We will be performing a major infrastructure and web site ugprades on Friday, September 27th 8:00 am - 5:00 pm MST (UTC -7).

The website will be down during that time.

Stochastic analysis

  • ecasati
  • Topic Author
More
12 Aug 2019 08:52 #7608 by ecasati
Stochastic analysis was created by ecasati
Dear all,
I find the "stochastic capabilities" of SAM are very interesting and I'd like to present these to my students. However, I was not able to find a proper manual documenting them. I just found a recent  post with short video-tutorials .
Beside setting input and output, which is quite straightforward, there are unclear things:
  1. what is the "enable weather file analysis" option doing?
  2. there are calculated figures, likely regression results, named "Delta R^2" and "Beta" (calculated for each uncertain input variable). What is the definition of these quantities?
Is there a general user manual elucidating on these aspects?

Thanks

Emiliano 

Please Log in or Create an account to join the conversation.

  • pgilman
More
26 Aug 2019 23:09 - 26 Aug 2019 23:10 #7630 by pgilman
Replied by pgilman on topic Stochastic analysis
Hi Emiliano,

The "enable weather file analysis" option requires a folder containing set of weather files. When you enable the option, instead of creating a set of values for an input variable, it creates a set of weather files chosen from the files you provide, and then runs simulations using those files to calculate the distribution parameters for results variables you choose. For example, you could use this technique to determine a statistical range of values for annual energy or net present value based on the variation in a series of weather files over multiple years.

You can see what it does by clicking Compute samples to generate an example distribution of weather files. However, this table does not match the table SAM uses for the actual stochastic simulations when you click Run simulations, and unfortunately, the table of results does not show you the computed samples that SAM used for the simulations. This is a known issue that we will address by adding a right-click tool that shows you a table of computed samples used to calculate the results, but that feature will not be available until the fall 2019 version in a couple of months.

Best regards,
Paul.
Last edit: 26 Aug 2019 23:10 by pgilman.

Please Log in or Create an account to join the conversation.

  • ecasati
  • Topic Author
More
17 Dec 2019 12:32 #7845 by ecasati
Replied by ecasati on topic Stochastic analysis
Dear Dr Gilman,
thanks for your precise feedback about my first point!
Could you please give me some info on the second point of my question?

Thanks in advance

Emiliano 

Please Log in or Create an account to join the conversation.

  • pgilman
More
19 Dec 2019 23:42 #7847 by pgilman
Replied by pgilman on topic Stochastic analysis
Hi Emiliano,

SAM's stochastic simulations use the Sandia Latin Hypercube Sampling (LHS) package. The references for that package are provided in the SAM Help system:

SAM uses the Sandia Latin Hypercube Sampling (LHS) package ( dakota.sandia.gov/content/packages/lhs ) to calculate statistical distributions for stochastic simulations. The LHS sampling method is described in Wyss, G; Jorgensen, K. (1998). "A user's Guide To LHS: Sandia's Latin Hypercube Sampling Software." Sandia National Laboratories. SAND98-0210. 140 pp. ( PDF 559 KB ) For a more general discussion of sampling-based uncertainty analysis, see Helton, J.; Davis, F.; (2000). "Sampling-Based Methods for Uncertainty and Sensitivity Analysis." SAND99-2240. 121 pp. ( PDF 5 MB ) For a basic description of the Latin hypercube sampling method, see the Wikipedia article " Latin hypercube sampling ."

The PDF documents describe the different distribution inputs, but don't seem to describe those regression outputs, and, unfortunately I don't have specific documentation for those.

Best regards,
Paul.

Please Log in or Create an account to join the conversation.

Moderators: pgilman
Powered by Kunena Forum