[Python] Problems with MultiLCA and monte carlo simulation in brightway25

I am trying to perform a Monte Carlo simulation on my LCIA results with brightway25. Unfortunately, I am facing multiple problems.But step by step.

If I try to perform a Monte Carlo simulation using the LCA class I use the following code:

lca_results = {}
relevant_contributions = []
relevant_flows = []
first_level_impact = []
iter_monteCarlo = 100
for method in LCIA_methods:
lca = bc.LCA(functional_unit, method)
try:
lca.lci()
lca.lcia()
lca_results[method] = lca.score

## Get relevant processes
relevant_processes = bw2analyzer.ContributionAnalysis().annotated_top_processes(lca,limit=0.01,limit_type='percent')
## Get relevant emissions
relevant_emission = bw2analyzer.ContributionAnalysis().annotated_top_emissions(lca,limit=0.01,limit_type='percent')
## sum the relevant data up
relevant_contributions.append({
"Methode": method,
"Total score": lca.score,
"Processes": relevant_processes,
"CutOff": 0.01,
})
relevant_flows.append({
"Methode": method,
"Total score": lca.score,
"Processes": relevant_emission,
"CutOff": 0.01,
})

## perform Monte carlo simulation
lca = bc.LCA(functional_unit, method,selective_use={
'technosphere_matrix': {'use_distributions': True},
'biosphere_matrix': {'use_distributions': True},
'characterization_matrix': {'use_distributions': False}},seed_override=None) # Critical for random sampling)
lca.lci()
monte_carloResults = np.zeros(iter_monteCarlo)
for i in range(iter_monteCarlo):
lca.lci()
lca.lcia()
monte_carloResults = lca.score
except Exception as e:
print("Error loading LCI data: ", e)


The problem is that for every iteration of the Monte Carlo simulation, I obtain the same value. Also, the value is, compared to the static LCIA result (~24 kg CO2-eq), much higher, 4000-40000kg CO2-eq. After calling the constructor of the LCA class I get different results, but as I understand, this is not the correct way.

Also if I use the MultiLCA class I do not obtain different results between the iterations. Here is my code with the MultiLCA class:

test_func_unit = {"CellProduction": {activity.id: 1}}
method_config = {"impact_categories": LCIA_methods}
data_objs = bd.get_multilca_data_objs(test_func_unit,method_config=method_config)
mlca = bc.MultiLCA(demands=test_func_unit,method_config=method_config,data_objs=data_objs,use_distributions=True)
# monte carlo
for i in range(iter_monteCarlo):
mlca.lci()
mlca.lcia()
monte_carloResults = next(iter(mlca.scores.values()))


Of course, some probabilites have to be wrongly defined, otherwise I would not get results multiple orders of magnitude higher than the static LCIA results. Nevertheless, the results should at least differ from iteration to iteration. What did I do wrong?

Thanks already a lot for your help. I hope I could describe everything in an understandable way.

Continue reading...