Structure Analysis And Laminate Optimization For Passenger Vehicle Composite Subframe

In order to achieve Energy-Saving and Emission Reduction for vehicle,we designed a vehicle rear subframe using Carbon Fiber Reinforced Plastic(CFRP),and optimized laminate structure.However,the major problems of optimization are the non-linear and high-dimension,and difficult to establish a surrogate model based on design of experiment(Do E)due to the non-rectangular feasible region of design variables.Therefore,to reduce the influence of non-linear and high-dimension in optimization model,the optimization process was divided into two stages including thickness optimization and sequence optimization;to reduce the calculation cost,we established a surrogate model for high-dimensional problem using the Gaussian Process Regression(GPR),the clustering-based Do E and the adaptive sampling method.The main research contents are as follows:Firstly,we summarized the structural design requirements of laminate composites,then designed the CFRP subframe with reference to the steel subframe,and determined the initial laminate structure for the‘super-ply’scheme.Moreover,we established a finite element model of subframe of composite and steel respectively by ABAQUS.To obtain the hard point load of the subframe in five typical conditions,we calculated the tire ground force for five typical conditions,and inputted the load into multi-body dynamic model of the rear suspension established by ADAMS for simulation.Then,we obtained the safety factor,strain energy of each condition and free modal frequency by finite element analysis.The results show that the performance of the initial laminate structure composite subframe meets the design requirements.Next,to maximize the first-order free modal frequency,minimize the sum of the strain energy of the five conditions,and minimize the mass,we optimized the layer thickness of each orientation for the initial laminate with a total of 16 design variables,and constraint is the proportion of each orientation is not less than 10%.The optimization process was based on the surrogate model established by the GPR,and the clustering-based Do E was used to select the training sample points in non-rectangular feasible region,then verified and updated surrogate model with adaptive sampling method.Finally,the correlation factor R~2 is above 0.9,and maximum error is within10%.Using the NSGA-II,we obtained Pareto’s frontier of three optimization objective successfully.Furthermore,we selected a compromise optimal solution from Pareto’s frontier,and rounded the result to a multiple of 0.1mm of the manufacturable layer thickness.After the optimization,the mass of the composite saving is 8.7%,the first-order modal frequency is increased by 5.4%,and a reduction of 10.5%for total strain energy.Finally,to further improve the total strain energy and the first three orders free modal frequency,we optimized the layup sequence based on the use of bending stiffness parameter.For finite element analysis,the corresponding layup sequence of bending stiffness parameter was obtained by genetic algorithm.The optimization process is based on the surrogate model with the correlation factor R~2 is above 0.95,and the compromise programming methods was used to weight the four objectives into a single objective function.Finally,using the Particle Swarm Optimization to solve the problem,the first three order modal frequencies are significantly increased,and the total strain energy is reduced by 6.5%again.Compared with the steel subframe,the mass is reduced by 53.7%and most performance improvements are obvious.The application of composite materials on the subframe has achieved significant lightweight with better performance.