Multifield coupling axial flow turbine performance prediction model and multi-objective optimization design method

Abstract

The simulation of thermal–fluid–solid coupling in turbines is critically important for design optimization. Historically, most research on thermal–fluid–solid coupling has been conducted in three-dimensional, often with computational speeds that do not meet practical expectations. This study proposes a one-dimensional performance prediction and multi-objective optimization design methodology for turbines, integrating aerothermodynamics and structural strength, to facilitate rapid multidisciplinary coupling design optimization at a low-dimensional level. Initially, a multidisciplinary coupled turbine performance prediction model is established, incorporating the combined effects of turbine aerothermodynamics and structural mechanics. This model links the thermodynamics of the blade passage with the stress and strain of the blade. The predictive accuracy of this model is validated against experimental data from a four-stage axial flow turbine, showing a maximum error of 1.56% for the total temperature ratio and 1.69% for the total expansion ratio. Subsequently, using blade parameters, degree of reaction, load coefficient, and flow coefficient as optimization variables and targeting the turbine's overall isentropic efficiency and power as optimization objectives, a rapid Non-dominated Sorting Genetic Algorithm II and the Technique for Order Preference by Similarity to an Ideal Solution are employed to optimize the last stage of the four-stage axial flow turbine. The optimized turbine demonstrates an increase in overall isentropic efficiency by 1.333% and an increase in overall power by 3.329%, while satisfying structural strength requirements. The novelty of this study lies in its rapid optimization design and performance prediction method for the coupled aerothermodynamics and structural mechanics at a one-dimensional level.