Veda2.0 Released!

Dummy technologies (IMP*Z) outperforming available technologies

We are modeling decarbonization scenarios for ethylene production by considering several low-carbon technologies to reach net-zero emission in 2050. However, we noticed that some technologies with high production costs and lower production efficiency outperform more efficient available technologies. We tried to understand why this happens but unfortunately we could not figure it out. We would be grateful if you can have a look on our case study:
For the considered net-zero scenario, the model predicts the Biomass-MTO pathway as the dominant technology in 2050, while there are other technologies e.g. WE-MTO which have lower production cost (and higher production efficiency and also lower emission intensity), but the model predicts only minor contributions for them. For example, we tried to deactivate Biomass-MTO pathway to see how other technologies contribute for the production, but we noticed that the model rather implement Dummy technologies (IMP*Z) instead of other available technologies which we are sure can fulfill the demand and emission reduction constraint. Could you please let us know your thoughts on that? Thanks a lot.

Ok, I looked at your model, and I could readily identify three clear modeling problems, plus a more general good-practice issue.

The modeling problems:

  • There were dummy imports in the model, which should be corrected in the first place. Basically all of those dummy imports were related to the DynACT_* constraints, which were badly formulated. As you may remember, I suggested the following formulation for them (See Box 1 in this thread):

    VAR_ACT(t+1) ≥ VAR_ACT(t)×(0.9^5) + [VAR_CAP(t+1) − VAR_CAP(t)]×0.5

However, you apparently did not want to follow my suggestion, but have removed the VAR_CAP terms.  Consequently, your constraints do not work well (due to the reasons explained in the earlier thread).

  • You have negative emissions for ETHYCO2 in the model.  I am not exactly sure how the process WE-MTO produces its negative CO2 stream, because it only seems to consume electricity, but anyway, according to your model it can produce lots of negative emissions. However, you have not enabled negative emissions balances in your model, and by not doing so there are limitations implicitly imposed on utilizing this technology. (See Negative emissions.)

  • You have defined meaningless interpolation/extrapolation options for FLO_SHAR parameters. The interpolation options should be integer numbers; for example, the option 3 will request full interpolation and extrapolation. But you have used the share values as the values of the IE options, which does not make sense.

The good/bad practice issue:  Your model is very badly scaled. For example, the objective function has a value of 1453155713077, which is 1.5×10^12, a huge number. And likewise, many flows are also very large, for example the flow of NAPHTHA is 460023146 in 2020 (460 million GJ?). I would suggest to scale the model to use larger units, at least kt (kilo-tonnes) instead of tonnes and TJ instead of GJ, kUSD instead of USD (or Mt, PJ and MUSD).

I tested by correcting all of the issues mentioned above, and then the model worked a lot better (probably more or less like you had expected, if the role of WE-MTO as a carbon sink is intended). There were no dummy imports in the results after the corrections, and the Biomass-MTO technology was not being utilized at all.

.zip (Size: 12.87 KB / Downloads: 4)
A small additional remark: For my corrections, I enabled negative emissions only for ETHYCO2 in your model, because it appeared to be the only emission commodity that needed to have a negative balance. However, in other models it is often the total CO2 emission commodity (usually named TOTCO2 or the like) that needs a negative balance enabled, especially when modeling scenarios for the whole energy systems, where negative total net CO2 emissions are required to comply with the target of at most 1.5°C global temperature increase. Likewise, if you would create a scenario where the emission target for TOTCO2 would be negative (and not just zero), you would thus need to enable a negative balance also for TOTCO2.

Many TIMES models also make use of special commodities (often named SNK*) for the modeling of negative emission streams, such that the processes which capture CO2 (or carbon) produce positive amounts of those SNK* commodities, and then there are other processes representing the actual sinks (e.g. geological storage or some other type of a carbon sink), consuming those SNK* commodities while also reducing the actual net emissions. That kind of an approach has been found useful in many models for managing the net emission balances adequately.
Hi Antti,

Thank you so much for all your helpful advises. We have corrected our modeling mistakes accordingly and thus all the problems we had before is now resolved.


Possibly Related Threads…
Thread Author Replies Views Last Post
  Help with modelling CCS technologies for cement industry farzin 3 217 26-04-2024, 08:16 PM
Last Post: Antti-L
  Costs for CCS and DAC technologies Kristina.Haaskjold 2 103 26-04-2024, 04:24 PM
Last Post: Kristina.Haaskjold
  Tax credits implementation for low-carbon technologies jabarivelisdeh 18 4,610 27-07-2023, 01:12 AM
Last Post: olexandr
  Negative dummy import variable Hesam 6 1,167 15-06-2023, 08:30 PM
Last Post: Hesam
  dummy imports dtsintsk 2 806 16-03-2023, 09:17 PM
Last Post: dtsintsk
  PCG for multiple commodity technologies jabarivelisdeh 2 962 01-03-2023, 10:18 AM
Last Post: jabarivelisdeh
  Cheaper Dummy Imports for selected UC Lukas 3 2,503 01-02-2022, 09:53 PM
Last Post: Antti-L
  Dummy import for slack variables Mahmoud 5 3,462 26-10-2021, 03:04 PM
Last Post: Antti-L
  Modeling New Vehicle Technologies eorear86 2 2,368 23-03-2021, 10:20 PM
Last Post: Pernille.S
Question Dummy Imports JozefO 10 9,223 18-01-2021, 05:52 PM
Last Post: JozefO

Forum Jump:

Users browsing this thread: 1 Guest(s)