SCR Installation at Dynegy’s Baldwin plant. Photo courtesy of TVA.
Drive around just about any medium to large city these days and you’re bound to see the physical fruits of a booming economy: new housing developments, new corporate campuses, modernized airports, new sporting facilities, etc. Cranes dot the landscape, stretching their steel arms to erect another tangible reminder of the vitality of the U.S. economy.
What does this vitality have to do with power generation, other than the electricity that will be required to power these facilities? More specifically, what does it have to do with the focus of this article, NOx emissions control? Ironically, the strength of the economy may significantly impact the ability of power plant owners to install the selective catalytic reduction systems (SCR) required at many facilities to comply with air emission regulations. Crane availability and labor availability are just two of the myriad factors that must be factored into planning and implementing NOx control measures. Just about every aspect of power plant operation is impacted by NOx control decisions: plant design/layout, personnel safety, downstream environmental impacts, byproduct usage, maintenance schedules, parasitic power loads, draft system capacity, air heater operation, etc. The list goes on and on. Familiarity with these issues is essential in devising a reasoned, effective compliance approach.
The only place to start an article on NOx emissions control is with the current state of regulations. Of primary concern is the NOx SIP Call. On March 3, an appeals court ruled in favor of EPA in a suit brought by various groups, lifting a stay on a rule requiring 22 states and the District of Columbia to develop plans to meet tighter NOx emissions levels. Under the SIP Call provisions, all affected states will have to implement emission control measures by May 2003 to reduce NOx emissions during the ozone season; emissions must be reduced to specific levels by 2007. On April 11, EPA asked the court to lift the stay of the submittal date. If that stay is lifted, SIP plans would be submitted in two phases. Phase I SIPs, covering the majority of states, would be due Sept. 1, 2000; Phase II SIPs, encompassing reductions in states where the court ruled more analysis was needed (Georgia and Missouri), would be due no earlier than Dec. 1, 2000. Wisconsin has been removed from the SIP Call list for the time being.
Section 126 petitions, which were filed by several Northeast states seeking regulatory action against upstream emission sources in the Midwest, also impact NOx control decisions. In January, EPA finalized a remedy for the Section 126 issue. A Federal NOx Budget Trading Program will be implemented and become effective in May 2003, providing a cap-and-trade mechanism for reducing NOx emissions in the eastern U.S. EPA is initially allocating NOx allowances based on heat input, but updated allocations will likely be based on electricity output.
Because they target similar emission sources, there is a defined relationship between Section 126 and the SIP Call, according to EPA’s Paul Tsirigotis, who gave a keynote presentation in May at the National Energy Technology Laboratory’s Conference on SCR and SNCR in Pittsburgh. If a state receives approval for its state implementation plan, that SIP can replace any Section 126 actions. Further, sources in states subject to the SIP Call can trade with sources in states affected by Section 126, but not outside this region.
Choose Your Weapon
Although there is still opportunity for additional appeals to the EPA rulings, it appears almost inevitable that fossil plant operators will face NOx emission levels on the order of 0.15 lb/MMBtu in the next several years. While there is enough flexibility built into the system to enable asset owners to consider a mix of compliance technologies, the extent of the reductions required is forcing many to choose SCR. Demonstrations of selective non-catalytic reduction (SNCR) technologies, particularly Fuel Tech’s NOxOUT system, have been promising, but the majority of utilities to-date have decided that the biggest bang for the buck comes from SCR, despite its high initial capital cost.
Primarily because of the site-specific nature of SCR retrofits, capital costs vary all across the map, anywhere from $55/kW up to $130-140/kW, according to Ed Healy of Southern
Company Services. Further, retrofit costs have been somewhat higher than original EPA estimates, reports TVA’s Jerry Golden, Manager of Production Technology, and SCR projects have yet to demonstrate significant economies of scale (Figure 1). Although the EPA data in Figure 1 are based on SCR systems for low-efficiency cyclone boilers, while the TVA data include all types of boilers, the general trend remains valid.
Southern Company, which has about 15,000 MW of SCR projects underway or being planned to meet emissions requirements primarily associated with Title I of the Clean Air Act, has adopted a “no regrets” NOx compliance policy. In other words, they’d prefer to implement proven NOx controls on units that are the most cost-effective to control the current and futre requirements.
