Building Maintenance Cost Estimation and Circular Economy: The Role of Machine-Learning

Abstract

The building industry generates large amounts of solid waste due to construction, maintenance, or demolition activities. Although the construction and demolition waste is well-documented in the literature, the maintenance phase could benefit from more academic research. Maintenance waste could be managed based on a linear economy or a circular economy. Since the existing linear economy poses significant problems, a transition to a circular economy is necessary. Such a transition is in its infancy stage and is hindered by multiple barriers. This paper developed a practical methodology for this transition and demonstrated its advantages to encourage its adoption by the building industry. These advantages included increasing the accuracy of building maintenance cost estimation, reducing the possibility of maintenance cost over-estimation, and reducing the waste of maintenance resources. These advantages were mostly due to replacing the deterministic and probabilistic maintenance cost estimation models with a machine-learning-based methodology. This methodology was shown to be versatile in several areas of application, i.e. anomaly detection, feature engineering, cost estimation, and validation. The methodology could be applied to any building type in any geographic location. It also facilitates achieving sustainable development goals #9 (through innovative building maintenance), #11 (via promotion of welfare-oriented building services), and #12 (by encouraging responsible consumption of maintenance resources).