Descriptif

Modeling the energy and climate transitions

PHY 657

J.-F. Guillemoles

&  D. Suchet

A-   Summary

Production and energy management systems are presently undergoing profound changes. An ever larger number of decisions are based on models and simulations, and especially models with a physical basis. This is often seen in particular in the context of the energy and climate transitions: energy systems models help to improve design and get better performances, better and more competitive fabrication processes, longer lifetimes, better informed investment decisions (in terms of how much energy could ultimately be expected), climate models help understand complex interaction between human activity and environment, scenarios for social transitions help steer political decisions. The models are used to improve our understanding of energy systems by their description (simulation) or by providing tentative explanations, based on our knowledge of physical laws and of the dynamics of the system, they are also expected to deliver some level of prediction, on which decisions could be based. Unfortunately, while the output and conclusions of the models are often readily used and debated, the methodology and the validity of hypothesis and results are not often enough closely scrutinized.

All these models describe complex situations. They can also be ill-used, challenged, or even rebutted. In this context it is essential to understand how these models are built, and how reliable they are. Physics is the science that arguably gave the most accurate models available. It is very powerful in terms of analyzing the key components of an evolving system.

 The educational objective of the course is to show how to use physicist’s tools of the trade to answer the upcoming challenges on energy and climate change: how to build a model in the energy and climate change contexts? How can the key elements be selected? How is it validated? How is its reliability explored? How are uncertainty and incomplete data dealt with?

This is not a course in applied mathematics: efficient programming is important, as is a careful choice of algorithms, but this comes after the general framework is chosen and the direction set.  Nor is it a course on big data analysis. An efficient algorithm can produce irrelevant or wrong data, and using up to date data analysis tool will not help to produce sensible conclusions. This course will focus on reliable and relevant model production.

 The first 3 lectures will provide general methodological basis. Following courses, that will include presentations by experts of the fields covered, will dig into prominent cases of application.  Participants will complete their training through a case study that will be proposed to them. This work will be guided, and will include a report followed by a discussion for the final examination.

B-   Proposed Planning