KDHE  -  Bureau of Air Response to the National Park Conservation Association Comments on the Kansas Regional Haze SIP
                                December 1, 2011
The NPCA comments can essentially be summarized as stating that KDHE did not perform an acceptable five factor analysis for the Kansas BART sources, thus resulting in emission rates in the agreements that were not low enough to be considered BART. In some cases, the comments stated that additional control equipment should be required to be installed. In other cases, the comments suggest that the control equipment to be installed should be capable of achieving a lower emission rate.  It is KDHE's position that in all cases an acceptable five factor analysis was performed for the BART affected units.  KDHE believes these statutory factors have been met for these BART sources in accordance with the provisions of 40 CFR 51, Regional Haze Regulations and Guidelines for BART. 
KDHE does not agree with these comments; however, to address them, dispersion modeling using the CALPUFF model was conducted. The modeling used the same inputs that were used when the original Regional Haze SIP modeling was conducted as part of the BART technical analysis, except for changes to the emission rates to address the NPCA comments.  Each of the original modeling runs was repeated with lower rates in order to evaluate the incremental visibility benefits that would result from lower SO2 and NOX rates than those currently in the Regional Haze agreements with KCP&L and Westar Energy. The rates that were selected were based on NPCA comments and on recent BACT determinations. These rates are more stringent than required by the BART process. BACT and BART are not equivalent. BACT is applied to new sources while BART is applied to existing sources. The emission rates that can be achieved at a new source will be more stringent than those that can be achieved by retrofitting an existing source.  The modeled emission rates, the model results, and the cost of supplemental controls are included in an attached table.
Upon completion of the current modeling, the differences in the modeled visibility impacts on the five Class I areas that were evaluated in the SIP were determined. KDHE also obtained or developed annualized costs for the additional equipment that would be required to be installed in order to achieve the lower emission rates. For NOX, the modeling and associated costs were determined for SNCR and SCR at Jeffrey Energy Center (JEC) Units 1 and 2 as well as La Cygne Unit 2. For SO2 , the modeling and associated costs were determined for installation of new scrubbers at JEC Units 1 and 2, and for operation of new scrubbers at a lower rate at La Cygne. The metric suggested by NPCA in their comments, cost per deciview of visibility improvement, was used to evaluate the results of the modeling. The results are included in the attached table.   
Jeffrey Energy Center Units 1 and 2 - NOX
The BART analysis for JEC Units 1 and 2 identified various potential controls that were feasible for retrofit on these units.  The control technologies were ranked and it was determined based on this ranking that low NOX burner systems and SCR systems were the most effective at NOX reductions.  Because these two technologies were ranked highest for removal efficiencies, their cost effectiveness was evaluated first.  During the cost evaluation it was determined that LNB costs were reasonable and this was ultimately the technology chosen as BART.  SCR was further evaluated for incremental cost improvements and this technology was excluded as BART based on the high incremental cost and low incremental visibility improvements.  The cost analysis for SCR was performed based on the BART emission limit of 0.10 lb/MMBtu, which is within the range of effectiveness one could expect as a retrofit control on these older units.  The analysis assumed a removal rate of approximately 80%.  KDHE believes this is reasonable for a retrofit control.  
The NPCA comments that the costs used in the analysis are inflated and that the cost used in the JEC analysis should reflect the best possible reductions that SCR can achieve.  The cost is reflective of the assumed emission rate limit of 0.10 lb/MMBtu.  The NPCA comments suggest that the cost used in this analysis should reflect an emission rate of 0.05 lb/MMBtu.  The 0.05 lb/MMBtu rate is consistent with current BACT rates for new facilities, not retrofitting 35 year old existing facilities. The incremental cost associated with an SCR at tangential-fired units such as JEC Units 1 and 2 is significantly higher than for a wall-fired boiler, as reflected in the significantly different EPA presumptive rates of 0.15 lb/MMBtu versus 0.23 lb/MMBtu. In response to the NPCA comments, KDHE obtained an annualized cost for SNCR control on Jeffrey Units 1 and 2 from Westar Energy. The annualized cost for SCR control on Jeffrey Units 1 and 2 was obtained from the original Westar BART analysis, Tables 5.4 and 5.5. 
KDHE does not agree with the NPCA comments; however, to evaluate the incremental benefit of an SCR installed at JEC Units 1 and 2, a CALPUFF run was performed with a rate of 0.08 lb/MMBtu. The base SO2 rate used when conducting the SNCR and the SCR modeling runs for NOX at Jeffrey Unit 1 was kept at the same rate as the original BART modeling rate (0.86 lb/MMBtu). The base SO2 rate used when conducting the SNCR and the SCR modeling runs for NOX at Jeffrey Unit 2 was kept at the same rate as the original BART modeling rate (0.88 lb/MMBtu). The change in modeled impact between the SCR scenario (0.08 lbs NOx /MMBtu) and the LNB scenario (0.15 lbs NOx /MMBtu) at Hercules Glade was less than 0.05 deciview. The difference in the cumulative impact of all five Class I Areas for this scenario is 0.16 deciview. The differences in visibility impact between an SCR and LNB for Jeffrey Unit 1 are insignificant when compared to the differences in impact between the LNB scenario and the pre-BART scenario. The differences for Jeffrey Unit 2 under these scenarios are even lower than for Unit 1.
