Skip to Main content Skip to Navigation
Theses

Towards seamless value-oriented forecasting and data-driven market valorisation of photovoltaic production

Abstract : The decarbonation of electricity production on a global scale is a key element in responding to the pressures of different environmental issues. In addition, the decrease in the costs of the photovoltaic (PV) sector is paving the way for a significant increase in PV production worldwide. The main objective of this thesis is then to maximize the income of a PV energy producer under uncertainty of market prices and production. For this purpose, a probabilistic forecast model of short (5 minutes) and medium (24 hours) term PV production is proposed. This model is coupled with a market participation method that maximizes income expectation. In a second step, the coupling between a PV plant and a battery is studied, and a sensitivity analysis of the results is carried out to study the profitability and sizing of such systems. An alternative participation method is proposed, for which an artificial neural network learns to participate with or without batteries in the electricity market, thus simplifying the process of PV energy valuation by reducing the number of models required.
Document type :
Theses
Complete list of metadatas

https://pastel.archives-ouvertes.fr/tel-02988233
Contributor : Abes Star :  Contact
Submitted on : Wednesday, November 4, 2020 - 3:48:11 PM
Last modification on : Friday, November 6, 2020 - 3:13:48 AM

File

2020UPSLM019_archivage.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-02988233, version 1

Citation

Thomas Carriere. Towards seamless value-oriented forecasting and data-driven market valorisation of photovoltaic production. Electric power. Université Paris sciences et lettres, 2020. English. ⟨NNT : 2020UPSLM019⟩. ⟨tel-02988233⟩

Share

Metrics

Record views

82

Files downloads

28