By: Brian K. Schimmoller, Managing Editor
Environmental issues never stray far from the center of discussion in the electric power industry. Despite an industry-friendly Administration, and repeated calls for "environmental certainty," however, the only certainty that exists is the certainty of tighter limits and the need for additional emissions control equipment. Power plant owners and operators must deal with this amorphous patchwork of emissions regulations and devise emission control strategies as best they can.
Multi-Pollutant Waiting Game
In recent years, multi-pollutant control, or integrated emissions control, has experienced a revival, primarily due to pending multi-pollutant legislation, but also due to a growing belief among power plant owners and operators that integrated control should prove less expensive than separate control. The list of multi-pollutant control technologies is long, and continues to grow as researchers seek financial support and/or host sites for demonstration projects.1 These technologies rely on various mechanisms to effect emissions reduction, from gas phase oxidation and plasma/electron beam to combustion modifications, wet scrubbing and dry injection.
Very few integrated emissions control technologies have advanced to the point where reliable cost data is available to make informed capital investment decisions, but several are moving in that direction. Powerspan's Electro-Catalytic Oxidation (ECO) process has been in operation on a 1 MW pilot slipstream at FirstEnergy's R.E. Burger plant since February 2002, and a 50 MW commercial demonstration unit at Burger is scheduled to come on-line in early 2004. At the pilot scale, emission reductions have been promising: 90 percent for NOx, 98 percent for SO2, 95 percent for particulate matter, and 80-90 percent for mercury. Power consumption for a 500 MW unit is estimated at 3-5 percent of plant power, comparable to existing pollution control systems.
The ECO Process relies on a barrier discharge reactor, situated downstream of a power plant's particulate control device, to generate high-energy electrons that lead to the formation of oxygen and hydroxyl radicals. These radicals then oxidize emission species into particulate matter and mists that are captured downstream in an ammonia scrubber and wet electrostatic precipitator.
Although pilot-scale results are promising, testing under various conditions through the addition of elemental mercury has proven difficult, primarily due to operation of the mercury semi-continuous emission monitoring systems (CEMS).2 In addition to numerous component failures, inlet mercury concentration measurements have been biased toward oxidized mercury due to reaction of elemental mercury with ash captured in the sample conditioning process. Powerspan installed an Apogee Scientific inertial separation probe to sample the inlet flue gas for mercury. Initial results indicated a greater fraction of injected mercury reports as elemental mercury than was measured using the probe supplied with the CEM system. Independent stack tests conducted by Air Compliance Testing using the Ontario-Hydro method showed average mercury reductions of 88 percent.
BOC Gases' low-temperature oxidation process, LoTOx, originally developed as a high-efficiency NOx control technology, has also shown multi-pollutant reduction capabilities. The LoTOx process relies on ozone injected into the flue gas at temperatures below 300 F to react with NO and NO2 to form soluble higher oxides that can be removed with a wet scrubber. NOx reduction efficiencies greater than 90 percent have been achieved, and because the ozone also oxidizes elemental mercury to soluble oxidized species, mercury removal is possible as well. Laboratory test data indicate that mercury reductions greater than 90 percent can be achieved for a range of coal types when LoTOx is used along with a wet scrubber.
At a demonstration project at the Medical College of Ohio, LoTOx used in conjunction with a simple wet scrubber provided 90-95 percent NOx removal.3 Mercury reduction was not measured. A separate engineering analysis for a lignite-fired cyclone boiler estimated NOx removal costs of $990/ton when linking LoTOx with other NOx technologies such as overfire air and modified SNCR to achieve a cumulative 90 percent reduction. At similar ozone dosage rates, and if a wet scrubber is present, LoTOx will simultaneously remove 90 percent of the mercury at little additional cost.
Mobotec USA recently evaluated the furnace injection of limestone and trona for SO2 and mercury removal. The project, performed at Progress Energy's Cape Fear Unit 5, was performed in conjunction with the ROFA and ROTAMIX NOx reduction systems in place at Cape Fear.4 Mixing in the ROFA and ROTAMIX systems create optimal conditions for achieving multi-pollutant reduction by providing ample turbulence and residence time within a specific temperature window. For the short-term tests at Cape Fear, SO2 reductions of 69 percent for trona and 64 percent with limestone were achieved; mercury reductions were 67 percent for trona and 89 percent for limestone. Trona provided better SO2, HCl, NOx and particulate matter reductions, while limestone provided better mercury reductions.
