[M-IS02] Geomaterials in cultural heritage: weathering, investigation techniques, and conservation

[MIS02-P04] Predicting Long-Term Weathering Effects on Historical Buildings Using IoT and Machine Learning

Baptiste PETIT¹、Emily HUBY³、Ciryl RABAT¹、Celine Elise SCHNEIDER²、*Patricia VAZQUEZ²、Hacène FOUCHAL¹ (1.LAB - I, University of Reims Champagne Ardenne、2.GEGENA UR3795, University of Reims Champagne Ardenne、3.Laboratoire de Recherche des Monuments Historiques (LRMH), CRC – MNHN, CNRS, Ministère de la Culture – UAR 322)

キーワード：Monitoring, Microclimate, IoT, Machine learning, Cultural Heritage, Weathering Prediction

A precise and realistic understanding of the microclimate of buildings is crucial for effectively managing and preserving historical monuments. The primary objective of this study is to model the long-term weather and the related behavior of stone materials under various climate change scenarios.

The data were collected from an adapted IoT architecture deployed on an emblematic monument in the city of Reims, the Basilique St-Remi, over a two-year period. An ad-hoc wireless sensor network was installed in different orientations and heights on the building walls.
A substantial amount of data recorder on-site, alongside general weather data from Météo France, was analyzed. Features related to data variations were extracted and categorized into different clusters. The behavior of stones concerning humidity and temperature was modeled. The final step involved predicting the behavior of the entire building in the future (over the next 50 to 100 years) based on weather expectations in three climate change scenarios.

This study represents an initial effort to provide precise insights into the behavior of historical buildings, enabling decision-makers to select relevant preservation measures. The obtained results demonstrate precision comparable to previous general predictions studied in the past.