Application And Research Of Kriging Model In Reliability And Reliability-Based Design Optimization

The definition of reliability is:The ability of structure to complete the required functions under the prescribed time and conditions.Reliability-Based Design Optimization is the optimization design of the product volume,quality,stiffness and other parameters under the premise of ensuring the products meet the safety indicators.Although the domestic and foreign actively carry out the practical engineering application research in this aspect,there are still a lot of problems of Reliability-Based Design Optimization,such as the process of reliability analysis is complex,the sampling efficiency of Reliability-Based Design Optimization is low.This paper focuses on the above issues to carry out in-depth and systematic research,in order to make the reliability method and Reliability-Based Design Optimization method are more effective in the design of complex mechanical products.In the paper,firstly,in light of the situation of the traditional reliability method is limited by the design variables,the limit state function and the limit state function as implicit function,we propose a reliability solution method based on Kriging model,and in light of the situation that the low sampling efficiency of Reliability-Based Design Optimization,we put this reliability solution method based on Kriging model into it.The main contents are as follows:(1)The reliability method based on Kriging model.In light of the situation of the traditional reliability method is limited by the design variables,the limit state function and the limit state function as implicit function,we propose a reliability solution method based on Kriging model,by using a learning function H based on Kriging model predictive value,and combined with Monte Carlo sampling,and find out the best sampling point in the design variable region which is corresponding to the sample point with the maximum value of the learning function H,and then the efficiency of calculating reliability of Kriging model is improved.Finally,the proposed method is compared with the response surface method based on importance sampling method and the reliability method based on Kriging model and learning function U and EFF.(2)The Reliability-Based Design Optimization method based on Kriging model.In light of the situation that the low sampling efficiency of Reliability-Based Design Optimization,we put the reliability solution method based on Kriging model and learning function into the process of reliability analysis.Finally,the correctness and efficiency of the proposed method are verified by an example of the optimization design of simple supported beam and Reliability-Based Design Optimization of the gear.