Nonlinear damping identification using extended continuous wavelet transform and long-short term memory: Application to a spur gear pair system

Nourhaine Yousfi; Fatma  Mejdoub; Ali  Akrout; Lassaad Walha; Mohamed  Haddar

doi:10.59400/sv3869

Nonlinear damping identification using extended continuous wavelet transform and long-short term memory: Application to a spur gear pair system

Nourhaine Yousfi
Mechanics, Modelling and Production Laboratory (LA2MP), National Engineering School of Sfax, University of Sfax, Sfax 3038, Tunisia
Fatma Mejdoub
Laboratory of Electromechanical Systems (LASEM), National School of Engineers of Sfax, University of Sfax, Sfax 3038, Tunisia
Ali Akrout
Mechanics, Modelling and Production Laboratory (LA2MP), National Engineering School of Sfax, University of Sfax, Sfax 3038, Tunisia
Lassaad Walha
Mechanics, Modelling and Production Laboratory (LA2MP), National Engineering School of Sfax, University of Sfax, Sfax 3038, Tunisia
Mohamed Haddar
Mechanics, Modelling and Production Laboratory (LA2MP), National Engineering School of Sfax, University of Sfax, Sfax 3038, Tunisia

Article ID: 3869

Keywords: nonlinear damping; continuous wavelet transform (CWT); long short-term memory (LSTM); system stability; time-frequency feature

Abstract

Damping significantly affects the dynamic analysis of spur gear pair systems. The identified damping ratios may suffer from instability owing to many reasons, such as time-varying conditions and nonlinear effects. Long-short-term memory (LSTM) has been developed for damping model identification in systems with nonlinear behavior. However, owing to the poor quality of the input data, the identification results may not be reliable. In this regard, this study proposes a hybrid technique based on the Continuous Wavelet Transform (CWT) and LSTM methods. The major novelty of this study is the utilization of the time-frequency information provided by the CWT as the input data. Therefore, the LSTM network was fed these extracted features to enhance noise attenuation and improve the robustness and stability of nonlinear damping identification. Thus, the CWT technique is used as a preprocessing tool for the observed signals, enhancing the data quality by reducing the influence of noise. To verify the effectiveness of the proposed CWT–LSTM approach for damping identification, CWT representations of the simulated spur gear pair system response were used in a series of analyses. The numerical results indicate that the combined CWT–LSTM approach provides more reliable and accurate nonlinear damping estimation than the conventional LSTM model. This methodology has a strong potential for the accurate identification of damping in gear transmission systems.

Published

2026-03-27

How to Cite

Yousfi, N., Mejdoub , F., Akrout, A., Walha, L., & Haddar, M. (2026). Nonlinear damping identification using extended continuous wavelet transform and long-short term memory: Application to a spur gear pair system. Sound & Vibration, 60(2). https://doi.org/10.59400/sv3869

Download Citation

Issue

Vol. 60 No. 2 (2026): in progress

Section

Article

This work is licensed under a Creative Commons Attribution 4.0 International License.

References

[1]Al-hababi T, Cao M, Saleh B, et al. A Critical Review of Nonlinear Damping Identification in Structural Dynamics: Methods, Applications, and Challenges. Sensors. 2020; 20(24): 7303. doi: 10.3390/s20247303

[2]Ghahari F, Sargsyan K, Taciroglu E. Quantification of modeling uncertainty in the Rayleigh damping model. Earthquake Engineering & Structural Dynamics. 2024; 53(9): 2950–2956. doi: 10.1002/eqe.4143

[3]Zahid FB, Ong ZC, Khoo SY. A review of operational modal analysis techniques for in-service modal identification. Journal of the Brazilian Society of Mechanical Sciences and Engineering. 2020; 42(8): 398. doi: 10.1007/s40430-020-02470-8

[4]Lisitano D, Bonisoli E, Mottershead JE. Experimental direct spatial damping identification by the Stabilised Layers Method. Journal of Sound and Vibration. 2018; 437: 325–339. doi: 10.1016/j.jsv.2018.08.055

[5]Bonisoli E, Lisitano D, Vigliani A. Damping identification and localisation via Layer Method: Experimental application to a vehicle chassis focused on shock absorbers effects. Mechanical Systems and Signal Processing. 2019; 116: 194–216. doi: 10.1016/j.ymssp.2018.06.013

[6]Amabili M, Alijani F, Delannoy J. Damping for large-amplitude vibrations of plates and curved panels, part 2: Identification and comparisons. International Journal of Non-Linear Mechanics. 2016; 85: 226–240. doi: 10.1016/j.ijnonlinmec.2016.05.004

[7]Balasubramanian P, Ferrari G, Amabili M. Identification of the viscoelastic response and nonlinear damping of a rubber plate in nonlinear vibration regime. Mechanical Systems and Signal Processing. 2018; 111: 376–398. doi: 10.1016/j.ymssp.2018.03.061

