Subgood layer

More
4 years 6 months ago #438 by adminST
Replied by adminST on topic Subgood layer
You are completely right: subgoods are modelled in the same way as substances. But as soon as the option "Use subgood layers" is set under Extras > Options, the "sustances" are treated as subgoods.

Actually, there is only one substancial difference between gubgoods and substances:
When working with subgoods, for each flow in a model, the sum of the mass flows of all subgood layers must be equal to the mass flow on the layer of goods:

m_good = m_subgood_1 + m_subgood_2 + ... + m_subgood_n

So, we actually assume that all subgoods contained in goods are considered. Note that the equation above will be added to the equation system of the STAN model.

When working with substances we normally do not consider all of the substances contained in a good, but only those of interest. Thus, for each flow, the sum of the mass flows of all substance layers is smaller than (or equal to) the mass flow on the layer of goods:

m_good <= m_substance_1 + m_substance_2 + ... + m_substance_n

Note that the equation above will NOT be added to the equation system of the STAN model.

Hope, this makes things a little bit clearer.

A possible way how to model regions of a nation can be found in the attached example. In this file, I have modelled regions as subsystems and additionally used subgood layers.
Attachments:

Please Log in or Create an account to join the conversation.

More
4 years 5 months ago #439 by pa_bontinck
Replied by pa_bontinck on topic Subgood layer
Thanks that's really helpful. So you can have a series of nested subsystems for instance looking like: Country < Region < Commodity, right?

For the project I am working on, I would need to report results for all the different commodities covered, at the national and regional level. So for instance, quantities of potatoes produced and wasted through the supply chain, in each state and nationally. Same thing for all food commodities, the final aim being to get to a national figure that balances out. It will be a really complex model which we need to be able to look at from different angles, slice, extract results, and potentially update in subsequent years easily. We could probably do it on Excel, but I thought Stan would be helpful for the balancing capability and for allowing to show data graphically. But I'm not entirely sure that the tool is flexible enough for what we are trying to do. What do you think? I'm happy to take that discussion offline if that's easier.

Please Log in or Create an account to join the conversation.

More
4 years 5 months ago #440 by adminST
Replied by adminST on topic Subgood layer
Let’s keep the discussion online, so that other users can profit from the exchange of thoughts.

Your stated hierarchy of Country < Region < Commodity is not completely right. You have to distinguish between the spatial and temporal boundaries of your model.

The spatial boundary encompasses all the processes (e.g. production) and flows (e.g. sold products) that you want to consider in your model. In your case, the country is represented by the main system, while the regions are represented by the subsystems inside of main system. The main system shows what is happening in your country without showing the details what is going on in your subsystems. However, the latter can be opened and displayed as individual regional systems.

Which numbers are shown in your flow graphs (main system or subsystems) is influenced by your choice of the respective layer (individual food commodities = layer of “subgoods”, sum of food commodities = layer of goods) and the period of time you were collecting/entering data for (e.g. 1 year, 1 month, …). The latter represents the temporal boundary of your model.

The angles and slices you can look at your model are limited to those elements mentioned above (main system, subsystem, different layers, different periods). Here, Excel might be more flexible. But the main advantage of STAN is that it can deal with uncertain data (using statistical tools like data reconciliation, gross error detection and error propagation). For more information about this topic have a look at our help videos under www.stan2web.net/support/videos/ (especially video 3).

Please Log in or Create an account to join the conversation.

More
4 years 5 months ago #441 by pa_bontinck
Replied by pa_bontinck on topic Subgood layer
Thanks for that, I'll definitely check out the video. I'll get back to you if I need any extra clarifications

Please Log in or Create an account to join the conversation.

More
2 years 8 months ago #496 by jabbakhc
Replied by jabbakhc on topic Subgood layer
hi,
i am reading the discussion and i am unfortunately not getting, how can i differentiate the layer of subgood and layer of substance from each other in STAN. in my case i should model a water system of rice farming with all the pesticide and fertilizer and rice etc. on the other hand i should consider N and P as layer. but my focus is on water system and how the nutrient flows in- and out of my system. i must also define this layers:
Goods
Water (as subgood layer)
other goods (as subgood layer)
N (as substance)
P (as substance)
my problem is you said that sum of the subgoods should be equal to goods. if i define the layer of Goods, Water, other goods, N and P within my system, how should i tell STAN that N and P are just substances but Water and other goods layer are actually my subgood layers? Is there a way to demonstrate that?
that would be a huge help if you could answer me or suggest me a source to understand this because i saw all of tutorials and forum posts and couldn't find anything about this.
Thanks

Please Log in or Create an account to join the conversation.

More
2 years 8 months ago #497 by adminST
Replied by adminST on topic Subgood layer
In STAN, it is not possible to consider subgood layers and substance layers simultaneously!
Either you are modeling subgoods inside of goods or you are modeling substances inside of goods. Both at the same time is unfortunately not possible.
The following user(s) said Thank You: jabbakhc

Please Log in or Create an account to join the conversation.

Time to create page: 0.263 seconds
Powered by Kunena Forum