Accurately diagnosing bearing faults is crucial for maintaining the efficient operation of rotating machinery. However, traditional diagnosis methods face challenges due to the diversification of application environments, including cross-condition adaptability, small-sample learning difficulties, and cross-dataset generalization. These challenges have hindered the effectiveness and limited the application of existing approaches. Large language models (LLMs) offer new possibilities for improving the generalization of diagnosis models. However, the integration of LLMs with traditional diagnosis techniques for optimal generalization remains underexplored. This paper proposed an LLM-based bearing fault diagnosis framework to tackle these challenges. First, a signal feature quantification method was put forward to address the issue of extracting semantic information from vibration data, which integrated time and frequency domain feature extraction based on a statistical analysis framework. This method textualized time-series data, aiming to efficiently learn cross-condition and small-sample common features through concise feature selection. Fine-tuning methods based on LoRA and QLoRA were employed to enhance the generalization capability of LLMs in analyzing vibration data features. In addition, the two innovations (textualizing vibration features and fine-tuning pre-trained models) were validated by single-dataset cross-condition and cross-dataset transfer experiment with complete and limited data. The results demonstrated the ability of the proposed framework to perform three types of generalization tasks simultaneously. Trained cross-dataset models got approximately a 10% improvement in accuracy, proving the adaptability of LLMs to input patterns. Ultimately, the results effectively enhance the generalization capability and fill the research gap in using LLMs for bearing fault diagnosis.
{"title":"LLM-based framework for bearing fault diagnosis","authors":"Laifa Tao, Haifei Liu, Guoao Ning, Wenyan Cao, Bohao Huang, Chen Lu","doi":"10.1016/j.ymssp.2024.112127","DOIUrl":"https://doi.org/10.1016/j.ymssp.2024.112127","url":null,"abstract":"Accurately diagnosing bearing faults is crucial for maintaining the efficient operation of rotating machinery. However, traditional diagnosis methods face challenges due to the diversification of application environments, including cross-condition adaptability, small-sample learning difficulties, and cross-dataset generalization. These challenges have hindered the effectiveness and limited the application of existing approaches. Large language models (LLMs) offer new possibilities for improving the generalization of diagnosis models. However, the integration of LLMs with traditional diagnosis techniques for optimal generalization remains underexplored. This paper proposed an LLM-based bearing fault diagnosis framework to tackle these challenges. First, a signal feature quantification method was put forward to address the issue of extracting semantic information from vibration data, which integrated time and frequency domain feature extraction based on a statistical analysis framework. This method textualized time-series data, aiming to efficiently learn cross-condition and small-sample common features through concise feature selection. Fine-tuning methods based on LoRA and QLoRA were employed to enhance the generalization capability of LLMs in analyzing vibration data features. In addition, the two innovations (textualizing vibration features and fine-tuning pre-trained models) were validated by single-dataset cross-condition and cross-dataset transfer experiment with complete and limited data. The results demonstrated the ability of the proposed framework to perform three types of generalization tasks simultaneously. Trained cross-dataset models got approximately a 10% improvement in accuracy, proving the adaptability of LLMs to input patterns. Ultimately, the results effectively enhance the generalization capability and fill the research gap in using LLMs for bearing fault diagnosis.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"12 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142665623","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-13DOI: 10.1016/j.ymssp.2024.112112
Yeongtae Jang, Eunho Kim, Jinkyu Yang, Junsuk Rho
A novel approach to metamaterial design is introduced through the development of a stable 3D woodpile structure composed of slender cylindrical beams. These beam elements possess diverse bending vibration modes, intricately coupled with propagating waves, leading to complex wave dynamics within the structure. For the efficient analysis of various architectures, an extended discrete element model (DEM) is introduced to accurately emulate the local resonance caused by the beam’s bending vibration modes. The high level of accuracy achieved by the DEM is attributed to the utilization of a physics-informed discrete element modeling approach, rooted in continuum beam theory and wave dynamics within periodic structures. Utilizing the extended DEM, the interplay between propagating waves and local resonance within the beams is investigated, and the adjustability of mode coupling is confirmed by altering the interacting positions of neighboring beams. Subsequent to this, a graded 3D woodpile architecture is designed to progressively superimpose multiple frequency band structures. By adjusting mode coupling, it is shown that the graded woodpile is capable of displaying either a broad frequency passband or a broad frequency bandgap. Further demonstration reveals that the broad frequency bandgap facilitates high-frequency filtering, which effectively attenuates impact waves without the need for additional damping. The stable 3D woodpile architecture proposed in this study shows great potential for practical applications in vibration filtering and impact mitigation across various domains, ranging from small-scale material design to large-scale constructions.
