General sound auscultation and structure knock fault diagnosis method for bearing equipment in coal mine underground

Abstract

The present application relates to the technical field of coal mine underground equipment fault diagnosis, and specifically discloses a coal mine underground bearing equipment general sound auscultation and structure knocking fault diagnosis method, mixed acoustic vibration data composed of bearing fault signals and background turbulent noise is picked up under the equipment running state; the data is injected into a pre-constructed asymmetric bistable potential field, the background noise is used as driving energy, the potential well depth and the asymmetric factor parameters are adjusted, the fault signal is subjected to stochastic resonance transition, and the enhanced fault impact sequence is output; the sequence is subjected to envelope demodulation, time domain averaging and spectrum transformation, the periodic amplitude spectrum line is extracted and compared with a health baseline, an auditory abnormality early warning mark is generated, the knocking position is determined according to the early warning mark after shutdown, the echo waveform is collected and the vibration attenuation duration is analyzed, and the structure loosening or the excessive clearance is determined by comparing with a standard threshold value; the present application realizes online enhancement and quantitative diagnosis of bearing weak faults in a strong noise environment.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology for underground equipment in coal mines, specifically to a general method for diagnosing faults in underground bearing equipment through sound auscultation and structural impact. Background Technology

[0002] In underground coal mine production, rolling bearings are widely used as slewing support components in key equipment such as belt conveyors, main ventilation fans, coal mining machines, and scraper conveyors. Bearings operate under heavy loads, in humid, dusty, and confined roadway environments, making them highly susceptible to pitting, spalling, and cage breakage. In severe cases, this can lead to equipment shutdowns or even safety accidents. Currently, the main methods for detecting bearing failures in coal mines include manual auscultation and structural tapping. Manual auscultation involves maintenance personnel using a metal stethoscope or a long-handled screwdriver against the bearing housing, relying on experience to determine if the bearing's operating sound is abnormal. Structural tapping involves gently tapping the bearing housing or casing with a hammer after the equipment has stopped, inferring whether the structure is loose or cracked based on the crispness or dullness of the echo. These methods are simple to operate, require no complex instruments, and meet underground explosion-proof requirements.

[0003] The existing technology has the following shortcomings: In the harsh acoustic environment of underground coal mines where strong background turbulent noise completely drowns out the weak high-frequency fault signals of bearings, how can we abandon the conventional "filtering and noise reduction" approach and instead use this strong noise as driving energy to actively enhance the fault signal through the asymmetric random resonance effect, thereby achieving early fault warning without stopping the equipment; at the same time, after the sound diagnosis issues a warning, how can we further use quantitative multi-point tapping echo waveform analysis to accurately distinguish between internal rolling wear of the bearing and structural loosening or excessive clearance between the bearing housing mating surfaces, so as to avoid false shutdowns or missed detections due to fuzzy diagnosis. Summary of the Invention

[0004] The purpose of this invention is to provide a universal method for diagnosing bearing equipment in underground coal mines through sound auscultation and structural impact fault diagnosis, so as to solve the problems mentioned above.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A general method for diagnosing bearing equipment failures through sound auscultation and structural impact testing in coal mines includes the following steps:

[0007] S1: Under the operating state of the bearing equipment in the coal mine, continuously collect mixed acoustic and vibration data at predetermined measuring points on the bearing housing shell. The mixed acoustic and vibration data is composed of the superposition of bearing fault signal and background turbulence noise.

[0008] S2: The mixed acoustic and vibration data are injected into a pre-constructed asymmetric bistable potential field. The background turbulence noise is used as the driving energy. The potential well depth parameter and asymmetric factor parameter of the asymmetric bistable potential field are adjusted synchronously to drive the bearing fault signal to undergo random resonance transition and output the fault impact sequence after noise energy enhancement.

[0009] S3: Perform envelope demodulation on the fault impact sequence to obtain the envelope signal, and extract the periodic amplitude spectrum corresponding to the bearing rotation passing frequency and its harmonics from the envelope signal;

[0010] S4: Compare the periodic amplitude spectrum with the baseline spectrum under the same equipment health condition in a frequency band. When the amplitude at any fault characteristic frequency exceeds the preset multiple of the baseline amplitude, generate an auditory abnormality warning sign.

[0011] S5: After the equipment is shut down, the same bearing housing is tapped at multiple points according to the auditory abnormality warning sign. The time-domain waveform of the echo generated by each tap is collected, and the duration of vibration attenuation is analyzed from it. When the duration of vibration attenuation is shorter than the lower limit threshold of the standard attenuation duration under the condition of structural integrity, it is determined that the bearing equipment has a structural loosening or excessive fit clearance fault.