For SNCR, no further evaluation was performed in the original analysis due to the low overall ranking of estimated control efficiency.  To address this NPCA comment and evaluate the incremental benefit that would result from requiring that an SNCR be installed at JEC Units 1 and 2, a CALPUFF run was performed with a NOX rate of 0.10 lb/MMBtu. For Jeffrey Unit 1, the change in modeled impact between the SNCR scenario (0.10 lb NOx /MMBtu) and the LNB scenario (0.15 lb NOx /MMBtu) at Hercules Glade, was 0.03 deciview. The difference in the cumulative impact of all five Class I areas for this scenario is only 0.09 deciview. The differences in visibility impact between an SNCR and LNB for Jeffrey Unit 1 are even lower than those for the SCR scenario, as expected. The results for Unit 2 are nearly identical. It is KDHE's conclusion that the additional NOX reduction costs for the incremental visibility improvements for Jeffrey Units 1 and 2 are too high and do not warrant changes to the Kansas Regional Haze Westar Agreement or the SIP.
Jeffrey Energy Center Units 1 and 2  -  SO2
The BART analysis for JEC Units 1 and 2 for SO2 resulted in the decision that rebuilding existing wet scrubber units and meeting a rate of 0.15 lb/MMBtu was BART.  The NPCA comments suggest that retrofit technology should be required to meet BACT level emission rates. The 0.05 lb/MMBtu rate suggested by NPCA is consistent with current BACT rates for new facilities, not retrofitting 35 year old existing facilities.  This is not how the BART rates are established and the NPCA has offered no justification for lower rates specific to these units. Although KDHE does not agree with these NPCA comments; to evaluate the incremental cost and visibility improvement of requiring a lower SO2 emission rate, KDHE conducted a CALPUFF run for the installation of a new scrubber for each of the two JEC units instead of the rebuilt scrubbers determined by KDHE to be BART.
The CALPUFF runs were performed with a rate of 0.05 lb/MMBtu for a new scrubber. The base NOX rate used when conducting the new scrubber modeling run for SO2 at Jeffrey Unit 1 was kept at the same rate as the original BART modeling rate (0.49 lb/MMBtu). The base NOX rate used when conducting the new scrubber modeling run for SO2 at Jeffrey Unit 2 was kept at the same rate as the original BART modeling rate (0.48 lb/MMBtu). In response to the NPCA comments, KDHE obtained annualized costs for new scrubbers on Jeffrey Units 1 and 2 from Westar Energy. 
The change in modeled impact between the new wet scrubber scenario (0.05 lb SO2/MMBtu) and the BART Agreement rebuilt scrubber (0.15 lb SO2/MMBtu) was less than 0.05 deciviews for Jeffrey Unit 1 at Wichita Mountains. The difference in the cumulative impact of all five Class I Areas for this scenario is 0.17 deciview. The differences in visibility impact between the new scrubber and the BART Agreement scrubber for Jeffrey Unit 1 are insignificant when compared to the differences in impact between the BART Agreement scrubber and the pre-BART scenario. The differences for Jeffrey Unit 2 under these scenarios are even lower than for Unit 1.  It is KDHE's conclusion that the additional SO2 reduction costs for the incremental visibility improvements for Jeffrey Units 1 and 2 are too high and do not warrant changes to the Kansas Regional Haze Westar Agreement or the SIP.
La Cygne Units 1 and 2  -  NOX
The BART analysis for La Cygne Units 1 and 2 for NOX resulted in the decision that establishing a combined emissions limit for both units with a rate of 0.13 lb/MMBtu was BART.  The SCR at Unit 1 was constructed while the Regional Haze SIP development process was underway and was an existing unit when the SIP was adopted and submitted to EPA.  Unit 1 is a cyclone design, which has a much higher initial NOX emission rate. This is reflected in the fact that the BART presumptive limits were not based on boiler modifications such as low NOX boiler systems, but rather an SCR. It is not technically feasible for an older cyclone boiler to be retrofit to achieve NOX rates that are consistent with current BACT rates for wall fired boilers as NPCA suggests. KDHE reviewed the CAMD and EIA databases at the time the Regional Haze SIP was being developed to review emissions data for cyclone boilers equipped with SCR technology. A relatively small number of cyclone boilers were so equipped at that time and their emission rates varied both above and below the presumptive NOX rate.
NPCA commented that a lower NOX emission rate for Unit 2 should be required and could be achieved at the same cost provided in the analysis.  The costs will be directly associated with the designed emissions rate of the SCR and the lower the rate the higher the cost.   NPCA commented that KDHE afforded KCP&L flexibility in the technology used to reach the rates agreed to in the Regional Haze Agreement.  In the La Cygne BART analysis, pages 5-10, KCP&L proposes SCR on Unit 2 along with further study on combustion controls.  The combined rate for the two units cannot be achieved without the installation of an SCR on Unit 2.  KCP&L has applied to KDHE for a permit to install an SCR on Unit 2. For SNCR, no further evaluation was performed due to the low overall ranking of estimated control efficiency and because of the low combined NOX emission rate in the KCP&L agreement.  
