Autonomous Generation of Equipment and Piping System Layouts With Routing Optimization

Design and construction of nuclear piping systems rely on traditional workflows that are labor-intensive, error-prone, and poorly integrated across software platforms. Current practices depend on manual translation of 2D drawings into 3D models, causing inefficiencies, errors, and poor software interoperability. This work presents an autonomous framework that transforms 2D piping designs into 3D building information models (BIMs) for integrated equipment and piping layouts. The framework integrates model-based systems engineering and computational modeling to autonomously generate geometric models from structured data schemas. Equipment definitions, connectivity, spatial constraints, and routing requirements are represented systematically and transformed into Industry Foundation Classes (IFC)-based BIM geometry. Artificial intelligence (AI)-assisted piping optimization was developed to maintain connectivity and layout rules. Open-source technologies and standardized formats were employed to ensure extensibility and reduce proprietary software dependence for 3D BIM development, and a custom web-based application was designed to support input creation, validation, and model generation through visualization. A case study using a one-loop reactor coolant system and secondary coolant system was conducted to demonstrate successful autonomous generation of 3D equipment and piping system models. Compared to manual BIM modeling workflows, this proposed framework can significantly reduce modeling effort during early design phases.