Fluid Catalytic Cracking Units (FCCU) October 1995 TNRCC Rule 116.111(3) in Regulation VI requires that Best Available Control Technology (BACT) be applied to all facilities that must obtain a permit. BACT determinations are made on a case-by-case basis. Current BACT guidelines for FCCU is: CEMS Control Technology PM NSPS no electrostatic precipitator (EPS) scrubber Opacity 15-20% yes same as PM CO 500 ppmv yes hot regenerator CO boiler CO combustion promoters SO2 200-300 ppmv yes scrubber (or alternative method of flue gas desulfurization) feedstock hydrodesulfurization and/or blending SOx reduction catalysts NOx 200 ppmv yes minimize use of CO combustion promoters VOC <10 ppmv no complete combustion Comments on BACT: (a) The PM emission rate is obtained from the maximum coke burn-off rate. PM emission rates have also been calculated using the AP-42 emission factor of 19 lbs per 1,000 bbl of fresh feed. This is acceptable but does not obviate the need to demonstrate compliance with the NSPS PM standard if NSPS is applicable. When the NSPS PM standard was promulgated by EPA, their test method measured the filterable portion only; that is, the impinger catch (or condensible PM) was not included. TNRCC practice has always added both to determine PM; that is, PM always means total suspended particulate (TSP) in the state program. Because of the nature of the NSPS PM standard, some state permits have included two limits, one for filterable PM based on NSPS, and one for TSP based on Reg I. It should be noted that Reg I is reasonably available control technology (RACT) and not BACT. Accordingly, Reg I does not constitute a BACT review for TSP so that the applicant still has the burden of demonstrating BACT for the project. The TSP issue is more likely to be raised by the applicant if an ESP is the final PM control device. (b) PM control usually consists of two or three stage cyclones, followed by either an ESP or a scrubber designed for both PM and SO2 removal. TNRCC staff believe that scrubbers may have a higher degree of reliability. (c) Occasionally an operator will propose ammonia injection as a means of improving ESP opacity control. Ammonia injection for opacity control is not necessary in a properly designed and operated system and is not considered BACT. (d) When a grandfathered FCCU is brought into the permit process, that act does not necessarily trigger a BACT review with regard to all pollutants. For example, if the operator proposes to use CO combustion promoters in a grandfathered FCCU, SO2 would not normally be an affected pollutant. In this example, the BACT review would not be extended to SO2, instead the grandfathered SO2 emission rate would be determined and incorporated in the permit. Grandfathered emission rates are based on actual operating data, not on nameplate or design capacity. (e) A CEMS is required with regard to opacity, CO, SO2, and NOx, but not with regard to VOC. For CO, concentration is measured on a ppmvd basis averaged over a one-hour period. For SO2, concentration is measured on a ppmvd, air-free basis, averaged over a one-hour period. CO and SO2 follow NSPS Subpart J practice. There is no set standard practice with regard to NOx. Each CEMS must comply with all applicable state and federal requirements. An oxygen CEMS will also be needed. (f) The most direct means of complying with the CO limit is to operate the regenerator at a higher temperature. (Hot regenerator means an operating temperature of about 1,350øF.) Sometimes the metallurgy of an existing unit does not allow this. However, many of the newer units can operate hot. The operator usually selects one or more of these options: hot regenerator, CO combustion promoters, CO boiler(s). The CO boiler has the advantage of contributing process steam. (g) Regenerator SO2 (and NOx) emissions will reflect to some degree the composition of the crude. Approximately 10 to 20% of the total sulfur in the FCCU feedstocks remains in the coke on the catalyst. (In order to be conservative, use 20%.) (h) The design and operation of an add-on scrubber must assure at least 90% SO2 removal, a NSPS requirement. NSPS contains provisions governing continuous demonstration of compliance with Subpart J, SO2 limits. The add-on "scrubber" is actually an absorber. A caustic solution is utilized. They are often referred to as "sodium scrubbers." The amount of sodium hydroxide (SOx) consumed by these scrubbers is approximately 1.15 times the amount of SOx in the inlet gas. The pH of the circulating scrubbing liquid should be maintained above 6.0 at all times. Venturi or tray tower scrubbers may be used. A tray tower is less attractive if simultaneous PM removal is desired. Jet-ejector and high energy venturi scrubbers have been used successfully in Texas. Downstream of a CO boiler, a jet- ejector venturi scrubber would normally be used. With a hot regenerator only, a high energy venturi scrubber would normally be used. All currently existing systems in Texas are venturi scrubbers. The addition of water is usually controlled by a water level monitor on the separation tank. The addition of caustic is usually controlled by a pH monitor. Each permit will require continuous