Methodology of the quantitative assessment of the moisture content of saline brick walls in historic buildings using machine learning

Abstract

AbstractConducting moisture tests of brick walls in buildings under conservation protection is associated with many difficulties that result from the inability to freely interfere with historic tissue. The current paradigm of conducting such research, which assumes the use of just one non-destructive method, has many limitations that affect the accuracy of obtained results. Up-to-date research concerning an alternative non-invasive method, which allows reliable test results to be obtained in the case of the quantitative assessment of the moisture content of saline brick walls in historic buildings, has shown that it is possible to reliably assess such a moisture content using machine learning and two complementary non-destructive methods. In the article, the original methodology of such a quantitative assessment is described and presented in the form of block diagrams. The methodology consists of two stages. The first stage includes carrying out experimental and archival research in selected historical buildings to create a data set. The second stage involves generating a machine learning model for assessing the moisture content based on algorithms and the data collected in the first stage. The article is illustrated with an example of the application of the developed methodology to assess the moisture content of the brick walls of the Golden Gate building in Gdańsk. The presented example shows the reliability and practical usefulness of the developed methodology.