Machine Learning-Based Building Life-Cycle Cost Prediction: A Framework and Ontology

Numerous costs are associated with the design, construction, installation, operation, maintenance, and deconstruction of a building or building system. One of the challenges usually faced by an organization's capital planning department and/or facility management department is that they do not have an effective means to quickly estimate a new facility's whole life-cycle costs (LCC) during the programming phase when no building design is available. To provide facility managers and owners with an effective and reliable means to assess the total cost of the facility ownership, the authors are developing an approach that uses the historical data stored in multiple building systems and building information models (BIM) as basis to predict facilities' LCC - initial design and construction cost, utility cost, and operation and maintenance cost. In this paper, the authors propose a machine learning-enabled facility LCC analysis framework using data provided by building systems. The corresponding domain ontology - LCCA-Onto - is also presented. The proposed approach provides organizations who own multiple facilities with an innovative solution to the LCC prediction issue.