A Projected Gradient Descent-Based Distributed Optimal Control Method of Medium-Voltage DC Distribution System Considering Line Loss

Abstract

Real-time minimization of line loss is a great challenge for conventional distributed control methods in medium-voltage DC distribution system (MVDC-DS), which may lead to low efficiency and high energy consumption. To solve this problem, a novel projected gradient descent-based distributed optimal control method is proposed by combining the operation scheduling layer and coordination control layer. Firstly, the limitations of conventional consensus-based distributed control methods are analyzed. Then, the projection operator is introduced to modify the optimality conditions of the optimization problem so that variables with constant derivatives, such as line loss, can be integrated into the objective function of the operation scheduling layer. On this basis, the proposed distributed control is presented, and the iterative solution process based on projected gradient descent is realized by integral controller and feedback mechanism. Finally, the convergence and stability of the proposed method are analyzed. Results of some different scale case studies verify that the proposed method not only improves the system economy by real-time line-loss optimization, but also realizes voltage and power management.