Fault Diagnosis For A Class Of Complex Dynamical Networks With Unknown Parameters

Fault diagnosis for a class of output-coupling complex dynamical networks with unknown node parameters is studied in this dissertation. The online evolution of topology is monitored by designing a proper response network and controller. The main contents and contributions of the dissertation are as follows:Firstly, complex dynamical network models are reviewed briefly together with the methods of inner synchronization and outer synchronization. Fault diagnosis is analyzed and studied for output-coupling complex dynamical networks with known node parameters.Secondly, considering an output-coupling network with unknown node parameters, an approach of fault diagnosis is proposed based on outer synchronization theory, and then is extended to the networks with time-varying coupling delay. Output variables are used to construct controllers in response network. The conditions for choosing the controller parameters are also studied from Lyapunov stability theory for fault diagnosis of networks.Thirdly, considering an output-coupling network with unknown node parameters, an approach for fault diagnosis is proposed based on parameter identification. First, parameter identification for complex network is studied and two methods for parameter identification are proposed based on outer synchronization. Pinning control is applied to parameter identification. Then, the identified parameters are used to design response network directly, which can simplify the design of response network for fault diagnosis. Simulations are given to verify the effectiveness of the proposed approach.