Abstract
High penetration of distributed generation units in microgrids has led to challenges of incorporating the generation resources while maintaining the frequency and voltage stability. Traditionally, the coordination of converter-based generation is achieved by autonomous droop control, which suffers from dependence on the grid topology and power line impedances. Adopting the communication infrastructure that supports the microgrid operations, the coordinated secondary control can resolve this problem by judiciously dispatching active power resources and reactive power support. Nevertheless, the conventional centralized secondary control architecture requires high-cost communication infrastructure and is low in scalability. Due to inherently distributed and heterogeneous nature of the microgrid, it becomes an ideal platform for applications of distributed control algorithms. To this end, this dissertation aims to develop novel distributed secondary control for power converters in microgrids. The control algorithm is based on multi-agent consensus control and will be able to enhance both the real-time frequency and voltage stability. First, the consensus-based droop control with sparse communication network is proposed to overcome the drawback of existing droop control methods. In particular, when line impedances of the power grid are either lossy with the uniform R/X ratio or even pure resistive, the consensus droop control is still an effective method for autonomous real and reactive power sharing. In addition, closed-loop system stability of the proposed consensus-based droop control method can be ensured by the energy function approach under certain mild conditions. Real-time simulations of two microgrid systems are studied to validate the feasibility of the proposed consensus-based droop control method. Second, the conventional distributed interface converters (DICs) used in the AC microgrid have raised a major concern since they do not have a rotating mass. The almost inertia-less DIC-based microgrid configuration may result in poor frequency and voltage response during large disturbance. In order to overcome this difficulty, the virtual synchronous generator (VSGs) was proposed recently in which the DIC mimics conventional synchronous generators (SGs) by designing proper parameters of the SG into each local droop control mechanism of the DIC. Under this framework, we deploy the distributed consensus-based control algorithm in the secondary control level of VSGs. With only neighboring information exchanged between VSGs, both the restoration of nominal operating point and the accurate generation coordination of resources can be achieved. The stability of the closed-loop system is ensured by the transient energy function under certain mild conditions. Numerical experiments of a 14-bus/6-DIC microgrid system on real-time simulators are performed to validate the effectiveness of the proposed control mechanism. Third, although the average consensus algorithm can ensure the stability of operation points while achieving accurate power sharing among VSGs in microgrid, its convergence rate and the resilience to noises are still not satisfactory. Therefore, two implementations of consensus-based distributed droop control by the Alternating Direction Multipliers Method (ADMM) are proposed. By employing the ADMMs in the secondary control level, it can be shown that the closed-loop system is described by the second-order vector differential equation, and the stability of every trajectory can be analyzed by the energy function approach under certain mild conditions. Real-time digital simulations are performed to validate the effectiveness of the proposed distributed droop control mechanism. Moreover, a communication-based distributed frequency control framework also exposes the microgrid assets to potential malicious cyber-attacks. To study this problem, a distributed secondary control design for isolated microgrids is first developed, and the countermeasures for malicious attacks on the communication network are then investigated. The proposed design architecture consists of the local droop control at the primary level and a distributed node-to-node update at the secondary level. The latter aims to achieve a proportional power sharing while maintaining the nominal system frequency. By casting it as a consensus optimization problem, we adopt the partial primal-dual (PPD) algorithm for a totally distributed update requiring only neighboring information exchange. Interestingly, the specially designed PPD-based control rules would mimic the network power flow dynamics. Furthermore, two types of malicious attacks on the communication network, namely, the link and node attacks, are studied. Model-based anomaly detection and localization strategies are developed based on the dual variable related metrics. Numerical simulations have been performed under a real-time simulation environment to demonstrate the effectiveness of the proposed control design and countermeasure metrics.