{"title":"Sculpt wave propagation in 3D woodpile architecture through vibrational mode coupling","authors":"Yeongtae Jang, Eunho Kim, Jinkyu Yang, Junsuk Rho","doi":"10.1016/j.ymssp.2024.112112","DOIUrl":"https://doi.org/10.1016/j.ymssp.2024.112112","url":null,"abstract":"A novel approach to metamaterial design is introduced through the development of a stable 3D woodpile structure composed of slender cylindrical beams. These beam elements possess diverse bending vibration modes, intricately coupled with propagating waves, leading to complex wave dynamics within the structure. For the efficient analysis of various architectures, an extended discrete element model (DEM) is introduced to accurately emulate the local resonance caused by the beam’s bending vibration modes. The high level of accuracy achieved by the DEM is attributed to the utilization of a physics-informed discrete element modeling approach, rooted in continuum beam theory and wave dynamics within periodic structures. Utilizing the extended DEM, the interplay between propagating waves and local resonance within the beams is investigated, and the adjustability of mode coupling is confirmed by altering the interacting positions of neighboring beams. Subsequent to this, a graded 3D woodpile architecture is designed to progressively superimpose multiple frequency band structures. By adjusting mode coupling, it is shown that the graded woodpile is capable of displaying either a broad frequency passband or a broad frequency bandgap. Further demonstration reveals that the broad frequency bandgap facilitates high-frequency filtering, which effectively attenuates impact waves without the need for additional damping. The stable 3D woodpile architecture proposed in this study shows great potential for practical applications in vibration filtering and impact mitigation across various domains, ranging from small-scale material design to large-scale constructions.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"1 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142665638","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-12DOI: 10.1016/j.ymssp.2024.112132
Yingying Fan, Xin Liu, Dong F. Wang
A new concept of vibration synergized energy harvesting is proposed for ultra-low frequency scenarios. A dual-oscillator synergized piezoelectric energy harvester (DOS-PEH), inspired by hummingbirds, is designed to demonstrate the new concept, both theoretically and experimentally. Mimicking the synergy mechanism of hummingbird muscles and wings, the DOS-PEH adopts a supporting oscillator (SO) and a buckled beam designated as the dominating oscillator (DO) to synergize the vibrations through magnetic coupling. SO engenders a hinge-support-like configuration at the beam midspan, by which DO exhibits three stable equilibrium positions while taking on four stable equilibrium states, including two second buckling modes that lower snapping force to facilitate snap-through oscillations. The non-contact magnetic force, introduced by SO, acts as a link that cohesively connects the dual oscillators. It enables continuous vibration transmission from the ambient environment, through SO, and ultimately to DO. A fresh bandwidth, extending from 2.5 to 10 Hz, of 7.5 Hz emerges under 0.4 g excitation. The DOS-PEH, in general, achieves the broadband, stable, and progressively improving voltage output across the ultra-low frequency range. Further, the output voltage of the DOS-PEH is about 70 times higher than that of the collision-based piezoelectric energy harvester (C-PEH), and the operational bandwidth is broadened to 136 %. It highlights the contribution of synergistic vibration to the ultra-low-frequency energy harvesting.
