Metodologias automatizadas para geração de modelos geométricos a partir de nuvem de pontos: Estudo de caso

Abstract

This work addresses the challenge posed by the lack of project documentation, such as original plans and "as-built" drawings, for old dams like the Mascarenhas de Moraes Hydroelectric Power Plant (1957), a situation that compromises the safety and maintenance management required by current legislation. Scan-to-BIM technology is recognized as a solution for the reverse engineering of these structures, however, manual modeling methods are notoriously laborious and the automation methodologies present in the literature do not cover the geometric structures of dams. It was proposed to analyze Scan-to-BIM automation methodologies and, from this analysis, to develop and present a hybrid and optimized workflow for the strictly geometric reconstruction of the dam, seeking greater efficiency and precision in comparison with manual approaches. A hybrid strategy is adopted that correlates two distinct methodologies to specific geometries: for the arch, with a continuous and extrudable geometry, the Scan-to-BIM-to-Sim methodology is adapted, based on the extraction of cross-sections; for the spillway, with a segmented geometry analogous to a building, the Cloud2BIM methodology is applied, focused on the detection of elements (slabs, walls, openings) through point density analysis. The main result confirms that the intrinsic geometry of the structural component is the overriding factor in the selection of the most appropriate automation technique. The theoretical application of the hybrid approach resulted in the proposal of a cohesive framework, which demonstrates greater efficiency than the application of a single method for the complete structure, by directing the most appropriate reconstruction logic for each part. It is concluded that the hybrid approach constitutes the most efficient and precise strategy for the digital reconstruction of dams with complex geometries, with the proposed workflow serving as a methodological guide that optimizes resources, reduces the need for manual intervention, and ensures the generation of a high-fidelity geometric model, fundamental for future safety, operation, and maintenance analyses.