Fault diagnosis can be achieved by the cluster analysis process. We give a complete similarity measurement method and a cluster criterion function defined by modularity increment in complex network community detection. An agglomerative cluster method is proposed and the diagnosis rules are extracted to fault diagnosis for the marine engine system. Using the samples collected from self-manufactured marine engine room simulator, fault diagnosis simulation experiment is carried out to verify the algorithm performance. The results show that the method is accurate and less time-consuming, and is able to recognize the fault pattern which does not exist in the fault history data.
We generalize the label propagation algorithm in complex networks to weighted networks by weighting the label propagation rule and the termination condition of label propagation algorithm. Experiments on computer-generated networks and real-world networks are carried out to compare the generalized algorithm with original algorithm. The results show the generalized algorithm is accurate, and the number of iteration is less. The weighed label propagation algorithm is applied to fault classification, and the good classification results are obtained.
Fault diagnosis for marine engine system can be achieved by cluster process. We propose a multiple clusters method using label propagation model repeatedly to cluster for fault samples. Due to randomness in label propagation, some classes which are not easy to be found in single cluster can be found. To verify the performance of this method, the samples from self-manufactured marine engine room simulator are used in fault diagnosis simulation experiment. Both fault analysis results and fault diagnosis results are completely correct, and fault pattern which is not in fault history data can be recognized.
Cluster methods can be used for fault diagnosis of marine diesel engine. We propose a cluster method by assigning a class label to each sample and updating it according to the class label probability. Due to the random walk process in cluster process, multiple results are obtained, and some classes can be found more easily. Fault diagnosis simulation experiment is executed on a self-manufactured marine engine room simulator to verify the method performance, and we obtain the correct graphical and number results. Both fault patterns in fault history and new fault pattern can be recognized.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.