The paper describes a technique for estimating the parameters of the second-order electrical model of lithium battery (LB), based on the analysis of its response to a step-like load change. The advantage of this approach is the implementation with simple and compact measuring equipment. Taking into account the limitations for the maximum LB load, the solution of a system of nonlinear equations when estimating the model parameters can be reduced to solving a system of linear algebraic equations. This finding makes it possible to implement the approach with inexpensive general-purpose microcontrollers. Thus, the paper proposes a method for reducing computational costs, which consists in sequential fitting of experimental relaxation curves with two single-exponential models instead of using a complex two-exponential model. It provides more than twofold reduction in the time for identifying the parameters of the LB’s electrical model. The experimental studies show that the obtained dependences of the equivalent circuit parameters can be explained by the general properties of physicochemical reactions occurring in LB. From a practical point of view, the proposed method has great advantages. First, it has better time localization compared to electrochemical impedance spectrometry. This advantage is especially important for LB due to the effect of the discharge processes on the results. In addition, the response time of diagnostic systems is significantly reduced. Secondly, the proposed method requires simpler and more inexpensive test equipment, which potentially makes it possible to integrate controls into the LB at the manufacturing and exploitation stages.
Read full abstract