Nitrogen Oxide (NOx) Emission Requirements for Gas Turbines are Becoming Tighter and Tighter.
Last year, California adopted a new standard for best available control technology (BACT) for gas turbines that calls for 5 ppmvd of NOx emissions from new simple cycle gas turbine plants and 2.5 ppmvd NOx for new combined cycles.1 Meanwhile, the Massachusetts Department of Environmental Protection mandated that NOx emissions in new combined-cycle and cogeneration plants should not exceed 2.0 ppmvd.
Even for existing machines, the need to purchase NOx emission credits creates an incentive to find ways to lower NOx emissions. The U.S. EPA now requires that every major new source of emissions must be offset by a reduction of an equal or greater amount of emissions from an existing source. The price in California in 1998 for NOx offset credits averaged $11,750 for one ton per year of emissions. For power generators, the most feasible method of creating these emissions ‘offsets’ is often reducing emissions at their older power plants.
In this regulatory climate, gas turbine operators are searching for cost-effective means to reduce NOx emissions. An analysis of laboratory data and a screening level feasibility study have indicated that enriching natural gas with 10 to 15 percent hydrogen (H2) could be an economically attractive way to reduce NOx emissions from some gas turbines.
Since the 1970s, gas turbine operators have used steam or water injection into the combustor to control NOx emissions. NOx formation increases exponentially with flame temperature, so by adding water or steam, the flame temperature decreases and NOx emissions fall as well.
A drawback to steam and water injection is that a reduction in flame temperature also tends to increase CO emissions. With Pratt & Whitney’s FT-4 gas turbine, for example, when using water injection, the FT-4 can only decrease NOx to 75 ppmv without having the CO emissions exceed 100 ppmv, and at 25 ppmv NOx, the CO emissions are at 250 ppmv.2
Since the mid-1980s, gas turbine manufacturers have been offering dry, low NOx (DLN) combustors, which produce low NOx emissions without the addition of water or steam. DLN combustors are designed to produce pre-mixed flames, in which the fuel and air are mixed together in a region where ignition does not begin. The advantage of a pre-mixed flame is that lean air-fuel ratios can be used, where more air is present than is needed to completely burn the fuel. The presence of the excess air serves the same purpose as water or steam injection: it provides added thermal mass which decreases the flame temperature.
One problem faced by DLN combustors is the ability to operate at part-loads. Because the flow of air through a gas turbine is essentially constant, at part-loads the air-fuel ratio is leaner than it is at full load. At some point the lean flammability limit of natural gas is approached and problems arise with flame stability. DLN designers avoid this problem by adding equipment to adjust the amount of air going to the pre-mixing chamber (i.e., variable geometry), or by switching to a diffusion flame at low loads. Both of these techniques, however, tend to increase the complexity and cost of DLN combustors and lead to higher NOx levels.
Pratt & Whitney’s FT-4 gas turbine is one candidate for the use of hydrogen-enriched natural gas for emissions reduction.
Currently, most gas turbine manufacturers cannot guarantee NOx exhaust of less than about 25 ppmvd (corrected to 15 percent O2 in the exhaust) when firing natural gas. Consequently, to meet the single digit limits now being demanded by regulatory authorities, a selective catalytic reduction (SCR) system must be installed downstream of the turbine.
When looking at the options for achieving NOx emission reductions from existing gas turbines, it becomes clear that few cost-effective options are available. The one exception is the use of water injection. If demineralized water is already available, water injection can generally be implemented on a 20-30 MW gas turbine for less than $50,000. If a water demineralization system has to be added, the cost jumps to $150,000. However, because of the inverse relationship between CO and NOx, the amount of NOx reduction that can be achieved before running into CO emissions limits may be small.
Because of the complexity of DLN designs, retrofitting an older gas turbine with a DLN combustion system is a costly option. In addition to the combustors themselves, new fuel valves and fuel control systems must be provided. Retrofitting DLN combustors to a 20-30 MW gas turbine will typically cost $1 million, which translates to $33-50/kW. Consequently, DLN combustors are typically only installed in new machines.
Adding SCR to an existing combined-cycle unit is prohibitively expensive unless the HRSG was designed with room for the SCR catalyst rack, in which case the cost will be on the order of $25/kW. The SCR will also result in a slight increase in heat rate due to the added pressure drop through the exhaust system.
Adding SCR to an existing simple-cycle plant is also an expensive undertaking. In addition to the SCR, the existing exhaust stack must be moved to allow room for the SCR. The exhaust duct also typically must be widened considerably to reduce the flow velocity to levels required in SCR systems.
Hydrogen-Enriched Natural Gas
Various studies have examined the feasibility of using hydrogen-enriched natural gas as a method for achieving NOx reductions in gas turbines. GE has conducted a series of tests on H2-rich fuels at their combustion lab in Schenectady, NY.3 These tests have confirmed the superior flame stability of H2-rich fuels. The tests have also shown that natural gas with 10-20 volume percent H2 will also have improved emissions performance. Specifically, H2-enriched natural gas was shown to allow lower NOx emissions with steam injection (for a given CO limit), and to provide a greater flame out operating range with a pre-mixed combustor.
