基于模糊神经网络的绗缝加工变形量预测

Deformation Prediction of Quilting Based on Fuzzy Neural Network

根据非刚性材料高速加工中材料容易变形的特点,引入模糊神经网络和计算机视觉测量技术,在传统PID控制基础上,增加一个模糊神经网络预测器,对加工轨迹进行预补偿。建立单针绗缝加工的变形量模糊神经网络预测模型,并进行仿真。实验和仿真结果表明:模糊神经网络的误差补偿,在满足精度高的前提下,具有较快的收敛速度,为绗缝加工提供一种较好的可行方法。

According to the characteristics of the deformation of non-rigid materials on high-speed machining,this paper introduces the fuzzy neural networks and computer vision measurement technology.Based on the traditional PID control,a fuzzy neural network predictor is increased for pre-processing of trajectory compensation.A prediction model of fuzzy neural network deformation of the single needle quilting is established and simulated.Experimental and simulation results show that error compensation based on fuzzy neural network has a good real-time and allows fast and accurate automated processing of quilting,which provides a better method of quilting processing.