Ni-MH Battery SOC Estimation Based On Intelligent Algorithm

The working status of the power battery during charge and discharge is an important part of the development of the electric energy vehicle.The NiMH battery was used in this paper.The following items are specifically developed.The first is the transmission and acceptance of test data.This experiment is based on the wireless charging platform,so we chose the wireless network;the collected charge and discharge data are used to screen the equivalent circuit model,and the second-order RC circuit model is determined through data analysis;then the model parameters are completed.Real-time estimation;based on model parameters,battery SOC estimation is achieved.Faced with the construction and data transmission of wireless network,ZigBee technology chosed from the commonly wireless technology is the application technology in this article.ZigBee technology is mainly used in the short-range wireless connection.It can complete the establishment of a wireless network between multiple nodes and achieve wireless communication between nodes.This article chooses ZigBee technology to complete the data receiving on the secondary side of the wireless charging system.The data is charging current and the battery terminal voltage of the battery,and then the data can be used to analyze and process.The mathematical model of Ni-MH battery is the equivalent circuit model.The most important part of the equivalent circuit model is the RC link,which is the selection of the number of RC links.We chose three models to be candidate models.The experimental data of the battery are used to identify the parameters of the model and the performance is compared.From the calculation amount and the estimation accuracy,we select the appropriate model.The evaluation standard of the equivalent circuit based on the terminal voltage error was established,and the model structure for the NiMH battery was specifically determined.The parameter identification of the power battery model,based on the data obtained from the wireless network,combined with the established circuit model,using a recursive least squares method to estimate the circuit model parameters under constant current charge and discharge conditions in real time,and the terminal voltage of the battery is an algorithm.In the feedback loop,the output voltage value of the model is very good to track and estimate the experimental value,and then the polarization voltage and polarization capacitance of the equivalent circuit of the power battery at each moment are calculated.Power battery model according to the data gained by the wireless network,combined with the established circuit model,using recursive least squares method and KF algorithm of constant current charge and discharge condition and working condition of converter circuit model parameters for real-time estimation,voltage of the battery as the feedback part of algorithm,the model of output voltage to track to the experimental value estimation,and then calculate the power battery in every moment of the equivalent circuit of the polarization resistance and capacitance value.Under the precondition of solving the model parameters in real-time,in order to get accurate SOC estimation requires the accurate real-time model parameters,using double kalman filtering method(DKF)for power battery model parameters for real-time estimation and state of charge at the same time.After analyzing the estimation error of SOC,the genetic optimization double kalman filter(GA-DKF)is proposed to optimize the process noise and measurement noise to obtain a more accurate estimate of charge state.