Teaching

How I teach the method: start from the signal, compare simple references, measure what a model adds, and add complexity only if it pays.

Forecasting the intermittent solar resource

The course, slide by slide

A 59-slide course on forecasting solar irradiance one hour ahead, taught in the UE14 applied research project of the Mines Paris – PSL engineering curriculum, built on ten years of hourly measurements from Ajaccio (Corsica).

Open the 59 slides with their text · PowerPoint · PDF

Teaching and research supervision

I supervise and co-supervise doctoral research on solar forecasting, satellite imagery, parsimonious models, energy-system control and radioiodine modeling. Previous doctoral work includes solar forecasting, air quality and solar-potential estimation.

At Mines Paris–PSL, my teaching includes solar resource forecasting and academic tutoring of international internships and gap years (UE45). Earlier teaching spans biostatistics, thermodynamics, biophysics and numerical modeling at the University of Corsica.

The course starts from the signal

Start from the signal: what is predictable in solar irradiance, and what is not. The figure shows why the course works on the deviation from clear sky.

Trois journées à Ajaccio : irradiation mesurée et ciel clair, puis écart additif. Ciel clair, variable et couvert. Figure pédagogique du cours UE14 disponible sur GitHub.
From signal to methods: three days in Ajaccio (2007), measured under clear, variable and overcast skies. Bottom: additive deviation from clear sky, not a ratio. Teaching illustration from the UE14 course, separate from the papers' results. Source: cyrilvoyant.github.io/teaching/solar-forecasting-basics.html

The course page · Solar forecasting basics, with equations · BLEND · ARTU · Stochastic persistence

Talks and teaching

Research Excellence in 2026, watch my talk · YouTube channel
Summary, references and associated preprint.

Forecasting the intermittent solar resource, Mines Paris–PSL course UE14
A guided route through solar variability, statistical models, evaluation and frugal forecasting.

Academic profile · Methods and code · Google Scholar