Pub Date : 2026-02-11DOI: 10.1016/j.ymssp.2026.113998
Zhuhong Wang, Hang Zhou, Hanlong Liu
Accurate measurement of three-dimensional deformation behavior is critical for understanding material mechanical properties. However, traditional Digital Volume Correlation (DVC) methods are limited by discrete sub-volume discretization, lack of physical constraints, and low computational efficiency. Data-driven approaches cannot guarantee physical plausibility and depend on large quantities of densely sampled data. This study proposes a novel physics-informed deep learning method for DVC (PiNetDVC). The method takes spatial coordinates as inputs and simultaneously predicts displacement and strain fields through continuous function representation, overcoming spatial resolution limitations and data dependency. The strain field is directly incorporated as a network output, with strain–displacement compatibility enforced by comparing network-predicted strains with strains derived from displacement gradients. A unified loss function integrates image consistency constraints with physics-informed regularization. Validation on six scenarios demonstrates superior performance over traditional ALDVC, achieving accuracy improvements of 81%, 83%, and over 95% for rigid body translation, uniaxial tension, and shear band deformation, respectively. For complex deformation patterns such as sinusoidal and non-uniform star-shaped modes, errors are maintained at the order of 10-3. Stable accuracy is maintained under 20 dB noise, with robust performance across different architectures and loss configurations. PiNetDVC provides an effective solution for 3D deformation measurement in aerospace, mechanical, and civil engineering applications.
{"title":"Physics-informed neural networks based digital volume correlation for displacement and strain measurements","authors":"Zhuhong Wang, Hang Zhou, Hanlong Liu","doi":"10.1016/j.ymssp.2026.113998","DOIUrl":"10.1016/j.ymssp.2026.113998","url":null,"abstract":"<div><div>Accurate measurement of three-dimensional deformation behavior is critical for understanding material mechanical properties. However, traditional Digital Volume Correlation (DVC) methods are limited by discrete sub-volume discretization, lack of physical constraints, and low computational efficiency. Data-driven approaches cannot guarantee physical plausibility and depend on large quantities of densely sampled data. This study proposes a novel physics-informed deep learning method for DVC (PiNetDVC). The method takes spatial coordinates as inputs and simultaneously predicts displacement and strain fields through continuous function representation, overcoming spatial resolution limitations and data dependency. The strain field is directly incorporated as a network output, with strain–displacement compatibility enforced by comparing network-predicted strains with strains derived from displacement gradients. A unified loss function integrates image consistency constraints with physics-informed regularization. Validation on six scenarios demonstrates superior performance over traditional ALDVC, achieving accuracy improvements of 81%, 83%, and over 95% for rigid body translation, uniaxial tension, and shear band deformation, respectively. For complex deformation patterns such as sinusoidal and non-uniform star-shaped modes, errors are maintained at the order of 10<sup>-3</sup>. Stable accuracy is maintained under 20 dB noise, with robust performance across different architectures and loss configurations. PiNetDVC provides an effective solution for 3D deformation measurement in aerospace, mechanical, and civil engineering applications.</div></div>","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"248 ","pages":"Article 113998"},"PeriodicalIF":8.9,"publicationDate":"2026-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147128","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 : 2026-02-11DOI: 10.1016/j.ymssp.2026.113979
Bohao Xu, Ling Yu, Zhenhua Nie
As one of the challenging topics in structural health monitoring, the identification of multiple moving vehicle loads remains largely unexplored owing to the large differences in load magnitudes. Even though a recent study introduced multiple regularization parameters (MRP) within a two-stage framework to distinguish the properties of different loads, its performance is highly sensitive to the initial estimates and deteriorates as the number of loads increases. To address this, the original two-stage work is extended into an alternating iterative framework (AIF), which iteratively updates the static load, dynamic load, and the variance of the dynamic loads. This extension follows the conclusion in the previous study that the regularization parameters chosen within the reasonable range of residual noise are close. Furthermore, Anderson acceleration is introduced only to the static load and the variance of dynamic load to enhance effectiveness. A safeguard strategy is incorporated to ensure the local convergence of the AIF. Finally, the proposed method is validated in both numerical simulations and laboratory experiments. The comparative cases under different response combinations, different numbers of loads and different initial estimates in the numerical simulations show that the proposed method achieves a higher accuracy, especially in comparison with the previous study. The SNR threshold required for maintaining reliable identification decreases from 25 dB to 20 dB, even when the noise variance is inaccurately estimated. Moreover, the weight of the model vehicle can be reasonably estimated by the proposed method in the validation of experiment.
