Selecting input variables

7 posts / 0 new
Last post
cspchathu
Selecting input variables

Dear All,
When we are selecting input variables, should there be correlation between the variables?If so how to select them?
I did a parametric analysis using PPA price, SM, number of field subsections, and TES (hrs) for my thesis on CSP. my supervisor asked how do you select these variables as the inputs? what are the correlation between them? what are the statistics behind selecting of inputs? are they have linear relationship or another relationship? How do I answer those questions? please help me.
Other thin is when I click on some links on NREL web site and some google links (of search results of SAM tutorials etc) they are not working. error massages come as "This page isn’t working" or "URL does not have a redirection page ...". Can you help me on these issues?

Paul Gilman

Hello,

You are correct that there may be relationships between input variables, which means that when you change the value of one, you should also change the value of the other. I simple example is that when you change the properties of a component that makes it more expensive, you should change both inputs for those properties and the cost. SAM will not automatically change the values of related input variables -- that is your responsibility as the analyst.

When you encounter a broken link on the SAM website or in SAM's Help system, please let us know so that we can fix them.

Best regards,
Paul.

cspchathu

Dear Mr.paul,
Do you have any guideline or tutorial on this matter. It should include with characteristics of parameters like independent variable, dependent variables, correlated variables etc, and also limitations of parametric analysis which is carried out in SAM.

Paul Gilman

Hello,

We do not have specific guidelines, in part because the relationships depend on the goals of your specific analysis.

The webinars listed under Concentrating Solar Power walk through some examples and include discussions of some of these relationships between inputs, but are not exhaustive:

https://sam.nrel.gov/webinars

Best regards,
Paul.

cspchathu

Dear paul,
I went through the webinar and the technical manual for parabolic trough physical model. But couldn't get clear answer.
My problems are
1. Why do we do a parametric analysis? why don't we do a non-parametric analysis?
2. When we do a parametric analysis what are the assumptions and limitations? here I want statistical answer. that means, that should address correlations between inputs,sample sizes, hypothesis based on, Statistical power, different spread of each group,skewed and non-normal distributions and p value etc...

cspchathu

Dear paul,
I went through the webinar and the technical manual for parabolic trough physical model. But couldn't get clear answer.
My problems are
1. Why do we do a parametric analysis? why don't we do a non-parametric analysis?
2. When we do a parametric analysis what are the assumptions and limitations? here I want statistical answer. that means, that should address correlations between inputs,sample sizes, hypothesis based on, Statistical power, different spread of each group,skewed and non-normal distributions and p value etc...

Paul Gilman

Hello,

SAM's parametric simulations allow you to assign more than one value to one or more input variables to calculate values of output variables. That is useful for optimizing values of input variables, investigating sensitivity of an output variable to inputs, or other analyses involving understanding how output variables vary for a range of input variable values.

The stochastic simulation has more sophisticated statistical tools that allow you to define ranges of input variable values as statistical distributions, and to define correlations between input variables.

Best regards,
Paul.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer