Subspace-based system identification algorithms have been developed as an advanced technique for performing modal analysis. We introduce a novel tensor subspace-based algorithm to identify the time-varying modal parameters of bridge structures. A new time dimension is introduced in the traditional Hankel matrix, and a mathematical model of tensor subspace decomposition is established. Combined with the stabilization diagram, tensor parallel factor decomposition is used to estimate the frequencies, mode shapes, and modal damping ratios. The effectiveness of the proposed algorithm is validated by comparing it with the classical sliding-window–based stochastic subspace algorithm on a model cable-stayed bridge dynamic test. The proposed algorithm is further applied to process the dynamic responses of a real bridge health monitoring system to identify its time-varying modal frequencies. Our results demonstrated that the proposed algorithm significantly reduces computational efforts and extends the range of solution ideas for future out-only time-varying system identification problems.
AdeliHJiangX. Dynamic fuzzy wavelet neural network model for structural system identification. J Struct Eng2006; 132(1): 102–111.
3.
WordenKTomlinsonGR. Nonlinearity in structural dynamics: detection, identification and modelling. Boca Raton: CRC Press, 2019, 51.
4.
NoëlJPKerschenG. Nonlinear system identification in structural dynamics: 10 more years of progress. Mech Syst Signal Pr2017; 83: 2–35.
5.
PoulimenosAGFassoisSD. Parametric time-domain methods for non-stationary random vibration modelling and analysis—a critical survey and comparison. Mech Syst Signal Pr2006; 20(4): 763–816.
6.
SpiridonakosMDFassoisSD. Parametric identification of a time-varying structure based on vector vibration response measurements. Mech Syst Signal Pr2009; 23(6): 2029–2048.
7.
MuTZhouLYangJN. Experimental study of an adaptive sequential nonlinear LSE with unknown inputs for structural damage tracking. Shock Vib2014; 2014: 1–14.
8.
SuWCLiuCYHuangCS. Identification of instantaneous modal parameter of time‐varying systems via a wavelet‐based approach and its application. Computer‐aided Civ Inf2014; 29(4): 279–298.
9.
ReyndersEWurstenGDe RoeckG. Output-only structural health monitoring in changing environmental conditions by means of nonlinear system identification. Struct Health Monit2014; 13(1): 82–93.
10.
SpiridonakosMDFassoisSD. Non-stationary random vibration modelling and analysis via functional series time-dependent ARMA (FS-TARMA) models–a critical survey. Mech Syst Signal Pr2014; 47(1–2): 175–224.
11.
MuTZhouLYangJN. Comparison of adaptive structural damage identification techniques in nonlinear hysteretic vibration isolation systems. Earthq Eng Eng Vib2013; 12(4): 659–667.
12.
CohenL. Time-frequency distribution-a review. A. instantaneous frequency and analytic signal. Proc IEEE1989; 77(7): 968.
13.
LiJZhaoCSuH. A mode mixing elimination method of HHT in fault detection. In: 2nd International Conference on Cloud Computing and Internet of Things (CCIOT), Dalian, China, 22–23 October 2016, 120–123. Piscataway: IEEE. 2016.
14.
KlepkaAUhlT. Identification of modal parameters of non-stationary systems with the use of wavelet based adaptive filtering. Mech Syst Signal Pr2014; 47(1−2): 21–34.
15.
NoëlJPKerschenG. Frequency-domain subspace identification for nonlinear mechanical systems. Mech Syst Signal Pr2013; 40(2): 701–717.
16.
PaiPFNguyenBASundaresanMJ. Nonlinearity identification by time-domain-only signal processing. Int J Non Linear Mech2013; 54: 85–98.
17.
Van OverscheePDe MoorBL. Subspace Identification for Linear Systems: Theory—Implementation—Applications. Berlin: Springer Science and Business Media, 2012.
YangYNagarajaiahS. Time-frequency blind source separation using independent component analysis for output-only modal identification of highly damped structures. J Struct Eng2012; 139(10): 1780–1793.
20.
LiHTaoDHuangY, et al.A data‐driven approach for seismic damage detection of shear‐type building structures using the fractal dimension of time–frequency features. Struct Control Health Monit2013; 20(9): 1191–1210.
21.
WordenKStaszewskiWJHensmanJJ. Natural computing for mechanical systems research: a tutorial overview. Mech Syst Signal Pr2011; 25(1): 4–111.
22.
FeldmanMBraunS. Nonlinear vibrating system identification via Hilbert decomposition. Mech Syst Signal Pr2017; 84: 65–96.
23.
ReyndersEMaesKLombaertG, et al.Uncertainty quantification in operational modal analysis with stochastic subspace identification: validation and applications. Mech Syst Signal Pr2016; 66: 13–30.
24.
Haji Agha Mohammad ZarbafSENorouziMAllemangRJ, et al.Stay force estimation in cable-stayed bridges using stochastic subspace identification methods. J Bridge Eng2017; 22(9): 04017055.
25.
