A sound source positioning method for transmission device fault identification

Abstract

The application discloses a sound source positioning method for transmission device fault identification, and relates to the technical field of fault identification, and the technical solution points of the method comprise the following steps: acquiring multi-path sound signal data of a transmission device, performing PTP hardware clock synchronization processing on the multi-path sound signal data to obtain synchronized sound signal data; performing noise reduction processing on the synchronized sound signal data to obtain a fault feature enhanced signal, setting a sound source positioning reference point according to the fault feature enhanced signal; and judging the fault feature enhanced signal according to the sound source positioning reference point to obtain a time delay deviation coefficient, a spatial distribution deviation coefficient, a time delay abnormal reference point and a spatial abnormal reference point; the method solves the time delay calculation deviation problem caused by the millisecond-level error of software synchronization, provides an accurate time reference for subsequent sound source positioning, improves the positioning accuracy from the meter level to the centimeter level, and meets the core requirement of accurate positioning of transmission device faults.

Description

Technical Field

[0001] This invention relates to the field of fault identification technology, and more specifically, to a sound source localization method for fault identification in transmission devices. Background Technology

[0002] As a key component of core equipment in industries such as metallurgy, mining, and power, the operating status of transmission devices directly determines the safety and stability of the entire production line. Gears and bearings within transmission devices operate under harsh conditions of high speed, heavy load, high temperature, high dust, and strong vibration for extended periods, making them prone to wear, cracks, and tooth breakage. Failure to identify and locate these faults in a timely manner can lead to equipment downtime or even safety accidents, resulting in significant economic losses. Therefore, accurate identification and location of transmission device faults is a core requirement for ensuring the safe operation and maintenance of industrial equipment. With the development of acoustic detection technology, fault location methods based on acoustic signals are increasingly being applied to transmission device fault identification. These methods achieve non-contact detection by collecting the acoustic signals of equipment operation, eliminating the need for downtime installation and adapting to harsh operating conditions.

[0003] However, existing sound source localization technologies suffer from insufficient clock synchronization accuracy at each acquisition node during the acquisition of multiple sound signals. Software synchronization is often employed, resulting in millisecond-level synchronization errors. This leads to deviations in the calculation of the time difference of arrival of sound signals, directly reducing the accuracy of sound source localization. Simultaneously, industrial sites are characterized by strong background interference such as power frequency noise and fan noise. Traditional noise reduction methods struggle to effectively filter out steady-state interference and non-steady-state random noise, causing fault characteristic signals to be submerged and making it impossible to extract effective fault information. Furthermore, existing sound source localization methods rely solely on time delay estimation for location calculation, failing to consider the impact of spatial distribution deviations on the localization results. They lack a graded correction mechanism for localization deviations, making the localization results susceptible to environmental interference and unable to achieve accurate calibration. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a sound source localization method for fault identification of transmission devices.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A sound source localization method for fault identification in a transmission device, the method comprising the following steps: Acquire multi-channel acoustic signal data from the transmission device, and perform PTP hardware clock synchronization processing on the multi-channel acoustic signal data to obtain synchronized acoustic signal data; The synchronous acoustic signal data is denoised to obtain the fault feature enhancement signal, and the sound source localization reference point is set based on the fault feature enhancement signal. Based on the sound source localization reference point, the fault feature enhancement signal is judged to obtain the time delay deviation coefficient, spatial distribution deviation coefficient, time delay anomaly reference point, and spatial anomaly reference point; The fault sound source abnormal reference point is constructed based on the time delay abnormal reference point and the spatial abnormal reference point; the positioning deviation of the fault sound source abnormal reference point is obtained based on the preset time delay deviation weight and time delay deviation coefficient, and the preset spatial deviation weight and spatial distribution deviation coefficient. Based on the positioning deviation, a positioning correction intensity of the abnormal reference point of the fault sound source is generated; based on the positioning correction intensity, a positioning correction signal of the sound source positioning reference point is generated; and the signal transmission distance is obtained based on the abnormal reference point of the fault sound source and the preset positioning correction node. The correction priority of the abnormal reference point of the fault sound source is obtained based on the positioning correction intensity and signal transmission distance; the positioning correction node performs three-dimensional positioning calibration on the abnormal reference point of the fault sound source according to the correction priority and the positioning correction signal. The three-dimensional positioning calibration results are mapped to the three-dimensional model of the transmission device to complete the visualization output of the fault sound source.

