Fault Tolerant Control For Singularly Perturbed Systems Based On Controller Reconfiguration

Singularly perturbed system is a dynamic system with small parameters,which is characterized by multiple time scales.It covers many complex industrial systems,such as aerospace field,power electronics and complex mechanics.Since the existence of small perturbation parameters,the system is more complex.Because of the rapid development of computer intelligent technology and power electronics technology,the integration,complexity and sophistication of the system have been improved unprecedentedly.Therefore,it is essential to enhance the reliability and safety of singularly perturbed systems.In the actual industry,the common faults of the system mainly include actuator faults and sensor faults.Due to the randomness of the faults,it is impossible to establish a unified system model.Then,the research on fault tolerant control of complex systems has certain research significance.On account of this,the main contents of fault-tolerant control of singularly perturbed systems based on controller reconfiguration are discussed in this paper:（1）An active adaptive fault tolerant controller design method for a linear singularly perturbed system with actuator faults is presented.It is founded on the adaptive fault estimation and Youla controller parameterization.The controller can be divided into two parts by Youla parameter coprime decomposition.When the system is not disturbed or the system fails,the controllerK₀ can used to be ensure the constancy of the system.When the system fails,the controllerK_f works,which can guarantee the robustness of the system.The controllerK_f can be derived from the adaptive control theory.By designing the observer of singularly perturbed system,the actuator fault signal can be estimated accurately.At last,a numerical example is given to illustrate the feasibility and effectiveness of the method.（2）Fault tolerant control of singularly perturbed systems based on data driven is analyzed.First,the equivalent space is constructed according to the input and output data of the system,and then a multi-channel observer type residual generator is designed.An extended internal model controller（EIMC）based on residual generator is proposed to estimate the fault of the system.Finally,a numerical example shows the effectiveness and feasibility of the proposed method..（3）Fault tolerant control of singularly perturbed systems is put forward in this part,which is based on model predictive control.In the case of actuator failure,the model predictive control method is applied to design the controller.The prediction model is substituted,which contains the fault parameters.And then,the objective function is designed to achieve the purpose of rolling optimization.The system can function properly even though it suffer from failures.Finally,the correctness and feasibility of the conclusions are verified by numerical examples.