基于BP与GA对非标机床结构对比研究

Comparative Study of Non-standard Machine Tool Structure Based on BP and GA

针对多工位非标机床既节约成本又运行稳定困难的问题,非常有必要对非标机床进行优化设计研究。以非标机床—齿轮淬火自动上料机为载体,提取非标机床关键部件进行结构设计并对此建立多目标数学模型,运用神经网络与遗传算法对关键部件优化分析,最终通过ANSYS Workbench对两种优化结果进行理论对比以及仿真分析。其结果表明:在本案例中,通过遗传算法优化的数据结果比神经网络优化出的数据更合理,其总体质量相差5.1%;同时所优化结果中最大位移变形值分别为0.00017968mm和1.2249mm;最大应力分别为0.6863MPa和0.057739MPa;前6阶模态值在4Hz至50Hz以内,同时为以后类似研究提供基础。

Aiming at the problem that multi-station non-standard machine tool not only saves cost but also runs stably, it is very necessary to carry on the optimization design research to the non-standard machine tool. Taking the non-standard machine tool-gear quenching automatic feeder as the carrier, the key parts of the non-standard machine tool are extracted for structural design and multi-objective mathematical model is established. The key parts are optimized and analyzed by neural network and genetic algorithm. Finally, the theoretical comparison and simulation analysis of the two optimization results are carried out by ANSYS Workbench. The results show that: in this case, the data optimized by genetic algorithm is more reasonable than the data optimized by neural network, and the overall quality difference is 5.1%;At that same time, the maximum displacement and deformation values in the optimized results are 0.00017968 mm and 1.2249 mm respectively. The maximum stresses were 0.6863 MPa and 0.057739 MPa respectively. The first six modes are in the range of 4 Hz to 50 Hz, which provides a basis for similar research in the future.