By Scot Pritchard, Cormetech
Selective catalytic reduction (SCR) for NOx removal from utility boilers has been shown to have extremely high removal efficiencies. In recent years, there has been increased focus on a potentially significant pollution control co-benefit of the SCR process, specifically the capability to oxidize vapor phase mercury in coal-fired power plants.
A test program, performed in cooperation with Dominion Power and The Babcock and Wilcox Co. (B&W), was executed at Dominion Power’s Mount Storm power plant in Grant County, W. Va. The program was focused on both the SCR catalyst capability to oxidize mercury as well as the scrubber’s capability to capture and retain the oxidized mercury. This article focuses on the SCR catalyst performance aspects. Further information is available from the presentation provided at the 2005 Institute of Clean Air Companies Forum.
The Mount Storm site consists of three units totaling approximately 1,660 MW. All units are equipped with SCR systems for NOx control. A full-scale test to evaluate the effect of the SCR was performed on Unit 2, a 550 MW T-fired boiler firing a medium sulfur bituminous coal. This test program demonstrated that the presence of an SCR catalyst can significantly affect the mercury speciation profile. Observation showed that in the absence of an SCR catalyst, the extent of oxidation of elemental mercury (Hg0) at the inlet of the flue gas desulfurization (FGD) system was about 64 percent. The presence of a Cormetech SCR catalyst improved this oxidation to levels greater than 95 percent, almost all of which was captured by the downstream wet FGD system. Cormetech’s proprietary SCR Hg oxidation model was used to accurately predict the field results.
Field Test Results
The SCR system at the Mount Storm unit consists of two parallel reactors containing V2O5•WO3•TiO2 SCR catalyst. The catalytic reactors are designed to treat 5.82 million lb/hr of flue gas at 700 F using two installed layers of SCR catalyst and achieves 90 percent-plus NOx reduction. After passing through the SCR, the flue gas passes through an air heater and then enters an electrostatic precipitator to remove particulate matter. The division of flue gas flow between two ducts starting from the economizer outlet and continuing until the electrostatic precipitators (ESP) outlet is shown in Figure 1, which is a schematic of the Unit 2 duct configurations, flue gas flow direction and sampling points. At the ESP outlet, the flue gas trains are recombined in one common duct and are directed to a B&W wet flue gas desulfurization (FGD) system. The wet FGD system is a limestone forced-oxidation system using dibasic acid for improved lime solubility and SO2 capture.
Effect of SCR on Hg Speciation
The effect of SCR on Hg speciation was determined by comparing the results obtained during two Hg sampling programs. During the first program, the SCR was bypassed and during the second program the SCR was on-line. For each program, Hg speciation profiles at the economizer outlet, FGD inlet and FGD outlet were measured by the Ontario Hydro Method (OHM) and a number of on-line mercury speciation analyzers. In the absence of the SCR catalyst, the natural oxidation of Hg0 from the economizer outlet (upstream of the SCR) to the inlet of the FGD unit averaged about 64 percent. The presence of the SCR catalyst increased this oxidation to approximately 98 percent. Figure 2 displays the effect of the SCR catalyst on Hg speciation and control across the air pollution control devices. Observation showed that the presence of the SCR catalyst significantly affected the Hg speciation profile at the inlet of the FGD system.
Previous studies have shown that the presence of HCl in the flue gas is an essential component for the heterogeneous, catalytic oxidation of Hg0 by the SCR system. The concentration of HCl or other halogens should be high enough to thermodynamically allow high levels of heterogeneous oxidation of Hg0 by the SCR catalyst at the SCR temperature and expected flue gas composition. The medium sulfur bituminous coal used in the Mount Storm test program had sufficient chlorine content to produce significant quantities of HCl in the flue gas. HCl concentration at the inlet of the FGD was 34±1 ppmvdc as measured by the EPA Method 26A.
