PAT Algorithmic Mode DOE not creating all expected cases

I am using "run on cloud" in PAT 2.9.1 and Latin Hypercube Sampling in the pursuit of running combinations of a range of values for the following variables - 4 different wall insulations, 4 different roof insulations, 6 values for window u-value, 5 values for window SHGC - which total to 19 different variables.

I ran a sample LHS with 1000 "number of samples". But I see only 33 different runs. I don't really understand what the number of samples is supposed to do. I am sure I am misunderstanding what it means and would love someone to enlighten me! image description

I tried DOE algorithmic method, with the "number of samples = 19" which ran 400 cases, which are half of what I am expecting. What should be my Algorithm Settings to make all the cases run? The roof and ext wall measures didn't work. I have attached the PAT Project File if you wish to check it out.

I have another follow-up question posted a while ago. Take a look if you can, hopefully, you know the answer to it as well!- https://unmethours.com/question/43078...

I appreciate any and all help! Thanks

sborle

asked 2020-03-05 13:18:02 -0500, updated 2020-03-05 13:19:27 -0500

edit flag offensive 0 remove flag close merge delete

Comments

add a comment see more comments

@sborle Your sample method is set to individual_variables which only lets one variable change at a time, with all other variables using the static variable value. I'm not sure where the 33 runs are from vs. 20 (19 variable values plus baseline - all static values). If you change the sample method to all_variables you should have more datapoints, with all variables being altered for each datapoint. However 446*5 equals 480 so you still won't get 1000 runs, but it doesn't hurt to list a larger number of samples than possible. If you made your variables continuous instead of discrete, then you would get 1000 datapoints.