Research on a collision detection method for large-scale tubular structures driven by dimensionality reduction

Efficient collision detection is of great significance for the design and analysis of large-scale complex tubular structures. However, traditional methods rely on 3D surface calculations or bounding boxes, which suffer from problems such as high computational costs, poor accuracy, and difficulties in balancing efficiency and accuracy. These issues are likely to cause design delays and safety risks. To address this, this paper constructs a centerline-parameterized collision detection dimensionality reduction method. Its core logic is derived from the “simplifying complexity” concept of “using numbers to govern shapes”, disassembling the three-dimensional geometric information of tubular components into two-dimensional center-line topologies and scalar radius parameters, thus achieving the separation of geometric forms and calculation rules. The proposed centerline-parameterized collision detection dimensionality reduction method is based on the parametric representation of tubular geometric features. Through the “center-line radius” parameterized strategy, it transforms 3D collision detection into 2D geometric distance calculations. It integrates a hierarchical Bounding Volume Hierarchy (BVH) acceleration structure to achieve coarse-grained filtering, and combines parameter-domain classification and systematic analysis to complete precise verification, solving the bottlenecks of traditional methods in high-dimensional calculations and iterative convergence. Experimental results in rebar collision scenarios show that the detection accuracy of this algorithm reaches 100 %, and its efficiency is an order of magnitude higher than that of traditional methods such as Open Cascade(OCC). It is applicable to mature tasks in fields such as construction and railways. Furthermore, in the future, consideration can be given to integrating the centerline-parameterized collision detection method—while retaining its main steps and core ideas—with AI models such as Transformer and NIF to achieve efficient collision detection in complex scenarios across multiple domains.