Fault impact deconvolution tracking method for adaptive period detection of rail transit intelligent bearing

Abstract

The invention specifically discloses a fault impact deconvolution tracking method for adaptive period detection of a rail transit intelligent bearing, and the method comprises the following steps: obtaining a vibration acceleration signal of the intelligent bearing, setting an expected filter length range, dividing the filter length range into two subintervals, determining three special points, searching an optimal filter coefficient; and calculating and comparing target harmonic energy ratios THER in the blind deconvolution signal envelope spectrums of the three special points, determining the position of the maximum value, determining a filter length search interval, if the interval length reaches a threshold condition, outputting the filter length and the optimal filter coefficient corresponding to the maximum index, and outputting the final filtered blind deconvolution signal. And obtaining an envelope spectrum of the blind deconvolution signal, and evaluating the degradation state of the intelligent bearing. By adopting the technical scheme, vibration signal acquisition, cyclic frequency estimation, recursive search and fault diagnosis are performed, the fault period of the intelligent bearing is positioned, and the calculation efficiency is remarkably improved while the diagnosis precision is ensured.

Description

Technical Field

[0001] The present invention belongs to the technical field of bearing fault diagnosis, and relates to a fault impact deconvolution tracking method for adaptive cycle detection of intelligent bearings in rail transit. Background Art

[0002] As a core device in the ventilation system, the refrigeration fan of rail transit ensures the exhaust air and fire smoke exhaust in the platform, concourse and subway tunnel, and plays a crucial role in the rail transit environment and safe operation. The bearing is the core component of the fan's mechanical transmission system, and most of the fan faults are caused by bearing faults or can be reflected in the operating state of the bearing. Therefore, detecting and analyzing the operating state of the fan bearing is of great significance for the fault diagnosis and operation and maintenance of the entire intelligent bearing in rail transit.

[0003] Currently, vibration-based methods have been widely used in the fault diagnosis of rotating machinery. However, such methods follow the principle of eliminating interference or highlighting fault features during feature extraction and are easily affected by components unrelated to faults. In addition, the actual operation of the refrigeration fan in rail transit is often accompanied by multiple cyclic frequency components, and these frequencies may show a high degree of diversity and dynamics due to different equipment structures, working conditions and fault types. This unpredictability and diversity of the fault cycle greatly increase the difficulty of locating the fault cycle of the fan. Summary of the Invention

[0004] The purpose of the present invention is to provide a fault impact deconvolution tracking method for adaptive cycle detection of intelligent bearings in rail transit to solve the problem of difficult fault cycle location of intelligent bearings in rail transit.

[0005] To achieve the above purpose, the basic solution of the present invention is: A fault impact deconvolution tracking method for adaptive cycle detection of intelligent bearings in rail transit, comprising the following steps:

[0006] S1, deploy vibration acceleration sensors on the intelligent bearings of the rail transit fan, and collect the vibration acceleration signals of the intelligent bearings;

[0007] S2, according to prior knowledge, obtain the rotation frequency of the rail transit fan and the energy distribution interval of its resonance band, and set the expected filter length range;

[0008] S3, according to the trichotomy method, evenly divide the filter length range into two sub-intervals and obtain three special points;

[0009] S4, estimate the cyclic frequency of the vibration acceleration signal through the envelope product spectrum, use the second-order cyclic stationary degree as the objective function of blind deconvolution, and find the optimal filter coefficients;

[0010] S5. Calculate and compare the proportion THER of the target harmonic energy in the envelope spectrum of the blind deconvolution signals at three special points, determine the position where the maximum value is located, perform dynamic pruning using the Alpha-beta strategy, determine the search interval for the filter length. If the interval length reaches the threshold condition, execute step S6; otherwise, return to step S3.

[0011] S6. Output the filter length corresponding to the maximum index and the optimal filter coefficients, output the finally filtered blind deconvolution signal, obtain the envelope spectrum of the blind deconvolution signal, that is, obtain the fault characteristics of the amplified estimated cyclic frequency, diagnose the fault type of the intelligent bearing, and evaluate the degradation state of the intelligent bearing according to the calculated THER.

[0012] The working principle and beneficial effects of this basic solution are as follows: This technical solution uses a vibration acceleration sensor to obtain the vibration acceleration signal of the intelligent bearing. Within the expected filter length range, the blind deconvolution signal is obtained by using the ternary method strategy guided by EHPS and the second-order cyclic stationary degree respectively. Calculate and compare the proportion (THER) of the target harmonic energy in the envelope spectrum of the blind deconvolution signal. THER can be used as an index to evaluate the degradation state of the intelligent bearing, determine the search direction, and thus gradually narrow the search range of the filter length.