{"title":"A hummingbird-inspired dual-oscillator synergized piezoelectric energy harvester for ultra-low frequency","authors":"Yingying Fan, Xin Liu, Dong F. Wang","doi":"10.1016/j.ymssp.2024.112132","DOIUrl":"https://doi.org/10.1016/j.ymssp.2024.112132","url":null,"abstract":"A new concept of vibration synergized energy harvesting is proposed for ultra-low frequency scenarios. A dual-oscillator synergized piezoelectric energy harvester (DOS-PEH), inspired by hummingbirds, is designed to demonstrate the new concept, both theoretically and experimentally. Mimicking the synergy mechanism of hummingbird muscles and wings, the DOS-PEH adopts a supporting oscillator (SO) and a buckled beam designated as the dominating oscillator (DO) to synergize the vibrations through magnetic coupling. SO engenders a hinge-support-like configuration at the beam midspan, by which DO exhibits three stable equilibrium positions while taking on four stable equilibrium states, including two second buckling modes that lower snapping force to facilitate snap-through oscillations. The non-contact magnetic force, introduced by SO, acts as a link that cohesively connects the dual oscillators. It enables continuous vibration transmission from the ambient environment, through SO, and ultimately to DO. A fresh bandwidth, extending from 2.5 to 10 Hz, of 7.5 Hz emerges under 0.4 g excitation. The DOS-PEH, in general, achieves the broadband, stable, and progressively improving voltage output across the ultra-low frequency range. Further, the output voltage of the DOS-PEH is about 70 times higher than that of the collision-based piezoelectric energy harvester (C-PEH), and the operational bandwidth is broadened to 136 %. It highlights the contribution of synergistic vibration to the ultra-low-frequency energy harvesting.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"7 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142665639","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Monitoring and nowcasting of mooring line are of paramount significance for maintaining the stability of floating structure. Recently, data-driven approaches for mooring monitoring have been proposed to identify potential mooring damage, aiming to achieve digital real-time integrity management. This paper proposes a framework for detection and nowcasting of health status of mooring line. The framework can identify multiple damage locations and degrees of mooring line, as well as various complicated multi-coupled scenarios. Our proposed method does not rely on experience-based manual feature extraction in all existing studies, but instead uses fully automatic sequence input, retaining complete series information and pattern recognition, which helps the model comprehensively grasp mooring deterioration patterns. Most existing methods simplify the problem by ignoring randomness and inherent noise in environments. In this paper, we account for the potential randomness and uncertainty of the data source during model construction, enhancing generalizability and noise resistance. Given the time series nature of the input variables, we have designed a novel ConvMamba architecture that integrates the convolutional layers and Mamba block, which includes multiple modules and selective state space model. This design ensures the architecture maintains the recurrent framework characteristic of RNNs while also benefiting from the parallel computing capabilities of CNNs. After ablation experiments and comparisons with other existing sequence models, the superiority of proposed architecture is demonstrated in both accuracy and efficiency. Furthermore, model maintains impressive noise-resistant accuracy under high interference from three different types of noise experiments, attributable to the robust model design. For the practical applications, two strategies are proposed to improve the original model and bolster noise resistance. While these strategies have certain limitations, they offer potential for further optimization.
{"title":"Local damage identification and nowcasting of mooring system using a noise-robust ConvMamba architecture","authors":"Yixuan Mao, Menglan Duan, Hongyuan Men, Miaozi Zheng","doi":"10.1016/j.ymssp.2024.112092","DOIUrl":"https://doi.org/10.1016/j.ymssp.2024.112092","url":null,"abstract":"Monitoring and nowcasting of mooring line are of paramount significance for maintaining the stability of floating structure. Recently, data-driven approaches for mooring monitoring have been proposed to identify potential mooring damage, aiming to achieve digital real-time integrity management. This paper proposes a framework for detection and nowcasting of health status of mooring line. The framework can identify multiple damage locations and degrees of mooring line, as well as various complicated multi-coupled scenarios. Our proposed method does not rely on experience-based manual feature extraction in all existing studies, but instead uses fully automatic sequence input, retaining complete series information and pattern recognition, which helps the model comprehensively grasp mooring deterioration patterns. Most existing methods simplify the problem by ignoring randomness and inherent noise in environments. In this paper, we account for the potential randomness and uncertainty of the data source during model construction, enhancing generalizability and noise resistance. Given the time series nature of the input variables, we have designed a novel ConvMamba architecture that integrates the