Abstract
Computerized maintenance management systems (CMMSs) of large commercial and institutional buildings archive work order logs. These logs contain invaluable operational performance indices pertaining to occupants and building systems. However, there are barriers to extract and visualize meaningful metrics from CMMSs. This study proposes a three-part framework to facilitate this process. The first part includes text mining techniques to preprocess and normalize the textual and other types of data. The second part develops benchmarking methods and metrics to evaluate the maintenance performance of buildings as well as to identify the buildings with most work orders on a complex. The last part adopts visualization methods to transform valuable information such as total work orders, the intensity of work orders of buildings and their change rates into more intuitive forms. To do so, a visualization method is implemented via building information model (BIM). Finally, the framework is applied to a case study to demonstrate how effectively the proposed methods perform in practice. The application of the framework can be used to quantify the distribution of WOs among buildings. For example, the case study revealed that 80 percent of all WOs originated from fewer than half of the buildings of a university campus.
Acknowledgment
The authors thank the Natural Sciences and Engineering Research Council of Canada (NSERC), Ontario Centres of Excellence (OCE), and Brookfield Global Integrated Solutions (BGIS) for funding this work. The authors would like to acknowledge Carleton Facilities Management and Planning for providing the CMMS database.