Research On Optimal Scheduling And Islanding Division Of Active Distribution Network With Volatility Energy

In the traditional distribution network,the high-permeability access of volatility energy generation such as wind and light forms an active distribution network.The uncertainty of wind and light power generation brings new challenges to the dispatching of power systems.This paper first realizes the transformation of power generation uncertainty to deterministic problem through multi-scene technology.Then the research on distribution network scheduling problem is carried out.The scheduling concerns in this paper mainly include the economic optimization scheduling during normal operation and the island division after failure,and thus improve the economical operation of the power grid and the important load of restoring power supply after the fault.The research on the optimal scheduling and islanding strategy in the active distribution network with wind power generation and photovoltaic power generation carried out the following work:1)The power generation characteristics of wind and photovoltaic are analyzed,including the conversion characteristics of power generation and the uncertainty of active output;then the short-term load forecasting is studied.Firstly,the two parameters of the visual field and the step size are adaptively set to dynamically change,and then the elite reverse learning mechanism is introduced to obtain the improved artificial fish swarm algorithm.The kernel width coefficient and normalization parameters in the least square support vector machine are optimized by the improved algorithm.In this paper,the optimized least squares support vector machine is used to predict the load.2)Economic optimization scheduling of active distribution network: The improved K-means clustering algorithm is used to divide the wind and light power generation scenes.For load uncertainty,the optimal scene division number is determined by information entropy,and the uncertainty to deterministic problem is realized.Then,the multi-time scale optimization scheduling model is established,which is the day-to-day rolling-real-time scheduling model,and propose the corresponding scheduling strategy;the traditional fireworks algorithm is improved by introducing chaotic search,adaptively setting the blast radius and the application of elite retention and elite crossover strategy.The improved fireworks algorithm is used to solve the scheduling model to get the optimal scheduling scheme.3)Active distribution network islanding after fault: Based on solving the uncertainty problem,a multi-period islanding strategy is proposed.Firstly,the exact distributed power output and load prediction data in different scenarios are obtained.Then,under the constraint conditions,the goal of minimizing network loss and restoring the value of the power supply load is established.Finally,using the depth-first search method,a dynamic multi-time islanding scheme is obtained through the "search + adjustment" strategy in the non-faulty area.Finally,the improved PG&E69 node power distribution system is used to simulate and verify the practicability of the proposed strategy.