Weak Fault Feature Extraction and Intelligent Fault Diagnosis Method for Key Components of Rail Transit Equipment

Abstract

The present invention relates to the technical field of component fault diagnosis. Specifically, it relates to a method for extracting weak fault characteristics and intelligent fault diagnosis of key components of rail transit equipment, which includes the following steps: calculating the vehicle force parameters corresponding to the track curvature; using the dynamic modeling method and introducing the impact force of raindrops and hailstones on the components and the additional viscous resistance of snow accumulation, calculating the coupling relationship between the two components; using the variational mode decomposition method to decompose the coupled fault feature signal into multiple modal components, analyzing the track curvature modulation feature spectrum of each modal component, and calculating the Lyapunov exponent of the modal component; calculating the Lyapunov exponents of all modal components to determine whether the two components are faulty. The method for extracting weak fault characteristics and intelligent fault diagnosis of key components of rail transit equipment combines the modal coupling analysis and signal synthesis of the two components under rainy and snowy weather conditions to improve the detection of weak fault signals in the wheel pair bearing-gear system.

Description

Technical Field

[0001] The present invention relates to the technical field of component fault diagnosis, and more specifically, to a method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment. Background Art

[0002] The method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment aims to improve the early fault detection ability of key components of rail transit equipment and enhance the fault recognition accuracy under complex working conditions and rain and snow weather. By considering the influencing factors of rain and snow weather through dynamic modeling methods, a dual-component coupling dynamic model is constructed. Combining the track curvature modulation feature spectrum and intelligent classification algorithms, it controls the loss and misjudgment of fault signals caused by factors such as track curvature, load changes, and dynamic impacts, and realizes the accurate identification of weak faults under different operating states, improving the operating safety and maintenance efficiency of rail transit equipment.

[0003] Existing component fault diagnosis methods usually have difficulty in accurately extracting early fault signals of wheel pair bearings and gear transmission systems. Moreover, due to the change in track curvature, the impact force of raindrops and hailstones on parts during rain and snow weather, and the additional resistance of snow adhesion, faults such as fatigue spalling of wheel pair bearings and gear pitting are difficult to detect on straight sections and in good weather conditions, but only appear on curved sections or under specific weather conditions. Therefore, a method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment is provided. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment to solve the problem mentioned in the above background art that due to the change in track curvature, uneven load, and environmental noise interference, faults such as fatigue spalling of wheel pair bearings and gear pitting are difficult to detect on straight sections, but only appear on curved sections or under specific loads.

[0005] To achieve the above purpose, the present invention aims to provide a method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment, including the following steps:

[0006] S1. Use track measurement equipment to collect track geometric data and vehicle operation data, calculate the vehicle force parameters corresponding to the track curvature, and establish a force relationship matrix of vehicle components;

[0007] S2. Use sensors to collect the component state data in real time during vehicle operation, calculate the coupling relationship between dual components by using dynamic modeling methods, and use signal synthesis methods to superimpose the vibration signals of dual components to generate coupled fault feature signals;

[0008] S3. Use the variational mode decomposition method to decompose the coupled fault feature signal into multiple modal components, analyze and calculate the orbital curvature modulation characteristic spectra of each modal component under different orbital curvature conditions, calculate the Lyapunov exponents of the modal components, analyze the dynamic stability of the modal components under different orbital curvatures and operating states, and store the analyzed orbital curvature modulation characteristic spectra and Lyapunov exponents;

[0009] S4. Calculate the Lyapunov exponents of all modal components to determine whether the two components are faulty.

[0010] As a further improvement of this technical solution, the track geometric data includes track curvature, track gradient, track superelevation, and track irregularity;

[0011] The vehicle operation data includes operating speed, acceleration, axle load, wheel-rail contact force, and suspension system parameters;

[0012] In S1, calculate the vehicle force parameters corresponding to the track curvature, and establish a force relationship matrix for vehicle components. The specific method steps are as follows:

[0013] S1.1. Establish a curvature calculation model based on the track centerline equation:

[0014] ;

[0015] where, is the track coordinate; is the curvature at the track coordinate 𝑠; is the track tangent angle; is the arc length increment on the track;

[0016] S1.2. Based on the curvature calculation model, calculate the vehicle force parameters, including normal force, lateral force, and component forces;

[0017] Among them, the component forces are the bearing force and the gear force;

