Real-Time Cascading Failures Prevention for Multiple Contingencies in Smart Grids Through a Multi-Agent System

Cascading failures (CF) is complex in nature and difficult to accurately model or solve mathematically. The current industry approach to preventing CF, which leads to blackout event, involves incurring losses. In this paper, a technique based on an adaptive multi-agent system algorithm is implemented to prevent CF without loss incurrence. The algorithm uses mathematical combinations heuristically selected through the use of sensitivities obtained from the economic dispatch history of the power system to redispatch the power from the generators. This approach enables the implementation of the algorithm on systems of any size. The algorithm is experimentally applied in real-time with the consideration of necessary constraints as it halts the occurrence of CF. The test system is an experimental set up of the generation and transmission side of the IEEE 30-bus system using a reconfigurable smart grid laboratory hardware developed for testing algorithms requiring two-way communication capabilities. It was first showed that the test system will experience CF if nothing is done to prevent CF after the occurrence of a contingency. A detailed experimental analysis of the ensuing blackout event is given. The algorithm was used to prevent CF in the system after the occurrence of N-1 and N-1-1 contingencies. The algorithm was also tested on a larger system, the IEEE 118-bus system, through simulation. The experimental and simulation results affirm the efficacy of the proposed algorithm for systems of any size. In fact, it was discovered from our results that the large number of generators in large systems helps the algorithm converge faster than it does for small systems, which have restricted resources and combinations.