[Paper Review] Voltage stabilization in DC microgrids: an approach based on line-independent plug-and-play controllers

Michele Tucci, Stefano Riverso+1 more|arXiv (Cornell University)|Sep 8, 2016

Microgrid Control and Optimization4 citations

TL;DR

This paper proposes a line-independent, plug-and-play control design for DC microgrids that enables seamless integration of new distributed generation units without retuning neighboring controllers or requiring precise line parameter knowledge. Using a structured Lyapunov function and graph Laplacian analysis, the method ensures asymptotic voltage stability via solvable linear matrix inequalities (LMIs), validated through PSCAD simulations on a 5-DGU system with successful plug-in of an additional unit.

ABSTRACT

We consider the problem of stabilizing voltages in DC microGrids (mGs) given by the interconnection of Distributed Generation Units (DGUs), power lines and loads. We propose a decentralized control architecture where the primary controller of each DGU can be designed in a Plug-and-Play (PnP) fashion, allowing the seamless addition of new DGUs. Differently from several other approaches to primary control, local design is independent of the parameters of power lines. Moreover, differently from the PnP control scheme in [1], the plug-in of a DGU does not require to update controllers of neighboring DGUs. Local control design is cast into a Linear Matrix Inequality (LMI) problem that, if unfeasible, allows one to deny plug-in requests that might be dangerous for mG stability. The proof of closed-loop stability of voltages exploits structured Lyapunov functions, the LaSalle invariance theorem and properties of graph Laplacians. Theoretical results are backed up by simulations in PSCAD.

Motivation & Objective

• To address voltage stability in DC microgrids under decentralized control with minimal communication and no reliance on line parameter knowledge.
• To enable seamless plug-and-play integration of new distributed generation units (DGUs) without retuning neighboring controllers.
• To develop a control design procedure that is both scalable and privacy-preserving, suitable for multi-owner energy systems.
• To prove asymptotic stability of the closed-loop system using structured Lyapunov functions and graph Laplacian properties.
• To provide a feasibility test via LMIs that can reject unsafe plug-in requests before deployment.

Proposed method

• The control architecture uses a decentralized, primary-level controller for each DGU, designed independently using a Linear Matrix Inequality (LMI) problem.
• The LMI formulation depends only on local DGU parameters and a global scalar parameter σ̄, not on line resistance or inductance values.
• Stability is proven using a structured Lyapunov function and the LaSalle invariance principle, leveraging the graph Laplacian representation of electrical coupling under Quasi Stationary Line (QSL) approximations.
• The method allows plug-in requests to be validated via LMI feasibility: if infeasible, the request is denied to prevent instability.
• The controller structure is identical to prior PnP work, but the design is line-independent, eliminating the need for neighbor controller updates upon DGU connection.
• Simulations in PSCAD validate the approach on a 5-DGU microgrid with successful plug-in of an additional DGU, showing stable voltage dynamics.

Experimental results

Research questions

• RQ1 Can a plug-and-play control design for DC microgrids be made independent of power line parameters while preserving stability?
• RQ2 Does the proposed method allow new DGUs to be added without updating controllers of neighboring units?
• RQ3 Can LMI-based feasibility tests reliably predict whether a new DGU can be safely integrated into the microgrid?
• RQ4 Is asymptotic voltage stability guaranteed under the proposed line-independent control architecture?
• RQ5 How does the proposed method compare in feasibility and scalability to line-dependent PnP designs?

Key findings

• The proposed line-independent LMI-based controller design is feasible for typical low-voltage DC microgrid parameters, as confirmed by extensive numerical testing across a wide range of converter parameters.
• For the same DGU parameters, the line-independent LMI (58) remains feasible even when the line-dependent LMI in [1] becomes infeasible, indicating broader applicability.
• The maximum number of DGUs that can be connected to a single PCC before failure in the line-dependent approach was found to be three, failing at the fourth plug-in request.
• The closed-loop system achieves asymptotic voltage stability, proven via structured Lyapunov functions and the LaSalle invariance theorem applied to the graph Laplacian model.
• Simulations in PSCAD demonstrated stable voltage dynamics during the plug-in of an additional DGU into a 5-DGU microgrid, confirming the practical viability of the method.
• The method supports privacy-preserving operation, as new DGU integration does not require disclosure of neighboring DGU models or controller updates.