Abstract:For the manufacture and assembly of the mechanical products, quality management is the key feature to improve the service performance, reduce the overall costs and enhance the sustainability of the manufacturing systems. While, with the emergence of advanced data-driven and digital twin technologies, the Zero-Defect Manufacturing (ZDM), which corrects, predicts, and prevents product defects based on multi-sources of data, is of great significance in improving assembly quality. As the specific application for the utilization of ZDM in product assembly, the critical performance index of the assembly results is predicted by taking into account the multi-source factors such as geometric deviation of the parts, material properties, assembly sequences and process boundary conditions during assembly. In this paper, we address the high computational cost and low computational efficiency of numerical simulation methods under multi-source factors, and propose a data-driven approach named DTA-VIT based on the fusion of heterogeneous variables for digital twin assembly modeling of products. Firstly, the geometric and performance variables of the assembly process are analyzed and modelled. Secondly, a multi-source assembly data fusion network under the Vision Transformer framework is developed. This network takes the parameter space, which fuses multi-source variables from the assembly process as input and the assembly result as output. Finally, a case study of the assembly process of composite bolted joint structures in aircraft assembly is conducted to verify the effectiveness and feasibility of the proposed method. The methodology provides a solid foundation for subsequent assembly quality control and prevent by predicting assembly performance efficiently, ultimately enabling the production of high-quality products.

Combination modeling of auto body assembly dimension propagation considering multi-source information for variation reduction

Propagation Analysis of Variation for Fuselage Structures in Multi-Station Aircraft Assembly

Variation modeling and analysis with interval approach for the assembly of compliant aeronautical structures

Variation modeling for fuselage structures in large aircraft digital assembly

Assembly Variation Prediction for Compliant Aeronautical Structures Using Fuzzy Interval Analysis

Variation analysis driven by small-sample data for compliant aero-structure assembly

A Multi-Object Posture Coordination Method with Tolerance Constraints for Aircraft Components Assembly

A Review on Variation Modeling of Aircraft Assembly

A generic approach for analysis of mechanical assembly

Auto-body Assembly Process Fault Diagnosis Based on a Dynamic Variation Modeling Approach

A Digital Twin-Based Method for Assembly Deviations Analysis

Realistic Geometry Based Feature Modeling of Complex Part and Its Application in Assembly Quality Analysis

A Digital Twin-Based Assembly Model for Multi-Source Variation Fusion on Vision Transformer

Simulation and Analysis of Assembly Processes Considering Compliant, Non-Ideal Parts and Tooling Variations

Active Variation Compensation in Vehicle Body Conceptual Assembly

Multiscale Cost-Sensitive Learning-Based Assembly Quality Prediction Approach under Imbalanced Data

Deviation Propagation Model of Assembly Sequence and Quality Evaluation Approach Based on Degree of Dimensional Variation

Variation Analysis for Multi-Flexible Autobody Assembly

Assembly Tolerance Analysis Based on the Jacobian Model and Skin Model Shapes

Assembly accuracy prediction based on CAD model

A Modification of DMVs Based State Space Model of Variation Propagation for Multistage Machining Processes