基于自适应粒子群优化BP神经网络的高速电主轴温度预测模型

Temperature prediction model for a high-speed motorized spindle based on back-propagation neural network optimized by adaptive particle swarm optimization

为精确地预测电主轴高速运转时内部温升的变化情况,提出了一种基于自适应粒子群优化BP神经网络的电主轴温度预测模型(APSO-BPNN).该模型在PSO-BPNN算法的基础上,引入自适应惯性权重,使权重跟随粒子适应度的变化而变化;采用自适应学习因子,在算法的初期和后期获得不同的搜索能力;融入变异算子,增加种群的多样性,避免算法的早熟收敛等缺点.然后,分别采用BPNN、PSO-BPNN和APSO-BPNN预测模型对电主轴不同测温点的温度进行预测.实验结果表明,与传统的BP神经网络和PSO-BPNN预测方法相比,所提APSO-BPNN模型预测精度最高,鲁棒性最强,可为电主轴及机床温升的智能控制和早期预警系统开发提供理论依据.

To predict the temperature of a motorized spindle more accurately,a novel temperature prediction model based on the back-propagation neural network optimized by adaptive particle swarm optimization(APSO-BPNN)is proposed.First,on the basis of the PSO-BPNN algorithm,the adaptive inertia weight is introduced to make the weight change with the fitness of the particle,the adaptive learning factor is used to obtain different search abilities in the early and later stages of the algorithm,the mutation operator is incorporated to increase the diversity of the population and avoid premature convergence,and the APSO-BPNN model is constructed.Then,the temperature of different measurement points of the motorized spindle is forecasted by the BPNN,PSO-BPNN,and APSO-BPNN models.The experimental results demonstrate that the APSO-BPNN model has a significant advantage over the other two methods regarding prediction precision and robustness.The presented algorithm can provide a theoretical basis for intelligently controlling temperature and developing an early warning system for high-speed motorized spindles and machine tools.