New Page 2010

New Page 2010
ES-10-105
CERTIFIED MAIL
RETURN RECEIPT REQUESTED
July 28,2010
Mr. Brian J Hug
Deputy Program Manager
Air Quality Planning Program
Air and Radiation Management Administration
Maryland Department of the Environment
1800 Washington Boulevard
Baltimore, MD 21230
Dear Mr. Hug:
Enclosed please find our BART five factor analysis for each BART-eligible unit at the Luke mill. Modeling
has demonstrated that the control (2001-2003 data) emissions result in a 3-year average eighth highest
delta deciview impact of 2.36dv at Shenandoah National Park (cumulative impact of the three BART
units). With the emission control levels proposed in the analysis, the deciview impact drops to 0.78 dv.
The emission control levels represent a 90% drop in S02 emissions from No. 25 Boiler and NOx controls,
which keep the NOx emissions to an annual average of 0.40 Ib.lMMBtu.
These conclusions are based on the five statutory factors required by the Clean Air Act and represent the
best BART controls. In a letter dated January 26, 2010, MDE supported these controls and levels by
proposing a formal regulation. We support the terms and conditions proposed in those regulations and
hope that they will be maintained as MDE moves forward with the BART program.
If you have any questions about the analysis, please contact me at (301) 359-3311, Extension 3262.
Sincerely.
onald E. Paugh
Asst. Environmental Manager
REP:plt
Enclosure
NewPage Corporation, Coated Paper
300 Pratt Street, Luke MD 21540 T 301 359 3311, ext. 3001 F 301 3592001
A:COM
Environment
Prepared for:
NewPage Corporation
NewPage Luke Mill
Luke, MD
Prepared by:
AECOM
Westford, MA
60158068-1
July 30,2010
Five Factor BART Analysis for NewPage
Luke Mill
AS'COM
Environment
Prepared for.
NewPage Corporation
NewPage Luke Mill
Luke. MD
Prepared by:
AECOM
Westford, MA
60158068-1
July 30,2010
Five Factor BART Analysis for NewPage
Luke Mill
eemantini Deshpande. Robert Hall, Olga Kostrova
Environment
AECOM
Contents
ES Executive Summary
ES·1
1.0 Introduction
1·1
2.0 Baseline Data
2·1
2.1
Overview of BART Emission Units
2-1
2.2
Current Control Technologies
2-1
2.3
Baseline Emissions
2.3.1
Power Boiler No. 25
2.3.2
Power Boiler No. 26
2.3.3
NO.3 Recovery Boiler
2-1
2-1
2-2
2-2
3.0 Emission Control Alternatives
3·1
3.1
Power Boiler No. 25
3.1.1
S02 Emission Controls
3.1.2
NOx Emission Controls
3.1.3
PM Emission Control
3-1
3-1
3-4
3-5
3.2
Power Boiler No. 26
3.2.1
S02 Emission Controls
3.2.2
NOx Emission Controls
3.2.3
PM Emission Control
3-6
3-6
3-6
3-7
3.3
NO.3 Recovery Boiler
3.3.1
S02 Emission Controls
3.3.2
NOx Emission Controls
3.3.3
PM Emission Control
3-7
3-7
3-8
3-9
4.0 CALPUFF Modeling Inputs and Procedures
4-1
4.1
Location of Source vs. Relevant Class I Areas
4-1
4.2
General Modeling Procedures
4-1
4.3
Model Version
4-1
4.4
Background Air Quality Data
4-1
4.5
Light Extinction and Haze Impact Calculations
4-1
5.0 CALPLIFF Modeling and BART Determination Results
5.1
Baseline CALPUFF Modeling Results
Five Factor BART Analysis lor NewPage Luke Mill
5·1
5-1
JUly 2010
Environment
AECOM
ii
5.2
Modeling Results for the BART Control Case
5-1
5.3
BART Results and Discussion
5-2
6.0 References
6-1
List of Appendices
Appendix A Particulate Emissions for Pulp and Paper Industry Specific Sources
Appendix B NCASI - Infonnation on Retrofit Control Measures for Kraft Pulp Mill Sources and Boilers
for NOx, S02 and PM Emissions
Appendix C Modeling Archive (CD attached)
List of Tables
Table 2-1
NewPage Luke Mill - Baseline Emissions for Power Boiler No. 25, Power Boiler No. 26
and No.3 Recovery Boiler
2-4
Table 2-2
NewPage Luke Mill - Baseline Stack Parameters for Power Boiler No. 25. Power Boiler
No. 26 and No.3 Recovery Boiler
2-4
Table 3-1
NewPage Luke Mill - Emissions Control Case
3-3
Table 3-2
NewPage Luke Mill - Stack Parameters Control Case
3-3
Table 4-1
References to the New IMPROVE Equation CALPOST Inputs
4-3
Table 5-1
Regional Haze Impacts Due to Baseline Emissions
5-1
Table 5-2
Regional Haze Impacts Due to the Future Controlled Emissions
5-2
List of Figures
Figure 4-1
Location of Class I Areas in Relation to the NewPage Luke Mill
FIVe Facto< BART Analysis lor NewPage Luke Mill
4-4
July 2010
AECOM
Environment
ES-1
Executive Summary
Federal regulations under Title 40 of the Code of Federal Regulations (CFR) Part 51 Appendix Y
provide guidance and regulatory authority for the application of Best Available Retrofit Technology
(BART) to those existing eligible sources in order to help meet the targets for visibility improvement at
designated Class I areas. The Maryland Department of the Environment (MOE) has identified the
coal-fired boiler (Power Power Boiler No. 25), the natural gas-fired boiler (Power Boiler No.26) as well
as the No.3 Recovery Boiler. at NewPage Luke Mill as BART-eligible emission units. The BART
rules require that sources that are subject to BART perform a site-specific BART analysis including a
control technology review and CALPUFF modeling to assess the visibility impact of the emission units.
This report documents the case-by-case BART analysis conducted for NOx, S02, and PM 10 emissions
from Power Boilers No. 25 and 26 and No.3 Recovery Boiler. This analysis addresses the five
statutory factors required by Section 169A (g) (7) of the Clean Air Act that states must consider in
making BART determinations:
(1) the costs of compliance,
(2) the energy and non-air quality environmental impacts of compliance,
(3) any existing pollution control technology in use at the source,
(4) the remaining useful life of the source, and
(5) the degree of improvement in visibility which may reasonably be anticipated to result from the
use of such technology
The follOWing emission scenarios were evaluated for the Luke Mill BART analysis:
\
1
•
Baseline Case (2001-2003 period) - Maximum daily emissi~ns of S02. NOx as well as
particulate matter for Power Boiler No. 25 were provided by NewPage. The maximum daily
heat inputs to Power Boiler No. 25 and 26 were also provided. Daily maximum emissions of
S02 and particulate matter were based on CEMS data and AP-42 emission factors. Maximum
daily black liquor solids (BLS) firing rate for No.3 Recovery Boiler was also provided by
NewPage and standard emission factors from the National Council for Air and Stream
Improvement (NCASI) were used to calculate emissions of S02, NOx and particulate matter.
•
Control Case - NOx• S02, and PM 10 emissions signature of Power Boiler No. 25 which
assumes a 90% reduction in S02 emissions from the baseline (via the installation of either a
Spray Dryer Absorber or a Circulating Dry Scrubber), reduction in NOx emissions from 0.99
Ib/MMBtu to 0.40 Ib/MMBtu (via year-round operation of the existing SNCR\ The particulate
matter emissions for Power Boiler No. 25 remain the same as the baseline as do the S02,
NOx and particulate matter emissions for Power Boiler No. 26 and No.3 Recovery Boiler.
According to Luke Mill's current TiUe V Operating Pennit, an SNCR was installed on Power Boiler No. 25 in the year 2006.
Fove Factor BART Analysis for NewPage Luke Mill
July 2010
,/
AECOM
Environment
ES-2
r---..
CALPUFF modeling of baseline emissions showed that Power Boiler No. 25, Power Boiler No. 26 and
No.3 Recovery Boiler are subject to BART based on a 3-year average eighth highest delta deciview
impact of 2.35 dv at Shenandoah National Park (cumulative impact of the three units). CALPUFF
modeling results show that substantial visibility improvement occurs with the implementation of
Control Case emission controls. For the control case, the 3-year average eighth highest delta
deciview impact at Shenandoah National Park is 0.78 dv.
Therefore, the recommended BART for Power Boiler No. 25 is the installation of add-on S02 controls
(either a Spray Dryer Absorber or a Circulating Dry Scrubber), year-round operation of the existing
SNCR for NOx control and multicyclones and baghouse for PM control. Burning natural gas, which
inherently has low nitrogen, sulfur and ash content, constitutes BART for Power Boiler No. 26. The
currently installed two level staged combustion air control system with ESPs constitutes BART for the
No.3 Recovery Boiler.
FIVe F8C1Ilr BART Analysis for NewPage Luke "'"
July 2010
Environment
AECOM
1.0
.#
1-1
Introduction
Federal regulations under Title 40 of the Code of Federal Regulations (CFR) Part 51 Appendix Y
provide guidance and regulatory authority for conducting a visibility impairment analysis for designated
eligible sources. The program requires the application of Best Available Retrofit Technology (BART)
to those existing eligible sources in order to help meet the targets for visibility improvement at
designated Class I areas. The BART analysis will be revieWed and used by the Maryland Department
of the Environment (MOE) for development of the state's Regional Haze State Implementation Plan
(SIP). The MOE has identified the coal-fired boiler (Power Power Boiler No. 25), the natural gas-fired
boiler (Power Boiler No.26) and the No.3 Recovery Boiler at the NewPage Luke Mill as BART-eligible
emission units.
The BART rules require that sources that are subject to BART perform a site-specific BART analysis
including a control technology review and CALPUFF modeling to assess the visibility impact of the
emission units.
The BART analysis was conducted in accordance with the procedures contained in the Final BART
Guidelines published by the USEPA on July 6,2005 (Federal Register Volume 70, No. 128).
Consistent with the BART Guidelines, the five steps for a case-by-case BART analysis were followed:
1. Step 1 - Identify all available control technologies for the affected units including
improvements to eXisting control equipment or installation of new add-on control equipment.
2. Step 2 - Eliminate technically infeasible options considering the commercial availability of the
technology, space constraints, operating problems and reliability, and adverse side effects on
the rest of the facility.
3. Step 3 - Evaluate the control effectiveness of the remaining technologies based on current
pollutant concentrations, flue gas properties and composition, control technology
performance, and other factors.
4. Step 4 - Evaluate the annual and incremental costs of each feasible option in accordance
with approved EPA methods, as well as the associated energy and non-air quality
environmental impacts.
5. Step 5 - Determine the visibility impairment associated with baseline emissions and the
visibility improvements provided by the control technologies considered in the engineering
analysis (Steps 1 - 4).
The baseline period for BART analysis as specified in 40 CFR 51 is 2001-2003.
r "..
The regUlation further requires a formal choice of BART based on the above data, plus the degree of
improvement in visibility (impacts), which may be reasonably anticipated to result from the installation
or implementation of the proposed BART. Economic analysis, remaining useful life of the plant, and
impacts on facility operation that are a cost consequence of air pollution control equipment may be
considered in the final BART decision-making process.
Five Fac:tor BART Analysis for NewPage Luke Mill
July 2010
AECOM
Environment
1-2
This report documents the case-by-case BART analysis conducted for S02. NOx and PM emissions
from Power Boiler No. 25 and 26, and the No.3 Recovery Boiler at the NewPage Luke Mill. Section
2.0 provides a description of the BART-eligible units and their baseline emissions. Section 3.0
provides a discussion of available S02, NOx and PM control technologies and improvements in
emissions of S02, NOx and PM. The available meteorological data and the CALPUFF modeling
procedures are described in Sections 4.0 and 5.0, respectively. The results of the visibility
improvement modeling using CALPUFF are also presented in Section 5.0, along with the BART
recommendation. References are listed in Section 6.0.
Five Factor ~T Analysis for NewPage Luka Mill
July 2010
Environment
AECOM
2.0
2.1
2-1
Baseline Data
Overview of BART Emission Units
The BART-affected emission units at the Luke Mill are the coal-fired boiler (Power Power Boiler No.
25), the natural gas-fired boiler (Power Boiler No.26) and the No.3 Recovery Boiler.
Power Boiler No. 25 - Power Boiler No. 25 bums coal as a primary fuel with natural gas used as a
secondary fuel. Built in 1965, this boiler has a nominal rating of 785 MMBtulhr. The boiler is used as a
backup system for incineration of emissions from non-condensible gas (NCG) and stripper off gas
(SOG) systems.
Power Boiler No. 26 - Power Boiler No. 26 was installed in 1970 and was converted to natural gas in
1982. With a nominal rating of 338 MMBtulhr, the boiler is also used as a backup system for
incineration of emissions from the NCG and SOG systems.
No.3 Recovery Boiler - The No.3 Recovery Boiler is used to recover chemicals from spent pulping
liquors and to produce steam for the mill. It fires black liquor as the primary fuel with NO.4 oil used for
startup purposes. Installed in 1969, this boiler has a nominal rating of 287,500 pounds of 50% black
liquor solids per hour.
2.2
Current Control Technologies
Power Boiler No. 25 - Power Boiler No. 25 has a multi-cyclone mechanical collector in series with a
baghouse for control of particulate matter. The boiler is also equipped with an over-fire air system,
10w-NOx burners and a selective non-catalytic reduction system for controlling NOx emissions.
Emissions from Power Boiler No. 25 exhaust into a single tall stack which serves as the common
emission point for the exhaust streams from Power Boiler Nos. 24, 25 and 26. The combined stack is
equipped with CEMS for NOx, SOx and flow, and a continuous opacity monitor.
Power Boiler No. 26 - There are currently no emissions controls installed on Power Boiler No. 26.
Emissions from Power Boiler No. 26 exhaust into a single tall stack which serves as the common
emission point for the exhaust streams from Power Boiler Nos. 24, 25 and 26. The combined stack is
equipped with CEMS for NOx, SOx and flow, and a continuous opacity monitor.
