First time here? Check our help page!
0

What reporting measure should I use in a sensitivity analysis with the Morris method on the PAT to have boxing plots showing me which variable is most influential?

I was trying to see how Morris's algorithmic analysis works on Openstudio's PAT and I was reading on the guide that as Reporting measure adds "Calibration reports enhanced" but when I go to run the simulation as output I only have a plotvthat compares 2 values, the one that used in calibration. Instead I would need a plot that shows me which of the variables the model is less sensitive to.

erminia.v's avatar
1
erminia.v
asked 2022-09-25 04:03:21 -0500
Aaron Boranian's avatar
14.1k
Aaron Boranian
updated 2022-09-26 09:37:25 -0500
edit flag offensive 0 remove flag close merge delete

Comments

add a comment see more comments

1 Answer

1

you can use any reporting measure you want. The important part is to define the variable as an 'Objective Function', as seen in the schema here. Then the Morris algorithm will compute the sensitivity relations between your input variables and output variables (Objective Function = True)

BrianLBall's avatar
2k
BrianLBall
answered 2022-09-27 17:17:05 -0500
edit flag offensive 0 remove flag delete link

Comments

Hi! I am a bit confused. So we need to define output variables (energy, temperature,etc) as objective function (Objective Function = True) and them Morris will compute the relations for all the input variables that are defined in measures (infiltration, insulation, etc)? Also, the results from the Morris algorithm are located in the output csv file from the experiment? Thanks

Raul Teixeira's avatar Raul Teixeira (2022-12-13 12:46:02 -0500) edit

That sounds correct. An example OSA can be found here. The output variable section defines the output variables and which ones are objective functions. 'name' should match the 'measure.argument' that is found in the workflow or OSW part of the OSA. The input variables are defined here

BrianLBall's avatar BrianLBall (2022-12-13 13:04:55 -0500) edit
add a comment see more comments