Optimization Of Microgrid System Considering Wind And Solar Power Forecast

The gradually exhausted fossil fuel and pollution to environment and economy make renewable energy utilization a more urgent and vital task.Therefore,distributed power generation and microgrid technology have attracted increasing attention and have been widely applied because of their unique advantages.As the key part of smart grid,advanced technologies of distributed power generation and microgrid system will be quickly developing and more widely applied in the future.This thesis first performs a literature survey on wind and PV energy forecast and microgrid system economic dispatch.Then an improved forecast model for wind and PV energy is proposed by blending the Long Short-Term Memory(LSTM)with Support Vector Regression(SVR),to improve the accuracy of output forecast of wind and PV energy.In addition,the daily dispatch model is designed for the microgrid system for renewable energy coordination and microgrid economic dispatch by using the intelligent optimized algorithm,for system cost minimization.The key contributions of this work as follows.Firstly,based on the analysis on background of microgrid,the wind-PV power forecast model is introduced.By using the real wind-PV data in an Australian area,the wind-PV forecast scheme is proposed.Moreover,the key modules of microgrid and their mathematical expressions are provided to formulate the dispatch model and economic optimization.Secondly,three typical methods,i.e.,the support vector machine,BP network,and Long Short-Term Memory network are used for wind-PV power forecast.We then blend the LSTM with SVR to formulate a novel hybrid forecast method to improve the prediction accuracy.Comparative studies are performed for the four methods to demonstrate the superiorities of the LSTM-SVR method.Finally,to tackle the deficit that typical Cuckoo Searching Algorithm may fall into the local optima with insufficient mechanism of communication and coordination,a hybrid Cuckoo Searching Algorithm is proposed for microgrid power dispatch optimization.In simulations,using the predicted wind-PV power output,daily dispatch optimization is realized,demonstrating the advantages of the improved Cuckoo algorithm.