Process Pattern Recognition for Building Information Models:A Case Study

Abstract

Background: The fourth dimension of building information modeling (BIM) plays a significant role in construction planning by linking the construction schedule to the existing building information model. However, a difficulty may arise concerning the ease with which a link can be made and modified. Pattern-based techniques that search for recurring processes can help eliminate this drawback by producing predefined process templates. Methods: This paper critically examines the applicability of generating pre-defined process templates for BIM-based schedules using process pattern recognition techniques to reduce the effort of defining the construction schedules and integrating the templates into the BIM data. The technique estimates the level of recurrence of certain tasks within a schedule by applying several metrics. A real-life construction schedule from a housing project in the State of Kuwait was decomposed into numerous sub-schedules based on a set criterion to estimate the level of recurrence, the sub-schedules were compared based on structural and contextual similarities. Results and Discussion: The generated process templates produced demonstrated that the approach is ideal for projects with repetitive processes and that by utilizing the templates, linking building elements to tasks in 4D modeling can be greatly facilitated, thereby reducing the planning time. Conclusion: This study corroborates the results of previous literature, which found that the improvement in the efficiency of construction planning could be achieved by applying reusable process templates. The generated templates should enable the idea of storing the process templates in data banks for detecting undesired regularities in previous schedules in preparation for future use.