Greybox system identification , regarding the input and output expected samples?

Here is the MATLAB code:

Model = idnlgrey('mymodel',[1;1;2],[15;1;1],[]);
y = pressure;
u= pwms;
data = iddata(y, u, 0.1, 'Name', "Testing_Data");
Estimated_Model = nlgreyest(data,RovModel);

It's used to estimate the Model using to CSV Files where i obtained 2 readings from 2 different sensors (both are taken at certain condition) , and i want to get a transfer funcution "system model" between them.

below, you will find the 2 files i used + the C Files i wrote + the .mex file after compiling the Cfile.

Each file (Pressure-PWM) is about (64*1).

=> the problem is that the result i got are super inaccurate, is this due to the small data given or the actual logic of the greybox.