A wheel polygon recognition method and system based on angle domain averaging

Abstract

The present application relates to a kind of wheel polygon identification method and system based on angle domain average, method includes the following steps: the axle box vibration acceleration signal and vehicle speed signal of train in running process are collected;Determine wheel rotation period, according to angle domain average method to axle box vibration acceleration signal is signal resampling, superposition and average, obtain the axle box vibration acceleration signal after processing;Wheel polygon state includes the order of wheel polygon and corresponding amplitude, based on the wheel polygon signal after processing, the order of wheel polygon and amplitude identification.Compared with prior art, the axle box vibration acceleration and vehicle speed generated in the running process of railway vehicle are collected, and wheel polygon information in acceleration signal is extracted using angle domain average method, and then the order of wheel polygon and amplitude are further processed, wheel polygon characteristic signal can be extracted from complex acceleration signal and accurately identified.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of detection and identification of wheel polygonal state of rail vehicles, and particularly relates to a wheel polygonal identification method and system based on angle domain average. BACKGROUND

[0002] Wheel polygon is a common form of wheel non-circular, which specifically shows periodic uneven wear of wheel radius along the circumferential direction. Wheel polygon significantly increases the vertical force between the wheel and the track, aggravates the dynamic response between the wheel and the track, and affects the stability and safety of vehicle operation and the fatigue life of system components. Therefore, timely detection and maintenance of the wheel state are of great significance for the safe and stable service of the vehicle.

[0003] Currently, the detection of wheel polygon is mostly divided into direct measurement and indirect measurement. The traditional manual direct measurement method can only measure when the train is stopped, which consumes a lot of manpower and material resources and is low in efficiency. The indirect wheel polygon measurement method refers to not directly measuring the wheel, but arranging sensors at various positions of the vehicle to collect information related to the wheel polygon, so as to indirectly identify the wheel polygon. In actual application, the axle box vibration acceleration signal can obviously represent the vibration characteristics of the rail vehicle, and can detect the wheel polygon during vehicle operation, which has the advantages of reducing cost and improving efficiency, and is also helpful for monitoring and timely warning of the wheel condition.

[0004] However, the axle box vibration is affected by the whole vehicle-track coupling system during the operation of the train, and the vibration acceleration signal collected on site is inevitably affected by white noise and the like, which limits the accuracy of the identification result. Therefore, it is necessary to improve the existing wheel polygon detection scheme. SUMMARY

[0005] The purpose of the present application is to overcome the defects of the prior art and provide a wheel polygon identification method and system based on angle domain average.

[0006] The purpose of the present application can be achieved by the following technical solutions:

[0007] According to a first aspect of the present application, a wheel polygon identification method based on angle domain average is provided, comprising the following steps:

[0008] (1) collecting the axle box vibration acceleration signal and the vehicle speed signal of the train during operation;

[0009] (2) determining the wheel rotation period, and performing signal resampling, superposition and average on the axle box vibration acceleration signal according to the angle domain average method to obtain the processed axle box vibration acceleration signal;

[0010] (3) The wheel polygon state includes the order and the corresponding amplitude of the wheel polygon, and the order and the amplitude of the wheel polygon are identified based on the processed wheel polygon signal.

[0011] Further, the step (1) is specifically: collecting the axle box vibration acceleration signal of the train in operation through the acceleration sensor installed on the axle box of the train, and collecting the vehicle speed signal of the train in operation through the speed sensor installed on the bogie of the train.

[0012] Further, the step (2) includes:

[0013] S1, assuming that the collected axle box vibration acceleration signal is x(t), which includes the periodic signal wheel out-of-roundness signal f(t) and the non-periodic signal g(t), that is:

[0014] x(t)=f(t)+g(t)

[0015] S2, determining the vehicle rotation period based on the vehicle speed signal, and intercepting the axle box vibration acceleration signal x(t) according to the vehicle rotation period, so as to divide the axle box vibration acceleration signal x(t) into P segment signals;

[0016] S3, superimposing the P segment signals to obtain:

[0017]

[0018] wherein i=1, 2, …, P, x(t i ) represents the axle box vibration acceleration signal of the i-th wheel rotation;

[0019] S4, averaging the superimposed signal to obtain the processed axle box vibration acceleration signal:

[0020]

[0021] wherein y(t) represents the processed axle box vibration acceleration signal.

