Peidaee, Pejman (2021) Integration of Heuristic Multi-Agent Protection System into Distribution Network Interconnected with Distributed Energy Resources. PhD thesis, Victoria University.

Abstract

Power system operation is undergoing rapid changes due to market deregulations and interconnection of Distributed Energy Resources (DERs) such as wind power and solar panels. The power flow complexities arising from interconnection of DERs into the distribution network have adverse effect on protection systems which can degrade the reliability and power quality in power systems and lead to cascading failures or blackouts. However, with the prospect of integrating Information and Communication Technologies (ICT) infrastructures into power system operation, the ability to utilize advanced protection strategies has become realizable. Given the size and complexities in operation of the future power systems the need for distributed and resilient protection system is inevitable to address the difficulties and inaccuracies introduced in protection settings. Multi-Agent Systems (MAS), as a branch of Distributed Artificial Intelligence (DAI) are capable to deal with complex and large scale systems such as power distribution networks. Moreover, in MAS agents can be deployed in power system to engage with interdependencies between various components while pursuing global goals through supervisory function or behaviours specific to each agent types. In this research, a Multi-Agent Protection System (MAPS) consists of different agent types with certain tasks has been developed to effectively cooperate with other Intelligent Electronic Devices (IEDs) within the protection communication network. A heuristic approach based on exchanging information between different IEDs in the system is utilized to adjust the IED settings according to fault current level in the protected zone. Additionally, to validate the outcomes of the research under real-world scenario, an experiment setup based on Power Hardware in the Loop (PHIL) methodology has been developed to verify the outcomes of the research. The simulation results are discussed to emphasise MAS as a distributed and scalable approach to deal with complexities in future power systems and specifically in relation to protection systems which is crucial for reliability and efficiency of the interconnected distribution networks.

Search Google Scholar

Repository staff login