH, Germany
Bernath described his design of a new jet for use in FID detectors in GCs used for
hydrocarbon monitoring. Discussion on different standard gases and the importance of
matrix matching were discussed. The problem of low and variable oxygen levels in
stack gases was mentioned and the effect on accuracy in determining hydrocarbons.
Continuous monitoring of hazardous smokestack metal emissions using
inductively coupled argon plasma atomic emission spectroscopy
Michael Selzer, Naval Air Warfare Centre Weapons Division, USA
Selzer described an ICP-AES system used for the “continuous” monitoring of the
emission of trace elements during combustion. The system samples the gas stream
every 60 seconds. It consists of a heated transfer probe, a vacuum eductor, a sub-
sampling device connected to a conventional Thermo Jarrell Ash Trace 61E
51
spectrometer. Problems with calibration were discussed. This requires ultrasonic
nebulisation to form a dry aerosol that is diluted with ambient air. This instrument was
designed to monitor furnaces to destroy explosives at military installations.
Instruments have been installed at waste combustors.
It is apparent that sensitivity would be a problem in applications such as power
stations where the levels of emitted trace elements are extremely low.
New developments in long-term sampling of dioxins
Christian Philipp, ESI GmbH, Germany (presented by Johannes Mayer)
Mayer presented this paper on the topic of the ongoing demand for dioxin sampling
particularly long-term sampling. The EU requires three samplings per year of 6-8 h
duration on specified plants. In Flanders and Belgium, continuous dioxin sampling
will be required with analysis every two weeks by 2000. He indicated that there was
no on-line techniques available and described a technique of automatic sampling and
adsorption onto XAD resin. The sampling device includes a RAM card (data logger)
which is removed and sent to the laboratory with the resin cartridges.
Dioxin sampling - a long term view
Mike Smurthwaite, Westech, UK
Smurthwaite described WESTED’s sampling system which complies with EN 1948.
A quartz-lined titanium probe (iso-kinetic) linked to a sampling train containing twin
cartridges of CXAD-2 resin on a polyurethane support. Problems with sampling were
discussed, eg. location of probe and condensation. Smurthwaite indicated that the
adsorbed dioxin compounds were retained on the resin for periods of months at
ambient temperatures and the dioxins could be measured after this interval. The
system is designed to allow sampling of the stack for a period of one month before
analysis.
Perspect UV - an open path ambient gas monitor for industry
Clive Lee, Sci-Tec Kipp and Zonen, UK
This was a description of the company’s long path length (1000 m) UV spectrometer
to measure gaseous pollutants (non-range resolved and reported as a path integrated
concentration).
US EPA's evaluation and applicability determination process for
particulate matter continuous emission monitoring systems
Dan Bivins, US EPA, USA
Bivins reported on the appraisal of particulate matter continuous emissions
monitoring systems (PM CEMS) with a view towards their application in future
national regulations.
The US Clean Air Act requires the EPA to write standards to limit emissions of
hazardous air pollutants (HAPS) from major industrial sources.
52
This paper was a summary of the expansive EPA report on PM CEMS. Bivins
touched on the variability of results, the requirement for QA/QC and the cost of
instrumentation.
This was a long presentation on the relative merits of different instrument types and
their suitability to accurately monitor the emissions of particulate matter.
Experiences with calibration of dust AMS according to VDI 2066 and
prEN 13284
Lars Johansen, dk-Teknik; Denmark
Johansen of dk TEKNIK (an independent Danish research laboratory) compared
various methods for determining the particulate matter concentrations in gas stream.
The reported results indicate the variability between methods. This must be of
concern to legislators.
Monitoring stack gas emissions of volatile trace elements, sulphur
dioxide and PM10 particulates from coal-fired stoker systems
R Whitney, CRL Energy, New Zealand
This paper was similar to that given by Clemensen at the last Australian Coal Science
Conference. It detailed the methods used in estimating the emissions from plants (in
NZ) burning high calcium sub-bituminous coals. Whitney reported that the approach
used by CRL Energy in their test work gave good recoveries for most trace elements,
except Se.
The author was interested to learn that boron leached from coal ash is of concern in
New Zealand, particularly to the Kiwi fruit industry.
