Research On Structural Reliability Design Optimization Based On Intelligent Response Surface Method

In order to improve the accuracy and efficiency of reliability analysis and optimization design, the dynamic characteristics in complex mechanical operation and the highly nonlinear characteristics of limit state function are thoroughly discussed in this paper, and the intelligent response surface method of mechanical reliability analysis is proposed by integrating the intelligent algorithm, such as neural network algorithm and particle swarm algorithm, with high nonlinear mapping ability and response surface method with simplified calculation ability. The basic idea of the method is establish an extremum or multiple response surface model by using the neural network theory and particle swarm algorithm; Then the linkage sampling on response surface model are completed by using Monte Carlo method, and the mechanical reliability could be carried out. Based on the reliability analysis, this paper proposes a new optimization method: particle swarm-intelligent response surface method, which is combining the particle swarm optimization algorithm and the intelligent response surface method. In this method, the sensitivity of random variables are calculated firstly, and then regarding the high sensitivity random variables as design variables, reliability and other constraints as constraint conditions, the reliability-based optimization design mathematical model is established. Finally, the mechanical reliability optimization design are commendably completed by using the intelligent response surface method to calculate mechanical reliability and the particle swarm optimization algorithm to solve the reliability-based optimization mathematical model. And the engineering simulation calculation examples are given:According to the time varying characteristics of flexible manipulator, the dynamic reliability analysis is completed by using the intelligent extreme response surface method considering time variation. Based on the reliability analysis, using intelligent extreme response surface calculates the flexible manipulator reliability, and particle swarm optimization searches optimal size to reduce section area. This method can reduce the use of materials to increase the economic benefit. At the same time, the reliability requirements of mechanical system are met.According to the multiple field coupling characteristics of aeroengine blisk structure, the multiple response surface method is proposed to complete the multiple failure mode structure reliability analysis. On the basis of reliability analysis, using multiple response surface method to calculate the failure mode reliability, multiple objective particle swarm optimization algorithm to search the optimal design point sets to reduce the turbine blisk max imum radial deformation and maximum stress. This method can reduce the structural load size and improve the safety performance of the structure. At the same time, the reliability requirements of mechanical structure are met.