#Product Trends
Single Neuron Adaptive Hysteresis Compensation of Piezoelectric Actuator Based on Hebb Learning Rule
Single neuron adaptive hysteresis compensation of piezoelectric ceramic actuator based on Hebb learning rule Research direction: micro nano positioning.
Experiment Name:
Single neuron adaptive hysteresis compensation of piezoelectric ceramic actuator based on Hebb learning rule
Research direction: micro nano positioning
Experimental content:
The piezoelectric actuator has hysteresis nonlinearity, which greatly reduces its motion accuracy. Because of the time-varying and asymmetric characteristics of its hysteresis, it increases the difficulty of Hysteresis Modeling and compensation. In this experiment, a single neuron adaptive control method is used to compensate the hysteresis nonlinearity of piezoelectric actuator on-line, so as to improve the tracking performance of piezoelectric actuator.
Test purpose:
To verify the performance of the hysteresis compensation algorithm.
Test equipment:
DSPACE real-time acquisition module, dynamic bridge strain gauge, high frequency power amplifier ATA-4052.
Experiment process: hardware connection and software operation interface.
The tested object is pzs001 piezoelectric ceramic driver produced by thorlabs company. The maximum displacement is 12.925 μm under the maximum driving voltage of 100V. Using the ATA-4052 amplifier, the control piezoelectricity is amplified into the driving voltage of the piezoelectric ceramic driver. The piezoelectric ceramic driver is equipped with four resistance strain gauges, which form a 4-bridge resistance strain gauge. Sdy2105 bridge amplifier produced by Beidaihe Institute of practical electronic technology is used to measure the deformation of piezoelectric ceramic actuator.The hardware connection diagram of the test system （Figure 1）
The test program is written under Matlab / Simulink, and runs through the micro labbox real-time controller produced by dSPACE company. The test flow is as follows: firstly, a sinusoidal signal of 0-10V is generated in the control program. After being amplified by an amplifier, the piezoelectric ceramic is driven to move forward and backward. The real-time measurement of the control signal and displacement signal is completed by using microlabbox. According to the characteristics of the piezoelectric ceramic, a single neuron adaptive compensation algorithm is written, and the performance test of the algorithm is completed by using equipment.
The Simulink code of the test program is shown as follows:
Test results:
The effect of the control algorithm in tracking sinusoidal and triangular trajectories is tested respectively. For sinusoidal trajectories, the single neuron adaptive compensation algorithm can effectively eliminate the influence of hysteresis nonlinearity. Compared with the traditional PID control, the single neuron adaptive compensation algorithm has higher adaptability and robustness. For sinusoidal trajectories within 50Hz, it can well eliminate hysteresis nonlinearity. For triangular trajectory, single neuron adaptive compensation algorithm can achieve similar results. The experimental results（Figure 2、3）
The performance of the amplifier in this experiment:
The control signal is weak current, and its voltage range is 0-10V, which is not enough to drive the piezoelectric ceramic driver. The high frequency power amplifier is used to amplify the control signal and generate the driving voltage to drive the piezoelectric ceramics.
