Stochastic Model for Prediction of Capacity Loss and Remaining Useful Life of Lithium-ion Batteries

Although the global market for hybrid and electric vehicles is growing, a key improvement necessary to accelerate the adoption of these vehicles is to optimize the efficiency with which they are used. One of the highest component costs of any hybrid or electric vehicle is its lithium-ion battery, which can cost tens of thousands of dollars and place a significant financial burden on consumers. Moreover, batteries are often replaced too early and used inefficiently as there is no accurate way to evaluate when they have reached end of life. The industry benchmark is that end of life occurs when the battery's capacity is at 80% of its original capacity. The universal standard to measure capacity loss is by completely discharging the battery after using the CC-CV protocol to charge it, which can take up to twenty-four hours and can only be done in a laboratory setting with technical equipment. Due to this challenge, battery management systems in cars use a crude estimate of when a battery has reached end of life by simply counting up to a pre-determined number of days or discharge cycles. This method, in addition to being an inaccurate reflection of capacity, does not factor in how drivers use their vehicles with differing frequency and intensity.

We propose a novel stochastic model to predict capacity loss and remaining useful life from various battery operating conditions and other aging data from a previous experimental campaign. The stochastic nature of the model allows it to account for non-deterministic operating conditions that affect aging. The experimental data used tests the effects of these various operating conditions on the aging process. Successful application of this model could maximize battery usage and possibly extend use by several years (through ensuring that batteries are only replaced when end of life has been reached), provide individualized and immediate feedback to each driver as to how to mitigate battery aging by changing driving behavior, and increase the safety of the vehicle (through preventing use past end of life).