NO.3 Recovery Boiler - The No.3 Recovery Boiler has two level staged combustion air control
system for the control of S02 and NOx emissions. The boiler flue gases are routed through
electrostatic precipitators (ESP1, ESP2 and ESP3) for control of particulates.
2.3
Baseline Emissions
2.3.1
Power Boiler No. 25
Maximum daily S02 and NOx baseline emissions for Power Boiler No. 25 for the baseline period were
provided by NewPage. Maximum daily S02 emissions during the baseline period were estimated by
NewPage to be 40 tons per day. The maximum daily NOx emissions rate was calculated using an
emission factor of 0.99 Ib/MMBtu and a daily maximum heat input rate of 761 MMBtu/hr. Based on
the stack test conducted on the combined stack in April 2002. AECOM calculated a filterable PM
Five Factor BART Analysis lor N8wPllge luke Mill
JUly 2010
AECOM
Environment
2-2
emission rate of 1251blhr from the combined stack and apportioned it to Power Boiler No. 25 using a
scaling factor proportional to its rated heat input capacity. Therefore, baseline filterable PM emission
rate from the Power Boiler No. 25 was calculated to be 70.8 Iblhr.
Speciation of the particulate matter emissions from Power Boiler No. 25 into filterable and
condensable PM 10 components was conducted using the following approach:
•
Power Boiler No. 25 was equipped with multicyclone back in 2001; therefore, filterable PM
was subdivided by size category consistent with the default approach cited in AP-42, Table
1.1-6. For a coal-fired boiler equipped with a multicyclone, 29% of PM emissions are PM 10,
and 3% are fine PM 10 Le. PM 2.5.
•
For coal-fired boilers, elemental carbon is expected to be 3.7% of fine filterable PM 10 based
on the best estimate for electric utility coal combustion in Table 6 of "Catalog of Global
Emissions Inventories and Emission Inventory Tools for Black Carbon", William Battye and
Kathy Boyer, EPA Contract No. 68-0-98-046, January 2002.
•
Condensable PM 10 emissions from coal fired boilers are based on AP-42 Table 1.1-5. Total
condensable PM 10 = 0.01 S-o.03 Ib/MMBtu, where S is the % sulfur content of coal. Inorganic
condensable PM 10 is 80% and organic is 20% of total condensable PM 10. Inorganic fraction
of condensable PM 10 emissions is assumed to consist entirely of sulfates (i.e. no soil
component).
2.3.2
Power Boiler No. 26
Maximum daily NOx baseline emission rate of 741blhr was provided by NewPage. Maximum daily
S02 emission rate was calculated using the AP-42 emission factor given in Table 1.4-2 (0.0006
Ib/MMBtu) and the maximum daily heat input rate provided by NewPage. AECOM estimated the
filterable PM emission rate using the emission factor given in AP-42 Table 1.4-2 (1.9 Ib/MMcf) and the
maximum daily heat input to the boiler.
Speciation of the particulate matter emissions from Power Boiler No. 26 into filterable and
condensable PM 10 components was conducted using the following approach:
•
In accordance with AP-42 Section 1.4, for a natural gas fired boiler, 100% of all filterable PM
is PM 2.5 or smaller.
•
For natural gas-fired boilers, elemental carbon is expected to be 6.7% of fine filterable PM 10
based on the best estimate for natural gas combustion in Table 6 of "Catalog of Global
Emissions Inventories and Emission Inventory Tools for Black Carbon", William Battye and
Kathy Boyer, EPA Contract No. 68-0-98-046, January 2002.
•
Condensable PM 10 emissions from natural gas-fired boilers are based on AP-42 Table 1.4-2.
Total condensable PM 10 5.7IbIMMcf. Inorganic condensable PM lO is 50% and organic is
50% of total condensable PM 10. Inorganic fraction of condensable PM10 emissions is
assumed to consist entirely of sulfates (i.e. no soil component).
2.3.3
=
NO.3 Recovery Boller
Maximum daily firing rate of black liquor solids for the baseline period was provided by NewPage. The
No.3 Recovery Boiler is a direct contact evaporator. 502, NOx and PM emissions were calculated
using the emission factors given in NCASI Technical Bulletin 884 Table 4.11 .
.r--~
/
"
Rve Factor BART Analysis for NewPage Luke Mil
JUly 2010
Environment
AECOM
2-3
Speciation of the particulate matter emissions from No.3 Recovery Boiler into filterable and
condensable PM lO components was conducted using the following approach:
•
For a recovery boiler equipped with an ESP, 76% of filterable PM is PM 10 consistent with the
default approach cited in AP-42, Table 10.2-2. Fine PM 10 (PM 2.5 ) is 53.8 % of filterable PM 10.
•
Elemental carbon percentage was not available for a black liquor solids fired boiler.
Therefore, an oil fired boiler was used as a surrogate. For oil boilers, elemental carbon is
expected to be 7.4% of fine filterable PM10 based on the best estimate for oil combustion in
Table 6 of ·Catalog of Global Emissions Inventories and Emission Inventory Tools for Black
Carbon", William Battye and Kathy Boyer, EPA Contract No. 68-D-98-046, January 2002.
•
Condensable PM 10 emissions from No.3 Recovery Boiler are based on NCASI emission
factors documented in Appendix A entitled ·Particulate Emissions for Pulp and Paper Industry
Specific Sources". Inorganic condensable PM 10 emission factor is 1.36 Ibl ton BlS and
organic condensable PM 10 emission factor is 0.148 Iblton BlS. Sulfates, based on ion
chromatography, are 35% of condensable PM 10.
Table 2-1 provides a summary of the S02, NOx, and PM emissions that were used in the modeling
analysis for baseline conditions. Table 2-2 provides the stack parameters that were used in the
baseline modeling analysis.
Fi.. Factor BART MSysis for NllwPege Luke Mill
July 2010
')
')
Environment
AECOM
e_..)
lIu'",_
DIoII)'Heal
Ilnput Ra.'
IILSF. .d
Ra.""
UnIt
Power Beiler 125
Power Beiler #26
UnIta
MMBluIhr
781
(.)
CoeI
1,_.
- '.=.. U_.fZ,
0.05 (a)
NG
I'Ol
_~'
11.42
__ ~_ .~ __..
111.01
011
,
~
Pblhrl
3,333.33 (a) 753.39 (a)
lIM8IuIIlr
In Raeowry bollerl
F1.........
PM t1b1hr)
NOx
so, Pblhrj
F_I
MMBluIhr
III
79
Comlllnad Slack
..~"._. __ ~
2-4
NewPage Luke Mill - Baseline Emissions for Power Boiler No. 25, Power Boiler No. 26 and No.3 Recovery Boiler
Table 2·1
~_ ~~.
')
--cIIt1­
__ __ ~. __ .....:. __ _._.__ ,' __' _ .:...._.
I
T....'P....
FIIonblo p....
~
(Mr)
I-I
T....' fPII
Coars.
_T....'
197.85
20.53
19.92 (b)
0.62 (b)
-­ 0.59 (e)
0.02 (e)
(ilion.... +
70.8 (a)
',CondI..... PM...
I-I
CPMIOft
Be_nltl ToIalCPII TottlCPII
0 .._
10ft
In.31
(d) 141.85 (d)
ToIalCPII
Sol
aur•• s
OR
0.00 Id)
141.85 (d)
35.46 (d)
74.00 (a)
0.15
0.59
0.15
0.00 (b)
0.15 (b)
0.14 (e)
0.01 Ie)
O.44ld)
0.22 (d)
0.00 (d)
0.22 (d)
0.22 (d)
121••
7lI..
,.43
20."
11.12
0.71
0.73
0.03
177.71
142.07
0.00
142.07
31.1.
I 41..011111 I
12.84 Igl
1:1.10
I
_._~~_~._ ~_.~.l_
~21
,,_. "
I
17L"
I
4L04
114.03 (a I
I
34.01 la' 1
...;. . __..__.__ ._.~_,, ..,, .. _--'­ __ _._ .. _~_L __.__ ~~ __l~.
~_.,,_"--
'1.4'111 / 2.12111
_~_"
__,_ -_.__.. ~_" _
,,~_
1121.11 101/118.17 Igll 71.08(111
.i _
'
"·
cIIt1
A.··· ..
~
'
'"",_ . .
~"',
,-,
' __
~"
' . ' ,"
_.
..,
_OlinG period _ ... 2001 "'2003. MIX. _ hoot_r... f..... boIon f25 .... m_ldtdbf NowPoto Lullo MIX.
lood r• •, '*Y bIoek......_ (2,050tpd) prov_
_eel on40tmw per diIy Ind NOr: . . .Ions.e _ _ on'" perrrllltlor 0.98 .......... tnd nuirnlmhMt . . . For Boler 2e, "'eiona ofS02 .... bneelonh AF42 errWaionfltCtDr~en.,
0f11M r - * AprtI2002e.k_tcOftliJctecf on 1M tIIetack. Fa. ........Iona werellppCrignedlo h 8oIer21_~on. rDdhnti1Put
)iiiiii tt2$ ha<f. mikyclaM'Nc:k.n~iii(~&iit·diliilUiiOft1lbaed onAF42T""1]:~n.-.b*COIirCiiiidhfti1i~.~~COf*iildWb.-AC»Orci.G IoA~2 Ta. f;·:.e.>i.-.ca.i·"cfbOllt-~wih"ft1jkyclone, 29% "f ..·. ._PM.
O..d 3% II fne FIll 0 (l•. AI1.51.Ace"""'AI42 _
1.4, I
po ' d
'OO% "'oI_AlII FIIlO .... '00%.1 .. _.bIoI'MlD II file I'MIO (
1 _ ...).
c) 3.7%OiiiO·iO.i,;iI·ihOiiO·...,,;,...-W. . . .,h
hiHT..~ ~ ;"ii..,.. .801-~dOiObliiirii._lWoniiiritI _ionlWonIory T_ ' IiiiC.C";O;;;;;,M;;, ..... ....dKoohyBaYe;.,·~Contiaci,.,. 6i-i>-M:04li, .IoIV.Ilol:z002:
.7% oIfi'teFM10il the a:"eeI on"'lIV.ege
belt
forn."'gD contutlon iI T" 601 -C8IIIbgof GIc*.. Entla'" hv
endErril_ion hventoly Took for EIIckCllrbon·, MlmBdye Mel I<dty Boyer. ~ Contract NJ.6e-D-H-04e,.IlII'lIarV
_.oIla ... -..."'filePMlO_
I.) Itghoolhourty
bot
Rbn ~ n an ..... dIbtd-Utt 1,2010. For BollI' 25. S02
1IbIit1.4-21nd N:1IC III 74 ...... For Boler 25. taw. . All
cU\'Ig
_eel
rA".
••
d)'con<teM"·~D-i.·COi~~IiIlb8aiidon'AFCi2:-".~c~"·Piifii'.o:otS:c[i3~~"
....*ttiQ-i("Qiia.ir..crOr_··il20%--of_~hi~>.'UiNcnDIM' -(Le:-no"GI~~).'·~"t~'fo'iOr'~QU-Bing,,"bneciOn
fL42, r. . 1.4-.2. Tatalcondlnl" PM10. 5.7btrA1:f.
.ellrlltdbJ 1M 50% cAtoIIIaworgMic: III .eUfl'lld 10 be 5O%otlDllt. horgRc••ll8urNldlobe ...-.
(le. na salCOfI'I)(lnent)•
-0-*"
•)F.<.,....;;;y........ biiSOd"-Afi:42, TIiiiOTli.2-iliidiijinlO}:2. 7111cifiiWliiiO AlII PM;;:.~ "",,;;o~aiiIoJlM,-;COOiuJlM" ;0-;". cIff""IiiOOiiOtw iiii.rtiOribiOAA;;..,rhPM;,:· ....
f) 7.4'lIii7h-FMiOiiM.iiiiii·........Ofii01iOOiOO~f«peo;
. .m
........1iOn»ToilIe6.i'COiiiiiiiiDI ClIabOI8riiiionalwointorioa....liEiriiiSiiiOlW'onIory·f...·iiir""BiOCk'COiiiOil':ViilliiinIloll'i.lii1i1KiilY·IiOY..,e¥.Canii... ,.,. n-o.oa-o.e, .......... 2002: Fin.
aI ill the b~c. of,.. MO tobIL
IIfCCiiiCiiiMiiiI·FMiD~ -.ecovo;y-liOiiiilijibiiS..i;,· iCASi. ......'. .·(AppOiMi,iA. ~ EIria.... fQ; ~ .... Jiiip. ndUe.ySp.ck sGUr...: F".......,8.2006).
OR
bot...... ion
_ .... :15% at ... CI'M.
SuI_.
Table 2·2
ThO·........ itCb,ii,-•• (56 _ . , BLSI... Cl'MDR ..d 0.141_.' illS I..
NewPage Luke MIII- Baseline Stack Parameters for Power Boller No. 25, Power Boiler No. 26 and No.3 Recovery Boller
Stack Information
Boilers Combined
Stack
One stack for the 2 BART
eligible boilers (Boilers 25 & 26)
Recovery Boiler (3
stacks)
ESP 1
ESP2
ESP 3
Five Factor BART Analysis for NewPage Luke Mil
Stack round
or
rectangular
Diameter
(ft)
Stack
height
(ft)
Exhaust
flow
(acfm)
Exhaust
temp
(OF)
Exhaust
velocity
(ftls)
Round
16.50
623
394,000
373
30.71
Rectangle
Rectangle
Rectangle
7.94
8.88
9.03
294
308.5
300
280,000
280,000
140,000
345
345
345
94.28
75.29
36.46
July 2010
AECOM
3.0
Environment
3-1
Emission Control Alternatives
The control alternatives pertaining to visibility impairing pollutants (NOx, S02, and PM) are discussed
in this section. Information on control of these pollutants through application of a control device,
combination of devices, and/or operational change is provided in this section.
3.1
Power Boiler No. 25
The following BART control scenario was evaluated for the No. 25 Power Boiler:
•
3.1.1
Control Case - NOx, S02, and PM 10 emissions signature of Power Boiler No. 25 which
assumes a 90% reduction in S02 emissions from the baseline (via the installation of either a
Spray Dryer Absorber or a Circulating Dry Scrubber), reduction in NOx emissions from 0.99
Ib/MMBtu to 0.40 Ib/MMBtu (via operation of the existing SNCR which was installed in 2006)
and no change in particulate matter emissions compared to the baseline.
