Research On Health Evaluation Technology Of Key Components Of Machining Center Based On Random Forest Algorithm

As one of the most widely used equipment in machinery manufacturing,detecting and evaluating the health status of the machining center timely,effectively carring out predictive maintenance of the equipment effectively,and preventing the machining center from malfunctioning and downtime are an urgent problem that machinery manufacturing enterprises need to solve.This paper has conducted systematic and in-depth research around the extraction of health features of key components of the machining center,the applicability of modeling methods,health evaluation methods,evaluation system development and verification.The specific work completed is as follows:First,this paper uses the command domain analysis method to analyze the process of given tasks performed by the machining center,and then proposes a methon of health evaluation based on fixed actions.Then the performance requirements and kinematic characteristics of key components of the main kinematic system,feed system and automatic tool change system of the machining center were analyzed separately,and the fixed physical examination program was designed combined.The command domain features for performance evaluation were extracted to provide technical basis for research evaluation algorithm.Second,this paper divides the health status of each component based on the ability to execute instructions of the moving parts and the trend of component.Different data-driven algorithms are studied to realize the health evaluation process.A performance degradation experiment for the feed system simulation was designed.The extracted command domain features were used as input to compare the performance and performance requirements of each algorithm.Finally,the random forest and similarity analysis algorithms are selected to perform health evaluation in the upper and lower computer environments.Then,based on the previous research,this article designed and developed the health evaluation software for key components of the machining center.The software is integrated in the CNC system and can complete the tasks of physical examination program generation,data sampling,health assessment,and abnormal state recording.Through custom configuration,memory recycling,evaluation visualization and other designs,the software can finally be applied to the health assessment of key components of any type of machining center.Finally,this paper designs a physical examination program for a five-axis machining center machine tool group of a user company and carries out long-term operation tracking,and establishes a random forest evaluation model of this type of machine tool by collecting data,and obtains a 90.74% accuracy rate with axis Z as an example.Then,the actual cases evaluated as "degraded" were investigated on the spot,and it was found that the Z-axis assembly accuracy declined,which proved the feasibility of the research content of this article to evaluate the health status of key components of the machining center in the actual industrial environment.