基于神经网络的数控线切割加工状态建模技术研究

Research on modeling of wire electronic discharge machine status-manufacturing based on artificial neural networks

研制了一种数控线切割机床智能状态监测系统,实现了特征信号的实时采集与处理,基于人工神经网络技术建立了数控线切割加工状态模型,提出了运行状态综合劣化度的概念及其量化方法,实现了数控线切割加工状态在线辨识与运行状态在线监控.实验证明,该系统能够快速采集特征信号并进行去噪声处理,所建立的加工状态模型能够正确地识别加工状态,运行状态劣化度实时客观地评价了数控线切割机床的运行状态,从而有效地避免了机床严重故障的发生.

The intelligent status monitoring system for wire electronic discharge machine (WEDM) is developed,the sampling and filtering of the characteristic signals are realized at real-time,the WEDM status recognizing module is developed based on artificial neural networks (ANN),the concept of competitive status severity factor is proposed,the measurement of severity factor is also probed into,and the recognizing and monitoring of the running state are realized at real-time.Experiments show that this system can sample the characteristic signals and filter out the noise,and can recognize the WEDM status correctly,and the competitive status severity factor can evaluate the WEDEM running state objectively with the characteristic signals online,so the occurrence of serious fault on the WEDM is also avoided.