Research On Management Technology Of Lithium Battery Remaining Power In Electric Vehicle

With the increasing of environmental pollution and the appearance of energyshortage, the pure electric vehicle, regarding as the green and environment vehicle, iscoming the hot spot of research for various countries’ automobile manufacturers.Lithium iron phosphate battery with long life, low cost, high effective, safety, rapidcharging rate and environmental protection, becomes one of the best choices. Theaccurate predicting of the status of battery is the key in the battery management system.This research comes from the national "863" project, mainly studying the forecastingmethod of the remaining power of lithium ion battery. And the value of SOC isdescribed the number of remaining power.This article focuses on analysis of electric vehicles, especially pure electricvehicles and the battery management system existence subject matter. Embarks from theiron phosphate lithium power battery’s electrochemistry characteristic, with itslaboratory static environment charger and discharge data as the basis, analysis of theinfluence of lithium ion battery state of charge precise predict factors, and has carriedon the definition to the current time battery’s state of charge.Profits from the domestically produced of some pure electric automobile technicalparameter and the target request, measured for60Ah lithium iron phosphate batterycharge and discharge static laboratory data. It has established the complete vehicleSimulink simulation model using the ADVISOR automobile simulation software, and inthe domestic and foreign8kind of typical state of roads carry on the combination underthe condition, gained the lithium ion battery voltage, current, temperature when theydischarged, using the Ah method to determine the SOC values under the condition ofanti-error-accumulation.Based on this, in view of the tradition forecast technique’s insufficiency, hasdesigned the intelligent algorithm which one kind of the heredity-support vectorregression algorithm (GA-SVR). Using European light car test state of roads NEDC andJapanese Prius automobile test state of roads1015, separately has carried on thesimulation to the SVR algorithm and the GA-SVR algorithm, and has carried on thecontrastive analysis to the MSE and R parameters, satisfie the request of MSE≤0.3%，R ≥90%，for the complete vehicles energy’s intelligence forecast that has laid therationale.The research has built takes the master control chip by MSP430the lithium ionbattery management system platform, has carried on the laboratory direct currentmachine bench test and gathering data, to mixed the algorithm the feasibility and theversatility has carried on the confirmation. The research indicated that, GA-SVRalgorithm has a battery SOC prediction accuracy, robustness and so on.