[0012] As a further aspect of the present invention: the pre-construction process of the asymmetric bistable potential field is as follows:

[0013] Collect mixed acoustic and vibration data of the bearing under test when it is running under known health conditions, and use the mixed acoustic and vibration data as the reference input;

[0014] Multiple stochastic resonance simulations were performed on the reference input. In each resonance simulation, the potential well depth parameter and the asymmetry factor parameter were adjusted respectively, and the signal-to-noise ratio of the output signal was recorded.

[0015] Select the potential well depth parameter value and the asymmetric factor parameter value that maximize the signal-to-noise ratio, and fix the potential well depth parameter value and the asymmetric factor parameter value as the initial construction parameters of the asymmetric bistable potential field.

[0016] As a further aspect of the present invention: based on the initial construction parameters, a fault impact sequence enhanced with noise energy is output, specifically including:

[0017] Based on the potential well depth parameter value and the asymmetric factor parameter value, the currently acquired mixed acoustic and vibration data are injected into the asymmetric bistable potential field to obtain a preliminary fault impact sequence.

[0018] Calculate the time-domain kurtosis value of the initial fault impact sequence. Based on the deviation between the time-domain kurtosis value and the preset target kurtosis interval, fine-tune the potential well depth parameter in both directions near the benchmark until the kurtosis value falls into the target interval.

[0019] While maintaining the fine-tuned potential well depth parameter, the asymmetric factor parameter is adjusted by a variable step size iterative optimization method until the spectral line amplitude at the expected fault characteristic frequency in the power spectrum of the output signal reaches a local maximum.

[0020] The fine-tuned potential well depth parameter and the adjusted asymmetry factor parameter are locked together, and the fault impact sequence after noise energy enhancement is continuously output.

[0021] As a further aspect of the present invention: the method of adjusting the asymmetric factor parameters using a variable step-size iterative optimization specifically includes:

[0022] Starting from the current asymmetric factor parameter value, set the initial step size;

[0023] Apply a step-size perturbation in the direction of parameter increase, calculate the spectral amplitude at the expected fault characteristic frequency in the power spectrum of the output signal after perturbation, if the spectral amplitude increases, maintain the perturbation direction and increase the step size, if it decreases, reverse the perturbation and decrease the step size;

[0024] Repeat the process of applying perturbations and comparing amplitudes until the step size is reduced to the preset lower limit, and two consecutive perturbations do not cause an increase in the spectral amplitude. At this point, lock the final value of the asymmetric factor parameter.

[0025] As a further aspect of the present invention: the extraction process of the periodic amplitude spectral lines is as follows:

[0026] The number of sampling points in a complete cycle is calculated based on the bearing rotation frequency, and the fault impact sequence is divided into continuous multi-frame data according to the number of sampling points.

[0027] The periodic average waveform is obtained by summing the corresponding points of multiple frames of data in the time domain and dividing by the number of frames.

[0028] Perform a discrete Fourier transform on the periodic average waveform to extract the amplitude at the corresponding bearing rotation frequency and its harmonics, which are then used as periodic amplitude spectral lines.

[0029] As a further aspect of the present invention: the generation of the hearing abnormality warning identifier specifically includes:

[0030] Divide the amplitude at each fault characteristic frequency in the periodic amplitude spectrum by the amplitude of its corresponding baseline spectrum to obtain a normalized ratio sequence.

[0031] An exponentially weighted moving average is applied to each ratio in the normalized ratio sequence to obtain a smoothed ratio sequence.

[0032] The frequency bands corresponding to each ratio in the smooth ratio sequence that exceeds a preset multiple are accumulated and counted. When the accumulated count reaches a preset threshold, a hearing abnormality warning sign is generated.

[0033] As a further aspect of the present invention: the determination of the duration of vibration attenuation specifically includes:

[0034] Based on the fault frequency band direction indicated by the hearing abnormality warning sign, select three mutually perpendicular tapping points on the end face and side of the bearing housing.

[0035] Collect the peak amplitude of the first half-cycle and the peak amplitude of the second half-cycle of the echo time-domain waveform at each impact point, and calculate the ratio of the two as the attenuation characteristic value.

[0036] The duration of vibration attenuation is obtained by taking the arithmetic mean of the attenuation characteristic values ​​of the three impact points and multiplying it by the total duration of the echo waveform.

[0037] As a further aspect of the present invention: the selection of three mutually perpendicular striking points on the end face and side face of the bearing housing based on the fault frequency band direction indicated by the hearing abnormality warning sign specifically includes:

[0038] The fault characteristic frequency corresponding to the auditory abnormality warning sign is analyzed, and the fault characteristic frequency is compared with the pre-stored bearing housing structure modal frequency to obtain the fault excitation direction vector.

[0039] The radial orientation of the first impact point is determined by projecting the fault excitation direction vector onto the bearing housing end face, the tangential orientation of the second impact point is determined in the vertical direction of the end face, and the position of the third impact point is determined along the axial direction on the side of the bearing housing.

