by Robert Bland, Gas Turbine Efficiency
As traditional global sources of natural gas are exhausted, new nontraditional sources such as coal bed methane and imported liquefied natural gas (LNG) will supply a bigger fraction of the total demand.
Consequently, the variety and variability of compositions of natural gas in pipelines will increase and differ by region. These changes in composition might result in increased concentrations of higher hydrocarbons, producing changes to heating value and hydrocarbon dew point over relatively short time scales. The variations in fuel compositions, if large enough, can have a significant impact on the performance and operability of gas turbines, particularly for those with Dry Low NOX (DLN) combustion systems.
Gas interchangeability traditionally has been based on the gas Wobbe number, which is a measure of the volumetric energy density. Thus, if the gas is pure methane or a mixture of methane, inerts and heavier hydrocarbons, as long as the Wobbe number is the same, then the gases are considered interchangeable. Gas turbine manufacturers typically set other limitations on the levels of the heavier hydrocarbons, keeping the delivered gas within a relatively narrow range of compositions, which ensures that the systems meet performance and operability requirements. This approach historically has succeeded in maintaining a workable level of gas consistency, but that might change soon as these new nontraditional sources impact pipeline gas composition.
New gas sources, such as LNG, often have higher levels of the heavier hydrocarbons, ethane and above than classically have been allowed in the pipeline. The LNG supplier wants to maximize the energy content of the gas and has no local use for the higher hydrocarbons, such as chemical plants, at the liquefaction facility. Conventional diffusion combustion systems on gas turbines are relatively insensitive to these changes in fuel composition and probably will be largely unaffected by this change in fuel supply. The lower NOX premixed combustors, specifically DLN combustors, are considerably more sensitive to these factors.
DLN combustors by nature are not robust devices. To meet the low-emissions levels required, the systems operate close to the lean flame extinction limit because NOX output from a DLN combustor is primarily determined by the maximum flame temperature. For example, the required emissions output of a simple-cycle turbine with a turbine entry temperature in the range of 2,600-2,700 F, significantly above the melting point of metals, is essentially the same as for a domestic gas water heater, which runs at a significantly lower temperature. Historically, any improvements in robustness have been sacrificed to attain ever-lower emissions levels.
As a consequence of operating close to the physical flammability limit, DLN combustors are sensitive to wide parameters. Figure 1 shows the effect of fuel composition on the speed at which the flame propagates. Flame speed affects where the flame can exist in the combustor and how it ignites the fresh fuel air mixture continually entering the combustor. The impact of composition changes on this, and other fundamental characteristics of the combustion process can impact many performance characteristics of the gas turbine combustion system, some of which are:
- NOX and CO emissions,
- Combustion dynamics, which can impact the durability of combustion systems components,
- Turndown and stability of the combustion system, which can lead to blowouts and machine trips, and
- Auto-ignition or flashback of flame into premixers, which can lead to catastrophic damage to combustor and hot section components.
Constant Wobbe Index
Figure 2 shows such an incident on an ABB combustor where fuel composition changes produced a significant impact on system performance. The heating value of the fuel suddenly drops, and the emissions and dynamics respond adversely to the change.
With many DLN combustors, such as the GE 7FA+e DLN2.6, the fuel distribution can be modified spatially by the use of a series of independently controllable fuel circuits. This allows the system to be tuned for a specific set of circumstances, e.g., ambient conditions or fuel composition. If the combustor encounters fuel compositions outside the narrow range for which it was tuned, emissions or combustor dynamics can be affected adversely. If the variations are large enough, it might not be possible to tune the systems to meet performance and regulatory requirements.
There are two ways to address the issue. The first is for the original equipment manufacturers (OEMs) or third parties to design combustors that are more robust to fuel quality. Siemens has achieved this with its SGT6-5000F 9 ppm combustion system (ULN or Ultra Low NOX). Figure 3 shows the NOX output of the ULN compared with that of older 15/25 ppm NOX DLN combustion systems, as the fuel Wobbe number (energy density) is increased. The ULN system is insensitive to Wobbe number variation as a result of a more robust premixing technology.
This approach might not be a practical or cost-effective solution for some pre-existing DLN combustion systems. In this case, it is necessary to expand the range of fuel compositions over which the system can operate by a combination of fuel conditioning and continuous real-timing tuning of the combustion systems to maintain system performance when faced with variations in fuel compositions. This approach takes advantage of the combustion system’s ability to operate across a moderate range of conditions if correctly tuned for those conditions. In the past, tuning was performed by a specialist coming to site and modifying the control system settings. This was slow and expensive. In response, OEMs such as GE, Siemens and MHI, as well as third parties such as GTE, have introduced local, near real-time monitoring and control systems. These systems control the combustors, either based directly on the real-time emissions/dynamics data—Siemens, MHI and GTE—or on synthetic models—GE—by modulating the combustor fuel flow splits in response to excessive emissions or combustor dynamics. By continually monitoring the system, changes can be made as frequently as necessary and the performance of the system optimized. Figure 3 shows a number of the drivers and parameters affecting optimization.
In addition to real-time tuning, fuel conditioning and heating capabilities might be a requirement to keep systems operating properly as pipeline fuel compositions change. If the fuel contains heavier hydrocarbons greater than C4, then these can condense out in the fuel system. Liquids that are not captured or evaporated prior to entering the combustor can easily auto-ignite when mixed with the more than 700 F air present in an F-class gas turbine. Combustion will occur in areas that were not designed for it, significantly damaging combustor and potentially downstream hot gas path hardware.
To minimize such problems, an effective fuel-conditioning system is necessary. This takes the fuel though the multiple stages of separating the majority of the liquid, coalescing out any airborne droplets and finally heating the fuel to ensure it has an adequate dew point margin, to pass through the fuel delivery system without cooling to a level where condensation of any of the heavier constituents can take place.
Placing the fuel heater under the control of the auto-tuning system can provide another means of controlling the fuel characteristics to address performance impacts arising from variations in fuel composition. A possible addition to the fuel conditioning system is a fuel analyzer that can allow the fuel’s composition and heating value to be measured in real time. This information can either be fed directly into the control system and used to define the combustor fuel splits or recorded in the historian to help understand factors affecting turbine operation.
The variability in fuel composition that will occur with increased use of nonstandard gas supplies can adversely impact the performance of many gas turbines.
In some cases, new combustion hardware might be available to deal with these issues.
In others, available control and fuel-conditioning technologies exist that can mitigate the majority of the effects on turbine performance.
Robert Bland is chief technologist, combustion architectures, at Gas Turbine Efficiency. He obtained a doctorate from the University of Sheffield and has 20 years of computational fluid experience and 25 gas turbine patents.