Reliability Analysis And Optimization Design Of Blisk Plate Based On Intelligent Distributed Collaborative Response Surface Method

As the key component of the aero-engine,the turbine disk has many failure modes such as high and low cycle fatigue,creep deformation,and their coupling effects in actual working conditions.In order to improve the calculation accuracy and efficiency of multi-failure mode reliability analysis method for turbine disk,the extreme response surface method,distributed cooperative response surface method,generalized regression neural network and multi-population genetic algorithm are combined to propose the intelligent distributed collaborative response surface method for reliability analysis and optimization design of turbine disk.(1)Reliability analysis of low cycle fatigue life of leaf disk based on generalized regression extreme response surface method.In order to study the influence of the thermal-structural coupling on the reliability of the low cycle fatigue life of the turbine disk,the generalized regression neural network method(GRNN)and the extreme response surface method(ERSM)are combined,so as to propose a generalized regression extreme response surface method(GRERSM).Considering the material parameters and working parameters of the leaf disk as the input random variables,the life value of the minimum fatigue life point is taken as the analysis target through thermal solid coupling method.Using the Latin hypercubic sampling technique(LHS)to extract the small batch sample values from the input random variables,the GRERSM mathematical model is established,and the large batch samples are extracted from the input random variables by the Monte Carlo method(MCM),and the sample points are brought into the GRERSM model and analyzed.The reliability and sensitivity of low cycle fatigue life of leaf disk under thermal-structural coupling are obtained,the results show that the reliability of the minimum fatigue life of the disc is P_r=0.99848,when the allowable value of low cycle fatigue life is y*=6000 cycles.(2)Reliability analysis of fatigue-creep coupling damage of disk and blade based on distributed cooperative generalized regression response surface method.In order to study the reliability analysis of multi-objective coupling failure mode,the generalized regression neural network(GRNN)and distributed cooperative response surface method(DCRSM)are combined to propose distributed Collaborative Generalized Regression Response Surface Method(DCGRRSM).Taking the maximum stress point,maximum strain point and minimum fatigue life point of blade-disk as the research object,the fatigue-creep coupling damage of blade-disk is solved by MATLAB,a small batch of samples is obtained by using LHS,and the DCGRRSM mathematical model is established.The fatigue-creep coupling reliability of high pressure turbine blade-disk is obtained by large batch sampling analysis of DCGRRSM with MCM.The results show that when the allowable values of fatigue-creep coupling damage of disk and blade are 0.2272 and 0.2476,respectively,the reliability of fatigue-creep coupling damage is 0.9956.(3)Reliability analysis of radial running clearance of blade tip based on distributed cooperative generalized regression extreme response surface method.In order to study the effect of high temperature creep on the radial running clearance of the tip of the high pressure turbine,the generalized regression extreme response surface method(GRERSM)and the distributed cooperative response surface method(DCRSM)are combined,and the distributed cooperative generalized regression extreme response surface method(DCGRERSM)is proposed.considering the randomness of temperature,rotational speed,material parameters and convection heat transfer coefficient,the maximum radial creep deformation point of disc,blade and casing is taken as the research object,and the mathematical model of DCGRERSM is established,and the reliability analysis of the radial running gap of blade tip is performed by large batch sampling of DCGRERSM model by MCM.The results show that when the steady-state tip clearance is set to?=2.2 mm,the reliability of the radial running clearance of the tip of the high-pressure turbine is0.9909.(4)Reliability optimization design of blade tip radial running gap based on multi-group genetic algorithm.In order to solve the problem of multi-objective collaborative reliability optimization design,the multi-group genetic algorithm(MPGA)is combined with the distributed cooperative generalized regression extremum response surface method(DCGRERSM).In this paper,a Multiple population genetic algorithm-Distributed Collaborative Generalized Regression Extreme Response Surface Method,MPGA-DCGRERSM,is proposed.Firstly,the DCGRERSM mathematical model is established,using MCM to analyze dynamic reliability and sensitivity of DCGRERSM,And the influence degree of each input random variable on the reliability of tip radial running clearance is obtained.the multi-objective collaborative reliability optimization design(MOSRBDO)model is established.Finally,MPGA is used to optimize the MOSRBDO model,and the optimal solution of the optimization objective is obtained.The optimized design results show that the creep deformation of the disc,the blade and the casing is reduced by 0.1258,0.0536,and 0.0710 respectively,and the deformation of the tip radial running clearance is reduced by 0.1084,and the reliability is improved by 0.77%.