Multi leader particle swarm optimization for optimal placement and sizing of multiple distributed generation for a micro grid

• Main Author: Ariya Sinhalage Buddhika Eshan Karunarathne
• Format: text::Thesis
• Language: English
• Published: 2023

The growing concerns on climate change, depletion of fossil fuel, and few other reverberations such as technical, economic, and environmental issues have erupted with the deployment of centralized power stations. In order to tackle and overcome these problems, the distributed generators (DG) have been introduced to the power systems as an integral remedy. The DGs could be integrated at any bus in the power system to assist the demand locally while enabling a portion of power to transfer to another bus in the system. However, the DGs have to be strategically sized and placed in the distribution system to maximize their benefits, since the unplanned and inappropriate integration of DGs exposes the distribution system to severe challenges. Existing literature has been proposed techniques to optimally size and place the DGs in a distribution system. However, these techniques are weaker in active power loss reduction and still suffer from several drawbacks such as high computational burden and complexity, difficulty in implementation, and premature convergence. Thus, the solutions to these techniques might not be able to leverage the benefits of DGs. Hence, the aim of this thesis was to investigate the impacts of DG integrations into the distribution systems in terms of active power loss and voltage variation and to develop an optimization technique to overcome the aforementioned drawbacks along with maximizing the benefits concerning the active power loss reduction and voltage variation of the system. In this thesis, detailed network models of two medium voltage distribution systems have been developed considering the topological and electrical parameters of the systems. A comprehensive analysis was performed on a developed system in order to demonstrate the effect of a single and multitude of DGs on active power loss and the variation of the voltage profile. In addition, the influence of multiple DG integrations was evaluated focusing on successive DG integrations. A voltage stability index-based approach was adopted to establish the optimal locations of the DGs in the system. Initially, the Particle Swarm Optimization was implemented to establish the optimal sizes of DGs and the performance of the implemented algorithm was analyzed and quantified. Thereafter, the Multi-Leader Particle Swarm Optimization algorithm (MLPSO), which is a novel evolutionary optimization technique in the field of power systems was developed and employed in the optimization process. This algorithm is capable of surmounting the aforementioned drawbacks especially premature convergence, through its reward-based dynamic leader assignment and self-learning strategies. This optimization framework was carried out to identify the optimal sizes and placements of multiple DGs while contemplating the active power loss minimization of the system as the objective function. In addition, the solutions have been evaluated based on pre-defined performance metrics and the outcomes of the optimization framework were compared with the other existing optimization techniques to evaluate the potency and the productivity of the developed MLPSO algorithm. The findings reveal that the formulated MLPSO methodology could substantially increase the percentage of active power loss reduction while satisfying the statutory voltage limits in both IEEE 33 bus system and real Malaysian bus system up to 67.40% and 80.32% respectively. Ultimately, this framework can assist the distribution network operators to optimally locate and size DGs to minimize the active power loss in the distribution network.