Comparison of a networked multi-head AC coupled triboelectric digital
dust transmitter system to other traditional dust monitoring systems
Jeff Hayhurst, Goyen Controls, UK
This was another paper on advances in instrumentation to meet the requirements of
legislation. In this paper, the components of an AC coupled networked triboelectric
dust ‘transmitter’ system was compared with alternative system, ie. traditional optical
instrumentation and DC coupled triboelectric systems. The Goya instrumentation was
described as offering an economical solution in those installations where a number of
emission locations have to be monitored.
Using new electrodynamic cross-correlation technology for flow
measurement in stacks
William Averdieck, PCME Ltd, UK
Averdieck reported on the importance of flow measurement which is necessary
(obviously) to calculate total mass emissions from stack gas concentration data. The
electrodynamic cross-correlation technique was described and its performance
compared with the traditional approaches, pitot, ultrasonic transit time and thermal
53
mass. It was stated that these techniques could be unreliable where high levels of
particulate or moisture occur.
Characterisation of utility boiler contributions to PM2.5
Glenn England, EERC, USA
England presented an interesting paper. He stated that although emissions from coal-
fired boilers may contribute significantly to ambient particulate concentrations,
speciation studies often show that geological matter, which makes up the majority of
fly ash produced by coal combustion, makes up only a small fraction of the fine
particulate material. Sulfates, nitrates and organic compounds often comprise more
than 80 % of the ambient PM2.5. Secondary aerosol precursors (NOx, SOx, NH3, and
volatile organic compounds) contribute significantly to ambient PM2.5. England
stressed that speciation of primary particles emitted from coal-fired boilers provides
important markers that can be used to identify the relative contributions of specific
categories of sources to ambient PM2.5. He reported on the development of a test
protocol for estimating PM2.5 emission factors.
Development of a real-time stack particulate mass monitor
Jim Mills, ETI Group Ltd, UK
Gas standards relevant to stack emission measurements at the National
Physical Laboratory
P Holland, NPL, UK
This presentation was a description of the role of the National Physical Laboratory
(NPL) which produces and maintains primary standards of gas mixtures necessary for
accurate monitoring of stack emissions. The standards are prepared by absolute
gravimetric methods and are used to certify secondary standards to be used as
working standards by UK industry and government laboratories.
Centralised QA for CEM applications using AQM software
Andrew Newman, Signal Group Ltd., UK
This presentation was on Signal Group’s software designed to provide a system of
preparing compliance reports from different sources of test data.
The value of accurate calibration in emissions trading
Bob Davis, Scott Specialty Gases, UK
Davis presented examples of the need to have accurate calibrations and thus the need
for accurate calibration standards. An emission trading of SO2 was used to illustrate
the cost of inaccuracy.
54
Facilities for the testing of gas monitoring instruments
B Goody, NPL, UK 1520-1600
The facilities available at the National Physical Laboratory (NPL) for the laboratory
testing and calibration of gas monitoring instruments were described. Laboratory
testing to the requirements of the MCERTS is available.
Plant operator problems with CEM's - a common evaluation
Mike Yeoman, AEA Technology, UK
Yeoman (a consultant) discussed the problems that can occur when monitoring
equipment, supplied to the plant operator, does not meet the requirements of the
legislators or fails to work in the environment in which it is installed. Case studies
were given.
Specifications and quality measures for CEMs calibration curves
Wolfgang Jockel, TUV Rheinland, Germany
Jockel discussed the calibration and performance of emission monitors, ie. the use of
ISO Standards and approved tests for validation. The need to conduct long term field
tests under varying operating conditions was stressed.
Modern gas flow monitoring practices: a review of ultra-sonic systems
within CEMs applications
Volker Herrmann, Sick Engineering GmbH, Germany
Herrmann discussed the development by Sick Engineering of continuous flow
monitors necessary for the accurate determination of mass emissions, eg. particulate
matter or SOx.
It is obvious that the development of such instrumentation is driven by legislation.
Chlorine and hydrogen chloride monitoring utilising ion mobility
spectroscopy (IMS)
Kurt Webber, Molecular Analytics, UK
Webber reported on the company’s instrument, of which there are some 800 units in
the USA. The instrument is similar to a time of flight mass spectrometer. It uses a Ni63
source to produce singly charged ions of HF, HCl and Cl. Using a chiller, it is
possible to remove HCl and determine species of Cl.