502 Emission Controls
Sulfur dioxide emissions are generated in fossil fuel-fired combustion units as a result of the oxidation
of sulfur present in the fuel. Approximately 98% of the sulfur in coal is emitted upon combustion as
gaseous sulfur oxides, S02 and S03' Uncontrolled emissions of S02 are directly related to the fuel
sulfur content, and not by the firing mechanism, boiler size, or operation. Many coal-fired boilers in
the U.S. limit emissions of S02 through the use of low sulfur western coals, including Powder River
Basin Coal. Compared with higher sulfur eastern bituminous coal that may contain as much as 4%
sulfur, the practice of burning western coal can reduce S02 emissions by approximately 70% to 90%.
However, control equipment such as wet and dry scrubbers can generally remove a higher
percentage of the S02 from higher sulfur coal than lower sulfur coal. The selection of coal type and
sulfur content, therefore, is an important aspect of the determination of BART for S02 and needs to be
considered in conjunction with add-on control alternatives when performing the BART analysis.
The following S02 control option was evaluated for this BART analysis:
•
S02 Control Case - S02 emissions signature of Power Boiler No. 25 which assumes a 90%
reduction in S02 emissions from the baseline (via the installation of either a Spray Dryer
Absorber or a Circulating Dry Scrubber). Emission levels and stack parameters
corresponding to this case are shown in Table 3-1 and Table 3-2, respectively.
Alternative add-on control technologies such as Wet Flue Gas Desulfurization (capable of achieving
90-95% control), Lime Dry Scrubber/Fabric Filter (capable of achieving 80 to 90% control) and Dry
Sorbent Injection with Trona (capable of achieving approximately 60% control) were not evaluated
since they offer relatively low control advantage compared to their annual cost. Moreover, due to
NewPage's commitment to the installation of an S02 control technology (either a Spray Dryer
Absorber (SDA) or a Circulating Dry Scrubber (CDS)); S02 emissions are reduced by as much as
90% from baseline levels.
Since this S02 control measure has already been committed to be implemented at the facility (see
October 31, 2007 letter from Gary Curtis, VP, Luke Operations to Brian Hug of MOE), no further
Five Factor BART Analysis for NewPage Luke Mill
July 2010
AECOM
Environment
3-2
review of the performance, and economic, energy. and environmental impacts of the control option
was necessary.
Five Factor BART Analysis for NewPage Luke Mill
July 2010
)
'I
)
Environment
AECOM
Table 3-1
3-3
NewPage Luke MiII- Emissions Control Case
Mulmum
IlllllyHeet
IInplltRalel
BLS Feed
Ralele)
Un/Ill
Power Boiler #25
761
MMBluIhr
Power Boiler #26
79
Unit
Combined Sleek
13 Raeo...ry bollerl
(0)
MMBIu/1v
(01
MMBtulhr
1140
85.42
lal
I
tone
BLS/hr(OI
so.
NO"
Ilb/hr)
(lb/hr,
Em••ions
FllI8rabie
PM (Ib/hr)
Condenstie PM1D
(lbIhr)
Rllarablo PM,.
(1JIhr1
TOlaIPM,.
Fuel
(fillarablo +
eonde.-ibIol
(1JIhr1
TotalFPM
CPMIOR
Rna
Coor. .
R.- Total
R.-Sol
Bememal TolaICPM TOlaICPM
cartlon
lOR
TOlaICPM
Sol
Sureteo
OR
53.3 (a)
64.23
49.01
23.03 (b)
25.97 (b)
25.01 (e)
0.96 (e)
15.22 (d)
12.18 (d)
0.00 (d)
12.18 (d)
3.04 (d)
Coal
333.3 (a) 304.40 (a)
NG
0.05 (a)
74.00 (a)
0.15
0.59
0.15
0.00 (b)
0.15 (b)
0.14 (e)
0.01 (e)
0.44 (d)
0.22 (d)
0.00 (d)
0.22 (d)
0.22 (d)
-
333.31
378.40
53.42
.....82
48.15
23.03
26.12
25.15
G.87
1...6
12.40
0.00
12.40
3.25
48.04
I 14.03 Ie) I
34.01(e)
U q o;:.,'2
I
181.08
I
83.10
I
63.21
I
176.85
I
I
31.48 '"
I
2.52 '" 1128.81 (1111115.17 Ill) I 71.08 11111 45.08 Ill) 1 12.84 Ill)
~thourty acbJII"'aiorw baed on the "'imJrndltf heetr.p.rtungthe opernng period between 2001 and 2003. Mar. dBIy heathput rite for the boIers 1'25 and 1'26 provided by NewAilge Luke MI. Max. daly feed rate of liy blllcIl;lquor.~(2,l)5Qtpd)
Boler 2S. S02 8I'T'Dslona und8rVO. reduction of 90% froml'le baI'*'e c..e 8nd N:>x erriuions .re baed on O.4l»Jt.U31u. For Bole' 26, errilsiona of S02 8nd N:>x rel'Tlli1 unch.-.ged cOfJl)lred to the b._e. For
Boler 25. future elTilalona of tier" PM baed on 0.07 ....... (a docun-ented i'I tie BART Corrmlm!=nt LeUer from fit. GIry CurtiB of New PllQII to fit. Brian .....a of t.«E).
(I)
'provtded by Ron PlIuWi'" en emIII dIted~ 1,2010. For
tiereforethe tteat
of
(b) A b8.,.,.ewa-~stBled on-BOi8ii25~200j;
.~iZe ciltrbitJon ~ h-.eci'o.;'AF4-2 f"'"-1.1-8"fOiil1K.r·cciitrOiedh8Y'ilO".iM!~e·_<~. etlntroi'l:ieVfe:AccOrdhg tiAF42~TBbIe"('-6.for' co.frredbok ecpJ~ed with--. baUhouse, 92% d-flterllb6e FMis
FMlOand 53% is file PM10.Accardi'lgto AF42 Section 1.4. far e""""" gas fi'ed~. 100% of alfbreble FMiI PM10 and 100% ofal flenlble PM10. fhe PM10(lesslh., 1 nicron).
{ej 3.1% Of~fi1e Ft4,~'. ~eEc b8aecfo.,'the ~er. of
"~~fOr~cD.i ~uon' ~ TIbte-i~oT~c.iti.k;goTGiOt)Ij-&riS.iOM hVenhirift;.niiE;ri;slOn 'hVentDrY iOOil-far'B.ck carbOn": VAlm BdVe~.nd i<achY &yer, EPA eomr·.~ r..:,.6&-D-9&-00t6, Jaooary 2002.
6.7% of file FM10is ffle B:b.. ed on"" lIVer. of the beetettirrlltes far nldur"g.. eormus.... n Table 6 of ~ of Q»b8I ems.ions hvsnh:riel and Erri..ion hventory Took far Bllickc.rbon-, ~mBdre and Kathy Boyer, SIA Contract rt:I. 6&-0.98-046, ....uary
2002. Fi1e .01 is the blllnce of me PYlO teal.
(dfCondensIbie -PM.o "or eOli'-rtTIg". based "on AF42, TIlbte 1.1-5.- TOt.! condenal:ll8 fIM1 ii • 0.02 M.Miu lshceh""bOier wi have elher s SQ6. or c.-y .crUbber i'I the future cae). horgar'lc iii ..sunwd to be ao% 01 tOtal and arganic • 20% of tdtal. horganic...surred to
be .ulates (le. no.oIco~. Conden8. . PM10 for n...... gas fritg II based on AF42. TIIb6e 1.4-2. TomIcondensllle PM10 -5.1 bUit:f. hargenic . . .surredto be 50% of totIIIend argllnic Is .ssurrRd to be SO% of totlll. harganice ..surredto be alsurBles (i.e. no
,01 corrponent)"
"(_iF« -. reCOvery baler, based on AF42.
'fc).,2::i Md Figure"10.i~2, 18% 0'- tleilbte FM iii ~o' Fl1e AA,~ II; 53.8% fhlble- PM,IO' eo.se ~o is is the
between total tier. . FlM,o ei.d '..Ie FM,o·
{ij 1.4"-01 &N PM1 0 tNtIecffin
of tie-beef..tiImiei
petr~OIeum contKlBliOrl'~
oiCabdoij'Of 'Ciob8i"i:nis.-iOri."'hveirtOrieJl' a,dAErrii.iO" iWentarY-foOilfOr Eilcii Carbon';,-VilBm
.lAd 1<8ihy" Boyer, EPA Contract ~. ~D-98-046, Jenullry 2002.
Fi1e solis the balance of t"e PM10 tota
(g)' Conden.IiIe RdfO eins•
trom recOvIrY balerS "Ire bee.8d' on N:ASI-erril.1or1 ilCtors (APtienCD~A~" Fw1IcL8te Etris.lorM'- tor R4i ~d FraPfi-hcill,iry Sptcif.e-"SOUrces-, FebrU.,.y 9, -2006). The efril.ions fectors ..e 1.361:H'1on of BLS tot a:t.A JJR end O.141b'ton of BlS
for CFMOfl Sulfates, based on Ion chrotnlwwaphy, are 35% of the a:Y.
r.,.
'Of
hllVerage'
t.li*-eo(
of
cit'ference
Bdye
.on.
Table 3-2
NewPage Luke Mill - Stack Parameters Control Case
Stack Information
Boilers
Combined
Recovery Boiler
(3 stacks)
One stack for
Boilers 25 & 26
ESP 1
ESP 2
ESP 3
Five Factor BART Analysis lor NewPage Luke
M~I
Stack round or
rectangular
Diameter
(tt)
Stack
height (tt)
Exhaust
flow (acfm)
Exhaust
temp (OF)
Exhaust
velocity (ftIs)
Round
16.50
623
394,000
373
30.71
Rectangle
Rectangle
Rectangle
7.94
8.88
9.03
294
308.5
300
280,000
280,000
140,000
345
345
345
94.28
75.29
36.46
July 2010
AECOM
3.1.1.1
Environment
3-4
Discussion of Candidate S02 Control Technology
Given the commitment to install either a Spray Dryer Absorber or a Circulating Dry Scrubber to
achieve 90% control of S02 emissions, no alternative add-on controls were evaluated. Therefore,
BART for S02 is 90% control compared to the baseline levels Le. an S02 emission rate of 0.44
Ib/MMBtu.
3.1.2
NOx Emission Controls
Nitrogen oxides formed during the combustion of coal are generally classified as either thermal NOx
or fuel-bound NOx. Thermal NOx is formed when elemental nitrogen in the combustion air is oxidized
at the high temperatures in the primary combustion zone yielding nitrogen oxide (NO) and nitrogen
dioxide (N0 2). The rate of formation of thermal NOx is a function of residence time and free oxygen,
and increases exponentially with peak flame temperatures. Thermal NOx from coal combustion can
be effectively controlled by techniques that limit available oxygen or reduce peak flame temperatures
in the primary combustion zone. Fuel-bound NOx is formed by the oxidation of chemically bound
nitrogen in the fuel. The rate of formation of fuel-bound NOx is primarily a function of fuel bound
nitrogen content, but may also be affected by fuel/air mixing.
The technologies available to control NO x from coal-fired boilers include combustion controls, such as
10w-NOx burners (LNB) and overfire air (OFA), and post-combustion control techniques, such as
selective non-catalytic reduction (SNCR) and selective catalytic reduction (SCR).
Power Boiler No. 25 currently has an SNCR (installed in 2006), low NOx burners, and overfire air for
the control of NOx emissions during the ozone season.
The technical feasibility and performance levels of the alternative NOx control technologies are
evaluated below in terms of their application to Power Boiler No. 25.
Selective Catalytic Reduction
Selective Catalytic Reduction (SCR) is a process that involves post-combustion removal of NOx from
flue gas utilizing a catalytic reactor. In the SCR process, ammonia injected into the flue gas reacts
with NOx and oxygen to form nitrogen and water vapor. The SCR process converts NOx to nitrogen
and water by the following general reactions:
The reactions take place on the surface of a catalyst. The function of the catalyst is to effectively
lower the activation energy of the NOx decomposition reaction to about 375 to 750°F, depending on
the specific catalyst and other contaminants in the flue gas. The factors affecting SCR performance
are catalyst reactor design, optimum operating temperature, sulfur content of the fuel, catalyst
deactivation due to aging or poisoning, ammonia slip emissions, and design of the ammonia injection
system.
The SCR system is comprised of a number of subsystems, including the SCR reactor, ammonia
injection system, and ammonia storage and delivery system. Typically, the SCR reactor is located
downstream of the economizer and upstream of the air pre-heater and the particulate control system.
Five Factor BART Analysis for NewPage Luke Mill
July 2010
Environment
AECOM
3-5
From the economizer outlet, the flue gas would first pass through a low-pressure ammonia/air
injection grid designed to provide optimal mixing of ammonia with flue gas. The ammonia treated flue
gas would then flow through the catalyst bed and exit to the air pre-heater. The SCR system for a
coal boiler typically uses a fixed bed catalyst in a vertical down-flow, multi-stage reactor.
Reduction catalysts are divided into two groups: base metal, primarily vanadium, platinum or titanium,
(lower temperature), and zeolite (higher temperature). Both groups exhibit advantages and
disadvantages in tenns of operating temperature, ammonia-NOx ratio, and optimum oxygen
concentration. The optimum operating temperature for a vanadium-titanium catalyst system is in the
range of 550 0 to 750°F, which is significantly higher than for platinum catalyst systems. However, the
vanadium-titanium catalyst systems begin to break down when operating at temperatures above this
range. Operation above the maximum temperature results in oxidation of ammonia to ammonium
sulfate and NOx, thereby actually increasing NOx emissions.
SCR with ammonia injection technology is a demonstrated, commercially available technology. SCR
has been used with other coal-fired boilers; therefore, it is a technically feasible technology for the
control of NOx emissions from Power Boiler No. 25.
