Author Topic: ecoinvent v3 Basics: Activities  (Read 19805 times)

Offline pbeilschmidt

  • Administrator
  • Hero Member
  • *****
  • Posts: 586
    • View Profile
ecoinvent v3 Basics: Activities
« on: November 05, 2012, 16:01 »
Activity Dataset / vulgo: 'Unit Process'
The ecoinvent glossary (Data Quality Guidelines, p.155) defines "Unit process [as the] smallest element considered in the life cycle inventory for which input and output data are quantified. ISO 14040:2006, definition 3.34.". Fair enough. But let's get more practical: When building a life cycle model, you typically define the processes (transforming activities) of your production steps as processes with inputs of intermediate flows (see above) that are required to produce one unit of the reference product on the output side. Additionally there might be direct inputs from and direct outputs to the natural environment (elementary exchanges, see above). All flows are equipped with a coefficient and have a unit. This is the description of the process (the "recipe", "definition" or "specification"). Such a proces - a unit process - is a building block of the life cycle model. It will be connected to other processes that deliver the input flows and possibly to other processes that take up the product, by-products or waste. 

Ecoinvent supplies most (=almost all, with the exception of some that are not publicly available in disaggregated form) processes as unit processes. Unit processes allow to unroll (and understand) all upstream supply chains of any process, by connecting - one tier after another - the respective delivering unit process in front of each intermediate flow input. In Umberto NXT LCA there is an automatic support to do this: the Expand feature in the process specification allows expanding the delivering process upstream (or the consuming treatment process downstream) from the process specification. See this post for more details.

In this screenshot of a unit process you can see the reference product on the output side (right) at the top. The unit process refers to 1 m*year (see post on units), and is identified by the Green material type (this is a specific of Umberto, there are no such material types in ecoinvent). Also on the output side are some direct emissions (elementary exchanges, see above). On the input side (left) of the process one can identify only intermediate exchanges (input flows that are delivered by other processes).

More details on activity datasets (unit processes) can be found in the ecoinvent Overview and Methodology report (Data Quality Guidelines) in chapter 4.1 (pp. 18-20).

Accumulated System Dataset / System Terminated / vulgo: 'Result Process'
The ecoinvent glossary (Data Quality Guidelines, p.155) defines "Accumulated system dataset [as] an activity dataset showing the aggregated environmental exchanges and impacts of the product system related to one specific product from the activity".

In this screenshot of an accumulated system dataset (result process) the reference product "road" on the output side (right) at the top is the only intermediate flow. This result process has the same name as the unit process described above and also refers to 1 m*year. The reference flow on the output side is identified by the Green material type. All other entries on the input and on the output side are elementary exchanges. To calculate these exchanges all upstream process chains of this process have been taken into account. So, there are many - possibly hundreds - of direct inputs to and direct outputs from processes that have been accumulated ("encapsulated") and all these elementary exchanges are provided in one result process.  Such a process can be used in the Life Cycel model and does not have to be followed further. This is why this type of process was formerly also called "System Terminated". This is a big advantage for modelling, however, the detailed ("disaggregated") information from which unit process of the upstream supply chain processes certain emissions originate is not available any more.

A few sentences on accumulated system datasets (result processes) can be found in the ecoinvent Overview and Methodology report (Data Quality Guidelines) in chapter 4.15 (p. 35). Much of chapter 14 describes how linking mechanisms and system models (see below) are used to compute accumulated system datasets.

Activities are structured in groups. This grouping is called "Classification". Up to version ecoinvent v2.2 the grouping for activities (formerly called "processes") was a self-defined two-level group/subgroup structure. This way of structuring has been abandoned and ecoinvent v3 now uses the ISIC Revision v4 classification. See UNSTATS webpage for ISIC.

ecoinvent also added some additional sub-divisions for datasets that would not fit into any group.
More details on the classification as well as a table with the unofficial supplemented groups can be found in the ecoinvent Overview and Methodology report (Data Quality Guidelines) in chapter 9.7 (p. 72).

Hint: One side-effect of using the ISIC Rev.4 classification is that many of the actual activities can only be found in the group on the fourth level (e.g. activity "viscose production (Global)" in group 2030 "Manufacture of man-made fibres", under group 203 "Manufacture of man-made fibres", under group 20 "Manufacture of chemicals and chemical products" under top level group 2 "C - Manufacturing".

Activity datasets (both unit as well as accumulated system) are identified in regard to a  geographical localisation. This can be for a specific country (e.g. CH, DE, US), to a region (e.g. RER, RAS, RNA). Additionally you will find datasets that do not refer to a specific geography but are valid globally (GLO). A special geography marker for activity datasets is "Rest-of-World" (ROW).

