Subspace-based multivariable system identification from frequency response data

Abstract

Two noniterative subspace-based algorithms which identify linear, time-invariant MIMO (multi-inpuUmultioutput) systems from frequency response data are presented. The algorithms are related to the recent time-domain subspace identification techniques. The first algorithm uses equidistantly, in frequency, spaced data and is strongly consistent under weak noise assumptions. The second algorithm uses arbitrary frequency spacing and is strongly consistent under more restrictive noise assumptions. Promising results are obtained when the algorithms are applied to real frequency data originating from a large flexible structure.

State of the art in linear system identification: time and frequency domain methods

“…[2], [16], [1], and [9]. The aspects of the weights are the same as on the weights in (8), relevance and reliability. The relevance is typically related to the distance x−x k as mentioned above, while the reliability should reflect the variance of the additive noise in (2).…”

“…The prediction error approaches (oe, pem etc) implement the routines of Section VI, while the subspace estimation command n4sid is described in [8] for frequency domain data. (See also [7].…”

Frequency-Domain Subspace Identification of Nonlinear Mechanical Systems: Application to a Solar Array Structure

“…The techniques available today are really quite sophisticated and advanced. In particular, subspace algorithms [1,2] are powerful identification methods that are routinely used for experimental and operational modal analysis [3] , but also for advanced processing such as damage detection and structural health monitoring [4] . However, nonlinearity is a frequent occurrence in engineering structures and, for this reason, subspace-based methods have recently been generalised to handle nonlinear systems.…”

An exploration of active hard mount vibration isolation for precision equipment

“…For the estimation of the state-space model parameters, a frequency domain subspace-based algorithm is used [McKelvey et al (1996), Algorithm 2]. The interested reader is referred to the original paper for further details and proofs.…”

“…For the estimation of the state-space model parameters, a frequency domain subspace identification algorithm is used [McKelvey et al (1996), Algorithm 2] (see also appendix E.5). The algorithm requires the user to select only one parameter q, which determines the size of the data matrices that are involved in the subspace identification.…”