Research On Fault Diagnosis Technology Of Military Aero-engine Based On Neural Network And Manifold Learning

With the development of science and technology,the reliability requirements of military aircraft engines is increasing.However,because the functions are expanding and the conditions of service are more difficult,the structure of military aircraft engine mechanical parts become more complex,the probability of failure is also getting higher and higher.The causes of military aircraft engine failure are complex,they are affected by many factors such as human error,material defects,changes of using environmental,and fatigue,wear and aging effects.This makes the military aircraft engine fault diagnosis and analysis more and more difficult,needing more aviation maintenance and protection work and a lot of manpower and material resources to guarantee the reliability of the aircraft.Although the traditional fault diagnosis technology can diagnose fault,but the diagnosis of efficiency and accuracy are not high.In this paper,due to the urgent need of military aeroengine fault diagnosis,we apply data mining theory and technology and apply the neural network and manifold learning algorithm to the diagnosis of military aeroengine.This paper introduces the current situation of fault diagnosis technology and data mining technology at home and abroad,and the background knowledge of military aircraft engine rolling bearing fault diagnosis,including the basic vibration mechanism,the common fault types and the traditional diagnostic methods,and then explains the rolling bearing vibration signal characteristics in detail.The PNN and SOM,which are commonly used in the field of fault diagnosis,are explained and compared in detail.The experimental results show that the SOM classifier is more effective in fault diagnosis applications.This paper studies of manifold learning theory and methods,analyzes the common three manifold learning algorithms: ISOMAP,LLE and LE.We compare the traditional linear dimensionality reduction method PCA and the nonlinear dimensionality reduction method LE,propose the improvement scheme of the LE algorithm,use the improved distance function to replace the original European distance function,which can better reflect the local structure information of the high dimensional data set and improve the performance of the LE algorithm.This paper mainly studies the fault diagnosis of military aviation engine rolling bearing.Rolling bearing is one of the core components of a military aero engine,in which it plays a role in bearing and delivering loads,and its operating state has a great influence on the working state of the engine.And once the rolling bearing failures,it will make the vibration of engine rotor increasing,more seriously it will make engine damaged and serious flight accidents.Therefore,the state of the aircraft engine rolling bearing status monitoring and fault diagnosis is essential to ensure its normal work on the engine and even the aircraft's safe and reliable work.The wavelet signal is used to denoise the vibration signal.This paper uses the time domain and frequency domain analysis method to extract the characteristic attribute of the vibration signal,and constructs the fault sample set.In this paper,a new model of fault diagnosis based on SOM and improved LE algorithm is proposed,and the general steps of fault diagnosis of military aeroengine rolling bearings are set up and used in the field of military aeroengine fault diagnosis for the first time.The experimental results show that this model can effectively improve the efficiency and accuracy of fault diagnosis of military aircraft engine rolling bearing.