[0006] Preferably, acquiring multi-channel acoustic signal data from the transmission device and performing PTP hardware clock synchronization processing on the multi-channel acoustic signal data to obtain synchronized acoustic signal data includes: A master-slave clock architecture is constructed, and the synchronization signal data is obtained by synchronizing the multi-channel audio signal data based on the hardware timestamp in the master-slave clock architecture of the PTP hardware. The synchronization signal data is pre-amplified and anti-aliasing filtered to obtain standardized synchronization sound signal data.

[0007] Preferably, the synchronous acoustic signal data is subjected to noise reduction processing to obtain a fault feature enhancement signal, including: Short-time Fourier transform is performed on the synchronous acoustic signal data to extract the steady-state noise spectrum features; The target signal is obtained by filtering out the power frequency and fan steady-state interference signals in the steady-state noise spectrum characteristics based on adaptive notch filtering. The fault feature enhancement signal is obtained by calculating the target signal based on Kalman filtering.

[0008] Preferably, the time delay deviation coefficient, spatial distribution deviation coefficient, time delay anomaly reference point, and spatial anomaly reference point are obtained by judging the fault feature enhancement signal based on the sound source localization reference point, including: Based on the sound source localization reference point, the time delay of the fault feature enhancement signal is calculated to obtain the time delay deviation coefficient, and the time delay deviation coefficient is marked as the time delay anomaly reference point; The spatial distribution deviation coefficient is obtained by judging the distribution based on the spatial coordinates of the sound source localization reference point and the acquisition node, and the spatial abnormal reference point is marked based on the spatial distribution deviation coefficient.

[0009] Preferably, the time delay deviation coefficient is obtained by calculating the time delay of the fault feature enhancement signal based on the sound source localization reference point, and the time delay deviation coefficient is marked as the time delay anomaly reference point, including: The time difference of the signal arriving at each acquisition node is calculated from the enhanced fault characteristic signal; The time difference is compared with the preset time delay threshold, and the time delay deviation coefficient is obtained based on the time difference and the time delay threshold. The corresponding sound source localization reference point is marked as the time delay abnormal reference point.

[0010] Preferably, a spatial distribution deviation coefficient is obtained by determining the distribution based on the spatial coordinates of the sound source localization reference point and the acquisition node, and spatial anomaly reference points are marked based on the spatial distribution deviation coefficient, including: Collect data from nodes designed for harsh working conditions to construct three-dimensional spatial calibration coordinates; Calculate the spatial distance deviation between the sound source localization reference point and each acquisition node; The spatial distance deviation is compared with the preset spatial threshold, and the spatial distribution deviation coefficient is obtained based on the spatial distance deviation and the spatial threshold. The sound source localization reference point corresponding to the spatial distribution deviation coefficient is marked as the spatial anomaly reference point.

[0011] Preferably, the signal transmission distance is obtained based on the abnormal reference point of the fault sound source and the preset positioning correction node, including: The first moment when the abnormal reference point of the fault sound source sends the positioning correction signal is obtained, and the second moment when the positioning correction node receives the positioning correction signal is obtained. The signal transmission duration is obtained based on the first and second time points; The signal transmission distance is obtained by combining the ambient temperature-compensated sound velocity, the signal transmission time, and the compensated sound velocity.

[0012] Preferably, the correction priority of the abnormal reference point of the fault sound source is obtained based on the positioning correction intensity and the signal transmission distance, including: Preset correction strength weights and transmission distance weights; The positioning correction coefficient is obtained based on the positioning correction strength and the correction strength weight. The spatial distance coefficient is obtained based on the signal transmission distance and the transmission distance weight. A comprehensive correction coefficient is obtained based on the positioning correction coefficient and the spatial distance coefficient, and the correction priority is set from high to low based on the comprehensive correction coefficient.