[0040] The coordinate deviations of the three impact points from the original reference point of the bearing housing were measured respectively to complete the selection of the three points.

[0041] The beneficial effects of this invention are:

[0042] (1) This invention utilizes a pre-constructed asymmetric bistable potential field to drive the inherent strong background turbulent noise during the operation of underground equipment, causing weak bearing fault signals to undergo random resonant transitions and be effectively amplified. This allows for the extraction of bearing fault characteristic frequencies from strong noise masking without requiring shutdown or relying on expensive filtering equipment. This method is particularly suitable for key coal mine equipment that cannot be easily shut down, such as main ventilation fans and coal mining machines, enabling early fault warnings during operation and reducing unplanned downtime caused by sudden faults.

[0043] (2) After issuing an abnormal warning through sound auscultation, this invention further performs multi-point structural tapping while the machine is stopped. By analyzing the attenuation characteristic value and vibration attenuation duration of the echo waveform and comparing it with the baseline under good equipment conditions, it can accurately distinguish between rolling friction wear inside the bearing and structural loosening or excessive clearance between the bearing housing mating surfaces. This combined diagnostic method of "sound diagnosis + tapping" effectively compensates for the inability of a single auscultation method to determine structural loosening defects, improves the accuracy of fault location, and provides a more reliable basis for maintenance decisions in coal mines. Attached Figure Description

[0044] The invention will now be further described with reference to the accompanying drawings.

[0045] Figure 1 This is a flowchart of the method of the present invention;

[0046] Figure 2 This is a flowchart of the pre-construction process of the asymmetric bistable potential field in this invention;

[0047] Figure 3 This is a flowchart of the extraction process of periodic amplitude spectral lines in this invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Please see Figure 1 As shown, this invention provides a general method for diagnosing bearing equipment in underground coal mines through sound auscultation and structural impact testing, comprising the following steps:

[0050] S1: Under the operating state of the bearing equipment in the coal mine, continuously collect mixed acoustic and vibration data at predetermined measuring points on the bearing housing shell. The mixed acoustic and vibration data is composed of the superposition of bearing fault signal and background turbulence noise.

[0051] S2: The mixed acoustic and vibration data are injected into a pre-constructed asymmetric bistable potential field. The background turbulence noise is used as the driving energy. The potential well depth parameter and asymmetric factor parameter of the asymmetric bistable potential field are adjusted synchronously to drive the bearing fault signal to undergo random resonance transition and output the fault impact sequence after noise energy enhancement.

[0052] S3: Perform envelope demodulation on the fault impact sequence to obtain the envelope signal, and extract the periodic amplitude spectrum corresponding to the bearing rotation passing frequency and its harmonics from the envelope signal;

[0053] S4: Compare the periodic amplitude spectrum with the baseline spectrum under the same equipment health condition in a frequency band. When the amplitude at any fault characteristic frequency exceeds the preset multiple of the baseline amplitude, generate an auditory abnormality warning sign.

[0054] S5: After the equipment is shut down, the same bearing housing is tapped at multiple points according to the auditory abnormality warning sign. The time-domain waveform of the echo generated by each tap is collected, and the duration of vibration attenuation is analyzed from it. When the duration of vibration attenuation is shorter than the lower limit threshold of the standard attenuation duration under the condition of structural integrity, it is determined that the bearing equipment has a structural loosening or excessive fit clearance fault.

[0055] In S1, under the operating condition of the bearing equipment in the underground coal mine, mixed acoustic and vibration data are continuously collected at predetermined measuring points on the bearing housing shell. The mixed acoustic and vibration data consists of the superposition of bearing fault signals and background turbulent noise, specifically including:

[0056] Pre-treatment of the measuring point and preparation of tools: In the underground roadway of the coal mine, the bearing equipment to be tested is kept in normal load operation. The operator uses a copper scraper to remove the coal sludge and rust layer from the surface of the bearing housing, exposing the metal substrate. Then, using a handheld angle grinder fitted with a 60-mesh abrasive wheel, a bright circular measuring point with a diameter of 20 mm is ground on the flat area corresponding to the radial center of the outer ring of the bearing on the housing. After grinding, the surface of the measuring point is wiped with a cotton cloth soaked in anhydrous ethanol to remove oil and grinding debris.

[0057] Sensor Installation and Connection: Take an intrinsically safe piezoelectric accelerometer with a nominal sensitivity of 100 mV / g and a frequency response range of 0.5 Hz to 10 kHz. Apply a uniform layer of industrial-grade silicone grease (approximately 1 mm thick) to the bottom surface of the sensor. Attach the bottom surface of the sensor vertically to the center of the bright measurement point and secure it using a magnetic mount or quick-drying adhesive, ensuring there is no relative looseness between the sensor and the bearing housing.

