Mehdi Fadaie Abras, Karim Abbaszadeh, Alireza Siadatan,
Volume 1, Issue 1 (Journal of Control (English Edition), VOL. 01, NO. 01, 2022)
Abstract
Economic approach and optimization in rail transportation systems led to the introduction of the mono inverter dual parallel motor (MIDP) system. Most researchers introduce the model predictive control (MPC) method to drive this system in order to overcome the problem of load torque inequality on the wheels. But the obtained control signals do not result in the proper operation of the MIDP system, because the cost function is solved online or evaluated by the limited number of control signals. The present paper introduces an energy-based predictive speed control instead of the conventional proportional-integral controller in the outer loop and uses Pontryagin’s maximum principle to regulate electrical currents in the inner loop. Since this method solved the quadratic-linear cost functions offline, the control signals of the MIDP system are obtained as linear-parametric functions. After modeling and simplifying the mathematical equations, the introduced method is simulated and compared with conventional Finite and Infinite Control Set-MPC methods. The results indicate the agility and high accuracy of the controllers in both transient and steady states.
Hasan Zamani, Karim Abbaszadeh, Mohammadhadi Karimi, Johan Gyselinck,
Volume 1, Issue 2 (Journal of Control (English Edition), VOL. 01, NO. 02, 2022)
Abstract
For synchronization with the grid and controlling the injected active and reactive currents of the LCL-filter based grid-tied inverters, capacitor voltages can be sampled. An LCL filter attenuates the switching harmonics effectively but needs an extra sensor for the LCL filter resonance damping. Popular methods use capacitor currents for the LCL filter resonance damping. Theoretically, the derivative of capacitor voltage, which is proportional to the capacitor current, damps the resonance, and the extra sensor is avoided. However, traditional discretization methods for digital implementation of the derivative operator are not valid when the resonance frequency is high. Indeed, they don't preserve the phase and magnitude of the ’s’ function in the resonance frequency region. This paper introduces an effective method for discretizing the ’s’ function in the desired frequency range. The capacitor voltages of the LCL filter are sampled and the proposed function makes their derivative. The output of the derivative function with a tuned gain is added to the controller’s output for damping the LCL filter resonance. The simulation results show the effectiveness of the proposed method.
