ATTACHMENT 1

To NPS/FWS Comments - Maine Draft Regional Haze SIP

NPS Comments Regarding Verso Androscoggin Paper Mill BART Evaluation

July 23, 2010

Power Boilers #1 & #2: NOx

The following statement by Verso is misleading:

The Androscoggin Mill followed the guidance and procedures outlined in
40 CFR Part 51, Appendix Y and the OAQPS Air Pollution Cost Control
Manual. Supporting cost evaluation spreadsheets are provided in
Attachment C, Table Nos. C-1, C-2, C-3, and C-4.

While we applaud Verso’s intent to use the Cost Manual, the actual
Verso approach appears to have borrowed the Cost Manual method for
evaluating wet scrubbers and applied it to SCR and SNCR, which we
believe is inappropriate.

 

In actuality, there is no OAQPS Air Pollution Cost Control Manual (Cost
Manual) procedure for evaluating costs for SCR or SNCR for oil-fired
EGUs. The procedures described by the Cost Manual are intended for use
with coal-fired boilers > 250 mmBtu/hr. So we adapted them to oil-fired
boilers, but the cost algorithms for the Direct Capital Costs are from
the Cost Manual coal-boiler method and therefore questionable. 

Even if we accept the Verso approach as a default, it still contains
some highly questionable estimates for SCR, and Verso clearly did not
follow the Cost Manual:

If we assume that Power Boilers #1 & #2 are capable of producing about
68 MW each, then the Total Capital Investment (TCI) per kW is about $115
for SCR, which is in the middle of the   $50 - $260/kW range for
coal-fired EGUs. The attached Excel workbook estimates a slightly higher
TCI by applying our adapted Cost Manual approach.

Verso has estimated an annual reagent cost of $414,000/boiler. This
exceeds the $54,000 annual reagent cost that the Cost Manual procedure
estimates. Verso must justify this estimate.

Verso has estimated an annual catalyst replacement cost of
$155,000/boiler. Since this exceeds the $92,000 annual catalyst
replacement cost that the Cost Manual procedure estimates for the 330 MW
Naughton Unit #3 (that Wyoming is requiring to install SCR as BART), the
Verso estimate appears to be very high. Our adapted Cost Manual method
estimates catalyst volume at 88 m3, a 24,000 hour catalyst life, and an
annual Catalyst Replacement Cost = $41,000/boiler. Furthermore, because
most catalyst vendors do not charge for recovery of the spent catalyst,
that $30,000 annual cost also appears unfounded.

Verso's Capital Recovery Factor (12.4% interest over a 10-year SCR life)
is inflated. The Cost Manual recommends 7% interest over a 20-year SCR
life.

Verso estimates an annual cost of $5.1 million to control both boilers
versus our estimate of $1.1 million for each boiler, and Verso estimates
$7,361/ton versus our $3,070/ton.

According to Maine Department of Environmental Protection (MEDEP):

The cost effectiveness numbers in the table above are based on
controlling NOx emissions from Power Boilers #1 and #2 at the control
effectiveness rates indicated in the table from the highest estimated
two year average annual emissions between 2002 and 2008.  In recent
years (2008 and 2009) these boilers have been operating close to only
20% of the time, which for example, would result in an actual cost
effectiveness of $16,313 per ton of NOx removed with the installation of
SCR. 

MEDEP estimates cost-effectiveness at $5,271/ton versus the $7,361/ton
estimated by Verso; we request an explanation for this difference.
Furthermore, if MEDEP intends to consider the reduced operation of these
boilers in the economic analysis, those reduced operational parameters
should be made federally enforceable if they affect the outcome of the
analysis.

Because BART is a visibility improvement program, we believe that
cost/deciview ($/dv) is a very important parameter. In this case, for
the four Class I areas evaluated by Verso, SCR would improve visibility
by a total of 4.6 dv. (We would also like to see the visibility
improvements that would occur in the other two Class I areas.) This
results in a cost-effectiveness value of less than 0.5 million/dv, which
is quite reasonable compared to the average $13 - $20 million/dv that we
are seeing accepted by states and sources that are proposing reductions
under BART. Even if one considers only the visibility improvement at
Acadia National Park, the addition of SCR results in a
cost-effectiveness value of $1.3 million/dv. This leads to the
conclusion that SCR is BART for the Androscoggin power boilers.

