• Lyapunov-based MRAC with decoupling improves control of a nonlinear MIMO coal • mill. • Decoupling converts the MIMO system into three independent SISO loops. • MRAC outperforms PID in rise time, settling time, and overshoot. • Adaptive control ensures zero steady-state error and strong disturbance rejection. This research is concerned with the application of a Lyapunov-based Model Reference Adaptive Control system to a complex, nonlinear, and coupled MIMO model of a pulverized coal mill. The system has three controlled variables (outlet temperature, inlet negative pressure, and pressure difference) which are managed by manipulating three inputs: coal feed rate, air flow rate, and recycled warm air flow rate. The control of pulverizing a coal mill using a conventional Proportional-Integral- Derivative (PID) controller is a challenging task. Because the PID controller can work well in one operating condition, it exhibits poor handling of MIMO coupling, often leading to instability or sluggish response when applied directly to multivariable systems. In this context, the approach proposed here will include a decoupling strategy that changes the complex MIMO system into three individual SISO subsystems and optimizes controller parameters using an adaptive law derived from Lyapunov stability theory to minimize tracking error. A MATLAB/Simulink model and simulation were performed for the proposed MRAC controller, and a comparison was carried out with a conventional PID controller. The results of this simulation demonstrated significant improvements in the dynamic responses across all variables. In particular, the MRAC scheme improved the rise time, settling time, and overshoot of the outlet temperature by 30.78%, 84.3%, and 50.9%, respectively. Similar improvements were observed for the inlet negative pressure and pressure difference, where the rise time, settling time, and overshoot improved by 82.8%, 84.6%, and 44.2%, and by 92.5% and 70%, respectively. These results indicate that the MRAC system outperforms the standard PID controller in terms of performance, response time, and stability.
Gebregewergs et al. (Sun,) studied this question.