| Aerospace facilities normally work in the vibration-centrifuge compound environment. Limited by the experimental facility, vibration or centrifugal force are supplied asynchronously to test the experimental object. But the potential reliability of the objects in the compound environment can not be forecast accurately by the asynchronous experimental results. So the increasing significance is attached to the vibration-centrifuge compound environment experiment.Working under the vibration-centrifuge compound environment, the components of the vibrator are unbalanced because of the centrifugal force, which prevents the vibrator from working properly. So the active control on the vibrator is necessary. The stiffness and the dump of the vibrator change with the rotation speed, whose parameters alter sharply. Obviously the traditional control methods are hardly to work. The computational intelligence may be a good choice.On the basis of the discrete state equation, which is got by the means of governing motion equation of the vibrator, the genetic algorithms is applied to control the acceleration of the countertop under the sine excitation. Furthermore, a new active control method is proposed, which uses the genetic algorithms and neural network system. In the active control system, the neural network is used as the emulator to simulate the movement of the vibrator; the genetic algorithms is applied to determinethe optimal control force. The numerical examples prove the effectiveness of the proposed methods.The dominant motion equations are obtained on the basis of the hypotheses that the centrifuge arm is rigid. The imprecise conclusion of the vibrator motion may be induced. In this paper, take a further step to discuss the governing motion equations supposing the centrifuge arm is soft.This paper provides with the foundation to solve the self-adaptation inverse control of motion of the vibrator under the stochastic excitation. |
PDF Full Text Request