[0018] ;

[0019] ;

[0020] where, is the centrifugal force caused by the track curvature; is the vehicle mass; is the vehicle speed; is the total normal force of the vehicle; is the acceleration due to gravity; is the track gradient;

[0021] ;

[0022] Among them, is the lateral force caused by the track curvature; is the track superelevation; is the track curvature radius;

[0023] ;

[0024] ;

[0025] ;

[0026] Among them, is the contact force between the wheel and the track; is the wheel-rail contact area; is the bearing force; is the gear force; is the transmission torque; is the gear radius;

[0027] S1.3. Establish the force relationship matrix of vehicle components:

[0028] ;

[0029] Among them, is the force relationship matrix of vehicle components.

[0030] As a further improvement of this technical solution, in the above S2, the component state data includes vibration, stress, temperature, and current;

[0031] In the above S2, the dynamic modeling method is used to calculate the coupling relationship between two components, and the signal synthesis method is used to superimpose the vibration signals of the two components to generate a coupled fault characteristic signal. The specific method steps are as follows:

[0032] S2.1. Based on the force relationship matrix of vehicle components, construct a coupled dynamic equation for two components and calculate the modal coupling relationship between the two components;

[0033] S2.2. Based on the modal coupling relationship between two components, determine whether the two components can perform signal synthesis, and extract the fault impact characteristics and construct a coupled fault characteristic signal.

[0034] As a further improvement of this technical solution, in the above S2.1, based on the force relationship matrix of vehicle components, construct a coupled dynamic equation for two components and calculate the modal coupling relationship between the two components. The specific method is as follows:

[0035] S2.1.1. Based on the force relationship matrix of vehicle components, incorporate the impact force of raindrops and hailstones, the additional resistance of snow, and the track humidity into the state relationship of the double components, and construct the dynamic equation of the modal coupling relationship of the double components:

[0036] ;

[0037] Among them, is the mass matrix; is the damping matrix; is the stiffness matrix; is the displacement vector; is the velocity vector; is the acceleration vector; is the damping correction term affected by humidity; is the damping correction term for raindrop and hail impact; is the stiffness correction term affected by track slipperiness; is the raindrop and hail impact force; is the influence matrix of the raindrop and hail impact force; is the additional resistance of snow accumulation; is the influence matrix of the additional resistance of snow accumulation;

[0038] ;

[0039] ;

[0040] ;

[0041] ;

[0042] ;

[0043] Among them, is the humidity influence coefficient; is the humidity; is the impact influence coefficient; is the impact frequency of raindrops and hailstones on the components; is the raindrop and hail impact force; is the correction coefficient; is the wheel-rail friction coefficient under the slippery state; is the stiffness matrix; is the raindrop and hail mass; is the impact velocity; is the snow accumulation density; is the snow accumulation contact area;

[0044] S2.1.2. Based on the dynamic equation of the modal coupling relationship of the double components, calculate the modal coupling relationship of the double components:

[0045] ;

[0046] Among them, is the component index; is the component index; is the component and the component the coupling degree between; is the time; is the component at time the vibration displacement signal; is the component at time the vibration displacement signal; is the component and the component the signal cross-integration between; is the component the energy calculation of the vibration signal; is the component the energy calculation of the vibration signal. As a further improvement of this technical solution, in S2.2, based on the dual-component modal coupling relationship, it is judged whether the two components can perform signal synthesis, and the fault impact characteristics are extracted and the coupled fault characteristic signal is constructed. The specific method is as follows:

[0047] If the coupling degree between the component and the component , then the signals of the two components are strongly correlated and signal synthesis can be performed:

[0048] ;

[0049] Among them, is the dual-component synthesized signal; is the vibration signal of the component ; is the vibration signal of the component ; is the weight coefficient of the component ; is the weight coefficient of the component ;

[0050] Adaptive filtering is used to remove the noise of the vibration signals of the component and the component , and the cross-correlation function of the vibration signals of the component and the component is calculated:

[0051] ;

[0052] Among them, is the time delay; is the component and the component cross-correlation function of the vibration signals;

[0053] Calculate the instantaneous energy and envelope signal of the combined signal of two components :

[0054] ;

[0055] ;

[0056] Among them, is the instantaneous energy; is the envelope signal; is the Hilbert transform;

[0057] Construct the coupled fault feature signal:

[0058] ;

[0059] Among them, is the coupled fault feature signal.

