search

NREL / BioproductTransitionDynamics

A system dynamics decision support tool for bioproduct industry stakeholders who want to investigate how their decisions can impact the process of bioproducts gaining U.S. market share.
https://bioenergymodels.nrel.gov/models/14/
1 stars 0 forks source link

Investigate errors in sensitivity runs #79

Open bwbush opened 4 years ago

bwbush commented 4 years ago

The model sometimes fails to reach 2050 in the sensitivity analysis runs. I'm storing the .cin files for these failed runs in Y:\6A20\Public\BushData\BTD\failed\. The version of BTD.mdl used for the runs is also in this directory.

It's likely that there are just a few causes of failure, but there are many failed runs. There are two approaches to seeing what is wrong:

  1. Manually load the .cin file into the simulation control, press set, run the model, and look at the error message in the log window.
  2. Create an interactive Vensim script to make each failed run.
bwbush commented 4 years ago

P.S., the runs occur in set of 83 (one per input variable), so we're likely to see series of failed runs that occur once a variable is set to a error-provoking value. Thus it is probably sufficient to examine the .cin files that start series of failures, instead of every .cin file.

bwbush commented 4 years ago

It looks like floating point errors in bioproduct market share mass and idealized regulatory wait account for nearly all of the failed sensitivity runs.

bwbush commented 4 years ago

The latest fixes and input ranges seem to prevent the model from crashing.