[0022] Further, in step S2, the vehicle rotation period is determined based on the vehicle speed signal, and specifically:

[0023] The wheel circumference is calculated based on the wheel radius, the vehicle speed signal is integrated with respect to time, and the integral calculation result is equal to the integral interval of the wheel circumference corresponding to one vehicle rotation period, as follows:

[0024]

[0025] wherein s(t) is the vehicle displacement at time t, R is the wheel radius, [t1, t2] is one vehicle rotation period, and v(t) is the vehicle speed signal.

[0026] Repeat the above steps to obtain multiple vehicle rotation periods, and the axle box vibration acceleration signal x(t) is intercepted according to the vehicle rotation period, and the axle box vibration acceleration signal x(t) is divided into P segment signals.

[0027] Further, in step (3), the excitation frequency of each order of wheel polygon is determined according to the relationship between the wheel polygon order and the wheel-rail excitation frequency, and the relationship between the wheel polygon order and the wheel-rail excitation frequency is as follows:

[0028]

[0029] Wherein, l represents the wheel polygon order; f w represents the wheel-rail excitation frequency generated by the wheel polygon wear; v represents the vehicle speed; d represents the wheel diameter.

[0030] Further, in step (3), the amplitude of the corresponding order of the wheel polygon is calculated according to the inertia reference method, including:

[0031] 1) The processed axle box vibration acceleration signal y(t) is subjected to Fourier transform, and the discrete time domain signal is converted into a frequency domain signal, as follows:

[0032]

[0033] Wherein, F(k) represents the result of transferring to the frequency domain after Fourier transform, s represents the sampling point, S represents the total number of sampling points, j represents the virtual unit, and k represents the frequency;

[0034] 2) In the frequency domain, the acceleration is converted to displacement, and frequency domain filtering processing is carried out and divided by the square of the angular frequency to obtain the result of twice frequency domain integration:

[0035]

[0036] Wherein, A(k) represents the result obtained after integration, f l and f h are the lower and upper cut-off frequencies respectively, and ω is the angular frequency;

[0037] 3) According to the excitation frequency f w calculated in step 1), the amplitude A(f w ) of each order of wheel polygon is determined, the amplitude is converted into roughness level, and the wheel roughness level calculation formula is:

[0038]

[0039] Wherein, L ris the amplitude of the wheel surface roughness grade; r0 is the reference value of the short-wave roughness of the rail surface; r rms is the amplitude A(f w ) of the roughness of the rail surface.

[0040] According to a second aspect of the present application, a wheel polygon identification system based on angle domain averaging is provided, comprising:

[0041] A signal acquisition module is configured to acquire an axle box vibration acceleration signal and a vehicle speed signal of a train during operation.

[0042] A signal processing module is configured to determine a wheel rotation period, and perform signal resampling, superposition and averaging on the axle box vibration acceleration signal according to an angle domain averaging method to obtain a processed axle box vibration acceleration signal.

[0043] An identification module is configured to identify a wheel polygon order and an amplitude based on the processed wheel polygon signal, wherein the wheel polygon state comprises the order and the corresponding amplitude of the wheel polygon.

[0044] Further, the step (2) comprises:

[0045] S1, setting the acquired axle box vibration acceleration signal as x(t), wherein the axle box vibration acceleration signal comprises a periodic signal wheel polygon signal f(t) and a non-periodic signal g(t), i.e.

[0046] x(t) = f(t) + g(t)

[0047] S2, determining a vehicle rotation period based on a vehicle speed signal, and cutting the axle box vibration acceleration signal x(t) according to the vehicle rotation period to divide the axle box vibration acceleration signal x(t) into P segments of signals.

[0048] S3, superimposing the P segments of signals to obtain:

[0049]

[0050] wherein i = 1, 2, …, P, x(t i ) represents the axle box vibration acceleration signal of the i-th wheel rotation.

[0051] S4, averaging the superimposed signal to obtain a processed axle box vibration acceleration signal:

[0052]

[0053] wherein y(t) represents the processed axle box vibration acceleration signal.

[0054] Further, in step (3), the excitation frequency of each order of wheel polygon is determined according to the relationship between the wheel polygon order and the wheel-rail excitation frequency, and the relationship between the wheel polygon order and the wheel-rail excitation frequency is as follows:

[0055]

[0056] wherein, l represents the wheel polygon order; f w represents the wheel-rail excitation frequency generated by the wheel polygon wear; v represents the vehicle speed; and d represents the wheel diameter.

