Design Of Photovoltaic DC Charging Pile And Prediction Of Remaining Quantity Of Energy Storage Battery

Increasingly serious environmental problems have triggered mankind's urgent need for clean energy.Electric vehicles have emerged and developed rapidly.At the same time,it brought an urgent need for charging piles.At present,traditional charging piles are frequently connected to the power grid,which has an impact on the stability and power quality of the power grid.The emergence of photovoltaic charging piles not only solves the above problems,but also avoids the waste of cable resources and land resources.It can achieve independent operation with the help of energy storage batteries.Therefore,in an independent photovoltaic power generation system,an energy storage battery that stores energy is an indispensable component.The energy storage battery is affected by factors such as environmental temperature changes,overcharge and discharge,etc.,which shortens its life.In the case of no improvement in the production technology of energy storage batteries,in order to prolong the service life of energy storage batteries,it is necessary to effectively prevent overcharge or discharge of energy storage batteries.Therefore,the central issue of energy storage batteries is concentrated on accurately determining the remaining battery(State of Charge,SOC).This thesis takes the photovoltaic DC charging pile as the research object and completes the determination of the overall structure of the photovoltaic DC charging pile system.The structure is composed of a photovoltaic system,a control system and a protection system,and the photovoltaic system is designed in detail.The working principle and output characteristics of the photovoltaic array in the photovoltaic system are analyzed.A unidirectional converter is used to achieve maximum power tracking,and a bidirectional DC/DC converter is used to achieve a bidirectional flow of energy between the energy storage battery and the DC bus.The charging method of the energy storage battery is analyzed,the MPPT function is combined with the remaining amount of the energy storage battery,and a three-stage charging method is used to charge the energy storage battery.The remaining battery capacity prediction method is introduced and compared.In the first stage,the maximum power charging is performed to make the energy storage battery capacity reach 70%?80%.In the second stage,constant voltage charging is charged into the remaining 20%?30%.In the final stage,the floating charge is used to compensate for the self-discharge of the energy storage battery.The KELM method in the neural network method is used to predict the remaining capacity of the energy storage battery,and it is compared with the battery SOC estimated by the ELM algorithm.At the end of this thesis,the three parts of photovoltaic array,maximum power tracking and bidirectional DC-DC were used for simulation.The Simulink simulation system in Matlab was used to verify the rationality of the experimental design.Matlab software and public lithium ion battery charge and discharge data are used toestablish a nuclear limit learning machine model,set appropriate operating parameters,and constitute the skeleton ofthe prediction model.Then through multiple tests and error analysis of the test set.The results show that the prediction results and error range of KELM neural network are better than those of ELM neural network.This shows that the remaining energy storage battery prediction method described in this paper has a good prediction effect.