Battery parameter identification method and system based on robust recursive least squares

Abstract

This disclosure proposes a battery parameter identification method and system based on robust recursive least squares. The method includes the following steps: establishing the battery system equation; determining the relationship between the open circuit voltage and the remaining power of the battery offline; Based on the battery current and voltage data and the open circuit voltage of the battery, the least squares iterative solution is used based on the robust loss function and the adaptive forgetting factor, and the battery parameter estimates based on the robust recursive least squares method are obtained. By setting a robust loss function, in the iterative process, the delay noise problem is generated according to the interference caused by the asynchronous delay of the acquisition, and the adaptive abnormal data judgment limit is established, thereby reducing the impact of the acquisition delay on the identification results, and at the same time increasing the deviation compensation link To reduce the influence of measurement noise and improve the accuracy of parameter identification.

Description

technical field [0001] The present disclosure relates to the technical field related to vehicle batteries, and in particular, relates to a battery parameter identification method and system based on robust recursive least squares. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art. [0003] The online parameter identification of the power battery model can track the changes of the battery model parameters in real time, so as to provide an accurate battery model for the model-based state estimation algorithm, thereby improving the accuracy of state estimation. Among many power battery models, such as equivalent circuit model, data-driven model, electrochemical model, etc., the equivalent circuit model is the most suitable battery model for actual vehicle applications because of its relatively low complexity and relatively appropriate accuracy. [000...

AI Technical Summary

Problems solved by technology

Sensitivity analysis experiments show that 5mV acquisition noise will cause 6%-11% identification error of different model parameters

[0011] 2. In the battery management system, the asynchronous phenomenon of voltage and current acquisition has a serious impact

For the existing linear minimum variance class identification algorithm, this problem will introduce a large prediction error in each iteration of the algorithm, that is, the residual error in the parameter identification process is too large

At the same time, the RC link in the equivalent circuit model has a relatively large time constant, resulting in large fluctuations in the identification parameters, and in extreme cases, it will also cause the failure of the identification algorithm

[0013] To sum up, although the existing least squares identification algorithm has the advantages of iterative operation and small amount of calculation, it is very suitable for online applications, but it cannot effectively suppress the two problems of measurement noise and acquisition asynchrony in practical applications. question

Images