convolutional layers and Mamba block, which includes multiple modules and selective state space model. This design ensures the architecture maintains the recurrent framework characteristic of RNNs while also benefiting from the parallel computing capabilities of CNNs. After ablation experiments and comparisons with other existing sequence models, the superiority of proposed architecture is demonstrated in both accuracy and efficiency. Furthermore, model maintains impressive noise-resistant accuracy under high interference from three different types of noise experiments, attributable to the robust model design. For the practical applications, two strategies are proposed to improve the original model and bolster noise resistance. While these strategies have certain limitations, they offer potential for further optimization.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"226 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142665646","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-12DOI: 10.1016/j.ymssp.2024.112113
Fangqi Hong, Pengfei Wei, Sifeng Bi, Michael Beer
As a main task of inverse problem, model updating has received more and more attention in the area of inspection, sensing, and monitoring technologies during the recent decades, where the estimation of posterior probability density function (PDF) of unknown model parameters is still challenging for expensive-to-evaluate models of interest. In this paper, a novel variational Bayesian inference method is proposed to approximate the real posterior PDF of unknown model parameters by using Gaussian mixture model and measurement responses. A Gaussian process regression model is first trained for approximating the logarithm of the product of likelihood function and prior PDF, with which, another Gaussian process model is induced for approximating the expensive evidence lower bound (ELBO). Then, two Bayesian numerical methods, i.e., Bayesian optimization and Bayesian quadrature, are combined sequentially as a novel Bayesian active learning method for searching the global optima of the parameters of the variational posterior density. The proposed method inherits the advantages of both Bayesian numerical methods, which includes good global convergence, much less number of simulator calls, etc. Three examples, including the dynamic model of a two degrees of freedom structures, the lubrication model of a hybrid journal bearing, and the dynamic model of an airplane structure, are introduced for demonstrating the relative merits of the proposed method. Results show that, given desired requirement of numerical accuracy, the proposed method is more efficient than the parallel methods.
近几十年来，作为逆问题的一项主要任务，模型更新在检测、传感和监控技术领域受到越来越多的关注，而对于昂贵的相关模型而言，未知模型参数的后验概率密度函数（PDF）估计仍是一项挑战。本文提出了一种新颖的变分贝叶斯推理方法，利用高斯混合模型和测量响应来逼近未知模型参数的真实后验概率密度函数。首先训练一个高斯过程回归模型来逼近似然函数与先验 PDF 乘积的对数，然后诱导另一个高斯过程模型来逼近昂贵的证据下限（ELBO）。然后，两种贝叶斯数值方法，即贝叶斯优化和贝叶斯正交，被依次组合成一种新的贝叶斯主动学习方法，用于搜索变分后验密度参数的全局最优值。所提出的方法继承了这两种贝叶斯数值方法的优点，包括良好的全局收敛性、更少的模拟器调用次数等。本文介绍了三个实例，包括双自由度结构动态模型、混合轴颈轴承润滑模型和飞机结构动态模型，以展示所提方法的相对优势。结果表明，在数值精度要求较高的情况下，建议的方法比并行方法更有效。
{"title":"Efficient variational Bayesian model updating by Bayesian active learning","authors":"Fangqi Hong, Pengfei Wei, Sifeng Bi, Michael Beer","doi":"10.1016/j.ymssp.2024.112113","DOIUrl":"https://doi.org/10.1016/j.ymssp.2024.112113","url":null,"abstract":"As a main task of inverse problem, model updating has received more and more attention in the area of inspection, sensing, and monitoring technologies during the recent decades, where the estimation of posterior probability density function (PDF) of unknown model parameters is still challenging for expensive-to-evaluate models of interest. In this paper, a novel variational Bayesian inference method is proposed to approximate the real posterior PDF of unknown model parameters by using Gaussian mixture model and measurement responses. A Gaussian process regression model is first trained for approximating the logarithm of the product of likelihood function and prior PDF, with which, another Gaussian process model is induced for approximating the expensive evidence lower bound (ELBO). Then, two Bayesian numerical methods, i.e., Bayesian optimization and Bayesian quadrature, are combined sequentially as a novel Bayesian active learning method for searching the global optima of the parameters of the variational posterior density. The proposed method inherits the advantages of both Bayesian numerical methods, which includes good global convergence, much less number of simulator calls, etc. Three examples, including the dynamic model of a two degrees of freedom structures, the lubrication model of a hybrid journal bearing, and the dynamic model of an airplane structure, are introduced for demonstrating the relative merits of the proposed method. Results show that, given desired requirement of numerical accuracy, the proposed method is more efficient than the parallel methods.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"50 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142665644","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-12DOI: 10.1016/j.ymssp.2024.112077
Xiaodong Zhang, Yilong Wang, Jipeng Li, Shuai Chen, Bo Fang, Jinpeng Wang, Dengqing Cao
