Open-Source Workflows for Energy System Optimization Models: Learning from Experiments
0 likes116 views
The document outlines the Indian Institute of Technology Roorkee's open-source workflows for energy system optimization models, focusing on the IDEEA initiative aimed at enhancing renewable energy integration and facilitating energy policy discussions in India. It highlights the development of open-source modeling tools, community engagement, and collaboration with international organizations to address the challenges of decarbonizing the energy sector. Key goals include building a population of trained modellers, aligning energy scenarios with national priorities, and creating adaptable energy models to contribute to India’s clean energy transition.
Open-Source Workflows for Energy System Optimization Models: Learning from Experiments
- 1. INDIAN INSTITUTE OF TECHNOLOGY ROORKEE Open-Source Workflows for Energy System Optimization Models: Learning from Experiments 24 Jun, 2024| Summer 2024 Semi-Annual ETSAP meeting Tarun Sharma, IIT Roorkee Oleg Lugovoy, Open Energy Transition & OptimalSolution, USA
- 2. 2 Workstreams in our research group: • TIMES • energyRt • SWITCH • Data pipelines
- 3. Modelling workflow: A critique • Requirements/Questions/Functionalities: • Decarbonizing energy system • Abatement curve, Policy levers, R&D and Investments • Inter-sectoral coupling: Electrification of Transport • Requirements of storage and transmission • Renewable integration • Assessment of flexibility • Modelling decisions: • Spatial, technical and temporal resolution • Planning horizon • Selection of modelling environment • Validation (Trutnevyte,2016; Pfenninger, 2017) Fundamental transition: Assets lasting multiple decades • Joint development within research networks, e.g., IEA- ETSAP, openmod, IIASA, IDEEA • Open code and open data: scrutiny and replicability
- 4. Mission: “IDEEA aims to build a community of trained energy modellers using an open access modelling framework for engagement in larger energy policy discussions nationally and globally.” • IDEEA aims: 1. To help increase RE share in India’s energy portfolio and clean-energy adoption in a sustainable manner and understand the transition challenges 2. To develop scenarios that are in alignment with national priorities as well as adaptable to India’s widely varying contexts at the sub-national level 3. To create an indigenous open-source energy model for the country, which can evolve with India’s developments over time
- 5. 6 How is IDEEA different from the ongoing efforts? • Use open-source modeling tools and data • Strive for transparency • Documentation of the source code • User Manual • Embed database queries into the documentation for dynamically interacting with the data • Building a community around the effort • Limited to electricity system thus far
- 6. 7 Choices faced as energy system Modellers • Weather to write our own code and/or model generator? • Selection of language • Solvers • Modularity • Is the data pipeline automated? • RE potentials • Generation capacities • Capabilities for advanced visualization and quick diagnostics? • Capability for scenario management?
- 7. 8 Choices made with energyRt and IDEEA The gap addressed by writing energyRt • Minimize time of development and application of Reference Energy System(RES) models • Byproduct of independent research by several researchers with prior experience TIMES, OSEMOSYS plus own insights • The initial goal was to post-process the results and produce reports- Using R • R Scripts provide flexibility, ease in shareability, and reproducibility compared to the then experience of researchers • Why R? • (+) Data handling and visualization • (+) GIS, and more • (-) Not much optimization options • (+) Many scientists and economists use R (statistics, econometrics) • GAMS for replication; Not Open Source • The goal was capacity building: sharing of tools, initial projects in China. • Need for assimilating input data => Development of the full code leading to the package. • Choice of implementation in Julia, Python, GLPK to users • Cross-validation/verification • Embed database queries into the documentation for dynamically interacting with the data
- 8. 9 energyRt: main classes • technology – stores technology parameters • commodity – stores commodity information • supply – stores data on resources supply • demand – stores final demand data • constraint – constraint on endogenous variables • repository – stores commodities, technologies, supply, demand • model – stores main model parameters and all info above • scenario – stores all info above and results from the model run
- 9. 10 Work supported by EDF • ODPP-MCET 1 (Oct 2022-Sep 2023) – Participation in modeling training and regular meetings about the project. – Adaptation of the models to India, execution of the models, and validation of results – Decarbonization scenarios of the Indian electric power sector • ODPP-MCET 1 Amendment ( Apr 2023 – Oct 2023) – Lead organizer for the IDEEA Offsite, Workshop and the Launch. • MCET 2 (Oct 2023 – Sep 2024) – Adapt and extend open electricity planning models to meet the needs of stakeholders in India. – Participate in the International Network on the Economics of the Energy Transition (INEET), discussing methods and results for decarbonizing electricity production and related activities. – Work with EDF staff to refine the analysis, share results and plan future work.
