The anti-rolling tank swing bench is a typical hydraulic position servo system. Its function is to simulate the ship roll motion and to verify the performance of the anti-rolling tank. Generally, the PID control method can achieve an ideal control effect. But the uncertainties, such as the couplings and the nonlinearty of the swing bench, make it difficult to establish the precise mathematical model of system. During experiments, the different disturbance torques are generated according to the different dimension scales of the anti-rolling tanks. Changes of the natural characteristics of the system are caused by changes of load, and the fixed parameters of PID controller are difficult to continue to ensure a good control effect. In order to reduce the influences caused by the factors mentioned above, a BP neural network PID control system based on virtual instrument LabVIEW is designed. By simulating the responses of the system in real time control, the verification tests using actual system are avoided. Finally, the experiments of swing bench are carried out. The results show that the controller can ensure the control accuracy and stability. The parameters tuning process efficiency is improved. The operation is convenient. The data acquisition process is simple and practical.
Read full abstract