Praxair's oxygen-enhanced low-NOx technology creates conditions that promote the formation of molecular nitrogen rather than NOx. Photo courtesy of Praxair.
During the testing, excessive slagging occurred on the superheater tubes, necessitating a shutdown for cleaning. Although limestone has a lower ash fusion temperature than trona, trona sticks more tenaciously due to the nature of the chemical bond. Mobotec believes the slagging problem can be dealt with by injecting the sorbent in a furnace location where the combustion gases are below the ash fusion temperature and by operating the sootblowers more frequently.
A group of companies — Airborne Pollution Control, Babcock & Wilcox, USFilter, and LG&E Energy — are exploring dry sodium bicarbonate injection, coupled with enhanced wet sodium bicarbonate scrubbing, to provide SO2, NOx, mercury, and heavy metal reductions.5 Although sodium bicarbonate scrubbing is well known as an effective flue gas cleanup process, commercial application has been prevented by the high cost of sodium bicarbonate, the limited economic value of the scrubber product (sodium sulfate), and the economic and environmental issues associated with sodium sulfate disposal. Airborne Pollution Control has developed a recycling process — which will regenerate sodium sulfate back into sodium bicarbonate and a sulfate-based fertilizer product — that may eliminate the financial and disposal barriers. A 5 MW field verification facility has been built at Kentucky Utilities' Ghent Generation Station to evaluate the process.
Combining sorbent injection for mercury control with other technologies for NOx and/or SOx removal represents another multi-pollutant control option. Various companies are exploring sorbent and chemical injection techniques that can remove mercury at reasonable costs, including Sorbent Technolgoies Corp., ADA-Environmental Solutions, URS Corp., EPRI and Alstom Power. Results are promising — some novel sorbents outperform commercial activated carbon products — but additional feield testing is needed.
Washington Group International and EPRI1 have examined a number of integrated emissions control technologies and have performed detailed cost estimates of the LoTOx and ECO systems based on published information and interviews with the technology developers. Table 1 compares the capital, O&M, and total annual cost of these two processes with those of the base case. These figures are obviously open to debate; Powerspan, for example, estimates capital costs for ECO at $150-200/kW and levelized O&M costs at 2.0-2.5 mills/kWh, much lower than the $267/kW and 7.30 mills/kWh figures given in Table 1. Discrepancies in the capital cost estimates can be attributed to the retrofit factor assumed and to various other cost adders that were similarly applied to all technologies in WGI/EPRI's analysis. The large discrepancy in O&M costs is due primarily to a difference in the value of the credit for producing the ammonium sulfate fertilizer byproduct. Still, the fact that the costs are within the same order of magnitude as those of current commercial emissions control systems is encouraging. As more data is collected from test and demonstration projects, a better picture of the overall value of integrated emissions control technologies will emerge.
Keeping Pace with NOx
Despite the growing interest in multi-pollutant emissions control, the more immediate area of interest for power plant owners and operators remains NOx control. Bringing power plants into compliance with various NOx emission limits, such as those stipulated by the NOx SIP Call, has nearer term financial implications. Selective catalytic reduction (SCR) technology vendors have benefited tremendously from legislative NOx mandates, but a host of other technologies have succeeded as well, typically offering plant owners lesser reductions but at less cost. Power Engineering has covered many of these technologies in recent years, including combustion optimization, low-NOx burner upgrades, advanced overfire air, forced and induced flue gas recirculation, reburn and rich reagent injection.
Newcomers continue to join the NOx parade, however. Mitsui Babcock's patented NOxStar technology will be installed at TVA's 200 MW Colbert Unit 4 this October, providing NOx reductions of 65-70 percent at much lower cost than SCR. The technology, which was originally developed for NOx reduction in stationary diesel engines, has undergone extensive engineering development and testing at Mitsui Babcock's technology center in the United Kingdom and a full-scale utility application at TVA's 200 MW Kingston Unit 9. The NOxStar system features a permanent reagent grid positioned within a cavity of the boiler convective pass. A reagent consisting of ammonia and a small amount of a hydrocarbon (natural gas or propane) is injected into the flue gas, creating a gas-phase autocatalytic reaction and reducing NOx to water vapor and elemental nitrogen. The NOxStar system does not require any changes to existing plant layout, ductwork or the draft plant.
