The Research On Remote Monitoring Of CNC Machine Tool Vibration And Its Prediction

CNC machine tool is an advanced manufacturing equipment with high degree of automation and complex structure,which is the key equipment in the production.If it failures,it won't work,and then it will cause tremendous losses to the enterprise because of long maintenance time and high maintenance costs.Although CNC machine tool numerical control system built-in can complete simple fault diagnosis for the components which are associated with it,it can't make prediction and diagnosis to the failures in the mechanical system with high rate emergence.If we want to predict mechanical failure to the CNC machine tools,first we should predict the information shown from the diagnosis target(such as vibration,cutting force, power,acoustic emission and cutting temperature,etc.).Gernerally,the failure of the mechanical system is shown from vibration,so the accuracy of the vibration prediction has a great influence to the results of the failure prediction.Therefore,this dissertation has developed a remote monitoring system for CNC machine tool to predict the trend of the vibration.Firstly,the dissertation makes an analysis on the nowadays' research situation at home and abroad,and then organizes the structure of remote monitoring for the CNC machine tools and analycises all levels to determine the content of the main study;after that proposes the design goal of remote monitoring system for CNC machine tool, developes the system design process according to the designed goal and makes an analysis on the key technologies.Secondly,based on the designed goal of remote monitoring,we will choose appropriate hardware according to the system design process to design the hardware systems;and make a development of data acquisition modules and remote monitoring modules using the LabVIEW software development platform;then complete the development of the prototype system by the means of integrating of various hardware and software systems.Thirdly,it will introduce the BP neural network for trends prediction and improve the BP neural network by the use of Levenberg-Marquardt algorithm based on the deep study on the BP neural network for its shortcomings,finally build the vibration trend prediction model which is based on the improved BP neural network by applying time-scrolling method. Last,the dissertation carries on the test to the CNC machine tool remote monitoring prototype system and predictes the vibration tendency of CNC machine tool,carries on the real-time online long-distance monitoring to tool holder and headstock vibrational state,and makes nd the feature extraction after colleceting the vibration signals.The vibration characteristic values are put into the improved BP neural network model and standard BP network model.Compareing with the standard BP network model,the improved BP neural network model has higher convergence speed and forecast precision.The system will predict the vibration curves of tool holder in the future,which can provides a basis for the predictive maintenance of CNC machine tool in the next step.The system operates normally in testing.It can meet the demand of remote monitoring of CNC machine tool vibration and its prediction,with a friendly interface, high-speed stability processing unit,powerful data analysis functionality,easy maintenance and good scalability,high reliability and stability,etc.characteristics.So it has a promising application foreground.This system provides basis for predictive maintenance of CNC machine tool,and can make up for the shortage of numerical control system whose self-diagnosis function can't diagnose mechanical fault.