[0057] Further, in step (3), the amplitude of the wheel polygon corresponding order is calculated according to the inertia reference method, including:

[0058] 1) The processed axle box vibration acceleration signal y(t) is subjected to Fourier transform, and the discrete time domain signal is converted into a frequency domain signal, as follows:

[0059]

[0060] wherein, F(k) represents the result of the Fourier transform transferred into the frequency domain, s represents a sampling point, S represents the total number of sampling points, j represents a virtual unit, and k represents a frequency;

[0061] 2) The acceleration is converted into displacement in the frequency domain, and frequency domain filtering processing is performed and divided by the square of the angular frequency, to obtain the result of twice frequency domain integration:

[0062]

[0063] wherein, A(k) represents the result obtained after integration, f l and f h are the lower limit cutoff frequency and the upper limit cutoff frequency, respectively, and ω is the angular frequency;

[0064] 3) The amplitude A(f w ) of each order of wheel polygon is determined according to the excitation frequency f w of each order of wheel polygon obtained in step 1), and the amplitude is converted into a roughness level, and the wheel roughness level calculation formula is as follows:

[0065]

[0066] wherein, L r is the wheel surface roughness level amplitude; r0 is the reference value of the rail surface short wave roughness; and r rms is the effective value of the rail surface roughness amplitude A(f w ).

[0067] Compared with the prior art, the present application has the following beneficial effects:

[0068] (1) The axle box vibration acceleration generated by the rail vehicle in the running process and the vehicle speed are collected, and the angle domain average method is used to extract the wheel polygon information in the acceleration signal, and then the order and amplitude of the wheel polygon are further processed, so that the wheel polygon characteristic signal can be extracted from the complex acceleration signal and accurately identified.

[0069] (2) Compared with other wheel polygon identification methods, the present application fully considers the speed fluctuation of the train in the running process, and the signal is segmented according to the rotation period of the vehicle, so that the wheel polygon amplitude can be accurately identified under the condition of the change of the train speed, and has great practical application value.

[0070] (3) The present application separates the wheel polygon signal according to the angle domain average method for signal resampling, superposition and average, which can effectively alleviate the problem of interference of other noise signals when using indirect method to measure the vehicle polygon, and improves the identification accuracy.

[0071] (4) The present application combines the inertia reference method with the frequency domain secondary integration method, avoids the error accumulation problem in time domain integration, and further improves the accuracy of identifying the wheel polygon by using the axle box vibration acceleration. BRIEF DESCRIPTION OF DRAWINGS

[0072] Figure 1 The flowchart of the wheel polygon identification method based on the angle domain average of the present application is shown in the figure.

[0073] Figure 2 The comparison chart of the measured value and the identified value is shown in the figure. DETAILED DESCRIPTION

[0074] The present application will be described in detail below in combination with the drawings and specific embodiments. The present embodiment is implemented on the basis of the technical solution of the present application, and gives a detailed implementation manner and specific operation process. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. The protection scope of the present application is not limited to the following embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.

[0075] As used in this document, "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one implementation of the application. The appearances of the phrase "in one embodiment" or "an embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily referring to a single embodiment. As such, the terms "first," "second," and the like within the specification, do not necessarily imply a particular order or sequence. Also, the terms "comprises," "comprising," and the like, are intended to cover a non-exclusive inclusion, such that any process, method, article, or apparatus that includes a list of elements is not necessarily limited to those elements, but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. Further, the terms "may" and "might" are used under their permissive senses, rather than their mandatory senses, such that these terms do not necessarily indicate that a particular step or action is required.

[0076] The specification provides method operation steps as an example or flowchart, but can include more or fewer operations than shown or described, without departing from the scope of the application. The steps listed in the examples are merely one way of executing the steps, and are not necessarily the only way of executing the steps. The steps can be executed in the order listed, in parallel, in a different order, or in an adjusted order without departing from the scope of the application. The system or server product can be executed in the order listed in the example or in parallel (for example, in a parallel processor or multi-threaded processing environment) or in an adjusted order without departing from the scope of the application.

[0077] The application provides a wheel polygon recognition method based on angle domain average, and the implementation flow is as shown in Figure 1 The method comprises the following steps.

