Based On Approximate Model Thermo-mechanic Coupled Multi-objective Optimization In Diesel Engine Piston

Taking into account the incommensurability and contradictory among the weight, themaximum equivalent stress, maximum temperature of the third ring groove, the maximumtemperature gradient of the piston, therefore the paper do the thermo-mechanic coupledmulti-objective optimization to the diesel engine piston. If doing directly, it often encounterslow computational efficiency, no inaccurate simulation results for lacking of the experimentalanalysis and correction, difficult to solve the non-conductivity and the black box by usingtraditional optimization algorithms based on gradient, low optimization efficiency if usingglobal optimization algorithm and so on. So the paper has carried out the parameterizedpretreatment research of diesel engine piston based on the "block", accuratethermal-mechanical coupling research, primary and secondary factors research and MIGAglobal multi-objective optimization based on approximation model.Aiming at the parameterized pretreatment research of diesel engine piston based on the"block", the paper will introduce two concepts that are "body model" and "block model",propose and compare with three new parametric design methods, finally successfully applythe second method to the piston, and reduce the number of grid, and improve the efficiency ofthermal-mechanical coupling analysis and subsequent optimization.Aiming at the accurate thermal-mechanical coupling research, firstly the paper establishthe piston’s finite element model that is used the parametric pretreatment, relying on thetemperature field test’s results and constantly revising the thermal boundary conditions untilsatisfying the requirement, and carry out the static and dynamic thermal-mechanical couplinganalysis of the diesel engine piston, finally choose the typical study points and compare withthe static and transient thermal-mechanical coupling in equivalent stress and deformationaspects, thus getting the former equivalent stress and the deformation that is along the piston’smovement direction is larger than the latter, and illustrate the multi-objective optimizationbased on the latter have greater optimization space than the former. but considering in the process of primary and secondary factor analysis and construct an approximate model basedon the latter, because of low computational efficiency, therefore, the primary and secondaryfactor’s analysis and following MIGA global multi-objective optimization based onapproximate model are based on a static thermal-mechanical coupling analysis.Aiming at the primary and secondary problem’s research between the maximumequivalent stress, maximum temperature gradients, the third ring groove maximumtemperature and piston’s weight, paper relying on ISIGHT software, uses optimal Latinhypercube design, and ensure the accuracy by variance analysis; get the results that themaximum equivalent stress is primarily related to X1and X4, the third ring groove maximumtemperature, the maximum temperature gradient piston and piston weight is mainly related tothe X2, X4by Pareto graph, interaction diagram analysis and the main effect analysis, andpave the way for obtaining the three-dimensional RBF fitting between the main factor and thecorresponding response and solution spatial distribution.Aiming at the MIGA global multi-objective optimization’s study based on approximationmodel, paper on the basis of static thermal-mechanical coupling analysis of the diesel enginepiston, relying ISIGHT software, establish a RBF neural network approximation model, andsuccessfully use the MIGA to achieve the global multi-objective optimization of the dieselengine piston, finally the piston weight decreases4.2%, the third ring groove maximumtemperature decreases10.56%, the piston’s maximum temperature gradient decreases4.01%,although the maximum equivalent stress increase3.05%.