[0013] Preferably, the three-dimensional positioning calibration results are mapped to the three-dimensional model of the transmission device to complete the visualization output of the fault sound source, including: Import the 3D model of the transmission device; Map the fault coordinates in the 3D positioning calibration results to the corresponding positions in the 3D model; Display the fault location and probability in the form of marked points or heat maps, and output fault location information.

[0014] Compared with existing technologies, this invention has the following advantages: It achieves nanosecond-level synchronous marking of multiple acoustic signals through hardware timestamps, solving the problem of time delay calculation deviation caused by millisecond-level errors in software synchronization. This provides a precise time reference for subsequent sound source localization, improving the localization accuracy from meter-level to centimeter-level, meeting the core requirement of precise fault localization in transmission devices. Furthermore, it extracts steady-state noise spectrum features through short-time Fourier transform, combines adaptive notch filtering to remove steady-state interference from power frequency and fans, and then uses Kalman filtering to suppress non-steady-state random noise, forming a multi-level progressive noise reduction mechanism. This effectively solves the problem of strong background noise in industrial sites drowning out fault features, allowing for the extraction of clear fault feature signals even in low signal-to-noise ratio environments, significantly improving fault identification. The system ensures accuracy and reliability; by using both time delay deviation coefficient and spatial distribution deviation coefficient, it identifies time delay anomaly reference points and spatial anomaly reference points respectively, constructs fault sound source anomaly reference points, calculates positioning deviation degree by combining preset weights, generates positioning correction intensity and correction priority, realizes hierarchical correction of positioning deviation, effectively offsets positioning deviation caused by environmental interference, and ensures the accuracy of positioning results; by mapping the three-dimensional positioning calibration results to the three-dimensional model of the transmission device, it intuitively displays the fault location and fault probability in the form of marked points or heat maps, solving the problem that existing technologies only output coordinate data and lack intuitive presentation. Maintenance personnel can quickly locate faulty components, significantly shorten fault diagnosis and repair time, and improve the operation and maintenance efficiency of industrial equipment. Attached Figure Description

[0015] Figure 1 This is a schematic diagram illustrating the steps of a sound source localization method for fault identification in a transmission device according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the steps of obtaining the signal transmission distance in a sound source localization method for identifying faults in a transmission device, as provided in an embodiment of the present invention. Detailed Implementation

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0017] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0018] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0019] Reference Figures 1-2 As shown.

[0020] This embodiment further illustrates the sound source localization method for fault identification of transmission devices proposed in this invention.

[0021] A sound source localization method for fault identification in a transmission device, the method comprising the following steps: Acquire multi-channel acoustic signal data from the transmission device, and perform PTP hardware clock synchronization processing on the multi-channel acoustic signal data to obtain synchronized acoustic signal data; The synchronous acoustic signal data is denoised to obtain the fault feature enhancement signal, and the sound source localization reference point is set based on the fault feature enhancement signal. Based on the sound source localization reference point, the fault feature enhancement signal is judged to obtain the time delay deviation coefficient, spatial distribution deviation coefficient, time delay anomaly reference point, and spatial anomaly reference point; The abnormal reference point of the fault sound source is constructed based on the time delay abnormal reference point and the spatial abnormal reference point; the positioning deviation of the abnormal reference point of the fault sound source is obtained based on the preset time delay deviation weight and time delay deviation coefficient, and the preset spatial deviation weight and spatial distribution deviation coefficient.

[0022] First, by combining the acquisition timing, propagation path characteristics, and spatial distribution information of multiple acoustic signals, corresponding time delay anomaly reference points and spatial anomaly reference points are identified. The time delay anomaly reference point reflects the deviation of the time delay from normal operation when different acquisition nodes receive the fault acoustic signal. This value reflects the degree of time delay deviation caused by changes in fault location and media interference during the propagation process. The spatial anomaly reference point, based on the spatial coordinate relationship of each acquisition node, determines the characteristic points where the distribution of the fault acoustic signal deviates from normal operating conditions in the spatial propagation dimension. It encompasses the abnormal information of signal amplitude distribution differences between spatial nodes and the magnitude of propagation direction offset.