The same situation applies to SNCR. The actual Verso approach appears to
have borrowed the Cost Manual method for evaluating wet scrubbers and
applied it to SNCR, which we believe is inappropriate. Even if we accept
the Verso approach as a default, it still contains some highly
questionable estimates for SNCR:

If we assume that Power Boilers #1 & #2 are capable of producing about
68 MW each, then the Total Capital Investment (TCI) per kW is about $47
for SNCR, which is on the high end of the   $29 - $45/kW range we are
seeing in proposals to install SNCR on coal-fired EGUs (See  HYPERLINK
"http://www.wrapair.org/forums/ssjf/bart.html"
http://www.wrapair.org/forums/ssjf/bart.html ). The attached Excel
workbook estimates $26/kW. Verso should provide vendor quotes to support
its higher-than expected estimates.

Verso has estimated a Direct Annual Cost (DAC) of $0.55 million/boiler.
Since this exceeds the $0.12 million DAC that the Cost Manual procedure
estimates, the Verso estimate appears to be very high. The biggest
difference is in Verso's estimate of almost $0.5 million/year/boiler for
reagent versus the Cost Manual estimate of $0.06 million/yr.

Verso's Capital Recovery Factor (12.4% interest over a 10-year SNCR
life) is inflated. The Cost Manual recommends 7% interest over a 20-year
SCR life.

Verso estimates an annual cost of $2.6 million to control both boilers
versus our estimate of $0.29 million for each boiler, or Verso's
$9,758/ton versus our $2,128/ton.

MEDEP estimates cost-effectiveness at $5,973/ton versus the $9,758/ton
estimated by Verso; we request an explanation for this difference. 

In this case, for the four Class I areas evaluated by Verso, SNCR would
improve visibility by a total of 4.3 dv. (We would also like to see the
visibility improvements that would occur in the other two Class I
areas.) This results in a cost-effectiveness value of less than 0.13
million/dv, which is quite reasonable compared to the average $13 - $20
million/dv that we are seeing accepted by states and sources that are
proposing reductions under BART. Even if one considers only the
visibility improvement at Acadia National Park, the addition of SCR
results in a cost-effectiveness value of $0.41 million/dv. This leads to
the conclusion that SNCR could also be a candidate for BART for the
Androscoggin power boilers if SCR is rejected.

Power Boilers #1 & #2: SO2

Some comments on Verso's BART analysis for SO2 from the Androscoggin
mill Power Boilers #1 & #2.

Power Boilers #1 & #2 wet scrubber cost analysis

Verso's Purchased Equipment Costs are not supported or justified.

Is there a state sales tax exemption for pollution control equipment?

Verso's Maintenance costs are not supported or justified.

Verso's Utilities costs are not supported or justified.

Can Verso use waste caustic from the mill to augment caustic purchases?
(We are seeing this at other mills.)

Verso's annualized costs do not make sense--the numbers do not work out
as presented.

Verso overestimated the interest rate and underestimated equipment life.
According to the OAQPS Control Cost Manual, the correct interest rate is
7% and the correct equipment life is 15 years.

Verso's Power Boilers #1 & #2 lower sulfur fuels analysis is incomplete.
For example, FPL evaluated 1%S residual, 0.5% S residual and 0.3% S
residual fuel oils for its Wyman facility, Verso should at least
evaluate the lower sulfur residual oils.

Waste Fuel Incinerator (WFI): NOx

We adapted the OAQPS Air Pollution Cost Control Manual (Cost Manual)
procedure for evaluating costs for SCR or SNCR for oil-fired EGUs to
oil-fired EGUs (see electronic attachment), but the cost algorithms for
the Direct Capital Costs are from the Cost Manual coal-boiler method and
therefore questionable. So, even if we accept the Verso approach as a
default, it still contains some highly questionable estimates for SCR:

If we assume that the WFI is capable of producing about 48 MW, then the
Total Capital Investment (TCI) per kW is about $165 for SCR, which is in
the middle of the   $50 - $260/kW range for coal-fired EGUs. The
attached Excel workbook estimates a slightly lower TCI.