Rotating and Length-Varying Flexible Manipulators (RLVFMs) benefit from the ability to transform their length to adapt to complex and demanding workspaces but suffer from increased complexity in nonlinear dynamical characteristics and thus difficulties in modeling. To provide an in-depth understanding of the RLVFMs, this paper proposes a novel dynamical modeling approach for the RLVFMs, called the Parametric Global Modal Method (PGMM), and presents a framework to study their nonlinear responses. It is capable of addressing time-varying boundary conditions and describing the elastic deformation of all flexible components with only one set of modal coordinates. A low-dimensional dynamical model of a RLVFM is developed. The natural characteristic results obtained from the models developed by the PGMM and the finite element method (FEM) are compared for verifications of the PGMM. Via a convergence analysis of responses, the high precision of the model developed by the PGMM is verified to be achieved by using only the first two modes. On this basis, the dynamic responses and computational efficiency of the low-dimensional model are validated through experiments and finite element method (FEM) simulations. Moreover, the responses of the RLVFM under operations of rapid maneuvering are studied and a potential vibration control strategy for the RLVFM is preliminarily demonstrated. This work provides a new way of developing advanced dynamical modeling methods of reconfigurable and deformable multi-component mechanisms for their dynamical design, response analysis, and system control.
{"title":"Parametric global mode method for dynamical modeling and response analysis of a rotating and length-varying flexible manipulator","authors":"Xiaodong Zhang, Yilong Wang, Jipeng Li, Shuai Chen, Bo Fang, Jinpeng Wang, Dengqing Cao","doi":"10.1016/j.ymssp.2024.112077","DOIUrl":"https://doi.org/10.1016/j.ymssp.2024.112077","url":null,"abstract":"Rotating and Length-Varying Flexible Manipulators (RLVFMs) benefit from the ability to transform their length to adapt to complex and demanding workspaces but suffer from increased complexity in nonlinear dynamical characteristics and thus difficulties in modeling. To provide an in-depth understanding of the RLVFMs, this paper proposes a novel dynamical modeling approach for the RLVFMs, called the Parametric Global Modal Method (PGMM), and presents a framework to study their nonlinear responses. It is capable of addressing time-varying boundary conditions and describing the elastic deformation of all flexible components with only one set of modal coordinates. A low-dimensional dynamical model of a RLVFM is developed. The natural characteristic results obtained from the models developed by the PGMM and the finite element method (FEM) are compared for verifications of the PGMM. Via a convergence analysis of responses, the high precision of the model developed by the PGMM is verified to be achieved by using only the first two modes. On this basis, the dynamic responses and computational efficiency of the low-dimensional model are validated through experiments and finite element method (FEM) simulations. Moreover, the responses of the RLVFM under operations of rapid maneuvering are studied and a potential vibration control strategy for the RLVFM is preliminarily demonstrated. This work provides a new way of developing advanced dynamical modeling methods of reconfigurable and deformable multi-component mechanisms for their dynamical design, response analysis, and system control.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"13 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142665656","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-12DOI: 10.1016/j.ymssp.2024.112123
En-Guo Liu, Meng Li, Hu Ding
The broadband vibration reduction performance of nonlinear energy sink (NES) has attracted wide attention. However, the impact of the NES’s additional mass other than the oscillator and how it is connected to the primary structure has been ignored. More recently, it has been discovered that vibration attenuation through the cellular application of NES can achieve greater efficiency. However, the connection between NES cells and the primary structure, as well as between cells, has not been studied. In this study, by considering the additional mass of the NES cells, the influence of the connection modes of NES cells on the vibration reduction efficiency is investigated theoretically, optimally and experimentally for the first time. The forced vibration models of linear oscillator coupled with NES cells are established by viscoelastic connection and rigid connection respectively. The approximate analysis and numerical analysis show that the vibration reduction efficiency of NES cells is affected by the resonance frequency of the primary structure and the external excitation intensity and shows a nonlinear trend. With the change of the resonant frequency of the primary structure, the viscoelastic connection NES cells can almost always obtain higher vibration reduction efficiency than the rigid connection NES cells. The global bifurcation results show that the strongly modulated responses of the structure can be triggered by the viscoelastic connection. Moreover, the connection modes between NES cells also affect the vibration reduction efficiency. The optimal parameters of the connection damping and connection stiffness are obtained by the particle swarm optimization algorithm. Finally, the viscoelastic connection and rigid connection, and the effect of the connection mode between NES cells on the vibration reduction efficiency are compared by experiments. The conclusions of theoretical research are verified. This work can provide theoretical guidance for the engineering application of NES cells.
