Study On The Theory And Method Of Online Chatter Prediction For MR Intelligent Boring Bar Based On HMM-SVM

Due to the low structural stiffness of boring bar, chatter occurs frequently during precision hole boring process, it results in low quality of finished surface and even damages the cutting tool. To solve this problem, the theory and method of online chatter prediction for MR intelligent boring bar was studied systematically in this dissertatioa The research work is supported by the Natural Science Foundation of Zhejiang Province (Grant No. Y104462) and the National Natural Science Foundation of China (Grant No.50405036).In Chapter1, the background and significance of the research were introduced, the research status and development trend of the technology of online chatter prediction were described in detail. Then, the main research contents of this dissertation were put forward.In Chapter2, firstly, the generation mechanism of cutting chatter was studied, the stability of cutting machine tools was analyzed, and the Stability Lobe Diagram of cutting machine tools was derived. Secondly, in order to find out appropriate methods for the chatter suppression, the effect of structural stiffness and damping on system stability was analyzed. Then a magnetorheological fluid based intelligent boring bar was proposed.In chapter3, firstly, the Empirical Mode Decomposition (EMD) and Hilbert-Huang Transform (HHT) based method for chatter feature extraction was proposed, and the basic theories of EMD and HHT were investigated. Then, the vibration signal of boring bar was decomposed by EMD and then transformed by HHT. Finally, the features of chatter symptom were extracted by analyzing the Hilbert spectrum of each Intrinsic Mode Function (IMF), which can provide sufficient guarantee for follow-up chatter recognition and prediction.In chapter4, based on the study of feature fusion technology, FastICA theory was introduced into boring chatter recognition and prediction field, then the IMF virtual channels and FastICA based chatter symptom separation system was established. The signal-noised separation was accomplished by ICA, and the chatter symptom was gained. The simulation results showed that the EMD-ICA based vibration signal processing could separate the chatter symptom signal rapidly and effectively.In chapter5, based on the study of Hidden Markov Model (HMM) and Support Vector Machine (SVM), an HMM-SVM based method for boring chatter recognition method was proposed, and the chatter identification system was established. In the system, HMM was used for the boring status preliminary selection to get two possible boring status, and then SVM was used for the deeper classification, and gain the final recognition results.In chapter6, to verify the proposed theory and method, the experimental setup of MR fluid intelligent boring bar was built. Firstly, the boring vibration signal was descomposed by EMD. Then, IMFs were separated by FastICA and gained the chatter symptom signal. Finally, the boring statuses were classified by the HMM-SVM chatter recognition system. Series of experiments were carried out, and the results showed that the method could recognize boring chatter accurately and rapidly, and the intelligent boring bar could suppress chatter efficiently.In Chapter7, the main research work of this dissertation was summarized and prospected.