The two largest cost components of SCR retrofit projects are construction costs and catalyst costs. For the AES Somerset SCR system, construction costs comprised about 50 percent of the total capital costs and catalyst accounted for a sizeable 20 percent, according to Walter Nischt with Babcock & Wilcox. To reduce initial costs, emphasis must be placed up front on designing for constructability, using modular components for example. Additional cost-saving techniques, being applied by Southern Company, AEP and others, include volume procurement of SCR components (catalyst, sootblowers, dampers, ammonia systems, etc.) and system-wide alliances among utilities, engineering firms, catalyst suppliers and other equipment vendors.
The majority of SCR systems designed for U.S. boilers have selected a high-dust configuration (ahead of particulate removal), which lowers installation costs and eliminates the need for expensive flue gas reheating. For economic reasons, anhydrous ammonia is the reagent of choice, accounting for 92 percent of world SCR applications, according to Daniel Ott with Environex. Catalyst selection is fairly evenly split between plate and monolith designs, emphasizing the importance of carefully matching catalyst to the fuel being used.
If, as many expect, ozone season NOx limits extend across the country and also extend to year-round applicability (as is happening in certain parts of Texas), effective SCR performance with Powder River Basin coals will become critical, particularly since European and Japanese SCR experience has been limited primarily to bituminous coals. Preliminary results from Associated Electric’s New Madrid SCR system, therefore, are being watched closely. Catalyst deactivation is one of the biggest concerns with respect to PRB coal. Coupon test results reported on at the Second EPRI SCR Seminar for a Hitachi catalyst exposed to PRB flue gas have indicated significant catalyst deactivation after only 3,000 hours of operation. However, according to David Harris with Black & Veatch, which designed the New Madrid SCR system, after 850 hours of full-scale operation, there has been no appreciable deactivation of the Cormetech catalyst, which is guaranteed for 20,000 hours of operation. The New Madrid SCR has now operated for more than 1,000 hours with NOx removal in excess of 93 percent.
Air heater operation and maintenance is a key concern when designing an SCR retrofit project. To accommodate the interactive effects of ammonia and the upstream catalyst bed in a high-dust SCR, many utilities are having to make major modifications to the air heater. This can involve complete replacement, such as the installation of a new “SCR ready” air heater at New Madrid, or heater upgrades, such as the replacement of carbon steel last stage, cool-end heat exchanger baskets with enamel-coated baskets that can withstand the effects of condensed sulfuric acid. In some cases, according to TVA’s Jerry Golden, air heater modification costs can exceed the cost of the catalyst bed.
Excessive ammonia slip levels can lead to air heater fouling and potentially impact the marketability of fly ash. Getting ammonia slip below required levels, and keeping it there, demands careful attention. Somewhat ironically, in analyzing the technical issues surrounding the NOx SIP Call, EPA failed to consider the downstream effects of ammonia, according to TVA’s Golden. Stack emissions of ammonia and leaching from byproducts, for example, were not factored into the regulatory determination. Subsequent analysis and regulation, therefore, may be required.
SCR system specifications for ammonia slip at coal-fired plants are typically set at or below 3 ppm. New Madrid’s limit is 3 ppm, Southern Company specifies 2 ppm for its coal retrofit projects, and AES set a 3 ppm limit for its Somerset plant. Lower limits may be required. “At 2 ppm, frequent outages are necessary for air heater cleaning, even with enameled baskets,” said Environex’s Ott. “Lower limits, in the 1.0-1.5 ppm range, are necessary to increase time between outages, as well as to keep the ammonia concentration in the fly ash below 90-150 ppm and maintain its marketability.”
Imperfect distribution is typically the biggest contributor to high ammonia slip, highlighting the importance of detailed flow modeling to ensure that the ammonia concentration profile matches that of the NOx profile across the catalyst surface. Many designers have gone to great lengths to ensure “even” distribution by adding mixing devices, only to find that they cannot adjust for the non-uniformity of the NOx profile during plant operation. The best designs are those that allow adjustment of the localized ammonia concentration in the duct.
