Research On Ultra-short-Term Load Forecasting And Optimal Dispatching Of Microgrid

In recent years,the problems of global environmental pollution and energy shortage have become increasingly serious,and countries around the world are increasing their research efforts on the development and utilization of new energy.Microgrid systems have been widely used due to the advantages of high utilization of renewable energy and flexible control.However,the power of distributed power and user loads in the microgrid system has strong uncertainty.In order to ensure the reliability of the power supply in the microgrid,ultra-short-term prediction of the power load is required to obtain more accurate data.At the same time,with the wide application of microgrid systems,how to achieve economic optimality on the basis of safe and stable operation has become a key issue in microgrid research.In view of the above problems,this paper conducts in-depth research on ultra-short-term load forecasting and multi-time-scale optimal scheduling of microgrids.The main research contents are as follows:(1)Aiming at the problem of low load forecasting accuracy in microgrid,a long-short-term memory neural network ultra-short-term load forecasting model based on particle swarm optimization is constructed.The particle swarm algorithm is used to optimize the important parameters in the neural network.The process of adjusting parameters is changed to the process of self-optimization by heuristic algorithm.The highest temperature,average temperature,lowest temperature,relative humidity,precipitation,and day type are the main influencing factors of load through correlation analysis,and they are used together with historical load data as the input variables of the model.The prediction effect of the neural network is tested in combination with an actual example.The results show that the long-short-term memory neural network based on particle swarm optimization can improve the prediction accuracy and save the time for parameter adjustment.(2)Aiming at the problem that particle swarm optimization algorithm has limited the ability of searching in microgrid optimization and scheduling,and it is easy to fall into local optimum,an improved immune particle swarm optimization algorithm is designed.On the basis of the immune particle swarm algorithm,the inertia weight of linear differential decrease and the learning factor of linear adjustment are introduced,so that the algorithm has a good search ability in the whole iterative process.According to a local microgrid system,a microgrid simulation model including photovoltaic arrays,wind turbines,micro-turbines and battery energy storage systems is established,and the algorithm is tested with the example data,compared with particle swarm optimization and immune particle swarm optimization,and improved The immune particle swarm optimization algorithm effectively improves the economic optimization effect of microgrid.(3)Aiming at the problem that the short-term prediction error of the day-a-day optimal scheduling is large and cannot deal with the impact of random events on the load,resulting in a certain error in the scheduling results,a multi-time-scale optimal scheduling strategy for microgrids is formulated,including two time scales:day-ahead and intra-day.A real example in a certain place is selected for simulation verification,and the output of each distributed power source is reasonably adjusted through intraday optimization,and the deviation of the scheduling plan before the day is corrected,which further improves the stability and optimization effect of the microgrid.Based on the above,this paper uses the long short-term memory neural network based on particle swarm optimization to perform ultra-short-term prediction of load,and uses the improved immune particle swarm algorithm to optimize scheduling according to the predicted data,and formulates a multi-time-scale optimization scheduling strategy.The simulation of the example verifies the feasibility of the dispatching strategy,effectively improves the economy and reliability of the microgrid,and provides a certain reference for the optimal operation of the microgrid.