PMSM伺服控制系统的改进单神经元PID控制

摘要

Permanent magnet synchronous motor (PMSM) AC servo system is a kind of complicated nonlinear,unknown system,containing unknown complicated interference.The quality of PID control is affected by the interference,and cannot meet the special demand of high accuracy servo system.An intelligent adaptive PID controller combined with traditional PID control and single neuron was designed by using the self learning and adaptive ability of neural network,and was improved on the online learning rule of single neuron adaptive parameters.Simulation result shows that the servo system with the single neuron PID control has simple structure and environmental adaptability.It can suppress interference,improve control accuracy,has strong robustness,and obtain ideal control effect.%永磁同步电动机(PMSM)交流伺服系统中含有未知复杂干扰,属于一类复杂的非线性、不确知系统,会影响PID控制算法的控制质量,无法满足伺服系统的高精度指标要求.利用神经网络对未知信息数据的自学习和自适应能力,设计出一种结合传统PID控制与单神经元自适应智能PID控制器,并针对单神经元自适应PID参数的在线学习规则进行了改进.仿真结果表明,应用了单神经元PID控制的伺服系统,不但结构简单,而且能适应环境变化,有效抑制干扰,控制精度明显改善,有较强的鲁棒性,达到了理想的控制效果.