Type: Semester project (10 credits) / Master project (30 credits)
Period: 2016 Spring (will be offered also in 2016 Fall)
Assistant: Stefano Moret
Long-term planning for urban and national energy systems is often based on deterministic modelling, economic optimization and forecasts of fuel prices. When fuel price evolution is underestimated, the consequence is a low penetration of renewables and more efficient technologies in favour of fossil alternatives. This project aims at overcoming this issue by assessing the impact of uncertainty on energy planning decisions.
The IPESE research group develops methodologies for optimization of energy systems. A methodology defining how to treat different types and degrees of uncertainty within strategic energy planning is a current research topic. Methods available in the literature are compared and tested on Mixed-Integer Linear Programming (MILP) models, representing the typical trade-offs in energy planning.
- Learning about optimization and familiarizing with the statement and coding of an optimization problem.
- Classifying and evaluating uncertainty within an energy planning problem.
- Reviewing approaches for optimization under uncertainty.
- Comparison of the methods: application to an energy planning optimization problem.
- Literature review: methods for optimization under uncertainty, with specific focus on energy planning applications. Key methods of interest are robust optimization and multi-stage stochastic programming.
- Comparison of the methods: highlight of key advantages and drawbacks.
- Analysis and further development of a Mixed-Integer Linear Programming (MILP) energy planning problem, representing typical trade-offs in energy planning.
- Application of one or more methods to the MILP problem. Results discussion and key conclusions.
- Optional: Development of a general uncertainty classification for energy planning problems.
Desired skills are:
- Energy conversion systems knowledge (EPFL courses: Thermo I/II, Energy conversion, Advanced Energetics, or equivalent);
- Programming skills: MATLAB, optimization (Modeling and Optimisation of Energy Systems EPFL course or equivalent)
- Statistics and optimization (linear programming)
If interested, please take contact with Stefano Moret attaching your CV and transcript of records (Bachelor’s and Master’s). Ideally, the candidate will continue the work over a longer period of time, possibly another semester project and/or a Master thesis.
As the IPESE research group is now located in Sion, students might need to commute. A travel expense compensation of 1600 CHF (Master project) or 600 CHF/month (Semester project) will be given to students. The frequency of travel to Sion will be agreed with the project supervisor.
 J. Koomey, P. Craig, A. Gadgil, and D. Lorenzetti. Improving long-range energy modeling: A plea for historical retrospectives. The Energy Journal, 24(4):75–91, 2003.
 Alireza Soroudi and Turaj Amraee. Decision making under uncertainty in energy systems: State of the art. Renewable and Sustainable Energy Reviews, 28:376–384, December 2013.
 S. Moret, M. Bierlaire, F. Maréchal. Robust optimization for strategic energy planning. 1st European Conference on Stochastic Programming and Energy Applications. Paris, France. September 2014.
 S. Moret, M. Bierlaire and F. Maréchal. Robust Optimization for Strategic Energy Planning. Technical Report TRANSP-OR EPFL, 2014.