Intelligent Recognition Of Low-pressure Turbine Shaft Connection Status Based On Vibration Response

The aero engine is known as "the crown jewel of industry".It is a complex high-precision equipment composed of 30,000-40,000 parts and 2000-3000 bolts.It has extremely high requirements for assembly accuracy and dynamic stability.The research object of this paper is the low-pressure turbine shaft,which is the core component of the aero engine to transmit power.It is assembled by the shaft disk and the cone wall through 36 sets of bolts,which is a typical multibolt connection flange structure.Because of long-term service in a high-temperature,high-pressure,high-speed working environment,its connection state directly affects the dynamic characteristics of an aircraft during service.Therefore,using effective methods to monitor the connection status of low-pressure turbine shafts,and achieving accurate assessment of assembly accuracy and service performance is particularly critical for improving the stability and reliability of turbine shafts.In this paper,the low-pressure turbine shaft is taken as the research object,and the intelligent identification model of the connection state is designed based on its vibration energy transmission characteristics.Use the multi-bolt structure connection state intelligent identification model to identify the position of the bolt at the joint,and the single-bolt structure connection state intelligent identification model to identify the specific decrease in the bolt pretension of the loose position,and finally achieve the connection status of each position on the joint intelligent recognition.The main research contents of this article are as follows:First,in Chapter 2,the dispersive characteristics of the preload force caused by the working load and elastic interaction of the bolted structure are analyzed;through the finite element simulation and vibration power flow theory,the relationship between the connection state of the joint and the vibration energy transfer characteristics of the structure is analyzed.On this basis,the feasibility of the method of "identifying the connection state of the joint based on the structural vibration energy transfer characteristics" is verified.Secondly,in Chapter 3,based on the vibration power flow transmission characteristics of the single-bolt connection structure,a single-bolt connection beam experiment was designed to obtain the vibration response signals of the structure under different connection states.The wavelet packet decomposition and frequency domain analysis method are used to process the experimental data and extract the signal features.Then,the sensitive features are selected as the input vector according to the distance evaluation theory.Finally,the KNN pattern recognition algorithm is used to train to obtain the single bolt structure connection state intelligent recognition model,and then through the vibration signal successfully identified the connection state of the single-bolt structure under different preloads.Then,in Chapter 4,a multi-bolt connection beam experiment was designed to collect vibration signals.Due to the elastic interaction between the bolts,the connection state of the joint and the vibration energy transmission process were more complicated.Therefore,EEMD was used to decompose the vibration signal to avoid mode aliasing.Then,combined with singular value decomposition,the main component analysis of the signal is performed,and the singular value entropy is used as the feature vector input SVM to train the classifier model.Finally,an intelligent recognition model for the connection state of the multi-bolt structure is obtained,and the connection state of the junction area is successfully identified.Finally,in Chapter 5,the low-pressure turbine shaft is taken as the research object,which is subjected to finite element simulation and vibration power flow energy transfer characteristic analysis.And according to the analysis results,the experiment is designed and the vibration signal measurement points are arranged.Through the technical route of Chapter 4,the intelligent identification model of the loosening position of the multi-bolt of the low-pressure turbine shaft is established;through the technical route of Chapter 3,the intelligent identification model of the connection state of the single-bolt pretension of the low-pressure turbine shaft is established.The two intelligent identification models cooperate with the test scheme of evenly distributed signal acquisition points on the shaft disk,which can quickly and accurately identify the connection status of various positions on the food section of the lowpressure turbine.