search

BioSTEAMDevelopmentGroup / Bioindustrial-Park

BioSTEAM's Premier Repository for Biorefinery Models and Results
MIT License
38 stars 18 forks source link

Persistent peculiarities in GWP and FEC TRY plots for the HP biorefinery (potentially cache-related) #24

Closed sarangbhagwat closed 3 years ago

sarangbhagwat commented 3 years ago

The plots (as well as the associated .npy array files) can be reproduced by downloading or cloning from https://github.com/sarangbhagwat/Bioindustrial-Park and running HP/TRY_analysis.py.

While the white spaces are not issues (they're a matter of color bar scales, which had to be set to these values for demonstration), the sharp differences in GWP and FEC values are. They seem to be following consistent trends, which may indicate cache issues at points with similar simulation parameters. The issue does not seem related to sugar-to-cell mass conversions in the Cofermentation reactor (R302), as changing these even to extremely low values does not get rid of the issue or change the apparent trend.

You can run HP_sys.diagram("cluster") in your console to fetch a digraph of the full system.

TRY_analysis_neutral_fermentation

yoelcortes commented 3 years ago

Solved! Please pip install thermosteam==0.24.5

I made some minor updates to the system file so that it works with the new biosteam (you'll want to copy and paste the code). system_light_lle_vacuum_distillation.txt

Among other things, I also updated the boiler turbogenerator to get a better estimate of the amount of lime (required to remove SO2 from the flue gas) and boiler chemicals needed.

It seems the plots reflect the three clear areas of operation in the TRY analysis related to adding dilution water, adding dilution water with evaporation to reduce inhibitors, and simple evaporation. Not so sure about this tho, but it looks like its supposed to be like this

It only took me like 11 hr non-stop to get these done jajaja, but happy its solved. Hopefully we'll be able to get really nice figures even for really complex biorefineries like these.

try_analysis

sarangbhagwat commented 3 years ago

Thanks for working on the cache-related part of the issue, Yoel! Sadly the peculiarities in the graph are still there, though (so persistent!). Yoel and I will continue working to solve this issue, and help from anyone else is also most welcome.