Distribution Network Planning With Distributed Generation Based On Bi-level Optimization Model

Smart grid is the development direction of modern power grid, which requires future distribution network to accept distributed generation (DG) freely. However, DG (particularly intermittent and renewable DG) has the feature of dispersion and randomness. It's large access to distribution networks will affect the whole system's safety and stability, such as power loss, voltage stability, power quality, relay protection, power supply reliability and so on.Based on research status at home and abroad, this paper focuses on studying distribution network expansion planning with DG. The mathematical model of expansion planning considers a variety of factors, such as DG cost, annual operating costs, loss of distribution network, environmental benefits and so on. This paper establishes a multi-objective optimization function based on minimum annual cost. To solve the objective function better, this paper introduces bi-level optimum theory to convert objective function to bi-level planning model:the upper planning etermines the line construction and the lower planning determines the access location and DG capacity. This paper uses improved single parent genetic algorithm to search optimal grid architecture to solute the upper planning's problem, and lower planning introduces a new evolutionary algorithm called cultural algorithm and analyzes details about how to apply the algorithm to solve optimization problem of lower DG.. Finally, verify the feasibility of conversion method about bi-level optimum model proposed in this paper and the effectiveness of the algorithm with numerical examples.