Selective Non-Catalytic Reduction
Selective non-catalytic reduction is a post-combustion control technology that involves ammonia or
urea injection into the flue gases without the presence of a catalyst. SNCR, similar to SCR, involves
the reaction of NOx with ammonia, where a portion of the NOx is converted to molecular nitrogen and
water. Without the use of a catalyst or supplemental fuel injection, the NOx reduction reaction
temperature must be tightly controlled between 1,600 and 2,200°F (between 1,600 and 1,800°F for
optimum efficiency). Below 1,600°F ammonia will not fully react, resulting in un-reacted ammonia that
is emitted into the atmosphere, (referred to as ammonia slip). If the temperature rises above 2,200°F,
the ammonia added will be oxidized resulting in an increased level of NOx emissions.
SNCR with ammonia injection technology is a demonstrated, commercially available technology.
SNCR has been used with other coal-fired boilers; therefore. SNCR is indeed technically feasible for
the control of NOx emissions from Power Boiler No. 25. However, NOx removal efficiencies with
SNCR are lower than those with SCR, typically ranging from 30 to 50% depending on the combustion
process and inlet NOx concentrations.
In 2006, SNCR was installed on Power Boiler No. 25 for the control of NOx emissions.
3.1.2.1
Discussion of Candidate NOx Control Technologies
The NOx post-combustion control technologies identified for evaluation are SCR and SNCR. Of these
technologies, SCR has been demonstrated to be the most effective technology in minimizing NOx
emissions from coal-fired boilers. However, Power Boiler No. 25 already has an SNCR in place for
controlling NOx emissions which offers a control efficiency of -60% compared to baseline NOx levels.
Therefore, continuous operation of the SNCR system with a rolling 30-day emission rate of 0.40
Ib/MMBtu is recommended as BART for Power Boiler No. 25.
3.1.3
PM Emission Control
Power Boiler No. 25 currently employs a multi-cyclones and a fabric filter to control PM emissions.
The baseline PM emission rate for Power Boiler No. 25 is 71.4lb/hr which is well below its permit limit.
Moreover, PM emissions are not a significant contributor to the visibility impacts as seen in the
Five Factor BART Analysis for NewPage Luke Mill
July 2010
AECOM
Environment
3-6
modeling analysis presented in Section 5. Visibility modeling shows that PM emissions have a
relatively minor contribution to the overall visibility impacts. Given the high performance level of the
existing multi-cyclones and baghouse, these PM control devices are considered BART for Power
Boiler No. 25 and no additional PM controls were considered as part of this analysis.
.'
3.2
Power Boiler No. 26
The following BART control scenario was evaluated for the No. 26 Power Boiler:
•
3.2.1
Control Case - The current (2009-2010 period) emissions signature of Boiler No. 26 which
assumes no change in NOx, S02, and PM 10 emissions compared to the baseline.
802 Emission Controls
Sulfur dioxide emissions are generated in fossil fuel-fired combustion units as a result of the oxidation
of sulfur present in the fuel. Uncontrolled emissions of S02 are directly related to the fuel sulfur
content, and are not affected by the firing mechanism, boiler size, or operation. Power Boiler No. 26 is
a natural gas fired boiler. Pipeline quality natural gas has a very low sulfur content and hence minimal
S02 emissions.
The following control option was evaluated for this BART analysis:
•
S02 Control Case - The current (2009-2010 period) S02 emissions signature of Boiler No. 26
which is 0.0006Ib/MMBtu (assumes no change from baseline levels). Emission levels and
stack parameters corresponding to this emissions scenario are shown in Table 3-1 and Table
3-2, respectively.
Since Power Boiler No. 26 is a natural gas fired boiler, it has very low emissions of S02 (less than 0.1
Ib/hr). No add-on emission controls were considered for this unit as any visibility improvement
(offered by an add-on control) to the already small visibility impacts would not be cost-effective and
hence would not constitute BART.
Hence, BART for S02 is the current emissions signature of Power Boiler No. 26, Le., an S02 emission
rate of 0.0006 Ib/MMBtu.
3.2.2
NOx Emission Controls
Nitrogen oxides formed during the combustion of fossil fuels are generally classified as either thermal
NOx or fuel-bound NOx. Thermal NOx is formed when elemental nitrogen in the combustion air is
oxidized at the high temperatures in the primary combustion zone yielding nitrogen oxide (NO) and
nitrogen dioxide (N02). The rate of formation of thermal NOx is a function of residence time and free
oxygen, and increases exponentially with peak flame temperatures. Fuel-bound NOx is formed by the
oxidation of chemically bound nitrogen in the fuel. The rate of formation of fuel-bound NOx is primarily
a function of fuel bound nitrogen content, but may also be affected by fuel/air mixing. Natural gas has
insignificant amounts of chemically bound nitrogen.
The NOx emissions from Power Boiler No. 26 are small (on the order of 74 Ib/hr). Power Boiler No.
26 burns a clean fuel, has a low annual capacity factor and produces relatively small visibility impacts.
No add-on emission controls were considered for this unit as any visibility improvement (offered by an
Five Factor BART Analysis for NewPage Luke Mill
July 2010
Environment
AECOM
3-7
add-on control) to the already small visibility impacts would not be cost-effective and hence would not
constitute BART.
Hence, BART for NOx is the current emissions signature of Power Boiler No. 26, Le., an emission rate
of 0.94 Ib/MMBtu.
3.2.3
PM Emission Control
Power Boiler No. 26 currently has no PM emissions control systems. The baseline PM emission rate
for Power Boiler No. 26 is very low (0.0018Ib/MMBtu). Moreover, the contribution of PM emissions
from Power Boiler No. 26 to visibility impairment is relatively small.
Since Power Boiler No. 26 is a natural gas fired boiler, it has very low emissions of PM (less than 0.15
Ib/hr). No add-on PM emission controls were considered for this unit as the small visibility
improvement offered by any add-on control would not be cost-effective and hence would not
constitute BART.
Hence, BART for PM for Power Boiler No. 26 is the current emissions signature of Power Boiler No.
26 Le., a PM emission rate ofO.0018Ib/MMBtu.
3.3
No.3 Recovery Boiler
A recovery boiler is used to recover pulping chemicals, evaporate residual moisture from the black
liquor solids, and bum the organic constituents and to produce steam. No.3 Recovery Boiler is a
straight fired unit burning black liquor. No.2 oil may be bumed during start-up and shutdown. No.3
Recovery Boiler at the Luke Mill is currently equipped with a two level staged combustion air control
system with ESPs.
The following BART control scenario was evaluated for the NO.3 Recovery Boiler:
•
3.3.1
Control Case - The current (2009-2010 period) emissions signature of the NO.3 Recovery
Boiler which assumes no change in NOx, S02, and PM 10 emissions compared to the baseline
levels. The recovery boiler currently has a two-level staged combustion air control system to
control NOx and S02 emissions and ESPs installed to control PM emissions. Emission levels
and stack parameters corresponding to this emissions scenario are shown in Table 3-1 and
Table 3-2, respectively.
502 Emission Controls
In general, black liquor contains a significant amount of sulfur, nominally 3% to 5% by weight of the
dissolved solids. S02 emissions from recovery boilers occur due to the volatilization and subsequent
oxidation of sulfur compounds present in black liquor and the occasional use of auxiliary fuel (in this
case NO.2 oil). Unlike conventional steam boilers, the vast majority of sulfur that is present in the
liquor is not converted to S02. Proper operation of the recovery boiler maximizes the conversion of
sulfur compounds in the liquor to the principal constituents of pulping chemicals through capture of
these compounds in the combustion zone of the boiler by sodium fume released from the smelt bed.
Consequently proper operation of the recovery boiler itself results in inherent control of S02
emissions.
The available retrofit technologies for control of S02 from Kraft mill Recovery Boilers are:
Five Factor BART Analysis for NewPage luke Mill
July 2010
AECOM
•
•
Environment
3-8
Staged Combustion Systems
Wet Scrubbers
The No.3 Recovery Boiler is currently equipped with a two-level staged combustion air control system
which provides control over the boiler operating parameters that minimize S02 emissions.
The only available alternative to proper recovery boiler operation with a staged combustion air system
for S02 control is the use of wet scrubbing. Dry scrubbing techniques, which seek to remove S02 via
direct injection of solid lime or limestone directly into the recovery boiler flue gas, are not technically
feasible for recovery boilers. The reaction products from dry scrubbing (calcium sulfate and other
calcium-based salts) would necessarily need to be collected in the PM control system (ESP), and
thereby would contaminate the sodium sulfate (salt cake) that the ESP collects and recycles back to
the pulping process.
There are only three recovery boilers in the U.S. that are equipped with wet scrubbers in addition to
ESPs. However, none of these scrubbers were installed to achieve S02 removal: two of the units
were installed for heat recovery reasons, and the third was installed on a unit that is equipped with a
direct contact evaporator to prevent liquid droplets from being entrained and discharged to the
atmosphere. Furthermore, the inherent S02 control provided via proper operation of a recovery boiler
results in a much lower S02 concentration in recovery boiler flue gas than in the flue gas of emission
units to which wet scrubbers are routinely applied (such as fossil fuel-fired steam generating units).
Therefore, BART for S02 emissions from No.3 Recovery Boiler is the currently installed two-level
staged combustion air control system with ESPs.
3.3.2
NOx Emission Controls
NOx emissions generally result from fuel NOx and thermal NOx. However, NOx formation in recovery
boilers is believed to be primarily from fuel NOx because the temperatures in the combustion zone of
the boiler are not high enough for significant thermal NOx formation. Fuel NOx emissions from
recovery furnaces are typically low due to the low nitrogen content of black liquor solids. In addition,
No.3 Recovery Boiler operates as a staged combustion boiler with independently operating primary,
secondary and tertiary air dampers. Total air and the distribution between the air stages can be
adjusted to control the Kraft recovery sodium sulfite reactions, to assure complete combustion of
organic compounds, and to control TRS, CO and NOx.
The available retrofit technologies for control of NOx from combustion sources are generally
considered to be:
•
•
•
•
•
Staged Combustion Systems
SNCR
SCR
Low NOx burners
Flue Gas Recirculation
Recovery boilers are complex systems, specifically designed for chemical recovery, that can not apply
the types of NOx emission controls used on typical coal, oil, and natural gas fired boilers such as low
NOx burner, flue gas recirculation, selective non-catalytic reduction, and selective catalytic reduction.
Five FactOI' BART Analysis for NewPage Luke Mill
July 2010
Environment
AECOM
3-9
Low NOx burners, which use staged combustion, have not been demonstrated on recovery boilers.
Black liquor has a large percentage of water that requires specifically designed black liquor burner
guns that provide long droplet trajectories for drying (see the NCASI retrofit control technology
assessment in Appendix B).
Flue gas recirculation is a technology used to control thermal NO x by reducing combustion
temperatures. However, NOx emissions from recovery furnaces are a result of fuel NOx oxidation.
Therefore, flue gas recirculation is not technically feasible for recovery furnaces.
Selective non-catalytic reduction has not been demonstrated on recovery boilers. There are concerns
that long term injection of ammonia or urea in a recovery furnace may adversely affect the chemical
recovery process. In addition, large variations in gas temperatures due to fluctuating loads and black
liquor quality would adversely affect the performance of selective non-catalytic reduction. Ammonia
slip and increased plume opacity are also issues.
Selective catalytic reduction is not technically feasible due to high particulate concentrations in the
economizer region and catalyst poisoning by alkali metals such as sodium.
In summary, there are no technically-feasible alternatives for control of NOx emissions from recovery
boilers other than the controls that are currently in place. In particular, emission controls which have
been demonstrated on conventional steam boilers (including low NOx burner, flue gas recirculation,
selective non-catalytic reduction, and selective catalytic reduction) cannot be applied to or have not
been demonstrated to be feasible on a recovery boiler.
Therefore, BART for NOx emissions from No.3 Recovery Boiler is the currently installed two-level
staged combustion air control system with ESPs.
3.3.3
PM Emission Control
Particulate matter emissions from the NO.3 Recovery Boiler are currently controlled using a three­
chamber electrostatic precipitator (ESP1, ESP2 and ESP3).
The available retrofit technologies for control of particulate matter from Kraft mill Recovery Boilers are:
•
•
•
Electrostatic Precipitators
Wet Scrubbers
Fabric Filters
Recovery boilers are designed and operated so that sodium fume released from the smelt bed is
present in the combustion chamber in order to capture the S02 that is generated as a result of
oxidation of the reduced sulfur compounds in black liquor. As a consequence, recovery boilers emit
relatively low levels of S02 emissions, but have higher levels of uncontrolled PM emissions.
Nonetheless. economical operation of the chemical recovery cycle in the Kraft pulping process
requires that the vast majority of this uncontrolled PM be captured in a control device and the
collected sodium salts (primarily sodium sulfate) be returned to the process.
Electrostatic precipitation is the only type of PM control technology used on modern recovery boilers
at Kraft pulp mills. At one time prior to the promulgation of New Source Performance Standards for
Kraft pulp mills, recovery boilers utilized venturi scrubbers for PM emissions control. However
because ESPs are capable of a greater degree of emissions control at a lower operating cost, venturi
scrubbers are no longer utilized.
Five Factor BART Analysis for NewPage Luke Mill
JUly 2010
AECOM
Environment
3-10
Fabric filters have been utilized on conventional coal-fired steam generating units and are generally
considered to be equivalent to ESPs on these types of sources in terms of PM control efficiency.
However, they have not been applied to recovery boilers at Kraft pulp mills and are thus not
considered to be an available BART alternative.
Consequently, there are no alternatives that offer a greater degree of PM emissions control than the
three-chamber ESP that is currently in use on this unit and hence constitutes BART for the No.3
Recovery Boiler.
Five Faclor BART Analysis for NewPage luke Mill
July 2010
AECOM
4.0
Environment
4-1
CALPUFF Modeling Inputs and Procedures
This section provides a summary of the modeling procedures that were used for the refined CALPUFF
analysis conducted for the BART units at the Luke Mill.
4.1
Location of Source vs. Relevant Class I Areas
Figure 4-1 shows the location of the Luke Mill relative to nearby Class I areas. There are four Class I
areas within 300 km of the facility: Shenandoah National Park (VA), Dolly Sods Wilderness Area
(WV), Otter Creek Wilderness Area (WV), and James River Face Wilderness Area (VA). The BART
modeling analysis has been conducted for all of these Class I areas in accordance with the referenced
Visibility Improvement State and Tribal Association of the Southwest (VISTAS) common BART
modeling protocol and FLAG 2008 guidance.