{"title":"Accelerated alternating iterative identification for multiple moving vehicle loads based on Anderson acceleration with safeguard strategy","authors":"Bohao Xu, Ling Yu, Zhenhua Nie","doi":"10.1016/j.ymssp.2026.113979","DOIUrl":"10.1016/j.ymssp.2026.113979","url":null,"abstract":"<div><div>As one of the challenging topics in structural health monitoring, the identification of multiple moving vehicle loads remains largely unexplored owing to the large differences in load magnitudes. Even though a recent study introduced multiple regularization parameters (MRP) within a two-stage framework to distinguish the properties of different loads, its performance is highly sensitive to the initial estimates and deteriorates as the number of loads increases. To address this, the original two-stage work is extended into an alternating iterative framework (AIF), which iteratively updates the static load, dynamic load, and the variance of the dynamic loads. This extension follows the conclusion in the previous study that the regularization parameters chosen within the reasonable range of residual noise are close. Furthermore, Anderson acceleration is introduced only to the static load and the variance of dynamic load to enhance effectiveness. A safeguard strategy is incorporated to ensure the local convergence of the AIF. Finally, the proposed method is validated in both numerical simulations and laboratory experiments. The comparative cases under different response combinations, different numbers of loads and different initial estimates in the numerical simulations show that the proposed method achieves a higher accuracy, especially in comparison with the previous study. The SNR threshold required for maintaining reliable identification decreases from 25 dB to 20 dB, even when the noise variance is inaccurately estimated. Moreover, the weight of the model vehicle can be reasonably estimated by the proposed method in the validation of experiment.</div></div>","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"248 ","pages":"Article 113979"},"PeriodicalIF":8.9,"publicationDate":"2026-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147127","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 : 2026-02-11DOI: 10.1016/j.ymssp.2026.113982
Chengyi Wu , Shijun Ji , Ji Zhao , Enzhong Zhang , Guang Yang
In the implementation of digital twin for ultra-precision machining (UPM) based on deep learning, conventional approaches suffer from limited interpretability of model and insufficient visualization capabilities. Moreover, their performance is significantly compromised by the coupling effects of multisource errors, making it difficult to achieve accurate position prediction and effective compensation. To address these limitations, this paper proposes a novel digital twin system which is driven by a hybrid model that integrates the Patch Time Series Transformer and multisource error coupling mechanism, and enables the visualization of the error compensation strategy. It achieves intelligent contour error compensation during machining by dynamically correcting the position commands along the trajectory. Based on an analysis of the theoretical error band arising from the multisource error coupling mechanism, the position prediction accuracy of each axis is improved through the self-supervised learning and hyperparameter fine-tuning methods. Furthermore, temporal stability is validated via time-effect analysis. Comprehensive case studies are conducted on a custom-built multi-axis ultra-precision machine tool, covering both single-axis and multi-axis motions under varying loads, feedrates, and ambient temperatures. The test results demonstrate that the proposed method improves single-axis positioning accuracy by 47.07% and multi-axis trajectory contour accuracy by 26.99%. In the micro-groove machining experiment, the compensated linear positioning error is reduced to 0.0393 μm, and the angular positioning error is 0.0013°, with the resultant cutting force indirectly reduced by up to 9.20%. The robustness and adaptability of the proposed method are validated under complex operating conditions, thereby enabling high-accuracy contour control in practical UPM applications.
