TL;DR
This paper introduces a novel distributed control architecture for AC islanded microgrids that allows for seamless "plug-and-play" integration and removal of distributed generation units while maintaining grid stability. The system enhances existing plug-and-play methods with virtual impedance and coordinated secondary control to ensure voltage tracking and reactive power sharing, validated through extensive experiments.
What Problem Does This Paper Address?
Islanded microgrids (ImGs) are crucial for reliable power supply in remote areas or as a backup for the main grid. A key challenge in ImGs is scalability: how to add or remove distributed generation units (like solar panels or wind turbines) without disrupting the grid's voltage and frequency stability. Traditional control methods often require a global model of the entire microgrid, which is impractical for dynamic systems. Furthermore, existing plug-and-play (PnP) approaches were primarily designed for "load-connected" topologies, where each generator has its own dedicated load. However, many real-world microgrids use "bus-connected" topologies, where multiple generators supply a common load bus. This paper addresses the need for a robust, decentralized PnP control strategy that works for bus-connected microgrids and also incorporates advanced features for better performance.
Methods
The paper builds upon a previously developed PnP control framework that uses convex optimization to design local controllers. This approach ensures stability by only requiring knowledge of local parameters and neighboring line impedances, avoiding the need for a global microgrid model.
Key methodological contributions include:
- Bus-Connected Topology Adaptation: The authors devise a method to map bus-connected microgrid topologies into equivalent load-connected ones, allowing the existing PnP control design to be applied.
- Virtual Impedance Loops: Local virtual impedance loops are integrated into the primary PnP controllers. This technique helps to mitigate circulating currents between generators and allows for better control over the output impedance of the voltage-source converters (VSCs).
- Distributed Secondary Control: A distributed secondary control layer is introduced. This layer focuses on two main objectives:
- Voltage Tracking at the Point of Common Coupling (PCC): It ensures the voltage at the common load bus is maintained at a desired reference level, even with varying loads.
- Reactive Power Sharing: It coordinates the reactive power injection from each generator to ensure equitable distribution.
- Convex Optimization for PnP: The core of the PnP design relies on solving Linear Matrix Inequality (LMI) optimization problems. When a unit wants to connect or disconnect, its local controller and those of its neighbors are re-tuned by solving these problems, ensuring stability is preserved.
Datasets
Not discussed in the paper. The paper focuses on experimental validation using a custom-built microgrid testbed.
Key Insights and Ideas
- Decentralization is Key for Scalability: The paper strongly advocates for decentralized control, where each generator has its own controller, as the most practical approach for achieving plug-and-play capabilities in microgrids.
- Topology Transformation is Feasible: The ability to transform a bus-connected topology into an equivalent load-connected one is a significant enabler for applying existing PnP control strategies to a wider range of microgrid configurations.
- Virtual Impedance Enhances Robustness: Adding virtual impedance to the primary control layer is shown to be effective in improving the performance and robustness of the microgrid, particularly in managing interactions between generators.
- Secondary Control Complements Primary Control: The research demonstrates that a distributed secondary control layer can effectively enhance the primary PnP control by providing voltage tracking and reactive power sharing, without compromising the stability achieved by the primary layer.
- PnP is Inertia-less: Unlike traditional droop controllers that rely on inertia, the PnP approach allows for faster transient responses when loads change or generators are added/removed.
Key Results
The paper presents extensive experimental validation on a three-generator bus-connected microgrid testbed. Key results include:
- Stable Plug-and-Play Operation: The system successfully demonstrated the addition and removal of generators without causing significant disturbances to voltage or frequency. Transients during these events were minimal.
- Effective Voltage Regulation: The primary PnP controllers, enhanced with virtual impedance, showed good voltage regulation at the PCCs under various load conditions, including linear, unbalanced, and nonlinear loads.
- Successful Secondary Control Implementation: The coordinated secondary control layer effectively achieved voltage tracking at the load bus and balanced reactive power sharing among the generators, as evidenced by experimental results.
- Harmonic Distortion and Imbalance: Experiments showed that the proposed controllers kept harmonic distortion and imbalance ratios within acceptable limits, even with nonlinear loads.
- Fast Transient Response: The PnP controllers exhibited faster transient responses compared to inertia-based methods when load conditions changed.
Limitations and Future Work
- Communication Network: The current implementation of the secondary control layer assumes a fully connected communication network for broadcasting measurements (e.g., voltages and currents) to all DGUs for averaging. This is a limitation for large-scale systems.
- Future Work: The authors suggest distributing these computations by exploiting only partial communication among DGUs and using consensus-like protocols for estimating global variables. This would make the system more scalable and robust to communication failures. They also mention that the current work focuses on AC microgrids and future research could extend to DC or hybrid AC/DC microgrids.
Summary updated