Degree Name

Master of Philosophy


Department of Electrical, Computer and Telecommunication Engineering


Most of the transmission networks in modern interconnected power systems are more heavily loaded than ever before to meet the growing demand. The continuing interconnection of bulk power systems due to economic and environmental pressures has led to an increasingly complex system that must operate closer to the stability limit. This is particularly worst during the peak demand of the year. Under such a stressed system, when catastrophic events due to unplanned multiple contingencies occur; the transmission grid cannot maintain its integrity to maintain the resilience of the network. As a result, power systems become vulnerable to various instability problems such as voltage instability, transient instability, dynamic instability etc. It is important to detect the causes of system breakdown and to actuate fast countermeasures to mitigate the impact of contingencies so that the power system, even under such catastrophic disturbance, can operate with sufficient security and reliability.

One type of system instabilities, which is usually experienced when the system is heavily loaded, is the voltage instability. This event is characterized by a slow variation in the voltage magnitudes followed by a rapid sharp disruptive change resulting in voltage collapse. Analysis of several voltage collapse incidents in the past few decades has revealed that the first impact of any critical disturbance occurs in a limited region of the transmission grid, gradually encompassing the entire grid if timely countermeasures are not taken. In this project, a novel approach based on the multi-agent technique is proposed to counteract the voltage instability and the resulting voltage collapse issues that arise from an unplanned multiple contingency. At first, the transmission network is divided into some local areas to take the benefit of the initial limited geographical effect of voltage instability. Several criteria such as bus effectiveness factor based on the reactive power injection capability and the electrical distance among the buses are considered to find the local zones. To determine the severity of the disturbance that can lead to voltage instability, performance indices have been formulated based on the local variations of load voltage magnitudes and generator reactive power outputs. Each area is assigned a team of intelligent agents to detect the occurrence of the instability and to initiate the appropriate and timely countermeasures to stabilize the system. A decentralized architecture of the multi-agent system is used so that the agents can take quick decision without any intervention from the central controller. For this purpose, various negotiation protocols among the agents have been researched to determine the proposed solution using Java Agent Development Framework (JADE). To determine the optimal amount of countermeasures, a sensitivity approach based on the linearized power flow equations has been proposed. Simulation results based on the IEEE benchmark systems have been used to validate the proposed methodology.