Optimization Based Support Vector Classifier for Vibration Fault Diagnosis of Steam Turbine-Generator Sets

AbstractThis paper presents particle swarm optimization (PSO)-based support vector classifier (SVC) to extract the optimal support vector from database for vibration fault diagnosis of steam turbine-generator sets (STGS). In this paper, the SVC is used to construct the vibration fault diagnosis model and the proposed PSO is then adopted to determine automatically the optimal parameters in the SVC. Test results demonstrated that the proposed PSO-based SVC method can achieve greater diagnostic accuracy than the SVC method. Thus, this paper has demonstrated the feasibility of applying the proposed PSO-based SVC method to the practical vibration fault diagnosis of the STG