Even as engine designs have been improved, reducing nearly 90% more emissions than they did a decade ago,  the need for aftertreatment isn’t expected  to diminish anytime soon.

By Diane McDilda

Older engines were not designed with today’s—or tomorrow’s—air emission standards in mind. This goes for both rich-burn and lean-burn engines. Appropriate emission treatments depend on the type of engine. And as regulations become more stringent, the challenge to keep these engines in compliance increases.

The California Energy Commission (CEC) estimates that approximately 10,000 MW of power must be added to the grid by 2013 to accommodate a growing population and resultant energy demands. This additional electricity is expected to come from both centralized and distributed generation (DG) facilities. Although DG plants comprise only a small portion of the power supply, air emissions per megawatt produced are higher than those for centralized plants. This is primarily because centralized plants have been stocked with cleaner gas turbines while, because of size, DG relies on reciprocating engines that haven’t yet experienced the full extent of tightened emission. But the times, they are a-changin’.

Even though DG was once touted in southern California as essential for a burdened centralized power system, it’s now considered part of the air-quality problem, meaning the DG systems are highly scrutinized by regulators, and air handling equipment must be spot-on to get permitted and meet specified operating conditions. This is primarily a concern where the South Coast Air Quality Management District (SCAQMD) is revamping plans and revising regulations to reduce stationary source emissions.

Not only are rule changes being proposed, but enforcement is more heavy-handed. Several years back, SCAQMD personnel armed with portable analyzers visited the DG facilities in Orange County and the urban portions of Los Angeles, Riverside, and San Bernardino counties. What they found was that over half of the engines tested were not in compliance with emission standards.

But not all DG systems are created equal. There’s the standard rich-burn engine, and then there is the relative newcomer to the field, the lean-burn engine. The majority of DG facilities use rich-burn generators that aim to operate at the optimal stoichiometric ratio of oxygen to fuel. These engines rely on an air-fuel ratio controller to ensure the proper mix is fueling the engine. Controlling the mixture can improve engine efficiency and lessen wear and tear, but at least as important is its responsibility in managing emissions. Rich-burn engines often include three-way catalytic converters to oxidize carbon monoxide and unburned hydrocarbons to form water and carbon dioxide, reducing nitrogen oxides to nitrogen and oxygen. When fuel ratios are optimized, so is the performance of the catalytic converter, meaning fewer pollutants escape from the exhaust. During the SCAQMD’s assessment, however, 50% of the rich-burn engines tested were found to be out of compliance with respect to nitrogen oxides and carbon monoxide alike.

The system includes a 1,500-kW Caterpillar G3516C LE lean-burn engine.

Lean-burn engines rely less on controlling the air-fuel ratio and depend more on controlling the additive used to initially reduce nitrogen-oxide emissions. Oxidation of carbon monoxide and unburned hydrocarbons then follows. This is accomplished through selective catalytic reduction (SCR), which treats the exhaust in two phases. Exhaust from lean-burn engines contains much less oxygen than that from rich-burn engines, and catalytic converters aren’t capable of reducing nitrogen oxide to acceptable levels. The SCR process incorporates the injection of a reducing agent, such as ammonia or urea, into the exhaust stream. The exhaust then runs through a catalytic converter’s ceramic honeycomb oxidation catalyst coated with precious metals, oxidizing carbon monoxide and hydrocarbons into carbon dioxide and water. A small portion of lean-burn engines was tested during the SCAQMD’s evaluation, and only 27% were found to be out of compliance for carbon monoxide and nitrogen oxide.

Tulsa, OK–based Miratech Corp. manufacturers SCR units used on lean-burn engines around the world. Nick Detor, regional manager for Miratech in California, is familiar with the emissions treatment for both rich- and lean-burn engines and how the tighter emission standards are apt to affect the industry. For newer engines, meeting emissions standards will be less of a burden, while bringing older engines up to today’s standards will be much more complicated.

