Advanced materials research · 2011 · 41 citations · 9 references
EngineeringAutomotive EngineeringElectrical EngineeringElectric Vehicle EngineeringHybrid Vehicle SystemsMachine LearningBattery SystemsBattery TechnologyKernel MethodsState EstimationState Of Charge EstimationSupport Vector MachinesAccuracyBattery State Of Charge ModelingHybrid Electric VehiclesLeast Squares Support Vector MachinesMachine Learning For State Of Charge EstimationState Of Charge MonitoringState Of Charge PredictionVoltage
This paper presents a new method to estimate the state of charge (SOC) of Ni-MH battery pack in hybrid electric vehicles (HEV). The proposed method establishes the relationship of the SOC to the battery’s voltage, current and temperature by using least square support vector machines (LS-SVM). According to the nonlinear characteristics of a battery pack system, the nonlinear SVM with polynomial kernel are developed for the estimation of the SOC with LS-SVM algorithm. To be more efficient in application, this method is also simplified in this paper. The results have conformed that the proposed method is able to estimate the SOC of Ni-MH battery with high accuracy and noise tolerating ability.
9