Font Size: a A A

The Control Theory Research Of Neural Network For The Condenser Cleaning Robot

Posted on:2013-05-31Degree:DoctorType:Dissertation
Country:ChinaCandidate:X R WuFull Text:PDF
GTID:1228330374491620Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
Condenser is a large heat transfer equipment in the thermal powerplant、chemical、machinery and other industries. It is the cold source of the ther-modynamic cycle of the large steam turbine. Good or bad performance of its work willdirectly affect the economics and safety of the entire steam turbine. For the chemicalreactions and unclear cooling water during heat exchange when the condenser is running,the fouling, which is not favorable for heat transfer, is accumulated in the inner wall ofcondenser tube. These fouling have produced such harm: reduces the efficiency of thesteam turbine, increases the cost of electricity and even leads to accidents because of theblockage and corrosion perforation of condenser tube. To solve the problem, this paperdesigns an intelligent condenser-cleaning mobile robot for large condenser in order tothe purpose of improving the heat transfer performance of condenser and increasing theefficiency of thermal cycle of turbo-generator unit. Thus, this dissertation focuses on themethods of neural network intelligent robust control for the condenser-cleaning mobilerobot. Main results and contributions of this dissertation can be list as the following sevenchapters.In the first chapter, a review on the theoretical research and practical significance ofthe dissertation is presented, the history and basic principles of the condenser-cleaningrobot, the mobile robot, the neural network control are described. The paper introduces themethods of the robust adaptive control, the fuzzy control, the intelligent control and thesliding-mode control. Finally, the structure arrangement of the dissertation is given.In the second chapter, the paper presents the speculative knowledge about the fuzzycontrol for a class of the delayed neural networks system. Based on the linear matrix in-equality(LMI), a robust fuzzy control to guarantee the global robust asymptotical stabilityof fuzzy neural networks with discontinuous activation functions is designed. Comparedwith the existing literature, this paper remove the assumptions on the neuron activationssuch as Lipschitz conditions, bounded, monotonic increasing property or the right limitvalue is bigger than the left one at the discontinuous point. Thus, the results of this pa-per are more general and widest. Finally, two numerical examples are given to show theeffectiveness and feasibility of the proposed fuzzy controller.In the third chapter, the paper presents the radial basis functions neural network-basedadaptive robust control(RBFNNARC) for condenser-cleaning mobile robot. First, a dy- namic modeling is obtained based on the practical condenser-cleaning mobile manipulatorsystem. Second, the RBF neural network is used to identify the unstructured system dy-namics directly due to its very good compensation nonlinear function ability. Using learn-ing ability of neural network, RBFNNARC can coordinately control the mobile platformand the mounted manipulator with different dynamics efficiently. The implementation ofthe control algorithm is dependent on the sliding mode control (SMC). Finally, based onthe Lyapunov stability theory, the stability of the whole control system, the boundednessof the neural network weight estimation errors, and the uniformly ultimately boundednessof the tracking error are all strictly guaranteed. Moreover, simulation and experiment aregiven to demonstrate that the proposed RBFNNARC approach can guarantee the wholesystem’s converges to the desired manifold with prescribed performance.In the fourth chapter, the paper studies the neural network fuzzy control for thecondenser-cleaning mobile robot. A neural network fuzzy controller for the condenser-cleaning mobile robot is designed using the fuzzy wavelet neural network and the robustadaptive control method. The fuzzy wavelet neural network is used to identify the uncer-tainties and disturbances of the robot system due to its very good compensation nonlinearfunction ability. Based on the Lyapunov stability theory, the stability of the condenser-cleaning mobile robot system is guaranteed. Finally, simulation is give to demonstrate theeffectiveness and feasibility of the proposed controller.In the fifth chapter, the paper researches the sliding mode control problem of a classof3-DOF condenser-cleaning mobile robot system with uncertainties and disturbances.First, the kinematic modeling of the mobile platform and the three joint manipulators aregiven. Then, the paper presents the dynamic modeling of the three joint mobile robotsystem. Second, the fuzzy wavelet neural network is used to identify the unstructuredsystem dynamics directly due to its very good compensation nonlinear function ability.Based on the Lyapunov stability theory, the paper designs a robust adaptive sliding modecontroller to guarantee the global robust asymptotical stability and the uniformly ultimatelyboundedness of the tracking error of the mobile robot system. Finally, simulation is giveto demonstrate the effectiveness and feasibility of the proposed controller.In the sixth chapter, the paper studies the intelligent neural networks control problemfor condenser-cleaning mobile robot. The RBF neural network is used to identify the localuncertainties and disturbances of the condenser-cleaning mobile robot system due to itsvery good compensation nonlinear function ability. A neural network controller for thecondenser-cleaning mobile robot is designed using the Lyapunov stability theory. Finally,simulation and experiment are give to demonstrate the robustness and the perfect tracking performance of the proposed controller.In the seventh chapter, a dynamic modeling is obtained based on the practicalcondenser-cleaning manipulator system. The fundamental structure and approximabil-ity of the fuzzy Gaussian function neural network are introduced. The fuzzy Gaussianfunction neural network is used to identify the unstructured system dynamics directly ofthe condenser-cleaning manipulator system due to its very good compensation nonlinearfunction ability. The paper designs an adaptive fuzzy controller for the condenser-cleaningmobile robot. Based on the Lyapunov stability theory, the stability theorem of the wholecontrol system is guaranteed. Moreover, simulation experiment is given to demonstrate therobustness and the perfect tracking performance of the proposed adaptive fuzzy controller.
Keywords/Search Tags:Neural network, Adaptive control, Fuzzy control, Mobile robot, Con-denser, Sliding-mode control, Robust, Intelligent control
PDF Full Text Request
Related items