Real-Time Control Strategy of Tractive Load Peak Clipping and Valley Filling Based on Model Predictive Control

Abstract

Access to energy storage devices (ESDs) is an effective way to solve the peak traction load shock and Regenerative Braking Energy (RBE) recycling. However, in the real-time operation of the system, there are problems of prediction errors affecting the control results and the problem of energy storage “dead time.” This article analyzes the mechanism of the energy storage “dead-time” problem and aims to solve this problem without increasing the energy storage capacity. Therefore, this article combines the ideas of rolling optimization and predictive control in the model predictive control (MPC) method and proposes an MPC method based on the adaptive correction (AC-MPC) of the State of Charge (SOC) of ESD. The predictive control can ensure the accuracy of power correction; the rolling optimization can reduce the global impact of each correction and reduce the complexity of the calculation. In addition, the rolling prediction in the rolling optimization link reduces the prediction error by reducing the prediction time scale, thus reducing the impact of the prediction error on the control results. The experimental data show that the proposed control strategy can effectively solve the two problems mentioned above, ensure the effect of peak reduction and RBE recovery, and improve the economy of system operation.