The problem about Simscape Battery MBC parameter estimate.

I am trying to calibrate my battery parameters using the MBC tool, but I encountered the following issue. I input the HPPC experimental data into the MBC tool and waited for it to fit the parameters, such as Em, R0, R1, etc. However, the results were disappointing. The fitting did not converge, and the curve could not be fitted at all. The result is shown in Figure 1 below. Could you please help me identify the issue? Is it a problem with my parameter settings in the MBC tool, or is there an issue with my HPPC experimental data?

[Mike Sasena]

Hi Geli,

It's pretty hard to debug the issue without more info. Can you publicly share the data you're trying to fit?

Mike

[Geli]

Hi Mike，

It is ok for me to share my experiment data.It is as the attachment named "experiment_data2.mat" which is a 3-column table with time,voltage and current.I am looking forward to your reply.And if there is anything more needed, please tell me,I am always ready to share more information with you.

sincerely thanks,

Geli

[Mike Sasena]

Thanks for sharing the data, Geli. I'll take a look. In the meantime, I'd recommend you review this example for the fitting process using MBC. It walks through the steps of pre-processing the data, then either fitting the models with MBC using either the UI or scripts, depending on your preference.

Answer 1

[Mike Sasena]

Hi Geli,

To start, I recommend you review the Time column in your data. If we zoom in one of the SOC sweeping step (where a slow rate discharge, for example C/10 was used to reduce SOC from for example 80% to 70%), this step only takes less than 50 seconds. We would expect this step to be much longer. If C/10 were used, it should be something like 3600 seconds. I suspect that the "Time" column is not time, but just row index. That might explain why the model is having a hard time matching a system response to a time-based signal.