[0058] Data acquisition parameter settings: Connect the sensor's output leads to an intrinsically safe data logger with 24-bit analog-to-digital conversion accuracy. Set the sampling frequency to 25.6 kHz, the sampling depth to continuous acquisition mode, and the sampling duration to 120 seconds on the logger. Enable the logger's pre-amplifier anti-aliasing filter with a cutoff frequency of 10 kHz to suppress high-frequency aliasing interference.

[0059] Continuous acquisition of hybrid acoustic and vibration data: The data logger is activated to continuously acquire vibration acceleration signals at the measuring points while the bearing equipment is in continuous operation. This signal is composed of the fault signal generated by the internal operation of the bearing and the background turbulence noise generated by the airflow impacting the casing of the fan duct or tunnel. The logger converts the acquired analog signal into digital format and stores it in the internal non-volatile memory as hybrid acoustic and vibration data for subsequent processing.

[0060] Please see Figure 2 As shown, in S2, the mixed acoustic and vibration data are injected into a pre-constructed asymmetric bistable potential field. Background turbulent noise is used as the driving energy, and the potential well depth parameter and asymmetry factor parameter of the asymmetric bistable potential field are adjusted synchronously to drive the bearing fault signal to undergo random resonant transitions. The output is a fault impact sequence enhanced by noise energy, specifically including:

[0061] Offline data preparation for constructing an asymmetric bistable potential field: The offline construction process is performed when the bearing under test is first put into use or confirmed to be in good condition after maintenance. Following the method described in step S1, the operator continuously collects 120 seconds of mixed acoustic and vibration data under normal operating conditions and marks this data segment as the health baseline data. In the health baseline data, the amplitude of the bearing fault signal is negligible, and background turbulence noise dominates.

[0062] Stochastic resonance simulation and signal-to-noise ratio calculation: Health baseline data is used as input to an asymmetric bistable potential field model for numerical simulation. This model is described by the Langevin equation and solved using the fourth-order Runge-Kutta method. A two-dimensional grid scan is performed with the potential well depth parameter 'a' ranging from 0.5 to 5.0 and the asymmetric factor parameter 'λ' ranging from -0.5 to 0.5, with step sizes of 0.1 and 0.02, respectively. In each scan, the power spectrum of the output signal is calculated, and the corresponding bearing rotation frequency is extracted from it. spectral line amplitude at and calculate with The average amplitude of the noise floor centered at a width of 5 times the frequency resolution The formula for calculating the signal-to-noise ratio is as follows: ;

[0063] in, The signal-to-noise ratio of the output signal. For the frequency of bearing rotation Spectral line amplitude at that location The total number of frequency points selected for the noise floor (value is 10). For the first The formula represents the spectral amplitude at each noise frequency point. It characterizes the ratio of fault characteristic energy to background noise energy.

[0064] Selection of initial construction parameters: For each set of parameters and Calculate the corresponding Value. After traversing all scan points, select the one that makes... The potential well depth parameter value corresponding to the maximum value. Asymmetric factor parameter values The parameters are fixed as the initial construction parameters for the asymmetric bistable potential field. The resulting potential field, under healthy conditions, best highlights the fault characteristics of the output signal, serving as a benchmark for subsequent online diagnosis.

[0065] Online injection and initial fault impact sequence acquisition: When the equipment is actually running and a potential fault exists, collect the current mixed acoustic and vibration data according to step S1, and record it as real-time data. Based on the initial construction parameters... and As a baseline, real-time data was injected into an asymmetric bistable potential field, and a preliminary fault impact sequence was obtained using the same numerical solution method as the offline simulation. The weak bearing fault signal in this sequence was initially amplified due to the stochastic resonance effect, but the potential field parameters have not yet been optimized for the current operating conditions.

[0066] Bidirectional fine-tuning of temporal kurtosis of potential well depth parameter: Calculation of temporal kurtosis values ​​for the initial fault impact sequence The formula for calculating temporal kurtosis is: ;

[0067] in, This represents the total number of sampling points in the sequence, with a value of 4096. For the first in the sequence The amplitude of each sampling point This is the arithmetic mean of the sequence. The preset target kurtosis interval is [2.8, 3.2]. Calculate the current... Deviation of value from the interval midpoint 3.0 .like If the value is greater than 0, then the potential well depth parameter will be... exist Based on this, adjust in decreasing steps of 0.05; if If the kurtosis is less than 0, adjust in the direction of increasing. After each adjustment, recalculate the kurtosis of the output sequence until the kurtosis value falls within the interval [2.8, 3.2]. Record the potential well depth parameter at this point. .

