Application Of Back Propagation Neural Network PID Control In Dynamic Positioning System

The ocean is rich in resources, the pace of the development and utilization of marine resources speed up step by step. In the continuously explore to the marine resources, marine engineering facilities had been applied more widely. In the field of deep sea engineering dynamic positioning system is the key technology to remain marine engineering facilities location. Dynamic positioning system consists of three systems, namely the position measurement system, control system and power propulsion system. control system is the core of the whole dynamic positioning system, this paper is mainly about the control system of dynamic positioning system.The research of ship dynamic positioning system is based on the specific marine environment, so in this article ship motion model is established as well as the basic model of the marine environment. Simplify the degrees of freedom of ship motion, the ship model is established on three degrees of freedom namely the sway, surge and yaw. The model of marine environment, which mainly include wind, wave and flow model is also established. To test and verify the accuracy of ship model, based on the MATLAB/Simulink simulation environment it is verified and confirmed that the model is accurate.To control dynamic positioning system, this article use the method which is based on the classical PID control, and adjust the PID parameters online by using neural network control to design the control system. PID control has become a mature and reliable control method after a longtime practice; Neural network has a powerful information processing ability and it has a great advantage in dealing with nonlinear problem. Combining PID and neural network, enhance the adaptivity of the controller for interference. Neural network control can be used in a variety of algorithms, this paper uses relatively mature BP algorithm, it is one of the most widely applied in neural network algorithm, mainly contains two process, namely the information spread forward and the error is back propagation. Combine the BP neural network law with incremental PID control law, and adjust the PID parameters online by the BP neural network control law. To verify the feasibility and stability of the control law, applying the MATLAB/Simulink, the designed controller is verified. Using the SQP method to optimize ship thrust allocation. Carry on the simulation of the whole system, the running results show that the controller can achieve better control of the system.