Demonstration of automated curative power flow control with a multi-agent system using distributed series reactors in a PHIL simulation

Abstract

This paper demonstrates the real-time applicability of an automated curative power flow control system in a High Voltage (HV) grid. The system is realized in a laboratory environment as an agent-based control algorithm implemented on commercially available hardware. Measurements and control signals are exchanged between a Real-Time Simulation (RTS) of an HV grid model and the agent system. To control power flows, the agents use simulated flexible loads and feeders as well as physical Distributed Series Reactors (DSRs), which are looped into the RTS via a Power Hardware-in-the-Loop (PHIL) interface. The control algorithm is tested in various scenarios and the agents’ performance is analyzed regarding overload reduction and reaction time. The results show that the heuristic agent algorithm can solve congestions ad-hoc and improve the overall system state. The agents’ reaction times depend on measurement availability: While a fast measurement sending rate can enable the agents’ base reaction time of approximately 11s, longer intervals between measurements - which may be favorable from a control center operator’s point of view - can cause longer delays until congestions are solved. These results put previous works in perspective, which did not consider real-time simulation and proposed possible agent reactions within milliseconds.