On the identification of Hammerstein systems in the presence of an input hysteretic nonlinearity with nonlocal memory: Piezoelectric actuators - an experimental case study

The identification problem of the linear dynamic part of piezo based actuators is addressed in this paper, exploiting the use of binary signals, specifically the pseudo random binary sequences (PRBS). Due to the presence of nonlocal memory hysteretic behavior in piezoelectric active materials, the dependence of the identification results on this strongly nonlinear effect is analyzed and useful guidelines for the design of the PRBS stimulating signal are derived. Moreover, some properties of hysteresis like cancellation and congruency are experimentally analyzed and their effects on the identification process are discussed. Finally, the use of a hysteresis compensation strategy in the identification process is evaluated and discussed.