[0013] Use the Alpha-beta strategy to achieve dynamic pruning, reduce unnecessary calculations, and finally output the signal corresponding to the maximum proportion of target harmonic energy.

[0014] The global optimal filter coefficients are the optimal filter coefficients at the filter length corresponding to the maximum THER, and then output the final blind deconvolution signal, the fault characteristics of the amplified estimated cyclic frequency. The estimated cyclic frequency can be matched with the characteristic frequency of the intelligent bearing, so that the fault type of the intelligent bearing can be diagnosed. This algorithm significantly improves the calculation efficiency while ensuring the diagnostic accuracy through an intelligent interval division and pruning strategy, calculates the period of the original vibration signal, diagnoses the fault type of the intelligent bearing, and evaluates the degradation state of the intelligent bearing.

[0015] Further, in step S4, the cyclic frequency H(w) of the vibration acceleration signal is estimated through the envelope product spectrum, which is:

[0016]

[0017] where F(w) is the amplitude of the Fourier transform of the vibration acceleration signal x(t); M is the number of harmonics considered, and m is the harmonic order.

[0018] When a local fault occurs in a rolling bearing, the periodic fault impact will excite resonance in the bearing system, thus generating a harmonic-related spectral structure (HRSS) in the envelope spectrum of the signal.

[0019] Further, in step S4, the second-order cyclostationarity is used as the objective function of blind deconvolution, and the method for finding the optimal filter coefficients is as follows:

[0020] Recovering the excitation s0 from the noisy signal x is expressed as:

[0021] s = x * h = (s0 * g) * h ≈ s0

[0022] where s is the estimated excitation source; h is the FIR inverse filter; g is the unknown impulse response of the transfer path;

[0023] The optimization of the inverse filter is expressed as:

[0024]

[0025] where O(f) is the objective function, that is, the ICS2 function; h0 represents the inverse filter, which is essentially an array;

[0026] Define a new index, the second-order cyclostationarity is:

[0027]

[0028] where ICS2 represents the second-order cyclostationarity, k is the sample index; is the second-order complex statistic of the estimated excitation source; e is the natural constant; j is the imaginary unit, is the second-order statistic of the estimated excitation source; N is the length of the signal, also known as the number of signal points, n is 1, 2, 3,..., N; L is the length of the inverse filter h; T s is the period when the fault impact occurs; α is the frequency corresponding to the fault impact in the harmonic product spectrum of the intelligent bearing vibration acceleration signal, that is, the cyclic frequency;

[0029] Calculate the optimal filter coefficients through the eigenvalue algorithm EVA, after discretizing and s, the optimal filter coefficient h is expressed as:

[0030] R XWX h = R XX hλ

[0031] where R XWX is the weighted correlation matrix, W is the weight correlation matrix, R XX is the correlation matrix, and λ is the maximum eigenvalue.

[0032] According to prior knowledge, obtain the rotational frequency of the rail transit fan and the energy distribution interval of its resonance frequency band, set the search interval for the filter length, use the second-order cyclostationarity as the objective function of blind deconvolution, and find the optimal filter coefficients.

[0033] Further, the steps for calculating the proportion THER of the target harmonic energy in the signal envelope spectrum are as follows:

[0034] The rotational frequency f of the fan rotating shaft is:

[0035]

[0036] where r is the rotational speed of the fan rotating shaft;

[0037] The proportion THER of the target harmonic energy K is expressed as:

[0038]

[0039] where Env(·) represents the amplitude of the spectral line in the envelope spectrum line, I is the order of interest, α is the estimated fault frequency, f is the frequency in the envelope spectrum, and f t is the frequency of interest.

[0040] Calculate the THER of the key points in the search interval, determine the search direction, thereby determine the filter parameters corresponding to the optimal index, and obtain the optimal filter.

[0041] Further, use the Alpha-beta strategy for dynamic pruning to determine the search interval of the filter length, specifically:

[0042] Use the Alpha-beta strategy for dynamic pruning to output the blind deconvolution signal corresponding to the maximum proportion of the target harmonic energy;

[0043] Compare the THER of three special points to determine the position where the maximum value is located. If the maximum value is in the left sub-interval, use the ternary search method to search the left length sub-interval; if the maximum value is in the middle, use the ternary search method to search both the left and right sub-length intervals simultaneously; if the maximum value is in the right sub-interval, use the ternary search method to search the right length sub-interval.