“For older technologies, it will be more difficult to pass modern emission regulations. As an engine ages, the parts age and the emission degrades,” explains Detor. “It’s not the same combination of technologies. Retrofitting older engines with modern ignition systems and updated fuel delivery systems will only take them so far. I suspect that it may not make economic sense to upgrade engines older than 20 years.” Detor does believe that while owners may consider replacing older engines, newer engines with up-to-date pollution control systems that are well-maintained should be able to stay in compliance.

The decision to implement DG or upgrade existing engines, however, is rarely based solely on economics. Like other institutions, the University of Redlands in California decided to implement its own DG after rolling blackouts and brownouts in 2000 and 2001. At the time of the power outages, the university purchased power from Southern California Edison (SCE) at an interruptible rate, which cost $0.09 per kilowatt-hour. But the lower price carried with it other costs. Being part of the interruptible program meant that the university could be requested to go offline within 30 minutes from the time it received the request. If the university chose not to comply with the request, it incurred excess energy charges of $9 per kilowatt-hour. The university, with over 4,000 students enrolled, soon realized there was no convenient time to lose power and the reduced electrical charges were costing money elsewhere. Even over the summer, when student numbers are low, the university sponsors conferences with paying attendees who often stay on campus.

The university installed backup generators, but after consulting with Goss Engineering of Corona, CA, it was decided the school would install its own onsite cogeneration system. Cogen would provide electricity needed to operate the campus, as well as excess heat to warm and cool buildings—all with the consistency and reliability the university needed. A sensitivity analysis was performed and resulted in a calculated payback period of seven to 10 years, an acceptable time frame, especially considering that the main objective was reliability.

Goss Engineering designed the system, which includes a 1,500-kW Caterpillar G3516C LE lean-burn engine. The engine is an extension of the Cat G3500C family of low-emission gensets designed for optimum fuel economy and nitrogen-oxide control. Even with this advanced technology, the engine still requires aftertreatment for emissions. It’s equipped with an SCR manufactured by Miratech to reduce nitrogen oxide, carbon monoxide, and hydrocarbons.

“We looked at different engines and their abilities,” says Lucas Hyman, president of Goss Engineering. “This suited the clients’ needs more than the others considered. It was less expensive and had a little more thermal output.”

The system runs under the supervision of two shifts.

Redlands still draws from the grid to meet its demands, which range from 2,000 kW during the winter to 3,100 kW over the summer. The university hasn’t ruled out becoming completely independent from the grid, but it is waiting to see how producing its own energy works out. Although designed to operate independently, the system runs only under the supervision of two shifts between 6 a.m. and 10:30 p.m.

Thermal output is an important factor in cogen. While the term cogen implies one fuel source and two outputs, Kyle Landis, an engineer with Goss, points out that the system really functions more as trigeneration, or “trigen”: one fuel source with three outputs. The primary output is electric power, and as is typical with cogen systems, excess heat from the production of electricity is used to heat water and warm buildings. Unlike other cogen facilities, however, the process at Redlands incorporates an absorption chiller used to produce chilled water that is then pumped to air-handling units at each building on campus to provide air conditioning.

“This plant operates at 70% to 80% efficiency,” Landis says. “Usual power plants run at 30% to 35%.” Redlands efficiency exceeds even the 60% efficiency seen with most cogen facilities.

As part of the design permitting process, different factors were considered. Because of the system’s proximity to student dormitories, noise had to be controlled. With the goal of increasing the noise level no more than 1 decibel, the system was constructed within a building, which was then surrounded by a parapet. The secondary benefit afforded by the structure is that of protecting the equipment from the environment.

As expected, a large portion of the permitting process involved air quality. The University of Redlands contracted with Karl Lany of SCEC, a southern California environmental consulting firm, to handle the air permitting through the SCAQMD. John Furlong and Bob Conklin of SCEC conducted many of the analyses, using the Industrial Source Complex Short Term air dispersion model, as required by the SCAQMD, to consider building locations in evaluating ambient air-quality standards for nitrogen oxides, carbon monoxide, and particulate matter. Hazardous air-pollutant emissions were considered as part of the risk assessment work performed to determine the actual health impacts from the emissions on the local employee and student populations. To design the SCR system, these data were partnered with the expected exhaust emissions.

