Nonlinear PI-Type Speed Control Design and Its Application to Monitoring of a Cooling Fan System With Perturbations

The increasing power density of modern electronic systems has intensified the need for efficient thermal management, where cooling fans play a pivotal role. However, fan performance is frequently compromised by external disturbances, aging effects, and sensor limitations. To address these challenges, this study presents a robust nonlinear proportional-integral (PI)-type speed control strategy. Unlike conventional PI or sliding mode-based designs, the proposed architecture integrates a nonlinear robustifying term within the integral path. This structural innovation filters the switching signal through the integrator and yields a continuous control input. Consequently, it effectively eliminates chattering while maintaining strong robustness against perturbations. Furthermore, a linear matrix inequality (LMI)-based design framework is established to handle system uncertainties. This systematic approach simultaneously guarantees global asymptotic stability and optimizes the control gain and convergence rate, even in the presence of bounded nonlinear aerodynamic uncertainties. Beyond robust control, the proposed method integrates physical modeling and parameter identification to enable sensorless real-time health monitoring via a digital twin. This facilitates early anomaly diagnosis without additional hardware sensors. Extensive experimental validation across various commercial cooling fan types demonstrates the robust generalizability and superior tracking accuracy of the proposed method compared to linear benchmarks. The developed framework supports real-world deployment in intelligent thermal management, fault-tolerant fan array, and condition-based maintenance applications.