Sulfur Recovery Units (SRU) March 1995 Technical Disclaimer References to abatement technologies are not intended to represent minimum or maximum levels of Best Available Control Technology (BACT). Determinations of BACT are made on a case by case basis. BACT determinations are always subject to adjustment in consideration of specific process requirements, air quality concerns and recent developments in abatement technology. Additionally, specific health effect concerns may indicate stricter abatement than required by the BACT determination. The represented calculation methods are intended as an aid in the completion of acceptable permit applications; alternative calculation methods may be equally acceptable if they are based upon, and adequately demonstrate, sound engineering assumptions or data. These guidelines are applicable as of the date of this document, but are subject to revision during the permit application preparation and review period. It is the responsibility of the applicants to remain abreast of any guideline or regulation developments which may affect their industries. BACT Guidelines Sulfur recovery units (SRU) maybe required at onshore natural gas plants and are, in general, expected at petroleum refineries. The typical SRU proposed is a three-stage Claus with a tail gas treating unit (TGTU) if greater than ninety-six percent recovery is required. Other types of sulfur recovery technology installed in Texas are Cold Bed Adsorption (CBA), MCRC, SulFerox, Clinsulf, Selectox and LoCat. Although much of the following discussion pertains to Claus units, the same issues: control effectiveness, reliability, on-stream time and enforceability must be addressed regardless of the type of SRU proposed. Sulfur Emissions Sulfur emissions are emitted from process fugitive components as hydrogen sulfide (H2S), sour water tanks, SRU as sulfur dioxide (SO2) and H2S, sulfur pits and loading operations. The thermal oxidizer or tail gas incinerator (TGI) is expected to oxidize H2S to SO2 with an efficiency of 99.9 percent. Acid gas flares are assumed to ninety-eight percent efficient in converting H2S to SO2. Any acid gas flaring must be handled by a flare that can meet the design and operation requirements of Title 40 Code of Federal Regulations Part 60.18 (40 CFR 60.18) for maximum tip velocities and minimum British Thermal Unit (BTU) values. The sulfur pit and sulfur loading operations are expected to be controlled either through a recovery system routing vapors back to the process or by routing the vapors to an incinerator or other control device. Much has been written by the TNRCC on sulfur recovery as a control technology. For further details regarding SRU in Texas, see the following documents: "BACT Criteria for SRU in Texas" presented to the Lawrence Reid Gas Conditioning in Norman, Oklahoma, March 1993. "Refinery SRU Permit Considerations in Texas", December 10, 1993. "Refinery Acid Gas Systems Air Permit Considerations", December 10, 1993. The remainder of the sulfur emissions discussion specifies the concerns which should be addressed when implementing a SRU. Guidelines have been provided at the end of this document, Table 1, relating the expected efficiency and the potential sulfur to be recovered. The TNRCC does not endorse the use of a particular sulfur recovery technology. However, the proposed technology should meet the target performance level consistently and have high reliability. Exemptions provided in New Source Performance Standards do not supersede the requirements of a permit to meet a specified recovery rate. An air quality permit will be issued based on an expected recovery efficiency which is not variable with lower inlet concentrations. The proposed technology must be able to consistently achieve the represented efficiency over the expected range of operating conditions. Additionally, the control system should have minimal downtime and be equipped with alternative means to handle the acid gas other than long term flaring. Lastly, the control system should provide measures to ensure enforceability. Starting with control effectiveness, Table 1, Sulfur Recovery Efficiency Guidelines, lists the expected efficiencies based on the size of the facility being permitted. At petroleum refineries, SRU greater than ten long tons per day (LTPD) will probably need sort of TGTU technology to achieve the guideline. In order to maintain ongoing reliability, measures should be incorporated into the facility design to insure proper operation. A high efficiency is meaningless if the SRU is not on-line. To this end, the design must incorporate two ideas, proper upstream handling of acid gas and design of the SRU. Due to potential fluctuations in inlet gas flows and H2S concentration, maintaining acid gas quality to the SRU can be a problem at onshore natural gas plants, but tends not to be a problem at petroleum refineries. Feed to a conventional straight through type Claus plant should maintain a concentration of at least fifty percent H2S. If concentrations of forty percent H2S or less are anticipated, additional measures such as split flow design, acid gas enrichment and feed/air preheat should be considered to maintain stable combustion in the Clause furnace. A conventional Claus unit is not recommended for concentrations of thirty percent H2S or less and other sulfur technologies should be explored. These values were obtained from available literature and do not necessarily reflect the position of the agency. A straight configuration is usually recommended for a Claus unit where the H2S feed concentration exceeds forty percent. With lower concentration feeds, stable combustion is difficult because a smaller portion of the entire feed stream will be oxidized in the furnace. The split-flow design allows part of the flow to be routed directly to the waste heat boiler. Utilizing split-flow allows an increase in the flame temperature in the front burner zone to properly destroy ammonia (NH3) and hydrocarbons (HC) and oxidize H2S. Problems with this design include the potential for some HC, NH3 and other contaminants to be fed directly from the amine acid gas to the first catalytic reactor. This may be eliminated as a concern through the use of burners utilizing high intensity, swirling action to aid the destruction of these contaminants. Contaminants in the acid gases being fed to the SRU can drastically affect not only performance but also reliability. The two main contaminants are NH3, which is not usually a problem at onshore natural gas plants and HC. NH3, if not handled properly, will cause NH3 salt deposits not only in the SRU but in any associated TGTU. Slugs of HC can upset the chemical balance in the thermal reactor causing decreased sulfur conversion. HC can also foul catalyst in the converter beds. In order to