Interpolating the Interval Where the Signal Goes to Infinity

I'm modelling a battery cell and while modelling the equation below:

Using the chain rule:

I, T_cell and OCV are known.

I modelled the equation like this:

While the cell temperature is a constant, the signal output is interrupted. I aim to interpolate those intervals. I tried to use Switch and Delay method, but it didn't work. Is this possible? If so, how can I do that?

[Mathieu NOE]

hello

you simply cannot do a division with zero at the denominator (when Tcell is constant)

[Can]

Hello Mathieu,

You're right, I disabled zero-crossing detection for that, and I'm trying to make the signal interpolate when T_cell is constant. Whatever can be done in this case is my question.

Maybe if I save the output to the workspace and use a lookup table.

Answer 1

[Can]

Now, I realized that it naturally goes to infinity, not zero. I can't change the title to "Interpolating an Interval Where the Signal goes to Infinity", but I did it. In a very stupid way, but it works. Let me explain.

1. First I ran the code and the model.

2. Then I wrote a code using isinf (to find infinite values) and delsample(to remove them).

Sim_time = 300;
sim('model');
%% INTERPOLATION
% isfinite is to determine which array elements are infinite(true if infinite)
% To compare, I defined dOCV_dT_inf, a timespace with infinite values.
dOCV_dT_inf = dOCV_dT;
% I defined infarray, a logical array that contains 1 (true) whenever the
% elements of the array dOCV_dT.Data is infinite.
infarray = isinf(dOCV_dT.Data); 
% Why does isinf take 97 of the data? How to change it?
% I ran the code below, number of infinite data went down everytime I ran
% it, after three times number of infinite values was zero.
x = 0;
while x < 3
for ii= 1:97
if infarray(ii) == 1
    dOCV_dT= delsample(dOCV_dT,'Index',ii);
end
end
x = x + 1;
end
dOCV_dT_noinf = dOCV_dT;
sim('model');

3. I ran this section and created a look-up table in the model, using this timeseries.

4. The input of this lookup table is time, so I put a clock as the input.

5. I ran the code and from the scope, I found that it works.

I'm sure there's a much easier way to do this. If anyone wants to explain it, I'm willing to learn.

[Mathieu NOE]

hello again

I'm not familiar with battery models , but I wonder if we can simply switch between models based on when Tcell is constant or not

My 2 cents

[Can]

Hello again Mathieu,

Thanks for the feedback. Code deletes inf values, if T_cell is not constant second model does nothing, thus there's no need to not run the second model.

Regards,

Can