[0068] Variable step-size iterative optimization of asymmetric factor parameters: keeping the potential well depth parameter constant. Remain unchanged, with the current asymmetric factor parameter value. As a starting point, set the initial step size. Apply a step-size perturbation along the direction of parameter increase, i.e. The real-time data is then injected back into the potential field, and the amplitude of the spectral line at the expected fault characteristic frequency f0 in the power spectrum of the output signal is calculated. .like If the value is greater than the previous recorded value, the perturbation direction remains positive, and the step size is updated to [value]. ;like If the value is less than the previous recorded value, the perturbation direction is changed to negative, and the step size is updated. Then, a perturbation is applied in the new direction, and the above comparison and step size adjustment process is repeated. When the step size is reduced to the preset lower limit of 0.0005, and two consecutive perturbations fail to increase the spectral amplitude, the iteration stops, and the current asymmetric factor parameter value is locked. .Will and As the final parameter, the fault impact sequence after noise energy enhancement is continuously output for subsequent processing steps.

[0069] Please see Figure 3 As shown, in S3, envelope demodulation is performed on the fault impact sequence to obtain the envelope signal, and periodic amplitude spectral lines corresponding to the bearing rotation passing frequency and its harmonics are extracted from the envelope signal, specifically including:

[0070] Envelope demodulation preprocessing: The fault impulse sequence output in step S2 is used as input and subjected to Hilbert transform to obtain the analytic signal of the sequence. Then, the magnitude of the analytic signal is calculated, which is the envelope signal of the fault impulse sequence. This envelope signal removes the high-frequency carrier component from the original signal and retains the low-frequency impulse envelope waveform generated by the bearing fault.

[0071] Determine frame length and segment frame data: Obtain the real-time rotational speed of the bearing under test and calculate the bearing's rotational frequency, which is equal to the number of revolutions per second of the bearing shaft. Divide the sampling frequency set during data acquisition by the aforementioned rotational frequency to obtain the number of sampling points corresponding to a single complete cycle, which is recorded as the length of that cycle. Starting from the beginning of the envelope signal, sequentially extract continuous data segments according to this length value; each segment is called a frame. If the number of data points remaining at the end is less than a complete cycle, discard that segment. Record the total number of frames obtained, assuming its value is a positive integer.

[0072] The periodic average waveform is obtained by time-domain superposition averaging: All segmented frame data are accumulated at corresponding points in the time domain. For the first sampling point in each frame, the amplitudes of the first sampling points in all frames are summed to obtain the cumulative sum. Similarly, the sums are accumulated for the second, third, and subsequent sampling points corresponding to the period length. Each cumulative sum is then divided by the number of frames to obtain the average amplitude at each corresponding position. These average amplitudes are arranged sequentially to form a new waveform sequence, called the periodic average waveform. This waveform reflects the average characteristics of the fault impact sequence within a complete cycle, significantly suppressing random noise that is asynchronous with the bearing rotation frequency.

[0073] Extracting the periodic amplitude spectrum: Perform a Discrete Fourier Transform on the periodic average waveform and calculate its spectrum. Locate the frequency point in the spectrum that corresponds to the bearing's rotational frequency and record its amplitude. Then, sequentially locate frequencies at 2, 3, up to 5 times the rotational frequency and record the amplitude at each harmonic. Combine these frequencies and their corresponding amplitudes into a data set, which is output as the periodic amplitude spectrum. This spectrum highlights the fault characteristic frequency components related to the bearing's rotational period.

[0074] In S4, the periodic amplitude spectrum is compared band-by-band with the baseline spectrum of the same equipment under healthy conditions. When the amplitude at any fault characteristic frequency exceeds a preset multiple of the baseline amplitude, an auditory abnormality warning indicator is generated, specifically including:

[0075] Obtain the normalized ratio sequence: With the equipment in a healthy state, a set of baseline spectra is collected and processed according to the same parameter settings as steps S1 to S3. These baseline spectra record the amplitude at the bearing's rotational frequency and five fault characteristic frequencies (2nd to 5th harmonics). The amplitude at each fault characteristic frequency in the periodic amplitude spectrum output in step S3 is divided by the amplitude at the corresponding frequency in the baseline spectrum, yielding five ratios. These five ratios are arranged in ascending order of frequency to form the normalized ratio sequence. For example, if the baseline amplitude at a certain frequency is zero or close to zero, the ratio corresponding to that frequency is set to a preset upper limit of 10.

[0076] Exponentially Weighted Moving Average Smoothing: Each ratio in the normalized ratio sequence is processed using an exponentially weighted moving average to obtain a smoothed ratio sequence. The specific calculation method is as follows: A smoothing coefficient of 0.3 is set. For the first ratio in the sequence, its smoothed value is simply the ratio itself. For the second and subsequent ratios, 0.3 times the current ratio is added to 0.7 times the previous smoothed value to obtain the smoothed value for the current ratio. This process is repeated for all five ratios, resulting in five smoothed values, which are then arranged in their original order to form a smoothed ratio sequence. This process can suppress instantaneous fluctuations and reflect the long-term trend of ratio changes.

