Advanced modeling and control of networked DC microgrid

Abstract

To meet the requirements of decentralized energy supply and low-carbon concerns, the DC microgrid with high efficiency and low control complexity will be a key tool to provide the flexible power distribution for the energy prospect. For the model and control of DC microgrid, this work will review the centralized secondary control and distributed secondary control with theoretical derivatives and results. Then, the improved adaptive virtual resistance control and event-triggering mechanism for secondary power and current allocation are proposed. Many optimization problems in DC microgrids can be solved by using the convex optimization-based methods. The high distribution power loss of DERs is a prominent issue that can deteriorate efficient operations of DC microgrids. The distribution power loss may not only degrade power transfer efficiency but also increase the cooling system costs. After literature review and measurement, it is found that the line loss and converter loss are quadratic functions of the output currents of DERs. Hence, the distribution power loss of the multi-DERs in DC microgrid is proofed as a convex function of the output currents. Given supply-demand balance, the distribution loss model is further modelled as a convex function with equality and in-equality constraints. Then, a hierarchical control is designed for loss minimization: convex optimization strategy in the tertiary layer, an adaptive droop control in the secondary layer, and the local dual-loop control in the primary layer. Generally, the distributed secondary control can eliminate the inherent drawbacks of the centralized counterparts, such as the high risk of facing single-point failure, high cost on data processing, and periodical updates on entire system models. However, it is also vulnerable to false data injection (FDI) attacks, which are considered to be the most frequent cyber-attacks in DC microgrids. Usually, the hackers are prone to falsify the variables of controllers to bring about incidents, such as bus voltage deviations, imbalanced power allocations, and even instability of the whole system. In practice, it is reasonable to assume that the dynamic response of primary control is much faster than that of higher control layer. By treating DERs as nodes, the DC microgrid can be modelled as a dynamic system. Thereby, the DC microgrid under FDI attack can be further written as a first-order equation with disturbance, which provides the theoretical basis for using observation method to solve the problem of FDI attack. Then, various extended state observers can be applied to estimate and compensate the attack signals. In this work, two observers, the distributed sliding mode observer (DSMO) and the distributed high-order differentiator (DHOD) are proposed with the convergence analysis. The results show that the dynamic response speed of the DHOD is better than that of the DSMO. However, the DHOD requires more computing resources of micro-processor.