Statistical models applied to service life prediction of rendered façades

An approach to the evaluation of the service life of rendered façades applying statistical tools is described. Using multiple linear regression analysis and artificial neural networks, mathematical models are established to estimate the degradation of this type of coating. To devise the models proposed, a sample of 100 rendered façades was subjected to meticulous field work to determine their condition. Some statistical parameters are used to evaluate the validity and efficacy of the models proposed. The service life of the sample of rendered façades is also evaluated, as estimated by the various models, and the result is expressed in histograms. The usefulness of these models to evaluate complex problems, such as the degradation phenomena of rendered façades, is thus demonstrated.

► This study presents statistical methodologies for service life prediction.
► The study is developed for the most common coating solution in Portugal.
► The methodology proposed is based on visual inspections of 100 external renders.
► The use of multiple linear regression and neural networks is discussed.
► The service life of renders is determined based on the numerical models proposed.