search

arfc / d3ploy

A collection of Cyclus manager archetypes for demand driven deployment
BSD 3-Clause "New" or "Revised" License
4 stars 11 forks source link

Installed cap should't be used with the back end #289

Closed robfairh closed 5 years ago

robfairh commented 5 years ago

When using installed cap for the back end of the cycle, the simulations fail. This is caused by the fact that calc_capacity takes the values of the capacity of the facilities. D3ploy doesn't see that the capacity of these facilities change as they get full.

The next input file shows that (change the extension to .xml to run it): simple.txt

The reactor produce 1e3 of waste every 2 timesteps. The sink can take up to 1e4. After 20 timesteps, the sink should get full, then d3ploy should be deploying a new one, but it doesn't.

I think we should enable the possibility of using installed cap in the supply driven deployment, because it will cause the simulations to fail. Then, the issue can be closed when those changes have taken place.

robfairh commented 5 years ago

What do you think @gwenchee ?

robfairh commented 5 years ago

@gwenchee do you think we can close this issue?

gwenchee commented 5 years ago

Yup. We shouldn't use installed capacity for the back end because it will cause no new facilities to deploy when there is a shortage of space in the back buffer since it is deploying based on installed capacity.

However, we are leaving the installed capacity capability in supply_driven_inst.