[0060] As a further improvement of this technical solution, in S3, the track curvature modulation feature spectrum is a set of features that describe the energy distribution, frequency characteristics, and nonlinear dynamic behavior of the coupled fault signal under different track curvature conditions, and is used to analyze how the weak fault signal is affected by the track curvature and extract the fault features that appear with the change of curvature;

[0061] In S3, the variational mode decomposition method is used to decompose the coupled fault feature signal into multiple modal components, analyze and calculate the track curvature modulation feature spectra of each modal component under different track curvature conditions, calculate the Lyapunov exponent of the modal component, analyze the dynamic stability of the modal component under different track curvatures and operating states, and store the analyzed track curvature modulation feature spectra and Lyapunov exponents. The specific method steps are as follows:

[0062] S3.1. Use variational mode decomposition for the combined signal of two components , construct the optimization problem of two components, and use the alternating direction multiplier method to solve the optimization problem of two components to obtain multiple modal components of two components;

[0063] S3.2. Based on the track curvature and modal components, calculate the instantaneous energy distribution, time-frequency energy density change, power spectrum entropy, and track curvature correlation of the modal components of two components;

[0064] S3.3. Calculate the Lyapunov exponent based on the dual-component modal components to determine whether the dual-components are stable and fault-free under the change of track curvature;

[0065] S3.4. Store the analyzed track curvature modulation feature spectrum and Lyapunov exponent.

[0066] As a further improvement of this technical solution, in the above S3.1, the variational mode decomposition is used for the dual-component composite signal , construct the dual-component optimization problem, and use the alternating direction multiplier method to solve the dual-component optimization problem to obtain multiple dual-component modal components. The specific method steps are as follows:

[0067] ;

[0068] Among them, is the index of the dual-component modal component; is the center frequency of the th dual-component modal component; is the unit impulse function; is the imaginary unit; is the partial derivative calculation with respect to time 𝑡; is the Fourier transform kernel function; is the two-norm calculation operation.

[0069] As a further improvement of this technical solution, in the above S3.2, based on the track curvature and modal components, calculate the instantaneous energy distribution, time-frequency energy density change, power spectrum entropy, and track curvature correlation of the dual-component modal components. The specific method steps are as follows:

[0070] ;

[0071] Among them, is the instantaneous energy distribution of the th dual-component modal component; is the instantaneous vibration signal of the th dual-component modal component; is the index of the dual-component modal component;

[0072] ;

[0073] Among them, is the frequency; is the time-frequency energy density of the th dual-component modal component at different times and frequencies; Short-time Fourier transform of a double-component modal component;

[0074] ;

[0075] wherein, is the power spectrum entropy of the th double-component modal component; is the frequency index; is the th double-component modal component at the th frequency point of the normalized power spectral density; is the total number of frequency points of the power spectrum;

[0076] ;

[0077] wherein, is the cross-correlation between the th double-component modal component and the track curvature;

[0078] wherein, due to the instantaneous energy distribution, time-frequency energy density variation, power spectrum entropy and track curvature correlation of the double-component modal component, an orbital curvature modulation characteristic spectrum is formed.

[0079] As a further improvement of this technical solution, in S3.3, based on the double-component modal component, calculate the Lyapunov exponent to determine whether the double-component is stable and fault-free under the change of track curvature. The specific method steps are as follows:

[0080] ;

[0081] wherein, is the Lyapunov exponent of the th double-component modal component; is the perturbation of the th double-component modal component at time ; is the perturbation of the th double-component modal component at time ;

[0082] As a further improvement of this technical solution, in S4, calculate the Lyapunov exponents of all modal components to determine whether the double-component is faulty. The specific method steps are as follows:

[0083] S4.1. If there exists the th double-component modal component such that , it means that the fault signal of the interaction between the double-components is amplified, there is a fault between the double-components, record this double-component and give an alarm;

[0084] If for all it is satisfied that , it indicates that there is no abnormality in the interaction between the two components, and the mutual influence of these two components is excluded.

[0085] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0086] 1. In the method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment, by using the dynamic modeling method and introducing the impact force of raindrops and hailstones on components and the additional viscous resistance of snow accumulation, the coupling relationship between two components is calculated, which can fully explore the dynamic coupling relationship between the wheel-set bearing - gear transmission system under rainy and snowy weather conditions, and extract weak fault features under the condition of signal mutual enhancement, improving the sensitivity and accuracy of fault detection.

[0087] 2. In the method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment, through the modulation modeling of the component fault signal by the track curvature modulation feature spectrum combined with the Lyapunov exponent stability analysis, it is ensured that under curved track sections or specific weather conditions, weak fault features can be accurately extracted. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] Figure 1 It is the overall method flow chart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0089] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0090] Embodiment:

[0091] Please refer to Figure 1 As shown, this embodiment provides a method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment, including the following steps:

[0092] S1. Use track measurement equipment to collect track geometric data and vehicle operation data, calculate the vehicle force parameters corresponding to the track curvature, and establish a vehicle component force relationship matrix;

[0093] The track geometric data includes track curvature, track gradient, track superelevation, and track irregularity;

[0094] The vehicle operation data includes running speed, acceleration, axle load, wheel-rail contact force, and suspension system parameters;

[0095] In this embodiment S1, the vehicle force parameters corresponding to the track curvature are calculated, and the force relationship matrix of vehicle components is established. The specific method steps are as follows:

[0096] S1.1. Establish a curvature calculation model based on the track centerline equation:

[0097] ;

[0098] where, is the track coordinate; is the curvature at the track coordinate 𝑠; is the track tangent angle; is the arc length increment on the track;

[0099] In this embodiment, the track centerline equation is a standard mathematical model in railway and rail transit engineering and can be expressed by a parametric equation or an implicit equation;

[0100] S1.2. Based on the curvature calculation model, calculate the vehicle force parameters, including the normal force, lateral force, and component forces;

[0101] where the component forces are the bearing force and the gear force;

[0102] ;

[0103] ;

[0104] where, is the centrifugal force caused by the track curvature; is the vehicle mass; is the vehicle speed; is the total normal force of the vehicle; is the acceleration due to gravity; is the track gradient;

[0105] ;

[0106] where, is the lateral force caused by the track curvature; is the track superelevation; is the track curvature radius;

[0107] ;

[0108] ;

[0109] ;

[0110] where, is the contact force between the wheel and the track; is the wheel-rail contact area; is the bearing force; is the gear force; is the transmission torque; is the gear radius;

[0111] S1.3. Establish the force relationship matrix of vehicle components:

[0112] ;

[0113] wherein, is the force relationship matrix of vehicle components.

[0114] S2. Use sensors to collect the state data of vehicle components in real time during vehicle operation, calculate the coupling relationship between double components by using the dynamic modeling method, and use the signal synthesis method to superimpose the vibration signals of double components to generate the coupling fault characteristic signal;

[0115] In step S2 of this embodiment, the component state data includes vibration, stress, temperature and current; The coupling relationship between double components is calculated by using the dynamic modeling method, and the vibration signals of double components are superimposed by using the signal synthesis method to generate the coupling fault characteristic signal. The specific method steps are as follows:

[0116] S2.1. Based on the force relationship matrix of vehicle components, construct the coupling dynamic equation of double components and calculate the modal coupling relationship of double components;

[0117] S2.2. Based on the modal coupling relationship of double components, judge whether the two components can perform signal synthesis, and extract the fault impact characteristics and construct the coupling fault characteristic signal.

[0118] In step S2.1 of this embodiment, based on the force relationship matrix of vehicle components, construct the coupling dynamic equation of double components and calculate the modal coupling relationship of double components. The specific method is as follows:

[0119] S2.1.1. On the basis of the force relationship matrix of vehicle components, incorporate the raindrop and hail impact force, the additional snow resistance and the track humidity into the state relationship of double components to construct the dynamic equation of the modal coupling relationship of double components:

[0120] ;

[0121] wherein, is the mass matrix; is the damping matrix; is the stiffness matrix; is the displacement vector; is the velocity vector; is the acceleration vector; is the damping correction term affected by humidity; is the damping correction term for raindrop and hail impact; is the stiffness correction term for the influence of track wet-slipping; is the raindrop and hail impact force; is the influence matrix of the raindrop and hail impact force; is the additional resistance of snow accumulation; is the influence matrix of the additional resistance of snow accumulation;

[0122] ;

[0123] ;

[0124] ;

[0125] ;

[0126] ;