Verso has estimated an annual reagent cost of $286,000. This exceeds the
$72,000 annual reagent cost that the Cost Manual procedure estimates. 

Verso's Capital Recovery Factor (12.4% interest over a 10-year SCR life)
is inflated. The Cost Manual recommends 7% interest over a 20-year SCR
life.

Verso estimates an annual cost of $2.4 million to control the WFI versus
our estimate of $0.9 million, or Verso's $5,092/ton versus our
$1,986/ton.

MEDEP estimates cost-effectiveness at $4,676/ton versus the $5,092/ton
estimated by Verso; we request an explanation for this difference. 

Because BART is a visibility improvement program, we believe that
cost/deciview ($/dv) is a very important parameter. In this case, for
the four Class I areas evaluated by Verso, SCR would improve visibility
by a total of 1.0 dv. (We would also like to see the visibility
improvements that would occur in the other two Class I areas.) This
results in a cost-effectiveness value of less than $1 million/dv, which
is quite reasonable compared to the average $13 - $20 million/dv that we
are seeing accepted by states and sources that are proposing reductions
under BART. Even if one considers only the visibility improvement at
Acadia National Park, the addition of SCR results in a
cost-effectiveness value of $2.3 million/dv. This leads to the
conclusion that SCR is BART for the Androscoggin WFI.

The same situation applies to SNCR. So, even if we accept the Verso
approach as a default, it still contains some highly questionable
estimates for SNCR:

Although Verso stated that SNCR could achieve 35% control, its cost
analysis is based upon 30% control.

If we assume that the WFI is capable of producing about 48 MW, then the
Total Capital Investment (TCI) per kW is about $65 for SNCR, which is
above the high end of the   $29 - $45/kW range we are seeing in
proposals to install SNCR on coal-fired EGUs. The attached Excel
workbook estimates $31/kW. Verso should provide vendor quotes to support
its higher-than expected estimates.

Verso has estimated a Direct Annual Cost (DAC) of $0.41 million. Since
this exceeds the $0.13 million DAC that the Cost Manual procedure
estimates, the Verso estimate appears to be high. The biggest difference
is in Verso's estimate of almost $0.34 million/year per boiler for
reagent versus the Cost Manual estimate of $0.07 million/yr.

Verso's Capital Recovery Factor (12.4% interest over a 10-year SNCR
life) is inflated. The Cost Manual recommends 7% interest over a 20-year
SNCR life.

Verso estimates an annual cost of $1.1 million to control the WFI versus
our estimate of $0.27 million, or Verso's $7,009/ton versus our
$1,757/ton.

MEDEP estimates cost-effectiveness at $5,944/ton versus the $7,009/ton
estimated by Verso; we request an explanation for this difference. 

In this case, for the four Class I areas evaluated by Verso, SNCR would
improve visibility by a total of 0.2 dv. (We would also like to see the
visibility improvements that would occur in the other two Class I
areas.) This results in a cost-effectiveness value of less than 1.4
million/dv, which is quite a bargain compared to the average $10 - $20
million/dv that we are seeing accepted by states and sources that are
proposing reductions under BART. Even if one considers only the
visibility improvement at Acadia National Park, the addition of SCR
results in a cost-effectiveness value of $2.7 million/dv. This leads to
the conclusion that SNCR could also be a candidate for BART for the
Androscoggin power boilers if SCR is ruled out.

Waste Fuel Incinerator (WFI): SO2

This is what Verso says about SO2 BART for the Androscoggin Waste Fuel
Incinerator:

When No. 6 fuel oil is fired at significant levels, the Mill adds
caustic to the wet scrubber to meet the SO2 emission limit for the WFI. 

