Machine Supporting Members Multi-objective Optimization Based On Uncertainty

The precision and ultra-precision machining in modern industry put forward higher requirements on the performance of the machine tool. Machine tool supports'structures have a direct influence on the machining quality whether the structures are reasonable or not. Traditional supports design methods frequently use the empirical approach and similar products analogy approach with poor efficiency, long cycle, high energy consumption and the final poor design effect, it is indispensable to explore a more efficient and effective supports design methods.Various kinds of uncertainty factors exist in the process of machine tool supports manufacturing and using, and the coupling of numerous uncertainty factors will generate a non-ignorable impact on the machining quality. Uncertainty multi-objective optimization methods become the key techniques to solve the engineering problem containing uncertainty factors, while conventional supports design methods cannot effectively handle such problems.In this paper, the machine tool supports (spindle box and column) uncertainty multi-objective optimization design was carried out using three different kinds of algorithms under the consideration of the supports density uncertainty, elasticity modulus uncertainty and cutting loads uncertainty. The optimization results were compared after which the final design scheme was selected. Job description in detail:(1)First, the supports uncertainty multi-objective optimization problem was established. The uncertainty variables were selected out, while the optimization variables were chosen by analyzing the sensitivity of the supports key feature sizes relative to the respective optimization objectives. The corresponding optimization objectives were calculated after the design points were sampled out by using latin hypercube method from the sample space formed by uncertainty variables and optimization variables. The approximation models of uncertainty variables and optimization variables to the optimization objective and constraint function were then established through using support vector machine (SVM) method, after which the approximation models'precisions were verified.(2)The spindle box and column uncertainty multi-objective optimization problems were solved by using nested genetic algorithm, with the cutting point and supports weight as the optimization objectives, the supports first inherent frequency as the constraint.(3)The same spindle box and column uncertainty multi-objective optimization problems were again solved by using nested particle swarm optimization algorithm, with the cutting point and supports weight as the optimization objectives, the supports first inherent frequency as the constraint.(4)Another solving algorithm that nested the genetic algorithm and the particle swarm optimization algorithm, was applied to the same uncertainty multi-objective optimization problems. Finally, the three solving results were compared after which the optimization scheme of the spindle box and column were selected.The results show that:The proposed uncertainty optimization methods can effectively solve the supports structure design issues. Relative to the old scheme, the optimized spindle box cutting point displacement mean value was reduced by 11.87%, the cutting point displacement interval size was decreased by 34.78% and the spindle box weight was reduced by 8.01%; the optimized column cutting point displacement mean value was reduced by 5.56%, the cutting point displacement interval size was decreased by 24.22% and the spindle box weight was reduced by 6.91%.