RESUMEN
Piezoelectric meso- and microactuator systems required for manipulation or assembly of microscale objects demand reliable force and/or displacement information. Available sensors are prone to dimension restrictions or precision limitation. Self-sensing method, based on the electric charge measurement, may represent a solution in terms of cost-effectiveness and integration, the actuator performing simultaneously as its own sensor. This paper presents a self-sensing method dedicated to free uni- and bimorph piezocantilevers but can also be adapted to other piezoactuator types. The integrated electric current, used to convert the charge, can be compensated against piezoelectric material nonlinearities to provide accurate displacement information. The advantages relative to existing self-sensing methods consist in the ability to keep this displacement information for long-term periods (more than a thousand seconds) and in the reduction in signal noise. After introductive issues related to the method the base principle allowing the estimation of tip displacement is presented. Then, the identification procedure of the estimator parameters is depicted and representative experimental results are shown. Finally, a series of aspects related to electronic circuits are discussed, useful for successful system implementation.
RESUMEN
This paper presents a new method of self-sensing both of the displacement and the external applied force at the tip of piezoelectric cantilevers. Integrated electric current across piezoelectric actuators is compensated against material nonlinearities (creep, hysteresis) to provide reliable information. We propose to compensate the hysteresis by using the Prandtl-Ishlinskii static approach while an auto regressive and moving average exogenous (ARMAX) model is used to minimize the creep influence. The quasistatic estimation, electronic circuit, and aspects related to long-term charge preservations are described or referenced. As an experiment, we tested the actuator entering in contact with a fixed force sensor. An input signal of 20 V peak-to-peak (10% of maximum range) led to force self-sensing errors inferior to +/-8%. A final discussion about method accuracy and its limitations is made.