SCR Mercury Oxidation Modeling
Cormetech’s proprietary reactor model, incorporating numerical integration of the governing differential equations for reaction kinetics, predicted concentration changes through the SCR catalyst. The effects of HCl on Hg0 oxidation by the SCR catalyst were also assessed by the model. The model results predicted Hg oxidation and NOx reduction with allowance for ammonia inhibition of SCR Hg oxidation, activity decline with catalyst age, thermodynamic limits on the extent of Hg oxidation (for example, due to low coal chloride content), as well as other key operating conditions. Regression constants for the model were based on extensive parametric studies on a pilot scale using both fresh and field-aged catalyst samples including variation of catalyst properties (for example, metals composition that is optimized for each plant site during the SCR design process). A quantitative statistical comparison was made between predicted SCR Hg oxidation performance and the observed field data at Mount Storm Unit 2. Key field data uncertainties relevant to SCR Hg oxidation performance prediction were identified through model sensitivity analysis.
For the SCR by-pass configuration, the observed non-SCR conversion of elemental Hg to oxidized Hg was around 68 percent from the economizer outlet to the FGD inlet (ESP outlet) based on OHM measurements at both locations. Similarly, for the SCR by-pass configuration with the use of an FGD sulfur donating compound additive, the observed non-SCR Hg conversion was around 59 percent. The average non-SCR Hg conversion of these two data sets is around 64 percent with a pooled standard uncertainty of around 8 percent absolute. This average value of 64 percent for the non-SCR conversion of elemental Hg is assumed to be valid for other portions of the field test program when the SCR is on-line. This assumption is based on field observations that the testing was done on full load with no change in coal composition and total Hg and no significant amount of Hg removed with the ESP fly ash.
Thus, for modeling of the case with the SCR on-line (not bypassed), it is assumed that an equal percentage as above (that is, around 64 percent) of the amount of elemental Hg leaving the SCR is oxidized by non-SCR mechanisms as the flue gas travels through the air preheater (APH)and ESP. The net observed fractional conversion of elemental Hg from SCR inlet to FGD inlet (denoted ηnet), may be related to the SCR and non-SCR fractional conversions of elemental Hg as follows:
Equation 1 is valid when the percent elemental Hg conversion downstream of the SCR remains constant as the configuration is changed from the SCR by-pass to SCR on-line.
The SCR portion of the Hg conversion may be calculated from rearrangement of Equation 1 using the OHM data from the second field test program when the SCR was operating normally with NH3 injection. The OHM data showed a net conversion of ηnet = 98.3 percent with a standard uncertainty of 0.5 percent (absolute).
The non-SCR contribution assumed is ηnon-SCR approximately 64 percent with a standard uncertainty of around 8 percent (absolute). Using these values, Equation 1 may be solved for the SCR contribution to Hg oxidation, yielding a value of ηSCR approximately 95 percent with a standard uncertainty of around 2 percent (absolute). This roughly 95 percent “observed” SCR Hg conversion and its uncertainty is compared to the model predictions in Table 1.
The model estimates for SCR conversion of elemental Hg were around 93 percent with a standard uncertainty of roughly 3 percent compared to the “observed” SCR Hg conversion of approximately 95 percent with a standard uncertainty of roughly 2 percent. The model prediction and “observed” SCR Hg conversion are not statistically different at an 80 percent confidence level based on a standard t-test. Thus, Cormetech’s Hg model prediction for SCR Hg oxidation in Mt. Storm Unit 2 is consistent with the field data.
The combination of an SCR system and the wet FGD constitutes a very good Hg control technology and should be considered when evaluating Hg control compliance plans.
- Predictions from Cormetech’s proprietary SCR Hg oxidation model accurately match the observed plant data thus allowing for guaranteed performance to the end user.
- Proves that model can be a valuable tool for SCR owners evaluating needs for sorbents for mercury control
- Hg oxidation performance capability of SCR catalyst is a function of many parameters including catalyst formulation, inlet boundary conditions, and so on and therefore results can vary widely. Utilities assessing Hg compliance should confer with qualified companies to assist in achieving guaranteed performance and assurance of meeting future compliance needs.
- In addition to the traditional Catalyst Management strategies associated with NOx control, Hg oxidation may need to be considered.
- Cormetech catalyst development work is ongoing with focus on optimizing Hg oxidation while enhancing NOx reduction capability and minimizing SO2 conversion.
Author: Scot Pritchard is vice president pf sales and marketing at Cormetech, Inc., a supplier of SCR catalysts and catalyst management services.