Chang et al (2021) energy system model framework survey

[Robbie Morrison]

Hello all, hello also Miguel

A recent paper surveys the energy modeling landscape and might be of particular interest to this community:

• Chang, Miguel, Jakob Zink Thellufsen, Behnam Zakeri, Bryn Pickering, Stefan Pfenninger, Henrik Lund, and Poul Alberg Østergaard (15 May 2021). "Trends in tools and approaches for modelling the energy transition". Applied Energy. 290: 116731. ISSN 0306-2619. doi:10.1016/j.apenergy.2021.116731. CC‑BY‑NC‑ND‑4.0 license.

The authors review 54 energy modeling frameworks (or "tools" in their usage) with varying software licensing/business models, comprising:

• Balmorel, Calliope, COMPOSE, DER‑CAM, DIETER, Dispa‑SET, E2M2 - European Electricity Market Model, EMLab‑Generation, EMMA, EMPIR, Enerallt, Energy Transition Model, EnergyPLAN, energyPRO, energyRt, EnergyScope, Enertile, ENTIGRIS, ESO‑XEL, EUCAD, EUPowerDispatch, Global Energy System Model (GENeSYS‑MOD), GridCal, Homer Grid, iHOGA, IMAGE, IMAKUS, Integrated Whole‑Energy System (IWES) model, INVERT/EE‑Lab, LIBEMOD, LIMES‑EU, LOADMATCH, LUSYM, Maon, MESSAGEix, National Energy Modeling system (NEMS), OpenDSS, OptEnGrid, POLES‑JRC, POTEnCIA, PRIMES, PSR – SDDP, Pymedeas, PyPSA, RamsesR, Regional Energy Deployment System (ReEDS), REMIND, Sifre, System Advisor Model, TIMES, TransiEnt Library, UniSyD5.0, WEGDYN, WITCH

Some open source frameworks not present would include DESSTinEE, GENESIS (the C++ variant), eomof, OnSSET, OSeMOSYS, SWITCH, and TEMOA (all listed here). The authors relied on knowledge of and survey responses from the various projects for inclusion.

On a legal level, the concept of "open access" — used several times in the paper but fortunately not in the survey — is not relevant to software. There is nothing in an "open access" definition, even under its strongest framing, that is equivalent to "open source". Indeed even the Creative Commons CC0‑1.0 waiver was rejected by the Open Source Initiative (OSI) for use with software when Creative Commons sought approval. Chang et al used the phrase "open source" in the survey, so the incoming data should be correct. The weak end of "open access" provides for full copyright but no paywall (usage as per Peter Suber). (I discussed that exact definition with Cambridge University after they made Stephen Hawking's unlicensed thesis public thus at doi:10.17863/CAM.11283 and then described the context as "open access".) Some background from me on "open" here.  And wikipedia lists the various colors of open access: gold, green, hybrid, bronze, diamond, and black (almost as bad as hydrogen in that regard). (Speculating, perhaps a reviewer ill‑advisedly requested the change from open source to open access?)

Figure 4 (page 11) summarizes the models based on their licensing arrangements and is worth studying:

Original caption: Comparison of tool types with user-interface among the 54 surveyed tools. Note that the sum of each bar and the total exceed 54 as some tools can fall under multiple licensing/availability and user interface categories.

It is hard to say which paradigm — open or non‑open— prevails without information on project uptake and use. But open source models are surely gaining ground. The concept of "freeware", for the record, is when one distributes suitably licensed binaries (usually *.exe files for Windows) at no cost but keeps the source code private. In any case, the open frameworks listed earlier but not included in the figure would bump the 14 total on the left to 21.

Continuing on the open/closed theme, Calliope, under the Apache‑2.0 license, was recently used by the National Renewable Energy Laboratories (NREL) in the United States to underpin a web‑based public engagement process for the state of Hawaii. The process was named "engage" with more here:

• https://engage.nrel.gov

I don't think the following will be an issue for our domain, but the notion of copyleft software licensing, essentially the GPL family, is falling down badly as software‑as‑a‑service (SaaS) becomes a common paradigm. For exercises similar to the NREL project just mentioned, the notion of transparency should prevail and I imagine the code will need to be available for download and inspection from a pubic code repository somewhere — to do otherwise will simply undermine one's public interest credibility.

Regarding model scope (and also covered well by Chang et al), one challenge that energy modelers might like to consider going forward is adding negative emissions technologies (NET) to offset difficult to avoid emissions in other sectors, primarily agriculture and aviation. Strictly speaking NET are not related to energy services supply but do they naturally align with energy systems. German researcher Oliver Geden has been particularly outspoken on this point, namely that energy modelers should properly support NET.  I understand that IAM scenarios are stacked full of NETs from 2050 onward and that energy models typically run to 2050 so maybe this is also a time horizon or window issue? (Perhaps IAM modelers could comment?)

with best wishes, Robbie

PS: I hope the Chang et al do not mind me pulling out a single figure and reproducing it here. The diagrams individually are standalone works under copyright law but the CC‑BY‑NC‑ND‑4.0 license used would also apply equally to the the entire compilation. The Google Groups server would operate under United States law, which is quite liberal on fair use.  Thanks, in any case, for the nice paper .. under open access!

-- 
Robbie Morrison
Address: Schillerstrasse 85, 10627 Berlin, Germany
Phone: +49.30.612-87617