{"title":"A data-mechanism-based digital twin system for intelligent contour error compensation of ultra-precision machining","authors":"Chengyi Wu , Shijun Ji , Ji Zhao , Enzhong Zhang , Guang Yang","doi":"10.1016/j.ymssp.2026.113982","DOIUrl":"10.1016/j.ymssp.2026.113982","url":null,"abstract":"<div><div>In the implementation of digital twin for ultra-precision machining (UPM) based on deep learning, conventional approaches suffer from limited interpretability of model and insufficient visualization capabilities. Moreover, their performance is significantly compromised by the coupling effects of multisource errors, making it difficult to achieve accurate position prediction and effective compensation. To address these limitations, this paper proposes a novel digital twin system which is driven by a hybrid model that integrates the Patch Time Series Transformer and multisource error coupling mechanism, and enables the visualization of the error compensation strategy. It achieves intelligent contour error compensation during machining by dynamically correcting the position commands along the trajectory. Based on an analysis of the theoretical error band arising from the multisource error coupling mechanism, the position prediction accuracy of each axis is improved through the self-supervised learning and hyperparameter fine-tuning methods. Furthermore, temporal stability is validated via time-effect analysis. Comprehensive case studies are conducted on a custom-built multi-axis ultra-precision machine tool, covering both single-axis and multi-axis motions under varying loads, feedrates, and ambient temperatures. The test results demonstrate that the proposed method improves single-axis positioning accuracy by 47.07% and multi-axis trajectory contour accuracy by 26.99%. In the micro-groove machining experiment, the compensated linear positioning error is reduced to 0.0393 μm, and the angular positioning error is 0.0013°, with the resultant cutting force indirectly reduced by up to 9.20%. The robustness and adaptability of the proposed method are validated under complex operating conditions, thereby enabling high-accuracy contour control in practical UPM applications.</div></div>","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"248 ","pages":"Article 113982"},"PeriodicalIF":8.9,"publicationDate":"2026-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147126","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}
Vibration isolation systems for ultra-precision instruments are strongly influenced by internal resonances, leading to an increase in vibration transmissibility of up to 10–30 dB at the resonance frequencies. The dual-chamber air-floating vibration isolation system exhibits an extremely low natural frequency. However, the presence of the expansion chamber introduces internal resonance problems at mid-to-high frequencies. To enhance the vibration isolation performance of the dual-chamber air-floated isolation system, this paper proposes an adaptive control strategy tailored to such systems to address internal resonance beyond the natural frequency. The dual-chamber air-floated isolation system is accurately modeled and systematically analyzed in this paper. The results reveal that the fundamental cause of internal resonance in the dual-chamber isolation system is Helmholtz resonance. To address this issue, a novel orthogonal basis function infinite impulse response (OBF-IIR) controller is designed in this paper to efficiently compensate for vibrations induced by the dual-chamber Helmholtz resonance effect. On this basis, a fast, accurate online adaptive algorithm is developed to update the controller zeros in real time, enabling adaptive, synchronous compensation of internal resonances in the dual-chamber isolation system. The proposed OBF-IIR controller not only suppresses internal resonances induced by the spring–damper model and the dual-chamber Helmholtz resonance effect, but also compensates for resonances arising from other sources. The proposed adaptive control strategy demonstrates faster convergence and higher accuracy, reducing the vibration transmissibility of the isolation system by 10–30 dB in the 2–100 Hz range and decreasing the cumulative power spectral density at 100 Hz by 23.8%–84.9%.
