基于BP神经网络的虚拟轴机床的实时误差补偿控制

Real Time Error Compensation Control Scheme for Virtual Axis Machine Tools Based on BP Neural Network

针对虚拟轴机床位置精度高、钢度大、机械结构简单等特点,提出了基于BP神经网络的实时误差补偿控制策略。应用BP神经网络对虚拟轴机床的运动学反解进行学习,利用训练好的神经网络对虚拟轴机床进行误差补偿;给出三坐标虚拟轴机床误差补偿的仿真结果。表明这种补偿控制方案十分有效,补偿后能较大地减小机床的位置误差。

According to the characteristics of Virtual Axis Machine Tools(VAMT), such as: very good stiffness to weight ratios, high position accuracy, simple mechanical structure etc. a real time error compensation control scheme based on BP neural network is proposed. The inverse kinematic problems for the VAMT is trained, and then error compensation problem of VAMT is solved by using this trained neural network. Finally, a simulation example for error compensation of 3coordinate VAMT is given. The results of simulation show that the control strategy is efficient, and the position error for VAMT can be reduced greatly after compensation.