Experience of multi-objective optimization of axial compressor stage

BACKGROUND: Developing a universal optimization approach can reduce the time needed to improve compressor geometry. Therefore, the issue of implementing this approach when performing optimization tasks is important one. AIMS: The development of an approach and testing of a spatial multi-criteria optimization method for the compressor stage. MATERIALS AND METHODS: The formation of an approach to optimization tasks is based on the experience of both research organizations and the methods used in compressor engineering. To test this approach, the IOSO algorithm is used in conjunction with the AutoGrid5 grid generator and the Ansys CFX solver. RESULTS: At this work, a general approach was developed to formulate a multi-objective optimization problem, which serves as the basis for this entire project. A complete cycle of verification and validation was performed for the mathematical model of the studied object, which was built in the ANSYS CFX system. A method for creating a parametric model of blades and their flow paths is described. Two approaches of the optimization problem are presented: using low-reynolds (SST) and high-reynolds (k-ε) turbulence models, in order to assess the qualitative impact of these models on the results. For the convenience of data processing, a program was written in Python. A complete list of the object functions, optimization parameters, constraints, and assumptions used in the study is provided. In total, six different geometries of the study object were considered. For each variant, a sample analysis was performed in each of the five design sections. The detailed description of these analyses is omitted from this work. Integral characteristics of each proposed variant were constructed. Based on the results of the analysis, the most suitable variant was selected, both in terms of geometry and problem formulation. CONCLUSIONS: The developed approach has been tested. The disadvantages of the used method of setting the multi-objective optimization problem and methods for their solution in subsequent works are noted.