Extra catalyst layers have been proposed as a possible remedy to excessive ammonia slip. According to Ott, test results indicate that extra layers don’t provide much benefit if the ammonia distribution is poor. A better alternative is to add an additional ammonia injection grid to form a series reactor that can improve distribution. In such an arrangement, two 60 percent reactors would provide comparable reductions at lower slip levels than a single 80 percent reactor.
SCR installation at AEC’s New Madrid plant. Phot courtesy of TVA
Returning to the introductory theme of the article, there is still much concern about the ability of electric utilities to complete the number of NOx control retrofit projects within the available outage windows before 2003. While the shortfalls in catalyst supply and structural steel that some anticipated have not materialized, other issues have emerged. There are approximately 200 SCR retrofits that need to be done by 2003, ABB project manager John Buschmann told Energy Argus, and that doesn’t include new gas-fired plants. With only five or six major engineering companies specializing in SCR retrofits, that total equates to 30-40 projects apiece, which will strain engineering firm capabilities.
Project managers are seriously worried about crane and craftperson availability. There are not only concerns about the number of cranes available to meet SCR retrofit project demand, but also about the ability to get them on-site at the necessary time, since it takes time to dismantle and relocate large cranes.
Further, with the U.S. economy achieving record low levels of unemployment, finding the skilled craft force necessary to install an SCR system is no simple matter (Figure 2). “We have experienced labor shortfalls the last few years,” said TVA’s Golden, “and we expect labor shortfalls for the next four years in all outages beginning in the Fall of 2000.” During the 1999 Fall outage, TVA had to stop work on the Paradise Unit 2 SCR system because the workers were needed to get the unit back on-line to meet electricity demand.
EPA maintains that there should be enough time for all units to complete retrofit projects by 2003 without endangering system reliability, pointing out that the actual tie-in period is less than five weeks and that SCR retrofits are expected on only about 10 percent of affected units. The current pace of SCR announcements, however, threatens to push the 10 percent figure much higher. As an added safeguard against reliability problems, the NOx SIP Call allows states to use a compliance supplement pool to assist facilities that experience unexpected difficulties. This pool contains credits that equal more than one-third of a single year’s NOx budget.
Since costs are a major concern to those both using and considering SCR systems, opportunities to reduce those costs always must be explored. One area that has received much attention is in-situ regeneration of the SCR catalyst. The ability to recover lost activity from in-bed catalyst layers-at a fraction of the cost of new layers-represents a quasi holy grail for SCR operators.
EnBW Ingenieure GmbH, a subsidiary of EnBw, Germany’s third largest electricity generator, has been researching and developing regeneration techniques for several years. Since environmental laws in Germany are as strict or more so than in the U.S., many of the same objectives have to be met in the design and operation of an in-situ regeneration process. It has to be useable without removing the catalyst modules, it has to be simple, there can be no potential damage to the catalyst, the cost has to be a fraction of the cost of a new layer, and the process can not employ any dangerous chemicals.
The process EnBW developed is called ReACT, and it utilizes a water-based regeneration solution that contains special additives for unplugging closed channels, removing solids (dissolving the “crust” that forms on catalyst layers), opening catalyst pores and washing out poisons. According to information presented by Herwig Maier at the SCR/SNCR Conference, for a 760 MW plant, the process can clean a complete catalyst layer in five working days, treating 35 cubic feet per hour with a crew of 5-7 workers.
The process requires plumbing and spray systems to deliver the regeneration solution to the layer, and a collection basin below the module being treated to collect the spent solution. Because the regeneration solution is water-based, it can be easily handled after use, by boiling it away in the boiler, by treating it with conventional wastewater equipment, or by employing it in the flue gas desulfurization unit (if so equipped).
The amount of activity level gained by the catalyst is a function of the catalyst age. The more a specific catalyst module has degraded, the greater the increase in activation. For a catalyst with 50 percent activity before regeneration, the process recovers about 20 percent absolute activity; for a catalyst with 90 percent activity before regeneration, the gain is only about 2 percent. The ReACT process has been used to reactivate catalyst layers multiple times (at least three) at a number of German power plants, demonstrating that deactivation does not occur at a higher rate after reactivation.
Cormetech now offers a commercial catalyst regeneration process as well, and TVA’s Golden believes such as system can provide catalyst savings on the order of 30 percent.