[0127] Among them, is the humidity influence coefficient; is the humidity; is the impact influence coefficient; is the impact frequency of raindrops and hailstones on components; is the raindrop and hail impact force; is the correction coefficient; is the wheel-rail friction coefficient under wet-slipping conditions; is the stiffness matrix; is the mass of raindrops and hailstones; is the impact velocity; is the snow accumulation density; is the snow accumulation contact area;

[0128] S2.1.2. Calculate the modal coupling relationship of the two components based on the dynamic equation of the modal coupling relationship of the two components:

[0129] ;

[0130] Among them, is the component index; is the component index; is the component and the component the coupling degree between; is the time; is the component at time the vibration displacement signal; is the component at time the vibration displacement signal; is the component and spare parts signal cross-integration between them; for spare parts energy calculation of vibration signals; for spare parts energy calculation of vibration signals.

[0131] In this embodiment, in rainy and snowy weather, the force and vibration characteristics of the track-wheel pair-bearing-gear system change, and weak fault characteristics will be amplified, mainly reflected in: snow or rain forms a lubricating layer on the track surface, reducing the wheel-rail adhesion coefficient and causing changes in wheel-rail forces (normal force, lateral force); the lubricating layer reduces the friction of the gear transmission system, reducing the torque transmission efficiency and affecting the gear force; the impact of raindrops or hailstones may cause additional impact loads, resulting in noise in the vibration signal and affecting the accuracy of the coupled signal; rainy and snowy weather increases air humidity, changing the damping coefficient of materials and affecting the vibration propagation characteristics;

[0132] In step S2.2 of this embodiment, based on the modal coupling relationship of two spare parts, it is judged whether the two spare parts can perform signal synthesis, and the fault impact characteristics are extracted and the coupled fault characteristic signal is constructed. The specific method is as follows:

[0133] If the spare parts and spare parts the coupling degree between them , then the signals of the two spare parts are strongly correlated and signal synthesis can be performed:

[0134] ;

[0135] Among them, is the synthesized signal of two spare parts; is the vibration signal of spare parts ; is the vibration signal of spare parts ; is the weight coefficient of spare parts ; is the weight coefficient of spare parts ;

[0136] Adaptive filtering is used to remove the noise of the vibration signals of spare parts and spare parts , and calculate the cross-correlation function of the vibration signals of spare parts and spare parts :

[0137] ;

[0138] Among them, is the time delay; For component parts and component parts cross - correlation function of vibration signals;

[0139] Calculate the instantaneous energy and envelope signal of the combined signal of two component parts :

[0140] ;

[0141] ;

[0142] wherein, is the instantaneous energy; is the envelope signal; is the Hilbert transform;

[0143] Construct the coupled fault feature signal:

[0144] ;

[0145] wherein, is the coupled fault feature signal.

[0146] S3. Use the variational mode decomposition method to decompose the coupled fault feature signal into multiple mode components, analyze and calculate the orbital curvature modulation feature spectra of each mode component under different orbital curvature conditions, calculate the Lyapunov exponents of the mode components, analyze the dynamic stability of the mode components under different orbital curvatures and operating states, and store the analyzed orbital curvature modulation feature spectra and Lyapunov exponents;

[0147] In this embodiment S3, the orbital curvature modulation feature spectrum is a set of features describing the energy distribution, frequency characteristics and nonlinear dynamic behavior of the coupled fault signal under different orbital curvature conditions, used to analyze how the weak fault signal is affected by the orbital curvature and extract the fault features manifested with the change of curvature;

[0148] In this embodiment S3, use the variational mode decomposition method to decompose the coupled fault feature signal into multiple mode components, analyze and calculate the orbital curvature modulation feature spectra of each mode component under different orbital curvature conditions, calculate the Lyapunov exponents of the mode components, analyze the dynamic stability of the mode components under different orbital curvatures and operating states, and store the analyzed orbital curvature modulation feature spectra and Lyapunov exponents. The specific method steps are as follows:

[0149] S3.1. Use the variational mode decomposition of the combined signal of two component parts , construct the optimization problem of two component parts, and use the alternating direction method of multipliers to solve the optimization problem of two component parts to obtain multiple two - component mode components;

[0150] S3.2. Calculate the instantaneous energy distribution, time-frequency energy density variation, power spectrum entropy, and track curvature correlation of the dual-component modal components based on the track curvature and modal components;

[0151] S3.3. Calculate the Lyapunov exponent based on the dual-component modal components to determine whether the dual components are stable and fault-free under the change of track curvature;

[0152] S3.4. Store the analyzed track curvature modulation feature spectrum and Lyapunov exponent.