非线性能量汇（NES）的宽带减振性能已引起广泛关注。然而，NES 除振荡器外的附加质量及其与主结构的连接方式所产生的影响一直被忽视。最近，人们发现通过 NES 单元应用来减弱振动可以实现更高的效率。然而，NES 单元与主结构之间以及单元与单元之间的连接尚未得到研究。在本研究中，通过考虑 NES 单元的附加质量，首次从理论、优化和实验方面研究了 NES 单元的连接模式对减振效率的影响。通过粘弹性连接和刚性连接，分别建立了与 NES 单元耦合的线性振子的受迫振动模型。近似分析和数值分析表明，NES 电池的减振效率受主结构共振频率和外部激励强度的影响，并呈现非线性趋势。随着主结构共振频率的变化，粘弹性连接 NES 单元几乎总能获得比刚性连接 NES 单元更高的减振效率。全局分岔结果表明，粘弹性连接可以触发结构的强调制响应。此外，NES 单元之间的连接模式也会影响减振效率。通过粒子群优化算法获得了连接阻尼和连接刚度的最佳参数。最后，通过实验比较了粘弹性连接和刚性连接以及 NES 单元之间的连接模式对减振效率的影响。验证了理论研究的结论。这项工作可为 NES 单元的工程应用提供理论指导。
{"title":"Influence of additional mass and connection of nonlinear energy sinks on vibration reduction performance","authors":"En-Guo Liu, Meng Li, Hu Ding","doi":"10.1016/j.ymssp.2024.112123","DOIUrl":"https://doi.org/10.1016/j.ymssp.2024.112123","url":null,"abstract":"The broadband vibration reduction performance of nonlinear energy sink (NES) has attracted wide attention. However, the impact of the NES’s additional mass other than the oscillator and how it is connected to the primary structure has been ignored. More recently, it has been discovered that vibration attenuation through the cellular application of NES can achieve greater efficiency. However, the connection between NES cells and the primary structure, as well as between cells, has not been studied. In this study, by considering the additional mass of the NES cells, the influence of the connection modes of NES cells on the vibration reduction efficiency is investigated theoretically, optimally and experimentally for the first time. The forced vibration models of linear oscillator coupled with NES cells are established by viscoelastic connection and rigid connection respectively. The approximate analysis and numerical analysis show that the vibration reduction efficiency of NES cells is affected by the resonance frequency of the primary structure and the external excitation intensity and shows a nonlinear trend. With the change of the resonant frequency of the primary structure, the viscoelastic connection NES cells can almost always obtain higher vibration reduction efficiency than the rigid connection NES cells. The global bifurcation results show that the strongly modulated responses of the structure can be triggered by the viscoelastic connection. Moreover, the connection modes between NES cells also affect the vibration reduction efficiency. The optimal parameters of the connection damping and connection stiffness are obtained by the particle swarm optimization algorithm. Finally, the viscoelastic connection and rigid connection, and the effect of the connection mode between NES cells on the vibration reduction efficiency are compared by experiments. The conclusions of theoretical research are verified. This work can provide theoretical guidance for the engineering application of NES cells.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"197 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142665641","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-12DOI: 10.1016/j.ymssp.2024.112074
Yuguang Fu, Zixin Wang, Amin Maghareh, Shirley Dyke, Mohammad Jahanshahi, Adnan Shahriar, Fan Zhang
Due to their unpredictable nature, many impact events (e.g., overheight vehicles striking on bridges) go unnoticed or get reported many hours later. However, they can induce structural failures or hidden damage that accelerates the structure’s long-term degradation. Therefore, prompt impact detection and localization strategies are essential for early warning of impact events and rapid maintenance of structures. Most existing impact detection strategies are developed for aircraft composite panels utilizing high-rate synchronized measurement from densely deployed sensors. Limited efforts have been made for infrastructure or human habitats which generally require large-scale but low-rate measurement. In particular, due to harsh environments (e.g., deep space habitats under meteoroids), structural impact localization must be robust to limited sensors (e.g., sensor damage during impacts) and multi-source errors (e.g., measurement errors). In this study, an effective impact detection and localization strategy is proposed using a limited number of vibration measurements, especially in harsh environments (e.g. in deep space). Convolutional neural networks are trained for each sensor node and are fused using Bayesian theory to improve the accuracy of impact localization. Special considerations are paid to evaluate the effect of both measurement error and modeling error in the analysis. The proposed strategy is illustrated using 1D structure, and further validated in 3D geodesic dome structure numerically. The results demonstrate that it can detect and localize impact events accurately and robustly on structures.
