The Study On SOC Estimation Strategies Of VRLA Batteries

The battery has a lot of advantages with large capacity, low cost, good safety, mature technology, abundant raw materials and maintenance free use that is both mobile and protable. Batteries will be widely used in the future development of Photovoltaic energy and wind powered electrical generation, uninterruptible power system (UPS), future lighting and electrical powered vehicle which are the most widely used fields for the two types of batteries for which there have not been any other kinds of batteries to replace their function completely. In order to improve battery’s efficiency and prolong their energy life effectiveness we must estimate the capacity or the state of charge (SOC) for batteries. SOC estimations need to be established the equivalent circuit model of batteries and selected appropriate methods. This paper focuses on the estimations of VRLA and SOC batteries in the following aspects:First by analyzing the working principles and characteristics of VRLA battery on their charge and discharge rate to include their temperature and battery state of health factors on SOC’s. Focusing on the estimation methods of SOC batteries which include the ampere hour method, the electric potential method, the neural network method, the fuzzy method and the Kalman filter algorithm method. In order to compere the several equivalent circuit models of batteries comprehensively to select an improved PNG V model as the focus of the study, which adopt the experiment of voltage response through current excitation in order to identify the parameter of the model using the cftool to fit it.This text adopts two schemes by comparing and analyzing the methods of SOC estimation.One is the improved ampere hour and the electric potential method, because only working currents, charges and discharge can effectively influence the Ampere hour method to make up the electric potential method potential models dependent defects and ANN method of error accumulation which can correct the electric potential method. Therefore the combination of electrical potential with the after correction of the ampere hour method by a weighted parallel structure to estimate the SOC is advantageous of the two methods and are complementary when the SOC estimation accuracy is improved.The other method is the EKF algorithm which according to the improvement of the PNG V model establishing the state and observation of the battery system, in order to determine the calculation process of the EKF for SOC resulting in the minimum variance estimations. In conclusion through using MATLAB off-line simulation calculations come very close to their theoretical values for the development high precision of batteries. EKF in SOC battery calculations have a very broad range of applications, further implementations based on EKF SOC estimation methods of electronic engineering are necessary.