NOxStar can achieve NOx levels of less than 0.15 lb/MMBtu alone. When combined with other combustion-related NOx reduction technologies such as low NOx burners, over-fire air or back-end polishing devices, it can achieve final NOx levels as low as 0.10 lb/MMBtu within acceptable ammonia slip limits. The technology has demonstrated no sensitivity to coal type or boiler type. According to Scott Affelt, VP of Sales and Marketing for Mitsui Babcock, capital costs for NOxStar are $35-40/kW, about one-third those of traditional SCR technology. Only a 2-3 week outage is required, and lead-times can be as short as nine months, meaning systems could still be designed and installed by next year's NOx season. A number of power plant asset owners have shown great interest in this technology.
Praxair's oxygen-enhanced low-NOx technology adds another wrinkle to the world of combustion-based NOx control. By replacing a small amount of the combustion air with a stoichiometrically equivalent amount of oxygen, conditions are created that promote the conversion of fuel-bound nitrogen into molecular nitrogen rather than NOx. Oxygen enhancement increases the flame temperature and promotes early pyrolysis. It also shifts the combustion kinetics in the fuel-rich region near the burner toward conditions unfavorable to NOx formation.
The technology is applicable to all types of coal boilers, but Praxair is focusing commercialization efforts on the U.S. population of more than 500 wall-fired units less than 500 MW in size, where economics are most favorable. "For wall-fired coal units, a gap exists between a good economic NOx control alternative and the available technology choices. Praxair NOx compliance technology is geared to address that gap," said Rich Jarrett, Director of Business Development.
System design depends on a complete analysis of the combustion environment, including the low-NOx burners, the overfire air system, the coal type, boiler capacity, and pulverizer performance. The technology requires the installation of proprietary oxygen injection lances, but these can be plumbed into the existing low-NOx burners quickly and without taking the burners out of service. Exact placement of the oxygen lances, and the amount of oxygen added, will be determined following a detailed boiler analysis.
Following laboratory and pilot test results that demonstrated NOx levels could consistently be reduced below 0.15 lb/MMBtu, Praxair scaled the technology up for use at the coal-fired James River Power Station, owned and operated by City Utilities of Springfield, Missouri, in late 2002. Praxair conducted more than 30 days of testing at various loads, fuel blends (including PRB), and oxygen injection concentrations. The tests showed consistent 40-50 percent NOx reductions from the staged combustion baseline (reductions relative to unstaged combustion levels would be much higher).
Additional benefits a power plant could achieve include: (i) Immediate improvements in flame stability and flame attachment; (ii) Reduced opacity compared to air-only staging; (iii) Improvements in unburned carbon levels in the fly ash; (iv) Restored generating capacity lost due to fan capacity limitations; and (v) Reduced burner-to-burner variation since oxygen injection can be controlled to each burner individually.
As the James River station is not subject to SIP Call provisions, City Utilities removed the oxygen-enhanced combustion system following the test program. A second installation is in place at Northeast Utilities' 147 MW Mt. Tom Station in Massachusetts. Initial results are positive, and ongoing testing is underway to further quantify the operational benefits and NOx emissions reductions associated with this technology.
Although the cost of oxygen-enhanced combustion is not insignificant, due to the use, handling, and sophisticated control required with oxygen, the technology offers a competitive alternative to boiler operators. "The lease and operating costs for the equipment are significantly less than those of SCR, and clearly lower than costs that would be incurred buying NOx emission credits," said Jarrett.
Simple-Cycle SCR Headaches
Controlling NOx emissions from gas turbines historically has been achieved through low-NOx burners. As emissions regulations have tightened in certain parts of the country, however, post-combustion emissions control is increasingly required. While selective catalytic reduction (SCR) for combined-cycle facilities is a proven technology, with a number of units achieving NOx emission levels as low as 2 ppm with 2 ppm ammonia slip, SCR has encountered greater difficulty when applied to gas turbines in simple-cycle configurations.