[8]Haghdoust P, Lo Conte A, Cinquemani S, et al. A Numerical Method to Model Non-linear Damping Behaviour of Martensitic Shape Memory Alloys. Materials. 2018; 11(11): 2178. doi: 10.3390/ma11112178

[9]Müller F, Woiwode L, Gross J, et al. Nonlinear damping quantification from phase-resonant tests under base excitation. Mechanical Systems and Signal Processing. 2022; 177: 109170. doi: 10.1016/j.ymssp.2022.109170

[10]Chatterjee A, Chintha HP. Identification and parameter estimation of cubic nonlinear damping using harmonic probing and volterra series. International Journal of Non-Linear Mechanics. 2020; 125: 103518. doi: 10.1016/j.ijnonlinmec.2020.103518

[11]Eberle R. Estimating Nonlinear Damping for Mechanical Single‐Degree‐of‐Freedom Systems: A Robust and Effective Approach. PAMM. 2025; 25(1): e202400158. doi: 10.1002/pamm.202400158

[12]Han SL, Kinoshita T. Nonlinear Damping Identification in Nonlinear Dynamic System Based on Stochastic Inverse Approach. Mathematical Problems in Engineering. 2012; 2012(1): 574291. doi: 10.1155/2012/574291

[13]Zaparoli Cunha B, Droz C, Zine AM, et al. A review of machine learning methods applied to structural dynamics and vibroacoustic. Mechanical Systems and Signal Processing. 2023; 200: 110535. doi: 10.1016/j.ymssp.2023.110535

[14]Yemets K, Izonin I, Dronyuk I. Enhancing the FFT-LSTM Time-Series Forecasting Model via a Novel FFT-Based Feature Extraction–Extension Scheme. Big Data and Cognitive Computing. 2025; 9(2): 35. doi: 10.3390/bdcc9020035

[15]Yun DY, Oh BK, Park K, et al. LSTM‐Based Approach for Stable Identification of Modal Damping Ratio in Building Structures. Structural Control and Health Monitoring. 2024; 2024(1): 6645626. doi: 10.1155/2024/6645626

[16]Liu D, Tang Z, Bao Y, et al. Machine‐learning‐based methods for output‐only structural modal identification. Structural Control and Health Monitoring. 2021; 28(12). doi: 10.1002/stc.2843

[17]Chen Y, Rao M, Feng K, et al. Physics-Informed LSTM hyperparameters selection for gearbox fault detection. Mechanical Systems and Signal Processing. 2022; 171: 108907. doi: 10.1016/j.ymssp.2022.108907

[18]Afridi YS, Hasan L, Ullah R, et al. LSTM-Based Condition Monitoring and Fault Prognostics of Rolling Element Bearings Using Raw Vibrational Data. Machines. 2023; 11(5): 531. doi: 10.3390/machines11050531

[19]Kulkarni A, Paul ASB. A hybrid EKF-LSTM framework with GA-tuned noise covariance for enhanced time series forecasting. Engineering Research Express. 2025; 7(3): 035266. doi: 10.1088/2631-8695/adf9be

[20]Zheng Y, Lee CL, Guo J, et al. Improving EFDD with Neural Networks in Damping Identification for Structural Health Monitoring. Sensors. 2025; 25(22): 6929. doi: 10.3390/s25226929

[21]Yun DY, Park HS. Modal identification of building structures under unknown input conditions using extended Kalman filter and long-short term memory. Integrated Computer-Aided Engineering. 2023; 30(2): 185–201. doi: 10.3233/ICA-220696

[22]Yun DY, Shim HB, Park HS. SSI-LSTM network for adaptive operational modal analysis of building structures. Mechanical Systems and Signal Processing. 2023; 195: 110306. doi: 10.1016/j.ymssp.2023.110306

[23]Han Z, She D, Liu J. LSTM-Based Fast Prediction of Seismic Response and Fragility for Bridge Pile-Group Foundations: A Data-Driven Design Approach. Designs. 2026; 10(2): 37. doi: 10.3390/designs10020037

[24]Chen Z, Li C, Sanchez RV. Gearbox Fault Identification and Classification with Convolutional Neural Networks. Shock and Vibration. 2015; 2015: 1–10. doi: 10.1155/2015/390134

[25]Yin A, Yan Y, Zhang Z, et al. Fault Diagnosis of Wind Turbine Gearbox Based on the Optimized LSTM Neural Network with Cosine Loss. Sensors. 2020; 20(8): 2339. doi: 10.3390/s20082339

[26]Zhang BC, Sun SQ, Yin XJ, et al. Research on Gearbox Fault Diagnosis Method Based on VMD and Optimized LSTM. Applied Sciences. 2023; 13(21): 11637. doi: 10.3390/app132111637

[27]He C, Yasenjiang J, Lv L, et al. Gearbox Fault Diagnosis Based on MSCNN-LSTM-CBAM-SE. Sensors. 2024; 24(14): 4682. doi: 10.3390/s24144682

