基于生物地理学优化算法的高速磨削电主轴定子电阻辨识  

Biogeography Based Optimization Algorithm of High Speed Griding Motorized Spindle Stator Resistance Identification

作　　者：吴玉厚 张云龙 张丽秀 WU Yuhou ZHANG Yunlong ZHANG Lixiu(School of Mechanical Engineering, Shenyang Jianzhu University, Shenyang, China, 110168)

机构地区：[1]沈阳建筑大学机械工程学院,辽宁沈阳110168

出　　处：《沈阳建筑大学学报（自然科学版）》2017年第5期898-905,共8页Journal of Shenyang Jianzhu University：Natural Science

基　　金：国家自然科学基金项目(51375317;51675353);辽宁省高效创新团队支持计划项目(LT2014011)

摘　　要：目的减小电主轴在运行过程中定子电阻变化导致的电主轴转矩脉动,进而提高高速磨削机床加工精度.方法基于直接转矩控制系统下的高速磨削电主轴温度检测试验数据,将BP神经元网络算法与随机扰动的生物地理学优化算法(MLBBO)相结合,利用MLBBO算法对传统的BP神经网络权值和阈值进行优化,建立基于MLBBO-BP的定子电阻辨识模型,并利用MATLAB进行仿真.结果利用MLBBO-BP模型方法对直接转矩控制系统下的电主轴定子电阻辨识精度可达±0.3%,模型辨识能力较强.和传统的BP神经网络辨识定子电阻方法相比,精度更高.结论利用MLBBO-BP方法可以有效地辨识定子电阻,辨识精度较高.In order to reduce the motorized spindle torque ripple caused by the changing of stator resistance in the process of running,which can increase the machining accurracy of high speed griding machine tool. Based on the motorized spindle temperature test data, the neural network algorithm are combined with random perturbation of biogeography（ MLBBO）,which is used to optimize the traditional BP neural network weights and thresholds and established the model of stator resistance identification based on MLBBO-BP,f inally simulated on Matlab. The simulation results show that the accuracy of stator resistance identification using MLBBO-BP model can reach 0. 3%, the capability of model identification is strong,which has higher precision than traditional BP neural network algorithm.

关 键 词：DTC MLBBO BP神经网络 电主轴 定子电阻辨识 

分 类 号：TH133.2[机械工程—机械制造及自动化] TP183[自动化与计算机技术—控制理论与控制工程]