[0077] Cumulative Counting and Warning Flag Generation: The warning multiplier is set to 1.8, and the preset threshold for cumulative counting is 3 times. Each ratio in the smoothed ratio sequence is checked one by one. If the ratio exceeds 1.8, the corresponding frequency band is marked as an abnormal event. The number of abnormal events occurring in the same frequency band within three consecutive acquisition cycles is accumulated to obtain the cumulative count value for that frequency band. When the cumulative count value of any frequency band reaches 3 times, an auditory abnormality warning flag is generated. This flag is stored in the data logger in the form of a digital code and is used to indicate the bearing frequency band direction with potential fault risk. If the cumulative count values ​​of all frequency bands do not reach the threshold, no warning flag is generated, and data monitoring continues in the next acquisition cycle.

[0078] In S5, after the equipment stops, multiple taps are performed on the same bearing housing based on the auditory abnormality warning sign. The time-domain waveform of the echo generated by each tap is collected, and the duration of vibration attenuation is analyzed from it. When the duration of vibration attenuation is shorter than the lower limit threshold of the standard attenuation duration under the condition of structural integrity, it is determined that the bearing equipment has a structural loosening or excessive fit clearance fault, specifically including:

[0079] Selection of striking point and preparation of tools: After the equipment is completely shut down and powered off, the operator obtains the auditory abnormality warning label generated in step S4. This label contains a fault characteristic frequency value, indicating the abnormal bearing frequency band direction. The operator obtains the mode shape diagrams of the bearing housing at five low-frequency modal frequencies (50 Hz, 80 Hz, 120 Hz, 160 Hz, 200 Hz) in advance through finite element simulation or experimental modal analysis, and stores the principal vibration direction vectors corresponding to each modal frequency in the recorder. The fault frequency in the warning label is compared with the above modal frequencies one by one, and the principal vibration direction vector corresponding to the mode with the smallest absolute value of frequency difference (less than 5 Hz) is selected as the fault excitation direction vector.

[0080] Specifically, after analyzing the fault characteristic frequency corresponding to the auditory abnormality warning sign, this frequency value is compared one by one with the modal frequency values ​​of the bearing housing obtained in advance through finite element simulation or experimental modal analysis. The modal order with the smallest absolute value of the frequency difference and less than five Hz is selected. For this modal order, the three-dimensional displacement component values ​​of multiple nodes on the bearing housing surface under the modal vibration mode are pre-stored. The displacement component of each node consists of radial, tangential, and axial components. The fault excitation direction vector is expressed as follows: taking the original reference point of the bearing housing as the origin, the displacement components of all nodes under the selected mode are vector synthesized to obtain the resultant displacement direction. Then, the resultant displacement direction is normalized to obtain the three directional component values, which correspond to the radial, tangential, and axial proportional relationships, respectively. This proportional relationship is used as the fault excitation direction vector.

[0081] The specific locations of the three impact points were determined as follows: Based on the projection of the fault excitation direction vector onto the bearing housing end face (a plane perpendicular to the bearing axis), the radial orientation of the first impact point was determined: the intersection of the projected direction line and the edge of the end face. A tangential line was drawn through this intersection on the end face to determine the tangential orientation of the second impact point, located at the edge of the end face and 90 degrees away from the first impact point. On the side of the bearing housing (a plane parallel to the bearing axis), the midpoint of the bearing housing length along the bearing axis was taken as the axial position of the third impact point. The three-dimensional coordinate deviations of the three impact points from the pre-set original reference points on the bearing housing (e.g., the center of the bearing housing mounting bolt holes) were measured using a steel ruler, and each deviation value was recorded to ensure subsequent reproduction.

[0082] Impact Operation and Echo Waveform Acquisition: The operator holds a 200-gram, brass-made explosion-proof hammer and applies a single transient impact to the first, second, and third impact points. The impact force is controlled to ensure that the hammer head rebounds no more than 5 centimeters. Using the same intrinsically safe accelerometer and data logger as in step S1, the echo time-domain waveform is continuously acquired for 0.5 seconds following each impact, with a sampling frequency maintained at 25.6 kHz. Each impact point is repeated three times, and the waveform with the best repeatability is used for subsequent analysis.

[0083] Calculation of attenuation characteristic value: For the echo time-domain waveform acquired at each impact point, starting from the impact trigger moment, search for the first positive peak and record its amplitude, called the first half-cycle peak amplitude. Then search for the first negative peak that appears after this peak and record its absolute value as the second half-cycle peak amplitude. Divide the second half-cycle peak amplitude by the first half-cycle peak amplitude; the ratio obtained is defined as the attenuation characteristic value of that impact point. The larger the ratio, the slower the vibration attenuation; the smaller the ratio, the faster the vibration attenuation. Calculate the attenuation characteristic values ​​for the three impact points respectively.

