基于CNC实时监测数据驱动方法的钛合金高速铣削刀具寿命预测

摘要

A large number of multi-source heterogeneous data existed in the monitoring data due to the inconsistency of the data structures of the different equipment in the workshops,which were difficult to collect and monitor through a single communication protocol.To solve this problem,a data model of machine-tools,communication frameworks and access strategies were developed to realize multi-source heterogeneous data acquisition based on the OPC-UA and MTConnect herein.Carbide tools were easily damaged and quick-wear in high speed machining(HSM)of titanium alloys,as it was difficult to predict the remaining useful tool life in real time.Therefore,a BP neural network model based PCA method was developed to reflect the relationship among machine conditions and the param-eters of the tools.The model realizes the monitoring and predicting of important informations of tool healthy conditions based on RUL.%针对车间内不同设备的数据结构不一致导致车间监控数据中存在大量的多源异构数据,难以通过单一的通信协议采集与监控的问题,研究了基于 OPC-UA技术与MTConnect协议的刀具、机床的数据模型、通信架构及访问策略,解决了多源异构数据的采集问题.针对硬质合金刀具在高速铣削钛合金工件时磨损较快、刀具剩余寿命的实时预测难度大的问题,建立了一种基于 PCA 前置处理数据的神经网络模型,实现了基于刀具剩余寿命的刀具健康状态信息的实时监测和预测.