Research On Fault Diagnosis Method Of CNC Tool Holder End-toothed Disc Locking System Based On Vibration Signal

As the "industrial mother machine" of the equipment manufacturing industry,CNC machine tools are not only the cornerstone of solving the major strategic needs of the country,but also the foundation to ensure industrial development.However,compared with the internationally advanced products of the same type,the manufacturing technology and reliability level of domestic CNC machine tools are still far behind,especially those equipped with domestic functional parts.As the key functional parts of CNC lathe,the technical level of CNC tool holder directly affects the comprehensive performance of CNC lathes.The end-toothed disc is the core structure to maintain the repeat positioning accuracy of CNC tool holder,and it is important to improve the reliability of the end-toothed disc locking system which is composed with the hydraulic system to enhance the reliability level of CNC tool holder.Fault diagnosis technology is an important means to improve the reliability of machinery and equipment.Timely detection of fault symptoms and troubleshooting can effectively prevent the occurrence of major accidents,and also extend the service life of products and reduce the operating cost of equipment to a certain extent.Therefore,in this paper,with the goal of improving the reliability of CNC tool holders,a fault diagnosis method study is carried out for the end-toothed disc locking system in CNC tool holders.The main research contents are as follows:(1)For the problem that there are many failure modes of the end-toothed disc locking system and the causes of each failure are still unclear,firstly,the structural composition of the end-toothed disc locking system is introduced and its working principle is analyzed.Secondly,according to the failure data obtained from on-site field tracking test and laboratory bench test,the typical failure modes of the end-toothed disc locking system are counted.Then the causes of each typical fault are analyzed in detail,and the corresponding early faults of each fault mode are summarized.Finally,the diagnostic performance of the commonly used monitoring signals of the tool holder is compared to determine the types of signals to be collected for the tool holder fault test.(2)In view of the problem that the vibration signal of end-toothed disc locking system has a lot of invalid information,the vibration signal endpoint detection method based on adaptive threshold judgment is proposed.Firstly,we analyze the association of each signal with the end-toothed disc locking system,and determine the signal extraction link: toothed disk locking and loading link.Secondly,to identify the signal endpoints of each link,an endpoint detection method based on adaptive threshold decision is proposed by improving the traditional double threshold method.Then we apply the variable scale double threshold method to change the detection parameters to solve the problem of refinement of each link signal sublevel.Finally,based on PCA method,feature extraction and dimensionality reduction are performed for each vibration signal obtained from the endpoint,and a vibration signal classification method based on multiple machine learning is proposed to achieve classification and storage of each link signal.(3)To address the problem of difficult feature extraction and state classification of fault signals,an ASO-VMD-MPE feature extraction method and an ASO-HKELM fault pattern recognition method are proposed.Firstly,the feature vector of the toothed disk locking signal is constructed using the time and frequency domain features of the PCA dimensionality reduction combined with the completion time of the locking action.Secondly,for the loading signal,the ASO-VMD-MPE-based feature extraction method is proposed and the superiority of the ASO-VMD method is demonstrated by using the simulated signal.Finally,by comparing the recognition effects of ANN,SVM,DT,and ELM,a fault recognition model based on ELM is established.Aiming at the problems of ELM,a fault pattern recognition method based on ASO-HKELM is proposed to accurately identify the fault of the locking system of the end-toothed disc.(4)Fault diagnosis test and method verification of the end-toothed disc locking system of a type of domestic CNC tool holder.Firstly,the reliability test bench and signal acquisition scheme of the CNC tool holder are introduced.Secondly,a typical fault test scheme is designed and carried out.Finally,according to the collected fault data,the vibration signal extraction algorithm and the fault identification algorithm are verified,which proves the accuracy of the proposed method.