TY - JOUR
T1 - Development of a charge estimator for piezoelectric actuators: a radial basis function approach
AU - Morteza Mohammadzaheri, Morteza
AU - Emadi, Mohammadreza
AU - Ghodsi, Mojtaba
AU - Saleem, Ashraf
PY - 2020/6/1
Y1 - 2020/6/1
N2 - Charge of a piezoelectric actuator is proportional to its displacement for a wide area of operating. Hence, a charge estimator can estimate displacement for such actuators. However, existing charge estimators take a sizable portion of the excitation voltage, i.e. voltage drop. Digital charge estimators have presented the smallest voltage drop. This paper first investigates digital charge estimators and suggests a design guideline to (i) maximise accuracy and (ii) minimise the voltage drop. Digital charge estimators has a sensing resistor; an estimator with a constant resistance is shown to violate the design guideline; while, all existing digital charge estimators use one or a few intuitively chosen resistors. That is, existing estimators witness unnecessarily large inaccuracy and/or voltage drop. This research develops charge estimators with varying resistors, fulfilling the design guideline. Several methods are tested to estimates the sensing resistance based on operating conditions, and radial basis function networks models excel in terms of accuracy.
AB - Charge of a piezoelectric actuator is proportional to its displacement for a wide area of operating. Hence, a charge estimator can estimate displacement for such actuators. However, existing charge estimators take a sizable portion of the excitation voltage, i.e. voltage drop. Digital charge estimators have presented the smallest voltage drop. This paper first investigates digital charge estimators and suggests a design guideline to (i) maximise accuracy and (ii) minimise the voltage drop. Digital charge estimators has a sensing resistor; an estimator with a constant resistance is shown to violate the design guideline; while, all existing digital charge estimators use one or a few intuitively chosen resistors. That is, existing estimators witness unnecessarily large inaccuracy and/or voltage drop. This research develops charge estimators with varying resistors, fulfilling the design guideline. Several methods are tested to estimates the sensing resistance based on operating conditions, and radial basis function networks models excel in terms of accuracy.
U2 - 10.4018/IJAIML.2020010103
DO - 10.4018/IJAIML.2020010103
M3 - Article
SN - 2642-1585
VL - 10
SP - 31
EP - 44
JO - International Journal of Artificial Intelligence and Machine Learning
JF - International Journal of Artificial Intelligence and Machine Learning
IS - 1
M1 - 3
ER -