Now I’m using Brightway2 and Activity browser on my MacBook Pro(M1 silicon). I built my own parameterized database with Ecoinvent-3.8 (apos). I found that the LCA results(score) are different via these two ways. Using Brightway2, I got 82439873.99596766, while in Activity Browser I got 65659359.41451239. Since I’ve already built the same model in Simapro, I think the smaller one(from Activity Browser) is better, though both are larger than that in simapro.

lca=bw.LCA({bw.get_activity(('ei38', 'fleet, vehicle cycle')):1},('CML v4.8 2016 no LT',
  'material resources: metals/minerals no LT',
  'abiotic depletion potential (ADP): elements (ultimate reserves) no LT')
)
lca.lci()

lca.lcia()
lca.score

I think the Brightway2 and Activity browser should use the same database, and I’ve already checked the LCIA method in both software, and they share the same parameters. So why do they give results that are totally different?

In fact, I’ve also checked the model with some other LCIA methods(GWP100, for example), but it seems that the output results are nearly the same. The results only show much difference in the abiotic depletion potential(ADP).

In addition, when I run that code in my Brightway2, I got the warning “UmfpackWarning: (almost) singular matrix! (estimated cond. number: 2.52e+13)”. I’m not sure if that matters.