| Marine steam power plant is the key equipment which ensures the ship can complete thetasks such as transport and sail. The steam turbine has the high single power， highreliability， economically， light weight， small size and easily maintenance. So the steamturbine is the main form of marine power equipment. In order to promote the intelligent levelof marine power system control， the thesis relies on the actual scientific research for thetopic research background and is based on the soft computing and advanced control methods.To aim at the intelligent control of steam turbine rotate speed control system， carried out thefollowing studies:First of all, combine with the characteristics of the marine steam turbine, the workingmechanism of the ship steam turbine speed control system is analysed deeply. The modelingmethod is used to model the steam turbine speed system, and the model is simplified. Thenthe dynamic respons model is established with changed load. The simulation tesults showthat，while the load changed，the dynamic response of steam turbin speed system in keepingwith basic operation rules, the reasonableness and accuracy of the model are proved, thefoundation is established for the follow-up research of control method.Secondly, the work condition of marine steam turbine changes frequently, so theparameters time-varying. And thus now commonly used in traditional PID control can notmeet increasing control requirements. Therefore, combine the advantages of fuzzy theory andneural networks，so a kind of fuzzy neural network based on T-S model is established. And thelearning method is given. To avoid the problem of network dimension explosion, the roughset is used to simplify the network structure. A kind of fuzzy neural network controller basedon the rough set is designed. The simulation results show that this control strategy to achieveintelligent control of the ship power system.Again, the study of marine steam turbine model shows that, the model is nonlinear and itis changed with work condition change. So the control method should have the ability tocontrol well while the model changes. Predictive function control is a control method basedon the predictive model for fast process. So the LS-SVM is used to identify the predictivemodle online. Design method of predictive fuction controller bansed on the LS-SVM is given.The simulation results show that, this control strategy is effective for the marine steam turbinecontrol, the intelligent control is come ture while the work conditions change and loaddisturbance. Finally, the special work environment of marine turbine is in the ocean. While the shipsailing, there are a large number of uncertain disturbance come from the load and boiler. Sothe steam turbine control system should have strong robust. Therefore, aim at thischaracteristic of turbine, the robust control is added, and the multi-scale neural network willalso be included to ensure the minimum size of the network, and the amount of computationis reduced. The simulation results show that，the control method can inhibit the disturbanceeffectively, and have the robust.In short, aim at the control difficulty for the marine steam turbine system, on the basis ofits dynamic model study, three kind of marine steam turbine controller designed, thesimulation results show that the three methods are able to achieve the intelligent control of themarine turbine system, and the difficulty control problems are solved. |
