Capacitance estimation for piezoelectric actuators, an artificial intelligence approach

This research aimed to investigate charge-based position estimation/control of piezo-actuated nanopositioning systems using a data-driven approach. In the analysis of these systems, piezoelectric actuators are widely approximated as capacitors with a fixed capacitance from an electrical viewpoint. This assumption was examined and found to be highly inaccurate. It was evidently demonstrated that the capacitance of piezoelectric actuators varies significantly with operating conditions (i.e., the frequency and amplitude of the excitation voltage). This paper also offers an alternative: considering the piezoelectric actuator as a capacitor with varying capacitance based on its operating conditions for analysis and design purposes. A linear model and an artificial intelligence (AI) model were developed to estimate the actuator capacitance on the basis of its operating conditions. The results demonstrate that the AI model outperforms the linear model and accurately estimates the capacitance of the piezoelectric actuator within the experimented range. Findings of this research pave the way to uplift the precision of piezo-actuated nanopositioning systems.