During the assembly process of the rear axle, the assembly quality and assembly efficiency decrease due to the accumulation errors of rear axle assembly torque. To deal with the problem, we proposed a rear axle assembly torque online control method based on digital twin. First, the gray wolf-based optimization variational modal decomposition and long short-term memory network (GWO-VMD-LSTM) algorithm was raised to predict the assembly torque of the rear axle, which solves the shortcomings of unpredictable non-stationarity and nonlinear assembly torque, and the prediction accuracy reaches 99.49% according to the experimental results. Next, the evaluation indexes of support vector machine (SVM), recurrent neural network (RNN), LSTM, and SVM, RNN, and LSTM based on gray wolf optimized variational modal decomposition (GWO-VMD) were compared, and the performance of the GWO-VMD-LSTM is the best. For the purpose of solving the insufficient information interaction capability problem of the assembly line, we developed a digital twin system for the rear axle assembly line to realize the visualization and monitoring of the assembly process. Finally, the assembly torque prediction model is coupled with the digital twin system to realize real-time prediction and online control of assembly torque, and the experimental testing manifests that the response time of the system is about 1 s. Consequently, the digital twin-based rear axle assembly torque prediction and online control method can significantly improve the assembly quality and assembly efficiency, which is of great significance to promote the construction of intelligent production line.

Multimodal sentiment analysis (MSA) has emerged as a prominent research frontier, enabling a comprehensive understanding of complex human emotions through the synergistic integration of heterogeneous multimodal signals. However, most existing approaches rely on idealized signal distribution assumptions, overlooking the detrimental impact of demographic bias on representation fairness and fusion robustness. This paper proposes a Label-Guided Contrastive Decoupling Fusion (LGCDF) framework that enhances model robustness to demographic bias by learning and fusing multimodal representations invariant to Sensitive Attributes (SAs). Specifically, the proposed LGCDF framework employs gender-sensitive attribute information as modality-level constraints to achieve language-centric cross-modal sentiment alignment, which is accomplished by computing contrastive losses between text–audio and text–visual feature pairs. Moreover, it introduces a SA-guided contrastive decoupling mechanism that decomposes multimodal representations into SA-related and -independent components. The SA-independent components are subsequently fused through a cross-modal attention fusion strategy, thereby facilitating fair sentiment representation and enabling efficient and robust multimodal information fusion. Extensive experimental results demonstrate that the proposed LGCDF framework achieves superior performance in fair representation learning and cross-modal information fusion while maintaining strong robustness under varying gender distribution biases.