Jenny Sorenson, environmental health and safety manager for the University of Redlands, serves as a liaison between the campus and the SCAQMD for permitting, monitoring, and reporting of the system. She admits her role has been challenging because the system is large enough to be permitted—but not under Title V, which has very specific requirements.

“We’re in between,” Sorenson says. “We don’t have a specific [SC]AQMD inspector assigned.” This means she has no one she can consistently turn to for permitting or compliance questions. “Prior to the cogen we had the emergency generators. Now we have continued emissions monitoring [CEM], and we have to submit reports every six months. It’s been a challenge.”

The high thermal output from a natural-gas engine benefits cogen and trigen operations, but the downside is that the higher temperatures generate a higher concentration of nitrogen-oxide emissions. Nitrogen-oxide emissions from the system at Redlands can reach upwards of 150 ppm. To stay in compliance, it must be reduced down to 13 ppm. With the CEM in place, any deviation is recorded for the SCAQMD to access.

When the system at Redlands was constructed, emissions were tested to verify compliance. “The system was set by a technician using a hand-held analyzer to make sure the emissions were where they needed to be,” explains Nathan Brahm, chief engineer with Johnson Power.

Once set, the emissions tend to vary only on the load of the lean-burn engine and not other conditions, such as outside temperature or fuel quality. This is in contrast to rich-burn engines, which can fluctuate more easily.

“This unit is affected more by load than temperature. The load can go down to about 60% without affecting emissions,” says Hal Wondolleck, vice president and general manager with Johnson Power. “If the engine got down to 700 kilowatts, say, at night, the engine wouldn’t be efficient and the emissions would be impacted.”

The engine is an extension of the Caterpillar family of low-emission gensets.

Emissions from a lean-burn engine are reduced through the SCR, which utilizes ammonia to covert nitrogen oxide to nitrogen and oxygen. Ammonia can be added to the exhaust in several different forms: anhydrous ammonia, aqueous ammonia, and urea. As a hazardous material, anhydrous ammonia is rarely used, as it’s much more difficult to handle. Of the two remaining options, the decision likely depends on cost. Gallon for gallon, aqueous ammonia is less expensive than urea but has a higher capital cost because of more stringent construction requirements. Storage and use of anhydrous ammonia also require that a hazardous response plan be prepared for the facility. Urea costs slightly more to purchase than aqueous ammonia, but savings are seen with the lower cost of storing the material. Because urea is benign it can be flushed down the drain and requires no cumbersome disposal plan.

The University of Redlands chose urea over aqueous ammonia and stores it in two 1,100-gallon tanks. Fred Weck is facilities director at the University of Redlands and is responsible for operating the system. “The SCR uses approximately 3 liters of urea per hour only when the engine is running. The delivery is more efficient if the trucks drop off a full 2,000 gallons at a time,” says Weck. For now, the tanks are refilled every one to two months, a realistic schedule with events and increased car and foot traffic regularly taking place across the street from the storage tanks.

Included with the SCR is an injection control system that monitors the concentration of nitrogen oxide in the exhaust and adjusts the amount of urea dosing accordingly. The injection feedback system ensures that the proper amount of urea is injected. If the concentration of urea is too low, there won’t be enough ammonia to react with nitrogen oxide. If the concentration of urea is too high, excess ammonia is produced. Detor explains that the catalytic converter in the SCR unit serves as a backup and will treat low concentrations, just a couple parts per million of ammonia, converting it to nitrogen and water. If the catalytic converter is overwhelmed, excess ammonia can convert back to nitrogen oxide. This is known as “ammonia slip.”

Another potential problem that can occur with SCR is fouling. “We had that problem at Redlands,” says Detor. “Scale in the piping broke loose and affected the control valve in the SCR. All pipes must be flushed and deburred prior to starting up a system. It’s a problem that can be easily avoided.” Detor stresses the importance of properly cleaning pipes between construction and operation.

The system began operating in 2005 and to date is in compliance with its permit. “The initial construction was based on expected emissions, and they’re in line with what we’re seeing,” says Sorenson. “We’re permitted at 13 parts per million nitrogen oxide, and we’ve been well below around 7 to 9 parts per million. Sometimes we have a spike up around 13 parts per million, but it’s always within 45 minutes of startup.” The permit currently allows concentrations to exceed allowable levels during the allotted 45-minute startup time.