{"title":"Effective structural impact detection and localization using convolutional neural network and Bayesian information fusion with limited sensors","authors":"Yuguang Fu, Zixin Wang, Amin Maghareh, Shirley Dyke, Mohammad Jahanshahi, Adnan Shahriar, Fan Zhang","doi":"10.1016/j.ymssp.2024.112074","DOIUrl":"https://doi.org/10.1016/j.ymssp.2024.112074","url":null,"abstract":"Due to their unpredictable nature, many impact events (e.g., overheight vehicles striking on bridges) go unnoticed or get reported many hours later. However, they can induce structural failures or hidden damage that accelerates the structure’s long-term degradation. Therefore, prompt impact detection and localization strategies are essential for early warning of impact events and rapid maintenance of structures. Most existing impact detection strategies are developed for aircraft composite panels utilizing high-rate synchronized measurement from densely deployed sensors. Limited efforts have been made for infrastructure or human habitats which generally require large-scale but low-rate measurement. In particular, due to harsh environments (e.g., deep space habitats under meteoroids), structural impact localization must be robust to limited sensors (e.g., sensor damage during impacts) and multi-source errors (e.g., measurement errors). In this study, an effective impact detection and localization strategy is proposed using a limited number of vibration measurements, especially in harsh environments (e.g. in deep space). Convolutional neural networks are trained for each sensor node and are fused using Bayesian theory to improve the accuracy of impact localization. Special considerations are paid to evaluate the effect of both measurement error and modeling error in the analysis. The proposed strategy is illustrated using 1D structure, and further validated in 3D geodesic dome structure numerically. The results demonstrate that it can detect and localize impact events accurately and robustly on structures.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"80 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142665647","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-12DOI: 10.1016/j.ymssp.2024.112125
Yang Liu, Zhanpeng Jiang, Ning Zhang, Jun Tang, Zijian Liu, Yingbing Sun, Fenghe Wu
As the actuator of the turbine speed control system, the performance and response characteristics of the speed control valve actuator directly affect the operational economy, maneuverability, and reliability of the turbine unit. When faults occur in scenarios where data scarcity is coupled with high noise levels, existing deep neural network models are limited by their inability to extract key discriminative features from noisy signals and by the lack of sufficient training information. This limitation hinders the development and application of highly reliable fault diagnosis systems. We propose a novel fault diagnosis framework, Resformer, which is designed to address the challenges posed by data scarcity and high noise coupling, as well as the highly coupled and complex fault modes in electro-hydraulic systems. The Resformer framework offers a highly interpretable feature selection and fusion strategy to identify key features. It also integrates the Local Binary Pattern algorithm to extract local features from grayscale images of multi-sensor data, significantly enhancing the representativeness and noise resistance of the dataset. Moreover, to strengthen the Resformer’s multi-scale feature extraction capability and noise robustness, a multi-kernel dilated convolutional residual network architecture is introduced, enabling the discovery of critical discriminative features under conditions of data scarcity and high noise coupling. The proposed efficient multi-scale self-attention mechanism effectively extracts important features at different scales, further improving the performance of Resformer. Experiments conducted on the GVA testbed have validated the effectiveness and robustness of Resformer.
