| According to the characteristics of cantilever beam with smart materials structures,based on the finite elements analysis,a mathematic model of controlling system is established.Making use of the finite elements analysis software-ANSYS, then the natural frequencies and natural mode shapes are gotten so as to establish dynamics functional analysis for following controllers.Through the dynamic analysis of structural vibration,combined with the characteristics of fuzzy control technology,the displacement and speed are identified as the input of fuzzy control,and control power imposed structure as output,fuzzy inference neural network controller is designed,that is the basic theory of fuzzy control as core and the neural network as the carrier.The use of neural network can train self-learning ability and memory of the experience of knowledge.Therefore mamdani inference method can be strengthened.BP algorithm through fuzzy control rules can be studied by himself,adjusting the connection weights,which can control algorithm to generate the rule base on the structure of the dynamic fuzzy control.Meanwhile,during the signai adoption and processing aspects,because of the complexity of the environment,measurement of signal interfere always carry random disturbance,the thesis designed noise cancellation filter based on A MATLAB adaptive neural-fuzzy inference system.Simulation results shows that this control system can eliminate noise in complex environment so as to provide a clear controller vibration data signals and ensure the reliability of control.Due to its limitations of low study efficiency,slow absorption speed and thesis to emerge parts,enlightened by the idea of self-adjusted,the particle swarm optimization algorithm optimizes factor。Adopting these neutral network and PSO analysis,it has achieved evident control results,the successful simulation of flexible structures and non-linear structure vibration have brought a practical significance.
|
PDF Full Text Request