SO2 Emission Factors for Fluidized Bed Combustion coal generation

[grgmiller]

The SO2 emission factor used by eGRID for SUB coal is 35*S, while the emission factor for BIT coal is 38*S for all boiler firing types (see Table C-3 of the eGRID2020 Technical Guide).

However, in Appendix A-1 of the EIA's 2020 electric power annual, they use a factor that is 10% of each of these for fluidized bed boiler types (3.5S for SUB, and 3.8S for BIT). These are the numbers that we have been using in OGE so far. The EIA notes that these emission factors come from the EPA's AP-42 document, specifically Table 1.1-3.

However, when we consult AP-42, there are actually no SO2 emissions factors for fluidized bed boilers listed in the table; instead there is a footnote which states:

SO 2 emission factors for fluidized bed combustion are a function of fuel sulfur content and calcium-to-sulfur ratio. For both bubbling bed and circulating bed design, use: lb SO 2/ton coal = 39.6(S)(Ca/S) -1.9. In this equation, S is the weight percent sulfur in the fuel and Ca/S is the molar calcium-to-sulfur ratio in the bed. This equation may be used when the Ca/S is between 1.5 and 7. When no calcium-based sorbents are used and the bed material is inert with respect to sulfur capture, the emission factor for underfeed stokers should be used to estimate the SO2 emissions. In this case, the emission factor ratings are E for both bubbling and circulating units.

If we assume that the bed is inert and we were to use the underfeed stoker factor, the emission factor would be 31*S.

It does not appear that EIA-860 includes information about the molar calcium-to-sulfur ratio of the fluidized bed at any specific plant, but if we calculate the factor using the valid range for the equation given in the footnote (1.5%-7%), we find that the factor could range anywhere from 0.98*S to 18.33*S.

The 3.8*S value used by the EIA falls within this range, but how did they choose that number? I'm not entirely sure, but I have a guess: the FGD table (Schedule 6F) of EIA-860, which includes information about the sorbents used in each plant, lists the "Removal efficiency of sulfur" for each sorbent type in each plant. If we examine all calcium-containing sorbents (LI and LS, which are some form of lime or limestone), the average removal efficiency is about 90%, and 38 - (38*90%) = 3.8.

This suggests that eGRID is systematically over-estimating SO2 emissions from fluidized bed coal plants, and that the EIA factor, while likely closer to the actual number, could still be improved by using plant-specific data. For now, we will continue using the EIA value of 3.8.

Proposed improvements

We could use the FGD table from EIA-860 to potentially identify boiler-specific SO2 emissions factors based on information about the sorbent used. Here are some thoughts:

• If the identified sorbent type is not one of the Calcium-containing sorbents (LI or LS), we would use the underfeed stoker factor of 31*S (assuming that without a calcium-based sorbent, the bed would otherwise be inert with respect to sulfur capture). If it uses LS or LI as a sorbent, we discount the 38*S factor by the sulfur removal efficiency column.
• however, even for non LS and LI sorbents, there is often a removal efficiency listed - should we always use this removal efficiency and only use the underfeed stoker rate if no removal efficiency data is listed?
• When there is sulfur removal efficiency data listed in EIA-860 Schedule 6F, and SO2 removal efficiency data listed in EIA-923 schedule 8C, how should this information be used together? For example, plant 50392 lists a sulfur removal efficiency of 99% in EIA-860, but an SO2 removal efficiency of 45-53% in EIA-923. Does this mean that the fluidized bed removes 99% of the sulfur from the fuel, and of the SO2 generated from that, a further ~50% is removed? Or should these be interpreted differently?