Parameter Estimation Battery at different temperatures

Hi Everybody,

I'm currently working on a battery model in simscape. I already got my measured data of the cell, but the parameter estimation is the problem right now. I am using a 1-RC model. The Parameter Estimation is done with the ParameterEstimation App.

The measured data is at different temperatures (-10,0,10,25). The previous capacity measurement at room temperature was 61Ah. As expected the capacity degrees with temperature so at -10°C the capacity was only 48Ah.

So my problems are

1. I did my parameter estimation at 25°C with the real measured capacity (59.6 Ah) and the previously measured capacity (61 Ah) (see Attachments at 25°C). There you can see that the Em with the adapted capacity doesn't seem right.
2. Which capacity do I have to use for Parameter estimation (48Ah or 61Ah) --> I tried both and only at 61Ah the Em got a plausible values as at 25°C
3. I did the measurement with 20 SoC steps from 61Ah (3,05Ah each step), so at 48Ah there are only 16 steps done until the final discharge voltage is reached. Therefore I did the simulation with SOC_LUT (0.2:0.05:1) is this right?

I hope somebody can help me.

Let me know if you need more data or infomations.

Thanks and kind regards

Felix

[KUNAL PATIL]

I do have the same doubt....if you found any answer to the question...please post in the comment...and i would also like to know that which prameter estimation app are you using ?

Answer 1

[Lorenzo]

Hello Felix,

If I were you I would start with the "most standard" capacity which would be the capacity at 25°C. You can start by characterizing the cell at 25°C using this capacity. To check if your measured capacity is correct you could for example integrate the current over time and see what value you obtain. By using the current over time of your data you can also plot the SOC over time, for example defining the SOC as:

SOC = 1 + cumtrapz(time,current)/(Capacity*3600)

And see that the SOC is not above 1. If it is, this would suggest probably some measuring error in the current or in the capacity estimation. Small deviation are tolerable (it can be sligthly abve 1) because you also have to take into account measurement error of your testbench for current and voltage.

I would then take this one value and first parametrize the 25°C case and compare with the measured results. If satisfactory, take the capacity over and use it to parametrize the other cases as well. The influence on the temperature will be modeled by the values of the RC block, which will change and therefore also change the internal resistance of the battery.

There is an exmaple that does this temperature-dependency analysis using Curve Fitting Toolbox instead of the Parameter Estimation App. You can take inspiration on the example and use it to implement your Parameter Estimation Solution!

Here the example: https://www.mathworks.com/help/simscape-battery/ug/battery-cell-characterization-for-ev.html

Hope this helps!

Lorenzo