[0078] (1) Collecting the axle box vibration acceleration signal and the vehicle speed signal of the train in the running process;

[0079] Specifically, the acceleration sensor is installed at the axle box position of the train, connected with the collection device through the data line, and the change information of the train axle box vibration acceleration when the train runs, i.e. the axle box vibration acceleration signal of the train when running, is collected. The speed sensor is installed at the position of the train bogie, connected with the collection device through the data line, and the change information of the speed when the train runs, i.e. the vehicle speed signal of the train when running, is collected.

[0080] In the actual implementation process, the sampling frequency of the axle box vibration acceleration sensor is 5120 Hz, and the sampling frequency of the speed sensor is 20 Hz.

[0081] (2) According to the angle domain average method, the signal is resampled, superimposed and averaged, and the wheel polygon signal is separated. In actual situations, due to the influence of track irregularities and other interference excitations, it is usually difficult to identify the frequency components of the wheel polygon. Therefore, the collected signal needs to be processed again to separate the wheel polygon signal from the complex integrated signal. Based on this, the angle domain average method is proposed. According to the measured vehicle speed signal, the wheel rotation period is determined; combined with the wheel rotation period, the original axle box vibration acceleration signal is resampled, superimposed and averaged, the acceleration time domain signal is converted into an angle domain signal, and the wheel polygon characteristic signal is extracted.

[0082] According to the periodic characteristics of the wheel polygon, the time domain signal is converted into an angle domain signal based on the wheel rotation angle, and is superimposed and averaged, and the specific implementation process is as follows:

[0083] S1, the collected axle box vibration acceleration signal is x(t), which includes periodic signal wheel polygon signal f(t) and non-periodic signal g(t), that is:

[0084] x(t)=f(t)+g(t)

[0085] S2, the vehicle rotation period is determined based on the vehicle speed signal, and the axle box vibration acceleration signal x(t) is intercepted according to the vehicle rotation period, and the axle box vibration acceleration signal x(t) is divided into P segment signals: x(t1), x(t2), …, x(tP), x(tP+1), x(tP+2), …, x(tP+P-1), x(tP+P) and x(tP+P+1), wherein x(tP+P+1) represents the axle box vibration acceleration signal of the i-th wheel rotation, which includes the wheel polygon signal and the noise signal: P i

[0086] x(t i )=f(t i )+g(t i )

[0087] Among them, the vehicle rotation period is determined based on the vehicle speed signal, and the specific implementation process is as follows:

[0088] According to the vehicle speed, the displacement of the vehicle changes with time. Based on the wheel radius R, the wheel circumference 2πR is calculated, and the vehicle speed signal is integrated with respect to time. Whenever the integral result is equal to the wheel circumference, it can be considered that the wheel rotates one revolution in the integral interval, that is, the integral interval corresponding to the integral result equal to the wheel circumference corresponds to one vehicle rotation period. As follows, the wheel rotation period can be obtained:

[0089]

[0090] ​​Where s(t) is the vehicle displacement at time t, R is the wheel radius, [t1, t2] is a vehicle rotation period, and v(t) is the vehicle speed signal.

[0091] The above steps can be repeated starting from t = 0 to obtain multiple vehicle rotation periods, and the axle box vibration acceleration signal x(t) is correspondingly intercepted according to the vehicle rotation periods, so that the axle box vibration acceleration signal x(t) is divided into P signals, each of which represents the axle box vibration acceleration signal for one rotation of the wheel:

[0092] x(t i ) = f(t i ) + g(t i )

[0093] S3, superimpose the P signals, and due to the incoherence of the non-periodic signals, obtain:

[0094]

[0095] Where i = 1, 2, …, P, and x(t i ) represents the axle box vibration acceleration signal for the i-th rotation of the wheel.

[0096] S4, average the superimposed signal to obtain an output signal, i.e., a processed axle box vibration acceleration signal:

[0097]

[0098] Where y(t) represents the processed axle box vibration acceleration signal, and the output signal y(t) at this time can be regarded as the separated wheel polygon signal.

[0099] From the above formula, after averaging, the wheel out-of-round signal remains unchanged, and the remaining non-periodic signals are attenuated by a factor of times after averaging. When P is large enough, it can be considered that the part of the non-periodic signal approaches 0, and at this time the output signal can be regarded as the angular domain signal of the axle box vibration acceleration change caused by the wheel out-of-round in one wheel rotation period. In this way, the separation process of the wheel out-of-round signal is completed.

