Fuzzy Logic-Based Fault Diagnosis for Independent Metering Control Systems

Independent metering control systems (IMCS) pose significant fault diagnosis challenges due to strong hydraulic– mechanical coupling, limited sensing, and the indirect mapping between internal faults and measurable signals. To address the limited fault coverage of existing IMCS diagnostic approaches, this paper presents a model-based fault diagnosis framework enhanced with fuzzy logic reasoning. First, a physical model is calibrated using time-domain similarity metrics to improve consistency between simulated and measured responses. Using this model, state- and input-based residuals are derived by exploiting IMCS analytical redundancy. These residuals capture the characteristic signatures of valve, displacement, and pressure sensor faults. To further enhance robustness, a trend-aware residual smoothing mechanism is introduced to suppress fluctuations caused by noise and disturbances while preserving fault-related signatures. The smoothed residuals are then evaluated by a fuzzy logic-based system, enabling interpretable fault classification and severity estimation. Experimental validation on a full-scale test platform demonstrates that the proposed method can accurately identify multiple fault types and severity levels, achieving an overall diagnostic accuracy of 96.67% without requiring additional sensors or hardware modifications. Future work will extend this framework from single-fault conditions to concurrent fault diagnosis.