MechE Seminar: Insights, not numbers: leveraging sector-coupled models for planning net-zero energy systems
Event details
| Date | 11.02.2026 |
| Hour | 13:00 › 14:00 |
| Speaker | Dr. Evren Mert Turan, Department of Mechanical and Process Engineering, ETH Zürich |
| Location | Online |
| Category | Conferences - Seminars |
| Event Language | English |
Abstract: Achieving a net-zero society is one of the most significant challenges of our time. Optimization-based predictive models have become central to decision support for the energy system because they translate techno-economic and environmental targets into concrete designs and operating decisions. However, these models are not crystal balls. The optimal solution of a model is rarely the optimal solution in the real world, and the greatest risk is not being wrong; it is being confidently wrong.
This seminar argues for a shift from reporting a single or a few deterministic least-cost options to systematically extracting insights from sector-coupled energy system models. First, a multi-model comparison under uncertainty for Switzerland is introduced, revealing where models agree and disagree and thereby exposing modelling biases and artifacts. As part of this, we show how machine learning can compress high-dimensional results into interpretable, decision-relevant “strategies” for stakeholders.
Second, we introduce ORACLE, a rigorous method for mapping the full space of near-optimal energy systems: solutions that are close to optimal but otherwise differ in technology choices and system structure. In the literature, near-optimal solutions are primarily found using heuristics that do not fully explore the space, leaving out important solutions and potentially skewing decision-making by leaving viable energy options. By using a metric that quantifies the extent of exploration, ORACLE enables systematic, comprehensive exploration of the near-optimal space, achieving equivalent coverage to other methods in significantly fewer iterations.
Together, these efforts reposition energy system models as tools to reveal the futures we can still choose from.
Biography: Evren Turan is a postdoctoral researcher and lecturer at ETH Zurich. His current research interests are the development of rigorous, interpretable, and uncertainty-aware optimization frameworks for the design and operation of complex systems, currently with a focus on energy and process systems. He earned his BSc and MSc in Chemical Engineering from the University of Cape Town and completed his PhD in Chemical Engineering at the Norwegian University of Science and Technology (NTNU), where he focused on process control, optimization, and decision-making under uncertainty.
This seminar argues for a shift from reporting a single or a few deterministic least-cost options to systematically extracting insights from sector-coupled energy system models. First, a multi-model comparison under uncertainty for Switzerland is introduced, revealing where models agree and disagree and thereby exposing modelling biases and artifacts. As part of this, we show how machine learning can compress high-dimensional results into interpretable, decision-relevant “strategies” for stakeholders.
Second, we introduce ORACLE, a rigorous method for mapping the full space of near-optimal energy systems: solutions that are close to optimal but otherwise differ in technology choices and system structure. In the literature, near-optimal solutions are primarily found using heuristics that do not fully explore the space, leaving out important solutions and potentially skewing decision-making by leaving viable energy options. By using a metric that quantifies the extent of exploration, ORACLE enables systematic, comprehensive exploration of the near-optimal space, achieving equivalent coverage to other methods in significantly fewer iterations.
Together, these efforts reposition energy system models as tools to reveal the futures we can still choose from.
Biography: Evren Turan is a postdoctoral researcher and lecturer at ETH Zurich. His current research interests are the development of rigorous, interpretable, and uncertainty-aware optimization frameworks for the design and operation of complex systems, currently with a focus on energy and process systems. He earned his BSc and MSc in Chemical Engineering from the University of Cape Town and completed his PhD in Chemical Engineering at the Norwegian University of Science and Technology (NTNU), where he focused on process control, optimization, and decision-making under uncertainty.
Practical information
- General public
- Free