The advancement in the development of high-precision motion systems in the nanometer/micrometer range has led to the increased utilization of smart actuators. The unique advantages of piezoelectric actuators (PEAs), such as high precision, high force-to-weight ratio, low weight, and fast response, make them suitable for applications requiring precise positioning. However, the inherent nonlinear behavior of piezoelectric actuators, known as hysteresis, is the primary factor limiting performance accuracy. Additionally, the presence of high-frequency dynamics in the actuator structure restricts its performance only in high frequencies. To obtain the desired performance, proper modeling and designing an effective controller is necessary. Advancing model-based control methods poses a critical challenge in accurately modeling and identifying the controlled system. Therefore, the primary objective here is to model-free control for piezoelectric actuators. Another significant hurdle lies in the digital implementation of controllers; digital control introduces noise and time delay, leading to a reduction in control system performance. Therefore, the second objective focuses on a sampled-data design approach and designing a suitable controller based on this method. A pivotal step toward achieving this objective involves developing a sampled-data model for the piezoelectric actuator. Finally, to enhance controller performance at high frequencies, the design of an appropriate controller, considering the time delay of the control system, is undertaken. One noteworthy technique for elevating control system performance involves using a suitable estimation method to estimate uncertainties and external disturbances. The extended state observer (ESO) is an effective and robust method. In this thesis, the ESO is employed to achieve heightened performance speed alongside acceptable tracking accuracy. Another noteworthy consideration in this thesis is maintaining the control signal within permissible bounds to safeguard the actuator. In this thesis, four comprehensive algorithms are introduced to address the specified control objectives: a continuous-time sliding mode control (SMC) based on ESO, a sampled-data model-free adaptive control (SDMFAC) based on sampled-data ESO (SDESO) considering input rate constraints, a sampled-data model-free adaptive sliding mode control (SDMFA-SMC) subject to time delay, a SDMFA-SMC based on SDESO considering both control system delays and input rate constraints. Furthermore, the thesis presents experimental results for each of the four algorithms. To eva‎luate the performance of the proposed control algorithms, experiments are conducted to track sinusoidal reference signals. To gauge the effectiveness of control algorithms, we assess their performance using key indexes: maximum absolute error (MAE), root mean square error (RMSE), peaking value, and convergence time of the error to a zone around zero. These indexes provide a clear eva‎luation of how well the algorithms minimize tracking errors, offering a comprehensive understanding of their performance in approaching high-frequency tracking.

ايمان ايزدي نجف آبادي

سيد فخرالدين عالم , حامد جلالي بيدگلي

جعفر قيصري , مرضيه كمالي , محسن اكراميان