The quadratic compression method for model updating and its noise filtering properties

The paper presents the quadratic compression method (QCM) method for model updating. The modelling error is defined in the parametric set-up, i.e. with pre-specified principal submatrices multiplied by unknown scalar parameters. The optimal parameters are obtained by minimising the error in a squared down version of the eigenvalue equation, thus with reduced computation. The method is shown to belong to the class of Minimisation of the Error in the Characteristic Equation, with a particular choice of the weighting matrix. Analysis of the propagation of the noise into the identified parameters reveals that QCM has desirable noise filtering properties. The paper presents also a weighted version of the method, called WQCM, which is motivated by further reducing the effect of measurement noise. In addition to the theoretical analysis, the superior robustness to noise properties of QCM and WQCM are demonstrated by simulations and experimentally.