Search

Ultrasonic machining is a hybrid machining process which combination of the additional rotary and ultrasonic vibration energy. It has been successfully applied in hard and brittle materials processing. In the ultrasonic processing application, the vibration energy will be different in the same driving current under the conditions of different parameters of the transducer. This paper proposes a method to predict the driving current using Elman neural network. According to the different parameters of the transducer, the different driving current required by the same vibration energy is obtained by the Elman neural network prediction model. A number of experimental data have been collected to be used by training the network model. Training results and predictive results show that it can be well adjusted by the driving current to obtain the same vibration energy and further to get the stable ultrasonic energy.

A design method for ultrasonic torsional transducer is presented based on the one-dimensional wave equation and the equivalent circuit method. Firstly, the equivalent circuits of torsional transducer and horn are constructed. The mounting flange can be placed right at the nodal plane, where no angular displacement occurs. An ultrasonic torsional tool is fabricated according to the presented method, and the resonance frequency and torsional vibration amplitude are measured by Laser. The comparison between the measured and predicted results shows that the present method is reliable, simple and straight.