[0100] (3) The wheel polygon state includes the order and corresponding amplitude of the wheel polygon, the order of the wheel polygon is identified according to the characteristic frequency of the wheel polygon, and the amplitude of the wheel polygon is calculated by frequency domain quadratic integration according to the inertia reference method.

[0101] ① According to the relationship between the order of the wheel polygon and the wheel-rail excitation frequency, the excitation frequency of each order of the wheel polygon is determined, and the relationship between the order of the wheel polygon and the wheel-rail excitation frequency is as follows:

[0102]

[0103] Where l represents the order of the wheel polygon; f w The value represents the wheel-rail excitation frequency caused by polygonal wear of the wheel; v represents the vehicle speed in km / h; and d represents the wheel diameter in m.

[0104] ②The amplitude of the wheel polygon corresponding to the order is calculated using the inertial reference method.

[0105] The inertial reference method is a commonly used method for determining the amplitude of a wheel polygon. Its main principle is that when the wheel travels on the track, the vertical displacement H of the wheel axle box is the wheel-rail roughness η, which is the surface roughness of the wheel under the combined action of various wheel polygon orders. Based on the relationship between acceleration and displacement, the vertical displacement of the wheel can be obtained by performing a second integration on the vertical acceleration x(t) of the vehicle body measured by the accelerometer, as shown in the following formula:

[0106] η=H=∫∫x(t)dtdt

[0107] Furthermore, to avoid the cumulative bias caused by direct calculation using the inertial reference method, a frequency-domain quadratic integration method is proposed. This method requires first transforming the time-domain array to the frequency domain using a Fast Fourier Transform (FFT), then performing integration and filtering in the frequency domain using the integral property of the Fourier Transform, and finally returning to the time domain using an Inverse Fourier Transform (IFFT). The resulting array is the integral result of the time-domain array. Its main process is as follows:

[0108] 1) The processed axle box vibration acceleration signal y(t) is subjected to Fourier transform to convert the discrete time-domain signal into a frequency-domain signal, as follows:

[0109]

[0110] Where F(k) represents the result after Fourier transform and transfer to the frequency domain, s represents the sampling point, S represents the total number of sampling points, j represents the imaginary unit, and k represents the frequency;

[0111] 2) Convert the acceleration to displacement in the frequency domain, perform frequency domain filtering, and divide by the square of the angular frequency to obtain the result of two frequency domain integrals:

[0112]

[0113] Where A(k) represents the result obtained after integration, f l and f h These are the lower cutoff frequency and the upper cutoff frequency, respectively, and ω is the angular frequency;

[0114] 3) The excitation frequency f is obtained from the wheel polygons of each order calculated in step 1). wDetermine the amplitude A(f) of each order of wheel polygon. w The amplitude is converted into a roughness grade, and the formula for calculating the wheel roughness grade is as follows:

[0115]

[0116] Among them, L r r0 is the wheel surface roughness grade amplitude, in dB; r0 is the rail surface short-wave roughness reference value, generally, the rail surface short-wave roughness reference value r0 = 1 μm; r rms The surface roughness amplitude A(f) of the rail w ) effective value.

[0117] Compared to traditional wheel polygon detection technology, this invention can effectively identify the polygonal state of wheels during train operation, saving manpower and material costs. Compared to previous wheel polygon recognition methods, it can effectively reduce the impact of noise and also consider the impact of speed fluctuations during vehicle operation, improving the accuracy of recognition. Furthermore, it facilitates tracking the evolution of wheel polygons, enabling timely scheduling of wheel maintenance.

[0118] like Figure 2 As shown in the figure, this figure is the identification result of the order and amplitude of the wheel polygon in this embodiment. The identification result shows that the wheel polygon detection method proposed in this invention can accurately identify the main order of the wheel polygon and has high detection accuracy.

[0119] The present invention also provides a wheel polygon recognition system based on angle domain averaging, comprising:

[0120] The signal acquisition module is used to acquire axle box vibration acceleration signals and vehicle speed signals during train operation;

[0121] The signal processing module is used to determine the wheel rotation period and to resample, superimpose, and average the axle box vibration acceleration signal according to the angle domain averaging method to obtain the processed axle box vibration acceleration signal.

[0122] The recognition module is used to identify the order and amplitude of the wheel polygon based on the processed wheel polygon signal, wherein the wheel polygon state includes the order of the wheel polygon and the corresponding amplitude.

