Li Xiang Duan’s research while affiliated with China University of Petroleum - Beijing and other places

In oil-delivery pumps, impeller failure is a common cause leading to excessive vibration. This paper is aimed to analyze the fault causation of impeller failure in oil-delivery pumps by using fuzzy Petri net (FPN) theory. The longest path algorithm based on the forward reasoning was put forward and introduced into fuzzy Petri net. First, on the basis of various factors causing impeller failure, an FPN model of impeller failure in oil-delivery pumps was constructed. Then, by using the proposed algorithm, fault causation analysis of impeller failure was completed to calculate the credibility of impeller failure. Finally, the corresponding preventive measure was presented. The results indicate the key causing factor of impeller failure is mechanical impurities, and the credibility of impeller failure is 0.7342, which is consistent with the actual situation. The research finding demonstrates the flexibility and effectiveness of the FPN in fault causation analysis.

The helath condition of rotor has been greatly concerned in rotating machinery. But for the lack of information, it is very difficult to judge the actual conditon. Based on the fuzzy and grey characteristics between faults and symptoms, a new method integrated with fuzzy clustering and grey relation analysis was put forward to identify the condition of rotor system. Firstly, eight features, such as average value, peak-peak value, variance value, virtual value and etc., were extracted from the vibration signal of rotor system. Then, fuzzy C-means algorithm was used to cluster forty samples into 4 clusters, meanwhile, the clustering center was acquired and regarded as standard pattern matrix. Finally, the grey relation degree was calculated between pattern to be inspected and the standard pattern matrix. Using this method, the unbalanced conditions of rotor system was precisely identified, which shows that the integrated method is valid and practicable.