{"title":"Ultra-low frequency air flotation vibration isolation system with a dual-chamber structure using adaptive control strategy","authors":"Tianyi Li, Shilong Guo, Zhendong Lan, Bo Zhao, Jiubin Tan, Chenglong Yu","doi":"10.1016/j.ymssp.2026.113987","DOIUrl":"10.1016/j.ymssp.2026.113987","url":null,"abstract":"<div><div>Vibration isolation systems for ultra-precision instruments are strongly influenced by internal resonances, leading to an increase in vibration transmissibility of up to 10–30 dB at the resonance frequencies. The dual-chamber air-floating vibration isolation system exhibits an extremely low natural frequency. However, the presence of the expansion chamber introduces internal resonance problems at mid-to-high frequencies. To enhance the vibration isolation performance of the dual-chamber air-floated isolation system, this paper proposes an adaptive control strategy tailored to such systems to address internal resonance beyond the natural frequency. The dual-chamber air-floated isolation system is accurately modeled and systematically analyzed in this paper. The results reveal that the fundamental cause of internal resonance in the dual-chamber isolation system is Helmholtz resonance. To address this issue, a novel orthogonal basis function infinite impulse response (OBF-IIR) controller is designed in this paper to efficiently compensate for vibrations induced by the dual-chamber Helmholtz resonance effect. On this basis, a fast, accurate online adaptive algorithm is developed to update the controller zeros in real time, enabling adaptive, synchronous compensation of internal resonances in the dual-chamber isolation system. The proposed OBF-IIR controller not only suppresses internal resonances induced by the spring–damper model and the dual-chamber Helmholtz resonance effect, but also compensates for resonances arising from other sources. The proposed adaptive control strategy demonstrates faster convergence and higher accuracy, reducing the vibration transmissibility of the isolation system by 10–30 dB in the 2–100 Hz range and decreasing the cumulative power spectral density at 100 Hz by 23.8%–84.9%.</div></div>","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"248 ","pages":"Article 113987"},"PeriodicalIF":8.9,"publicationDate":"2026-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146147125","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 : 2026-02-10DOI: 10.1016/j.ymssp.2026.113984
Xiaowei Zhang, Xiaopeng Wang, Yingrui Ye
Symmetry is commonly used in engineering design for its simplicity and structural stability. Conventional locally resonant metastructures with strict spatial symmetry exhibit only one active mode, limiting modal diversity and dynamic performance. To overcome this constraint, we introduce spatial stiffness asymmetry, enabling three-dimensional dynamic responses. Such asymmetric design induces coupling between translational and rotational degrees of freedom, allowing multiple resonant modes to be excited by a single-directional input. Leveraging this mechanism, we design a metastructure that achieves vertical vibration isolation through three distinct coupled modes generated by a single resonator. A theoretical model is developed to describe the asymmetric self-coupling behavior, and vibration-table experiments confirm the predicted multi-band isolation performance. This work provides a new strategy for enhancing modal utilization in resonant systems and offers practical guidance for compact, multi-band vibration control.
