In this paper, a novel integrated wavelet-machine learning approach is presented to identify critical dynamic parameters of damped dynamical systems based on wavelet transform method (WTM) and support vector regression (SVR). First, a novel data preprocessing technique based on WTM was introduced to extract important features from the raw data of time series with high-dimensional and highly nonlinear features. Moreover, a computational tool based on SVR was developed to excavate the predictive models from dimension-reduced data of the time series generated from damped dynamical systems. The proposed integrated wavelet-SVR method is capable of learning data from given damped dynamical systems with known dynamic parameters and then making parametric identification for the unknown damped dynamical systems. With the help of data preprocessing by WTM, the computational efficiency and identification accuracy of the integrated wavelet-SVR method could be improved greatly. Finally, extensive numerical experiments were conducted to validate the computational performance of the proposed wavelet-SVR approach on randomly generated data. Furthermore, this new approach has the best performance for resisting noise compared with other methods.
Citation: Hao Dong, Jiale Linghu, Yiheng Lou. Integrated wavelet-machine learning approach for effective identification of dynamic parameters in damped dynamical systems[J]. Big Data and Information Analytics, 2025, 9: 372-384. doi: 10.3934/bdia.2025018
In this paper, a novel integrated wavelet-machine learning approach is presented to identify critical dynamic parameters of damped dynamical systems based on wavelet transform method (WTM) and support vector regression (SVR). First, a novel data preprocessing technique based on WTM was introduced to extract important features from the raw data of time series with high-dimensional and highly nonlinear features. Moreover, a computational tool based on SVR was developed to excavate the predictive models from dimension-reduced data of the time series generated from damped dynamical systems. The proposed integrated wavelet-SVR method is capable of learning data from given damped dynamical systems with known dynamic parameters and then making parametric identification for the unknown damped dynamical systems. With the help of data preprocessing by WTM, the computational efficiency and identification accuracy of the integrated wavelet-SVR method could be improved greatly. Finally, extensive numerical experiments were conducted to validate the computational performance of the proposed wavelet-SVR approach on randomly generated data. Furthermore, this new approach has the best performance for resisting noise compared with other methods.
| [1] |
Tan JB, Liu Y, Wang L, Yang WG, (2008) Identification of modal parameters of a system with high damping and closely spaced modes by combining continuous wavelet transform with pattern search. Mech Syst Signal Process 22: 1055–1060. https://doi.org/10.1016/j.ymssp.2007.11.017 doi: 10.1016/j.ymssp.2007.11.017
|
| [2] |
Wong YS, Lee BHK, Wong TKS, (2001) Parameter extraction by parallel neural networks. Intell Data Anal 5: 59–71. https://doi.org/10.3233/IDA-2001-5105 doi: 10.3233/IDA-2001-5105
|
| [3] |
Chun N, Moon J, Lee HE, (2017) Alternative numerical method for identification of flutter on free vibration. Wind Struct 24: 351–365. https://doi.org/10.12989/was.2017.24.4.351 doi: 10.12989/was.2017.24.4.351
|
| [4] |
Barros-Rodríguez J, Fernandez-Fructuoso, JM, Flores Le Roux R, Sánchez Prieto S, Rodríguez Polo O, (2018) Unveiling modal parameters with forced response using SVD and QR during flutter flight testing. Proc Inst Mech Eng Part G 232: 68–76. https://doi.org/10.1177/0954410016673393 doi: 10.1177/0954410016673393
|
| [5] |
Smola AJ, Schölkopf B, (2004) A tutorial on support vector regression. Stat Comput 14: 199–222. https://doi.org/10.1023/B:STCO.0000035301.49549.88 doi: 10.1023/B:STCO.0000035301.49549.88
|
| [6] | Zhou ZH, (2016) Machine Learning, Tsinghua University Press. |
| [7] | Pathak RS, (2009) The Wavelet Transform, Atlantis Press/World Scientific. https://doi.org/10.2991/978-94-91216-24-4 |
| [8] | Gu X, Lü G, (2010) The multiscale decomposition of emg based on wavelet transform. Math Pract Theory, 40: 104–109. |
| [9] |
Boltežar M, Slavič J, (2004) Enhancements to the continuous wavelet transform for damping identifications on short signals. Mech Syst Signal Process 18: 1065–1076. https://doi.org/10.1016/j.ymssp.2004.01.004 doi: 10.1016/j.ymssp.2004.01.004
|
| [10] |
Dziedziech K, Staszewski WJ, Uhl T, (2015) Wavelet-based modal analysis for time-variant systems. Mech Syst Signal Process 50: 323–337. https://doi.org/10.1016/j.ymssp.2014.05.003 doi: 10.1016/j.ymssp.2014.05.003
|
| [11] |
Zhang Y, Wong YS, Deng J, Anton C, Gabos S, Zhang W, et al. (2016) Machine learning algorithms for mode-of-action classification in toxicity assessment. BioData Min 9: 19. https://doi.org/10.1186/s13040-016-0098-0 doi: 10.1186/s13040-016-0098-0
|
| [12] |
Bento PMR, Pombo JAN, Calado MRA, Mariano SJPS, (2018) A bat optimized neural network and wavelet transform approach for short-term price forecasting. Appl Energy 210: 88–97. https://doi.org/10.1016/j.apenergy.2017.10.058 doi: 10.1016/j.apenergy.2017.10.058
|
| [13] |
Kalteh AM, (2013) Monthly river flow forecasting using artificial neural network and support vector regression models coupled with wavelet transform. Comput Geosci 54: 1–8. https://doi.org/10.1016/j.cageo.2012.11.015 doi: 10.1016/j.cageo.2012.11.015
|
| [14] |
Rezaiy R, Shabri A, (2025) Integrating wavelet transform and support vector machine for improved drought forecasting based on standardized precipitation index. J Hydroinform 27: 320–337. https://doi.org/10.2166/hydro.2025.292 doi: 10.2166/hydro.2025.292
|
| [15] |
Popescu CA, Wong YS, Lee BHK, (2009) An expert system for predicting nonlinear aeroelastic behavior of an airfoil. J Sound Vib 319: 1312–1329. https://doi.org/10.1016/j.jsv.2008.06.035 doi: 10.1016/j.jsv.2008.06.035
|
| [16] |
Kou WB, Dong H, Zou MQ, Han JY, Jia XX, (2021) Hybrid wavelet-based learning method of predicting effective thermal conductivities of hybrid composite materials. Acta Phys Sin 70. https://doi.org/10.7498/aps.70.20201085 doi: 10.7498/aps.70.20201085
|
Figures(4) / Tables(5)
Hao Dong, Jiale Linghu, Yiheng Lou. Integrated wavelet-machine learning approach for effective identification of dynamic parameters in damped dynamical systems[J]. Big Data and Information Analytics, 2025, 9: 372-384. doi: 10.3934/bdia.2025018
DownLoad: