Research On Battery Capacity Testing Method Based On Soft Measurement

At present, pure electric vehicle and hybrid electric vehicle become the trend of thedevelopment of automobile industry in the future when energy is shortage and environmentalpollution problem is serious. Lead-acid battery has the characteristics of great energy and highperformance-cost ratio, so it is widely used as the power source of electric car. The effectivemanagement of power energy becomes the focus of electric car, and the electric vehicleenergy management system is on the basis of the residual capacity of lead acid battery.Therefore，the accurate estimates of the residual capacity of lead-acid battery become thehotspot problems in the research ifeld of electric vehicle. The dissertation puts forward theestimation method of lead-acid battery residual capacity based on the soft measurementtechnology of adaptive fuzzy neural network. The combination of neural network and fuzzyreasoning can play their respective advantages effectively and complement each other. Usingneural network technology to deal with the fuzzy information can generate fuzzy rules andfuzzy membership functions automatically, so the fuzzy inference system becomes adaptivefuzzy inference system. Neural network is used in the fuzzy reasoning, which can greatlywiden the scope and capacity of neural network information processing. The method ofadaptive fuzzy neural network can implement precise estimates of the residual capacity oflead acid battery.Main contents of the dissertation are as follows: Firstly, the dissertation analyzes anddiscusses the various estimation methods of residual capacity of lead-acid battery at home andabroad, and compare the advantages and disadvantages of various estimation methods, thedissertation put forward the soft measurement technology methods based on adaptive fuzzyneural network; Secondly，according to the basic thought of soft measurement technologymodeling，the dissertation built lead-acid battery discharge circuit based on the MATLAB2011version, and realize the data acquisition. According to reusing the new clusteringalgorithm of aiffne propagation clustering to preprocess the collected data, and analysis andresearch this clustering algorithm at the same time; Finally, the dissertation establish theestimation model of lead-acid battery residual capacity by using soft measurement modelingmethod based on the adaptive fuzzy neural network, and a simulation analysis was carried out.Through the simulation analysis, we can see that estimation accuracy is higher and thegeneralization ability of the model is betetr, and it proved that the method of this article iscorrect and feasible.