Multi-Agent Collaborative Control of Photovoltaic Microgrid Clusters Considering Virtual Inertia Support

Abstract

The increasing penetration of photovoltaic generation in modern power systems has precipitated a transition from centralized, high-inertia grids to distributed, low-inertia microgrid clusters. This paradigm shift introduces significant stability challenges, particularly regarding frequency resilience during islanding or load fluctuations. This paper proposes a novel multi-agent collaborative control framework designed to enhance the frequency stability of photovoltaic microgrid clusters through coordinated virtual inertia support. Unlike traditional centralized control schemes, which suffer from single points of failure and communication latency, the proposed method utilizes a distributed multi-agent system architecture. Each distributed generation unit functions as an autonomous agent, employing virtual synchronous generator control strategies to emulate the electromechanical behavior of traditional synchronous machines. We introduce an adaptive consensus algorithm that enables agents to dynamically allocate active power and inertia support responsibilities based on their real-time state of charge and generation capacity. Theoretical analysis and time-domain simulations demonstrate that this collaborative approach significantly reduces the rate of change of frequency and steady-state frequency deviation compared to conventional droop control methods. The results indicate that incorporating virtual inertia support within a multi-agent framework provides a robust solution for maintaining the stability and reliability of low-inertia renewable energy systems.