Release | 01
Overview
TIMES is a long‑standing energy system modeling ecosystem, normally applied to national and supranational investigations. Its genesis dates back some twenty years — and was built upon the even older MARKAL and EFOM modeling frameworks. TIMES has been applied to many important studies and covered in numerous peer‑reviewed publications.
The purpose of this posting is to briefly review measures over the past three years or so to move the TIMES ecosystem towards open modeling. Only more recent work involving the open‑source release of the TIMES Model Generator will be covered here. Earlier development, in which the TIMES framework was made available under a restrictive bilateral contract, will not be described.
The TIMES ecosystem falls under the Energy Technology Systems Analysis Programme (ETSAP), one of 38 Technology Collaboration Programmes (TCP) within the International Energy Agency (IEA). The current official host for the project — known as the operating agent — is the Oslo‑based Institute for Energy Technology (IFE). IFE replaced the Technical University of Denmark (DTU) in this role about three years ago. The official ETSAP website is here:
This report is entirely independent and voluntary and was not prompted by the developers of TIMES. The TIMES team were able to offer comments on a late draft — their feedback was minor and technical and has been incorporated. All remaining errors are mine.
TIMES ecosystem
The TIMES ecosystem consists of several components. The core is the TIMES Model Generator — which openmodders would more likely describe as a framework — together with two different data processing pipelines with GUI‑based user interfaces. The model generator itself is written in GAMS, a proprietary high‑level algebraic/mathematical programming language (A/MLP). The codebase was licensed solely under GPL‑3.0‑or‑later in January 2020. The master copy is maintained on GitHub and is regularly updated. In terms of implementation, the core TIMES framework shares similarities with OSeMOSYS — also coded in a comparable yet open A/MPL called GNU MathProg.
In line with other modeling projects, the TIMES ecosystem offers three user‑friendly support applications to manage the data pipeline and ease some of the tedium of processing low‑level inputs and outputs. VEDA (VErsatile Data Analyst) is the long‑established Windows‑only commercial model management system developed specifically for TIMES (Kanudia and Goldstein 2006). While MIRO is a much newer GUI application written in GAMS and R Shiny, currently being developed on GitHub and licensed under MIT. And finally, the TIMES‑excel‑reader project on GitHub, again under MIT but written in Python, should provide a convenient way to convert XLSX spreadsheets — including LibreOffice Calc files — to the DD GAMS format.
The following diagram indicates this broad architecture. XLSX is the preferred format for interfacing and archiving datasets. XLSX is essentially an open standard.
Figure 1: Diagram showing key components and associated data flows for the TIMES ecosystem. MIRO is under development and TIMES‑excel‑reader, being somewhat specialist, is not shown.
GAMS supports open‑source optimization solvers, including HiGHS — although these solvers will not be as performant as the commercial solvers normally used and will doubtless struggle to complete when presented with full‑scale TIMES models.
The terminology for energy system modeling can become problematic. So I will be careful to refer to the framework specifically as the TIMES Model Generator. And I will describe an individual TIMES “model”, in this more restricted sense at least, as a set of input files and a particular release of the TIMES Model Generator. This bundle is currently not sufficient for third parties to easily repeat model runs, but that gap is closing. The concept of an “open” TIMES model is a little more vague but the presence of a proprietary GAMS interpreter and a commercial solver does not invalidate the notion of open source. Users can prepare and feed text data files directly to the TIMES Model Generator should they wish, but this process is laborious and rarely done.
You can find official background on the various TIMES ecosystem tooling here:
- IEA-ETSAP (ongoing). Acquiring ETSAP tools. International Energy Agency (IEA) Energy Technology Systems Analysis Programme (ETSAP). Paris, France. Website.
And here a summary of current progress towards an open tool chain:
| Role | Name | Code license | Repo | Comment |
|---|---|---|---|---|
| modeling framework | TIMES Model Generator | GPL‑3.0‑or‑later | GitHub | information |
| data management | VEDA | proprietary | GitHub | installation information |
| data management | MIRO | MIT | GitHub | under development |
| data manipulation | TIMES‑excel‑reader | MIT | GitHub | under development |
| A/MPL interpreter | GAMS | proprietary | — | |
| solver | various | usually proprietary | — | mixed‑integer linear |
For completeness, VEDA now utilizes the PostgresSQL relational database. See here for background and latest documentation.