4.2
General Modeling Procedures
Class I modeling was conducted using three years (2001-2003) of CALMET meteorological database.
The database was developed for use in BART assessment in VISTAS. VISTAS has developed five
sub-regional 4-km CALMET meteorological databases. Class I modeling for the Luke Mill was done
using sub-domain #5.
CALMET processing procedures are fully described in the VISTAS common BART modeling protocol,
available at http://www.vistas-sesarm .0rq/documents/BARTModelingProtocol rev3.2 31 Aug06.pdf.
The receptors used for each of the Class I areas are based on the National Park Service database of
Class I receptors, available at http://www.nature.nps.qov/air/maps/Receptors/index.cfm.
4.3
Model Version
The EPA-approved version of CALPUFF was used to model the emissions and Version 6 of
CALPOST was used to process the regional haze impacts with Method 8 (New IMPROVE equation).
CALPUFF Version 5.8, Level 070623 and CALPOST Version 6.221, Level 080724 were used.
These programs are available at http://www.src.com/calpuff/calpuff1.htm.
4.4
Background Air Quality Data
CALPUFF modeling was conducted with the hourly background ozone data that was developed for
VISTAS sub domain #5 and a monthly ambient ammonia background of 0.5 ppb. This ammonia
background corresponds to the value listed in the VISTAS BART protocol.
4.5
Light Extinction and Haze Impact Calculations
The FLAG 2008 document (dated June 26, 2008) provides guidance on the recommended new
IMPROVE equation application. CALPOST Version 6.221 defines this application as Method 8. Mode
5. The assessment of visibility impacts at the Class I areas used CALPOST Method 8.
Five Factor BART Analysis for NewPage Luke Mill
July 2010
Environment
AECOM
4-2
The CALPOST postprocessor was used for the calculation of the impact of the modeled source's
primary and secondary particulate matter concentrations on light extinction. In the new IMPROVE
equation, the total sulfate, nitrate, and organic carbon compound concentrations are each split into
two fractions, representing small and large size distributions of those components. New terms, such
as sea salt (important for coastal locations), absorption by N02 (only used where N0 2 data are
available), and site-specific Rayleigh scattering have been added to the equation. The new
IMPROVE equation for calculating light extinction is shown below.
2.2· Is(lU 1)- (Small
1
SUlf~lh:) ~
2.4 / Is(IUI) - [Small Nitrate]
+ 2.8'. (Small Organic tvla.'is)
+ to
t
1
x
)(
5.1- liJRH)
x
(Large Nitmte]
6.1· [Large Organic ~'la'is)
[Elemental Carbon]
'Fine Soil]
+ 0.6
x
[Coarse Mass)
1.7
x
fss(RH)"· [Sea Salt]
t·
t
+
4.8- 1i-(lUI)' (Large Sulfate)
+ Rayleigh Scattering (Site Specific)
.. 0,33
x
[N02 (ppb)] {or as: 0.1755 )( (N02 (l-lglm3 )/l
Where:
) indicates concentrations in ~tg/nl
Is(RH) '" Relative humidity acljustment factor lor small sulfate and nitrate
1i-(lUt) "'" Relative humidity
adju~1ment factor
fi:>r large sulfate tmd nitrate
Iss(RH) '''' Relative humidity adjuslment factor for sea salt
For Total Sulfate < 20 ~tg/m3:
[Large Sulfate]
([Total Sulfate) ! 20 ~tglm3)
',0
[Total Sulfate]
For Total Sulf.1te ::: 20 l-1g1m3:
(Large Sulfate] "" [Total Sulfate]
And:
[Small Sulfate) =c (Total Sulfate] - [Large Sulfate]
To calculate large and small nitrate tmd organic mass. substitute ( I Large, SmaIL -rotaI:
'Nitrate. Organic Mass}) for Sulfate.
Five Factor BART Analysis for NewPage Luke Mill
July 2010
Environment
AECOM
4-3
The FLAG 2008 document provides inputs to the new IMPROVE equation that are based on either
the 20% best or annual average natural conditions. AECOM elected to use the more conservative
inputs that are based on the 20% best days natural conditions.
Inputs to the CALPOST Method 8 calculations for each Class J area were obtained from the FLAG
2008 document tables referenced below.
Table 4-1
References to the New IMPROVE Equation CALPOST Inputs
Sea salt concentration
FLAG 2008 Table V.1-2
Rayleigh scattering
FLAG 2008 Table V.1-2
Monthly f L (RH)
FLAG 2008 Table V.1-3
Monthly f s (RH)
FLAG 2008 Table V.1-4
Monthly f ss (RH)
FLAG 2008 Table V.1-5
Five Factor BART Analysis for NewPage Luke Mill
July 2010
AECOM
Figure 4-1
Environment
4-4
Location of Class I Areas in Relation to the NewPage Luke Mill
Class I Areas within
300 km of the
NPS Class IAreas
New Page Luke Mill
100
Five Facto< BART Analysis for NewPage Luke Mill
:J NewPage'
200
July 2010
Environment
AECOM
5.0
5-1
CALPUFF Modeling and BART Determination Results
This section presents the recommended BART determination and provides a summary of the
modeled visibility improvement as a result of applying BART to Power Boiler No. 25, Power Boiler No.
26 and No.3 Recovery Boiler at the Luke Mill.
5.1
Baseline CALPUFF Modeling Results
CALPUFF modeling results of the baseline emissions at four Class I areas are presented in Table 5-1.
Modeling was conducted for all three years of CALMET meteorological data (2001-2003). Emission
rates that were used in modeling the baseline emissions are listed in Table 2-1.
For each Class I area and year, Table 5-1 lists the 98th percentile (8th highest day's) delta-deciview.
The results indicate that the higher visibility impacts generally occur at Shenandoah National Park and
Dolly Sods Wilderness. Higher impacts at these Class I area are due to their proximity to the site and
local meteorological conditions.
th
EPA recommends in its BART Guidelines that the 98 percentile value of the modeling results should
be compared to the threshold of 0.5 deciviews to determine if a source contributes to visibility
impairment. The Guidelines also recommend using the 98th -percentile statistic for comparing visibility
improvements due to BART control options.
The results of the baseline emissions modeling indicate that the cumulative visibility impacts from the
Power Boiler Nos. 25 and 26 and No.3 Recovery Boiler exceed 0.5 deciviews in at least one Class I
area (see Table 5-1). Therefore, per 40 CFR Part 51, Appendix Y, Power Boiler No. 25, Power Boiler
No. 26 and No.3 Recovery Boiler at the Luke Mill are presumed to be subject to BART because their
emissions may reasonably be anticipated to cause or contribute to visibility impairment at a nearby
Class I area.
Table 5·1
Regional Haze Impacts Due to Baseline Emissions
2001
Cla_IArea
Ca.
I
2002
2001·2003
Avg
2003
day,,> days>
days> days>
days> days>
8th
8th
8th
MAXdv
MAXdv
MAXdv
8th Highest
0.5dv 1.0 dv
Highest 0.5dv 1.0 dv
Highest 0.5dv 1.0 dv
Hlghell
A B...
AB...
ABa.
dVA B...
AB... AB...
dv A B.., AB... ABa'
dvA Box AB... AB...
dVAB...
Shenandoah NP
Baseline
149
85
3.90
2.40
110
60
2.96
2.20
110
68
3.27
2.44
2.35
Olter Creek W
Baseline
31
11
2.34
1.32
28
10
2.19
1.12
22
12
2.68
1.22
1.22
Dolly Sods W
Baseline
40
22
3.39
2.11
40
18
3.66
1.27
27
18
3.96
1.70
1.69
James Riler Face W
Baseline
11
1
1.24
0.65
12
2
1.61
0.64
8
2
1.48
0.55
0.61
5.2
Modeling Results for the BART Control Case
CALPUFF modeling result of control case is presented in Tables 5-2. Modeling was conducted for all
three years of CALMET meteorological data (2001-2003) for the four Class I areas to determine the
effects of the proposed controls on the three BART-eligible units at the Luke Mill. Emission rates that
were used in modeling the BART control case are listed in Tables 3-1. Stack parameters associated
with the control case are given in Tables 3-2.
Five Factor BART Analysis for NewPage Luke MIll
July 2010
AECOM
Environment
5-2
For each Class I area and year, the tables below list the 98 th percentile delta-deciview values, number
of days above 0.5 and 1.0 delta-deciview due to the BART emission controls.
Class I modeling results show that the 3-year average regional haze impacts are reduced by about
1.57 delta-dv (67% reduction) at Shenandoah, by 0.83 delta~v (-68% improvement) at Otter Creek,
by 1.11 delta-dvat Dolly Sods (-66% improvement) and by 0.47 delta-dv at James River Face (- 77%
improvement) relative to the baseline case with the future controlled emissions signature of the Power
Boiler No. 25, Power Boiler No. 26 and NO.3 Recovery Boiler.
Table 5-2
Regional Haze Impacts Due to the Future Controlled Emissions
2001
Class I Area
2002
2001·2003
Avg
2003
Ca. days> days>
a" days> days> MAXdv a'" days> days> MAXdv a
MAXdv
0.5dv 1.0 dv
Higheot 0.5dv 1.0 dv
Higheot 0.5dv 1.0 dv
Higheot
A Boo.
AB...
A B...
A B... AB...
dvA B. A B... AB...
dvAB.. AB... A B...
dvA Boo
th
ath Hlgheot
dvA Boo,
Shenandoah NP
Future
43
5
1.39
0.89
21
0
0.97
0.69
38
3
1.06
0.76
0.78
Otter Creel< W
Fu(ure
4
0
0.85
0.34
4
0
0.83
0.34
6
0
0.88
0.48
0.39
Dolly Sods W
Future
10
4
1.18
0.73
7
2
1.31
0.48
9
2
1.29
0.51
0.58
James Ri",r Face W
Future
0
0
0.26
0.14
1
0
0.58
0.16
0
0
0.47
0.12
0.14
5.3
BART Results and Discussion
As discussed earlier in this section, visibility improvements resulting from the future controlled
emissions level of Power Boiler No. 25 are on the order of 66-77% compared to the baseline.
Therefore, we conclude that the recommended BART for Power Boiler No. 25 is the installation of an
add-on S02 control (either a Spray Dryer Absorber or a Circulating Dry Scrubber), year-round
operation of the existing SNCR for NOx control and multicyclones and baghouse for PM control.
Burning natural gas, which inherently has low nitrogen, sulfur and ash content, constitutes BART for
Power Boiler No. 26. The currently installed two-level staged combustion air control system with
ESPs constitutes BART for the NO.3 Recovery Boiler.
Five Factor BART Analysis for NewPage luke Mill
July 2010
Environment
AECOM
6.0
6-1
References
Environmental Protection Agency (EPA), AP 42, Fifth Edition, Compilation of Air Pollutant Emission
Factors, Volume 1: Stationary Point and Area Sources, January, 1995
Environmental Protection Agency (EPA), Guidance for Tracking Progress Under the Regional Haze
Rule, EPA-454/B-03-003, Appendix A, Table A-3, September, 2003a
Environmental Protection Agency (EPA), Guidance for Estimating Natural Visibility Conditions Under
the Regional Haze Program, EPA 454/B-03-005, September 2003b
Environmental Protection Agency (EPA), Interagency Workgroup on Air Quality Modeling (IWAQM)
Phase 2 Summary Report and Recommendations for Modeling Long Range Transport Impacts, EPA­
454/R-98-019, December, 1998
EPRI. Estimating Total Sulfuric Acid Emissions from Stationary Power Plants, EPRI, Palo Alto, CA:
2008. 1016384.
Regional Haze Regulations and Guidelines for Best Available Retrofit Technology (BART)
Determinations; Final Rule (FR Vol. 70, No. 128 published July 6, 2005).
Regional Haze Regulations; Revisions to Provisions Governing Alternative to Source-Specific Best
Available Retrofit Technology (BART) Determinations; Final Rule (FR Vol. 71, NO. 198 published
October 13, 2006).
Federal Land Managers' Air Quality Related Values Workgroup (FLAG). Phase I Report Revised
Draft, June 2008.
Visibility Improvement State and Tribal Association of the Southeast (VISTAS), Revision 3, Protocol
for the Application of the CALPUFF Model for Analyses of Best Available Retrofit Technology (BART),
updated July 18,2006.
Five Factor BART Analysis for NewPage luke Mill
July 2010
Environment
AECOM
Appendix A
Particulate Emissions for Pulp
and Paper Industry Specific
Sources
Five Factor BART Analysis lor NewPage Luke Mill
July 2010
February 9,2006
Particulate Emissions Data for Pulp and Paper Industry Specific Sources
The following tables contain summarized particulate emissions data for sources that are specific to the
pulp and paper industry. The source categories addressed in this document are smelt dissolving tanks,
lime kilns, and recovery furnaces. Boilers are not addressed since AP-42 emission factors for boiler
emissions are well documented and readily available including in NCASI Technical Bulletin No. 884.
Smelt Dissolving Tanks
Data for smelt dissolving tanks were compiled from NCASI Technical Bulletins Nos. 884 and 898. This
data set includes test results from the use of a dilution tunnel method which quantifies total PM IO and
PM 2.S particulate matter. Total PM IO and PM2.S particulate matter are the sum of filterable and
condensible PM IO and PM2.S particulate matter. All smelt dissolving tanks in the data set have wet
particulate control devices.
The filterable PM numbers are obtained from combining the data set of 36 sources listed in NCASI
Technical Bulletin No. 884, Table A 15c, and the data set of 6 sources listed in NCASI Technical Bulletin
No. 898. The data for "Total PM IO" and "Total PM2.S" are from the 8 sources listed in NCASI Technical
Bulletin No. 884, Table A15d. All of the CPM data was from the 6 sources listed in NCASI Technical
Bulletin No. 898. The CPM data listed in Technical Bulletin No. 884 was not used as that data is an
estimate of CPM and not results from EPA Method 202. All of the sulfate data is from the 3 sources
tested by NCASI, and listed in Technical Bulletin No. 898.
