A control strategy for electro-magneto-mechanical system based on virtual system model

作者：Kim Hong youn; Yun Young min; Heo Hoon^*

来源：Journal of Mechanical Science and Technology, 2016, 30(9): 4075-4081.

DOI：10.1007/s12206-016-0820-3

• Electro-magneto-mechanical system
• Solenoid valve
• Magnetic force
• Random current
• Autotuning PID control
• Noise
• SSID (Stochastic system, identification)
• Virtual 2nd order system
• Maxwell system
• PWM (Pulse width modulation)

A new approach to the control of electro-magneto-mechanical system is proposed in this paper. Conventionally, these systems are controlled based on the Maxwell system model via an on-off or PID control technique, which displays acceptable performance in the low frequency region, but not in the high frequency region where position control performance is greatly degraded. In order to improve the performance, a newly developed virtual 2(nd) order system modeling technique, SSID, is adopted for a complex electro-magnetomechanical system in the study. This technique states that any unknown system exposed to a random disturbance with unknown intensity can be identified in terms of a virtual 2(nd) order system model via the inverse process of a certain stochastic analysis. As a typical hybrid system, a solenoid valve is used as the target electro-magneto-mechanical system to study the modeling of the virtual 2(nd) order system. In order to confirm the performance of the proposed control strategy, autotuning PID controller in PWM mode is utilized. Simulations based on the conventional Maxwell system model with control via the bang-bang, autotuning PID, and the proposed virtual 2(nd) order system model approaches are conducted using MATLAB Simulink. Performance of these three systems in the low and high frequency bands is also compared. The simulation results reveal that the control performance of the virtual 2(nd) order system model is much improved compared with that of the Maxwell system model under autotuning PID and bang-bang controls in both low and high frequency regions, where the error is drastically reduced to approximately 1/5 of the original value.

• 出版日期2016-9