An automated multi-objective optimization approach for aerodynamic 3D compressor blade design

The aerodynamic design of compressor blades, as a subtask of the overall aerodynamic compressor design, is a highly sophisticated and complex process based upon long-term experiences and knowledge of the design engineer as well as the interaction of various design tools developed and matured during the last few decades. Even at the beginning of the 21 st century this blading process is mostly performed manually. Undoubtedly, intuition and the talent of comprehension are human strengths, which nobody can deny but sometimes it is beyond our capacities to cope with complicated dependencies fast enough without ending in suboptimal solutions. By the constantly ascending pressure of competition and rising performance demands on the aero-engine manufacturers, improved blade designs time have to be generated in shortest time. From this point of view, this work intends to address the power and effectiveness of optimization-driven blade design processes including smart parameterization concepts via freeform curves and surfaces for axial and radial smooth blade shapes in order to support engineering design work in optimal solution finding. Firstly, a quasi-3D blading process is introduced and validated, which contains a full 3D blade parameterization approach and an automated work flow for obtaining the aerodynamic blade quality from 2D-CFD flow calculations on different radial blade heights. Secondly, an automated 3D blade design process based upon Q3D optimizations results is demonstrated, in which the optimal 3D blade stacking, parameterized by a radial dependent common shift function in axial and circumferential direction, is considered as optimization goal. Both blading approaches are handled by the application of multi-objective optimization strategies with conflicting design criteria subject to aerodynamic constraints leading to multiple Pareto-optimal solutions from which the design engineer can choose trade-offs for his particular design problem.