[0084] Vibration attenuation duration and fault determination: The attenuation characteristic values ​​of the three impact points are summed and divided by 3 to obtain the average attenuation characteristic value. The total duration of the echo waveform in the time domain, measured from the impact trigger moment until the amplitude attenuates to 5% of its maximum value, is measured in milliseconds. The average attenuation characteristic value is multiplied by the total duration of the echo waveform to obtain the vibration attenuation duration. The vibration attenuation duration measured in the same way beforehand under good equipment conditions (e.g., after a new bearing is installed) is multiplied by 0.6 to obtain a lower threshold. If the currently calculated vibration attenuation duration is less than this lower threshold, the bearing equipment is determined to have structural loosening or excessive clearance; otherwise, the structure is considered intact. The determination result is displayed in text form on the data logger screen.

[0085] The working principle of this invention is as follows: During equipment operation, mixed acoustic and vibration data, consisting of a superposition of bearing fault signals and background turbulent noise, is continuously acquired at measurement points on the bearing housing housing. This data is injected into an asymmetric bistable potential field pre-constructed using health status data to maximize the signal-to-noise ratio. Using background turbulent noise as the driving energy, the potential well depth parameter is first finely adjusted bidirectionally based on time-domain kurtosis. Then, the asymmetric factor parameter is iteratively optimized using variable step size to induce random resonant transitions in the bearing fault signal, outputting a fault impact sequence with enhanced noise energy. After envelope demodulation of this sequence, it is divided by bearing rotation frequency and time-domain superposition averaging is performed to extract periodic amplitude spectra. The frequency band normalized ratio and exponentially weighted moving average of the line and the healthy baseline spectrum are compared. When the smoothing ratio at any fault characteristic frequency exceeds the preset multiple three times in a row, an auditory abnormality warning sign is generated. After the equipment is shut down, the bearing seat mode shape is compared with the fault frequency in the warning sign. Three mutually perpendicular impact points are determined on the end face and side. The echo waveforms at each point are collected and the ratio of the peak amplitude of the first half cycle to the second half cycle is calculated as the attenuation characteristic value. The arithmetic mean of the three attenuation characteristic values ​​is multiplied by the total echo duration to obtain the vibration attenuation duration. If this duration is less than the lower limit threshold of the standard attenuation duration under good condition, it is determined that there is a structural loosening or excessive fit clearance fault.

[0086] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A general method for diagnosing bearing equipment in underground coal mines using sound auscultation and structural impact testing, characterized in that... Includes the following steps: S1: Under the operating state of the bearing equipment in the coal mine, continuously collect mixed acoustic and vibration data at predetermined measuring points on the bearing housing shell. The mixed acoustic and vibration data is composed of the superposition of bearing fault signal and background turbulence noise. S2: The mixed acoustic and vibration data are injected into a pre-constructed asymmetric bistable potential field. The background turbulence noise is used as the driving energy. The potential well depth parameter and asymmetric factor parameter of the asymmetric bistable potential field are adjusted synchronously to drive the bearing fault signal to undergo random resonance transition and output the fault impact sequence after noise energy enhancement. S3: Perform envelope demodulation on the fault impact sequence to obtain the envelope signal, and extract the periodic amplitude spectrum corresponding to the bearing rotation passing frequency and its harmonics from the envelope signal; S4: Compare the periodic amplitude spectrum with the baseline spectrum under the same equipment health condition in a frequency band. When the amplitude at any fault characteristic frequency exceeds the preset multiple of the baseline amplitude, generate an auditory abnormality warning sign. S5: After the equipment is shut down, the same bearing housing is tapped at multiple points according to the auditory abnormality warning sign. The time-domain waveform of the echo generated by each tap is collected, and the duration of vibration attenuation is analyzed from it. When the duration of vibration attenuation is shorter than the lower limit threshold of the standard attenuation duration under the condition of structural integrity, it is determined that the bearing equipment has a structural loosening or excessive fit clearance fault.

2. The general sound auscultation and structural impact fault diagnosis method for bearing equipment in coal mines according to claim 1, characterized in that, The pre-construction process of the asymmetric bistable potential field is as follows: Collect mixed acoustic and vibration data of the bearing under test when it is running under known health conditions, and use the mixed acoustic and vibration data as the reference input; Multiple stochastic resonance simulations were performed on the reference input. In each resonance simulation, the potential well depth parameter and the asymmetry factor parameter were adjusted respectively, and the signal-to-noise ratio of the output signal was recorded. Select the potential well depth parameter value and the asymmetric factor parameter value that maximize the signal-to-noise ratio, and fix the potential well depth parameter value and the asymmetric factor parameter value as the initial construction parameters of the asymmetric bistable potential field.