Although KDHE does not agree with these NPCA comments, CALPUFF modeling was conducted to evaluate the incremental visibility improvement that would result from requiring a lower emission rate in the KCP&L agreement. The change in modeled impact between the BART agreement NOx controls (0.23 lb NOx /MMBtu) and the Unit 2 SCR scenario (0.08 lb NOx /MMBtu) was 0.08 deciview for Upper Buffalo. The difference in the cumulative impact of all five Class I Areas for this scenario is 0.25 deciview. These differences in impact are very low, as expected. 
The change in modeled impact between the BART agreement NOx controls (0.23 lb NOx/MMBtu) and the Unit 2 SNCR scenario (0.14 lb NOx /MMBtu) was 0.03 deciview for Upper Buffalo. The difference in the cumulative impact of all five Class I Areas for this scenario is 0.12 deciview. These differences in impact are very low, as expected. 
For KCPL, the change in modeled impact between the BART agreement controls and the Unit 2 SCR (0.08 lb NOx /MMBtu) Unit 1 and 2 new wet scrubbers (0.05 lb SO2/MMBtu) scenario was 0.17 deciview for Upper Buffalo. The difference in the cumulative impact of all five Class I Areas for this scenario is 0.65 deciview.  The cumulative differences in visibility impact for this scenario are larger than the JEC scenarios because the modeling was for two units combined. 
The annualized cost for SNCR control on KCP&L La Cygne Unit 2 was determined by using SNCR costs obtained from Jeffrey Energy Center Unit 1, and scaling the dollar amount using heat input and NOx rates. The annualized cost for SCR on KCP&L La Cygne Unit 2 was obtained from Table 5.5 of the BART analysis. These costs were multiplied by 20% to estimate the incremental costs associated with achieving a substantially lower emission rate for the analysis. It is KDHE's conclusion that the additional NOX reduction costs for the incremental visibility improvements for La Cygne Units 1 and 2 are too high and do not warrant changes to the Kansas Regional Haze KCP&L Agreement or the SIP.
La Cygne Units 1 and 2  -  SO2
The BART analysis for La Cygne Units 1 and 2 for SO2 resulted in the decision that establishing a combined emissions limit for both units with a rate of 0.10 lb/MMBtu was BART. The NPCA comments suggest that retrofit technology should be required to meet BACT level emission rates. The rate suggested by NPCA is consistent with current BACT rates for new facilities, not retrofitting 35 year old existing facilities. Unit 1 has an existing scrubber that will be modified to separate the PM control from the SO2 control with increased SO2 removal efficiency as the goal.  Unit 2, which did not have an existing scrubber, will be fitted with a new scrubber.   The combined BART emission rate chosen for SO2 controls are within the range of expected removal efficiencies, considering one unit is a retrofitted scrubber.  
Although KDHE does not agree with these NPCA comments, CALPUFF modeling was conducted to evaluate the incremental visibility improvement that would result from requiring a lower SO2 emission rate in the KCP&L agreement. For Unit 1 scrubber at KCP&L, the change in modeled impact from the BART Agreement scenario (0.15 lb SO2/MMBtu) to a lower rate consistent with current BACT limits (0.05 lb SO2/MMBtu) was 0.01 deciview at Upper Buffalo. The difference in the cumulative impact of all five Class I Areas for this scenario is 0.12 deciview.  These differences in impact are very low as expected. The Unit 2 scenario impact comparison is even lower. The incremental costs used to evaluate the cost effectiveness of a lower SO2 emission rate at La Cygne Units 1 and 2 were based on EPA AirControlNET costs. These costs were multiplied by 20% to estimate the incremental costs associated with achieving a substantially lower emission rate for the analysis. It is KDHE's conclusion that the additional SO2 reduction costs for the incremental visibility improvements for La Cygne Units 1 and 2 are too high and do not warrant changes to the Kansas Regional Haze KCP&L Agreement or the SIP.
The CALPUFF modeling conducted in response to NPCA's comments in all cases shows that the modeled visibility improvements for the five Class I areas are extremely small relative to the visibility improvements achieved with the BART rates established for the four units under question. In addition, it is important to note that other electrical generating units in Kansas were also included in the Kansas Regional Haze SIP and significant reductions were achieved at these as well. The majority of the issues raised by NPCA relate to a state's ability to justify BART thresholds and limits while following the five statutory factors.  KDHE acted within its authority in making the BART determinations in question and the NPCA has not provided any direct evidence that the BART determinations are not within the flexibility provided by the SIP development process.  The attached table with the annualized cost factors for achieving emission rates equivalent to BACT rates shows the incremental costs to be extremely high when gauged by either of the two metrics used for the analysis. KDHE affirms that the BART determinations for the units in question follow the statutory factors under the Clean Air Act and the regulatory requirements of the Clean Air Visibility Rule. They will result in significant, cost-effective visibility improvements to Class I areas affected by BART sources in Kansas.
