Research On Hybrid Energy Storage To Stabilize Prediction Error Of Photovoltaic Power Generation

Compared with traditional energy sources such as coal,photovoltaic energy has the advantages of cleanness,no pollution and zero emission of harmful gases.However,there are some limitations in the use of photovoltaic energy,which is deeply influenced by weather and other factors.Because of the variability of weather information,the prediction results of photovoltaic output are not ideal.The instability of photovoltaic power generation will bring danger and challenge to large-scale grid-connected photovoltaic power generation and power grid scheduling.To realize the universal use of photovoltaic power generation,photovoltaic power generation system and energy storage technology can be combined while optimizing photovoltaic power generation technology,and the prediction error in photovoltaic power generation prediction can be compensated by energy storage devices to ensure the stability of photovoltaic power generation and realize large-scale grid-connected photovoltaic power generation,thus improving the utilization rate of photovoltaic energy.The application of energy storage technology in photovoltaic power generation forecasting system makes use of its own storable and controllable characteristics to compensate the power error in photovoltaic power generation forecasting system and stabilize the error between the predicted value and the actual value of photovoltaic power generation.However,the energy density and power density of a single energy storage unit do not meet the energy storage requirements of the power grid.Therefore,the hybrid energy storage system composed of supercapacitors and batteries can not only solve the problems of insufficient service life and capacity loss of batteries due to frequent charging and discharging,but also give full play to the advantages of fast charging and discharging power and long service life of supercapacitors.The main research process and achievements of this paper are as follows:Firstly,in order to predict photovoltaic power accurately and quickly,a short-term photovoltaic power prediction method based on Cloud Genetic Algorithm(CGA)optimizing initial weights and thresholds in BP neural network is proposed.Then,according to the weather characteristics,similar days are selected,and the crossover probability and mutation probability in genetic algorithm are adjusted adaptively.The weights and thresholds optimized by CGA are assigned to BP neural network,and the optimized values are used as initial weights and thresholds to construct a brand-new prediction model.Simulation results show that compared with the other two simulation models,CGA-BP neural network model is used to predict photovoltaic output,and its accuracy is improved,which greatly reduces the error of photovoltaic power prediction.Secondly,according to the analysis of the error distribution characteristics of photovoltaic output prediction,the optimized prediction model still has some limitations,and there is still a gap between the predicted value and the actual value,so it is proposed to add a hybrid energy storage device to stabilize the errors in photovoltaic power generation prediction.According to the predicted power error,the spectral characteristics of the photovoltaic output prediction error are obtained by fast Fourier transform.According to the spectrum characteristics,the power error is divided into high-frequency part and low-frequency part by high-pass filter,and then according to the characteristics of each energy storage device,the high-frequency error is distributed to the supercapacitor,which is charged and discharged,and the remaining part is taken charge of by the storage battery,thus making reasonable capacity allocation for the energy storage battery.Then,in order to prevent energy waste and economic loss caused by unreasonable distribution of charging and discharging power in hybrid energy storage system,fuzzy control is adopted to control the charging and discharging power of hybrid energy storage system according to its non-linearity,so as to stabilize the error value of photovoltaic output prediction.Simulation results show that the output power accuracy of the system with hybrid energy storage is higher than that of the photovoltaic system with only photovoltaic output prediction.Finally,LabVIWE programming software is used to realize the software platform of hybrid energy storage to stabilize photovoltaic power prediction error,which improves the practical application value and provides convenience for operators.