Concentrations of gaseous trace elements such as mercury in the
ambient air of the Netherlands
Ruud Meij, KEMA, Netherlands
Meij reported on over 20 years of research at KEMA (a private company) into the fate
of trace elements following the combustion of coal in utilities in the Netherlands. This
presentation was generally restricted to mercury. Meij gave a paper on trace elements
in general in TraceElements99. Meij uses a carbon adsorber to collect the Hg in flue
gases. He discussed the pitfalls of sampling.
55
Field studies of combustion aerosols from Danish power stations
Dipl. Ing. Morten Thellefsen Nielsen, The Aerosol Laboratory, Denmark
Nielsen reported on studies completed at the Avedore power station. This plant burns
coal, straw/wood, oil emulsion (70% Venezuelan crude oil/30% water). He made the
observation that although the coarse fraction of the ash comprises 95% of the mass,
the 5% fines contains 90% of the particles. In the reported study, sampling locations
were before the ESP, the FGD and at the exit of the stack. Iso-kinetic sampling was
used with a low impact cascade impactor. EDAX was used for analysis.
Receptor modelling of air quality data using neural networks
John Grubert, East London University, UK
This was a most interesting talk. Grubert described the use of Neural networks in
preference to factor analysis. The application used as an example was a study around
Birmingham done over 70 days. The neural network was used to identify the sources
of the pollutants.
On-line analysis of copper foundry's stack gas composition by low
resolution FT-IR gas analyser
Petri Jaakkola, Temet Instruments Oy, Finland
This was a discussion of the matrix effects, which can affect the FT-IR gas analyser
used in a butane-fired copper foundry.
Air pollution in Deva-Hunedoara industrial area
Matei, Popescu and Sorim, University of Petrosani, Romania
Mining activity impact on air quality in Jiu Valley, Romania
Matei, Popescu and Sorim, University of Petrosani, Romania
These presentations were combined. Sorim reported on environmental studies in what
appears to be a very polluted area in Romania. The aim of the study was to identify
and resolve the major sources of pollution in the area, which contains mines, power
plants, coke plants and steel production.
CO2 -Monitoring in German industry: target achievement in 1997
Hans Georg Buttermann, RWI Essen, Germany
This was report on Germany efforts to obtain its target of CO2 reduction and the
monitoring that will be done to confirm this
TraceElements99
TraceElements99 was held on the 9th September. What is evident from the papers is
the amount of work that has been completed overseas on the emissions of trace
elements, in particular, the studies completed in the USA and the Netherlands. Meij
(KEMA) has written a significant amount on the distribution of trace elements during
combustion of coals sourced worldwide (including Australia).
56
Legislation on mercury and other trace element emissions
Lesley Sloss, IEA Coal Research, UK
Sloss gave a summary of the legislation pertaining to trace element emissions from
combustion plants. The emphasis was on coal-fired power plants but reference was
made to waste incinerators. Much of the paper was related to mercury. Worldwide,
with the exception of USA, there appears to be little concern over the emission of
mercury from coal-fired power stations, although the comment was made that Poland
has a “mercury problem” (high sulfur coals containing pyrite and associated
chalcophillic elements including mercury). There was much discussion following this
presentation. DeVito queried whether there was any information on mercury
emissions from China. There was a general consensus that coal firing was not a
significant source of mercury; Meij asked whether Hg in natural gas was a concern –
Chung (EPRI) replied that it was very low.
The author has not looked into this but recalls that significant Hg may be found in
natural gas.
Measurement of the speciated mercury emissions from coal-fired power
plants with the Ontario Hydro Method: successes, problems and pitfalls
Keith Curtis, Ontario Hydro, Canada
Curtis described the development of his method to determine the species of Hg in the
stack gas of power stations. This in Curtis’ words is “not rocket science”. The
procedure requires a trap in the EPA gas sampling train to capture soluble mercury
(oxidised). In studies on two plants, Curtis found in one plant that 65% of Hg was
oxidised, 33% elemental and 2% particulate; in the other plant, 40% was oxidised,
36% elemental and 24% particulate. This latter plant was producing an ESP ash with a
large amount of unburnt carbon. The difficulties encountered in obtaining reliable
data were discussed, eg. unstable permanganate solutions, loss of Hg on high C
particulate matter on filters.
The discussion following this presentation was interesting. In particular, the problems
of sampling and analysis and the general findings of others, in particular Meij and
DeVito, on the fate of mercury – the amounts of chloride in the coal and unburned
carbon in the ash are important.