The geography information for which an activity is valid is shown behind the name of the activity. The ecoinvent database, aiming at a worldwide coverage of their activity datasets have created global datasets for all activities. "To ensure completeness, the ecoinvent database contains a global reference activity dataset (a dataset with the geographical setting “Global”) for each of the included activities." (p.21). When data is available for an activity that related to a specific country region, a "Rest-of-World" version of the dataset can be found. "When a global dataset is the only dataset in the database for a given activity, time period, and macro-economic scenario, this global dataset is included like any other dataset (...). When both a global dataset and one or more non-global datasets are available for the same activity, time period, and macro-economic scenario ... the global dataset is not included ... [and] a dataset with the geographical location Rest-Of-World (ROW)" is provided.

Use advice: If the activity you wish to add to your life cycle model as a background dataset from ecoinvent is located in a country or region for which a specific geographical dataset exists, use this one. If you are in another country, use the "Rest-of-World" (ROW) variant of the activity. If you don't know where this activity is located, or if only one geography dataset is available, use the "Global" (GLO) dataset.

Note: With the new EcoSpold v2 data format the geography concept can in the future be used to identify activities at a specific point (e.g. the coordinates of the factory), along a path (e.g. a transport route), or within any area delimited by a polygon (e.g. on an island, in an administrative, natural oder geographical area smaller or larger than a country). While the first activity datasets and impact assessment methods that include such geographical information are expected to be published and created within the near future, this concept is at present not implemented in any LCA software.

More details on the geographical localisation of a dataset can be found in the ecoinvent Overview and Methodology report (Data Quality Guidelines) in chapter 4.2.1 (p. 21/22).

(Waste Materials and) Treatment Activities
ecoinvent has introduced a new type of activity that is supposed to be used in the models when a material is a waste that needs to be treated. These activities (processes) can be identified by their name "treatment of...".

Here are some examples:
  • treatment of slaughterhouse waste to tallow (Rest-of-World)
  • treatment of automobile catalyst (Europe)
  • treatment of used toner module, laser printer, colour, recycling (Rest-of-World)
  • treatment of blast furnace gas, in power plant (Bulgaria)
  • and so on...

In the original ecoinvent database, these processe have a special feature, which is a negative reference flow output. The ecoinvent documentation explains: "A treatment activity is a transforming activity with a reference product with a negative sign, which effectively means that the activity is supplying the service of treating or disposing of the reference product." (p.30)

In Umberto we do not only rely solely on matrix calculation and have different or supplementary ways of handling such situations. We prefer to think in "flows" rather than handling waste with negative flows. Treatment processes are therefore transformed in such a way that the material for treatment receives a red material type ("Bad") and the negative sign is removed. The material can then be lead to a treatment activity, where it constitutes an input into the process rather than an negative output of the treatment process. The topic of handling negative flows is described in a separate post.

In addition to the waste treatment activities there are now also market processes for these waste treatment activities. They can typically be identified by their name, beginning with "market for treatment of..." or simply beginning with "market for.." followed by the name of the waste material. Examples are e.g. 'market for treatment of waste bitumen sheet (Switzerland)' or 'market for waste plastic plaster, final disposal (Global)'.

Waste treatment and waste treatment markets are described in chapters 4.8 and 4.9 (p. 30/31) of the ecoinvent Overview and Methodology report (Data Quality Guidelines).

Transport Activities
Transport services in ecoinvent v2 were modeled using a service input of a transport dataset (e.g. a truck of a certain síze) with the unit "tkm" (ton-kilometer). This basic unit has been replaced and is now known as 'metric ton*km' in the master data.

The basic unit "vkm" (vehicle kilometer) was used in ecoinvent v2 in datasets that represented the mere operation of the vehicle (=without production, maintenance and dismantling of the vehicle, without construction, maintenance and dismantling of the road infrastructure). These datasets (typically called "operation, ..." in ecoinvent v2) are now replaced by datasets named "energy use and combustion emissions,...." or "energy use and operation emissions,...." and have the basic unit 'km'.

In ecoinvent v3 average transport expenses have been included in the market datasets supplying certain products. Only "for intermediate inputs that are not provided via markets, i.e. where a using activity is directly linked to a supplying activity, the transport is added directly as an input to the receiving activity." (p. 29).

The issue of transport is addressed in chapter 4.6 (p. 29) of the ecoinvent Overview and Methodology report (Data Quality Guidelines).

Infrastructure / Capital Goods
Quoted from the ecoinvent report p. 32 (to be described more user friendly): "The activity datasets for infrastructure production (infrastructure datasets) normally include the maintenance of the infrastructure during its lifetime, its land occupation and land transformation, and its decommissioning and waste treatment. Since the mass of the infrastructure products thus leaves the infrastructure dataset as wastes, the reference products of these datasets do not have any mass, but must be regarded as services providing production capacity. Therefore the reference product of the infrastructure production activities have the property “capacity” or “lifetime_capacity”, and the wastes of the infrastructure have the property “lifetime” exceeding one year."

The topic infrastructure/capital goods is addressed in chapter 4.11 (p. 32/33) of the ecoinvent Overview and Methodology report (Data Quality Guidelines).

You can post your questions in regard to activities as reply to this post.
Peter Müller-Beilschmidt
myUmberto Moderator
ifu Hamburg GmbH