Research On Intelligent Monitoring Technology Of Deep Hole Boring Tool Condition

In the process of deep hole boring,the cutting area will be in the interior of deep hole parts for a long time.If the boring tool is abnormal in the interior of deep hole,and it can't be known and react in time,the whole deep hole part will be worthless or even the machine tool will be damaged.Therefore,it is of great significance to carry out the research of deep hole boring tool condition monitoring technology to ensure the high-quality and efficient machining of deep hole parts.For this reason,this paper has carried out the research on the intelligent monitoring technology of deep hole boring tool condition,using acceleration sensor and microphone to collect cutting vibration and sound,based on low-pass filtering and adaptive synthesis sampling method to preprocess the data,designed the tool condition monitoring model based on deep learning network and the tool condition monitoring framework based on transfer learning,and established the mapping relationship between the signal and boring tool condition.Finally,the intelligent monitoring of the condition of deep hole boring tool is realized.First of all,the failure mechanism of the tool is introduced,and the finite element simulation analysis of boring process is carried out.This paper briefly discusses the wear forms and causes of cutting tools,and introduces the wear process and life criterion of cutting tools.The wear process of boring tool is simulated by finite element method and analyzed.Secondly,the monitoring system of deep hole boring tool condition is built,and the data is collected and preprocessed.Comparing the advantages and disadvantages of various monitoring signals in deep hole processing application,combined with the processing experience of experienced workers,a multi-sensor data fusion monitoring scheme with vibration and sound signals as monitoring signals is designed,and a deep hole boring tool condition monitoring hardware system is built.In view of the data redundancy and imbalance in the collected data samples,the data preprocessing method based on low-pass filtering and oversampling technology is studied.The low-pass filtering method is used to remove the useless part of the data and improve the training speed of the model.Based on the adaptive synthesis sampling method,the data of minority class of broken and blunt samples are synthesized to weaken the over fitting phenomenon of model training and improve the generalization ability and classification accuracy of the model.Then,based on the deep learning network,the monitoring model of deep boring tool condition is established.Greedy layer-wise training method is used to train stacked sparse autoencoder network,and autoencoder network is used for feature extraction.The supervised learning method is used to train the deep long short-term memory network,and the mapping relationship between the data and the tool condition is established by using the advantages of the long short-term memory network to process the sequence data.Based on stacked sparse autoencoder network and long short-term memory network,a monitoring model of deep hole boring tool condition is established.Next,to solve the problem that the trained deep learning model is difficult to be used in the monitoring of similar machine tools,a transfer learning framework is proposed.Model transfer is used to transfer the structure and parameters of the pre-trained deep long short-term memory network.Based on the domain adaptation,the distribution difference between the source domain feature and the target domain feature is reduced.The fine-tune method of deep learning is used to fit the classification layer to the target domain sample,and finally the deep learning model transfer is realized.Finally,the boring tool condition monitoring software system is designed and developed,and the deep learning monitoring model and transfer learning framework are tested and verified.The test results show that the monitoring accuracy of the deep learning monitoring model and the transfer learning framework proposed in this paper is great,which can realize the real-time monitoring of the deep boring tool condition,and provide a guarantee for the high-quality and efficient processing of deep hole parts.