"We're still on a learning curve with respect to SCR for simple-cycle machines, especially for F and E class turbines," said Joel Chalfin, GT/CC Environmental Compliance Manager with GE Power Systems during a panel session at the IGTI Turbo Expo in June. "The higher exhaust temperature and higher flows make SCR application on heavy duty gas turbines very difficult and costly. When you look at cost effectiveness numbers ($/ton of NOx removed) and the technology risk to install and operate simple-cycle SCRs on heavy duty gas turbines, the benefits do not appear to justify the expense. SCR is more appropriate for aeroderivative engines because of the lower exhaust gas temperatures and higher uncontrolled NOx emissions, but even there we've seen many problems."
Austin Energy and PPL Generation have both experienced some of the growing pains associated with SCR on simple-cycle gas turbine installations. In 2001, Austin Energy installed four LM6000 gas turbines for peaking duty to satisfy residential load swings. To achieve the 5 ppm permitted NOx level, the SCR systems were designed for 3.5 ppm NOx. Austin Energy installed air tempering fans — common for simple-cycle SCR installations — to reduce the exhaust gas temperature to the 830 F range desired for effective SCR operation.
After less than 1,000 hours of operation, the NOx emissions began to creep up over 5 ppm, with ammonia slip drifting above 7 ppm. Ammonia injection increased to maintain the desired emissions level, but this action caused the ammonia flow meter maximum to be exceeded. Moreover, the air heaters were unable to maintain temperature at the higher flow rate. As a result, the SCR provided less than 75 percent NOx reduction instead of the 80 percent design reduction, according to Oscar Backus, Division Manager of Environmental and Plant Technical Support Services with Austin Energy.
An engineering analysis indicated that the auxiliary equipment was undersized by about 25 percent based on the amount of ammonia needed. Austin Energy replaced the ammonia flow meter, the ammonia control valve, and the air heaters. Because the analysis revealed some premature SCR catalyst deactivation, Engelhard, which supplied the SCR system and the catalyst, replaced it as well.
Since making these changes, Austin Energy has been able to achieve the permitted NOx level, but careful control is required and questions still remain about the effect of peaking operation and high ramp rates on catalyst deactivation. Backus offers several recommendations when considering a simple-cycle SCR installation: ensure that the catalyst vendor guarantees performance for peaking operation; research catalyst types and temperature excursion limitations; don't underestimate the importance of inlet gas temperature; check the conditions of the catalyst pack periodically, not just during scheduled outages; and ensure auxiliary equipment is properly sized.
Bradley Piatt, Manager of Peaking Power for PPL Generation, experienced similar problems when he assumed responsibility for PPL's simple-cycle fleet, which included nine LM6000 SPRINT turbines in Connecticut and on Long Island. From the time the turbines entered operation, the 2.5 ppm NOx permit level presented a challenge. A number of small factors — including catalyst with low activity; imperfect packing, gasketing and sealing; catalyst frames that weren't straight; and uneven ammonia injection — quickly added up, leading to out-of-compliance NOx levels.
"We had no design margin between the permit level and the SCR guarantee level," said Piatt. "The systems were 90 percent effective, but to achieve compliance, we need close to 100 percent effectiveness. At 2.5 ppm, we're really pushing the envelope in terms of what SCR technology can do on simple-cycle units." To reach compliance levels, PPL Generation added depth to the catalyst beds, replaced the catalyst, and tightened down hard on operations and maintenance procedures. The units, which have accumulated 1,500-1,600 hours of operation to date, can now reach full load in 10 minutes and meet the permitted NOx level within 18 minutes, according to Piatt. Slip has averaged less than 6 ppm, within the permitted limit.
To guard against simple-cycle SCR problems, Piatt recommends, not surprisingly, that an acceptable margin be included between the design NOx level and the permitted NOx level to accommodate system inefficiencies and unforeseen design and operational issues. Piatt suggests making allowances for bigger pumps, deeper catalyst beds, and additional catalyst surface to facilitate process improvements after initial start-up. Operational review of the system at the design stage — when significant low-cost improvements can be made and re-work eliminated — is also important.
1 Keeth, Robert J., Michael R. Reed, and Chuck E. Dene, "Integrated Emissions Control — Process Reviews and Comparisons," Proceedings of POWER-GEN International 2002, Las Vegas, Nev., Dec. 10-12, 2002.
