Closed MaximeAgez closed 5 months ago
Similar issues observed on AB and on brightway2 itself
@cmutel, any thoughts?
@romainsacchi thanks for answering!
I will close this for now. If there are more questions regarding MC results, I suggest we open a new issue on brightway-lca/brightway2-calc.
Updating AB
What happened?
I ran multiple Monte Carlo analyses (100 iterations) on the "market for electricity, low voltage | electricity, low voltage | CA-QC | ecoinvent3.9.1 cut-off" process. The deterministic value for this process is 0.026651 kgCO2eq (using the IPCC2021 method). Yet, when I run Monte Carlo on this process alone (not modified or anything) the mean for the Monte Carlo is between 0.0293 and 0.0294 (did 4 times 100 iterations MC) and the median between 0.0285 and 0.0293. This is extremely weird that neither the mean or the median is remotely close from the deterministic result.
Similar issues observed on AB and on brightway2 itself (did not test brightway2.5).
Just to be sure I also tried with another process. Same issue with the process "market for 1-propanol | 1-propanol | GLO | ecoinvent3.9.1 cut-off" whose deterministic result is 6.6185kgCO2eq and MC results turn around 7.52kgCO2eq (mean) and 7.37kgCO2eq (median).
Relevant errors
No response
Operating system
Windows 11
Conda environment