Gas Turbine Advanced Performance Simulation

Abstract

Current commercial `state of the art' engine simulation software is of a low fidelity. Individual component performance characteristics are typically represented via nondimensional maps with empirical adjustments for off-design effects. Component nondimensional characteristics are usually obtained through the averaging of experimental readings from rig test analyses carried out under nominal operating conditions. In those cases where actual component characteristics are not available and default maps are used instead, conventional simulation tools can offer a good prediction of the performance of the whole engine close to design point, but can deviate substantially at of design and transient conditions. On the other hand, even when real component characteristics are available, zero-dimensional engine cycle simulation tools can not predict the performance of the engine at other than nominal conditions satisfactorily. Low-fidelity simulation tools are generally incapable of analyzing the performance of individual engine components in detail, or capturing complex physical phenomena such as inlet flow distortion. Although the available computational power has increased exponentially over the last two decades, a detailed, three-dimensional analysis of an entire propulsion system still seems to be so complex and computationally intensive as to remain cost-prohibitive. For this reason, alternative methods of integrating different types and levels of analysis are necessary. The integration of simulation codes that model at different levels of fidelity into a single simulation provides the opportunity to reduce the overall computing resource needed, while retaining the desired level of analysis in specific engine components. The objective of this work was to investigate different simulation strategies for communicating the performance characteristics of an isolated gas turbine engine component, resolved from a detailed, high-fidelity analysis, to an engine system analysis carried out at a lower level of resolution. This would allow component-level, complex physical processes to be captured and analyzed in the context of the whole engine performance, at an affordable computing resource and time. More specifically, this work identified and thoroughly investigated several advanced simulation strategies in terms of their actual implementation and potential, by looking into relative changes in engine performance after integrating into the basic, nondimensional cycle analysis, the performance characteristics of i) two-dimensional Streamline Curvature (SLC) and ii) three-dimensional Computational Fluid Dynamics (CFD), engine component models. In the context of this work, several case studies were carried out, utilising different two-dimensional and three-dimensional component geometries, under different operating conditions, such as different types and extents of compressor inlet pressure distortion and turbine inlet temperature distortion. More importantly, this research effort established the necessary methodology and technology required for a full, twodimensional engine cycle analysis at an affordable computational resource.