An Optimal Strategy About Economical Operation Of Microgrid Based On Dynamic Forecasting Of Power And Load

With power demand growing, the traditional power grid whose critical features arelarge capacity of generator, large-scale system and high voltage has been the majorpower supplier in the past few years. However, many problems in the power systemhave became serious challenges faced by the power industry in the world, such asincreasing load, aging equipment, increasing demands of power quality and reliability,energy efficiency and environment problem. Facing the series of serious challenges, anew development mode in power system has been developed by distributed generationas its advantages of flexible, efficient, economical and environmental operation anddiverse energy. Microgrid is an aggregation of distributed generation and loadsoperating as a single system which makes most of renewable energy and improves thepower quality and reliability. The economic operation of microgrid is one of the keyissues of microgrid. A better economic operation strategy and accurate prediction of theoutput power of the uncontrolled microsources and load in microgrid are the basis ofpopularizing the microgrid technology.In this paper, the economic operation strategies based on the wind power forecastand microgrid forecast in connection mode and island mode are developed. On the levelof microsource, distributed generation is divided into controlled microsource anduncontrolled microsource. Controlled microsource is able to adjust its output power andthe output power of the uncontrolled microsource varies randomly with time due to thevariability of weather. As a typical uncontrolled microsource, the short dynamicprediction model of the wind power is developed. The major factors affecting windpower are analyzed and a radial basis function neural network with error discriminantfunction is proposed in this paper. Based on the neural network, the short dynamicprediction model of the wind power output is modeled and it has been found that windpower can be successfully predicted using this model. On the level of load in microgrid,the major factors affecting load in microgrid are analyzed and the short dynamicprediction model using the neural network with error discriminant function is developedand it has been found that load in microgrid can be accurately predicted.On the system operation level of microgrid, the research on the economic operationstrategies in connection and island modes is carried out. In connection mode,considering the cost of operation and maintenance of microsources and environmental cost and the electricity price, the economic operation strategies in different typical daysof different seasons are proposed based on the wind power prediction and loadprediction. And in island mode, the economic operation strategies in different typicaldays of different seasons are proposed based on the wind power prediction and sensitiveload prediction considering the cost of operation and maintenance of microsources andenvironmental cost.Finally, a simulation platform of microgrid economic operation is built based onMatlab software. The wind power prediction and load prediction are carried out in theplatform. Based on the simulation, the economic operation strategies in connectionmode and island mode are simulated to prove their validity and the simulation resultsare discussed in details.