{"title":"Asymmetric design enables self-coupled locally resonant metastructure for multi-modal vibration isolation","authors":"Xiaowei Zhang, Xiaopeng Wang, Yingrui Ye","doi":"10.1016/j.ymssp.2026.113984","DOIUrl":"10.1016/j.ymssp.2026.113984","url":null,"abstract":"<div><div>Symmetry is commonly used in engineering design for its simplicity and structural stability. Conventional locally resonant metastructures with strict spatial symmetry exhibit only one active mode, limiting modal diversity and dynamic performance. To overcome this constraint, we introduce spatial stiffness asymmetry, enabling three-dimensional dynamic responses. Such asymmetric design induces coupling between translational and rotational degrees of freedom, allowing multiple resonant modes to be excited by a single-directional input. Leveraging this mechanism, we design a metastructure that achieves vertical vibration isolation through three distinct coupled modes generated by a single resonator. A theoretical model is developed to describe the asymmetric self-coupling behavior, and vibration-table experiments confirm the predicted multi-band isolation performance. This work provides a new strategy for enhancing modal utilization in resonant systems and offers practical guidance for compact, multi-band vibration control.</div></div>","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"248 ","pages":"Article 113984"},"PeriodicalIF":8.9,"publicationDate":"2026-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146146825","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 : 2026-02-09DOI: 10.1016/j.ymssp.2026.113964
Weijia Liu, Changhai Zhai, Weiping Wen, Kun Liu
Traditional structural health monitoring relies on sensor deployment but is constrained by high installation costs, insufficient monitoring network coverage, and noise interference in complex seismic scenarios, limiting its application. Leveraging existing surveillance cameras in buildings for non-contact monitoring emerges as a promising solution. This study proposes a finite element model updating method integrating computer vision with time-domain signal autocorrelation sensitivity. This method deeply integrates visual displacement data from surveillance videos with structural mechanics models, employing the autocorrelation function of time-domain signals for effective noise reduction. It enhances the identification of local stiffness changes, thereby significantly improving the accuracy and robustness of model updating. This study first conducts model updating through numerical simulation methods. The displacement autocorrelation sensitivity method is employed, systematically accounting for measured response noise, seismic motion noise, and uncertainties in seismic motion (including spectral characteristics, duration, and peak ground acceleration). Numerical simulation results demonstrate that, under structural response and seismic motion noise conditions with a signal-to-noise ratio (SNR) as low as 20 dB, the displacement autocorrelation sensitivity method achieves a parameter updating error within 5%, validating its high adaptability and robustness in complex disturbance environments. For far-field non-impulsive seismic motions, the displacement autocorrelation sensitivity method exhibits higher precision and stability compared to traditional displacement sensitivity methods. For engineering feasibility assessment, shaking table tests were conducted on a three-story steel frame, integrating displacement time histories from indoor/outdoor camera videos with ground motion data from IMU sensors for model updating. Test results show Pearson correlation coefficients of 0.91, 0.94, and 0.97 for displacement time history predictions versus measured values from the top to the first story, with peak displacement relative errors below 6% for all stories. This method can efficiently utilize existing building surveillance videos to complete model updates within minutes in post-earthquake environment, providing reliable support for damage assessment and emergency response.
{"title":"Model updating method based on computer vision and autocorrelation sensitivity: Deep integration of visual information and physical mechanisms","authors":"Weijia Liu, Changhai Zhai, Weiping Wen, Kun Liu","doi":"10.1016/j.ymssp.2026.113964","DOIUrl":"https://doi.org/10.1016/j.ymssp.2026.113964","url":null,"abstract":"Traditional structural health monitoring relies on sensor deployment but is constrained by high installation costs, insufficient monitoring network coverage, and noise interference in complex seismic scenarios, limiting its application. Leveraging existing surveillance cameras in buildings for non-contact monitoring emerges as a promising solution. This study proposes a finite element model updating method integrating computer vision with time-domain signal autocorrelation sensitivity. This method deeply integrates visual displacement data from surveillance videos with structural mechanics models, employing the autocorrelation function of time-domain signals for effective noise reduction. It enhances the identification of local stiffness changes, thereby significantly improving the accuracy and robustness of model updating. This study first conducts model updating through numerical simulation methods. The displacement autocorrelation sensitivity method is employed, systematically accounting for measured response noise, seismic motion noise, and uncertainties in seismic motion (including spectral characteristics, duration, and peak ground acceleration). Numerical simulation results demonstrate that, under structural response and seismic motion noise conditions with a signal-to-noise ratio (SNR) as low as 20 dB, the displacement autocorrelation sensitivity method achieves a parameter updating error within 5%, validating its high adaptability and robustness in