[0153] In the present embodiment S3.1, the variational mode decomposition is used for the dual-component composite signal , construct the dual-component optimization problem, and use the alternating direction multiplier method to solve the dual-component optimization problem to obtain multiple dual-component modal components. The specific method steps are as follows:

[0154] ;

[0155] Among them, is the index of the dual-component modal component; is the th central frequency of the dual-component modal component; is the unit impulse function; is the imaginary unit; is the pi; is the partial derivative calculation with respect to time 𝑡; is the Fourier transform kernel function; is the two-norm calculation operation.

[0156] In the present embodiment S3.2, based on the track curvature and modal components, calculate the instantaneous energy distribution, time-frequency energy density variation, power spectrum entropy, and track curvature correlation of the dual-component modal components. The specific method steps are as follows:

[0157] ;

[0158] Among them, is the th instantaneous energy distribution of the dual-component modal component; is the th instantaneous vibration signal of the dual-component modal component; is the index of the dual-component modal component; is the total number of the dual-component modal components;

[0159] ;

[0160] Among them, is the frequency; is the The time-frequency energy density of a double-component modal component at different times and frequencies; is the short-time Fourier transform of the

[0161] -th double-component modal component;

[0162] wherein, is the power spectrum entropy of the -th double-component modal component; is the frequency index; is the -th double-component modal component at the -th frequency point of the normalized power spectrum density; is the total number of frequency points of the power spectrum;

[0163] ;

[0164] wherein, is the cross-correlation between the -th double-component modal component and the track curvature;

[0165] wherein, due to the instantaneous energy distribution, time-frequency energy density change, power spectrum entropy and track curvature correlation of the double-component modal component, an orbit curvature modulation characteristic spectrum is formed.

[0166] In the above S3.3, based on the double-component modal component, calculate the Lyapunov exponent to judge whether the double-component is stable and fault-free under the change of track curvature. The specific method steps are as follows:

[0167] ;

[0168] wherein, is the Lyapunov exponent of the -th double-component modal component; is the perturbation of the -th double-component modal component at time ; is the perturbation of the -th double-component modal component at time ;

[0169] S4. Calculate the Lyapunov exponents of all modal components to judge whether the double-component is faulty;

[0170] In this embodiment S4, calculate the Lyapunov exponents of all modal components to judge whether the double-component is faulty. The specific method steps are as follows:

[0171] S4.1. If there exists the -th double-component modal component such that It indicates that the fault signal representing the interaction between the two components is amplified, there is a fault between the two components, record these two components and give an alarm;

[0172] If for all all satisfy it indicates that the interaction between the two components is normal, and eliminate the mutual influence of these two components.

[0173] The above shows and describes the basic principle, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed.

Claims

1. A method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment, characterized in that It includes the following steps: S1. Collect track geometry data and vehicle operation data using track measurement equipment, calculate the vehicle force parameters corresponding to the track curvature, and establish a force relationship matrix for vehicle components; S2. Use sensors to collect the state data of vehicle components in real time during vehicle operation. Using the dynamic modeling method and introducing the impact force of raindrops and hailstones on the components and the additional viscous resistance of snow, calculate the coupling relationship between two components, and use the signal synthesis method to superimpose the vibration signals of the two components to generate a coupled fault characteristic signal; S3. Use the variational mode decomposition method to decompose the coupled fault characteristic signal into multiple modal components, analyze and calculate the track curvature modulation characteristic spectra of each modal component under different track curvature conditions, calculate the Lyapunov exponent of the modal component, analyze the dynamic stability of the modal component under different track curvatures and operating states, and store the analyzed track curvature modulation characteristic spectra and Lyapunov exponents; S4. Calculate the Lyapunov exponents of all modal components and determine whether the two components are faulty.

2. The method for extracting weak fault characteristics and intelligent fault diagnosis of key components of rail transit equipment according to claim 1, characterized in that: The track geometry data includes track curvature, track gradient, track superelevation, and track irregularity; The vehicle operation data includes operating speed, acceleration, axle load, wheel-rail contact force, and suspension system parameters; In S1, to calculate the vehicle force parameters corresponding to the track curvature and establish a force relationship matrix for vehicle components, the specific method steps are as follows: S1.

