Method for estimating SOC of vehicle battery by using unscented Kalman filter based on state detection mechanism

Abstract

The invention discloses a method for estimating an SOC of a vehicle battery by using unscented Kalman filter based on a state detection mechanism, and belongs to the technical field of vehicle power batteries. The method comprises the following steps: establishing a battery equivalent model, and determining a model equation of the vehicle battery by combining a composite model method with an ampere-hour method; performing model parameter identification, identifying relevant parameters of a battery model observation model of by a recursive least squares method, wherein the system input quantityis persistent excitation, and identifying the number of iterations, so that a final result of the parameters is converged and tends to be stable; and accurately estimating the SOC of the vehicle battery in real time, and estimating the SOC of the battery by adopting the unscented Kalman filter of the state detection mechanism based on residual information characteristics. According to the methodfor estimating the SOC of the vehicle battery by using the unscented Kalman filter based on the state detection mechanism provided by the invention, observed residual information is used as the basis,and abnormalities are detected with the state detection condition; under abnormal conditions, a self-adaptive attenuation factor is adopted to adaptively correct an associated covariance matrix of the unscented Kalman filter, so that the problems that the tracking convergence speed is low and the robust ability is weak are solved.

Description

technical field [0001] The invention relates to a vehicle battery SOC estimation method, in particular to an unscented Kalman filter vehicle battery SOC estimation method based on a state detection mechanism, which belongs to the technical field of vehicle power batteries. Background technique [0002] At present, in view of the problem of poor SOC estimation accuracy of vehicle power batteries, the unscented Kalman filter algorithm aims at the nonlinear characteristics of the battery, and uses the second-order or higher accuracy to approximate the posterior mean and variance of the Gaussian non-linear system state to improve the SOC estimation accuracy. And the convergence is good, which is a common solution for battery SOC estimation at present. However, when the algorithm has model errors, state initial value setting errors, or state mutations, it is easy to lead to poor robustness of the algorithm and estimation accuracy. Problems with reduced and reduced tracking capabi...

AI Technical Summary

Problems solved by technology

[0007] The main purpose of the present invention is to provide an unscented Kalman filter vehicle battery SOC estimation method based on the state detection mechanism, based on the observation residual information, the state detection conditions are abnormal, and the adaptive attenuation factor is used in abnormal cases Correct the correlation covariance matrix of the unscented Kalman filter to solve the problems of slow tracking convergence and weak tolerance

Images