Fernandez José M, Perriard Yves
Ecole Polytechnique Fédérale de Lausanne (EPFL) Institut de Production et Robotique (IPR), Laboratoire d'Actionneurs Intégrés (LAI) CH-1015 Lausanne, Switzerland.
IEEE Trans Ultrason Ferroelectr Freq Control. 2006 Jul;53(7):1352-61. doi: 10.1109/tuffc.2006.1665084.
This paper presents the sensitivity analysis of an ultrasonic linear motor using design of experiments (DOE) and the finite element (FE) optimization of its deformation amplitude. A first ultrasonic linear motor prototype has been built at the laboratory. A deformation amplitude of about 6.6 microm can be obtained by applying a 100 V voltage. The goal is to obtain a bigger deformation amplitude by varying the motor parameters, in particular the vibratory piece dimensions. First of all, a parametrization of the motor structure is carried out. Then, with the aim of reducing the variation ranges of the input parameters--but also to avoid performing a large number of simulations--a preoptimization stage is necessary. Thus, sensitivity analysis is carried out using design of experiments, which is a good way to obtain the influence of the input parameters on the objective function. Factorial designs have been chosen to find out the effects of each input factor but also the effect of their interactions. This method then is compared with Doehlert design technique, which is generally used for optimization approaches. The results show that it is absolutely necessary to take into account the quadratic terms in the model because they represent an important effect. The use of design of experiments revealed to be an interesting way to analyze numerically the ultrasonic motor as a preoptimization stage and already allows one to improve the deformation amplitude but also to reduce the input parameter variation ranges. Different FE optimization methods are then applied, and results show that the deformation amplitude can be increased by a factor higher than 10 compared to the initial design.
