Combustor design simulation requires the resolution of complex geometry, turbulent flow patterns, heat transfer, and detailed chemistry. Although computational fluid dynamics (CFD) can simulate the reacting flow in realistic geometries, it requires the use of severely restricted chemistry models that are too simple to accurately simulate emissions and operational stability. A new simulation tool is now available that eliminates this CFD shortcoming while significantly reducing computing time.
Combustion turbines are uniquely capable of burning a wide variety of fuels, including very low heat content blast furnace gas, medium heat content landfill and biofuels, and higher heat content liquefied natural gas (LNG) and Fischer-Tropsch fuels. This fuel flexibility can be a headache for engineers who are responsible for designing combustion systems capable of efficiently and reliably burning these fuels while accounting for the reality that regulators continue to ratchet down permissible emission limits. To complicate the task further, each fuel has unique characteristics that impact turbine operation, such as light-off and acceleration, compatibility with dry- and ultra-low-NOx (DLN and ULN) combustion systems, combustion dynamics, and flashback.
Ask a combustion engineer how long a complete analysis of a unique fuel will take, and chances are the answer will be, "It depends." For example, diffusion flame combustors are much more forgiving of fuel changes but are limited in the possible range of NOx reduction. DLN and ULN models can meet low-NOx emission requirements but are very fuel-finicky, especially during start-up and quick load changes. Each analysis must also consider flame-blowout tendency and combustion acoustics under every conceivable operating condition.
The critical path for any thorough analysis of how a combustor will perform on a new fuel is usually a function of the quality of computer tools, the skill of the analyst, and how fast the computer can crunch numbers. A "quick" estimate based solely on the designer’s experience might be possible by tomorrow. The "best" answer may take months of running several computation computer codes to mesh the combustor and use realistic turbulence and detailed chemistry, followed by weeks of data analysis and perhaps even expensive full-scale rig testing.
Turbine manufactures can’t afford the cost and time of finding the "best" answer when dozens of unique fuel inquiries arrive every week. They also can’t bet their business on quick estimates that may prove to be inaccurate and cost millions of dollars to correct. The "right" answer lies somewhere in the middle — it balances the desires of the technical designers to field a reliable power plant and the needs of business managers, who must turn a profit with minimum risk to the corporate treasury. The "right" answer may be defined in different ways, but most will agree that a good corporate electronic knowledge base coupled with the best computational tools are two essentials for deriving a right answer.
Email
Comments
Print
Reuse
