Mitigation of torsional vibration in large mill drive train system using state feedback control method

Torsional vibration limits the speed loop response of industrial drives and servo systems, deteriorating the transient response to speed commands and load disturbances. This thesis presents a damping method for torsional vibration produced by compliant components between the motor and the load in rolling mill applications. The proposed damping algorithm can solve the limitation of the classical damping approaches due to the large values of system time delay. The proposed algorithm is based on State Feedback Control (SFC) method with modified Linear Quadratic Gaussian (LQG) approach using a torque sensor as a feedback element. The result of modification is a low order single-input single-output compensator that mitigates the torsional vibration without affecting the speed loop. The verification of the design and the dynamic performance is accomplished by using time and frequency domain responses with real rolling mill system parameters. The performance of step commands, mitigation of torsional vibration and robustness to parameter variation is satisfied by using the proposed method. Also disturbance rejection performance is improved through load torque compensation. The experiment is performed on a 0.8 KW system with 24 Hz mechanical resonant frequency. Simulation and experimental results from the experimental system verify the proposed damping algorithm.