Energy Optimization Scheduling Of Islanded Microgrid System

The construction of islanded microgrid can not only meet the local load demand,but also alleviate the pressure of traditional energy crisis and environmental pollution by using clean energy such as scenery.However,compared with the large grid,due to the randomness and volatility of photovoltaic power generation and wind power generation,the islanded microgrid is relatively weak in the ability to withstand disturbances and faces higher operational risks.Therefore,improving the prediction accuracy of photovoltaic power generation and wind power generation will provide more accurate data for the safe and stable operation of islanded microgrid,improve the consumption level of new energy power generation and reduce the operation cost of islanded microgrid.In addition,because the energy optimization scheduling problem of islanded microgrid has typical multi-time scale characteristics,on the basis of obtaining high precision prediction data,reasonable multi-time scale energy optimization scheduling of controllable resources in microgrid can effectively improve the economy and stability of microgrid system.In this paper,the energy optimization scheduling technology of islanded microgrid system is studied.The main research contents are as follows:(1)Based on the introduction and analysis of the topology and energy scheduling architecture of islanded microgrid system,the energy scheduling model of islanded microgrid system is established,focusing on the detailed mathematical modeling of micro gas turbine,battery energy storage system and flexible load,which lays the foundation for the follow-up work of the paper.(2)The problems of wind power generation and photovoltaic power generation in microgrid are studied.Firstly,a comprehensive similarity day selection method based on difference similarity and gradient similarity is proposed.Secondly,the traditional particle swarm optimization method is combined with the LSTM network to iteratively optimize the parameter initialization method of the LSTM neural network.Finally,the PSO-LSTM power generation prediction model is built based on the Pytorch framework,and the effectiveness of the proposed model is verified by comparing the prediction error of the model on the test set.(3)A day-ahead-day two-stage energy optimization scheduling framework for islanded microgrid system is proposed,which combines day-ahead planned scheduling with day-ahead MPC rolling scheduling.In the day-ahead time scale,considering the adjustable resources of power supply,energy storage and flexible load,a day-ahead optimal scheduling model considering operation cost and decision penalty is established.Based on the model predictive control method,the basic framework of intraday rolling is constructed in the intraday time scale.At the same time,on the basis of the day-ahead optimal scheduling model,the modified penalty terms such as the state of charge of battery energy storage,the interruptible load interruption plan and the start-stop plan of micro gas turbine are introduced to coordinate the locality of intraday scheduling with the globality of day-ahead scheduling.Finally,an example is given to verify the effectiveness of the proposed two-stage energy optimization scheduling strategy for islanded microgrid.