Towards the generation of digital twins for facility management based on 3D point clouds

Advances and adaptation of Industry 4.0 practices in Facility Management (FM) have created demand for use of up-to-date digitized building assets. The use of Building Information Modelling (BIM) for FM in the Operation and Maintenance (O&M) stages of the building lifecycle is intended to bridge the gap between operations and digital data, but lacks the functionality of assessing and forecasting the state of the built environment in real-time. To accommodate this, BIM data needs to be constantly updated with the current state of the built environment. However, generation of <em>as-is</em> BIM data for a digital representation of a building is a labor intensive process. While some software applications offer a degree of automation for the generation of <em>as-is</em> BIM data, they can be impractical to use for routinely updating FM digital documentation. Current approaches for capturing the state of the built environment using remote sensing and photometry-based methods allow for the creation of point clouds that can be used as the basis for <em>digital twins, </em>along with existing BIM and FM documentation. Point clouds themselves do not contain any specific information about the building components they represent physically, but using machine learning methods they can be reconstructed as a BIM or used by themselves for assessment of the physical environment. Additional layers of information such as sensor data analytics can also be included. We present and describe our current research and development progress of a service-oriented platform that learns to interpret the contents presented by captured point clouds. We place a specific focus on the reconstruction and visualization of the captured state of the built environment for increasing FM stakeholder engagement and facilitating collaboration.