Research On Electrostatic Monitoring Technology For Gas-path Component Health Management

Aero engine gas path components belong to the core components of the engine,and the working conditions are very bad,so the monitoring of gas path components is particularly important,and it is one of the core contents of engine health management technology.In recent years,with the development of the state monitoring technology,the research on the static monitoring of aero engine gas path is gradually heating up.In this paper,the health management for aero engine gas path electrostatic monitoring technology research as the theme,carry out the electrostatic sensor model inference and verification,based on sparse decomposition denoising method,electrostatic signal engine gas path fault simulation experiment based on machine learning method of fault pattern recognition,turbofan engine gas path based on static state recognition,anomaly the data of gas path performance evaluation,gas path electrostatic monitoring system software and hardware implementation and so on.The main work and achievements are as follows:(1)Summarizes and analyzes the research status of the existing electrostatic monitoring of aero-engine,firstly analyzes the source of normal particles on the engine gas path components and abnormal charged particles,the gas charged particles damage and abnormal charged mechanism;the engine gas path electrostatic monitoring principle and related hardware and software the architecture is analyzed,based on the existing inlet annular electrostatic sensor and nozzle rod electrostatic sensor theory analysis object,combined with the actual needs of research on mathematical deduction for sensing model of two typical electrostatic sensor,and the titration bench on electrostatic sensor model verification experiment,obtained the electrostatic signals affect the relationship between factors and output the signal,through the accurate modeling of the sensing model provides a solid rationale for the subsequent in-depth study on the sensing signal On the basis and experimental support.(2)Aiming at the problem of noise electrostatic signal,the noise type and source analysis,summed up the shortcomings of previous denoising methods in different disturbance types,a signal denoising method based on sparse decomposition,the simulation signal and measured test electrostatic signal to verify the proposed method,and other classical denoising methods and compares denoising effect,ensure the denoising process flexibility and efficiency of the electrostatic signal.(3)In order to solve the problem of identifying typical gas path failure modes,four typical simulation experiments of gas path mechanical faults are put forward,and relevant gas path fault simulation experiments are carried out.According to the classic electrostatic signal time-frequency domain statistical feature extraction method,feature extraction and comparison analysis of sample data obtained by the experiment of four kinds of fault simulation,characteristics of failure modes in four typical gas path under the respective electrostatic signal characteristic parameters and differences in analysis.Simulation of experimental data obtained by the failure mode,further analysis of characteristic parameters in the classical results put forward several new characteristic index of electrostatic signal,the mode of sample extraction of new features,analysis of the new characteristics and put forward the initial fault logic theory.To verify the validity of the proposed method using the Fisher criterion characteristic indexes,this paper proposes a fault pattern recognition algorithm based on SOM neural network and the fault pattern recognition algorithm is verified,provides a better intelligent recognition method for gas path fault pattern recognition.(4)In order to verify the results of the practical application of electrostatic sensor,to carry out the actual test of turbofan engine static monitoring experiment,introduces the experimental test used in the process of electrostatic sensor and the experimental setup,and the electrostatic monitoring data,combined with the related performance parameters of the engine,to evaluate the effect of electrostatic monitoring,baseline the model and the establishment of a turbofan engine AL parameters.The abnormal state of engine gas path is detected by using permutation entropy and wavelet energy spectrum entropy respectively.Put forward a comprehensive assessment of fusion turbojet engine performance parameters of electrostatic monitoring data and health status on the engine gas path performance,analysis results show that the evaluation method of performance evaluation method of fusion electrostatic features and performance parameters is superior to the traditional,can win the maintenance warning window for maintenance personnel.(5)To solve the practical problem of subsequent electrostatic monitoring technology,design of a LABVIEW based graphic language of aero-engine monitoring and diagnosis software system and based on a set of portable static data acquisition and processing hardware system of embedded technology.In addition,based on the real-time main system,more signal processing methods are fused and machine learning is introduced to realize off-line diagnosis.The design of embedded hardware detection system based on,including data acquisition module and terminal based on WinCE 6 operating system through the touch screen display module,friendly man-machine interface,real-time display,signal acquisition and data storage and fault diagnosis function.