Research On Lithium Battery Power Estimation Method Based On Improved Recursive Principal Component Analysis

In the process of charging and discharging lithium battery,the estimation of peak power can evaluate the ability of lithium battery in the limit charge and discharge,and avoid the phenomenon of overcharge and overdischarge,which has an important impact on the rational use of lithium battery.Therefore,the research on the accurate estimation of the state of power of the lithium battery is extremely important for the efficient operation and safety of the lithium battery.Since the power characteristics of the lithium battery are related to parameters such as battery voltage,current,temperature,DC internal resistance,and SOC,a battery performance data acquisition system is designed to perform on-line detection of battery parameters.The hardware circuit of the system uses STM32 as the main control chip.Design voltage acquisition circuit,current acquisition circuit,temperature acquisition circuit and data communication circuit.Also design filter circuit to improve the anti-interference ability and stability of the circuit.The RS-485 bus is used to communicate with the host computer to realize real-time collection and transmission of current,voltage and temperature information of the lithium battery.Through verification and analysis of the test,the accuracies of voltage and current acquisition meet the design requirements.By improving the Thevenin equivalent circuit model,the battery equivalent circuit model is established,and the battery performance parameters voltage,current and SOC mathematical expressions are obtained.The validity of the model is verified by experiments and simulations.By testing the performance of the battery,the effects of battery discharge rate,temperature,DC internal resistance and other parameters on the SOC of the battery are analyzed.The principal component analysis method is used to determine the principal element factors affecting the battery SOC,and the battery SOC estimation model is established.The estimation model is verified by simulation and its estimation accuracy is 5.25%.In order to further improve the estimation accuracy of battery SOC,a recursive kernel principal component analysis algorithm is proposed to establish a battery SOC estimation model,and its accuracy is up to 3.64% by simulation comparison.Analyze and calculate the peak power under different conditions of voltage,current and SOC constraints,and determine the estimated model of lithium battery SOP by deriving several peak powers.The simulation proves that the estimation accuracy of the recursive kernel principal component analysis can reach 6%-18.5%.Finally,the SOP algorithm is transplanted into the upper computer program,and the design of lithium battery power test system is finished by combining the battery performance data acquisition system.It proves that the system can realize the estimation of battery SOP through the running verification of the system.