Analysis of stability and resilience of interacting physical systems to external forcing

The implementation of renewable energy resources into electricity pools, replacing the traditional use of fossil fuels, provides well-known advantages in terms of energy efficiency and reduction of carbon footprint. These, however, come at the expense of dangers and instabilities caused to the electricity grids. The unpredictability of the renewable energy systems is a potential source of damaging phenomena for these grids such as congestion and frequency/energy imbalances.

As a paradigmatic example, residential end-users are nowadays transitioning from simple energy consumers to active stakeholders actively managing their own energy requests (through smart-home systems) and production (e.g. through PV self-consumption). Understanding how this may compromise the stability and reliability of power grids requires a deep understanding and modeling of (i) consumption profiles, and (ii) their response to external inputs (like meteorological variables, day/night cycles, economic markets,…).

While these problems involve different (economic, social, technological,…) sides, the approach provided by statistical physics and complex systems will be presumably essential to reach a fundamental understanding. In this regard, the use of toy models (like Ising systems where individual spins represent consumers or groups of consumers, and interactions between them indicate information/energy flows) provides a powerful microscopic framework to explore their corresponding emergent (macroscopic) properties.

The present proposal will try then to explore (i) how spin systems and similar reference models in statistical physics respond to external forcing (like. noisy or periodic external fields), (ii) what possible mechanisms can be used to control/enhance predictability of the system in such situations, and (iii) how such understanding can be transformed into tools for efficient management of fluctuations/variability in real electricity grids.

The PhD workplan includes (i) the compilation of previous work/evidences for the statistical characterization of electrical load profiles of consumers/producers, as well as the corresponding external (e.g. meteorological) drivers of these profiles, (ii) the development of models and theoretical tools to interpret such profiles, and (iii) the computational work to implement the models and reach an understanding of their properties.

Accordingly, the academic background required should include both good analytical and computational skills. In particular, the candidates should hold a basic knowledge on reference models in statistical physics, complexity and population dynamics. For instance, Ising and spin-glass models, agent-based models, cellular automata are tools potentially related to the methodologies to be used in the project. Also, training in statistics and stochastic processes (both theoretical and computational) would be desirable. Finally, general programming skills (e.g. Python, R, parallel programming, …) would also represent an added value.

The group in Stochastic Processes in Biological and Social Systems (Daniel Campos and Javier Cristin) has its expertise in fields like nonlinear dynamics, stochastic phenomena, random walks, long-distance dispersal, and search and exploration patterns. It is part of the SGR ‘Ecoevolutionary responses of animals to climate change’ focused on understanding the response of biological populations to climate and other external forcing from a statistical physics approach.

The Biogeotrace Lab (Ariane Arias Ortiz) studies ecosystem–atmosphere interactions (carbon and energy fluxes). As part of the SGR-Marine and Environmental Biogeosciences Research Group at ICTA-UAB, it provides expertise in biogeochemical and gas flux dynamics across aquatic–terrestrial interfaces, supporting quantitative and statistical approaches to model and predict ecosystem responses to biophysical drivers. Both groups share then an interest on the impacts of environmental fluctuations/stochasticity on natural processes, so they plan to share tools and methodologies as a way to complement and improve their respective lines.