Raquel Pérez-Orozco assisted the 14th National and 5th International Conference in Engineering Thermodynamics (14CNIT) from 4th to 6th held at the University of Zaragoza (Spain).
Within the topic “Energy, exergy, environmental, and economic analysis of thermal systems” she presented the oral communication, entitled “Reconstruction of missing data in mural degradation studies: evaluating Bayesian networks vs. artificial neural networks”. This contribution presents some results of the SmArt project (CINTECX Challenge24) which aims to develop a predictive models to support sustainable conservation strategies of contemporary murals. The project focuses on creating a Machine Learning model to determine temperature distributions on paint mock-ups exposed outdoor and their relationship with their colorimetric changes, assessing murals’ sensitivity in different locations. In this contributions authors address the problem of the missing data during temperature recording. They planned two different approaches to solve this values’ gap using comparative approach: Bayesian networks and regression models (#MLP) in artificial neural networks.
Pains of different composition (PVA and alkyd) protected and unprotected were exposed in different locations (1-coastal area, sea level; 2-coastal area, 400 m altitude and 3-countryside, 400 m altitude) and varying orientations (north and south in 1 and 3) and west in location 2. The temperature monitoring consisted of sample recording with a10 min frequency.
Physical and chemical evaluation of the samples was carried out by using stereomicroscopy, colour spectrophotometry and ATR-FTIR. Results of this contribution constitute a significant starting point for filling gaps during temperature monitoring and exploring the use of ML for predictive models regarding deterioration of paints.
Acknowledgements: This study is part of the (SOS-Murals) Safeguarding the urban murals: cleaning and protection (CNS2022-135645) research project funded by MICIU/AEI/10.13039/501100011033 and, by “European Union NextGenerationEU/PRTR” and by the CINTECX Challenge2024 project: Estrategias sostenibles para la conservación del arte urbano. J.S. Pozo-Antonio was supported by the RYC2020-028902-I project funded by MICIU/AEI/10.13039/501100011033 and, by “ESF Investing in your future”. For more information: https://sosmurals.webs.uvigo.es/
