Data driven approach for fault detection and diagnosis of turbine in thermal power plant using Independent Component Analysis (ICA)

► Statistical signal processing method is used for fault detection and diagnosis. ► ICA is used for finding underlying components from multivariate statistical data. ► Combined statistical approach is applied for fault detection of a real power plant. ► Suggested method eliminates the requirement to detailed model of power plant. ► Results indicate that suggested method can distinguish main factors of abnormality.