complex disturbance environments. For far-field non-impulsive seismic motions, the displacement autocorrelation sensitivity method exhibits higher precision and stability compared to traditional displacement sensitivity methods. For engineering feasibility assessment, shaking table tests were conducted on a three-story steel frame, integrating displacement time histories from indoor/outdoor camera videos with ground motion data from IMU sensors for model updating. Test results show Pearson correlation coefficients of 0.91, 0.94, and 0.97 for displacement time history predictions versus measured values from the top to the first story, with peak displacement relative errors below 6% for all stories. This method can efficiently utilize existing building surveillance videos to complete model updates within minutes in post-earthquake environment, providing reliable support for damage assessment and emergency response.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"45 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2026-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146146780","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 : 2026-02-09DOI: 10.1016/j.ymssp.2026.113977
Lubing Wang, Ying Chen, Zhengbo Zhu, Xufeng Zhao
In prognostics and health management for mechanical systems, the uncertainty of remaining useful life (RUL) assessment caused by noise interference and measurement errors is often overlooked, which may lead to inaccurate maintenance results. To solve these challenges, this study presents a predictive maintenance framework that integrates uncertainty-aware RUL estimation to support maintenance decisions and spare parts management. We first introduce a hybrid model that combines bidirectional gated recurrent units with an integrated global and local multi-head sparse attention mechanism to capture long-term dependencies and transient patterns, while employing Monte Carlo dropout for quantifying RUL uncertainty. Using RUL uncertainty estimation, three distinct predictive maintenance models and spare parts ordering models are formulated. These models integrate estimated mean RUL, lower bounds, and maintenance costs to dynamically determine the optimal maintenance time and spare parts ordering time during periodic inspections. Validated on aero-engine and industrial machine datasets, the method outperforms existing strategies, achieving effective fault prevention and reducing the maintenance cost rate by over 50%. This work provides a practical solution for reliable and cost-effective mechanical systems by linking uncertainty-aware RUL estimation with maintenance decisions.
{"title":"Dynamic predictive maintenance framework for mechanical systems via uncertainty-aware RUL estimation","authors":"Lubing Wang, Ying Chen, Zhengbo Zhu, Xufeng Zhao","doi":"10.1016/j.ymssp.2026.113977","DOIUrl":"https://doi.org/10.1016/j.ymssp.2026.113977","url":null,"abstract":"In prognostics and health management for mechanical systems, the uncertainty of remaining useful life (RUL) assessment caused by noise interference and measurement errors is often overlooked, which may lead to inaccurate maintenance results. To solve these challenges, this study presents a predictive maintenance framework that integrates uncertainty-aware RUL estimation to support maintenance decisions and spare parts management. We first introduce a hybrid model that combines bidirectional gated recurrent units with an integrated global and local multi-head sparse attention mechanism to capture long-term dependencies and transient patterns, while employing Monte Carlo dropout for quantifying RUL uncertainty. Using RUL uncertainty estimation, three distinct predictive maintenance models and spare parts ordering models are formulated. These models integrate estimated mean RUL, lower bounds, and maintenance costs to dynamically determine the optimal maintenance time and spare parts ordering time during periodic inspections. Validated on aero-engine and industrial machine datasets, the method outperforms existing strategies, achieving effective fault prevention and reducing the maintenance cost rate by over 50%. This work provides a practical solution for reliable and cost-effective mechanical systems by linking uncertainty-aware RUL estimation with maintenance decisions.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"6 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2026-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146146779","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 : 2026-02-09DOI: 10.1016/j.ymssp.2026.113928
Ye Lu, Xiaomei Li, Zhijiang Xie, Haolan Jia, Zhenjun Su, Xiaoliang Hu
To investigate the coupling effects of clearance and flexible components on dynamic performance of pushrod-driven actuator, this study proposes a hybrid contact force model suitable for large loads with an adaptive restitution coefficient, and a modified transitional lubrication force model. Considering the influence of flexible components, a rigid-flexible coupling dynamics model of the actuator incorporating lubrication clearance is established. Subsequently, effects of clearance size, driving speed, dynamic viscosity and load on system’s dynamics and chaos are then analyzed. Finally, experimental validation confirms the model’s effectiveness. The results show that the choice of clearance size and drive speed significantly influences system stability, and that high dynamic viscosity lubricants can lower the output vibration frequency and amplitude. Under large loads, the lubricant film thickness at clearance approaches zero, intensifying clearance collisions and wear. This increases the output vibration frequency, and substantially reduces the lubricant’s mitigating effects on clearance and flexible factors. This study provides theoretical support for the design of high-performance rudder actuators.
