HI,
I want to upgrade an ecoinvent v2.2 product LCA with ecoinvent v3.1 allocation, cut-off by classification.
I think it is better if I make an example - followed by Questions.
For product A, a supplier delivers an intermediate product B of 1 kg copper. Both companies are in Germany.
ecoinvent v2.2:
with ecoinvent v2.2 the intermediate product B was calculated as follows:
1 kg “copper, at regional storage RER” (Material input at the supplier)
1 kg “copper product manufacturing, average metal working RER” (Production of the intermediate product B at the supplier)
0,1 tkm „ transport, lorry >16t, fleet average RER“ (Transport from the supplier to the manufacturer of the product A)
0,2 tkm “transport, freight, rail RER” (Transport from the supplier to the manufacturer of the product A)
With a transport from supplier to manufacturer with 100km by truck and 200km by train (according to ecoinvent report No. 1, "Overview and Methodology Data v2.0 (2007)" Table 4.2 / consumption in Europe / metals / copper)
ecoinvent v3.1:
ecoinvent v3.1 proposes market activities rather than transforming activities. I would choose the following data:
1 kg „market for copper, GLO“(Material input at the supplier)
1 kg “metal working, average for copper product manufacturing, RER” (Production of the intermediate product B at the supplier)
X tkm “transport, Y“ ???
1:
Because the intermediate product B is manufactured in Germany, is it better to use the direct linked transforming activity "metal working, average for copper product manufacturing, RER" instead of using the global market activity "market for metal working, average for copper product manufacturing, GLO "?
2:
How can I consider the transport from supplier to manufacturer?
Do I need assign the intermediate product B to a ISIC r4 number from "transport_default_20130722.xlsx" to get tkm and choose direct link activities for DE or RER transports such as "transport, freight train, DE", or "size-specific lorry transport to generic market for lorry transport, RER "?
Best regards and thanks
Julian
Dear Julian,
First remark: metal working datasets represent the losses and energy required to transform the metal. Here’s a quote from the documentation of the dataset: “This dataset encompasses manufacturing processes to make a semi-manufactured product into a final product. It includes average values for the processing by machines as well as the factory infrastructure and operation. Furthermore, an additional copper input is considered for the loss during processsing.” You will also have to add the quantity of copper that the product has in your study.
If you know which technology and where the metal comes from, you can connect directly to a transforming activity (i.e. not a market). It seems to be the case here. You have to make sure that the dataset you direct link to is representative of the processes in the supply chain you are modelling, and the general comments in the datasets can help you determine this.
However, if you use a direct link, the transport is not modeled. This is the advantage of using a market supplied by ecoinvent: it already has default assumptions about transport. So if you use a direct link, you have to collect data about transport distances and technology of transportation and connect them to your model. Transport is quantified in tonne*km in ecoinvent: multiply the mass of what is transported in ton by the distances in km it will be transported over, this gives you the quantity of transport you have to ask for each transportation means.
Note that you don’t need to add transport for metal working: it is a service. Transportation occurs for goods between transforming activities. Lets make an example. This is a bit tricky since you want to use direct links.
First, you have to figure out what is the proportion of loss in the metal working. The dataset you mention has a 22.7% loss. It means that if you ask for 1 kg of metal working, the metal working dataset will pull on the market of copper for 0.227 kg to compensate for the loss. Since you know where your copper comes from, you have to put this quantity to zero. Let’s imagine you have information about the loss and that it is 10%.
The product you are manufacturing contains 3 kg of copper. The processing took 10% more, so you have to ask for 3.3 kg. You can ask for 3.3 kg of copper to a unit process in Germany or any other country that produces it under the form that you need. Then, if the distance between this plant producing the copper and the plant you are modelling is 100 km, and you know that 90 km is made by train and 10 km by truck, you ask for .0033x90 = 0.297 t*km of train transport and 0.033 t*km of lorry transport. If you are not sure which one to choose from, you can inspire yourself from what the market do by default. Then, you have to ask for 3 kg of metal working (the amount of work is determined by the mass of copper in the finished good, not the total mass that flows through the machining).
There are other ways to model this situation, but I think this one is the less error-prone. I encourage you to define a parameter “copper_loss” and to use mathematical relations show that you are pulling on 3*(1+copper_loss) kg of copper.
I hope this answers your question.
Regards,
Guillaume Bourgault
Project Manager, ecoinvent
Dear Guillaume,
Many thanks for your detailed reply and understandable example for the case that we know where the copper comes from. The values of your 3 kg example is better than my 1 kg example (1 kg is often the amount of the reference flows in processes).
In my study, the focus of the LCA lies on the product A from the manufacturer with detailed specific data. The LCIA of the entire manufacturing phase should be presented separately for the amount of the main manufacturer and upstream processes (including suppliers). Specific information about the suppliers are not available, except that it is known that the production of the intermediate product B (now of 3kg copper) is in Germany. According to our previous posts I understand this situation as follows:
Because I don't know where the supplier purchases the copper for the intermediate product B, I choose 3 kg from the global market data for copper. The amount of copper for the finished intermediate product B and the transport of copper to supplier are included.
Because I know that the intermediate product B is manufactured in Germany, I choose no global market data for metalworking but instead 3 kg of a direct transforming activity linked metalworking data for RER (RER closest to DE). Any additional transport for the service of processing are not relevant.
I must make assumptions for further transport of the finished intermediate product B from supplier to manufacturer, (I choose 100km truck, train 200km). Because I know that this transport of 0.3 tkm truck and 0.6 tkm train is within Germany I choose not a global market transport datasets but a direct linked transport activity for DE or regional transport market datasets from RER (when RER is the next to DE). It is only the transport of the finished intermediate product B of 3kg. Not the copper loss.
In summary, the calculation of intermediate product B to the gate of the manufacturer:
3kg "market for copper, GLO" (material input at the supplier)
+
3 kg "metal working, average for copper product manufacturing, RER" (Production of the intermediate product B at the supplier)
+
0.3 tkm "transport, freight train, DE"
+
0.6 tkm "size-specific lorry transport to market generic for lorry transport, RER"
I hope this is the correct way
Regards,
Julian
Dear Julian,
This looks all very sound. Your model will contain emissions and consumption of intermediate products (for example electricity) from RER, which you know fits closely the real supply chain. You will have some transport distances that are also close to reality for a part of the supply chain. For the rest, it relies on default assumptions.
It was a pleasure helping you!
Regards,
Guillaume Bourgault
Project Manager, ecoinvent