A probabilistic model to quantify uncertainty in building inspection and diagnosis

Clara Pereira; Ana Silva; Cláudia Ferreira; Jorge de Brito; Inês Flores-Colen; José D. Silvestre

Key information

Authors:

Clara Pereira (Clara Isabel Fernandes Pereira); Ana Silva (Ana Filipa Ferreira da Silva Cigarro Matos); Cláudia Ferreira (Cláudia Alexandra Rocha Ferreira); Jorge de Brito (Jorge Manuel Caliço Lopes de Brito); Inês Flores-Colen (Inês Flores-Colen); José D. Silvestre (José Dinis Silvestre)

Published in

October 2021

Abstract

Uncertainty is intrinsic to the inspection and diagnosis process of building elements, even though it is seldom quantified. Uncertainty may derive from observation conditions, adopted methods and communication. The objective of this research is to raise awareness among surveyors about the weight of uncertainty on their activities. Thus, a quantification model is proposed, namely a Bayesian network representing the degradation mechanisms of rendered façades. The results of using such a model provide probabilities of occurrence of defects and their causes given input data. Low probabilities, such as those of mechanical action causes, may indicate the need of using additional tests to confirm or refute inferences deriving from visual observation only. In the future, the model may be expanded and similar models may be developed for other materials.