Prediction of Engine Performance Under Compressor Inlet Flow Distortion Using Streamline Curvature

Abstract

Traditionally, engine performance has been simulated based on nondimensional maps for compressors and turbines. Component characteristic maps assume by default a given state of inlet conditions that cannot be easily altered in order to simulate two- or three-dimensional flow phenomena. Inlet flow distortion, for example, is usually simulated by applying empirical correction factors and modifiers to default component characteristics. Alternatively, the parallel compressor theory may be applied. The accuracy of the above methods has been rather questionable over the years since they are unable to capture in sufficient fidelity component-level, complex physical processes and analyze them in the context of the whole engine performance. The technique described in this paper integrates a zero-dimensional (nondimensional) gas turbine modelling and performance simulation system and a two-dimensional, streamline curvature compressor software. The two-dimensional compressor software can fully define the characteristics of any compressor at several operating conditions and is subsequently used in the zero-dimensional cycle analysis to provide a more accurate, physics-based estimate of compressor performance under clean and distorted inlet conditions, replacing the default compressor maps. The high-fidelity, two-dimensional compressor component communicates with the lower fidelity cycle via a fully automatic and iterative process for the determination of the correct operating point. This manuscript firstly gives a brief overview of the development, validation, and integration of the two-dimensional, streamline curvature compressor software with the low-fidelity cycle code. It also discusses the relative changes in the performance of a two-stage, experimental compressor with different types of radial pressure distortion obtained by running the two-dimensional streamline curvature compressor software independently. Moreover, the performance of a notional engine model, utilizing the coupled, two-dimensional compressor, under distorted conditions is discussed in detail and compared against the engine performance under clean conditions. In the cases examined, the analysis carried out by this study demonstrated relative changes in the simulated engine performance larger than 1%. This analysis proves the potential of the simulation strategy presented in this paper to investigate relevant physical processes occurring in an engine component in more detail, and to assess the effects of various isolated flow phenomena on overall engine performance in a timely and affordable manner. Moreover, in contrast to commercial computational fluid dynamics tools, this simulation strategy allows in-house empiricism and expertise to be incorporated in the flow-field calculations in the form of deviation and loss models.