[0044] Use the Alpha-beta strategy to achieve dynamic pruning, reduce unnecessary calculations, and finally output the blind deconvolution signal corresponding to the maximum proportion of the target harmonic energy.

[0045] The present invention also provides a fault collection deconvolution tracking system for adaptive periodic detection of intelligent bearings in rail transit, including a data acquisition module and a processing module. The data acquisition module is used to collect the intelligent bearing vibration acceleration signal of the rail transit fan and transmit it to the processing module;

[0046] The processing module executes the method of the present invention to achieve the fault period positioning of the intelligent bearing.

[0047] This system is based on Alpha-beta pruning and uses a multi-scale adaptive maximum second-order cyclic stationary blind deconvolution algorithm to achieve intelligent fault cycle positioning of bearings.

[0048] Furthermore, the data acquisition module includes a vibration acceleration sensor and a DEWETRON multi-channel data acquisition device;

[0049] The vibration acceleration sensor is installed at the base of the rail transit fan and the rear end cover of the motor. The vibration acceleration sensor is connected to the processing module through the DEWETRON multi-channel data acquisition device.

[0050] The data acquisition module has a simple structure and is convenient to use. Description of the Drawings

[0051] Figure 1 is a schematic flow chart of the fault impact deconvolution tracking method for adaptive period detection of intelligent bearings in rail transit of the present invention. Detailed Embodiments

[0052] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the drawings, where the same or similar reference numerals indicate the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below with reference to the drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation of the present invention.

[0053] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "longitudinal", "transverse", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. are based on the orientation or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation of the present invention.

[0054] In the description of the present invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection", and "connection" should be understood in a broad sense. For example, it can be a mechanical connection or an electrical connection, or it can be the communication inside two elements. It can be directly connected or indirectly connected through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms can be understood according to specific situations.

[0055] The present invention discloses a fault impact deconvolution tracking method for adaptive periodic detection of intelligent bearings in rail transit. A multi-scale adaptive maximum second-order cyclic stationary blind deconvolution method based on Alpha-beta pruning is set up. Guided by the envelope product spectrum (EHPS), maximum second-order cyclic stationary blind deconvolution without prior knowledge is realized. At the same time, the target harmonic energy ratio is constructed as an index, and combined with the Alpha-beta pruning strategy and the trichotomy method to optimize and search for the optimal filter parameters, so as to realize the adaptive selection of multi-scale filter parameters.

[0056] As Figure 1 shown, the fault impact deconvolution tracking method for adaptive periodic detection of intelligent bearings in rail transit includes the following steps:

[0057] S1. Deploy vibration acceleration sensors on the intelligent bearings of the rail transit fan, and collect the vibration acceleration signals of the intelligent bearings;

[0058] S2. According to prior knowledge, obtain the rotation frequency of the rail transit fan and the energy distribution interval of its resonance frequency band, and set the expected filter length range; the rotation frequency of the machine = motor speed / 60, and the resonance frequency band is obtained through some dynamic simulation means. When the frequency of the fan system under excitation is close to a certain natural frequency of the fan, the vibration acceleration frequency amplitude of the fan increases significantly.

[0059] S3. According to the trichotomy method, evenly divide the filter length range into two sub-intervals and obtain three special points; according to the trichotomy method, for example, if the set iteration range is [20, 70], then it should be divided according to the starting point 20, the middle point 45, and the ending point 70, these three points.

[0060] S4. Estimate the cyclic frequency of the vibration acceleration signal through the envelope product spectrum (EHPS), use the second-order cyclic stationary degree as the objective function of blind deconvolution, and find the optimal filter coefficients;

[0061] S5. Calculate and compare the target harmonic energy ratio THER in the blind deconvolution envelope spectrum of the signals at the three special points (compare the three special lengths), determine the position where the maximum value is located, use the Alpha-beta strategy for dynamic pruning, determine the filter length search interval. If the interval length reaches the threshold condition, execute step S6, otherwise return to step S3;

