System modelling in the presence of nonlinear distortions

The aim of the study presented in this paper is to provide an experimental comparison between best linear approximation models and nonlinear models estimated for systems suffering nonlinear distortions. This is illustrated on a nonlinear mechanical resonating system. Parametric best linear approximation models are estimated from frequency response function measurements using random phase multisines. A full nonlinear model is also estimated for the system. The two types of models are then compared in terms of their performance in simulating the system and also when used as a basis for optimal controller design for the system. It is shown that in the presence of nonlinear distortions the choice between the two model types will depend on the intended use of the model.