monitoring of liquid pH to provide a continuous demonstration of compliance. A number of other approaches for monitoring scrubber operations have been suggested to enhance continuous demonstration of compliance. These include pressure drop through the venturi, scrubber liquid/flue gas ratio, and throat velocity ratio. One of these may also be incorporated in the permit. On-site wastewater facilities will be needed. At inland locations where wastewater disposal may be a problem, alternative scrubbing systems such as dual alkali or citrate scrubbing systems may be considered, or a different control technology such as SOx reduction catalysts can be selected. A citrate based system has been used in a cat cracker application in Texas (it was not considered a success). The agency is very supportive of the use of scrubbers in FCCU regenerator vent gas control because opacity and PM are also very effectively controlled. Opacity is reduced to less than 15%. As of October 1, 1994, eight scrubbers were in cat cracker catalyst regenerator flue gas vent service. The maximum BACT limit for SO2 emissions of 300 ppmv allows the operator to select control technologies other than scrubbers. It should be noted that an add-on scrubber can be designed and operated to limit SO2 emissions to 200 ppmv or less. (i) SOx reduction catalysts reduce FCCU SOx emissions by about 70%, most of which is transferred to the refinery sulfur plant as an increased sulfur load. Therefore, if SOx reduction catalysts are to be used, the sulfur plant must be carefully evaluated to assure that adequate capacity is available; and any related amendment to the sulfur plant permit must be secured. (j) Regenerator NOx forms by two mechanisms: thermal NOx as a result of the combustion process, and chemical NOx from nitrogen present in the coke. Regenerator temperatures are not high enough to create large amounts of thermal NOx. Even in hot regenerator operation, NOx emissions are generally less than 200 ppmv. Chemical NOx is the dominant mechanism. NOx emissions increase when CO combustion promoters or CO boilers are used. Burner design can help control thermal NOx emissions from CO boilers. NOx emissions do not seem to be significantly affected by hot regenerators. If add-on scrubbers reduce NOx emissions, the effects have not been quantified. There is no clear consensus on the effects of SOx reduction catalysts on NOx emissions (the effects are probably negligible). NOx may be adjusted to a value significantly higher than 200 ppmv on a case-by-case basis if the operator meets all of the other state and federal requirements and presents appropriate arguments together with supporting information. There is no FCCU flue gas NOx control technology per se. (k) Any process (CO boilers, CO incinerators, regenerator combustion promoters, hot regenerators) that pushes combustion to completion will significantly reduce VOC emissions. VOC emissions from a hot regenerator are less than 10 ppmv. It is unlikely that an add-on scrubber will significantly reduce VOC emissions. (l) A question arises on how to calculate BACT emission rates if a CO boiler is in the flue gas train given the fact that large amounts of refinery fuel gas are usually combined with the regenerator flue gas in order to assure good combustion. Comments follow: VOC resulting from refinery fuel gas combustion can be estimated by using the AP-42 emission factor for nonmethane volatile organics. Regenerator-related VOC in the CO boiler outlet can be estimated by applying the DRE of the combustion device (typically 99 to 99.9 percent) to the VOC waste load of the inlet flue gas (about 220 lbs of VOC per 1,000 bbl fresh feed). CO resulting from refinery fuel gas combustion can also be estimated using AP-42. With regard to the flue gas, CO cannot result from the combustion of CO, but can result from flue gas VOC combustion. Regenerator-related CO in the CO boiler outlet can be estimated by applying a destruction efficiency of 99 percent to the CO waste load of the inlet flue gas. The final (combined) vent should be converted to ppmv to assure that the value does not exceed 500 ppmv. Refinery fuel gas related SO2 emissions can be readily estimated. Regenerator-related SO2 emissions can be calculated using a BACT concentration of 300 ppmv and the flue gas flow rate prior to fuel gas addition. If an add-on scrubber is used, the vent of the SO2 control device is the final vent so the combined flow is the one which is used in the calculation; however, a value of 200 ppmv or less will be applied. In this case, refinery fuel gas related SO2 is not separately calculated. The amount of thermal NOx (typically much less than the chemical NOx) can be estimated using normal emission factors. The use of burners which reduce thermal NOx emissions have been required in some cases. CO boiler related products of combustion PM can be estimated using AP-42. However, if the CO boiler is followed by an add-on scrubber, normal practice has been to apply the NSPS emission rate to the scrubber vent and take the rate as representing TSP. This treatment is possible because of the ability of the scrubber to reduce PM emissions well below the NSPS rate.