The majority of Li-ion chemistry varieties consists of flammable materials and therefore may go to unpredictable catastrophic failures. One of the reasons for the battery thermal runaways is the development of low-resistive internal short circuits defined as hard shorts. Unfortunately, currently there are no fast and simple approaches for detecting the earliest signs of growing internal shorts and tracking battery health by this way. This impedes penetration of Li-ion batteries in electric vehicle and energy storage markets. The main objectives of our research are: elimination of thermal runaways and reduction of manufacturing and battery maintenance costs by early identification of potentially faulty cells. 

Internal shorts increase self-discharge (SD) of batteries. Therefore, the SD metric can serve for battery health estimation. However, presently there is no fast and at the same time direct methods to measure SD that doesn’t require full charge/discharge cycle with long rest times between charge and discharge steps. Conventional methods determine average SD rate for the specific long rest period. They are impractical for online SD measurement. Therefore, the focus and goal of our approach was on the development of a fast SD measurement method that can be incorporated in Battery Management System algorithms. The goal was reached with a suite of experimental electrochemical protocols combined with a newly developed model. The new metric, the SD current, which is determined under potentiostatic conditions at a slight discharge overvoltage, is proposed as a fast SD metric for detection of shorts and assessment of their severity. The slight discharge consumes only negligible capacity with minor state of charge disturbance. However, it allows equilibrating the battery quickly and subsequent determination of SD current as a pure metric at a stable state of charge corresponding to applied voltage. The method doesn’t require disconnection of the battery from the circuit and stabilization of battery voltage before measurement. Measurement can be done in 1–1.5 h or less in comparison with days or weeks for conventional SD measurement.

The method is validated using a cell comprising artificial external shorts of different severity and with no short. It precisely differentiates the shorts with different resistivity. In addition, the developed modeling approach allows further decrease in the time of measurement using only initial partial datasets instead of the full set of data. The mathematical description was validated for several runs with 6% deviation from experimental data. A prime advantage of the model lies in the ability to predict SD at different battery state of charges, temperatures, etc.

The method is non-invasive, and is battery chemistry and design agnostic. It is capable of the detection of nascent internal shorts as precursors of soft and then hard shorts that cause thermal runaway. Since the method detects early precursors of cell failures far before the point at which a thermal runaway occurs, it enables to provide advanced warning so that corrective action can be taken to avoid fires and cell explosions. The technology based on this patent pending method can be used in electric drive vehicles, stationary energy storage applications, military, aeronautic, as well as for portable electronic devices. It is easy adaptable by any battery management system to monitor battery state of health at any time and at any battery state of charge. Another promising application may be as a tool for the first responders on electric drive vehicles, grid storage and other battery related accidents. It also can be used at battery production, reducing inventory time to detect internal shorts not intentionally introduced to the battery during manufacturing process. Lastly, this technology could facilitate consistent analysis of SD behavior for previously used batteries and their cells that are repurposed within the secondary use industry.