There are also currently discussions at ETSAP to open source the ETSAP‑TIAM or TIMES Integrated Assessment Model codebase, a global integrated assessment model (IAM) developed using TIMES.
National/supranational models
Several national study teams have recently uploaded their XLSX files to either code development portals or content archives, such as GitHub and Zenodo respectively. Only those data archives that have recognized open data licenses are listed here. The Creative Commons CC‑BY‑4.0 attribution license arguably works best in this context.
| Study | Area | Data license | URL | Comment |
|---|---|---|---|---|
| JRC-EU-TIMES | Europe | CC‑BY‑4.0 | archive | snapshot on Zenodo (Nijs 2019) |
| TIMES-Ireland | Ireland | CC‑BY‑4.0 | archive | |
| ON-TIMES | Norway | ODbL‑1.0 | archive | deprecated share‑alike license |
Some literature
Giannakidis et al (2015) covers a number of TIMES applications, including an earlier version of the JRC‑EU‑TIMES model. Daly and Fais (2014) describe the UK TIMES Model (UKTM) — nonetheless yet to be made public. Balyk et al (2019) report on a Danish system model. Balyk et al (2022) report on the TIMES‑Ireland model. Hollands and Daly (2023) apply TIMES to examine residential cooking pathways in India, sub‑Saharan Africa, and elsewhere in the global south.
Closure
The TIMES ecosystem is technically open source and can be described as such without issue.
The allied question of whether the TIMES ecosystem, as currently implemented and managed, does strictly class as open analysis is somewhat secondary. The key point is that considerable effort has been made to shift towards open analysis over the last circa four years. Indeed, bringing legacy code and long established consortium projects into the realm of open science and transparent policy development is usually quite arduous.
Working releases of the MIRO user‑interface and the TIMES‑excel‑reader utility will add two important building blocks to the open tool chain being sought by the TIMES project.
The general question of providing convenient and repeatable workflows such that third parties can replicate the results applies to all open energy systems modeling projects at present.
Finally, my thanks to various people within the TIMES community who answered my sometimes blunt questions with patience and respect.
References
Balyk, Olexandr, Kristoffer S Andersen, Steffen Dockweiler, Maurizio Gargiulo, Kenneth Karlsson, Rikke Næraa, Stefan Petrović, Jacopo Tattini, Lars B Termansen, and Giada Venturini (1 January 2019). “TIMES-DK: technology-rich multi-sectoral optimisation model of the Danish energy system”. Energy Strategy Reviews. 23: 13–22. ISSN 2211-467X. doi:10.1016/j.esr.2018.11.003.
Balyk, Olexandr, James Glynn, Vahid Aryanpur, Ankita Gaur, Jason McGuire, Andrew Smith, Xiufeng Yue, and Hannah Daly (29 June 2022). “TIM: modelling pathways to meet Ireland’s long-term energy system challenges with the TIMES-Ireland Model (v1.0)”. Geoscientific Model Development. 15 (12): 4991–5019. ISSN 1991-959X. doi:10.5194/gmd-15-4991-2022.
Daly, Hannah E and Birgit Fais (January 2014). The UK TIMES Model: UKTM documentation. London, United Kingdom: UCL Energy Institute.
Giannakidis, George, Maryse Labriet, Brian Ó Gallachóir, and GianCarlo Tosato (editors) (2015). Informing energy and climate policies using energy systems models: insights from scenario analysis increasing the evidence base. Cham, Switzerland: Springer International Publishing. ISBN 978-3-319-16539-4. doi:10.1007/978-3-319-16540-0.
Hollands, AF and H Daly (1 March 2023). “Modelling the integrated achievement of clean cooking access and climate mitigation goals: an energy systems optimization approach”. Renewable and Sustainable Energy Reviews. 173: 113054. ISSN 1364-0321. doi:10.1016/j.rser.2022.113054.
Kanudia, Amit and Gary A Goldstein (2006). The VErsatile Data Analyst (VEDA) — Presentation. KanORS Consulting and IRG.
Nijs, Wouter (2 September 2019). JRC-EU-TIMES — JRC TIMES energy system model for the EU. Zenodo. doi:10.5281/zenodo.3383901.
Resources
TIMES ecosystem diagram as SVG: times-ecosystem.07.svg.tgz (372 kB)
▢
Copyright (c) 2023 Robbie Morrison.This posting is licensed under a Creative Commons Attribution 4.0 International (CC‑BY‑4.0) License.
If you want a local copy, simply print to PDF.