The importance of the quality of sampling in service life prediction

• Published in: Construction & building materials Vol. 66; pp. 19 - 29
• Main Authors: Galbusera, M.M., de Brito, J., Silva, A.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 15.09.2014; Elsevier B.V
• Subjects: Analysis; Architecture; Cladding; Construction costs; Construction materials; Design and construction; Durability; Estimates; Mathematical models; Metal cladding; Quality management; Quality of data; Sampling; Service life; Service life prediction models; Portugal; Service life prediction models; Quality of data; Durability
• ISSN: 0950-0618; 1879-0526
• DOI: 10.1016/j.conbuildmat.2014.05.045

•The research analyses the influence of the sample quality and size in service life prediction models.•The methodology adopted is based on field work data collection.•An empirical model is used that estimates the service life of the building elements.•The model is applied to evaluate the service life of ceramic external claddings.•The results confirm that more and better data conducts to more reliable results. In recent years the interest on the durability of buildings and their materials has been increasing. The reasons behind that attention are mainly economic and environmental. Service life prediction methods are an essential tool to support decision-making during the building’s life cycle. Service life prediction must be used to estimate the duration of a building component and allows a more rational definition of the maintenance operations, reducing unnecessary costs. However, service life prediction can be a complex issue. The results obtained are closely related to the quality of data used to define these models. In this study, the influence of the sample on the results of a service life prediction model applied to ceramic external walls is analysed. The methodology adopted is based on field work data collection, concerning the state of degradation of the façades. Two samples were used: the initial one, composed of 117 case studies, without a pre-selection of the cases analysed, and an improved sample, composed of 195 claddings, with more complete and reliable data. The results obtained confirm that more and better data conducts to more reliable results.