Question-and-Answer Resource for the Building Energy Modeling Community
Get started with the Help page
Ask Your Question
6

Best way to do sensitivity analysis in EP

asked 8 years ago

GauravC's avatar

updated 7 years ago

Would it be possible for anybody to share how exactly is sensitivity analysis done in Energy Plus? What are the exact steps that have to be taken?

Is there any Excel plugin for this or do we have to link up EP with Matlab using BCVTB or using a Python code for analysis of final result files?

Preview: (hide)

2 Answers

Sort by » oldest newest most voted
4

answered 8 years ago

updated 8 years ago

For sensitivity analysis in python I've been using SALib which is pretty easy to use.

I use the SQL output rather than CSV files, which I think is also the approach in jEPlus. There might be some useful snippets in my repository here which is a framework for dockerised client/queue/worker architecture for creating, running and analysing EnergyPlus jobs.

In particular:

Preview: (hide)
link

Comments

1

Hi, Thanks for your answer. I think I can dig into it work from there.

GauravC's avatar GauravC  ( 8 years ago )
3

answered 8 years ago

updated 8 years ago

All these tools can be used for be used for both sensitivity and optimization

Side note:

To some extent the Parametric Analysis Tool (OpenStudio PAT) can also be used, but it you have many cases it'll be tedious at best.

The EnergyPlus Parametric IDF objects can also be leveraged, as well as any scripting language you'd like.

Preview: (hide)
link

Comments

Hi,

Yes, I am aware of these tools, but I guess these are just meant for optimization or parametric simulations. I am able to make a Py script to make multiple idf files but am stuck after the simulation runs.

I wanted to know of a way to post process the data files from 1000s of simulation runs. If anybody can share Py script for post processing multiple CSV files, to calculate sensitivity of one parameter or even combo of multiple parameters towards certain output.

GauravC's avatar GauravC  ( 8 years ago )

The OpenStudio Analysis Framework is designed to perform a wide range of tasks including sampling analysis based on uncertain parameters within measures. PAT 2.0 will include much of the functionality of the analysis spreadsheet - removing the tedium that Julien mentions for certain classes of problems. The real strength of the framework is being able to use OS Measures to parameterize complex model transformations. If you want to analyze things with simple text substitution there are a number of options, but OpenStudio lets you do so much more very compactly.

ljbrackney's avatar ljbrackney  ( 8 years ago )

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Training Workshops

Careers

Question Tools

1 follower

Stats

Asked: 8 years ago

Seen: 1,191 times

Last updated: Jun 23 '16