Multi-objective Optimal Design for Carrier and Torque Frame Assembled Together in Planetary Gear Train of GTF by Integrated DOE, RSM and LSGRG

An approach of multi-objective optimal design for carrier and torque frame assembled together in planetary gear system of Geared Turbo Fan (GTF) is proposed in this paper. First, a program package is developed as parameterized model in Python language, the main function of which is to achieve automatic invoking of complete FEA during optimization. Therefore, development cycle of the product is reduced greatly. Second, significant factors are selected based on design of experiment (DOE) respectively for maximum von Mises stress of the frame and maximum displacement of the holes on the carrier. Third, sampling values of the two responses are obtained by central composite face-centered (CCF) design. Next, approximate quadratic polynomials are found through response surface method (RSM) to express the two objective functions explicitly. Finally, the Pareto efficient solution of three optimization objectives (total mass and the two) is acquired via large scale generalized reduced gradient (LSGRG) algorithm. The simulation results show that the relative error of each response is less than 5%, which indicates that the model proposed is quite correct. Furthermore, the comparison also reveals that the approach integrated DOE, RSM and LSGRG to determine the most favorable configuration of the assembly is appropriate and effective.