Digital twin-based high-precision assembly method for large-diameter cabin section docking

The docking of the nozzle and combustion chamber is a critical process in the assembly of solid rocket motors. Currently, docking assembly largely relies on manual operations, which constrains production efficiency. To address this, a digital twin-based high-quality assembly method for large-diameter cabin sections is proposed. First, a digital twin model is established using 3D scanning of the physical components. To tackle the challenge of detecting complex 3D assembly features, a feature target-guided digital modeling approach is introduced. By incorporating geometrically marked targets and constructing a digital model with identifiable feature points, the reconstruction of complex feature information is achieved. Subsequently, a collaborative virtual–physical assembly method is developed, employing a multi-camera system for online tracking of the physical components via the digital model. An elliptical cone model is constructed to resolve the misalignment between the center of a perspectively projected ellipse and the actual spatial circle center. The iterative closest point (ICP) based relative pose estimation method is applied to enhance the speed and accuracy of multi-camera vision pose measurement. Finally, a platform was built to validate the accuracy and efficiency of the proposed method. Experimental results demonstrate that the digital twin-based assembly approach enables high-efficiency assembly, while also allowing online monitoring and collision warning in visually obscured areas during the process.