How do I simulate battery performance in the next 5 years, using existing 5 years trend?

Hi y'all,

I have an enquiry about simulation. I have a set of data for battery performance (OCV,CCV,Ri,SCC,Capacity, etc) for a fixed set of amount of batteries. I have been monitoring the performances for this batch for the past 5 years. With this 5 years worth of data, how do I utilise MATLAB to simulate their performances in the next 5 years to come? Is it ANN or TSF? Are there any other methods? Are there some guide I can follow? Any advice on how battery quality depletion trend being non-constant,over the years, can affect my simulation?

Do appreciate any advice, thank you and have a great day!

Answer 1

[Javier Gazzarri]

Dear L. Sugu,

Thank you for your inquiry. If you have capacity fade and internal resistance increase as part of your last 5 years measurement, you can use them to train a Neural Network with V, I, and T as inputs and capacity and internal resistance as outputs. A LSTM is probably a good choice since it will capture different time frames and thus will be able do tell resistance variation due to temperature (short time span) vs. resistance change fue to aging (long time).

Please let me know if you need asisstance with this.

Javier Gazzarri