Optimization Research On Performance Of Finger Seal Based On Stackelberg Game

Parametric models of finger seal (FS) for finite element analysis (FEA) have been constructed in ANSYS software in this paper. Through analysis, optimization objects for finger seal's hysteresis performance and stiffness performance are proposed. Then Stackelberg Game is introduced into multi-objective optimization of FS and Stackelberg model of FS performance optimization is established. This paper proposes an optimization method for Fs based on leader-follower Stackelberg game. The function mapping models between the optimization objects and the configuration design variables are built by neural networks method fulfilled in matlab software. Using program design langue of Matlab software, the codes of Genetic algorithm based on finite element simulation and BP neural networks is programmed. Then, under the leader-follower Stackelberg game and certain working conditions, this program is used to study the performance optimization of FS. Finally, finite element simulation validates the optimization method and optimization result.The research of this paper reaches the goal of controllable multi-objective optimization of FS based on Stackelberg game. Meanwhile a structure of FS having best performances is obtained, and a new concave Shape-curve of FS is discovered. Furthermore the research also offers a new method for performance optimization of FS.Moreover, as a discussion of the method to multi-objective optimization, the method proposed in this paper also provides a new thought and method for complex structure with multi-variables, especially for those optimization problems in which the optimization object can hardly be explicitly expressed as the function of design variables.