Research And Implementation Of On-line Monitoring Method For Vibration State Of Machine Tool Spindle

In response to the "Made in China 2025" plan and to make China move towards the future development of intelligent manufacturing,CNC machine tools,as the basic equipment for "Made in China" machinery manufacturing,are undergoing comprehensive information upgrades to monitor the information of various machine tool components in the manufacturing process.As the key component of the machine tool,the spindle's health determines its performance.Spindle vibration is an important influencing factor of the spindle health status.Therefore,in order to improve the performance of the machine tool manufacturing process,it is necessary to perceive the spindle vibration information and carry out online monitoring of the spindle vibration.During the manufacturing process,the spindle of the CNC machine tool is under the multi-tool and multi-process conditions.The spindle vibration has different vibration states and the amount of long-term monitoring data is large.Therefore,to achieve the purpose of spindle vibration state monitoring,it is necessary to solve the original vibration Problems such as large data volume,many vibration states,and lack of abnormal vibration data samples.Aiming at the above problems,this paper studies from the vibration information feature extraction method,spindle vibration state monitoring method,machine tool spindle vibration signal acquisition and vibration state monitoring system implementation,and realizes that a large number of spindle vibrations generated during machine tool work are mapped to spindle vibration health information The main contents are as follows:(1)In this paper,the method of extracting the characteristics of the vibration signal of the machine tool spindle is studied to solve the problem of the large amount of raw data and large information content of the machine tool spindle vibration,which is not conducive to long-term online monitoring.According to the vibration characteristics of the main shaft,the FFT frequency domain transformation and trend term elimination processing of the vibration signal are combined with the spectral analysis and statistical value analysis to extract the local peak-center frequency spectrum parameter of the main shaft vibration signal,the total energy of the sub-band of the spectral segment-centroid frequency spectrum parameter and Statistical characteristic parameters,and the effectiveness of the feature extraction method is verified by the vibration simulation signal of the spindle shell.(2)In this paper,an online vibration state monitoring method based on the combination of probabilistic neural network(PNN)and similarity is studied,which solves the problem of spindles with many vibration states and lack of abnormal vibration data samples.In this paper,combined with the characteristics of the vibration state of the machine tool spindle,the structure of the vibration state model data collection is analyzed,and the model data collection is normalized and the abnormal data processing based on the DBSCAN clustering algorithm.The dispersion evaluation of vibration characteristics based on standard deviation is used,and the evaluation of the difference in state of vibration based on Euclidean distance can be defined.After evaluating the model data set combined with multiple PNN model structure,the multiple PNN vibration state model is constructed to perform state recognition and similarity calculation.Finally,the spindle shell vibration simulation signal verifies the correctness of the online monitoring method of vibration state based on multiple PNN-similarity(3)Developed a spindle vibration signal acquisition and condition monitoring system.Based on STM32,the vibration acquisition system hardware was built,and programs such as AD acquisition,SPI communication and data transmission based on WIFI communication were developed;based on the Qt platform,the online monitoring system software for vibration state was developed to realize feature testing,state construction,state expansion Four working modes of condition monitoring.Finally,the effectiveness of the system for vibration state monitoring was tested.