Using probability density function to evaluate the state of health of lithium-ion batteries

A new method, probability density function (PDF), is proposed for evaluating the state of health (SOH) of electric storage batteries by analyzing the charge/discharge (C/D) data. First, a comparison of the PDF method, the cyclic voltammogram (CV), incremental capacity analysis (ICA) and differential voltage analysis (DVA) is provided. The mathematical basis of the four methods are in agreement. Moreover, the PDF method and the ICA/DVA have an equivalence verified by mathematical derivation. Thus the results acquired by the PDF and the ICA/DVA are quite similar. LiFePO4 and LiMn2O4 batteries are tested to demonstrate the PDF method. Coin cells are tested by the PDF and the CV methods. Results show that the PDF curves and the CV curves have similar shapes. In addition, durability tests are conducted on four commercial batteries to analyze the aging regularities using the PDF method. The PDF results show that the height of the peak reduces as the battery capacity fades. Employing the regularity of peak height reduction with battery aging, an algorithm is proposed to evaluate the SOH online. The PDF method extends the application of the ICA/DVA method. The PDF algorithm is promising to be used in the online SOH evaluation of lithium-ion batteries. (C) 2013 Elsevier B.V. All rights reserved.