论文部分内容阅读
为了精确预测超声无磨料外圆抛光加工的效果,应用神经网络理论建立了超声无磨料外圆抛光加工件表面粗糙度的RBF神经网络模型。根据实际情况,进行了网络结构设计,比较了不同特征参数时网络模型的性能,并对所建模型进行了仿真验证,结果表明,预测数据与实测数据有较好的一致性,而且RBF网络性能稳定,说明可以使用神经网络模型对超声无磨料外圆抛光加工效果进行预测。
In order to accurately predict the effect of ultrasonic non-abrasive cylindrical polishing, a RBF neural network model of surface roughness of ultrasonic non-abrasive cylindrical polished workpiece is established based on neural network theory. According to the actual situation, the network structure is designed, the performance of the network model is compared with different characteristic parameters, and the model is verified by simulation. The results show that the predicted data has good consistency with the measured data, and the performance of RBF network Stable, indicating that the neural network model can be used to predict the effect of ultrasonic non-abrasive cylindrical polishing.