Model Based Gas Turbine Parameter Corrections
Ambient conditions have a significant impact on the temperatures and pressures in the flow path and on the fuel flow of any gas turbine. Making observed data comparable requires a correction of the raw data to sea level Standard Day conditions. The most widely applied gas turbine parameter correction method is based on keeping some dimensionless Mach number similarity parameters invariant. These similarity parameters are composed of the quantity to be corrected multiplied by temperature to the power ‘a’ and pressure to the power ‘b’ with exponent ‘a’ being theoretically either 0, +0.5 or −0.5 and ‘b’ either 0 or 1.0. To improve the accuracy of this approach it is common practice to empirically adapt the temperature and pressure exponents ‘a’ and ‘b’ in such a way that the correction process leads to a better correlation of the data. Finding empirical exponents requires either many consistently measured data that cover a wide range of ambient temperatures and pressures or a computer model of the engine. A high fidelity model is especially well suited for creating optimally matched exponents and for exploring the phenomena that make these exponents deviate from their theoretical value. This paper discusses the questions that arise when creating empirical exponents with a thermodynamic model of the gas turbine. The gas turbine parameter correction method based on Mach number similarity parameters can get complex if effects like humidity, bleed air or power off-take, free power turbines, switching between various fuel types (Diesel and natural gas), water respectively steam injection, variable geometry or afterburners have to be considered. In such a case it might be simpler — and certainly more accurate — to use the thermodynamic model for the gas turbine parameter correction. Computing power required for running a model is nowadays of no relevance and the better consistency of the data available for engine performance monitoring can yield a significantly improved performance diagnostic capability.