A historical perspective of industrial applications of system identification is provided with a short review of early applications of system identification techniques to full-scale industrial processes and nuclear power plants. More recent applications of identification techniques in industry will feature in the remainder of this book, as the techniques are described. The role of signal design in identification for control is then expounded. Time domain and frequency domain specifications for the identification of linear systems are explained. Performance measures for linear system identification are discussed where two important measures are introduced, namely the performance index for perturbation signals and the effective minimum ratio between the actual amplitude and the specified amplitude at any of the specified harmonics. The identification framework is then extended to nonlinear systems where the concept of harmonic suppression is explained. Finally, a comparison between periodic and non-periodic signals is provided. Three types of commonly used non-periodic signals are discussed. It is shown that periodic signals offer many advantages over non-periodic ones. Nevertheless, non-periodic signals remain attractive choices particularly for preliminary tests due to them being easy to generate and apply.
- Makino K, Saito Y, Fons P, Kolobov AV, Nakano T, Tominaga J, Hase M (2016) Anisotropic lattice response induced by a linearly-polarized femtosecond optical pulse excitation in interfacial phase change memory material. Sci Rep 6: Article 19758Google Scholar