1. Establish a curvature calculation model based on the track centerline equation: ； Among them, is the orbital coordinate; is the curvature at the orbital coordinate 𝑠; is the orbital tangent angle; is the arc length increment on the orbit; S1.

2. Based on the curvature calculation model, calculate the vehicle force parameters, including normal force, lateral force, and component forces; Among them, the component forces are bearing force and gear force; ； ； Among them, is the centrifugal force caused by the track curvature; is the vehicle mass; is the vehicle speed; is the total normal force of the vehicle; is the gravitational acceleration; is the track gradient; ； Among them, is the lateral force caused by the track curvature; is the track superelevation; is the track curvature radius; ； ； ； Wherein, is the contact force between the wheel and the track; is the wheel-rail contact area; is the bearing force; is the gear force; is the transmission torque; is the gear radius; S1.

3. Establish a force relationship matrix for vehicle components: ； Among them, is the force relationship matrix of vehicle components.

3. The method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment according to claim 2, characterized in that: In S2, the component state data includes vibration, stress, temperature, and current; In S2, use the dynamic modeling method to calculate the coupling relationship between two components, and use the signal synthesis method to superimpose the vibration signals of the two components to generate a coupled fault characteristic signal. The specific method steps are as follows: S2.

1. Based on the force relationship matrix of vehicle components, construct a coupled dynamic equation for two components and calculate the modal coupling relationship between the two components; S2.

2. Based on the modal coupling relationship between the two components, determine whether the two components can perform signal synthesis, extract the fault impact characteristics, and construct a coupled fault characteristic signal.

4. The method for extracting weak fault characteristics and intelligent fault diagnosis of key components of rail transit equipment according to claim 3, wherein: In S2.1, based on the force relationship matrix of vehicle components, construct a coupled dynamic equation for two components and calculate the modal coupling relationship between the two components. The specific method is as follows: S2.1.

1. On the basis of the force relationship matrix of vehicle components, incorporate the impact force of raindrops and hailstones, the additional resistance of snow, and the track humidity into the state relationship of the two components to construct a dynamic equation for the modal coupling relationship between the two components: ； Among them, is the mass matrix; is the damping matrix; is the stiffness matrix; is the displacement vector; is the velocity vector; is the acceleration vector; is the damping correction term affected by humidity; is the damping correction term affected by raindrop and hail impact; is the stiffness correction term affected by wet slippery track; is the raindrop and hail impact force; is the influence matrix of raindrop and hail impact force; is the additional resistance of snow accumulation; is the influence matrix of additional resistance of snow accumulation; ； ； ； ； ； Among them, is the humidity influence coefficient; is the humidity; is the impact influence coefficient; is the impact frequency of raindrops and hailstones on components; is the impact force of raindrops and hailstones; is the correction coefficient; is the wheel-rail friction coefficient under wet and slippery conditions; is the stiffness matrix; is the mass of raindrops and hailstones; is the impact velocity; is the snow density; is the snow contact area; S2.1.

2. Based on the dynamic equation for the modal coupling relationship between the two components, calculate the modal coupling relationship between the two components; ； Among them, is the component index; is the component index; is the component and the component coupling degree between; is the time; is the component at time vibration displacement signal; is the component at time vibration displacement signal; is the component and the component signal cross integration between; is the component vibration signal energy calculation; is the component vibration signal energy calculation.

5. The method for extracting weak fault characteristics and intelligent fault diagnosis of key components of rail transit equipment according to claim 4, characterized in that: In S2.2, based on the modal coupling relationship between the two components, determine whether the two components can perform signal synthesis, extract the fault impact characteristics, and construct a coupled fault characteristic signal. The specific method is as follows: If the components and the components have a coupling degree , the signals of the two components are strongly correlated and signal synthesis can be performed: ； Among them, is a composite signal of two components; is the vibration signal of component ; is the vibration signal of component ; is the weight coefficient of component ; is the weight coefficient of component ; Adaptive filtering is used to remove the noise of the vibration signals of components and components and calculate the cross - cross - correlation function of the vibration signals of components and components : ； Among them, is the time delay; is the cross-correlation function of the vibration signals of component and component ; Calculating the instantaneous energy and envelope signal of the composite signal of two components : ； ； Among them, is the instantaneous energy; is the envelope signal; is the Hilbert transform; Construct a coupled fault characteristic signal: ； Among them, is the coupled fault characteristic signal.

