A Novel Fault Diagnosis Method for the Plant with Min-Disturbance

摘要

For estimating sensor faults for the plant with min-disturbance, a novel fault diagnosis method based on support vector interval regression is proposed. This method can reduce the influence of min-disturbance. When faults occur in this plant, the interval of regression model can not contain outputs of the plant and the faults can be detected. Experiments are given to demonstrate the efficiency.