using monte carlo to determine uncertainty.

Question

In my thesis there are some uncertainties (for example in geometry: the diameter of cylinder, the height ,.. or the temperature of inlet fluid ) and I want to know what is the effect of them in my answer. it seems that a simple approach can be monte carlo. I tried to learn about it but the monte carlo in text books seems irrelevant to my aim.

assume there is two uncertainty vary from [-5 to 5] and I have a code that for each input can calculate the output.

I want to know why we don't use organized sample instead of random. why we don't use sample 1 instead ofsample 2 in figure below. and I will be grateful if you introduce me a source to study for using monte carlo in my thesis.

thanks

Answer 1

I don't think you want sample 1 or sample 2.

Do you have a uniform error on [-5,5], or is the error a Gaussian distribution? If it is Gaussian, you want to sample a Gaussian distribution, not random or uniform.

Note that this method of quantifying the error is often called a form of Uncertainty Quantification, or UQ for short.