Research On Equalization Topology And State Of Charge Estimation Of Lithium-ion Batteries Based On Neural Network And Filtering Algorithm

At present,countries all over the world are striving to build a low-carbon,environmentally friendly,clean and sustainable modern energy system.The development of renewable energy is continuously increasing and the development of new energy vehicles has begun.Lithium-ion batteries are the mainstream power batteries for new energy vehicles,and due to the trend of echelon utilization of lithium-ion batteries,energy storage power stations will also use a large number of lithium-ion batteries for energy storage.Lithium-ion batteries must be equzalized and managed during use.Although dissipative equalization,the most mature technology,is simple,it consumes a lot of energy during the equalizing process.Non-dissipative equalization can transfer energy between batteries to avoid excessive energy loss.Since the state of charge(SOC)of the battery cannot be directly measured,the non-dissipative equalization technology mostly uses voltage equalization,which will result in low equalization efficiency and loss of energy.In order to further improve the efficiency of battery equalization and reduce the energy loss,this paper studies the lithium-ion battery equlization technology with SOC equlization.Starting from the battery equalization topology and SOC estimation,this paper carries out on the Li Co O₂ battery and Li Ni Co Al O₂ battery with the highest energy density on the market,and the following are carried out the study:First of all,in view of the problem that the current research on the lithium-ion battery equalization topology cannot take into account the equalization circuit volume and the equalization speed at the same time,three equalization topology based on the principle of overall planning is proposed.By comparing the topology,the number of energy storage components and the equalizaion speed with the traditional equalization topology,it is proved that the proposed equalization topology can maintain the equalization speed while reducing the number of energy storage components.Then,to solve the problem that the same equalization topology cannot adapt to multiple application fields,an equalization topology based on the principle of adaptability is proposed.Most topologies have the most suitable application fields.Once the application fields are changed,manufacturers often need to re-select another equalization topology,which will greatly increase the workload of the entire battery management system development.The equalization topology based on the adaptability principle can select the appropriate number of energy storage components according to different application fields,and then flexibly adjust the volume of the equalization topology to adapt to different application fields.Afterwards,in view of the problem of the lithium-ion battery equivalent circuit model,a linear neural network equivalent circuit model based on the linearization of the terminal voltage rebound curve is proposed.The open circuit voltage(OCV)equation is established by ignoring the curve part of the terminal voltage rebound curve.A linear neural network is used to construct an equivalent circuit model,combined with the OCV-SOC relationship established by the Gaussian trinomial equation,and finally a simple SOC estimation method based on linearized equivalent model is realized.It is proved in experiments that the simple SOC estimation method based on linearized equivalent model can realize the equalization management with SOC equalization,and has less energy loss than the equalization with voltage equalization.Finally,aiming at further improving the estimation accuracy of SOC,an adaptive extended Kalman filter algorithm based on fuzzy algorithm is proposed.The algorithm uses fuzzy logic to switch between the second-order RC model and the third-order RC model,thereby improving the SOC estimation accuracy of the adaptive extended Kalman filter algorithm based on the second-order RC model.Experiments show that the proposed algorithm can achieve equalized management with SOC equalization,and has less energy consumption than the simple SOC estimation method based on linearized equivalent model proposed before.Based on the above work,this paper realizes the lithium-ion battery equalization technology with SOC equlization,so that the battery equalization technology can adapt to various fields while reducing the energy loss in the equalization process.The research results in this paper are easy to implement in engineering and can be applied to fields such as energy storage power stations and electric vehicles.