On the occasion of Mijail Littin’s thesis defense on December 11 and an ECOS exchange program between our two laboratories, we will welcome a Chilean delegation from the Universidad Técnica Federico Santa María in Valparaíso, Chile.
Félipé Escudero, Rodrigo Demarco, and Andres Fuentes will take advantage of their visit to give a seminar at CORIA on Tuesday, December 9, at 2 p.m. in the laboratory’s conference room. This seminar will be entitled «Combining Physics and Data: PIML Approaches for Detailed Flame Field Characterization », you will find the description below.
This talk explores how physics-informed machine learning (PIML) can enhance the characterization of flame flow fields, temperature, and soot properties beyond the limits of traditional diagnostics. I will present recent developments using sparse sensing and PINNs to infer velocity and temperature fields from sparse data, multispectral Telops imaging combined with radiative models for species and temperature reconstruction, and advanced neural network inversion methods such as ANNAbel and RGB-ANNAbel for retrieving soot volume fraction, temperature, and soot maturity, extending them to unsteady and oscillating flames. Together, these methods form a unified PIML framework for high-fidelity combustion diagnostics.