POD和DMD方法对燃料棒流致振动特性的分析

Analysis of Flow-induced Vibration Characteristics of Fuel Rods Using POD and DMD Methods

燃料棒的流致振动是造成燃料失效和反应堆停堆的重要因素之一,研究其流致振动机理非常必要。本文建立了燃料棒的高保真有限元模型,基于ANSYS Batch和随机振动理论批量计算了不同支撑刚度下燃料棒的流致振动响应,并搭建了1个流致振动响应数据库。该数据库采用不同支撑刚度值来模拟和检查燃料棒可能遇到的情况。基于数据库组装快照矩阵,采用两种数据驱动方法(POD方法和DMD方法)搭建了1个高保真降阶模型(ROM),该ROM能够实现流致振动响应的快速重构。对比了POD方法和DMD方法重构燃料棒流致振动响应的效果以及各自的优点和缺点。研究发现:使用相同数量的模态重构燃料棒的流致振动响应时,POD方法的重构效果优于DMD方法;基于相同数量的模态重构时,刚度为1 N/mm的重构效果最差,这主要是因为刚度越小,频率越小,系统的振幅越大,需要更多数量的模态才能捕捉大幅振动带来的影响;DMD方法不仅能够高效地重构燃料棒的振动响应,还能够判断每个DMD模态的稳定性。

The exploration of flow-induced vibration in fuel rods is of utmost importance as it plays a pivotal role in comprehending and mitigating factors that contribute to fuel failure and reactor shutdown.This understanding is crucial for advancing the nuclear energy industry.To unravel the intricacies of this complex phenomenon,a sophisticated high-fidelity finite element model of fuel rods was meticulously constructed.This model serves as the cornerstone for a computational analysis of flow-induced vibration responses utilizing ANSYS Batch,firmly roots in the principles of random vibration theory.In the pursuit of a profound understanding,support stiffness values to simulate and scrutinize diverse scenarios that fuel rods might encounter were systematically varied.The outcomes of these simulations were meticulously compiled to establish a comprehensive database of flow-induced vibration responses.This database stands as a valuable resource for future research endeavors and engineering applications within the nuclear energy domain.Taking a step forward,an innovative approach was adopted to streamline the analysis process.Leveraging the snapshot matrix derived from the extensive database,a high-fidelity reduced-order model(ROM) was developed.Two data-driven methodologies,namely proper orthogonal decomposition(POD) and dynamic mode decomposition(DMD) methods,were employed to construct the ROM.This ROM enables the rapid reconstruction of flow-induced vibration responses,facilitating efficient analysis and decision-making in real-world applications.A critical facet of this paper involves a comparative analysis of the reconstruction effectiveness of fuel rod flow-induced vibration responses using both the POD and DMD methods.The results of this comparison reveal that,when considering the reconstruction of vibration responses with the same number of modes,the POD method outperforms the DMD method.This finding underscores the importance of selecting appropriate methodologies based on specific objectives and computational efficiency.Furthermore,the results indicate that the DMD method excels not only in efficiently reconstructing the vibration responses of fuel rods but also offers the unique capability to assess the stability of each DMD mode.This dual functionality enhances the overall diagnostic capabilities of the DMD method,providing valuable insights into the dynamic behavior and potential instabilities of the fuel rod system.In conclusion,the exhaustive investigation outlined in this paper not only significantly contributes to the understanding of flow-induced vibration characteristics in fuel rods but also provides a robust framework for developing advanced models and methodologies for future research and practical applications in the nuclear energy industry.The comprehensive nature of our approach ensures that our findings are not only insightful but also applicable in shaping the future of nuclear energy technology.