Abstract
The hysteresis of piezoceramic actuators depends very much on the variation of the input rate. In order to describe this dynamic characteristic, the paper presents a new dynamic Preisach model as an extension of the existing dynamic Preisach model by introducing the history information of inputs and outputs into the Preisach function. Then a neural network is used for the identification of the model. The model is shown through experiments to offer high accuracy under band-limited white noise voltage excitations with different frequency bands. When the frequency band of the voltage is 50 Hz. the maximum of the absolute error as predicted by the new model is reduced by 66% compared with the existing dynamic Preisach model and by 91% compared with the classical Preisach model.
| Original language | English |
|---|---|
| Pages (from-to) | 205-209 |
| Number of pages | 5 |
| Journal | Beijing Ligong Daxue Xuebao/Transaction of Beijing Institute of Technology |
| Volume | 27 |
| Issue number | SUPPL. 1 |
| Publication status | Published - May 2007 |
Keywords
- Dynamic Preisach model
- Hysteresis
- Neural networks
- Piezoceramic actuators