磨料水射流切割可视化BP神经网络模型研究

针对磨料水射流切割性能与影响因素间存在复杂的非线性关系,无法用传统数学方法建模的问题,基于BP人工神经网络理论,结合典型材料的切割实验结果,在考虑射流压力、磨料流量、切割靶距、工件厚度、磨料喷嘴直径与切割速度6个因素情况下,建立了磨料水射流切割BP神经网络模型.同时,基于Delphi开发出了可移植的磨料水射流切割速度人工神经网络预测单元,实现了所建网络模型的可视化,为实现网络模型与数控系统的集成提供条件.研究结果表明,该网络模型能快速、准确、可靠地预测切割速度,与数控系统相集成可实现对磨料水射流切割质量的有效控制.

Study on Visual BP Neural Network Cutting Model for Abrasive Water Jet

It is difficult to establish a cutting model using traditional mathematical methods for the abrasive water jet because of the complex nonlinear relationship between the cutting performance and the influencing factors. A BP neural network cutting model of abrasive water jet which contains six influencing factors as jet pressure, abrasive flow, cutting target distance, work piece thickness, abrasive nozzle diameter and cutting speed was established based on BP artificial neural network theory and the results of the cutting experiment with typical material. And then, a portable abrasive water jet cutting speed artificial neural network prediction unit was developed to realize the visualization of the network model based on Delphi, and provided conditions for the integration of network model and NC system. The results show that the network model integrated with NC system can predict the cutting speed rapidly, accurately and reliably and realize the effective control of the cutting quality.