Automating Bridge Construction Scheduling Data with BIM and Machine Learning

Abstract

Conventional construction scheduling techniques often fall short due to inefficiencies, the prevalence of human error, and the lack of reliable scheduling data. Despite efforts to obtain better scheduling data and automate the process using Building Information Modelling (BIM) as a primary source of information, challenges persist. BIM models contain inconsistencies that can compromise their scheduling effectiveness. Moreover, existing automation efforts face obstacles, including complexity and scalability issues. These efforts frequently overlook the intricacies of accurately modelling construction tasks, neglecting resource considerations, subprocesses, and structured information management, which further complicates the creation of reliable and efficient construction schedule data. This research proposes a framework designed to systematically obtain scheduling data in an automated manner, aligning with a Work Breakdown Structure (WBS) to ensure organisational clarity. By incorporating subprocesses and focusing on resource constraint calculations and balancing. Furthermore, the framework structures scheduling data into a CSV format, facilitating easier analysis and integration with project management tools. The research follows the Design Science Research methodology. This multi-stage framework integrates BIM data extraction, automated element labelling, creation of a custom WBS, construction sequence, and resource balancing. The framework was developed using Python coding and libraries, facilitating seamless transitions between stages without manual intervention. The framework's performance is evaluated using 4D BIM software, assessing the generated data and logic in a girder bridge projects. This innovative framework enhances the accuracy and efficiency of construction data scheduling through automation. It reduces manual intervention, organises data effectively, and improves project timeline reliability. The use of 4D BIM further illustrates the practical application of this data in bridge projects, showcasing a scalable and robust contribution to construction automation.