[28]Xia X, Sun H, Wang A. Fault Diagnosis of Planetary Gearboxes Based on LSTM Improved via Feature Extraction Using VMD, Fusion Entropy, and Random Forest. Entropy. 2025; 27(9): 956. doi: 10.3390/e27090956

[29]Cheng X, Dou S, Du Y, et al. Gearbox fault diagnosis method based on lightweight channel attention mechanism and transfer learning. Scientific Reports. 2024; 14(1): 743. doi: 10.1038/s41598-023-50826-6

[30]Fang X, Luo H, Tang J. Structural damage detection using neural network with learning rate improvement. Computers & Structures. 2005; 83(25–26): 2150–2161. doi: 10.1016/j.compstruc.2005.02.029

[31]Chang CC, Chang TYP, Xu YG, et al. Structural Damage Detection Using an Iterative Neural Network. Journal of Intelligent Materials Systems and Structures. 2000; 11(1): 32–42. doi: 10.1177/104538900772664387

[32]Lei Y, Liu C, Liu LJ. Identification of multistory shear buildings under unknown earthquake excitation using partial output measurements: Numerical and experimental studies. Structural Control and Health Monitoring. 2014; 21: 774–783. doi: 10.1002/stc.1600

[33]Yousfi N, Zghal B, Akrout A, et al. Damping models identification of a spur gear pair. Mechanism and Machine Theory. 2018; 122: 371–388. doi: 10.1016/j.mechmachtheory.2018.01.002

[34]Dien NP. Damping identification in multi-degree-of-freedom systems using the continuous wavelet transform. Vietnam Journal of Mechanics. 2005; 27(1): 41–50. doi: 10.15625/0866-7136/27/1/5713

[35]Li H. Gear Damage Detection Using Impulse Response Wavelet Transform. Advanced Materials Research. 2011; 204–210: 1378–1381. doi: 10.4028/www.scientific.net/AMR.204-210.1378

[36]Nguyen TD, Nguyen PD. Improvements in the Wavelet Transform and Its Variations: Concepts and Applications in Diagnosing Gearbox in Non-Stationary Conditions. Applied Sciences. 2024; 14(11): 4642. doi: 10.3390/app14114642

This paper introduces a universal framework for understanding the vibration responses of systems subjected to harmonic excitation. By examining a simplified cylinder-spring-damper model, the study refurbishes traditional scaling methods for the excitation frequency ratio and critical damping ratio. The findings indicate that in damped systems, the maximum amplitude of vibration does not align with the natural frequency. This observation leads to the introduction of a new scaling method for reduced frequency. This new approach aligns resonance peaks at the new reduced velocity of 1.0 across different damping ratios, providing a consistent characterization of vibration behavior. A new critical damping ratio of 0.707 is identified for an excited system as opposed to the traditional damping ratio of 1.0 for an unexcited system. Key properties such as maximum amplitude, phase lag, bandwidth, and quality factor are analyzed, demonstrating that the proposed reduced frequency and critical damping ratio effectively capture the dynamics of both damped and undamped excited systems. The findings offer significant insights for practical applications in engineering and various scientific fields.

This study makes significant contributions to the field of ultrasonic testing (UT) by offering a novel approach to the identification of artificially introduced defects within Japanese cedar wood (Cryptomeria japonica). The findings are of particular relevance for the heritage conservation and construction sectors, where non-invasive defect detection is paramount. The study establishes a robust framework for assessing the structural integrity of timber by correlating ultrasonic wave velocity reductions with defect size and distribution. Big-sized defects led to more substantial decreases in wave velocity. The study establishes a robust framework for assessing the structural integrity of historical timber by correlating ultrasonic wave velocity reductions with defect size and distribution. This framework has the potential to be applicable to diverse wood species and defect types.

The control of vehicle interior noise has become a critical metric for assessing noise, vibration, and harshness (NVH) in vehicles. During the initial phases of vehicle development, accurately predicting the impact of road noise on interior noise is essential for reducing noise levels and expediting the product development cycle. In recent years, data-driven methods based on machine learning have gained significant attention due to their robust capability in navigating complex data mapping relationships. Notably, surrogate models have demonstrated exceptional performance in this domain. Numerous researchers have integrated diverse intelligent algorithms into the study of vehicle noise, leveraging advantages such as the elimination of precise modeling requirements, extensive solution space exploration, continuous learning from data, and robust algorithmic versatility. However, in NVH engineering applications, data-driven models face inherent limitations, particularly in interpretability and stability. To address these issues, this paper introduces an improved Long Short-Term Memory (LSTM) network that combines knowledge and data. Inspired by the physical information neural network concept, this approach incorporates values calculated through empirical formulas into the neural network as constraints. Comparative assessments with traditional LSTM networks highlight the advantages of this deep learning model. By integrating empirical formulas constraints, the model not only enhances interpretability but also achieves robust generalization with fewer data samples. The proposed method is validated on a specific vehicle model, showing significant improvements in prediction accuracy and efficiency.