As the production mode of aerospace equipment changes to intelligence, the aligning of cylindrical components increasingly adopts digital twin systems for assembly. However, measurement and transmission processes can affect the accuracy in pose adjustment and positioning. Notably, in the process of automatic aligning, conventional optimization methods fail to compensate for globally accumulated errors in multi-system collaboration, while the absence of real-time monitoring and online optimization further undermines aligning quality and efficiency. To address this challenge, a digital twin-driven online positioning accuracy optimization method is proposed for cylindrical components aligning. In this study, a digital twin system framework for cylindrical components aligning is first established, incorporating a closed-loop control methodology. Subsequently, the method systematically investigates error factors in multi-system coordination, including modeling and analysis of their impacts on positioning accuracy. Key innovations involve an iterative Singular Value Decomposition-based measurement pose optimization algorithm, a mechanism-data fusion-driven online prediction model for alignment errors, and a geometric-analytical cross-system transformation method for precise actuator parameter calculation. By integrating online precision refinement algorithms and predictive error compensation mechanisms, the cumulative positioning errors are effectively mitigated, enhancing dynamic aligning accuracy control capabilities. Experimental validation using a prototype system demonstrates that the method improve the aligning accuracy by 70.77% and shortens the docking cycle time by 53.10%, effectively improving both docking and efficiency, and verifying the correctness and effectiveness of the proposed method.

Facing the challenge of frequent changes on aircraft models, it is necessary to achieve rapid design for aircraft assembly tooling. Aircraft product assembly tooling design is a long-term and high-cost tasks. Designers invest significant time and effort in tooling design, heavily relying on personal experience, with limited support and inheritance of tooling design knowledge. To address this issue, this study proposed an assembly tooling design knowledge recommendation approach based on knowledge graph and user behavior feedback. First, a knowledge document repository and a domain knowledge label network are constructed, forming a knowledge graph. Then, by constructing and activating a knowledge push scenario model, the priority of knowledge documents is computed based on their relevance between the knowledge push scenario and knowledge tags. Additionally, a knowledge push adjustment mechanism is proposed, and the priority of the knowledge documents will be adjusted based on user behaviors. Using aircraft wall panel tooling design as an example, this study validates the effectiveness of the user behavior-based parameter adjustment mechanism in knowledge recommendation. The example demonstrates that the accuracy of the proposed method in pushing the top 5 relevant knowledge documents is 0.83. Results also show that tags with higher search frequency are ranked higher, and user ratings of “helpful” significantly increase the priority of recommended knowledge under the same recommendation frequency. Additionally, a prototype system for aircraft panel assembly tooling design knowledge push is developed on the CATIA platform, applying the method to facilitate designers' efficient access to relevant tooling knowledge.

Assembly interfaces, the joint surfaces between the vertical tail and rear fuselage of a large aircraft, are thin-wall components. Their machining quality are seriously restricted by the machining vibration. To address this problem, an in-process adaptive milling method is proposed for the large-scale assembly interface driven by real-time machining vibration data. Within this context, the milling operation is first divided into several process steps, and the machining vibration data in each process step is separated into some data segments. Second, based on the real-time machining vibration data in each data segment, a finite-element-unit-force approach and an optimized space–time domain method are adopted to estimate the time-varying in-operation frequency response functions of the assembly interface. These FRFs are in turn employed to calculate stability lobe diagrams. Thus, the three-dimensional stability lobe diagram considering material removal is acquired via interpolation of all stability lobe diagrams. Third, to restrain milling chatter and resonance, the cutting parameters for next process step, e.g., spindle speed and axial cutting depth, are optimized by genetic algorithm. Finally, the proposed method is validated by a milling test of the assembly interface on a vertical tail, and the experimental results demonstrate that the proposed method can improve the machining quality and efficiency of the assembly interface, i.e., the surface roughness reduced from 3.2 μm to 1.6 μm and the machining efficiency improved by 33%.