Abstract:Aiming at the problem of frequency response function (FRF) being difficult to be modeled and estimated due to light damping, based on the local rational model,a non-parametric identification method for FRF of a lightly damped multi-variable system was proposed.To eliminate the parametric estimation deviation caused by the least-squares method, a weight function was introduced to construct a new cost function, and realize parametric consistent estimation.A numerically stable parametric optimization algorithm was established, and an uncertain interval was constructed to estimate FRF. Then, the proposed FRF non-parametric identification method was extended to the case of variable with error system.Finally, the effectiveness of the proposed method was verified with simulation and tests.
[1] AVITABILE P. Modal testing: a practitioner's guide[M]. John Wiley & Sons, 2017.
[2] SCHOUKENS J, VANDERSTEEN G, BARBÉ K, et al. Nonparametric Preprocessing in System Identification: a Powerful Tool, European Journal of Control, 2009,15 (3–4):260-274.
[3] BRILLINGER D R. Time series: Data analysis and t¬h¬eory[J]. McGraw Hall, 1981.
[4] LJUNG L. System Identification-Theory for the User 2nd edition PTR Prentice-Hall[J]. Upper Saddle River, NJ, 1999.
[5] PINTELON R, SCHOUKENS J, VANDERSTEEN G. Frequency domain system identification using arbitrary signals[J]. IEEE Transactions on automatic control, 1997,42(12):1717-1720.
[6] MCKELVEY T. Frequency domain identification meth¬ods[J]. Circuits, Systems and Signal Processing, 2002,21(1):39-55.
[7] PINTELON R, SCHOUKENS J, VANDERSTEEN G, et al. Estimation of nonparametric noise and FRF mo¬d¬els for multivariable systems—Part I: Theory[J]. Me¬c¬hani¬cal Systems and Signal Processing, 2010,24(3):573-595.
[8] MCKELVEY T, GUÉRIN G. Non-parametric frequenc¬y response estimation using a local rational model1[J]. IFAC Proceedings Volumes, 2012,45(16):49-54.
[9] HÄGG P, SCHOUKENS J, GEVERS M, et al. The t¬r¬ansient impulse response modeling method for non-p¬a¬rametric system identification[J]. Automatica, 2016,68:314-328.
[10] LATAIRE J, CHEN T. Transfer function and transient estimation by Gaussian process regression in the frequ¬ency domain[J]. Automatica, 2016,72:217-229.
[11] DARWISH M A, LATAIRE J, TÓTH R. Bayesian fr¬e¬quency domain identification of LTI systems with O¬BFs kernels[J]. IFAC-PapersOnLine, 2017,50(1):6238-6243.
[12] GEERARDYN E, OOMEN T, SCHOUKENS J. Enha¬n¬c¬ing ℋ∞ Norm Estimation using Local LPM/LRM M¬odeling: Applied to an AVIS[J]. IFAC Proceedings Volumes, 2014,47(3):10856-10861.
[13] GEERARDYN E. Development of user-friendly system identification techniques[Z]. Ph. D. thesis, Vrije Univ¬ersiteit Brussel, 2016.
[14] EVERS E, de JAGER B, OOMEN T. Improved Local Rational Method by incorporating system knowledge: with application to mechanical and thermal dynamical systems[J]. IFAC-PapersOnLine, 2018,51(15):808-813.
[15] VOORHOEVE R, van der MAAS A, OOMEN T. Non-parametric identification of multivariable systems: A local rational modeling approach with application to a vibration isolation benchmark[J]. Mechanical Systems a¬nd Signal Processing, 2018,105:129-152.
[16] PINTELON R, SCHOUKENS J. System identification: a frequency domain approach[M]. John Wiley & Sons, 2012.
[17] PINTELON R, GUILLAUME P, ROLAIN Y, et al. Parametric identification of transfer functions in the fr¬e¬quency domain-a survey[J]. IEEE transactions on aut¬o¬matic control, 1994,39(11):2245-2260.
[18] ZHANG E, PINTELON R, SCHOUKENS J. Errors-in-v-a¬riables identification of dynamic systems excited by ar¬b¬i¬trary non-white input[J]. Automatica, 2013,49(10):3032-3041.
[19] DENNIS JR J E, SCHNABEL R B. Numerical meth¬o¬ds for unconstrained optimization and nonlinear equa¬t¬i¬ons[M]. Siam, 1996.