{"title":"Rigid-flexible coupling modeling and nonlinear analysis of rudder actuator with lubrication clearance","authors":"Ye Lu, Xiaomei Li, Zhijiang Xie, Haolan Jia, Zhenjun Su, Xiaoliang Hu","doi":"10.1016/j.ymssp.2026.113928","DOIUrl":"https://doi.org/10.1016/j.ymssp.2026.113928","url":null,"abstract":"To investigate the coupling effects of clearance and flexible components on dynamic performance of pushrod-driven actuator, this study proposes a hybrid contact force model suitable for large loads with an adaptive restitution coefficient, and a modified transitional lubrication force model. Considering the influence of flexible components, a rigid-flexible coupling dynamics model of the actuator incorporating lubrication clearance is established. Subsequently, effects of clearance size, driving speed, dynamic viscosity and load on system’s dynamics and chaos are then analyzed. Finally, experimental validation confirms the model’s effectiveness. The results show that the choice of clearance size and drive speed significantly influences system stability, and that high dynamic viscosity lubricants can lower the output vibration frequency and amplitude. Under large loads, the lubricant film thickness at clearance approaches zero, intensifying clearance collisions and wear. This increases the output vibration frequency, and substantially reduces the lubricant’s mitigating effects on clearance and flexible factors. This study provides theoretical support for the design of high-performance rudder actuators.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"46 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2026-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146146713","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 : 2026-02-09DOI: 10.1016/j.ymssp.2026.113978
Yuegang Luo, Ning Liu, Songsong Xiao, Wanlei Wang
Rolling bearings are assembled on the shaft through interference fit. The presence of shaft cracks directly affects bearing operation and may even induce defects. Conversely, bearing defects may also exacerbate shaft damage. Currently, research on the bearing defect-shaft crack coupled faults remains insufficient and requires further exploration. This paper proposes an inner raceway extension defect model that incorporates the motion trajectory of the rolling elements. A dynamic model of a rotor-bearing-pedestal system with bearing extension defect and shaft crack is established. The dynamic characteristics of defects, cracks, and coupled faults are systematically analyzed, and the coupling mechanism is further investigated. The simulation and experimental results indicate that for inner raceway defect-shaft crack coupled fault, an increase in crack depth amplifies the bearing fault characteristics, especially when the crack is located near the bearing support or at the midspan of the shaft. The extension of the defect also exacerbates the damage caused by the crack to the shaft. For outer raceway defect-crack coupled fault, shallow cracks suppress the bearing fault frequency. However, once the crack depth exceeds a certain threshold, this suppression transitions to amplification. Cracks located at the midspan of the shaft enhance the bearing fault characteristics. The extension of the outer raceway defect primarily affects the bearing fault frequency and the overall vibration amplitude. The findings of this study are expected to provide a valuable theoretical basis for diagnosing and predicting bearing defect-shaft crack coupled faults.
{"title":"Dynamic behaviors of a rolling bearing-rotor system with bearing extended defect and shaft crack: simulation and experimental investigation","authors":"Yuegang Luo, Ning Liu, Songsong Xiao, Wanlei Wang","doi":"10.1016/j.ymssp.2026.113978","DOIUrl":"https://doi.org/10.1016/j.ymssp.2026.113978","url":null,"abstract":"Rolling bearings are assembled on the shaft through interference fit. The presence of shaft cracks directly affects bearing operation and may even induce defects. Conversely, bearing defects may also exacerbate shaft damage. Currently, research on the bearing defect-shaft crack coupled faults remains insufficient and requires further exploration. This paper proposes an inner raceway extension defect model that incorporates the motion trajectory of the rolling elements. A dynamic model of a rotor-bearing-pedestal system with bearing extension defect and shaft crack is established. The dynamic characteristics of defects, cracks, and coupled faults are systematically analyzed, and the coupling mechanism is further investigated. The simulation and experimental results indicate that for inner raceway defect-shaft crack coupled fault, an increase in crack depth amplifies the bearing fault characteristics, especially when the crack is located near the bearing support or at the midspan of the shaft. The extension of the defect also exacerbates the damage caused by the crack to the shaft. For outer raceway defect-crack coupled fault, shallow cracks suppress the bearing fault frequency. However, once the crack depth exceeds a certain threshold, this suppression transitions to amplification. Cracks located at the midspan of the shaft enhance the bearing fault characteristics. The extension of the outer raceway defect primarily affects the bearing fault frequency and the overall vibration amplitude. The findings of this study are expected to provide a valuable theoretical basis for diagnosing and predicting bearing defect-shaft crack coupled faults.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"1 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2026-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146146778","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 : 2026-02-09DOI: 10.1016/j.ymssp.2026.113967
Hanqiao Sun, Jingfeng Lu, Jiawen Xu, Ruqiang Yan
Impedance signals for structural health monitoring are often sparse and difficult to acquire in damaged conditions. Increasing the damage categories would significantly reduce accuracy. In this study, we propose a Conv-Transformer model that is capable of multi-task structural health monitoring, addressing the complexities of small sample datasets while handling multiple fault detection tasks, including mass loss and bolt loosening. The model enhances feature extraction by combining convolutional layers and multi-head attention within the Transformer encoder, focusing on the relative location of the peaks and the local feature of each peak in the impedance signals. These advantages enable highly accurate multi-task SHM with small samples of impedance signals. The proposed model is first trained on a large amount of data in mixed conditions and then fine-tuned with small sample data for an eight-class fault classification task. Experimental results show that the model demonstrates strong learning ability and cross-condition transferability, achieving an accuracy of 92.12% for multi-task damage identification, a 4.49% improvement over a conventional Transformer baseline. The proposed method can be applied to health conditions identification of buildings, bridges, and trusses.
{"title":"Conv-Transformer based few-shot learning for highly accurate multi-task structural health monitoring via piezoelectric impedance","authors":"Hanqiao Sun, Jingfeng Lu, Jiawen Xu, Ruqiang Yan","doi":"10.1016/j.ymssp.2026.113967","DOIUrl":"https://doi.org/10.1016/j.ymssp.2026.113967","url":null,"abstract":"Impedance signals for structural health monitoring are often sparse and difficult to acquire in damaged conditions. Increasing the damage categories would significantly reduce accuracy. In this study, we propose a Conv-Transformer model that is capable of multi-task structural health monitoring, addressing the complexities of small sample datasets while handling multiple fault detection tasks, including mass loss and bolt loosening. The model enhances feature extraction by combining convolutional layers and multi-head attention within the Transformer encoder, focusing on the relative location of the peaks and the local feature of each peak in the impedance signals. These advantages enable highly accurate multi-task SHM with small samples of impedance signals. The proposed model is first trained on a large amount of data in mixed conditions and then fine-tuned with small sample data for an eight-class fault classification task. Experimental results show that the model demonstrates strong learning ability and cross-condition transferability, achieving an accuracy of 92.12% for multi-task damage identification, a 4.49% improvement over a conventional Transformer baseline. The proposed method can be applied to health conditions identification of buildings, bridges, and trusses.","PeriodicalId":51124,"journal":{"name":"Mechanical Systems and Signal Processing","volume":"60 1","pages":""},"PeriodicalIF":8.4,"publicationDate":"2026-02-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146146829","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}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号