| Summary: | 碩士 === 國立雲林科技大學 === 電機工程系碩士班 === 100 === The high speed and high accuracy requirements are required in recent year, so that, the high accuracy and excellent tracking performance servo position system is played important roles in industry development. The magnetic force is used at magnetic ball in magnetic ball suspension system which avoids the contact friction, noise and vibrating from the machine elements. However, this system is a nonlinear and open loop unstable system. In classical control, the controller is designed by using the operating point linearization method. It is suitable in a small region of operating point and sensitive to the parameter variations and external disturbances.
In this thesis, we build a mathematical model of a magnetic ball suspension system. The integral backstepping control is proposed in first part. The fuzzy estimator is designed to estimate the lumped uncertainties in real time. The genetic algorithm is used to search the center and width of Gauss function in the neural network by offline which promotes the convergence ability of proposed controller in second part. The stability of all control algorithms is approved by Lyapunov stability theory. Finally, the proposed controllers are tested through experiments to sure the estimator can be validated to estimate the lumped uncertainties that makes system eliminating steady state error and tracking the command successfully.
|
|---|