- 10. 11 Offsite, Workshop and the IDEEA launch Launch (21-23 Aug, 2023)
- 11. 12 Modelling workflow • IDEEA website The entire workflow and the data pipelines are described along with detailed code to reproduce the analysis.
- 12. 13 India 2050 Net Zero Electricity System Scenario Snapshot of results
- 13. 14 India 2050 Net Zero Electricity System Scenario Snapshot of results
- 14. 15 India 2050 Net Zero Electricity System Scenario Snapshot of results
- 15. 16 Output • Development of tools – Development and testing of the IDEEA package – Development of SWITCH India Model • Data Collation • Capacity building on Electricity System Modelling • Stakeholder outreach, engagement, and buy-in – Offsite – Workshop – Launch – Climate Week NYC – China Energy Modelling Forum
- 16. 17 Significance/Pathways to Impact • Developing roadmaps and policies • Identify technology and business opportunities for entrepreneurs and/or existing players in India over the upcoming decades arising from a low carbon energy future. • Assess in detail the value chain for these opportunities to determine genuine win-wins for India. • Developing an innovation system to enable delivery of these opportunities. • A platform for integrating models and scenario development. • Unlock the potentials in research and industry.
- 17. 18 Development choices • Attributes of Energy/Electricity Modelling exercise/analysis: • Open source for capacity building; GAMS might not always be affordable • Cross-validation/verification across languages • Reproducibility: Hard to make research reproducible with Visual User Interface; • Sharing is hard (need to show the clicks) • Doing it with GitHub and package is good practice for software building • Interactive rendering on website linked to the GitHub repository helps with outreach • Contrasts with standalone scripts, functions, and data • Detailed documentation: • Illustrations, use cases, case studies: Work in progress • Building sponsorship: • Ongoing development and provision of support and training will require sponsorship • Otherwise, it will continue to serve as a public repository and a useful resource for the researcher. • Stakeholder buy-in: • Interest in research teams like at IISc, Jadavpur, IIT Roorkee; Learn with handholding and troubleshooting • Policy experts: Traction, willing to engage and listen; Suggesting inputs for questions, capabilities etc.
- 18. 19 Ongoing work at our lab: - Compilation of hydro database including attributes on pond size (partially complete), hourly stream flows (Exploring approaches) - Capability to assimilate arbitrarily large number of time series each for wind and solar o Fetching the spatially resolved time series for Wind and Solar from MERRA-2, Global wind atlas o Integrating temporal and spatial clustering workflows - Prototypes to demonstrate the functionality for all complex objects e.g., long duration storage, transmission, hydro with dam, demand response. - Capability to build models for arbitrary number of regions (PowerGenome/PyPSA earth for India) - 2nd Indian Decarbonization Pathways Workshop
- 19. 20 Shunya Lab, IIT Roorkee Dr.Tarun Sharma (Assistant Professor) Principal Investigator Team leader Modeller, developer Malvika Bajpai (Project Associate) Project Coordinator Modeller Kartik Gupta (Btech 4th year student) Modeller, developer Ankit Sharma (Project Associate) Project Coordinator Modeller Dr. Arvind Bisht (Postdoctoral Project Fellow) Project Coordinator Modeller Anu Agarwal Doctoral student Prashant Giri Doctoral student Ashish Saraswat Doctoral student Sandeep Kumar Doctoral student Himanshu Badoni Doctoral student