Study on NN-Preisach hybrid model of piezoelectric actuator

Piezoelectric actuator is being widely used in vibration suppression and micro positioning applications for fast response, nanometer resolution, no backlash, no friction and bigger driving force. However, its inherent hysteresis characteristics between the input voltage and output displacement limit its control accuracy. An efficient way to eliminate this limitation is to model the nonlinearity. In order solve the problem, the characteristic of piezoelectric actuator was introduced, and its static hysteresis was studied by experiment. Artificial neural network was used to predict the output value that its input did not appear in the input numerical simulation form, and a NN-Preisach model was put forward to describing the hysteresis nonlinearity. Comparing previous models, the proposed model is more accurate to predict the output displacement, and overcomes some disadvantages of Preisach model.