Affiliation:
1. Guangxi Scientific Experiment Center of Mining, Metallurgy and Environment
Abstract
It is difficult to obtain precise mathematical model with the traditional methods because of the hysteresis nonlinearity of piezoceramics, and it affects the displacement output accuracy of piezoelectric actuator. The paper proposes an online modeling technique combined with neural network and genetic algorithm (GA). It make the initial parameters of BP neural network be optimized, and so improves the modeling accuracy. Experimental results show that displacement maximum error is 71nm, and the average error is 26nm in the whole trip. It meets the nanopositioning accuracy requirements of piezoelectric staging .
Publisher
Trans Tech Publications, Ltd.
Reference6 articles.
1. KIM B. WASHINGTON G N, YOON H S. Hysteresis-reduced dynamic displacement control of piezoceramic stack actuators using model predictive sliding mode control [J], Smart Materials and structures, 21(5): P. 055018. (2012).
2. Zhu Yulong, Wang Yong, Zhu Junhui and Sun Fujun. Feedforward control algorithm analysis of Piezoelectric ceramic hysteresis. Mechanical and electrical information[J], eighteenth period: P. 160-161. (2013).
3. Lai zhilin, Liu xiangdong and Geng jie, etc. Sliding mode control of hyteresis of piezoceramic actuators based on inverse Preisach compensation[J], Optics and precisi engineering, 19(6):P. 1281-1290. (2011).
4. MIELKE A, Generalized prandtl-ishlinskii operators arising from homogenization and dimension reduction [J], Phycica B: Condensed Matter, 407(9): 1330-1335. (2012).
5. BADEL A, QIU J H, SEBALD G, etc. Selfsensing high speed controller for piezoelectric actor [J], Journal of Intelligent Material Systems and Structures, 19(3):P. 395-405. (2008).