Condition Partition and Recognition of SOFC System Based on Fusion Clustering

The running state of the SOFC system determines whether it can provide normal power for other equipment. If there is a fault and can not be found in time, the fault will quickly spread to other components and make the system run abnormally. To accurately grasp the type of operating conditions is the key to ensure that the SOFC system is in a normal working state. In this paper, taking SOFC system as the research object, a method of condition division and identification of SOFC system based on fusion clustering is proposed. Take the typical air leakage fault as the main body, use the fusion clustering algorithm to get the clustering tags under different operating conditions; then use the RBF-SVM classification algorithm to train the clustering tags, and get the working condition identification model. The results show that this method can effectively identify the working condition type of air leakage in SOFC system, and can be extended to the fault diagnosis of fuel leakage.