SO2 BART ANALYSIS

Identify BART

The WFI has very low SO2 emissions due to the inherent alkalinity (i.e.,
SO2 control) of the primary fuel and the small amount of fuel oil used
in the WFI. In addition during the limited amount of time that No. 6
fuel oil is used to provide a significant amount of the heat for the
WFI, caustic is added to the wet scrubber. Since there are only 50 tons
of SO2 to control annually, the addition of caustic to the wet scrubber
would end up controlling a very small amount of emissions on an annual
basis. Considering visibility, the low, pre-control visibility impacts
from the WFI mean that any visibility reductions associated with
post-control of SO2 emissions would be imperceptible. Based on the
information developed in the Impacts Analysis, the Androscoggin Mill
believes that there is no SO2 BART determination for SO2 from the WFI.

Is Verso saying that it does not want its current procedure of adding
caustic to the wet scrubber when burning fuel oil to be considered BART,
but will keep doing it anyway? If so, that is clearly wrong because BART
would include this practice as a technically- and economically-feasible
option, as proven by Verso. Finally, a control option does not have to
produce a perceptible improvement to be viable.



ATTACHMENT 2

To NPS/FWS Comments - Maine Draft Regional Haze SIP

NPS Comments Regarding FPL Energy Wyman Station BART Evaluation

July 20, 2010

Beginning in 2006, capacity utilization of, and emissions from Units #3
& #4 dropped so much that, assuming that trend continues, it would
likely be cost-prohibitive to make any substantial capital expenditures
to reduce emissions. Furthermore, as noted by Maine Department of
Environmental Protection (MEDEP), NOX emissions are already so low as to
make any significant additional expenses economically infeasible. So, we
shall focus our comments on reducing SO2 emissions by switching to lower
sulfur fuels.

SO2

This appears to be the only BART analysis conducted by MEDEP in which
cost-effectiveness was not evaluated in terms of annual cost/ton of
pollutant removed. Instead, MEDEP appears to have relied solely upon
annual cost/deciviews (dv) of visibility improvement. While we encourage
the use of the $/dv metric, it was not properly calculated nor applied
in this case.

MEDEP also evaluated the BART strategies on the basis of incremental
cost/dv. While that is certainly a valid and useful parameter, it must
be used with caution and its results placed into the proper perspective.
The basic premise underlying the incremental cost analysis is to
identify those strategies that contribute relatively little
environmental benefit in proportion to their cost. Because, in most
cases, the cost of pollution control rises exponentially with control
efficiency, the slope of the cost curve will also increase. For this
reason, rigid use of incremental cost effectiveness will always result
in the choice of the cheapest option if carried to its ultimate extent.
(For example, if this approach were used to evaluate PM controls, it is
likely that all controls more expensive than a multiple cyclone would be
rejected.) According to the NSR Workshop manual, “As a precaution, the
difference in incremental costs among dominant alternatives cannot be
used by itself to argue one dominant alternative is preferred to
another.” Instead, it should be used to compare closely performing
options.

However, FPL did evaluate the costs and benefits of several SO2
reduction options, including the use of lower sulfur fuels. In doing so,
FPL included estimates of the annual costs and emission reductions for
each option, as well as the cost/ton for each of those options; those
results are contained in Tables 5-3 thru 5-5 of the FPL BART analysis.
We used the data from FPL’s Table 5-3 to generate the cost-benefit
data contained in the electronic attachment, and our results are
summarized below.

Wyman #3 (2007  - 2008)





Fuel Sulfur (%)	1	0.7	0.5	0.3

Increased Annual Fuel Cost 	 $     175,306 	 $     206,243 	 $    
835,283 	 $     1,722,127 

SO2 Emission Reductions (tpy)	270	351	405	459

SO2 Reductions Cost-Effectiveness ($/ton)	 $           650 	 $          
588 	 $        2,064 	 $           3,755 

Greatest Visibility Improvement (dv)	0.99	1.43	1.78	2.15

Cost-Effectiveness ($/dv)	 $     177,077 	 $     144,226 	 $     469,260
	 $        800,989 

Cumulative Visibility Improvement (dv)	2.61	4.26	5.26	6.28

Cumulative Cost-Effectiveness ($/dv)	 $      67,167 	 $      48,414 	 $ 
   158,799 	 $        274,224 













Wyman #4 (2007  - 2008)