Perhaps the most extensive experience burning an H2-rich fuel in gas turbines comes from the group of integrated gasification combined-cycle (IGCC) plants that were built in the 1990s. In an IGCC plant the gas turbines are fired with syngas that typically contains more than 20 volume percent H2 (dry basis) and even greater amounts of CO. Representatives of the Dutch IGCC project, Demkolec, have published data showing that NOx emissions for their Siemens V94.2 gas turbine range between 6 and 30 ppmv, with less than 5 ppmv of CO, when running on syngas with approximately 60 volume percent CO and 30 volume percent H2 (Figure 1).4 These remarkably low levels are achieved by diluting the fuel with both N2 and water vapor to the point that the adiabatic flame temperature of the diluted syngas is between 3,275 and 3,450 F. This is about 350 F cooler than the lowest stable natural gas flame temperature. Consequently, when running on natural gas, the turbine can only limit NOx emissions to 150 ppmv.
Adding H2 to natural gas would mitigate both the high CO emissions problem in steam and water injection applications and the part-load dynamics problem in DLN applications. When compared to natural gas, H2 has a higher flame speed and broader flammability limits (Table 1). The latter means H2 will continue to have a stable flame at leaner conditions than natural gas, while the former means that the kinetics of H2 combustion are much quicker than that of natural gas, and consequently, the addition of a thermal moderator such as steam or water will be less likely to quench the reactions before completion. This enhanced reaction rate occurs as a result of the increase in the radical pool that accompanies the addition of H2. Since the conversion of CO to CO2 is largely controlled by the OH radical, the CO reaction is more likely to continue to completion with added H2 even at somewhat lower temperatures. The Demkolec experience cited above is an example of this.
While the benefits of using H2 as a fuel supplement may be well-established, the practical aspects of supplying H2 to a gas turbine are not. Several studies have examined the feasibility of using steam reforming of methane in a so-called chemically recuperated gas turbine (CRGT) cycle.5-6 However, evaluations have revealed that a CRGT cycle would be much more expensive than a conventional combined cycle. The designers of CRGT cycles have tried to maximize methane conversion in order to reach efficiencies that are comparable to conventional combined cycles. As a result, the temperature pinch points of the steam reformer are quite small, which leads to the high capital cost.
There may be an opportunity to use steam reforming to partially convert natural gas to H2, and thereby receive the benefits of H2-rich fuel combustion without the cost of a full CRGT cycle. However, the simplest approach is to purchase bulk quantities of liquid hydrogen from an industrial gas supplier. An H2 supplier will generally lease equipment for storing and vaporizing the hydrogen so that the up front capital cost is minimized. This makes it an attractive option for simple cycle applications where the small number of operating hours per year cannot justify large capital expenditures. The price of liquid H2 depends on the region of the country and the quantity of H2 being purchased. Prices can range from $2-$6 per thousand scf. Based on a calorific value for H2 of 274 Btu/scf, these prices translate to $7.30-21.90/MMBtu, which is clearly more than the price of natural gas.
However, when burning a mixture of 12 volume percent H2 and 88 volume percent natural gas and using water injection to limit NOx to 25 ppmv, economic analyses indicate that the total cost of achieving the NOx reductions would be less than the cost of retrofitting a turbine with DLN combustors. This analysis was based on applying the technology to a simple-cycle FT-4 gas turbine operating 500 hours per year. The possibility of generating H2 on site by adding an HRSG and a steam reformer was also investigated, but the costs proved to be higher than trucking in liquid H2. The details can be found in an ASME technical paper,7 but the results are summarized in Table 2.
1 Chin, G. et al., “Guidance for Power Plant Siting and Best Available Control Technology,” California Air Resources Board, Sept. 1999.
2 Levine, P., “Gas Turbine and Combined-Cycle Capacity Enhancement,” EPRI Report TR-104612, 1995.
3 Maughan, J.R., J.H. Bowen, D.H. Cooke and J.J. Tuzson, “Reducing Gas Turbine Emissions through Hydrogen-Enhanced, Steam-Injected Combustion,” Proceedings of ASME Cogen-Turbo Conference, pp. 381-390, 1994.
4 DeWinter, H. and J. Eurlings, “IGCC Buggenum Commercial Operation,” 18th EPRI Gasification Technologies Conference, San Francisco, Oct. 1998.
5 Janes, J., “Chemically Recuperated Gas Turbine,” California Energy Commission Draft Staff Report P500-90-001, 1990.
6 Adelman, S.T., M.A. Hoffman and J.W. Baughn, “A Methane-Steam Reformer for a Basic Chemically Recuperated Gas Turbine,” ASME Journal of Engineering for Gas Turbines and Power, Vol. 117, pp. 16-23, 1995.
7 Phillips, J.N. and R.J. Roby, “Enhanced Gas Turbine Combustor Performance Using H2-Enriched Natural Gas,” ASME technical paper 99-GT-115, presented at ASME Turbo Expo ’99, Indianapolis, June 1999.
Jeffrey N. Phillips is a Project Manager with Fern Engineering, Inc., a consulting engineering firm that provides expert advice and design service to a worldwide clientele of gas turbine users.
Richard J. Roby is President and Technical Director of Combustion Science & Engineering, Inc., a research, development and consulting engineering firm that specializes in experimental and computational analysis of gas turbine combustion systems.