Unit vs System Process in EN15804 add-on for Ecoinvent 3.9.1 problem

Question

Greetings.

I am using the EN15804 add-on for Ecoinvent 3.9.1

I run calculation for provider "treatment of waste concrete, S" and treatment of waste concrete, U" with the impact method 15804+A2.

The system process game negative results and the unit process gave me positive results. The values were the same.

I think that the positive results of the unit process are correct cause there is not any kind of absorption of emission for waste concrete treatment apart from carbonation for carbon emissions.

Is the system process also correct but from a different point of view??

I thought that both processes should always provide the same result.

Can anyone help?

Thanks

Comments

May I ask why the usage of the system process database is still popular in openLCA in cases when there is the choice between unit and system process databases for exactly the same datasets?

Disadvantages of ecoinvent system process database compared to the unit process database:
- no supply chain visible, no sankey diagram and no contribution tree possible
- much slower in importing and exporting, much slower in making backups, much slower when searching flows or other database entries, since the system process version has millions of calculated inventory entries
- no possibility to customize system processes to its own needs/data as it can be done for unit processes
- nowadays no advantage anymore in calculation times, since the unit process database of ecoinvent has calculation times of a few seconds in the newest openLCA software
- hiding the supply chain for confidentiality is not an advantage, because it can be also done in the unit process database by saving the result as a system process if needed

The only very little advantage I can think of is that with the impact analysis tab several results from processes can be clicked through very quickly without pressing calculate first or in cases where the system process database is needed for other tools/software/tables where no matrix calculations can be performed. Sharing certain fully aggregated data sets might be easier, too, but again this can be also done with the unit process version by saving the results as a system process. I'm curious if there is use cases that I'm missing.

---

Greetings Conrad.

Your explanation covered a lot of my questions about the debate Unit vs System processes.

I will probably change to using unit processes unless someone proves me otherwise.

Thanks a lot

---

That's nice to hear Giannis. Of course there are a lot of databases out there where only system processes are available and there is no choice, unfortunately. But in the special case for ecoinvent, where unit and system processes are available, the usage of system processes is maybe a "thing of the past" when computers and LCA calculations where really slow and fully calculated system processes could speed up things. That's my guess.

---

Hello Conrad, to answer this question. At my workplace the computers we use are still basically bricks. We use system processes as default for calculations, but use unit processes to investigate impacts. Its double work, but its what the hardware can handle.

---

Hey Matias, thanks for letting me know. So it might be still an issue. Are you also using the Lazy/On-Demand calculations? They yield the same results, just that the process contributions or the Sankey Diagram are "calculated" on-demand when clicking on it. The Lazy calculations take a few seconds (standard and preferred way to calculate results) whereas the Eager calculations take a few minutes. SSD storage can improve the speed a lot, as well as sufficient RAM (at least 8GB).

Answer 1

Hello Giannis,

I raised this issue with EcoInvent before the update to 3.10. I hope they have updated it in the newest version, which i can see is available in Nexus.

If you cannot download it, i would agree that the disposal with an impact is the one to go with.

Good luck

Comments

The system process database from the EN15804 add-on by GreenDelta (based on ecoinvent 3.10) is online again on Nexus.

Nevertheless, I would always recommend the unit process database to see the supply chain, perform MCS, see and calculate data quality and so on, which is generally not possible with system process databases.

When using "Lazy/On-Demand" calculations, the unit process data sets can be also calculated very fast , within seconds,  in the newest openLCA and a computer that is maybe a few years old or newer.