Fuel Sulfur (%)	 	 	0.5	0.3

Increased Annual Fuel Cost 	 	 	 $  2,910,880 	 $     7,014,743 

SO2 Emission Reductions (tpy)	 	 	250	499

SO2 Reductions Cost-Effectiveness ($/ton)	 	 	 $      11,656 	 $      
  14,045 

Greatest Visibility Improvement (dv)	 	 	0.41	0.84

Cost-Effectiveness ($/dv)	 	 	 $  7,099,707 	 $     8,350,885 

Cumulative Visibility Improvement (dv)	 	 	1.60	3.38

Cumulative Cost-Effectiveness ($/dv)	 	 	 $  1,819,300 	 $    
2,075,368 



Our results differ from those presented by MEDEP because we used the
most-recent (2007 – 2008) average fuel use data provided by FPL
instead of the maximum two-year average. We did this because the
most-recent two years are much more representative of anticipated
reduced operation of these units. While use of the reduced-capacity
operation data did not affect the $/ton estimate (which MEDEP did not
include), it has a great effect on the $/dv estimate because of the
reduced annual costs.

Our results indicate that, on a $/ton basis, use of 0.7% sulfur oil is
the most cost-effective. However, BART is not necessarily the most
cost-effective solution. Instead, the $2,000/ton cost of switching Unit
#3 to 0.5 % sulfur oil would be considered reasonable by most states.

As noted above, MEDEP appears to have relied solely upon $/dv of
visibility improvement. However, the baseline for estimating the
increased costs of lower sulfur fuels (2% sulfur) is different from the
baseline for existing visibility impacts (1.6% S). Therefore, the
visibility benefits are underestimated because the baseline impacts are
underestimated. MEDEP has also presented 98th percentile visibility
values despite using only one year of meteorological data—that is
misleading because, when only one year is modeled, only the maximum
values are to be used.

Because BART is a visibility improvement program, we believe that
cost/deciview ($/dv) is a very important parameter. In this case, for
the six Class I areas evaluated by FPL, lower-sulfur (0.5% - 0.3% S)
fuels would improve visibility by a total of 6.9 – 9.7 dv. This
results in a cost-effectiveness value of $0.2 – 2.1 million/dv, which
is relatively inexpensive compared to the average $13 - $20 million/dv
that we are seeing accepted by states and sources that are proposing
reductions under BART. Even if one considers only the visibility
improvement at Acadia National Park, the lower-sulfur fuels result in
cost-effectiveness values of $0.5 – 8.4 million/dv. This leads to the
conclusion that 0.5% - 0.3% sulfur fuels are BART for the FPL boilers.



ATTACHMENT 3

To NPS/FWS Comments - Maine Draft Regional Haze SIP

NPS Comments Regarding SAPPI SD Warren Paper Mill BART Evaluation

July 20, 2010

Page 63 of the 2/06/09 draft of Maine Department of Environmental
Protection (MEDEP)  RH SIP contains Table 8-2 titled "Modeled
Impacts...of Maine BART-Eligible Sources..." That table shows a 0.75 dv
impact at Acadia and 0.78 dv at Moosehorn from Power Boiler #1 at the
SAPPI SD Warren Paper mill. 

The September 2009 company BART report did not evaluate Power Boiler #1.

The MEDEP BART analysis (1/21/10) listed Power Boiler #1 as a BART
source and included a BART determination for it. 

The MEDEP BART analysis (posted 6/29/10) did not mention Power Boiler
#1. 

Why was Power Boiler #1 omitted from the BART determination?

 Verso assumed a $0.08/kWh cost for electricity.

 Verso assumed a $0.08/kWh cost for electricity.

 Verso assumed a $0.08/kWh cost for electricity.

 In Massachusetts, sources evaluated 1%S residual, 0.5% S residual, 0.3%
S residual, 0.3% S distillate, 0.05% S distillate, and 0.0015% S
distillate.

 Verso assumed a $0.07/kWh cost for electricity.

 Verso assumed a $0.07/kWh cost for electricity.

BART Guidelines: “You should consider the incremental cost
effectiveness in combination with the average cost effectiveness when
considering whether to eliminate a control option” and “You should
exercise caution not to misuse these [average and incremental cost
effectiveness] techniques…[but consider them in situations where an
option shows]…slightly greater emission reductions…”

