Research On Fault Diagnosis Of Transmission System Of Hot Die Forging Press Based On Vibration Feature Extraction

The transmission system is an important part of the hot die forging press.Some components such as bearing and gear of transmission system fail frequently because of impact load and untimely maintenance and so on.In order to reduce component failures and improve the intelligent level of component fault diagnosis,this paper proposes to use the vibration feature extraction method to study the fault of the transmission system for hot die forging press.The main contents of this dissertation are as follows:1.Analysis of faults in the hot die forging press transmission system components.The main research object is the 2,500-ton hot die forging press produced by Jiangsu Yangli Group.Firstly,the structure and working principle of the hot die forging press was introduced,and then the common faults and fault characteristics of the hot die forging press transmission system was analyzed.Meanwhile,the analysis focuses on the characteristics of common faults and fault vibration characteristics of rolling bearings and gears in transmission systems.2.Vibration signal acquisition and experimental design of hot die forging press transmission system components.In order to obtain the vibration signal of the transmission system components,the fault simulation test bench was designed firstly,and the corresponding faults were processed on different self-aligning double row rolling bearings.Then the signal acquisition program was compiled by using the LabVIEW,and the fault vibration signal of the rolling bearing was collected by making use of the NI data acquisition Cards and other equipment.Finally,the vibration signal of different fault states of the gears was collected by using the QPZZ-II test rig.3.Fault feature extraction of hot die forging press transmission system.In order to construct the feature vector samples for fault pattern recognition,first of all,the noise and trend terms of the vibration signal were removed by the five-point cubic smoothing method and the polynomial least squares method.Secondly,the several fault feature extraction methods were analyzed.Finally,the time domain statistic characteristics,characteristic frequency and energy characteristics of wavelet packet decomposition nodes of self-aligning double row rolling bearings and gears were extracted by using the index method,the amplitude-frequency method and the nodal energy method respectively.4.Fault pattern recognition of hot die forging press transmission system components.Firstly,the classification idea and algorithm of support vector machine were expounded,and the sequence minimum optimization algorithm and sequence minimum support vector machine were introduced.Different faults of self-aligning double row rolling bearings and gears of hot die forging press transmission system were identified with sequence minimum support vector machine classifier and fusion features.In order to improve the accuracy of fault pattern recognition of components,a pattern recognition method based on the sequence minimum support vector machine optimized by grid search algorithm was proposed,and the method was used to identify the different faults of parts.The results show that the proposed method is effective.