Mechanical, electrical, and plumbing (MEP) works constitute a large portion of construction costs and thus need to be appropriately tracked. Assessment of the built status of MEP works in construction projects is however typically limited to subcontractor claims augmented and contrasted with periodic manual inspection. A more detailed manual inspection is costly and not considered worthwhile on most projects. Within a Scan-versus-BIM object recognition framework, three-dimensional laser scanning and project 3D/4D BIM models jointly offer the opportunity for a frequent, detailed, and semantically rich assessment of as-built status of construction projects at a cost that continues to decline. This potential has already been demonstrated for tracking structural works, but remains to be assessed in regard to other work sections, in particular MEP works. This paper explores that opportunity. A Scan-vs-BIM processing system is described with some enhancements over previous works. It is then tested with a representative and challenging case study of the construction of a utility corridor in a university engineering building. The results indicate that the proposed system is significantly challenged when tracking MEP systems constructed using traditional on-site fabrication, due to changes or adjustments made on-site that lead to actual component layouts varying in comparison to designed layouts. This performance could be revisited in cases where off-site prefabrication and preassembly is implemented. The results nonetheless lead the authors to propose a novel data processing system (conceptually described in this paper) integrating Scan-vs-BIM and Scan-to-BIM approaches. This system should provide superior performance over existing systems, enabling automated and robust quality control (including the estimation of the emerging performance metric percent built as designed ) and delivery of true as-built BIM models to facility owners and managers.
|Number of pages||28|
|Journal||Journal of Computing in Civil Engineering|
|Publication status||Published - Jul 2014|