[0062] S6 outputs the filter length and the optimal filter coefficients corresponding to the maximum index, outputs the finally filtered blind deconvolution signal, obtains the envelope spectrum of the blind deconvolution signal, that is, obtains the fault characteristics of the amplified estimated cyclic frequency, diagnoses the type of intelligent bearing fault, and evaluates the degradation state of the intelligent bearing according to the calculated THER (calculate THER according to the output signal, THER can be used as the proportion of the fault harmonic frequency, the larger the THER value, the more obvious the fault and the more serious the degradation of the bearing). Realize the fault frequency positioning. This algorithm can directly output the fault frequency. In addition, in the envelope spectrum of the obtained blind deconvolution result, the amplitude of the fault frequency can also reflect the type of bearing fault. Fault period positioning. At the same time, the calculated THER can be used as an index to evaluate the degradation state of the intelligent bearing.

[0063] In a preferred embodiment of the present invention, when a local fault occurs in a rolling bearing, periodic fault shocks will excite resonance in the bearing system, thereby generating a harmonic-related spectral structure (HRSS) in the envelope spectrum of the signal. In step S4, the cyclic frequency H(w) of the vibration acceleration signal is estimated through the envelope product spectrum, which is:

[0064]

[0065] where F(w) is the amplitude of the Fourier transform of the vibration acceleration signal x(t); M is the number of harmonics considered, and m is the harmonic order.

[0066] In HRSS, the amplitudes of the fundamental frequency w0 and its harmonics dominate, which leads to a sharp increase in the corresponding values in EHPS. The position of the global maximum of EHPS can be considered as the fundamental frequency in HRSS.

[0067] In a preferred embodiment of the present invention, in step S4, the second-order cyclostationarity degree is used as the objective function of blind deconvolution, and the method for finding the optimal filter coefficients is:

[0068] Using the index evaluating the cyclostationarity degree as the objective function of blind deconvolution to detect faults in rotating machinery and amplify the fault characteristics of the estimated cyclic frequency of the envelope product spectrum. Recovering the excitation s0 from the noisy signal x is expressed as:

[0069] s = x * h = (s0 * g) * h ≈ s0

[0070] where s is the estimated excitation source; h is the FIR inverse filter; g is the unknown impulse response of the transmission path;

[0071] The optimization of the inverse filter is expressed as:

[0072]

[0073] Among them, O(f) is the objective function, that is, the ICS2 function; h0 represents the inverse filter, which is essentially an array.

[0074] Define a new index, the second-order cyclostationarity degree is:

[0075]

[0076] Among them, ICS2 represents the second-order cyclostationarity degree, and k is the sample index. is to estimate the second-order complex statistical quantity of the excitation source; e is the natural constant; j is the imaginary unit. is to estimate the second-order statistical quantity of the excitation source; N is the length of the signal, also known as the number of signal points, n is 1, 2, 3,..., N; L is the length of the inverse filter h; T s is the period when the fault impact occurs; α is the frequency at which the fault impact corresponding to the harmonic product spectrum in the intelligent bearing vibration acceleration signal occurs, that is, the cyclic frequency.

[0077] Calculate the optimal filter coefficients through the eigenvalue algorithm EVA (which can be calculated using existing technologies). After discretizing and s, the optimal filter coefficient h is expressed as:

[0078] R XWX h = R XX hλ

[0079] Among them, R XWX is the weighted correlation matrix, W is the weight correlation matrix, R XX is the correlation matrix, and λ is the maximum eigenvalue.

[0080] Within the expected filter length range, adopt the trisection method strategy to obtain the blind deconvolution signal respectively guided by EHPS and the second-order cyclostationarity degree, calculate and compare the proportion of the target harmonic energy in the signal envelope spectrum (THER), determine the search direction, and thus gradually narrow the search range of the filter length.

[0081] In a preferred scheme of the present invention, the steps to calculate the proportion of the target harmonic energy in the signal envelope spectrum THER are:

[0082] The rotation frequency f of the fan rotating shaft is:

[0083]

[0084] Among them, r is the rotation speed of the fan rotating shaft.

[0085] The proportion of the target harmonic energy THER K is expressed as:

[0086]

[0087] Among them, Env(·) represents the amplitude of the spectral lines in the envelope spectrum, I is the order of interest, α is the estimated fault frequency, f is the frequency in the envelope spectrum, and f t is the frequency of interest.

