Axial Compressor Loss Sensitivity Analysis to Blade Row Design Parameters

Abstract In the preliminary design phase of an axial compressor stage, the selection of the most relevant design parameters within the shortest amount of time is paramount. An easier convergence toward the final design limits the use of costly optimization loops and extensive computational fluid dynamics simulations. By coupling a previously published physics-based loss decomposition tool, a mean line analysis, and radial equilibrium codes, the current study investigates the influence of design variables on blade row total pressure loss contributions. The core of the work presents a strategy to exploit the state-of-the-art design space. The compressor stage loss sensitivity to the design variables selected is evaluated using the probabilistic method of Sobol indices, quantifying parameters non-linear interactions inherent to the design process. As an example, the study synthesizes the relationship between loading, solidity, and row inlet boundary layer displacement thickness, to profile loss, end wall loss, and tip leakage loss. As a result, this work ranks the relative influence of the selected design parameters in the different loss sources, a process that may not be obvious from a term-to-term analysis of the loss models used.