{"title":"Resformer: An end-to-end framework for fault diagnosis of governor valve actuator in the coupled scenario of data scarcity and high noise","authors":"Yang Liu, Zhanpeng Jiang, Ning Zhang, Jun Tang, Zijian Liu, Yingbing Sun, Fenghe Wu","doi":"10.1016/j.ymssp.2024.112125","DOIUrl":"https://doi.org/10.1016/j.ymssp.2024.112125","url":null,"abstract":"As the actuator of the turbine speed control system, the performance and response characteristics of the speed control valve actuator directly affect the operational economy, maneuverability, and reliability of the turbine unit. When faults occur in scenarios where data scarcity is coupled with high noise levels, existing deep neural network models are limited by their inability to extract key discriminative features from noisy signals and by the lack of sufficient training information. This limitation hinders the development and application of highly reliable fault diagnosis systems. We propose a novel fault diagnosis framework, Resformer, which is designed to address the challenges posed by data scarcity and high noise coupling, as well as the highly coupled and complex fault modes in electro-hydraulic systems. The Resformer framework offers a highly interpretable feature selection and fusion strategy to identify key features. It also integrates the Local Binary Pattern algorithm to extract local features from grayscale images of multi-sensor data, significantly enhancing the representativeness and noise resistance of the dataset. Moreover, to strengthen the Resformer’s multi-scale feature extraction capability and noise robustness, a multi-kernel dilated convolutional residual network architecture is introduced, enabling the discovery of critical discriminative features under conditions of data scarcity and high noise coupling. The proposed efficient multi-scale self-attention mechanism effectively extracts important features at different scales, further improving the performance of Resformer. Experiments conducted on the GVA testbed have validated the effectiveness and robustness of Resformer.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"13 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142665626","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-11-12DOI: 10.1016/j.ymssp.2024.112128
Wen Hou, Jiachang Wang, Leilei Wang, Song Zhang
Reliable tool wear prediction is of great importance for the improvement of machining quality and efficiency. With the advent of the big data era, data-driven tool wear prediction methods have proven to be highly effective. However, these methods have also revealed issues such as shallow feature extraction and limited generalization of models across different machining processes. The objective of this research is to propose a tool wear prediction method based on multimodal information fusion and deep subdomain adaptation to solve the existing problems. First, the original one-dimensional time-series tool monitoring signals are encoded into images to generate a two-dimensional image dataset. Secondly, a two-channel prediction model combining Residual Network and Gated Recurrent Unit is constructed to extract features from the two-dimensional image signals and the one-dimensional time-series signals respectively, and the extracted spatial and temporal features are fused. Thirdly, the dataset is divided into subdomains based on wear values, and the generalization ability of the model is improved by reducing the feature differences between source and target domains through the subdomain adaptive method, thus achieving the prediction of the tool wear values under different situations. Finally, through the validation on two milling wear datasets and comparison with the prediction results of other models, the experimental results prove the accuracy and good generalization of the method, which can provide a reference to improve the machining quality and efficiency, and is suitable for practical industrial application scenarios.
{"title":"Novel tool wear prediction method based on multimodal information fusion and deep subdomain adaptation","authors":"Wen Hou, Jiachang Wang, Leilei Wang, Song Zhang","doi":"10.1016/j.ymssp.2024.112128","DOIUrl":"https://doi.org/10.1016/j.ymssp.2024.112128","url":null,"abstract":"Reliable tool wear prediction is of great importance for the improvement of machining quality and efficiency. With the advent of the big data era, data-driven tool wear prediction methods have proven to be highly effective. However, these methods have also revealed issues such as shallow feature extraction and limited generalization of models across different machining processes. The objective of this research is to propose a tool wear prediction method based on multimodal information fusion and deep subdomain adaptation to solve the existing problems. First, the original one-dimensional time-series tool monitoring signals are encoded into images to generate a two-dimensional image dataset. Secondly, a two-channel prediction model combining Residual Network and Gated Recurrent Unit is constructed to extract features from the two-dimensional image signals and the one-dimensional time-series signals respectively, and the extracted spatial and temporal features are fused. Thirdly, the dataset is divided into subdomains based on wear values, and the generalization ability of the model is improved by reducing the feature differences between source and target domains through the subdomain adaptive method, thus achieving the prediction of the tool wear values under different situations. Finally, through the validation on two milling wear datasets and comparison with the prediction results of other models, the experimental results prove the accuracy and good generalization of the method, which can provide a reference to improve the machining quality and efficiency, and is suitable for practical industrial application scenarios.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"99 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142665627","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：