关键词: 纺纱机械, 振动, 动力学建模, 数据建模, 动力学与数据建模融合

Abstract:

Significance Spinning, as the foundation of the textile industry, directly affects the quality of textile products, where the stability of spinning machinery is a key factor. Excessive vibration can deteriorate yarn evenness and strength, reduce process consistency, and accelerate wear of key components, thereby lowering overall production efficiency. Currently, spinning machinery faces the challenge of increasing speed and improving quality. As operating speed increases, rotating parts and transmission systems are more likely to trigger resonance, amplification of dynamic loads, and stability degradation, which makes vibration control increasingly difficult in practice. Although the domestic spinning machinery has witnessed progress in terms of intelligence and speed, there remains a gap in operating speed compared to high-end equipment manufactured by the leading developed countries, with vibration issues being the main bottleneck. Therefore, research on vibration in spinning machinery is crucial, not only for the realization of high-speed and intelligent capabilities but also as the basis for fault diagnosis and predictive maintenance. Given the current scarcity of documents, this paper reviews and summarizes existing research methods, aiming to provide theoretical support for in-depth studies, and to envisage application prospects of digital twins in vibration research of spinning machinery.

Progress Firstly, the paper elucidates the limitations and integration requirements of conventional dynamic modeling and data modeling. The modeling of dynamic mechanisms has strong interpretability and low data dependence, but the modeling cost is high, the parameters are uncertain, and it is difficult to fully characterize the real working condition disturbances. Vibration data modeling is suitable for complex nonlinear and efficient modeling, but has weak interpretability and is highly dependent on data quality and quantity. Secondly, the application practices of fusion methods in other fields are summarized from different stages such as dataset construction, model training, and model output. Based on two aspects of work, a dynamic-data modeling fusion architecture for vibration research of spinning machinery was proposed, and the fusion architecture combining digital twin technology was explored. Subsequently, the performance differences of different fusion strategies were discussed and their applicability in different devices and operating conditions was analyzed. Finally, the significance of fusion methods for the intelligent development of textile machinery was summarized, and the future development directions in small sample prediction, complex working condition adaptation, and digital twin integration were discussed.

Conclusion and Prospect The vibration of spinning machinery directly affects the operational efficiency and product quality of textile production. Due to multi-source excitations, nonlinear contact behaviors, and time-varying operating conditions, vibration phenomena in spinning machinery are often complex and difficult to model accurately using a single paradigm. In conventional research methods, pure dynamic modeling requires strong theoretical assumptions, while pure data-driven methods rely on a large amount of data and are difficult to meet the current requirements for vibration control in high-speed, efficient, and intelligent spinning production. The dynamic-data modeling fusion method combines physical mechanisms with data modeling to enhance the interpretability of the model and its adaptability to complex dynamics, providing a new research path for vibration analysis of textile machinery. The dynamic-data modeling fusion method will further promote the intelligent development of textile machinery in terms of small sample faults, adaptive complex working conditions, and digital twin integration. By integrating mechanism knowledge and operational data, a highly adaptive quality control and fault warning system will be constructed. The dynamic-data modeling fusion method will play an increasingly important role in fault diagnosis, quality improvement, and performance optimization of textile machinery, promoting the textile manufacturing industry to move towards a new stage of higher quality and intelligence.

Key words: spinning machinery, vibration, dynamic modeling, data modeling, dynamic-data model integration