Analysis And Optimization Of Mining Dump Truck’s Cab Based On High Dimensional Model Representation

Tumble and fall accidents often happen to mine dump trucks, due to drivingin bad and complex off-road conditions and the decline of vehicle stability whenthe truck is running. Therefore, it is necess ary to optimize the cab of mine dumptruck through ROPS&FOPS analysis in order to improve the safety of the driver.In nested optimization problems that are based on real models, calculatingthe objective function values needs to call the time-consuming simulation model.However, using the approximate model can solve the problem effectively.Currently, some widely-used approximate methods can effectively solve theproblem of low dimension and low level of nonlinearity. However, in terms ofcomplex nonlinear finite element model, the number of sample points and thecost time of calculation building approximation models increase exponentially,because of with the increase of dimension. This results in the decrease ofcomputational efficiency. High dimension mo del, on the other hand, can changethe exponentially growing computation cost to polynomial. Meanwhile, itunveils the contribution to approximate function of every design variables anddecreases the time of building the approximate model. Moreover, it rev eals thelinear or nonlinear relationship and their correlation of every input variable,thus solving high dimension nonlinear problems effectively. In this paper, twoapproximation model methods of BPNN-HDMR and TPS-HDMR areproposed and comparative analysis of several high dimensional model methodsare carried out in term of precision and efficiency.In the manufacturing process, there are inevitable errors or uncertainties inmaterial properties, boundary conditions and measuring errors. All theseuncertainties are of small values, but they may contribute to a large deviationwhen coupling together. Therefore, it has important theoretical and practicalvalues when optimizing the cab considering the parameters uncertainties.Considering the uncertainty of the cab parameters, the cab is carried outusing multi-objective optimization in this paper. Firstly, the uncertain materialparameters of the cab are treated as intervals and an approximate model betweendesign variables, uncertain variables and objec tive function was established bythe High Dimensional Model Representation based on Thin Plate Spline Interpolation (TPS-HDMR). Then, by using double-nested GA, the deformationand the weight of the ROPS are optimized in this paper. Specifically, the innerlayer adopts IP-GA to solve the interval of the objective function in uncertaindomain, while the outer layer implements elitist preserve strategy and removingduplicates individual non-dominate sorting genetic algorithm (NSGA-II).Finally, the optimization results verify the feasibility and the superiority of themethod proposed in this paper.