Operation Monitoring of Axial Flow Fans Based on Physics-Guided Support Vector Regression and Alarm Delay Timers

Axial flow fans play a critical role in the safety and efficiency of industrial processes. This paper proposes an operation monitoring approach for axial flow fans based on physics-guided support vector regression and alarm delay timer. The physical model of the axial flow fan is established based on aerodynamics to guide variable selection and feature construction. The physically augmented features are then used to train the support vector regression model for fan pressure prediction. The model residuals serve as the monitoring indicator for operation monitoring of the axial flow fan. The alarm delay timer is introduced to reduce false alarms caused by transient disturbances. The physically augmented features enhance the interpretability and generalization of the support vector regression model. The alarm delay timer improves the operation monitoring performance for the axial flow fan. Industrial case studies illustrate the effectiveness of the proposed approach.