The article discusses machine learning methods, their application areas, limitations and application possibilities. Additionally highlighted achievements in deep learning, which allow obtaining accurate results with optimal time and effort. The promising architecture of neural networks of transformers is also described in detail. As an alternative approach, it is proposed to use a generative adversarial network in the process of converting a scan into elements of a digital information model.
Keywords: scanning, point cloud, information model, construction, objects, representation, neural network, machine learning
The article introduces a methodology for verifying BIM models of capital construction facilities. This approach focuses on dynamic assessment of intersection collision weights, combining geometric analysis, statistical methods, and adaptive metric weighting. Key metrics considered include granularity, geometry errors, tessellation complexity, and fill factor. The proposed methodology utilizes Python implementation with IfcOpenShell, leveraging a multithreaded architecture to significantly reduce data processing time. Testing on 20 multidisciplinary models highlights critical problematic elements such as walls, beams, and air ducts. The results demonstrate that adaptive weight distribution effectively identifies and prioritizes potential errors, improving the accuracy and reliability of BIM models. The study's findings are crucial for enhancing design and construction processes. By accurately assessing and mitigating errors, the methodology reduces project delays, cost overruns, and safety risks. It also promotes better coordination among project stakeholders, streamlining workflows and improving project outcomes. In conclusion, the proposed methodology is a valuable tool for verifying BIM models, ensuring the integrity and quality of capital construction projects. Its application can lead to more efficient, cost-effective, and reliable construction processes, benefiting both developers and end-users.
Keywords: TIM, collisions, verification, dynamic weights, adaptive metrics, algorithms, IfcOpenShell, python, standard deviation