M.A.P.L.E. Lab

Optimal Battery Charging Protocol

To go directly to the video demonstration, click here.

Battery users typically desire two objectives when charging their batteries: 1)Store as much charge as possible 2)Store that charge in the quickest amount of time possible. While charging batteries at a high rate will push more charge into the battery, the answer to meeting both these demands is not as simple as applying the highest charging rate as possible. Increasing charging rates will give rise to additional phenomena that can be prohibitive to battery performance and life. High rates can lead to increased internal battery temperature leading to degradation and thermal runaway. Additionally, these rates can increase the side reactions occurring within the battery leading to increased resistive layer growth reducing the capacity of the battery. Also, because of diffusion and kinetic limitations, increasing the charge rates can lead to mechanical stress build up and failure within the battery. While there are many problems associated with fast charging, the biggest problem is current battery management systems inability to accurately measure the internal states of the battery during period of quick charging. If the internal states of the system were well known more aggressive charging rate could be applied as long as the monitored states remained within safe levels. Unfortunately measuring the internal states of the battery is very difficult in situ and therefore requires modeling to accurate monitor these in real time.

While there are many problems associated with fast charging, the biggest problem is current battery management systems inability to accurately measure the internal states of the battery during period of quick charging. If the internal states of the system were well known more aggressive charging rate could be applied as long as the monitored states remained within safe levels. Unfortunately measuring the internal states of the battery is very difficult in situ and therefore requires modeling to accurate monitor these in real time.

Battery models range from computationally fast empirical based models that can quickly accumulate error to extremely accurate kinetic Monte Carlo surface simulations which take weeks to run for only a few cycles. Currently most battery management systems use simplistic empirical based models because they can be solved in real time, but these models limit the battery’s ability to be fast charged. By implementing a higher order physics based model which will accurately track the internal states charging protocols that push the boundaries of fast charging can be implemented safely.

The model shown in the video is an optimization simulation that uses a porous electrode P2D model for a Lithium ion battery that has been mathematically reformulated for computational efficiency (For more about the reformulated model used in the simulation see our group’s paper here). The optimization seeks to obtain the greatest amount of stored charge over the given short charging time period. Constraints on temperature and capacity fade can be placed on the system so that the battery does not experience adverse effects from the fast charging protocol.

To view the demonstration video click here.

To contact the MAPLE group for access to the model and interface seen here please email the group’s PI: Dr. Venkat R Subramanian at vsubram@uw.edu