单神经元自学习PID在PMSM的S模块上的应用

Single nerve cell self-study PID application on the smodel of the PMSM

S曲线加减速是高档CNC系统中的一项重要功能,结合多层人工神经网,利用神经元的自学习功能及其控制灵敏度的快速近似求取方法,以神经元控制器作为伺服系统的速度调节器,与S曲线加减速算法构成交流伺服系统。所给出的曲线加减速算法克服了传统加减速算法中的缺点,速度在变化过程中十分平滑,是一种适合于高速切削的柔性加减速算法。利用神经网络的在线学习分析电机起、停的速度,以及硬件平台的传动距离。仿真与实验表明,神经元控制器具有较强的自适应性和较好的实时性。

The s-shape acceleration/deceleration（acc/dec） is an important function in advanced CNC machine tools. This paper presents a novel speed control system for servo motor. The neuron controller with its self-learning ability and its approximate solution algorithm of control sensitivity can realize the online learning algorithm. The neuron controllers as the speed controller and the s-shape acc/dec as the current controller together form servo system. Through simulation and experiments, it shows that the neuron controller has the strong adaptability and better online performance. Doing reliably and feasibility operated on the concrete hardware platform through the artificial verification Model＇s curve of the software finally.