Research On The Method Of Fault Diagnosis Of Mechanical Product Based On WFCM-GRA

Machinery products have a wide range of applications.Operating status of mechanical product directly affects the product process and quality of production.Mechanical product failure will produce secondary failure of equipment,leading to potential hazards to the devices as well as life.Therefore,the realization of mechanical product's real-time monitoring and quickly fault diagnosis efficiently has important practical significance to ensure safety,continuous operation of equipment in good performance,reduce the potential loss of mechanical product failures and reduce lifecycle cost.On the basis of process monitoring,a two-stage method of fault diagnosis based on Weighted Fuzzy C-Means clustering and Grey Relational Analysis(WFCM-GRA)is proposed in this paper,to realize fault diagnosis of mechanical product.The main content in this paper including:(1)Because importance of the different feature and sample to process monitoring and fault classification and recognition is different,a feature selection method based on J-index to select a key feature,a sample weighting method based on density is proposed in this paper.Feature selection and sample weighting are aim to highlight the influence of the key features and representative samples in the recognition and classification.(2)According to the different characteristic parameters of the sample data feature,selecting the appropriate joint control chart for process monitoring of mechanical product,and verifying the reasonableness of the proposed the program of joint control charts in the monitoring process of engine cylinder block slow fault.(3)According to features of mechanical products fault diagnosis,establishing two-stage fault diagnosis model based on WFCM-GRA: getting the historical data of Center using FCM integrated clustering validity function in first stage,using Grey Relational Analysis(GRA)to get integrated relational degrees of real-time data and the standard sequence,and result in fault diagnosis according to sequence of integrated associated degrees in second stage.And verifying effectiveness of the proposed model through several validation cases On the basis of method proposed in this paper,a process monitoring and fault diagnosis system of engine intended for consumer.is developed finally.