The multi-variable stepwise algorithm for internal short circuit detection in a serial battery pack with inconsistent state of health

Abstract

Internal short circuit of cells is one of the main causes of thermal runaway in electric vehicle battery systems. Therefore, one of the most effective ways to prevent thermal runaway is to detect and identify internal short-circuit lithium-ion batteries before thermal runaway using a battery management system. This paper investigates the detection and identification of internal short circuits in batteries by proposing a multi-variable stepwise analysis (MSA) method. The MSA method is proposed for detecting and identifying faulty batteries by combining horizontal and vertical comparison methods and aging cells' ohmic internal resistance variation characteristics. A less consistent pack containing aging cells was designed to perform internal short-circuit experiments. Based on setting the appropriate threshold and moving the window frame horizontally, comparing the deviation degree of the ohmic internal resistance of each cell in the battery pack and the average ohmic internal resistance of the normal battery, the aging battery in the battery pack can be effectively identified. State of health (SOH) is the percentage remaining of the battery's actual maximum capacity value. The deviation degree of ohmic internal resistance of aging batteries with SOH of 92% and 80% is maintained at more than 15% and 45%. For early internal shorts with an equivalent internal short-circuit resistance of 100 Ω, the internal short-circuit detection time is 3896 s. For the short circuit in the middle and later periods (<10$\Omega$), the MSA algorithm can achieve rapid internal short-circuit detection within the 50 s window, reducing the risk of thermal runaway. The results verified that the method could effectively identify aging cells within the battery pack and detect internal short circuits for other cells, reducing false positives and effectively preventing thermal runaway.