[0088] In a preferred embodiment of the present invention, the Alpha-beta strategy is used for dynamic pruning to determine the search range of the filter length, specifically as follows:

[0089] Use the Alpha-beta strategy for dynamic pruning to output the signal corresponding to the maximum proportion of the target harmonic energy;

[0090] Compare the THER of three special points (starting point, midpoint, ending point) to determine the position where the maximum value is located. If the maximum value is in the left sub-interval, use the ternary search method to search the left length sub-interval; if the maximum value is in the middle, use the ternary search method to search both the left and right sub-length intervals simultaneously; if the maximum value is in the right sub-interval, use the ternary search method to search the right length sub-interval.

[0091] Use the Alpha-beta strategy to achieve dynamic pruning, reduce unnecessary calculations, and finally output the signal corresponding to the maximum proportion of the target harmonic energy.

[0092] The globally optimal filter coefficient is the optimal filter coefficient at the filter length corresponding to the maximum THER, and then the final signal is output to amplify the fault characteristics of the estimated cyclic frequency, so that the fault type of the intelligent bearing can be diagnosed, and the THER can evaluate the degradation state of the intelligent bearing. This algorithm significantly improves the calculation efficiency while ensuring the diagnostic accuracy through an intelligent interval division and pruning strategy.

[0093] Use the actually measured vibration acceleration signal of the intelligent fault bearing of rail transit for verification. The bearing model is SKF 6310, and the fault characteristic frequencies at a rotational speed of 3000 r / min are calculated according to the corresponding parameters as shown in Table 1:

[0094] Table 1 Characteristic frequencies of rolling bearings (rotational speed 3000 r / min)

[0095] Bearing component Inner bearing ring Outer bearing ring Rolling element Characteristic frequency (Hz) 247.6 152.4 198.1

[0096] For the rolling element fault data of the intelligent bearing of rail transit, the comparison between the traditional traversal method and the method proposed in the present invention in terms of operation time and number of iterations is shown in Table 2:

[0097] Table 2 Comparison results

[0098] Evaluation index Proposed algorithm Traditional traversal Operation time (s) 601.24 1566.41 Number of iterations (times) 9 25

[0099] The present invention optimizes the selection of the filter length based on Alpha-beta pruning, effectively reducing the computational amount. This algorithm is applicable to a relatively large length range and can improve real-time performance.

[0100] The present invention also provides a fault impact deconvolution tracking system for adaptive periodic detection of intelligent bearings in rail transit, including a data acquisition module and a processing module. The data acquisition module is used to collect the vibration acceleration signals of the intelligent bearings and transmit them to the processing module.

[0101] The processing module executes the method of the present invention to achieve the positioning of the fault period of the intelligent bearing.

[0102] This system is based on Alpha-beta pruning and uses a multi-scale adaptive maximum second-order cyclostationary blind deconvolution algorithm to achieve the positioning of the fault period of the intelligent bearing.

[0103] In a preferred embodiment of the present invention, the data acquisition module includes a vibration acceleration sensor and a DEWETRON multi-channel data acquisition device.

[0104] The vibration acceleration sensor is installed at the base of the rail transit fan and the rear end cover of the motor. The vibration acceleration sensor is electrically connected to the processing module through the DEWETRON multi-channel data acquisition device. The vibration acceleration sensor is used to obtain the vibration acceleration signals of the intelligent bearing, and the DEWETRON multi-channel data acquisition device and the supporting upper computer software are used to monitor and record the multi-physical quantity information accordingly.

[0105] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0106] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the claims and their equivalents.

Claims

1. A fault impact deconvolution tracking method for adaptive periodic detection of intelligent bearings in rail transit, characterized in that It includes the following steps: S1. Deploy a vibration acceleration sensor on the intelligent bearing of the rail transit fan to collect the vibration acceleration signal of the intelligent bearing; S2. According to prior knowledge, obtain the rotation frequency of the rail transit fan and the energy distribution interval of its resonance frequency band, and set the expected filter length range; S3. According to the trichotomy method, evenly divide the filter length range into two sub-intervals and obtain three special points; S4. Estimate the cyclic frequency of the vibration acceleration signal through the envelope product spectrum, use the second-order cyclic stationary degree as the objective function of blind deconvolution, and find the optimal filter coefficients; S5. Calculate and compare the proportion THER of the target harmonic energy in the envelope spectrum of the blind deconvolution signals of the three special points, determine the position where the maximum value is located, use the Alpha-beta strategy for dynamic pruning to determine the filter length search interval. If the interval length reaches the threshold condition, execute step S6; otherwise, return to step S3; S6. Output the filter length corresponding to the maximum index and the optimal filter coefficients, output the finally filtered blind deconvolution signal, obtain the envelope spectrum of the blind deconvolution signal, that is, obtain the fault characteristics of the amplified estimated cyclic frequency, diagnose the fault type of the intelligent bearing, and evaluate the degradation state of the intelligent bearing according to the calculated THER.

