Research On Vibration Control Of Piezoelectric Cantilever Based On Bp Neural Network

Many modern engineering and mechanical operations involve unnecessary structural vibration. This kind of structural vibration will accelerate the fatigue and affect the normal work of the machine. To simplify the analysis model, this paper utilizes a cantilever beam structure to replace the complex mechanical structure for vibration analysis. Piezoelectric ceramic materials are used to excite and sense the vibration of the cantilever structure, and neural network algorithm is used to actively control the vibration. First, the neural network algorithm is used to identify the cantilever structure model offline, making output voltage of the network close to the actual target voltage value. When the required identification accuracy is reached, the identification is stopped. Then the identified model parameters are provided to the controller; the neural network in the controller calculates the real-time feedback signal, which is then loaded to the corresponding piezoelectric element for vibration suppression. The results indicate that the method can effectively realize the active vibration control of the cantilever beam.