-
Notifications
You must be signed in to change notification settings - Fork 94
Latin Hypercube Sampling
Santiago Peñate Vera edited this page May 14, 2017
·
2 revisions
Latin Hypercube is a sampling technique that enables the "meaningful" sampling of statistical distributions in high dimensions while keeping the number of samples relatively low.
In GridCal I have used the pyDOE library. A simple explanation is given here
Well, if you want to explore the loading in a grid you need to characterize each generator and load with its CDF (Cumulative Density Function) and then make combinations of CDF values and run power flows to see what happens (average voltage, voltage standard deviation, etc...)
You can do this with Monte Carlo, and Now with Latin Hypercube Sampling :)