Energy management of micro-grid using cooperative game theory

Oladejo, Ismaheel Oyeyemi

Energy management of micro-grid using cooperative game theory

Doctoral Thesis

2019

Abstract

Micro-grid (MG) has been introduced as a low voltage and a very small power system connected to a distribution grid through the point of common coupling. It consists of distributed energy resources (DERs) such as solar Photovoltaic (PV), wind turbine, fuel cell, etc.), interconnected load and energy storage sources. It can operate in grid-connected (i.e. when connected to the main grid) or islanded (i.e. when not connected to the main grid) mode. It has an advantage of utilizing low carbon sources and the possibility of its use in the remote local environment, which means that the transmission infrastructures and their associated costs may be deferred. Although there has been a proliferation of optimization methods of energy management in the MG, most of these methods consider self-interest of the players in profit distribution. Moreover, only a few of them consider a fair profit distribution using Nash bargaining solution (NBS) (i.e. when utility function is linear) leading to even profit distribution and high degree of dissatisfaction. For the MG to achieve better economic outcomes, a novel method based on weighted fair energy management among the participants (i.e. building of different types, such as residential buildings, schools, and shops) is proposed. The novelty of the proposed method lies in the new profit sharing method to favour certain participant by assigning a weight to each participant with cooperative game theory (CGT) approach using generalized Nash bargaining solution (GNBS). The proposed approach achieves a fair (reasonable or just) profit allocation with negotiating power indicator. In this work, a case study of six different participant sites is proposed using the CGT method of energy management. The proposed method is able to cope with the drawbacks of the existing independent method, which negotiate directly with other participants for selfish profit distribution. It is demonstrated that the independent method results in (1) a reduction in the profit of each participant of MG when compared with CGT approach and (2) the variation of transfer prices in some participants having profit below the specified lower bound profit since the method does not take into consideration the lower profit bounds. The use of CGT method (i.e. when participants form a coalition) to finding multi-partner profit level subject to specified lower bounds is demonstrated. This results in (1) increase in the profit of the MG participants (2) maintaining the profit level of all the participants above status-quo profit (lower specified profit bounds) with variation in transfer prices and (3) allowing certain participant to be favoured by assigning higher negotiating power to such participant. To achieve the optimal solution in the proposed method, a teaching-learning-based optimization (TLBO) algorithm is presented to efficiently solve the problem. For TLBO algorithm, no specific control parameters are needed except the number of generations and population size. This is in contrast with other heuristic algorithms such as genetic algorithm (GA) and particle swarm optimization (PSO) that require other control parameters (i.e. GA requires selection and crossover operation, while PSO makes use of social parameters and cognitive weight). To demonstrate the effectiveness of the proposed TLBO method, the profit allocations are tested in the grid-connected and the islanded mode using both the CGT and the independent method. In this work, the proposed TLBO method is compared with one traditional method, i.e. Lambda iteration method and two heuristic methods, i.e. PSO and GA. Thus, by using TLBO a considerable amount of computation time is saved. Using the same parameter setting for all the heuristic algorithms used, 20 trials are performed to be able to compare the quality of solution and convergence characteristics. The investigation reveals that TLBO gives the highest quality solutions and better convergence characteristics compared to PSO and GA.

Keywords

Electrical Engineering

Reference:

Collections

PhD / Doctoral

Full item page