Table 1. Smelt Tank Data Summary
Parameter
PM
6Total PM IO
6Total PM2.S
CPM MeCh Soluble
CPM Water Soluble
CPM
Sulfate
Measurement
Method
No. of
Sources
EPA Method 5
Dilution Tunnel
Dilution Tunnel
EPA 202
EPA 202
EPA Method 202
IC
42
8
7
6
6
6
3
Range
Mean
(lb/ton BLS)
0.03 - 0.64
10.031 - 0.666
10.027 - 0.570
10.0009 - 0.0192
10.0039 - 0.0832
10.0048 - 0.1024
30.0014 - 0.0297
0.148
20.154
20.132
20.0044
20.0192
20.0237
40.0069
sMean
Percent of PM
6104
689
3
13
16
%ofCPM=
29
IRange values were determined by applying the mean percent of PM to the range of values for PM. zMean values
were determined by applying the mean percent of PM to mean value for PM. 3Range values for sulfate were
determined by applying the mean percent of CPM to the range of CPM values. ~ean value for sulfate was
determined by applying the mean percent of CPM to the mean value for CPM. sMean percent of PM values are
derived from individual data sets. 6Values include filterable and condensible PM.
February 9, 2006
Recovery Furnaces
The recovery furnace data are a compilation of data in NCASI Technical Bulletins Nos. 852, 884, and as
yet unpublished NCASI data. All of the recovery furnaces in this data set use electrostatic precipitators
(ESP) for particulate control.
The PM data for DCE recovery furnaces is from the 23 sources listed in NCASI Technical Bulletin No.
884, Table AIle. The PM IO data for the DCE recovery furnaces is from the 4 DCE sources listed in
Technical Bulletin No. 884, Table Alld. The PM 2.5 data for DCE recovery furnaces is from the 4 DCE
sources listed in Technical Bulletin No. 884, Table Alld, plus a further two from as yet unpublished
NCASI data. The DCE CPM data is from two sources listed in Technical Bulletins Nos. 852 and 884,
and two sources from as yet unpublished NCASI data.
The PM data for the NDCE recovery furnaces is from the 20 sources listed in NCASI Technical Bulletin
No. 884, Table A12b. The PM 10 data for the NDCE recovery furnaces is from the 13 NDCE sources
listed in Technical Bulletin No. 884, Table A12c. The PM 2.5 data for NDCE recovery furnaces is from the
II DCE sources listed in Technical Bulletin No. 884, Table A12c, plus a further source from as yet
unpublished NCASI data. The NDCE CPM data is from 6 sources listed in Technical Bulletin No.884,
and one source from as yet unpublished NCASI data.
February 9, 2006
~-
Table 2. Recovery Furnace Data Summary
Kraft DCE Recovery Furnace
Parameter
Measurement
Method
No. of
Sources
PM
PM 10
PM2.5
CPM MeCh Soluble
CPM Water Soluble
CPM
Sulfate
EPA Method 5
EPACTM-040
EPACTM-040
EPA 202
EPA 202
EPA Method 202
IC
23
4
6
4
4
4
3
Range
Mean
(lb/ton BLS)
5Mean
Percent of PM
0.74
20.54
20.41
20.148
21.36
21.52
40.53
73
56
20
184
205
%ofCPM= 35
Range
Mean
(lb/ton BLS)
5Mean
Percent of PM
0.07 - 2.58
10.05 ­ 1.88
'0.04 - 1.44
10.014 - 0.516
10.13 -4.75
'0.14 - 5.29
30.05 - 1.85
Kraft NDCE Recovery Furnace
Parameter
Measurement
Method
No. of
Sources
PM
PM 10
PM 2.5
CPM MeCh Soluble
CPM Water Soluble
CPM
Sulfate
EPA Method 5
EPACTM-040
EPACTM-040
EPA 202
EPA 202
EPA Method 202
IC
20
13
11
3
3
7
2
0.02 - 3.50
'0.01 - 2.35
'0.01 - 1.82
'0.003 - 0.560
'0.016 - 2.84
10.02 - 3.40
30.007 - 1.16
0.65
20.44
20.34
20.104
20.53
20.63
40.21
67
52
16
81
97
% ofCPM = 34
[Range values were determined by applying the mean percent of PM to the range of values for PM. zMean values
were determined by applying the mean percent of PM to the mean value for PM. 3Range values for sulfate were
determined by applying the mean percent of CPM to the range of CPM values. ~ean value for sulfate was
determined by applying the mean percent of CPM to the mean value for CPM. 5Mean percent of PM values are
derived from individual data sets.
February 9, 2006
Lime Kilns
The lime kiln data are a compilation of data from NCASI Technical Bulletins Nos. 852, 884, and 898.
The emissions data are separated by control device type. The majority of lime kilns in this data set use
wet control devices for particulate control. Two of the lime kilns in this data set use an ESP for
particulate control, followed by a wet scrubber for S02 control. The remainder use an ESP for particulate
control.
The PM data for lime kilns using wet control devices is from 30 sources listed in NCASI Technical
Bulletin No. 884, Table Al3c. The PM IO and PM2.5 data for lime kilns using wet control devices is from
NCASI Technical Bulletin No. 884, Table Al3d. The CPM and sulfate data for lime kilns using wet
control devices is from Technical Bulletin No. 898.
All of the PM, CPM, and sulfate data for lime kilns using an ESP followed by a wet control device is
from two sources listed in NCASI Technical Bulletin No. 898.
The PM data for lime kilns using an ESP alone are from the 7 sources listed in NCASI Technical Bulletin
No. 884, Table Al3c. The PM 10 and PM2.5 data are from the 6 sources listed in Technical Bulletin No.
884, Table Al3d. The CPM and sulfate data are from 3 sources listed NCASI Technical Bulletin Nos.
852 and 884.
February 9, 2006
Table 3. Lime Kiln Data Summary
Lime Kilns with Wet Particulate Control Devices
11easurement
~o.of
Parameter
11ethod
Sources
Range
11ean
(gr/dsct) @10% O 2
511ean
Percent of P11
P11
EPA 11ethod 5
30
0.014 - 0.346
0.0995
6Tota1 PM IO
Dilution Tunnel
6
·0.014 - 0.349
20.100
101
6Total PM 2.5
Dilution Tunnel
7
·0.012 - 0.304
20.088
88
CPM 11eCh Soluble EPA 11ethod 202
3
14.2E-5 - 0.0010
20.0003
0.3
CPM Water Soluble
EPA 11ethod 202
3
10.0008 - 0.0208
20.0060
6
CPM
EPA 11ethod 202
3
10.0009 - 0.0218
20.0063
6.3
Sulfate
IC
2
30.0002 - 0.0046
40.0013
% ofCP11 = 21
"'These data are the result of dilution tunnel testing, therefore the P11 10 and P112.5 values reflect the sum of
filterable and condensible PM IO and P11 2.5 particulate.
Lime Kilns with a Dry ESP for Particulate Control Followed by a Wet Scrubber
Measurement
~o. of
Range
11ean
Parameter
Method
Sources
(gr/dsct) @1O% O 2
PM
PM IO
PM2.5
CP11 MeCh Soluble
CP11 Water Soluble
CP11
Sulfate
EPA Method 5
2
~oData
EPA 11ethod 202
EPA 11ethod 202
EPA 11ethod 202
IC
2
2
2
1
Lime Kilns with a Dry ESP for Particulate Control
Measurement
~o. of
Parameter
11ethod
Sources
PM
PM IO
PM2 .5
CP11 MeCh Soluble
CP11 Water Soluble
CP11
Sulfate
0.003 - 0.004
~o Data
EPA 11ethod 5
EPA CT11-040
EPA CT11-040
EPA Method 202
EPA Method 202
EPA Method 202
IC
7
6
6
3
3
3
3
0.0004 - 0.0081
0.0038 - 0.0054
0.006 - 0.012
0.002
0.004
Data
~o Data
0.0042
0.0046
0.009
0.002
~o
Range
11ean
(gr/dsct) @1O% O 2
0.002 - 0.033
10.001-0.211
10.0005 - 0.0079
10.0013 - 0.0208
10.003 - 0.045
10.004 - 0.066
30.0035 - 0.0581
511ean
Percent of P11
0.010
20 .0 06
20.0024
20.0063
20.014
20.020
40.0176
140
131
271
% ofCPM = 34
511ean
Percent of PM
64
24
63
137
200
%ofCPM=88
'Range values were determined by applying the mean percent of PM to the range of values for PM. 2Mean values
were determined by applying the mean percent of PM to the mean value for PM. 3Range values for sulfate were
determined by applying the mean percent of CPM to the range of CPM values. 4Mean value for sulfate was
determined by applying the mean percent of CPM to the mean value for CPM. 5Mean percent of PM values are
derived from individual data sets. 6Values include filterable and condensible PM.
Environment
AECOM
Appendix B
NCASI - Information on
Retrofit Control Measures for
Kraft Pulp Mill Sources and
Boilers for NOx, S02 and PM
Emissions
Five Factor BART Anal)'Sis for NewPage luke Mill
JUly 2010
ncasl
II
~..
NATIONAL COUNCIL FOR AIR AND STREAM IMPROVEMENT, INC.
P.O. Box 13318, Research Triangle Park, NC 27109-3318
Phone(919)941~400
Fax(919)941~1
Ronald A. Ye.ke, Ph.D.
President
(919) 941-6404
June 9,2006
TO:
Corporate Correspondents -- CC 06-014
Regional Managers
FROM:
RonaldA. Yeske
SUBJECT:
Information on Retrofit Control Measures for Kraft Pulp Mill Sources and Boilers
for NO x, S02 and PM Emis!!ions
~
The attached document "Retrofit Control Technology Assessment for NOx, S02 and PM
Emissions from Kraft Pulp and Paper Mill Unit Operations" was prepared to assist NCASI
member company personnel involved in conducting Best Available Retrofit Technology
(BARl) site-specific engineering analyses. It deals with the three main pollutants of concern
for BART analyses, namely NOx, S02 and particulate matter (PM). Potentially available control
technologies for these three pollutants for kraft recovery furnaces, lime kilns and boilers burning
wood, coal, gas, or oil are discussed. Also, control technologies for PM emissions from lime
slakers and smelt dissolving tanks are briefly reviewed.
Sources subject to BART analyses were generally built in the 1962 to 1977 time frame. Thus,
application of any control technologies to these sources will involve retrofits. Even though a
given technology may have been installed on newer more modem units, or may be theoretically
applicable, retrofitting the technology to an older existing unit requires consideration of unit­
specific and location-specific factors. In many situations, these factors would eliminate one or
more control technologies from consideration due to technical infeasibility or excessive costs.
As noted throughout this document, site-specific factors will playa critical role in BART
analyses.
...
.
.
This document does not directly address the cost-effectiveness ($Iton of pollutant removed)
of retrofit control measures. Site-specific information, including capital costs, operating and
maintenance costs, and annual capacity factors, must be considered in assessing the cost­
effectiveness of a given control technology to a particular emission source. Notsurprisingly,
the ranges in costs and potential emission reductions are expected to be very large.
For more information on this document, please contact Dr. Arun V. Someshwar, Principal
Research Engineer, at the Southern Regional Center office, phone (352) 331-1745, ext 226;
email [email protected]
Attachment
Retrofit Control Technology Assessment for NO", SOl and PM Emissions
From Kraft PUlp and Paper Mill Unit Operations
by Arun V. Someshwar, Ph. D., NCASI
1.0
Introduction
This document summarizes the general applicability of currently available emission control
technologies for NO", S~ and particulate matter (PM) to various pulp and paper mill sources.
. The three main unit operations in a kraft pulp mill that emit NO", S~ and PM are kraft recovery
furnaces, lime kilns and boilers. Boilers can be ofthe type which bum wood residues alone,
wood in combination with coal, gas or oil, or only fossil fuels. Particulate emissions can also
result from lime slakers and smelt dissolving tanks. Other pulp and paper mill sources for PM are
generally quite insignificant.
The origin and nature ofthe three pollutants in each relevant pulp mill unit operation is first
discussed. Such discussion should be useful in understanding why some control technologies,
while being suitable candidates for certain unit operations in other industries, may not be suitable
in the pulp and paper industry. It is hoped this document will be useful in the context of a Best
Available Retrofit Technology (BART) site-specific engineering analysis.. However, it must be
clearly noted that for any retrofit technology, site-specific considerations for a given emission
source may disqualifY a particular control technology from consideration, even though it might
theoretically be feasible or may even have been installed elsewhere on a new, modern unit or a
greenfield operation.
.
.
Cost and emission reduction estimates are specifically not covered in this document. However, it
is instructive to consider that a wide range in costs and potential emission reductions are expected
due to the fact that site-specific factors playa critical role in detennining how cost-effective
various technologies will be in practice. Many facilities are space-limited, have controls already
in place, or have older combustion equipment that cannot be retrofit to reach required conditions,
making installation of certain technologies problematic or very expensive.
2.0
Kraft Recovery Furnaces
2.1
NO" Control
Compared to coal- or residual oil-fired boilers of similar capacity, NO" emissions from kraft
recovery furnaces are generally quite low, typically in the 60 to 130 ppm range. These low NO"
emissions are due to several factors inherent to kraft recovery furnace operations which include
(a) low nitrogen concentrations in most "as-fired" black liquor solids (generally <0.2%), (b)
recovery furnace NO" fonnation resulting predominantly from "fuel NO,," mechanisms
(insufficient temperatures for "thennal NO,," fonnation), (c) the highly staged combustion design
of recovery furnaces, and (d) the existence of sodium fumes that might participate in "in-furnace"
NO" reduction or removal.
­
Researchers have concluded that nearly two-thirds to three-fourths ofthe liquor N is released
during pyrolysis or devolatilization, partly as NH3 and partly as N2 , the rest remaining with the
smelt product most likely as a reduced N species. The ammonia released from the black liquor
during pyrolysis partly oxidizes to NO and partly reduces to N2 • A review of the theoretical
kinetics governing the reactions between NH 3 , NO, and O2 suggests that, in the presence of
National Council for Air and Stream Improvement
June 4, 2006
2
~-
excess 02, a decrease in temperature decreases the degree of oxidation ofNH 3 to NO, thus
implying that fuel NOx generation during black liquor combustion is more temperature-dependent
than previously thought. However, a reduction in furnace temperatures, particularly in the lower
furnace, is generally expected to result in a sharp increase in S~ emissions from the furnace.
Most of the NO is formed by oxidation of the NH3 volatilized during pyrolysis of the liquor
droplets. Very little NO is formed from the N in the char bed. In certain instances, where the
liquor droplet dries completely before reaching the char bed, additional NO can be formed during
"in-flight" char combustion of the liquor droplet. The use of liquor sprays resulting in larger
droplet sizes avoids the problem of additional NO contribution from char burning.
Some have observed that NOx emissions increased when firing liquors with increasing liquor
solids contents. However, this may have had less to do with- thermal NOx or an "in-furnace"
capability ofalkali fume to capture NO x as suggested by some, but more to do with a possible
effect on increased conversion ofammonia to NO within the furnace due to an increase in lower
furnace temperatures resulting from firing higher solids liquors.
2.1.1
Low NO x Burners
The use of low-NO x burners (LNB) for black liquor combustion has not been demonstrated.
Unlike fossil fuels, black liquor has a large quantity of water and the drying, pyrolysis, and char
burning of liquor droplets occurs over a long flight trajectory from the iiquor guns to the char bed,
thus making unavailable the benefits of staged combustion inherent inLNB designs.
LNBs could however be applied to oil guns or gas burners in recovery furnaces that are used to
supply supplemental heat or for start-up/shut down purposes. However, for most recovery units,
the lise of auxiliary fuel is very limited; in such cases the benefit from conversion to LNB would
be marginal.
2.1.2
Staged Combustion
Recent research has concluded that to the extent "staged combustion" is allowed to take place in
the upper furnace during oxidation ofthe volatilized NH3 to NO, such oxidation can be
minimized. Limited short-term experience after installing "quaternary" air ports in two U.S.
furnaces showed that a 20 to 40% reduction in baseline NOx levels is feasible using such air
staging. However, to make it feasible to install a quaternary air system a recovery furnace
typically needs to be fairly large in size. Thus this option would not be feasible for most BART­
eligible recovery furnaces, since units built in the 1962 t 0 1977 time period were considerably
smaller than those installed in subsequent years.
2.1.3
Flue Gas Recirculation (FGR)
Flue gas recirculation (FOR) is also not a viable option for kraft recovery furnaces. In FOR, a
portion of the uncontrolled flue gases is routed back to the combustion zone, primarily with the
intention of reducing thermal NOx• Thermal NOx is, however, not a concern in recovery furnaces,
as discussed earlier. FORwould add additional gas volume in the furnace, increasing velocities
and potentially causing more liquor carryover, which would result in increased fouling ofthe
recovery furnace tubes.
National Council- for Air and Stream Improvement
June 4,2006
2.1.4
3
Oxygen Trim + Water Injection
Oxygen-trim + water injection, a NOx control technology generally utilized in natural gas-fired
boilers, would not be relevant to kraft recovery furnaces since (I) any injection of water into the
furnace would lead to an unacceptable explosive condition and (2) the oxygen trim technique
would have marginal effect due to the already existing highly staged combustion air configuration
in recovery furnaces.
2.1.5
Selective Non-Catalytic Reduction (SNCR)
At the current time, there is no published information on the extended use of SNCR on an
Qperating kraft recovery furnace. Short-term tests with the SNCR technology have been reported
in the literature on two furnaces in Japan and one in Sweden. There are a ntimber of critical,
unresolved issues surrounding the use of urea or ammonia injection in a kraft recovery furnace
for NOx control over a long-term basis. A kraft recovery furnace is the most expensive unit
operation in a pulp mill since its primary purpose is to recover chemicals from spent pulping
liquors in a safe and reliable manner. Although steam is generated from liquor combustion,
certain chemical recovery steps have to be accomplished inside the furnace. It is not known
whether the injection of NO,,-reducing chemicals into the furnace would have deleterious effects
on the kraft liquor recovery cycle on a long-term basis. Long-term tests would need to be carried
out to address this important issue. In addition, there are several other factors that make the use
ofSNCR in a kraft recovery furnace problematic such as (1) the impact of large variations in flue
gas temperatures at the superheater entrance due to fluctuating load and liquor quality, (2) limited
residence times for the NO,,-NH3 reactions available in smaller furnaces, (3) impact on fireside
deposit buildup due to reduced chloride purging from long-term NH:Jurea use and resulting
impact on tube corrosion and fouling, and (4) potential for significant NH3 slip and plume opacity
problems due to NH4Cl emissions. Unless these concerns are satisfactorily resolved, the use of
SNCR in a kraft recovery furnace should not be considered as a feasible technology.
2.1.6
Selective Catalytic Reduction (SCR)
The use of SCR on a kraft recovery furnace has never been demonstrated, even on a short-term
. basis. The impact of high particulate matter concentrations in the economizer region and fine
dust particles on cat.alyst effectiveness is a m~or impediment to the application of this technology
ahead of PM control, as is catalyst poisoning by soluble alkali metals in the gas stream. For SCR
installation after an ESP, the gas stream would be too cold for effective reaction with the NO". A
substantial energy penalty would have to be incurred to reheat the flue gas prior to the SCR
section which would be a major drawback.
2.1.7
Summary
In swnmary, optimization of the staged combustion principle within large, existing kraft recovery
furnaces to achieve lower NO" emissions might be the only technologically feasible option at the
present time for NO" reduction. However, the effect of such air staging on emissions of other
pollutants, chiefly SOI, CO, and TRS, and other furnace operational characteristics needs to be
examined with longer-term data on U.S. furnaces. Ultimately, the liquor nitrogen content, which
is dependent on the types of wood pulped, is the dominant factor affecting the level of NO"
emissions from black liquor combustion in a recovery furnace. Unfortunately, this factor is
beyond the control of pulp mill operators.
National Counal for Air and Stream Improvement
June 4,2006
4
2.2
SOl Control
Black liquor contains a significant amount of sulfur, nominally 3 to 5% by weight of the
dissolved solids. While the vast majority of this sulfur leaves the furnace in the smelt product, a
small fraction (generally under 1%) can escape in gaseous or particulate form. Average S02
concentrations in stack gases can range from nearly 0 to 500 ppm, although most furnaces
currently operate with <100 ppm S02 in stack emissions. Factors which influence S02 levels are
liquor sulfidity, liquor solids content, stack oxygen content, furnace load, auxiliary fuel use, and
furnace design. However, none of these factors has exhibited a consistent relationship with S02
emissions. At the present time, it is generally understood that conditions involving liquor quality
(such as high Btu, high solids liquors) and liquor firing patterns and conditions related to furnace
operations (air distribution, auxiliary fuel, etc.) that lead to maximizing temperatures in the lower
furnace result in minimizing S02 emissions from kraft recovery furnaces.
There is no experience in the pulp and paper industry with the use of dedicated, add-on flue gas
desulfurization technologies on kraft recovery furnaces. Although there are a few scrubbers on
U.S. kraft recovery furnaces, none of these were installed for S02 removal. Only one U.S.
recovery furnace does not use an ESP for particulate control; this unit has venturi scrubbers
instead. All of the other scrubbers follow an ESP. Two were installed for heat recovery reasons,
although some S02 scrubbing may also be occurring especially when caustic is added to the
scrubbing solution. One scrubber following an ESP was installed with the main purpose of
achieving incremental particulate matter removal. Another scrubber following an ESP was
installed on a furnace with a direct contact evaporator to control black liquor droplets being
entrained in the cascade and traveling all the way throughout the ESP and out the stack. Even if
these scrubbers had been installed to reduce S02 emissions, the removal costs in terms of dollars
per ton ofS02 removed would be large due to high gas flows and site-specific retrofit
considerations. Significant capital would be required for the large gas handling equipment and
additional induced fan capacity needed to overcome the increased pressure drop across the
scrubber.
2.3
Particulate MaUer Control
Recovery furnaces are designed and operated in a manner so as to ensure the presence of high
levels of sodium fumes in order to capture the sulfur dioxide produced as a result of oxidation of
reduced sulfur compounds. Consequently, uncontrolled recovery furnace flue gases contain high
levels of particulate matter. The uncontrolled particulate matter load from recovery furnaces is
highly variable and has been reported to range from 100 to 250 Ib/ODTP (oven dry ton pulp) for
direct contact evaporator (OCE) furnaces and 200 to 450 Ib/ODTP for non-direct contact
evaporator (NDCE) furnaces. The lower particulate loading from DCE furnaces is due to the
capture of some particulate matter in the direct contact evaporator. ESPs built for NOCE
furnaces are designed to compensate for the higher particulate loading.
Particulates generated in the recovery furnace are comprised mainly of sodium sulfate, with lesser
amounts of sodium carbonate and sodium chloride. Similar potassium compounds are also
generated. but in much lower amounts. Trace amounts of other metal compounds, e.g.
magnesium, calcium, and zinc, can be present. A significant portion of the particulate material is
sub-micron in size, which makes removal with additional add-on control devices more difficult.
Increasing liquor firing density (tonlday/fl?) increases recovery furnace particulate loading. Other
factors such as bed and furnace temperature, liquor solids, liquor composition, and air distribution
also affect uncontrolled particulate emissions from recovery furnaces.
National Council for Air and Stream Improvement
.-
.
5
June 4, 2006
ESPs are the device ofchoice for controlling PM emissions from kraft recovery furnaces. The
use oflarger ESPs is expected to result in better overall PM capture efficiencies. However, this
option is expected to be quite cost ineffective based on the high, site-specific, retrofit costs
incurred. Moreover, with the implementation of MACT IJ limitations in 2004, most recovery
furnaces are operating at or below NSPS levels (NCASI Corporate Correspondents Memo 01-0 I).
Any 8dditional benefit would thus be marginal.
3.0
Kraft Lime Kilns
3.1
NO. Control
NO. emissions from lime kilns result mainly from fossil fuel burning (natural gas and fuel oil). A
recent NCASI study involving NO. testing at IS lime kilns verified that ''thermal'' NO. was the
sole mechanism operative in gas-fired kilns, while the "fuel" NO. mechanism was mostly
operative in oil-fired kilns. Gas-fired kiln NO. emissions appeared to be strongly dependent on
the dry-end lime temperature. Oxygen availability in the combustion zone was determined to be
the key factor in oil-fired kilns. NO. emissions for gas-fired kilns also exhibited high short-term
variability, unlike for oil-fired kilns. Analysis of long-term daily average data from two lime
kilns showed no difference in NOx emissions between days with and without LVHC NCG
burning. An earlier NCASI study had shown that when stripper off-gases (SOGs) containing
ammonia were burned in lime kilns, a small fraction ofthe ammonia, up to 23%, converts to NOx'
A BACT analysis conducted for a new lime kiln in 1997 concluded that the use of low NOx
burners in lime kilns was technically infeasible due to complexities resulting in poor efficiency,
increased energy usage, and decreased calcining capacity ofthe lime kiln. The concept of'low
NOxburners' is considered a misnomer in the rotary kiln industry. In boiler burners where the
combustion air can be staged, 'low NOx' could be a genuine option. However, in rotary kilns it is
not possible to stage the mixing in the same way. There has to be sufficient primary (burner) air
to provide control in flame shaping although this can be limited to minimize NO. to some extent.
Effectively, the NO. can be reduced to some extent by 'de-tuning' the burner from optimized
combustion. However, the result is an energy penalty by way ofa higher heat input per ton
product and higher feed-end temperatures.
Post-combustion flue gas NO. control using SCNR or SCR is not feasible due to the
configuration of the kraft lime kiln. The necessary temperature window of 1500°F to 2000°F for
reagent injection in the SNCR process is unavailable in a kraft lime kiln. The very high PM load
prior to control would make SCR infeasible in advance of the controls and the requisite
temperature window of between 550°F and 750°F for applying SCR after a PM control device is
unavailable for a lime kiln, even for one equipped with an ESP.
Thus, NO. control in newer lime kilns may be achieved mainly by minimizing the hot end
temperatures in gas-fired kilns and by reducing the available oxygen in the combustion zone in
oil-fired kilns, both combustion related modifications. However, these modifications may be
difficult to achieve in certain existing kilns due to their inherent design. For example, in order to
complete the calcining reactions in kilns with short residence times, it is more difficult to control
hot end temperatures in shorter kilns than in longer ones.
National Council for Air and Stream Improvement
I
. . ~ .••••_• •
6
3.2
June 4, 2006
S02 Control
Sulfur dioxide is fonned in lime kilns when fuel oil or petroleum coke is burned as primary fuel.
S02 will also be fonned if non-condensible gases (NCGs) or stripper off-gases (SOGs) containing
sulfur are burned in the kiln. Lime muds also contain a small amount of sulfur, which when
oxidized, would form S02. Median sulfur content of concentrated NCGs and SOGs have been
reported as 1.1 and 4.2 Ib/ADTP (air dried ton pulp), respectively. Median sulfur contents of?
lime muds have been reported at 0.20/0, which translates to about 1.81b S/ADTP. Thus, fossil
fuels such as fuel oil, kraft mill NCG/SOGs, and soluble sulfides in lime mud can contribute a
significant amount of sulfur to the inputs of a lime kiln. Nevertheless, the regenerated quicklime
in the kiln acts as an excellent in-situ scrubbing agent, and venturi scrubbers following the kiln
can further augment this S02 removal process since the scrubbing solution becomes alkaline from
the captured lime dust. Consequently, even though the potential for S02 formation in a kiln that
burns sulfur-containing fuels with or without NCGs/SOGs is high, most lime kilns emit very low
levels ofS~ (~50 ppm). Some kilns do, however, occasionally emit higher levels ofS~ (50 to
200 ppm). Not much is known about why this happens.
Emission test data show that S02 concentrations do not appear to be related to either the fuel type
(oil, gas) or the presence or absence of concentrated NCG or SOG burning in the kiln. A
preliminary sulfur input-output balance carried out on 25 kilns with wet scrubbers and 7 kilns
with electrostatic precipitators (ESPs), with sulfur inputs from fuel oil, NCGs and SOGs, or just
lime mud, showed over 95% of the S02 generated from the oil, NCG/SOGs, or lime mud was
captured within the kiln. For kilns with wet scrubbers (majority) that have high S02 emissions,
. alkali addition to the scrubbing fluid could further reduce the S~ emissions.