2. The fault impact deconvolution tracking method for adaptive periodic detection of intelligent bearings in rail transit according to claim 1, wherein In step S4, the cyclic frequency H(w) of the vibration acceleration signal is estimated through the envelope product spectrum, and it is: where F(w) is the amplitude of the Fourier transform of the vibration acceleration signal x(t); M is the number of harmonics considered, and m is the harmonic order.

3. The fault impact deconvolution tracking method for adaptive periodic detection of intelligent bearings in rail transit according to claim 1, wherein, In step S4, the method of using the second-order cyclic stationary degree as the objective function of blind deconvolution to find the optimal filter coefficients is: The recovery of the excitation s0 from the noisy signal x is expressed as: s = x * h = (s0 * g) * h ≈ s0 where s is the estimated excitation source; h is the FIR inverse filter; g is the unknown impulse response of the transmission path; The optimization of the inverse filter is expressed as: where O(f) is the objective function, that is, the ICS2 function; h0 represents the inverse filter, which is actually an array; Define a new index, the second-order cyclic stationary degree is: Among them, ICS2 represents the second-order cyclic stationary degree, and k is the sample index; to estimate the second-order complex statistics of the excitation source; e is the natural constant; j is the imaginary unit, to estimate the second-order statistics of the excitation source; N is the length of the signal, also known as the number of signal points, n is 1, 2, 3, …, N; L is the length of the inverse filter h; T s is the period when the fault impact occurs; α is the frequency at which the fault impact corresponding to the harmonic product spectrum in the intelligent bearing vibration acceleration signal occurs, that is, the cyclic frequency; The optimal filter coefficients are calculated by the eigenvalue algorithm EVA, and after discretization and s, the optimal filter coefficients h are expressed as: R XWX h = R XX hλ Among them, R XWX is the weighted correlation matrix, W is the weight correlation matrix, and R XX is the correlation matrix, and λ is the maximum eigenvalue.

4. The fault impact deconvolution tracking method for adaptive periodic detection of intelligent bearings in rail transit according to claim 1, wherein The steps for calculating the proportion THER of the target harmonic energy in the signal envelope spectrum are: The rotation frequency f of the fan rotating shaft is: where r is the rotational speed of the fan rotating shaft; Target Harmonic Energy Ratio THER K Expressed as: Among them, Env(·) represents the amplitude of the spectral line in the envelope spectrum, I is the order of interest, α is the estimated fault frequency, f is the frequency in the envelope spectrum, and f t is the frequency of interest.

5. The fault impact deconvolution tracking method for adaptive periodic detection of intelligent bearings for rail transit according to claim 1, characterized in that Use the Alpha-beta strategy for dynamic pruning to determine the filter length search interval, specifically: Use the Alpha-beta strategy for dynamic pruning to output the signal corresponding to the maximum proportion of the target harmonic energy; Compare the THER of the three special points, determine the position where the maximum value is located. If the maximum value is in the left sub-interval, use the trichotomy method to search the left length sub-interval; if the maximum value is in the middle, use the trichotomy method to search both the left and right sub-length intervals; if the maximum value is in the right sub-interval, use the trichotomy method to search the right length sub-interval.

6. A fault impact deconvolution tracking system for adaptive periodic detection of intelligent bearings in rail transit, characterized in that, It includes a data acquisition module and a processing module. The data acquisition module is used to collect the vibration acceleration signal of the intelligent bearing of the rail transit fan and transmit it to the processing module; The processing module executes the method described in any one of claims 1-5 to achieve the positioning of the intelligent bearing fault cycle.

7. The fault impact deconvolution tracking system for adaptive periodic detection of intelligent bearings for rail transit according to claim 6, wherein The data acquisition module includes a vibration acceleration sensor and a DEWETRON multi-channel data acquisition device; The vibration acceleration sensor is installed at the base of the rail transit fan and the rear end cover of the motor. The vibration acceleration sensor is connected to the processing module through the DEWETRON multi-channel data acquisition device.

Images