On-line determination of heavy metals using chalcogenide glass
sensors in a flow-injection set-up
John Mortensen, Roskilde University Institute, Denmark
This was very much an academic presentation. Mortensen described the development
of sensors (ion selective electrodes) based on sulfide chemistry. These can be used to
estimate ions in solution.
The electrodes are still being developed and it is probably some years before these
will find wide application.
57
Development of a mobile analyser for the continuous determination of
mercury in the form of Hg(0) and Hg in flue gas
Marton Klein, L & C Steinmuller GmbH, Germany
Klein reported on the development of a continuous system for determining the
speciation of mercury occurring in flue gas. The instrument was tested on a waste
incineration plant where the transportable instrument was used to measure both Hg (0)
and Hg (II) at various locations in the ducting of the plant. The instrument utilises
conventional borohydride chemistry and atomic absorption spectrometry and has a
detection limit of 1 mg/m3.
Flue gas Hg measurements from coal-fired boilers equipped with wet
scrubbers
Matt Devito, Consol, USA
This was a comprehensive report on the measurement of mercury emissions at six
power stations in the USA. This was part of an EPA initiative (the aim is to collect
data on Hg emissions, including identification of species from 100 plants and the
concentration of Hg in the coals fired in over 1000 plants).
Devito reported on a number of factors that affect the emission of mercury. All of the
plants tested were equipped with FGD. On average, there was 54% removal of Hg
across the FGD; including what is adsorbed onto the flyash, there was 67% removal.
Of the oxidised Hg (as determined by Curtis’ procedure), 80-90% was removed in the
wet scrubber. There was no correlation of the proportion of oxidised Hg produced
with the chlorine concentration in the coal (it was indicated by Curtis and others,
during this conference, that the amount of Cl present was a factor in the formation of
oxidised Hg), although all coals had a similar concentration of Cl, approximately
0.1%. The relative amount of Hg (II) produced was related inversely to the oxygen
content of the coal (?). There was no correlation of Hg removal with S removal in the
FGD. The Hg in the FGD residues was neither leachable nor volatile at 1400 F.
Devito commented that the effect of coal quality or the species of Hg in the coal on
Hg emissions was speculative. Chu (EPRI) stated that methods used in the industry
were good for determining total Hg but results for Hg speciation were variable.
It appears that the coal industry and the utilities are concerned about the EPA
initiative on Hg emissions; EPA has identified coal combustion as the largest single
source contributor of anthropogenic Hg emissions in the USA. EPRI is undertaking a
study on the measurement of Hg concentrations. The EPA, DOE and EPRI are
undertaking an extensive study of emissions from coal-fired power stations.
Trace element emissions from a coal-fired power plant in Japan -
Legislation, emissions, environmental control technologies
Takahisa Yokoyama, CRIEPI, Tokyo, Japan
Yokoyama reported on the situation relating to trace emissions in Japan. A few facts
are interesting: only 10% of electricity (20000 MW) is generated in coal-fired plants.
Most of the other 90% is generated approximately equally by oil, natural gas, hydro
and nuclear means. Most of the coal-fired plants are PF systems (an IGCC system,
58
200 t /day was dismantled in 1997). Most (70%) are equipped with SCR and DeSOx
systems. Japan wide, there are 100 “types” of coal burnt; up to 30 “types” in each
power station.
The generators are required to monitor the emissions of mercury (this element is of
particular interest in Japan) and other trace metals such as arsenic, chromium,
cadmium, lead, beryllium, uranium, thorium, as well as the same elements in the to-
be-fired-coals. Yokoyama discussed the effect of plant design (temperatures at ESP,
DeSOx, DeNOx) on trace element emissions.
Mass balance studies of trace elements at coal-fired power plants
including co-combustion of waste and biomass
Ruud Meij, KEMA, The Netherlands
KEMA have extensively studied the trace element emissions from coal-fired power
stations. Meij has a number of papers in the English literature; there are also many
reports in Dutch (I am not sure whether these are readily available). The work
reported is impressive.
The concentrations and distributions of the elements (particularly the volatile
elements: As, B, Br, Cl, F, Hg, I, Se) in the coal feed, bottom ash, flyash, flue gases
up- and downstream of the FGD plant of coal-fired power plants in the Netherlands
were studied. The aim of the research was to establish the relationship between the
elements in the different process streams. It was found that most of the volatile
species are removed in wet FGD, ie.B, Br, Cl and I are largely removed and more
than half of the F, Hg and Se is removed. Hg is largely removed when a combination
of a SCR-installation and wet-FGD is present. F is largely removed in the absence of
a gas-gas reheater at the FGD. in ash. KEMA uses the data from these tests predict the
concentrations of trace elements in the ash, the leachate and the flue gases. Meij
reported that the relationships were also valid for co-combustion with less than 10%
waste and biomass.
