Fault Prediction Research Based On Fuzzy Methods

Now the development of equipment and systems become more and more complex, once the failure will result in significant losses. Therefore, it is very important to detect and predict the fault of equipment. As the fault complex equipment with characteristics of uncertainty, nonlinearity, fuzzy and so on. We design a method of fault prediction, which combines fuzzy mathematics, filter algorithm and grey theory model to predictive fault.The main contributions of this paper are as follows:(1) As fault prediction has the characteristics of uncertainty, design a method of fault prediction, which combines fuzzy mathematics membership function with particle filter algorithm to predictive fault. The new method using particle filter algorithm to calculate the future state of the device operation to predict, and then design the normal membership function and the abnormal membership function of the device operation state, calculate and compare the value of the normal and abnormal membership function by using the calculated results and based on the comparison result to predict potential failure. The feasibility of the proposed method is verified by experiments, which can predict the failure of the system in time.(2) As fault prediction has the characteristics of uncertainty, design a method of fault prediction, which combines fuzzy mathematics closeness degree with particle filter algorithm to predictive fault. The new method uses the membership function to describe the normal system with the normal fuzzy sets and the abnormal system with the abnormal fuzzy sets, and using particle filter algorithm to calculate predictive value, and using membership function to calculate the membership degree. Then calculate the closeness degree of predicted value of the normal membership degree with normal and abnormal fuzzy subset to implement a fault prediction. Compared with the first method, more reasonably to use the close degree to calculate the normal/abnormal fuzzy subsets of the sequence. And verified by experiments, which can predict the failure of the system in time.(3) In view of the above two methods, the particle filter algorithm is not fast and does not consider the non membership, we design a method of fault prediction, which combines intuitionistic fuzzy sets with grey model to predictive fault. The new method uses the membership function to describe the normal system with the normal intuitionistic fuzzy sets and the abnormal system with the abnormal intuitionistic fuzzy sets, and using grey model to calculate predictive value, and using membership function to calculate the membership degree. Then calculate the closeness degree of predicted value of the normal membership degree with normal and abnormal intuitionistic fuzzy subset to implement a fault prediction. Compared with the first and second methods, more reasonably to use normal/abnormal intuitionistic fuzzy subset to describe the predictive value, and this method is much faster.