Abstract This paper presents a gas turbine performance model based on streamline curvature analysis for axial flow turbomachinery. Unlike those traditional gas-turbine (GT) simulation programs used in the turbomachinery community, the present program is able to model whole GT performance at the design point and off design points where the performances of a compressor and a turbine are directly predicted using the throughflow (TF) models on the meridional plane. This feature is particularly attractive in managing secondary-air-system (SAS) for a high technology aero or industrial engine where the SAS between a compressor and a turbine is directly connected using the localized pressure and temperature predicted by the TF models, thereby ensuring more accurate predictions of the source and sink pressure driven flows. An additional benefit is the more accurate prediction in mixing between the SAS flow and the primary flow. Combustion was simplified using a relatively simple model where the combustion efficiency is derived from the combustor loading factor while the pressure loss includes the contributions from cold and hot parts. The TF treatment of component performances including a compressor and a turbine makes it feasible to remove the need for pre-generated component maps. Other advantages of this method are that the pressure loss coefficients and deviation angles in each blade row can now be easily adjusted in the TF models to match the CFD prediction or test data, thereby enabling the performance predictions to be more representative and physically meaningful to the real engine. The model can also provide the functionality of the digit-twin of a real gas turbine to simulate a virtual test, healthy operation status and performance of gas turbine. The developed program was used to predict performance at different operating conditions including part-load and part-speed. The validation of the code has been carried out using the test data at the design point (DP) and off-design-point (ODP) conditions of a production engine, indicating that a good agreement is achieved between the predictions and the test data.
Wu et al. (Mon,) studied this question.