KEMA has produced a model by which it is possible to predict the elemental
composition of all the outgoing streams of a coal-fired power plant.
Round table discussion
The formal presentations were followed by a round table discussion of the issues
relating to trace element monitoring and deportment of trace elements during and the
fate of trace elements after combustion.
Woodford (AEA Technology, UK) was concerned about the demands being made by
legislators and the need to use costly instrumentation. He also mentioned that standard
methods also require more elaborate techniques.
This is a fact but precise and sensitive techniques are required.
Chu (EPRI, USA) described the EPRI project on the investigations into method used
for Hg and Cl.
59
CSIRO was acknowledged for its contribution. Riley asked the origin of the need to
determine nickel carbonyl and nickel (II) subsulfide as required by Australia’s NPI;
the existence of these species was regarded as extremely unlikely in coal-fired power
stations.
He also asked whether Cr (VI) is produced in the ash. The general response was that
some 10% of the Cr may be in this valency state. This response was based on the
work of Huggins.
Gibb (Powergen, UK) presented work on the adsorption of Hg onto unburned carbon
in flyash. The question was posed as to whether poor combustion was a realistic
control technology. There was a general discussion on the factors effecting the
emission of Hg and whether Hg from coal-fired power stations was a legitimate
concern. Meij (KEMA) was of the opinion that testing for Hg emissions at a given
power station need only be done once. Chu (EPRI) stated that the US EPA was likely
to pressure the utilities. England (USA) made the valid point that Hg was biologically
persistent although it was generally thought that emissions from power stations were
not a problem.
Devito summed up the proceedings and thanked the attendees.
Concluding Remarks
The opportunity to attend these two conferences was dependent upon funding from
the CRC for Black Coal Utilisation. This funding is gratefully acknowledged.
Much of the proceedings directly relate to the CRC project on trace element
deportment in power stations. The extensive data generated by KEMA will be
invaluable. Whether the large database being compiled by EPRI can be assessed is
another matter. The cost of the report is approximately $US 25000. The information
likely to be contained in the report is relevant but the results will be exclusively on US
power stations probably equipped with FGD and burning US coals many of which
contain high sulfur.
60
APPENDIX 3
US Research in Trace Element Measurement
The focal point of the visits to US research establishments was to get a synopsis of the
results from the DOE, EPA and EPRI sponsored research programs for trace element
measurements and modelling in combustion systems. The individuals and
organisations visited during this trip are summarised in Table 1.
Table 1: Summary of organisations with whom discussions were held over the period 3 – 12 Dec
1999.
Individuals Organisation Topics Discussed
Dr Larry Baxter Sandia National Fine Particles
Dr Peter Walsh Laboratories
Prof. Jost Wendt University of Arizona Experimental methods
Mr Wayne Smears Fine particles
Mr Sheldon Davis Condensation mechanisms
Secondary reactions
Heavy metal sorbents
Health effects of fine
particles
Prof. Adel Sarofim University of Utah Trace element toxicology
Dr JoAnn Lighty Trace element emissions
Dr John Vernath Modelling vaporisation
Ms Sheree Swenson and condensation
Mr Alejandro Malina
Dr Eric Eddings Reaction Engineering Vaporisation modelling
Ms Christina Lee International
Mr David Wong
Dr Steve Benson Centre For Air Toxic CATM
and co workers Metals Combustion 2000
Energy And
Environmental Research
Centre
University of North
Dakota
Dr Constance Senior Physical Sciences Institute Major findings of DOE
funded program managed
by PSI
Dr Andrew Miller US Environmental Fine particle
Dr William Linak Protection Agency measurements
Sandia National Laboratories (SNL)
The fine particle research at SNL is focused on diesel engine emissions. In this work,
Peter Walsh uses laser induced breakdown spectroscopy (LIBS) to study the
composition of fine particles. The technique employs a YAG laser focused on a point
in a gas stream that destroys small particles (<5 micron) and produces excited states
for the elements in the sample. Characteristic emissions for particular elements are
61
detected as in flame emission spectroscopy, and the peak area provides a basis to
calculate the mass of the element present. Limitations of this technique are that it is
difficult to detect 2 species simultaneously and whether or not the results are
statistically significant in light of the relatively small area and small number of
particles analysed in the system. Ultimately the technique will provide a mass
distribution of small particles. Clearly this has potential to determine the composition
distribution of the fine particles in a gas stream.
University of Arizona (UoA)
Part of PSI managed project where condensation mechanisms are being studied and
identified. Primary reactor used is a 150kW, 6m, 140mm nominal diameter down
fired furnace. There is no electrical heating so there is a temperature profile
established down the length of the reactor. The flame temperature is around 1100°C,
with a residence time of approximately 6 seconds. Sampling is conducted at the base
of the flame at about 98% conversion and again near the base of the reactor to monitor
changes in the characteristics of the inorganic vapours. The collection system
comprises a Berner low pressure cascade impactor (as they state it blocks less than the
Anderson model). They manufactured their own impactor, which behaves the same
as the purchased one, however they have not calibrated any of their impactors.
Collection times for the particles are approximately 20 mins for fine particles in the
LPI and 3 – 5 minutes in the cyclone for coarse fractions.
From this apparatus, this group has identified the main As species in some cases to be
calcium arsenates. The implication is that some incorrect assumptions by some
research groups may have been made about the species present and selective leaching
procedures developed may be misleading. They will send all of their appropriate
publications, both published and in press. They have some data that shows how fine
particles destroy lung tissue in mice. Work is being conducted with NEDO and the
University of Stuttgart, which is concentrating on coal and sewerage sludge co-firing
issues.
Other interesting work in progress is related to the removal of metal emissions from
hot flue gas streams. CaO seems to absorb (albeit weakly) all As and Se as long as Ca
levels are high enough. They have extracted some kinetic data. Kaolinite seems to
absorb Cd, it is still a mobile form but at least it is not emitted. They seem to think
the concrete forming reactions will lock the trace elements up. The next phase of the
project aims to develop a multi metal capture system which utilises the kaolinite based
sludge from paper recycling, which is calcined. They seem to think anionic sites are
capturing the metals, and the sorbents only work in combustion systems.
Existing condensation models which may be commercially available are MAEROES
II and CONTAIN, which are based on predictions for nuclear accidents, developed by
Fred Gelbard from SNL, also Axelbomb from Washington State and Peter McMurray
from TSI have also developed computer codes.
Reaction Engineering International (REI)
REI are reanalysing the data of Quann and Sarofim 1990 (18 Symp. Int. Comb.) to
extract and model vaporisation of refractory oxides. The hypothesis is that trace
62
element emissions do not nucleate but heterogeneously condense onto condensed
refractory components, which would obviously undergo nucleating at much higher
temperatures. In particular, they are modelling the vaporisation of Al, Si, Ca, Mg and
Fe. Model is CFD based, looking at a 500MW boiler. Assumption is all ash particles
are of uniform composition and each coal particle contains ash as per proximate
analysis and particle size varies. Vaporisation from the model is dominated by
inclusion size and coal type (mineral form or particle burning temperature).
Bituminous coals appear to have a lower emission than other coals. There are no
comparisons of emissions with thermodynamic equilibrium calculations. Results of
vaporisation of the refractory components from model and experiments agree.
University of Utah (UU)
Some interesting work is being conducted here on modes of occurrence of trace or
vaporised species on human health. Some results to date which, while not statistically
significant yet, show that FeO in glass is very toxic to human lung cells, due to its
high mobility in amorphous phases, for particle sizes in the respirable range, while Fe
in the form of oxides is not. This group has provided a number of publications, which
postulate that not only particle size and mass are important but chemical composition.
The importance of this work is that an increase in mobile group II transition elements
in a given cell triggers a specific response and when these are lung cells, a whole body
response is triggered. At present the drop tube furnace facilities are not ready for the
laboratory studies.
Centre For Air Toxics Emissions (CATM) / University of North Dakota
(UoND)
The EERC appears to be becoming a private company in the business of technology
application so they are not so keen to discuss the details of their processes. The
USEPA has initiated an information collection request to around 75 power stations of
various sizes and configurations to establish a database of Hg emissions from these
plants. The choice of plant was somewhat selective in so far as a range of boiler types
and pollution control configurations have been selected. The data required is Btu, S,
Hg, Cl, in feed and Hg emissions. EERC is trying to become the central collection
point. Steve Benson is convinced that Hg concentrations are absolutely correlated
with pyrite levels. A source of data here is the work of Finkleman from USGS, which
is discussed, with the work of Connie Senior. Analytical procedures and
understanding have improved drastically in the past years and the USGS database is
being updated.
An interesting point is the political direction in the US presently. A report of the EPA
to congress states that they are not concerned with the toxicity of any trace element in
coal with the exception of Hg. So at present the EPA is only concerned with Hg
emissions and fine particulate matter (fine PM)
Not too much was provided in terms of data for trace element emissions. In the
TraceTran model, they use thermodynamic equilibrium to determine trace element
vaporisation, however they do apply a couple of correction factors to account for
differences in vaporisation from measurements and theory. This would account for
the mechanisms for trace element release such as absorption of some trace species by
63
silica and incomplete vaporisation found for non and intermediately volatile
components. They say they allow the refractory components to vaporise and allow
the trace species to condense onto these components. Fleming Frandsen generated the
condensation code. It was developed for gasification and may not be particularly
useful for combustion. The stable species are predicted from chemical equilibrium,
but as shown by Connie Senior at PSI, equilibrium is not necessarily the best
assumption for all conditions of temperature and pressure.
Physical Sciences Institute (PSI)
This was a very comprehensive discussion of the Phase 1 DOE funded project on
Toxic substances From Coal Combustion: a Comprehensive Assessment. The major
findings to date have been in relation to trace element mode of occurrence,
combustion zone transformations, post combustion zone transformations and
modelling. One interesting point was the observation of the non-equilibrium of
mercury gas phase species at temperatures less than 800K. This could have
significant impacts on the assumptions used in modelling trace element release and
subsequent speciation and deportment.
US Environmental Protection Agency (USEPA)
Whereas the DOE has a fuel based approach to fine particles and persistent bio-
accumulatable toxins (PBT), the EPA has an issue based research program. Major
issues at the moment are Hg emissions, fine particles, SOx and trans boundary
pollution. Mid west states were legislated to have a reduction in NOx of 75% and
eastern states were only required to have a reduction of 25% due to trans boundary
effects (related to ozone concentrations). This was recently taken to court due to the
economic impact this type of legislation would have on a state. Of interest were the
particle analysis systems. They use a combination of TSI aerodynamic particle sizer
and scanning mobility particle sizer to generate fine particle size distributions on line
in the range 0.01-20 microns. A Berner low pressure cascade impactor is used to
capture particle size bins for chemical analysis.
Conclusions and Recommendations
From the tour of US research facilities the following key conclusions can be drawn in
relation to studies on trace element release during combustion:
· The elemental form of trace species has an influence on the final
partitioning
· For the more volatile elements seem to vaporise independently of mode of
occurrence and the char particle temperature determines the degree of
vaporisation. Under reducing conditions however, some volatile elements
are significantly less volatile.
· More refractory elements show a greater dependence of mode of
occurrence and char particle temperature on vaporisation.
64
· Some post combustion zone transformations have been identified which
have an impact on model development. The interaction of Ca with As, SE
and Cd dictates that for coals which contain significant CaO the
partitioning behaviour of these elements is controlled by surface reactions
rather than condensation.
· Thermodynamic calculations indicate that the reactions of mercury in the
flue gas are frozen at 800K, the chlorine content of the flue gas is predicted
to have the largest impact on mercury gas phase speciation. Furthermore
there appears to be adsorption of gaseous mercury on char and ash at low
temperature.
The following recommendations are made:
Coal Selection
Because of the importance of the elemental form on partitioning a wide range of
elemental forms of the trace elements of interest be selected for the study.
Analytical Techniques
CCSEM, Mössbauer, XAFS, Microprobe
Bench Scale Studies
The effect of coal type and combustion conditions should be carried out in a laminar
flow drop tube furnace. A range of oxygen stoichiometries and temperatures could be
utilised. Samples could be collected using a number of captive sample and
continuous flow techniques. Captive samples would utilise cyclones and aerosol
filters, whist continuous flow would include low pressure cascade impactor, aerosol
particle sizers and scanning mobility particle sizers. Resulting ash samples would be
analysed for chemical composition. These studies would be used to develop a model
for vaporisation.
65
Related docs