3.3
Particulate Matter Control
While passing through the kiln, the combustion gases pick up a good deal of particulate matter
both from lime mud dust formation and from alkali vaporization. This PM must be removed
before the gases exit to the atmosphere. Mechanical devices such as dust chambers or cyclones
are generally used to remove larger particles, which are mainly calcium-containing. A wet
scrubber or electrostatic precipitator follows for removal of smaller particulates, which are mainly
sodium sulfate and sodium carbonate and have aerodynamic diameters less than 10 1J11l.
Kraft lime kiln PM emissions are typically controlled by venturi-type wet scrubbers. Scrubbers
with increasingly better PM removal efficiencies, such as the Ducon Dynamic Wet Scrubber,
have been installed up until the late 1980s. However, most of the PM control installations on
lime kilns since about 1990 have been ESPs. Replacing a wet scrubber with an ESP will most
likely reduce PM emissions, but may increase emissions ofS02. The wet scrubber acts as an
additional alkaline S0:2 scrubber since it captures alkaline PM leaving the kiln. Just as for
recovery furnaces, with the implementation of MACf II limitations in 2004, most lime kilns are
operating at or below NSPS levels. Any additional benefit would thus be marginal.
4.0
Boilen
The majority of pulp and paper industry boilers are combination boilers, in that they are designed to
burn more than one fuel. Thus, it should be noted that while a particular teclmology may be beneficial
for a particular pollutant, the same teclmology may not address the control of another pollutant. For
example, a wood-fired boiler with a wet scrubber for PM control may obtain better PM control with
an ESP. However, if the boiler also fires some sulfur-containing fuel (as is often the case), the
S02 removal capability of the wet scrubber will be sacrificed by the installation of an ESP.
National Council for Air and Stream Improvement
...........
June 4, 2006
4.1
7
Natural Gas-Fired Boilers
Gas-fired boilers are usually not equipped with particulate collectors. S02 emissions depend on
the sulfur content of the gas, which is typically negligible. NOx emissions are dependent on the
combustion temperature and the rate of cooling of the combustion products. There are several
combustion modification techniques available to reduce the amount of NOx formed in natural
gas-fired boilers and turbines. The two most prevalent ones are flue gas recirculation (FGR) and
10w-NOx burners. FGR reduces formation of thermal NOx by reducing peak temperatures and
limiting availability of oxygen. Low-NOx burners reduce formation of thermal NOx by delayed
combustion (staging) resulting in a cooler flame. In conjunction with FGR, the burners can
achieve NOx emission reductions of 60 to 90%. Other techniques include staged combustion and
gas rebuming. In general, these techniques have been incorporated in newer boilers and thus
their NOx emissions are lower than those of older units.
.
There are also add-on control technologies that can reduce NOx emissions from gas-fired boilers
such as selective non-catalytic reduction (SNCR) and selective catalytic reduction (SCR).
However, since most of the pulp and paper industry gas-fired boilers are of the package boiler
. type, cost considerations typically make the use of such technologies cost ineffective. Further,
both the SNCR and SCR technologies have not been proven to apply to industrial boilers with
frequent swing loads.
4.2
Fuel Oil-Fired Boilers
For fuel oil-fired boilers, criteria pollutants can be controlled by fuel SUbstitution/alteration,
combustion modification and post-combustion control. Fuel substitution reduces S~ and NO"
and involves burning an oil with lower S or N content, respectively. Particulate emissions are
lower when burning lower sulfur content oils, especially distillate oil.
4.2.1
NOx Control
For boilers burning residual oil, fuel NOx is the dominant mechanism for NOx formation and thus
the most common combustion modification technique is to suppress combustion air levels below
the theoretical amount required for complete combustion. There are several combustion
modification techniques available to reduce the amount of NO x formed in fuel oil-fired boilers,
including low excess air, burners out of service, biased-burner firing, flue gas recirculation,
overfire air, and 10w-NOx burners. NO x reductions that could range between 5 and 60% from
uncontrolled systems may be expected fr~m using these techniques.
Post-combustion controls include SNCR and SCR. NOx reductions from 25 to 0% and from 75 to
85% may be expected from use ofSNCR and SCR systems on oil-fired boilers, respectively.
However, just as for gas-fired boilers, most of the pulp and paper industry oil-fired boilers are of
the package boiler type, and cost considerations typically make the use of such technologies cost
ineffective. Furthermore, both the SNCR and SCR technologies have not been proven to apply to
industrial boilers with frequent swing loads.
4.2.2
SOl Control
S~
emissions are controlled by a number of commercialized post-combustion flue gas
desulfurization (FGD) processes which use an alkaline reagent to absorb S~ in the flue gas and
produce a sodium or calcium sulfate compound. The FGD technologies may be wet, semi-dry or
dry depending on the state ofthe reagent as it leaves the absorber vessel.
National Council for Air and Stream Improvement
........ -..
B
4.2.3
June 4. 2006
Particulate Matter Control
Due to the extremely low level of PM emissions, most residual oil-fired boilers do not have
particulate matter controls. A few boilers are, however, equipped with mechanical collectors or
ESPs.
4.3
Coal-Fired Boilers
4.3.1
NOs Control
NOx emissions from coal-fired boilers can be controlled by a) combustion controls and b) post­
combustion controls. Combustion controls involve a) reducing peak temperatures in the
combustion zone, b) reducing gas residence time in the high-temperature zone, and c) air or fuel
staging by operating at an off-stoichiometric ratio by using a rich fuel-air ratio in the primary
flame zone and lower overall excess air conditions. The use of combustion controls depends on
the type of boiler and the method ofcoal firing. Low-NOx burners and overfireair (OFA) have
been successfully applied to tangential- and wall-fired units, whereas rebuming is the only current
option for cyclone boilers. For large base-loaded coal-fired boilers, the most developed and
widely applied post-combustion NOx control technology is SCR. Catalyst deactivation and
residual NH 3 slip are the two key operating considerations in an SCR system. There is only
limited experience with the use ofSNCR systems on industrial coal-fired boilers. NOx reductions
from 30-70% and from 60-90% may be expected from use of SNCR and SCR systems on base­
loaded coal-fired boilers, respectively. SNCR has narrow temperature window in which it is
effective, in the 1500 to 1900°F range, and SCR has a similar, but lower temperature window of
550 to 750°F. When applied to industrial boilers, neither the SNCR nor the SCR technologies
have been proven to yield the same high NOx removal efficiencies expected when the boilers
operate at base loads as when they operate with frequent swing loads. The inability to maintain
good control within the required temperature window during swing loads is most likely
responsible for this reduction. Most coal-fired boilers in the pulp and paper industry operate in
the swing load mode, a function of supplying stearn as required to the various components ofthe
process.
a
4.3.2· SOl Control
Just as in fuel oil combustion, criteria ~Ilutants can be controlled by fuel SUbstitution/alteration,
combustion modification and post-combustion control. S~ reductions can be achieved by
burning a coal with lower S content. 802 emissions can be controlled by a number of
commercialized post-combustion flue gas desulfurization (FGD) processes which use an alkaline
reagent to absorb SOl in the flue gas and produce a sodium or calcium sulfate compound. The
FGD technologies may be wet, semi-dry or dry depending on the state of the reagent as it leaves
the absorber vessel. The pulp and paper industry has limited experience with operating FGD
systems on coal- or oil-fired boilers. Retrofit considerations include space restraints in many
facilities.
4.3.3
Particulate Matter Control
Particulate emissions from coal-fired boilers are controlled by using a) ESPs, b) fabric filters (FF)
or c) venturi scrubbers. Multi-cyclones are generally used as precleaners upstream of more
efficient ESPs or FFs. The key operating parameters that influence ESP performance include fly
ash mass loading, particle size distribution, fly ash resistivity (which is related to coal sulfur
content), and precipitator voltage and current. Data for ESPs applied to coal-fired boilers show
National Council for Air and Stream Improvement
......-..
,.
June 4, 2006
9
fractional collection efficiencies greater than 99% for fine (<0. I~) and coarse particles (> IO
I!m) and a reduction in collection efficiency for particles between O. I and 10 11m. Operational
parameters that affect fabric filter collection efficiency include air-to-cloth ratio, operating
pressure loss, cleaning sequence, interval between cleanings, cleaning method, and cleaning
intensity. Collection efficiencies of fabric filters can be as high as 99.9%. Scrubber collection
efficiency depends on particle size distribution, gas side pressure drop through the scrubber, and
water (or scrubbing liquor) pressure, and can range from 90 to 95% for a 2 ~ particle.
4.4
Wood-Fired Boiler Emissions
4.4.1
NO" Control
Most large wood-fired boilers used in the pulp and paper industry are of the spreader stoker
design. NO" control technologies effective for use on gas and oil burners are not applicable to
spreader-stoker design boilers. Furthermore, these boilers are often operated handling swing
loads, which makes add-on NO" controls difficult to implement. Spreader stoker boilers
inherently practice staged combustion, which lowers NO" emissions, but within limits.
Fuel NO" is the dominant NO~ formation mechanism operative during wood combustion. Fuel
NO" is most efficiently controlled by staged combustion. Overfire air ports inherent to most
spreader-stoker boilers provide for staged combustion. The underfire and overfire air are
balanced in most wood-fired spreader stokers to control No,..
As with other fuels, potential post-combustion controls include SNCR and SCR. SNCR has been
applied to a few base-loaded wood-fired boilers, mainly in the electric generating industry.
However, its long-tenn efficacy on wood-fired boilers with changing loads has not been
demonstrated. Experience in the pulp and paper industry to date has shown it has been used on
occasions for polishing, to get perhaps 10-20% NO" reduction during periods of air quality
problems. The problem with control of the required temperature window is an inherent difficulty
with use of SNCR for load-following boilers, whether wood or fossil fuel. Inadequate reagent
dispersion in the region of reagent injection in wood-fired boilers is also a factor mitigating
against the use of SNCR technology. At least one pulp mill wood-fired boiler met with
significant problems and had to abandon their SNCR system. Significant ammonia slip, caused
by inefficient dispersion of the reagent within the boiler, was to blame.
. The use of SCR on wood-fired boilers operating in the forest products industry has also never
been successfully demonstrated for spreader stoker boilers, and would face the same inherent
problem of requiring it to be post PM-control to protect the catalyst, and achieving and
maintaining the required temperature window for effective NO" control.
4.4.2
Particulate Matter Control
Particulate matter is the air pollutant of primary concern in wood-fired boilers. As for coal-fired
boilers, the most common devices used to control particulate emissions from wood-fired boilers
are wet scrubbers and electrostatic precipitators (ESPs). Fabric filters (FF) and the electrified
gravel bed filter (EGF) have been used on a few units. Wet scrubbers are widely used, operating
at gas pressure drops ranging from 6 to 25" H 20. Liquid to gas ratios in the venturi system
typically range from 8 to 10 gal H20/lOoo aefm saturated. Solids buildup in the recirculation
loop rarely is allowed to exceed 5% by weight. High carbon ash resulting from wood combustion
is more difficult to remove with an ESP due to its high conductivityllow resistivity. Thus,
specific collection areas (ratio of ESP plate area to gas flow volume through the ESP) for ESPs
National Council for Air and Stream Improvement
...........
June 4,2006
10
on wood-fired boilers are greater than for those for coal-fired boilers, ranging from about 300 to
500 ft2/l 000 acfm. Power requirements range from 150 to about 400 watts per acfm. To address
fire concerns, ESPs on wood-fired boilers are sometimes operated in the wet mode, where the
col1ection plates and internal parts are wetted continuously with water. A pre-quench is generally
used to saturate the gas stream. Fabric filters are rarely used on wood-fired boilers due to
concerns about bag flammability. Fabric filters have been successfully used where bark from
logs stored in salt water is burned and the salt reduces the fire hazard. In this situation, the fabric
filter is effective in removing the very small salt particulates exiting the boiler. Gravel-bed filters
have a slowly moving bed of granular "rock" as the filtration medium through which the flue gas
must travel. These systems are electrostatically augmented (10 to 20 watts/lOOO acfm). A high
voltage (about 50 kV) is applied to an electrical conductor positioned within the bed and this
creates an electrical field between the conductor and the inlet and outlet louvers. Particulate
collection efficiencies for wood-fired boilers range from 65 to 95% for two multiclones in series,
over 90% for wet scrubbers, from 93 to 99.8% for ESPs and FFs and about 95% forEGFs. Once
again, it should be noted that most wood-fired boilers are combination boilers that may burn other
sulfur-containing fuels. Thus, a change in the control device might affect the ability to control
other pollutants. For example, replacing a wet scrubber with an ESP for better PM control would
result in higher S02 emissionsfrom a boiler burning wood in combination with oil or coal.
5.0
Other Source Emissions
5.1
Siakers - PM emissions
Slakers are generally vented through a stack to discharge the large amounts of steam generated.
The stearn may contain particulate matter, which is largely calcium and sodium carbonates and
sulfates. Scrubbers are generally employed to capture this particulate matter. Other PM control
devices such as ESPs and fabric filters are both technologically infeasible (very high moisture
source)and not cost effective.
5.2
Smelt Dissolving Tanks - PM Emissions
As with the recovery furnace, particulate emissions from smelt tanks are comprised of mainly
sodium compounds with much lesser amounts of potassium compounds and some other trace
metal compounds. The dominant compound is sodium carbonate, followed by sodium sulfate.
Roughly 90% (by weight) of the particles have equivalent aerodynamic diameters under 10 IJm,
and 50% have diameters under I J.l.m. Most smelt tank PM emissions are controlled by wet
scrubbers, many of which are wetted fan scrubbers that are very effective in removing fine
particulate. A dry ESP is once again infeasible as an option due to the high moisture content of
the gases. The wet scrubber also serves to control total reduced sulfur compound emissions
through pH control, thus replacing it with a wet ESP is not an option. As noted for other kraft
mill sources, MACT II Implementation in 2004 has also resulted in significantly reduced
allowable PM emissions from smelt dissolving tanks.
National Council for Air and Stream Improvement
Environment
AECOM
Appendix C
Modeling Archive CO Available
on Request from MOE
July 2010
Five Fector !!ART Analysis for NewPage Luke Mill
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement