The lithium-ion battery has been widely used in distribution energy system and electric vehicles as its features of high density, long cycle, low self-discharge and energy friendliness.
To meet the requirements, hundreds and thousands of cells are always composed through series or parallel to make a battery system.
The state-of-charge as an important parameter for battery operation, which is related to the battery behavior in practice especially for electric vehicles and enables the protection of the battery pack from being over-discharged or over-charged , needs to be precisely predicted.
Unfortunately, even the battery group is rigorously selected in practice, different cells performance will lead to the battery inconsistency in the process of production and usage.
When any cell reaches the cut-off voltage, the charging/discharging process should be stopped to avoid cells over-charging/over-discharging, thus the pack cannot be fully charged/discharged.
The inconsistent cells in the process of production and usage make it difficult to directly estimate battery pack state-of-charge with conventional approaches.
To estimate the battery pack state-of-charge with a simple way, in a study newly published in energy, the researchers give the definition of battery pack state-of-charge, and then a lumped parameter battery pack model is proposed based on the data-driven model using the measured data to update the model parameters.
The algorithm of dual filters is employed to estimate the parameters and state-of-charge concurrently in analyzing the slow-varying parameters of battery pack and fast-varying state-of-charge under different situations. In the algorithm of dual filters, the extend Kalman filter is used to update parameters of battery pack on-line, while the unscented Kalman filter is used to estimate battery pack state-of-charge.
The proposed approach is verified by experiment under constant current condition and dynamic stress test conditions. The experiments results show that the proposed method approximates the true state with the mean absolute percentage error less than 0.04. The relationship between the usage efficiency of battery and the battery consistency is analyzed, and it can help us deeply realize the behavior of battery and improve the availability by some ways.
Our future works will focus on the quantization of the degree of battery inconsistency impacted on battery pack capacity and maximum available energy under various temperatures.
Co-authors: Xu Zhang, Yujie Wang, Duo Yang.
Citation: Zhang X, Wang Y, Yang D, Chen Z. An on-line estimation of battery pack parameters and state-of-charge using dual filters based on pack model. Energy, 2016, 115: 219-229. doi: 10.1016/j.energy.2016.08.109
Figure legend: This Knowridge.com image is derived from the cited article under CC-BY 4.0 license.
