The Neural Network Control Of The Magnetic Floating North Finder

This dissertation is based on the high-precise magnetic floating north finder, which is newly designed by the Center for Control Theory and Guidance Technology of Harbin Institute of Technology. The magnetic floating north finder is a new style system used to find north which is based on the magnetic floating bearing. Compared with the traditional north finding system, it realizes that the gyroscope's framework is suspended by magnetic force, so the friction is reduced to the lowest that greatly improving the speed and the precision of finding north. Magnetic floating bearing is a kind of new high-performance bearing in which rotor can be suspended by magnetic force. Compared with traditional ones, it has many advantages of long lifetime, low power loss and no lubrication. Therefore it can operate in various especial environments and has extensive applications in industrial field.Based on the introduction on the structure, working principle and the design of hardware in the part of electrical control of magnetic floating north finder, we introduce the structure of rotor and electromagnet, deduce the formula used to calculate the electromagnetic power of magnetic floating bearing, get the transfer function and state space equation of single-freedom-degree magnetic floating bearing, erect the model of the magnetic floating bearing north finder.It is so difficult to study the control of the magnetic floating system with the characteristic of open-loop unsteadiness and strongly nonlinear. At the same time, the performance of the controller directly impacts the dynamic performance of the north finder. So, the controller of neural network combined with PID method is designed utilizing the characteristic of associational memory and self-learning. Therefore the performance and robustness of whole system is improved greatly. At last we give the simulation and experiment of whole system.