Autonomous grid-forming inverters in microgrids

Abstract

Inverters can be programmed to behave as an interface between energy sources and consumers, creating the voltage and frequency references. Therefore, they are capable of avoiding power outages and meeting demand using available generation units within distribution power grids. This operational strategy is known as grid-forming (GFM), and enables clustered grid configurations, such as self-powered or networked microgrids. In these setups, a set of inverters can form a power grid or reconnect to an existing microgrid to supply the loads in an isolated subdivision or commercial buildings, enabling black-start capabilities for a deenergized microgrid. During the presence of the utility grid, the inverters should change the mode of operation to grid-following (GFL), following the voltage magnitude and frequency measured at the point-of-common-coupling, for optimum power management. However, grid detection requires communication with utility supervisory controllers to close the synchro-check relays, which hinders plug-and-play compatibility of inverters while escalating more challenges such as adaptability, communication disruptions or cyberattacks. Therefore, a decentralized (communication-less) grid-reconnection detection operation is crucial to provide flexibility for inverters for changing the mode of operation from GFM to GFL when communication signal is lost. In this dissertation, a novel synced-parallel control task concept is proposed for inverters that can autonomously detect grid-reconnection and islanding incidents, transitioning between GFL and GFM modes without external data communication or control units. The proposed controller, leverages a set of dual parallel control pathways, one path is dedicated for GFL operation while the other path is for GFM mode for both PWM voltage and angle reference generation. During the operation, one path is active (main path) and generates the reference signals, whereas the other path only runs in the background (inactive path) to match its outputs with the outputs of the active path. This active synchronization process ensures seamless transitions between different modes of operations. Furthermore, an innovative autonomous grid-reconnection detection algorithm is developed to identify the presence of utility grid and perform a smooth transition from GFM mode to GFL. The efficacy of the proposed autonomous control scheme is validated through a multi-inverter three-phase hardware setups forming two microgrids. One microgrid consists of a 30 kVA grid emulator, two 5 kVA and one 10 kVA, 208 V three-phase inverters, one 5 kW static load, and three-phase 208 V 4-pole 1/3 HP induction motor, while the other microgrid setup consists of a grid emulator with the same rating and two 5 kVA 208 V inverters as well as the a 5 kW static load and a three-phase 208 V 4-pole 1/2 HP induction motor. Experimental outcomes verify that inverters equipped with this unique control architecture can autonomously identify grid reconnection and islanding situations, seamlessly transitioning between GFL and GFM modes. Furthermore, the islanding and grid-reconnection instances are detected although inverters are connected to different nodes within a microgrid. This means that the developed method are capable of identifying the reconnection of utility grid with direct access to the utility-side of the main circuit breaker in the microgrid. This dissertation also comprehensively compares the performance of two common grid forming control strategies, i.e., direct and indirect control methods under static and dynamic loads, and different control parameter settings. To allow inrush current mitigation for the direct control method, a piece-wise linear virtual reactance technique is developed to mitigate inrush currents by allowing flexible voltage generation during a switched connection of an induction motor. On the other hand, outer loop coefficient settings of the indirect control method are changed to allow flexible voltage regulation at the output node, known as point of common coupling (PCC). This comparison procedures are completed experimentally and also using eigenvalue analysis. Statistical results on voltage recovery are also provided for comparison purposes. The outcomes verify that direct control method outperforms the indirect method while achieving simplicity in design, flexibility in controller gain selection, and more stable operation margin.