Data driven calibration model (python)

[Amine A]

Hello community,
I want to formulate a data-driven calibration model between physics and simulation results. I have an a model encapsulated in an executable program (data_generator.py) that for a load 𝑃 located at 𝑦 = 𝐿, generate 𝑁𝐵 random beam widths 𝑏(𝑥) and computed the beam displacement 𝑤(𝑥) at 𝑁 positions, at locations 𝑥(𝑖) = (𝑖 − 1) ⋅ 𝐿/(𝑁 − 1) ( file attached shows the beam )
I run the two cases (simulation results 'case 1' and physical model 'case 0')
Usage: data_generator.py <case> ...
* Case 0: Physical model
Usage: data_generator.py 0 <L> <h> <b> <P> <N> <y>
- L: Beam length (default 1)
- h: Beam height (default 0.01)
- b: Beam width (default 0.01)
- P: Load (default 30)
- N: Number of steps (default 100)
- y: Position to apply load on beam (default equal to L)
* Case 1: Simulation results
Usage: data_generator.py 1 <L> <h> <b> <P> <N> <y>
- L: Beam length (default 1)
- h: Beam height (default 0.01)
- b: Beam width (default 0.01)
- P: Load (default 30)
- N: Number of steps (default 100)
- y: Position to apply load on beam (default equal to L)
it gives me results in file attached. I don't know how to proceed to do a data-driven calibration model. anyone has an idea please ?

 

[Amine A]

IRstuff okat thanks I will see )
GregLocock that's in the second case that the beam width varies with x

 

[Amine A]

my first ideas are limited here

 

[Amine A]

if I will do a polynomial regression should I fit the polynomial equation with physics model or simulation results ?

 

[IRstuff]

Huh? The physics model is, by definition, already "fit"

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[Amine A]

I know and that's why I will take it as the reference. I don't know how to fit simulation results to physical model, I don't have any equation for simulation results that's what makes me confused. If I will do a regression, I will consider a polynome that I will fit by determining its coefficients. But if you look at the first question he asked to do a data driven calibration model between physics and simulation results.

 

[Amine A]

what is the problem if I just put (w1(x)+w2(x))/2 as a calibrated model please?? it give results between both models and give the minimal RMSE between both

 

[IRstuff]

It seems to me the whole point of parameterizing the equations is that the problem was given to you to fit a "black-box" physics phenomenological data to SOMETHING that can fit the data and predict the output of the black-box.

In fact, the data from the black box might not look anything like Roark's or Bernoulli's equations; we can only assume that the form is SOMEWHAT akin to the known equations for beam deflection

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[Amine A]

Exactly , the output is clearly "w(x)" but I don't if the inputs here should include physical parameters like (P,h,L,...) or just coefficiens (a,b,c..) related to a polynomial function

 

[IRstuff]

READ your problem statement; it tells you what to do.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm