Service life prediction of residential interior finishes for life cycle assessment

Purpose Service life of building products has an important influence on life cycle assessment (LCA) results of buildings. The goal of this study was to propose a systematic approach to estimate service life of building products by including both technical and social factors. Methods A hybrid service life prediction method, combining the statistical approach described in American Society for Testing and Materials standard G166 with the Factor Method adopted by International Organization for Standardization standard 15686, was proposed. In their current forms, the two methods are not suitable to provide reliable lifetime estimates for the wide variety of products that are used in buildings. Statistical analysis was preferred over a deterministic approach. Regression analysis was used to define Weibull distribution parameters for each product. These distributions were then used to calculate the mean estimated service life of products with an 80% confidence interval. Using actual lifetime observed from practice instead of design lifetime for reference service life was preferred. This enables the use of a smaller range of coefficients for each factor affecting service life, which decreases subjectivity and increases reliability of results. Results and discussion Example median service life estimates were demonstrated for common residential interior finishes that are replaced more frequently, and therefore require more maintenance planning and potentially have significant environmental impacts. Probability of renovation distributions was also presented for interior finishes. The proposed method inherently includes social factors in the dataset used to define lifetime distributions, which could be as important as durability for some building product categories. Another advantage is that choosing reference service life based on real-life conditions decreases the range of coefficients necessary for modifying factors in comparison to when design lifetime is used, thus decreasing the subjectivity of results due to variations in assigned values by different users. Conclusions The Factor Method is the most promising method available to estimate service life of products. Unless additional data points were gathered for investigated products, the presented lifetime distribution results can be directly applied to LCA studies. The proposed hybrid method can also be applied to other products that are studied within the Factor Method. Products whose lifetimes are influenced by social factors are prime candidates to apply this method.