Concentrations are higher during startup because the exhaust temperature is not warm enough for the treatment systems to adequately support the necessary reduction and oxidation reactions. For urea to be converted to ammonia, the temperature of the exhaust stream needs to be a minimum of 500°F. Once stabilized, Redlands exhaust usually exits at 900°F. Some facilities continuously run hot, and the exhaust must be cooled before the urea is injected.

For the ammonia produced by the urea to react with the nitrogen oxide in the exhaust stream, the temperature must be between 580°F and 930°F. At Redlands, there’s little room between the urea injection point and the first stage of the SCR, so no considerable heat is lost and the temperature is sufficient to support the reduction reaction. However, when the engine is throttled back to 70% to 75% power, the exhaust temperature increases up to 950°F. To lower the temperature and support the oxidation reaction, a blower system turns on until the temperature decreases to a range of 850°F to 900°F.

The system is still operating under its construction permit, as an operating permit has not been issued. Construction of the system took a little longer than originally planned, and now that the SCAQMD is in the process of changing the rules, it’s anticipated that the new permit won’t be issued until the rules are finalized. There’s concern that emission levels will be reduced and that carbon monoxide will be added to the CEM requirement. Carbon monoxide had been included in the CEM regulations but was eliminated with other sweeping rule changes in 1997.

It’s also anticipated that permitted concentrations of nitrogen oxide will drop to 11 ppm, or lower, as part of the future reduction to Best Available Control Technology (BACT) levels. If the facility is required to abide by the California Air Resources Board (CARB) 2007 standards, nitrogen oxide emissions will be calculated at a rate of 0.07 pound per megawatt-hour, with a credit given of 1 MWh per every 3.4 MWh of waste heat recovered.

Another example of changes to the regulations that reflects a return to pre-1997 requirements is source testing. Every three years, Redlands is currently required to test emissions with equipment other than its CEM. If the proposed rules take effect, the frequency will be increased to once every two years or 8,760 hours, whichever comes first. Pretests will no longer be allowed, and if equipment is found to be out of compliance the tests cannot be aborted. Tests are currently performed by a third-party contractor, approved by the SCAQMD, and this is not expected to change.

There has been discussion on reducing the allowable 45 minutes at startup for a system to be out of compliance down to 15 minutes. The SCAQMD asked for empirical data from industry, and based on results it is thought that the time may be revised to 30 minutes.

Other monitoring and reporting requirements for Redlands include quarterly linearity testing that compares CEM readings to a variety of gases. “It’s fancy for calibration,” says Sorenson. Redlands also submits a relative accuracy test audit (RATA), which looks at the accuracy of the CEM equipment. As with the source test, RATA must be performed by a third-party contractor.

Changes to the SCAQMD regulations are anticipated by everyone involved with DG in southern California and are expected to be finalized by the end of 2007 and implemented in 2008. While it’s understood that air quality in the basin must be improved, it’s expected that changes will make construction and operation of DG more costly. By lowering emission standards, the SCAQMD is lowering the bar for aftertreatment manufacturers and equipment operators.

For doing its part in supplying its own power, the University of Redlands received a $750,000 rebate, 30% of the project cost, from the State of California. The state funded the rebate program, and SCE was responsible for administering it.

The Self-Generation Incentive Program (SGIP) was enacted to lessen the burden of power demands, particularly during peak times, and to encourage renewable fuel–generating technologies. Ironically, the allowable emission concentration of 0.07 pound per megawatt-hour proposed by CARB had been based on the requirements of the SGIP.

The lower standard came about, Detor explains, because it was initially anticipated that smaller, home-based generator systems too small to come under the guise of the SCAQMD would be installed under this program. Funding the initiative would be a way to control emissions.

These lower standards caught on in an effort to improve overall air quality in the basin. For sites like Redlands, lower emission standards are not considered a threat, as their engines and aftertreatment systems will likely keep them in compliance for years to come.

Diane McDilda is an environmental engineer and writer living in Gainesville, FL.

DE - September/October 2007