Research On Intelligent Control Of Dynamic Positioning System For Surface Ships

At present people need increasing for high quality vessels with dynamic positioning (DP) systems,with exploiting sea's resources into deepwater. It impels researchers to study and develop hi-new technique continuously,so that the demand is satisfied in civil and military fields. Using neural network to form system models,filtering has been studied in the paper,and neuro-fuzzy with genetic algorithms (GA) to establish adaptive controllers for DP systems has also been studied in the paper. Now the studying of ship operating performance has some reports using neural network,but the study is only a little on neural network applying to DP systems. The reports that the studying methods with my studying are similar or near has not been found at home and abroad up to now. The main researching works in the paper are summarized as follows:1. Modularization methods are used to establish the mathematical models and simulation program for ship's DP systems. It included ship's moving models of low frequency moving and high frequency moving,thruster's models and environmental mathematical models (wind,wave and current). The models have been proved to be correct by simulation method.2. Firstly,studied feed-forward neural network and put forward a new algorithm on BP network,called BP Algorithm based on Robust Error Function (BPAREF),and the algorithm is proved to be effective for approaching nonlinear system. Afterwards,studied how to form neuro-fuzzy,it is testified that the fuzzy neural network is a all-purpose approximation for nonlinear system.3. Studied the method of GA optimizing BPAREF to achieve a optimal BP net in structure and parameters,it is used to identify ship moving model. Studied self-organization competition neural network to optimize fuzzy rules of neuro-fuzzy,so that the fuzzy net with optimal rules is obtained.4. Studied filtering method using neural network,and established filters to deal with high frequency moving of ships.5. Studied system identification methods using neural network,and set up models for low and high frequency moving of ships,in the same time the adaptive controllers of fuzzy neural network based on GA is also established. The simulation results showed the controllers have strong robust,good adaptability for environmental changing,reduced power consuming and improved the overall performance of ship's DP systems.The three main methods (neural network,fuzzy logic system and GA) of intelligent control are studied in the paper,and they are proved to be reasonable and practical. The research results and conclusions in the paper are theoretical signification and practical value to develop new DP systems'ship for our country.