2 McLarnon, Christopher R. and Dan Steen, "Combined SO2, NOx, PM, and Hg Removal from Coal-Fired Boilers," Proceedings of Combined Power Plant Air Pollutant Control Mega Symposium, Washington, D.C., May 19-22, 2003.
3 Jarvis, James B., Adrian T. Day, and Naresh J. Suchak, "LoTOx Process Flexibility and Multi-Pollutant Control Capability," Proceedings of Combined Power Plant Air Pollutant Control Mega Symposium, Washington, D.C., May 19-22, 2003.
4 Haddad, Edwin, John Ralston, Geoff Green, and Steven Castagnero, "Full-Scale Evaluation of a Multi-Pollutant Reduction Technology: SO2, Hg, and NOx," Proceedings of Combined Power Plant Air Pollutant Control Mega Symposium, Washington, D.C., May 19-22, 2003.
5 Mortson, Murray E. and Fred C. Owens II, "Multi-Pollutant Control with the Airborne Process," Proceedings of Combined Power Plant Air Pollutant Control Mega Symposium, Washington, D.C., May 19-22, 2003.
Environmental Tuning Services
Gas turbines equipped with Dry Low-NOx (DLN) combustion systems have enabled turbine owners to reach extremely low levels of NOx emissions, often allowing them to achieve compliance levels without the use of selective catalytic reduction. Since emissions and combustion dynamics are tied closely to each other in DLN systems, tuning for emissions optimization can impact combustion dynamics, possibly leading to high-frequency pressure oscillations and the notorious "humming" experienced by many advanced gas turbine models.
Emissions data from a gas turbine plant's continuous emissions monitoring system (CEMS) can present confusing information. For example, if a plant equipped with a CO catalyst experiences high CO readings at the stack, it is difficult to tell from the stack CEMS data whether the problem lies with the combustion system or with the CO catalyst. Salt River Project's Santan Generating Station outside Phoenix encountered such a situation after Spring 2003 outages when it installed new air inlet filtration systems and CO catalysts on two GE 7EA DLN-1 gas turbines. When the units came back on-line, the CO emissions from one unit were the same or higher than they had been prior to the outage, and power output was lower by approximately 2 MW. Santan engineers suspected a combustion problem, but because the plant only had a stack CEMS, it could not determine if the combustion system was at fault or if the CO catalyst was not performing properly.
Sermatech's Environmental Tuning Services (ETS) vehicle features dual sampling lines for on-site emissions testing and combustion tuning. Photo courtesy of Sermatech.
Santan considered several of the many reputable emissions testing companies to assess the CO catalyst performance, and selected Sermatech Power Solutions' Environmental Tuning Services (ETS) vehicle. "Because of the complicated nature of the turbine's DLN system, we wanted to be sure we worked with someone who had specific insight into the complex relationship between combustion dynamics and turbine emissions," said Bob LaRoche, Senior Engineer at Santan. "Sermatech was extremely responsive to our needs. Their on-site engineers explored issues not originally in the scope of work, and identified several areas where additional improvement is possible."
In late July 2003, Sermatech's mobile testing vehicle arrived at the Santan Generating Plant. Engineers set up parallel sampling lines on the stack and the gas turbine exhaust duct, and instrumented the combustors to measure combustion dynamic pressures. After traversing the duct upstream of the CO catalyst, Sermatech found that the average CO emissions were below the turbine OEM's guarantee, but that local spots were above the guarantee. This suggested problems with the combustion system. Santan is currently evaluating the test data to determine what corrective actions are necessary to eliminate the high CO concentrations. One possibility is that a fuel nozzle in one of the combustion cans is not functioning properly, according to LaRoche. To evaluate this theory, Santan hopes to perform in-place flow testing of the primary and secondary fuel nozzles. This testing uses air to flow-check the fuel and atomizing air passageways for compliance to flow specifications.
If natural gas prices continue to increase, power suppliers will be forced to depend more heavily on liquid fuel for their gas turbine fleets. Liquid fuel operation requires greater diligence in the maintenance and monitoring of the machines' vital systems. Sermatech's ETS is uniquely suited to perform a combustion system "health check" that can identify areas of concern before they turn into major problems. The vehicle's capabilities include emissions measurement, combustion dynamic pressure monitoring, turbine performance measurement, and vibration analysis.