3. The general sound auscultation and structural impact fault diagnosis method for underground bearing equipment in coal mines according to claim 2, characterized in that, Based on the initial construction parameters, the fault impact sequence after noise energy enhancement is output, specifically including: Based on the potential well depth parameter and asymmetric factor parameter, the currently acquired mixed acoustic and vibration data are injected into an asymmetric bistable potential field to obtain a preliminary fault impact sequence. Calculate the time-domain kurtosis value of the initial fault impact sequence. Based on the deviation between the time-domain kurtosis value and the preset target kurtosis interval, fine-tune the potential well depth parameter in both directions near the benchmark until the kurtosis value falls into the target interval. While maintaining the fine-tuned potential well depth parameter, the asymmetric factor parameter is adjusted by a variable step size iterative optimization method until the spectral line amplitude at the expected fault characteristic frequency in the power spectrum of the output signal reaches a local maximum. The fine-tuned potential well depth parameter and the adjusted asymmetry factor parameter are locked together, and the fault impact sequence after noise energy enhancement is continuously output.

4. The general sound auscultation and structural impact fault diagnosis method for underground bearing equipment in coal mines according to claim 3, characterized in that, The method of adjusting the asymmetric factor parameters using a variable step-size iterative optimization specifically includes: Starting from the current asymmetric factor parameter value, set the initial step size; Apply a step-size perturbation in the direction of parameter increase, calculate the spectral amplitude at the expected fault characteristic frequency in the power spectrum of the output signal after perturbation, if the spectral amplitude increases, maintain the perturbation direction and increase the step size, if it decreases, reverse the perturbation and decrease the step size; Repeat the process of applying perturbations and comparing amplitudes until the step size is reduced to the preset lower limit, and two consecutive perturbations do not cause an increase in the spectral amplitude. At this point, lock the final value of the asymmetric factor parameter.

5. The general method for diagnosing bearing equipment in underground coal mines by sound auscultation and structural impact as described in claim 1, characterized in that, The extraction process of the periodic amplitude spectral lines is as follows: The number of sampling points in a complete cycle is calculated based on the bearing rotation frequency, and the fault impact sequence is divided into continuous multi-frame data according to the number of sampling points. The periodic average waveform is obtained by summing the corresponding points of multiple frames of data in the time domain and dividing by the number of frames. Perform a discrete Fourier transform on the periodic average waveform to extract the amplitude at the corresponding bearing rotation frequency and its harmonics, which are then used as periodic amplitude spectral lines.

6. The general method for diagnosing bearing equipment in underground coal mines by sound auscultation and structural impact as described in claim 1, characterized in that, The generation of hearing abnormality warning indicators specifically includes: Divide the amplitude at each fault characteristic frequency in the periodic amplitude spectrum by the amplitude of its corresponding baseline spectrum to obtain a normalized ratio sequence. An exponentially weighted moving average is applied to each ratio in the normalized ratio sequence to obtain a smoothed ratio sequence. The frequency bands corresponding to each ratio in the smooth ratio sequence that exceeds a preset multiple are accumulated and counted. When the accumulated count reaches a preset threshold, a hearing abnormality warning sign is generated.

7. The general method for diagnosing bearing equipment in underground coal mines by sound auscultation and structural impact as described in claim 1, characterized in that, The analyzed vibration attenuation duration specifically includes: Based on the fault frequency band direction indicated by the hearing abnormality warning sign, select three mutually perpendicular tapping points on the end face and side of the bearing housing. Collect the peak amplitude of the first half-cycle and the peak amplitude of the second half-cycle of the echo time-domain waveform at each impact point, and calculate the ratio of the two as the attenuation characteristic value. The duration of vibration attenuation is obtained by taking the arithmetic mean of the attenuation characteristic values ​​of the three impact points and multiplying it by the total duration of the echo waveform.

8. The general method for diagnosing bearing equipment in underground coal mines by sound auscultation and structural impact as described in claim 7, characterized in that, Based on the fault frequency band direction indicated by the hearing abnormality warning sign, three mutually perpendicular tapping points are selected on the end face and side face of the bearing housing, specifically including: The fault characteristic frequency corresponding to the auditory abnormality warning sign is analyzed, and the fault characteristic frequency is compared with the pre-stored bearing housing structure modal frequency to obtain the fault excitation direction vector. The radial orientation of the first impact point is determined by projecting the fault excitation direction vector onto the bearing housing end face, the tangential orientation of the second impact point is determined in the vertical direction of the end face, and the position of the third impact point is determined along the axial direction on the side of the bearing housing. The coordinate deviations of the three impact points from the original reference point of the bearing housing were measured respectively to complete the selection of the three points.

Images