LiWVuVHLiuZ, et al.Extraction of modal parameters for identification of time-varying systems using data-driven stochastic subspace identification. J Vib Cont2018; 24(20): 4781–4796.
26.
LindquistAPicciG. Linear stochastic systems: a geometric approach to modelingestimation and identification. Berlin: Springer, 2015, Vol. 1.
27.
MellingerPDöhlerMMevelL. Variance estimation of modal parameters from output-only and input/output subspace-based system identification. J Sound Vib2016; 379: 1–27.
28.
LohCHChaoSHWengJH, et al.Application of subspace identification technique to long‐term seismic response monitoring of structures. Earthq Eng Struct Dyn2015; 44(3): 385–402.
29.
WuWHWangSWChenCC, et al.Application of stochastic subspace identification for stay cables with an alternative stabilization diagram and hierarchical sifting process. Struct Control Health Monit2016; 23(9): 1194–1213.
30.
BayraktarATürkerTAltunişikAC. Experimental frequencies and damping ratios for historical masonry arch bridges. Constr Build Mater2015; 75: 234–241.
31.
WeiXRussellJŽivanovićS, et al.Experimental investigation of the dynamic characteristics of a glass-FRP suspension footbridge. Dyn Civil Structures2017; 2: 37–47.
32.
LohCHChenJD. Tracking modal parameters from building seismic response data using recursive subspace identification algorithm. Earthquake Eng Struct Dyn2017; 46(13): 2163–2183.
33.
DaiKWangYLuW, et al.Investigation of the stochastic subspace identification method for on-line wind turbine tower monitoring. Int Soc Opt Photon2017; 10169: 101692F.
34.
CatinariFPierdiccaAClementiF, et al.Identification and calibration of the structural model of historical masonry building damaged during the 2016 Italian earthquakes: the case study of Palazzo del Podestà in Montelupone. In: AIP Conference Proceedings, China, Chongqing City, 27−28 May 2017. New York: AIP Publishing, 1: 090013.
35.
VibergM. Subspace-based state-space system identification. Circuits Syst Signal Process2002; 21(1): 23–37.
36.
FavoreelWMoorBDOverscheePV. Subspace state space system identification for industrial processes. J Process Control2000; 10(2−3): 149–155.
37.
QinSJ. An overview of subspace identification. Comput Chem Eng2006; 30(10): 1502–1513.
38.
Perez RamirezCAAmezquita SanchezJPAdeliH, et al.Time-frequency techniques for modal parameters identification of civil structures from acquired dynamic signals. J Vibroengineering2016; 18(5): 3164–3185.
39.
CandyJV. Model-based processing: an applied subspace identification approach. New Jersey: Wiley, 2019.
40.
WordenKTomlinsonGR. Nonlinearity in structural dynamics: detection, identification and modelling. Boca Raton: CRC Press, 2019.
41.
CichockiA. Era of big data processing: a new approach via tensor networks and tensor decompositions. arXiv preprint arXiv2048. 2014; 1403.
42.
KoldaTGBaderBW. Tensor decompositions and applications. SIAM Rev2009; 51(3): 455–500.
43.
SadhuAHazraBNarasimhanS. Decentralized modal identification of structures using parallel factor decomposition and sparse blind source separation. Mec Syst Signal Pr2013; 41(1−2): 396–419.
44.
AbazarsaFNateghiFGhahariSF, et al.Extended blind modal identification technique for nonstationary excitations and its verification and validation. J Eng Mech2015; 142(2): 04015078.
45.
FriesenPSadhuA. Performance of tensor decomposition-based modal identification under nonstationary vibration. Smart Mater Struct2017; 26(3): 035024.
46.
SadhuANarasimhanSAntoniJ. A review of output-only structural mode identification literature employing blind source separation methods. Mech Syst Signal Process2017; 94: 415–431.
47.
VirtaJNordhausenK. Blind source separation of tensor-valued time series. Signal Process2017; 141: 204–216.
48.
KoldaTG. Multilinear operators for higher-order decompositions, tech. Report SAND 2006-2081. Albuquerque, NM, Livermore, CA: Sandia National Laboratories, 2006.
49.
HitchcockF. The expression of a tensor or a polyadic as a sum of products. Stud Appl Math1927; 6(1−4): 164–189.
50.
CattellRB. “Parallel proportional profiles” and other principles for determining the choice of factors by rotation. Psychometrika1944; 9(4): 267–283.
51.
CarrollJDChangJJ. Analysis of individual differences in multidimensional scaling via an N-way generalization of “Eckart–Young” decomposition. Psychometrika1970; 35(3): 283–319.
52.
HarshmanRA. Foundations of the PARAFAC procedure: models and conditions for an “explanatory” multimodal factor analysis. UCLA Working Pap Phonetics1970; 16: 1–84.
53.
HarshmanRA. PARAFAC2: mathematical and technical notes. UCLA working Pap Phonetics1972; 22(3044): 122215.
54.
BakirPG. Automation of the stabilization diagrams for subspace based system identification. Expert Syst Appl2011; 38(12): 14390–14397.