minimize NH3 in the amine acid gas, adequate water wash should be provided at the upstream hydrotreating and hydrocracking units. The maximum amount of NH3 should be removed prior to treatment in the amine units. Otherwise, NH3 will carry through with the amine acid gases to the Claus thermal reactor. Since NH3 is not expected in the amine acid gas stream, the NH3 will not be properly destroyed in the Claus furnace, thereby contributing to the formation of NH3 salts in the catalyst beds. If thorough water wash is not feasible at upstream units, then the water draw-off from the amine regenerator overhead drum should be increased to assure minimum carryover of NH3 to the Claus reactors to avoid this problem. To prevent HC slugs from carrying over into the SRU, adequate retention time in the sour water system and the amine system should be provided upstream of the SRU. Typically, acceptable retention times for sour water have been three to five days to allow for adequate HC separation. A retention time of thirty minutes is an acceptable value for the amine surge tanks. Further, operation of the rich amine flash tanks at the lowest possible pressure should maximize the removal of light HC. This requires a low-pressure gas line returning flashed vapors back to the upstream processes or to a flare. SRU Design Now that upstream considerations have been considered, the SRU design should be evaluated for reliability. Reliability plays a major part in the review of a proposed technology. Consideration must be given to acid gas handling during SRU downtime. An adequate system will have an on-stream time of ninety-eight to ninety-nine percent. The TNRCC expects the applicant to limit acid gas flaring as much as possible. To that end, the preference is for no acid gas flaring during scheduled maintenance. For larger facilities, parallel trains will prevent acid gas flaring by allowing one train to process the acid gas if the other goes down. To put this in perspective, for onshore natural gas plants, parallel and redundant units are especially preferred if the amount of sulfur recovered is greater than eighty LTPD. For petroleum refineries, this level is ten LTPD. The ability to interchange should be built in the design such that each train has the ability to process seventy-five percent of design maximum load for the total sulfur complex. Therefore, if one SRU or TGTU goes down, the other can operate at 100 percent of its design maximum and production will only have to be cut back twenty-five percent of the combined total sulfur throughput for both trains or fifty percent of the throughput for each individual train. If the applicant can not propose redundant and parallel systems, a curtailment system should be proposed to allow for no continuous flaring in the event of an upset at the SRU. For partially redundant systems, acid gas curtailment will still be necessary as the applicant is still restricted to operating within the permitted maximum allowable emission limits. The applicant should consider excess capacity in the sour water and hydrotreating feed holdup to allow the petroleum refinery to eliminate or reduce processing during SRU downtime, thereby reducing the acid gas load to the SRU. Additional reliability issues are the proper design of the thermal reactor, the ability to by-pass the TGTU and feed the acid gas directly to the incinerator, the availability of installed or warehoused sulfur pit pumps and operation of the tail gas incinerator (TGI) at a minimum high temperature cutoff of 1800F. A high temperature cutoff should help prevent early shutdown of the TGI which would result in H2S emissions from a cold stack. The thermal reactor design should include considerations for spare equipment such as blowers, pumps, etc., proper burn technology to accomplish destruction of low concentrations of HC (less than three volume percent) and NH3 (about 300 ppmv) and adequate residence time. Process control considerations include the installation of a tail gas analyzer to insure that the H2S/SO2 ratio will be maintained at an optimum value of two to one. Typically, the analyzer sends a signal to adjust the air supply according to demand. Future permit reviews will emphasize more process control to aid in tracking NH3 and HC in amine and acid gas streams. The design of reheat between stages in the Claus unit has also been examined for its effects on sulfur recovery and reliability. In hot gas by-pass, boiler outlet gases usually contain large sources of uncondensed elemental sulfur which may result in a reversal of the Claus reaction. Since hot gas by-pass reheat results in a lower overall net sulfur recovery, this type of reheat is discouraged. Direct heat mixing of the reactor feed and combustion products may result in the formation of sulfur trioxide if the air control is not accurate. Low air rates may result in the production of carbon from HC in the feed. These problems may lead to a shorter catalyst life. Indirect reheat methods result in the best sulfur recovery rate and potentially longer catalyst life. The conclusion is a preference for an indirect heat transfer provided by direct fired heaters. As mentioned earlier, with the split flow design, HC slugs are a threat to the first catalyst bed, since a portion of the flow is not routed through the combustion zone. For a split flow design, the applicant should adequately address this problem. To ensure enforceability of permitted emission rates, the permit will include performance testing, continuous measurement of emissions and record keeping requirements. This will include continuous emissions monitoring systems (CEMS) for SO2 and oxygen and in some instances carbon monoxide (CO), stack testing to calibrate the CEMS and to demonstrate initial performance for all pollutants, calculations of sulfur recovery, record keeping of sulfur production, gas processing rates, start-up, shutdown, upset and major maintenance information. In summary, an acceptable BACT discussion of SRU will demonstrate that sulfur recoveries will be consistent with the guidelines, the overall design of the upstream units will provide good separation, the SRU will be designed for stable operation, redundancies that minimize downtime will be installed and during SRU downtime acid gas will be handled consistent with regulations. Other Pollutants Because the thermal oxidizer or TGI is a combustion device, the unit will also emit products of combustion such as CO and nitrogen oxides (NOx). For CO emissions, a maximum of 100 ppmv is expected. For NOx, low NOx technology should be used to achieve a maximum NOx emission rate of 0.06 pounds per million BTU fired. Sulfur Recovery Efficiency Guidelines March 1995 Table 1 Expected Control Efficiencies (in percent) Modified Facilities New Facilities Onshore Natural Gas Plants 50