Photovoltaic Power Forecasting And Ship Microgrid Optimization Scheduling Storage

With the large consumption of fossil energy,the increasing pollution of environmental pollution and the intensification of the greenhouse effect,it is extremely urgent to introduce new energy power generation technologies into microgrids and large power grids.With its abundant reserves,wide distribution,convenient access and simple conversion,solar energy has attracted the attention of photovoltaic power generation technology all over the world.However,the randomness and variability of photovoltaic power make micro-grid optimization scheduling a challenging study.The accuracy of PV power prediction and prediction determines the reliability and feasibility of microgrid optimization scheduling.In order to solve the problem of micro-grid optimization scheduling with optoelectronic power,this paper starts with the research of photovoltaic power prediction and micro-grid optimization scheduling based on prediction information.Firstly,according to the research of photovoltaic power prediction technology at home and abroad,it is concluded that the recurrent neural network plays an important role in the prediction of photovoltaic power due to its good nonlinear approximation ability and memory function.The solar radiation quantity prediction model was constructed by using the cyclic neural network,and its topology and related parameters were optimized.Secondly,based on the characteristics of the solar radiation quantity time series,the shortcomings of the prediction model based on the recurrent neural network are analyzed.In order to solve the shortcomings of this prediction model,this paper introduces the empirical mode decomposition algorithm and Kmeans algorithm,which can improve the prediction accuracy by decomposing and reconstructing the solar radiation time series,and the final reconstruction quantity is discussed.Thirdly,that methods can improve the accuracy of solar radiation prediction is explored.It is found that the Bagging model and the error correction model decomposition improve the accuracy of the prediction model from the two perspectives of sample and error.The Bagging model builds subsample sets in the form of put back samples,highlighting the different individuals to improve the generalization ability and prediction accuracy of the model.The error correction model corrects the predicted value of the model by historical error data,thereby improving the prediction accuracy.Finally,the prediction error of the prediction model is statistically analyzed,and the statistical information of the prediction error is extracted.Based on the prediction error information,the microgrid optimization scheduling model is established.The particle swarm optimization algorithm is used to solve the optimal scheduling problem of ship microgrid based on economic optimization and minimum carbon emission.