Design and Optimization of Tandem Arranged Cascade in a Transonic Compressor

This study proposed a design and optimization strategy for a tandem arranged cascade using the Non-dominated Sorting Genetic Algorithm (NSGA) II multi-objective optimization algorithm and Back Propagation (BP) neural network technology. The NASA Stage 35 was employed as the initial bench mark in the present study and five geometric control parameters were working as the optimization parameters aiming to enhance the aerodynamic performance in terms of total pressure rise and efficiency. Results showed that the feasibility and capability of the proposed optimization strategy was successfully examined. In view of the fact that the initial tandem cascade (directly scaling down from NASA Stage 35) cannot guarantee the aerodynamic performance, first optimization trial was conducted to optimize the initial design. Results showed that the optimum can improve the flow quality whereas the separation on the blade is decayed or even eliminated particularly at the tip and root regions. However, compared with the initial tandem design, the enhancement in total pressure ratio (0.47%) and efficiency (1%) are too small to be noticed. Second investigation was particularly emphasizing on a high turning tandem compressor with an increment by 28°. The pressure rise and efficiency were augmented by 1.44% and 2.34% (compared to the initial tandem design), respectively. An important conclusion can be drawn that the optimization strategy is worthy to be used in high turning compressors with a considerable performance improvement.