[0123] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0124] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for wheel polygon recognition based on angle domain averaging, characterized in that, Includes the following steps: (1) Collect axle box vibration acceleration signals and vehicle speed signals during train operation; (2) Determine the wheel rotation period, and perform signal resampling, superposition and averaging of the axle box vibration acceleration signal according to the angle domain averaging method to obtain the processed axle box vibration acceleration signal; (3) The wheel polygon state includes the order of the wheel polygon and the corresponding amplitude. The wheel polygon order and amplitude are identified based on the processed wheel polygon signal. Step (2) includes: S1. Let the collected axle box vibration acceleration signal be x(t), which includes the periodic signal f(t) indicating wheel out-of-roundness and the aperiodic signal g(t), i.e.: x(t)=f(t)+g(t) S2. Determine the vehicle rotation period based on the vehicle speed signal, extract the axle box vibration acceleration signal x(t) according to the vehicle rotation period, and divide the axle box vibration acceleration signal x(t) into P segments; S3. Superimpose the P-segment signals to obtain: Where i = 1, 2, ..., P, x(t) i This represents the axle box vibration acceleration signal during the i-th rotation of the wheel; S4. Averaging the superimposed signals yields the processed axle box vibration acceleration signal: Where y(t) represents the processed axle box vibration acceleration signal; In step S2, determining the vehicle rotation period based on the vehicle speed signal specifically involves: The wheel circumference is calculated based on the wheel radius by integrating the vehicle speed signal over time. The integral interval, where the wheel circumference is equal to the integral calculation result, corresponds to one vehicle rotation cycle, as follows: Where s(t) is the vehicle displacement at time t, R is the wheel radius, [t1,t2] is one vehicle rotation period, and v(t) is the vehicle speed signal. Repeat the above steps to obtain multiple vehicle rotation cycles. According to the vehicle rotation cycle, the axle box vibration acceleration signal x(t) is correspondingly truncated and divided into P segments. In step (3), the excitation frequency of each order of the wheel polygon is determined according to the relationship between the wheel-rail excitation frequency and the order of the wheel polygon. The relationship between the wheel polygon order and the wheel-rail excitation frequency is as follows: Where l represents the order of the wheel polygon; f w The wheel-rail excitation frequency caused by polygonal wear of the wheel is represented; v represents the vehicle speed; d represents the wheel diameter. In step (3), the amplitude of the wheel polygon corresponding to the order is calculated according to the inertial reference method, including: 1) The processed axle box vibration acceleration signal y(t) is subjected to Fourier transform to convert the discrete time-domain signal into a frequency-domain signal, as follows: Where F(k) represents the result after Fourier transform and transfer to the frequency domain, s represents the sampling point, S represents the total number of sampling points, j represents the imaginary unit, and k represents the frequency; 2) Convert the acceleration to displacement in the frequency domain, perform frequency domain filtering, and divide by the square of the angular frequency to obtain the result of two frequency domain integrals: Where A(k) represents the result obtained after integration, f l and f h These are the lower cutoff frequency and the upper cutoff frequency, respectively, and ω is the angular frequency; 3) The excitation frequency f is obtained from the wheel polygons of each order calculated in step 1). w Determine the amplitude A(f) of each order of wheel polygon. w The amplitude is converted into a roughness grade, and the formula for calculating the wheel roughness grade is as follows: Among them, L r r0 is the surface roughness grade amplitude of the wheel; r0 is the short-wave roughness reference value of the rail surface; r0 rms The surface roughness amplitude A(f) of the rail w ) effective value.

2. The wheel polygon recognition method based on angle domain averaging according to claim 1, characterized in that, The specific steps (1) are as follows: collecting the axle box vibration acceleration signal during train operation by using an acceleration sensor installed on the train axle box, and collecting the vehicle speed signal during train operation by using a speed sensor installed on the train bogie.

3. A wheel polygon recognition system based on angle domain averaging that implements the method as described in claim 1 or 2, characterized in that, include: The signal acquisition module is used to acquire axle box vibration acceleration signals and vehicle speed signals during train operation; The signal processing module is used to determine the wheel rotation period and to resample, superimpose, and average the axle box vibration acceleration signal according to the angle domain averaging method to obtain the processed axle box vibration acceleration signal. The recognition module is used to identify the order and amplitude of the wheel polygon based on the processed wheel polygon signal, wherein the wheel polygon state includes the order of the wheel polygon and the corresponding amplitude.

Images