6. The method for extracting weak fault characteristics and intelligent fault diagnosis of key components of rail transit equipment according to claim 5, characterized in that: In S3, the orbital curvature modulation characteristic spectrum is a set of characteristics that describe the energy distribution, frequency characteristics, and nonlinear dynamic behavior of the coupled fault signal under different orbital curvature conditions, and is used to analyze how weak fault signals are affected by the orbital curvature and extract fault characteristics that appear with the change of curvature. In S3, the variational mode decomposition method is used to decompose the coupled fault characteristic signal into multiple modal components, analyze and calculate the orbital curvature modulation characteristic spectra of each modal component under different orbital curvature conditions, calculate the Lyapunov exponents of the modal components, analyze the dynamic stability of the modal components under different orbital curvatures and operating states, and store the analyzed orbital curvature modulation characteristic spectra and Lyapunov exponents. The specific method steps are as follows: S3.

1. Use variational mode decomposition for two-component composite signals , construct a two-component optimization problem, and use the alternating direction method of multipliers to solve the two-component optimization problem to obtain multiple two-component mode components; S3.2: Calculate the instantaneous energy distribution, time-frequency energy density change, power spectrum entropy, and orbital curvature correlation of the two-component modal component based on the orbital curvature and the modal component. S3.3: Calculate the Lyapunov exponent based on the two-component modal component and determine whether the two components are stable and fault-free under the change of orbital curvature. S3.4: Store the analyzed orbital curvature modulation characteristic spectrum and Lyapunov exponent.

7. The method for extracting weak fault characteristics and intelligent fault diagnosis of key components of rail transit equipment according to claim 6, characterized in that: In the above S3.1, a variational mode decomposition is used to synthesize a two-component signal , and a two-component optimization problem is constructed, and the alternating direction multiplier method is used to solve the two-component optimization problem to obtain a plurality of two-component mode components. The specific method steps are as follows: ； Among them, is the double-component modal component index; is the center frequency of the nth double-component modal component; is the unit impulse function; is the imaginary unit; is the pi; is to calculate the partial derivative with respect to time 𝑡; is the Fourier transform kernel function; is the two-norm calculation operation.

8. The method for extracting weak fault characteristics and intelligent fault diagnosis of key components of rail transit equipment according to claim 7, characterized in that: In S3.2, calculate the instantaneous energy distribution, time-frequency energy density change, power spectrum entropy, and orbital curvature correlation of the two-component modal component based on the orbital curvature and the modal component. The specific method steps are as follows: ； Among them, is the instantaneous energy distribution of the th two-component modal component; is the instantaneous vibration signal of the th two-component modal component; is the two-component modal component index; is the total number of two-component modal components; ； Among them, is the frequency; is the time-frequency energy density of the th two-component modal component at different times and frequencies; is the short-time Fourier transform of the ； Among them, is the power spectral entropy of the -th two-component modal component; is the frequency index; is the -th two-component modal component at the -th frequency point of the normalized power spectral density; is the total number of frequency points of the power spectrum; ； Among them, is the cross-correlation between the nth two-component modal component and the track curvature; Among them, the instantaneous energy distribution, time-frequency energy density change, power spectrum entropy, and orbital curvature correlation of the two-component modal component constitute the orbital curvature modulation characteristic spectrum.

9. The method for extracting weak fault characteristics and intelligent fault diagnosis of key components of rail transit equipment according to claim 8, wherein: In S3.3, calculate the Lyapunov exponent based on the two-component modal component and determine whether the two components are stable and fault-free under the change of orbital curvature. The specific method steps are as follows: ； wherein, is the Lyapunov exponent of the -th two-component modal component; is the perturbation of the -th two-component modal component at time ; is the perturbation of the -th two-component modal component at time .

10. The method for extracting weak fault features and intelligent fault diagnosis of key components of rail transit equipment according to claim 9, characterized in that: In S4, calculate the Lyapunov exponents of all modal components and determine whether the two components are faulty. The specific method steps are as follows: S4.

1. If there is a -th two-component modal component such that , it means that the fault signal indicating the interaction between the two components is amplified, there is a fault between the two components, record these two components and issue an alarm; If for all it is satisfied that , it indicates that there is no abnormality in the interaction between the two components, excluding the mutual influence of these two components.

Images