Import Ecoinvent database into openLCA - unexpected links (different system model) in product system

Question

Dear openLCA community,

I have a problem with ecoinvent in openLCA would like to ask for your kind help.

I am using ecoinvent database 3.7.1 in openLCA 1.10.3.

The system model I chose is 'Consequential, long term', so I made sure that the flows I inserted in the input/output table are all consequential providers.

Then, I created the product system (Prefer default providers; Preferred process type: System process) and calculated the LCIA (IPCC 2013 GWP 100a).

I found that there are unexpected links of APOS flows in my product system, which is another system model. 

It has been confirmed by ecoinvent that if I choose the consequential system model, then it should only have consequential datasets in the linking. Therefore, it is not correct that I have APOS flows in my linking. 

One of the possible reasons is the incorrect way of importing database into openLCA.

What I have done is imported all three system models' database into the same ''folder'' in openLCA, but I make sure to only select the consequential providers for the flows I use as inventories in the input & output table. 

Since I click the option of 'Prefer default providers' to create the product system, could it be that openLCA automatically links the APOS flows as upstream activities that I couldn't see in my input/output table? Because APOS and Consequential database coexist in my folder? (I remember I read it somewhere that if I choose 'Prefer default providers' then openLCA will choose the first provider in the list if it has not been manually assigned by me)

Could you confirm if the above guess is right? or what could be the reason for the unexpected links from different system models?

Secondly, if it is confirmed that the problem is due to multiple system model LCI in the same folder, then I should create a new folder and import only consequential system model's LCI.

There are there files available from openLCA Lenux

If I understand correctly, the first file 'lci' is the algorithm to link flows under the consequential system model. And the second and third ones are unit processes that haven't applied the system model linking?

Should I import 1+2+3 or only 1+2 or 1+3 in the same folder?

(I understand the difference between regionalized and non-regionalized flows, but I am not sure if I can import them in the same folder)

Thank you very much for your help in advance!

Best

Answer 1

Hi, thank you for your long question, this helps me to understand. You are right, you should not mix different system models.

I am not certain what you mean by "APOS flows" since the ecoinvent system models, such as APOS, refer to processes, not to flows.

Since you will never mix different system models, you should not import different system models in one openLCA database, and therefore never will have this combination. Our ecoinvent databases always are modeled with default providers set, to processes (not flows..) of the same system model, so you can combine them in one openLCA database, but is has no benefit since you always need to keep them separate.

If you have modeled own processes, and not set providers to processes from the system model you want (consequential, here) but leave the providers empty, and create the product system then with "prefer default providers", then a random provider is selected if no provider is set, and this provider can then be from the APOS system for example.

Then, I do not understand "Folder"  ("into the same ''folder'' in openLCA") - I guess you mean an openLCA database.

To your question with the 1+2+3 import: regionalised is better for some impact categories that address regional impacts, you should import either the regionalised or the non-regionalised, both contain the same processes. For LCI, these are not regionalised and thus you should not combine it with the regionalised version.

Hope this helps!

Andreas