Multi-objective Optimization of a Simulating Specimen Design of a Turbine Rotor

This paper aims to perform an optimization of the simulating specimen to search for the optimal specimen design. A design flow chart was adopted and the basic guidelines of the simulating design for the turbine rotor were introduced. The main principle of simulating specimen design is to ensure that the maximum stress and stress gradient at the notch of the specimen are equivalent to the critical part of the component. Then, the genetic algorithm (GA) optimizer capable of optimizing multi-objective was developed combining the ANSYS APDL code. It was found that the vent hole of the turbine front baffle was the critical part. Stress analysis of each individual generated by the GA optimizer was performed and the maximum stress of the specimen was almost the same as the component, reaching 1024MPa. The stress gradient was also compared and showed good consistence with the component. Afterwards, the LCF test of the specimen was carried out and the specimen encountered failure after 150572 cycles. It is 2.18% and 3.36% lower than the predicted results of the specimen and component respectively. It offers a good reference for the LCF life prediction of the turbine rotor.