Research On Fault Diagnosis Method Of Hybrid System Based On Energy Model For Parameter Uncertainty

With the increasing development of science and technology,industrial production control system is more and more complex level and precision,which contains different energy domain subsystems and sensors,actuators,digital circuits and software constitute the electromechanical system.If a part of the system fails,it will cause the system to run the entire error,or even collapse,resulting in serious economic losses or significant casualties.In order to prevent their proliferation and ensure safe and reliable operation of the system,these faults must be detected and isolated,timely and effectively.Based on the energy domain,this paper discusses the fault detection and isolation(FDI)of the hybrid system based on the knowledge theory of the bond graph.Firstly,the basic knowledge and theory of the bond graph are expounded,and the method of allocating the causal relationship between the components is analyzed emphatically.By using the bond graph model of the system,the state space of the system can be obtained by using the node characteristic equation and the element characteristic expression,and the circuit system is taken as an example to illustrate the realization of the process.Then,the fault is not isolated for the system with the same characteristics.It is proposed that the system can be used to generate a new system fault feature matrix based on the bicausal bond graph and the sensor union method to increase the number of isolable faults.And the proposed method is verified by the simulation model of double tanks system.Finally,considering the influence of various uncertain factors in the actual system,such as the uncertainty of the parameters,the uncertainty of the measurement and the uncertainty of the modeling and so on.The developed method based on the linear fractional transformation and the bicausal bond graph is combined to reconstruct the global analytical redundancy relations,and the uncertain part is estimated to obtain the adaptive threshold for the residual evaluation,which can reduce the false alarm and improve the reliability.And the effective method is verified in the double tanks system.