Topic: result validification Subscribe | Previous | Next
Ali Tamoor 12/10/2019
as usually when we do some research or experimentation results are verified or compared with some threshold result data or validated in comparison to some old data. in case study of leap if we study a certain region how can we validate that our results are true. because if a certain region is being studied for 1st time or may be studied with a different data set how can we validate our results. please reply
Charlie Heaps 12/10/2019
2 Likes >>as usually when we do some research or experimentation results are verified or compared with some threshold result data or validated in comparison to some old data. in case study of leap if we study a certain region how can we validate that our results are true. because if a certain region is being studied for 1st time or may be studied with a different data set how can we validate our results. please reply
Hi Ali,
Good question! In general, its best to think of LEAP as a calculator (a slightly more specialized calculator than Excel, but similar in principle). LEAP is not a model of a specific system. Rather, its a model building framework, with which users can build their own models of particular energy systems. It includes a variety of methods that you can implement including bottom-up, engineering-based, activity/intensity calculations, and top-down, econometric-based projection methods. On the supply-side it also has a variety of methods including basic accounting and simulation approaches to sophisticated least-cost optimization approaches.
- Verification: The basic approaches built into LEAP have been widely tested and debugged and in general, most of the individual methods are fairly straightforward and transparent, making it quite easy for a user of the model to manually verify that the individual results calculated in LEAP are correct and accurate. There are a few exceptions to this. The optimization calculations in LEAP are quite opaque and quite hard to verify (which is the nature of an LP model). In addition the IBC module which is a reduced form version of a global atmospheric geochemistry model coupled to an exposure-response-impact model is very difficult to verify. But most of the other basic elements of LEAP are quite straightforward and easy to verify.
- Validation: However, the onus is on the user (the builder of the model) to check that a particular model is a reasonable representation of the country or region they are modeling. There are many steps in doing this, but an important one is comparing results against key historical statistics such as energy balances or emission inventories or (when developing projections, testing how well those methods are able to reproduce past trends). You can do all these things in LEAP, but the onus is on the developer of a particular model to do that. Here is a bit more on the general topic: https://www.tutorialspoint.com/modelling_and_simulation/modelling_and_simulation_verification_validation.htm
I hope this helps!
Charlie
Ali Tamoor 12/11/2019
thanks alot for such a detailed and guided response. yes as you said about the historical data set I am using is from organizational reports and the data sources are reliable and true. so i hope results will be true as well as per scenarios.
Naeem Haider 7/27/2020
Hi Charlie,
I found the link pasted above very informative. I want to ask a question:
I am carrying out "validation of the first time model" through subsystem validity method of a WEAP-LEAP model.
The model has three subsystems: 1) Water system, 2) Energy system and 3) GHG emissions system. I only have real/historic data of GHG emissions subsystem. Can validation of GHG emissions subsystem reflect the validation of the entire model ? because the model is working as a loop e.g:
Energy generation produces GHG emissions and requires water, thus, energy generation and emissions have direct relations and energy generation and water requirements have direct relations. So is it enough to just validate only GHG emissions system?