Research On Typical Fault Diagnosis And Location Method Of Aero-engine Rotor Based On Parameter Identification

The aero-engine is a complex system with high faults rate,complex adjustment,and large maintenance workload,Accordingly,it is important to have an effective fault diagnosis method to ensure its safety and reliability.However,due to the complex structure of the aero-engine,the traditional fault diagnosis methods are difficult to diagnose and locate the fault effectively.New methods based on big data and artificial intelligence rely on a large amount of fault data and experience,which is difficult for practical application.Considering the actual measurement conditions of the aero-engine,this paper uses the signal of the aero-engine casing,and proposes a method for diagnosing and locating the typical faults of the aero-engine rotor based on the faults feature parameter identification.The main contributions of the dissertation are as follows:Aiming at the actual aero-engine structure,we employed finite element method to establish a dual-rotor-support-casing dynamic model.We developed the dynamic model for the typical single rotor faults(unbalance,misalignment,crack,rubbing,bending,loosening)and compound faults.Based on the model,we further explored the dynamic response of the aero-engine with rotor faults.We applied the EKF algorithm to diagnose and locate the typical faults in the aero-engine rotor.In order to improve the convergence and stability of parameter identification,we introduced a weighted global iteration and attenuated memory filter algorithm to diagnose and locate the rotor faults based on fault feature parameter identification.By means of the identification of multiple fault parameters including unbalance,misalignment,crack and bending faults,the results validated the effectiveness of the method in fault diagnosis and location.Due to the accuracy and complex linearization of strong nonlinear systems of EKF,we introduced the UKF algorithm.And then we applied SRUKF to prevent the parametric failures during parameter estimation of UKF.With the identification of multiple fault parameters such as misalignment,rubbing,crack faults and complex faults,the results demonstrated that SRUKF has a higher accuracy of parameter estimation in strong nonlinear systems compared with EKF.Owing to the limitations of experimental conditions,aero-engine test data cannot be obtained.We performed the experiments including unbalance,misalignment,rubbing and crack faults on Bently rotor test stand,the results verified the effectiveness of finite element modeling methods as well as fault diagnosis and location methods based on EKF filtering.Based on C#,we developed a software for aero-engine condition monitoring and fault diagnosis software.Moreover,based on Sql Server we developed a database filling automatic processing script and designed a database of aero-engine vibration database.The system has the status monitoring module,the signal analysis module,the blind source separation module and the fault diagnosis module.