+1 vote


I have a model, which can be simplified in the below image. So, I have two products, z_A and z_B, they come from Process 2A and 2B respectively. 

Simplified diagram of a system with very basic allocation

Process 2A uses x from Process 1

Process 2B uses y from Process 1

Process 1 produces x and y, which have different weights and different values. 

So, I want to run a product systems for the Process 2A, and a seperate one for Process 2B. How do I make sure that the impacts for x and y are allocated correctly for Process 2A and Process 2B? I've tried using physical allocation within OpenLCA and the impacts of x and y are split equally, despite the different weights, meaning that the impacts of Process 2A and Process 2B hare identical, when they shouldn't by



in openLCA by (290 points)

2 Answers

0 votes
by (57.2k points)
edited by
After a longer explanation of the allocation in general I realise that the issue may be slightly different here - we'll come back to the question
0 votes
by (11k points)

`Process 1` is a multi-output process and you can define allocation factors related to the products `x` and `y` in `Process 1`. Say we have also an elementary flow `e` in `Process 1`:

We can allocate a fraction of the amount of `e` to each product `x` and `y` via allocation factors:

Now, when you calculate e.g. a product system with `Process 2A` which just takes `x` as an input:

you can select the allocation method in the calculation setup:

Depending which method you choose, the corresponding allocation factors are applied. For the example, `physical allocation` results in:

and `economic allocation` in:

Note that allocation factors are also applied to input product amounts so that you can allocate impacts of the supply chain too etc. The best is maybe just to try it out in openLCA. Also, in the next openLCA version it will be possible to use formulas for allocation factors so that you can bind allocation factors to product properties via parameters etc. 

by (290 points)
Brilliant, thanks, that all makes perfect sense, and works when I make simple models. However, when I apply this to my more complex model, the economic allocation works perfectly, but the physical does not, just giving a 50/50 equal split. I'm using Econinvent 3.3 within OpenLCA 1.10.2, are there any known bugs with the physical allocation?