[0023] By integrating time delay anomaly reference points with spatial anomaly reference points, the fault sound source anomaly reference point is determined. This reference point comprehensively reflects the combined anomaly characteristics of the fault sound source in both time and spatial dimensions. After constructing the fault sound source anomaly reference point, the system calculates the positioning deviation of the reference point based on preset time delay deviation weights, time delay deviation coefficients, preset spatial deviation weights, and spatial distribution deviation coefficients. The positioning deviation is calculated by quantifying the sum of the products of the anomaly weights and coefficients of different dimensions. The specific formula is: Positioning Deviation = Preset Time Delay Deviation Weight × Time Delay Deviation Coefficient + Preset Spatial Deviation Weight × Spatial Distribution Deviation Coefficient.

[0024] In the actual calculation process, the preset time delay deviation weight and preset spatial deviation weight are pre-set parameters according to the system's requirements for time synchronization accuracy and spatial positioning accuracy. For example, if the system pays more attention to the positioning accuracy in the time dimension, the preset time delay deviation weight can be set to 0.6, and the preset spatial deviation weight can be set to 0.4 accordingly. The time delay deviation coefficient is a specific value obtained by analyzing and fitting the time delay data of multiple sound signals. For example, the sound signal caused by a gear failure in a transmission device has a calculated time delay deviation coefficient of 0.82. The spatial distribution deviation coefficient is the result calculated based on the spatial coordinates of the spatial anomaly reference point and the signal distribution characteristics. For example, the analyzed spatial distribution deviation coefficient is 0.75. Substituting these parameters into the calculation formula, we can obtain the positioning deviation degree = 0.6 × 0.82 + 0.4 × 0.75, which is 0.792. This value is the positioning deviation degree of the abnormal reference point of the fault sound source.

[0025] The positioning deviation can intuitively reflect the comprehensive deviation of the abnormal reference point of the fault sound source in two key dimensions: time and space. The higher the deviation value, the more obvious the position of the fault sound source deviates from the normal theoretical positioning range. This provides a core quantitative basis for generating positioning correction intensity and performing three-dimensional positioning calibration, ensuring that the fault sound source positioning result can be accurately mapped to the actual structure of the transmission device, and realizing rapid and accurate fault identification.

[0026] Based on the positioning deviation, a positioning correction intensity of the abnormal reference point of the fault sound source is generated; based on the positioning correction intensity, a positioning correction signal of the sound source positioning reference point is generated; and the signal transmission distance is obtained based on the abnormal reference point of the fault sound source and the preset positioning correction node. The correction priority of the abnormal reference point of the fault sound source is obtained based on the positioning correction intensity and signal transmission distance; the positioning correction node performs three-dimensional positioning calibration on the abnormal reference point of the fault sound source according to the correction priority and the positioning correction signal. The three-dimensional positioning calibration results are mapped to the three-dimensional model of the transmission device to complete the visualization output of the fault sound source.

[0027] Acquire multi-channel acoustic signal data from the transmission device, and perform PTP hardware clock synchronization processing on the multi-channel acoustic signal data to obtain synchronized acoustic signal data, including: A master-slave clock architecture is constructed, and the synchronization signal data is obtained by synchronizing the multi-channel audio signal data based on the hardware timestamp in the master-slave clock architecture of the PTP hardware. The synchronization signal data is pre-amplified and anti-aliasing filtered to obtain standardized synchronization sound signal data.

[0028] First, a master-slave clock architecture must be constructed. In this architecture, the system sets a master clock node as the time base, and the remaining nodes that collect sound signals act as slave clock nodes. The master clock node periodically sends synchronization messages to each slave clock node. During the process of receiving messages, the slave clock nodes use hardware-level timestamp marking to mark the acquisition time of the sound signal data and the message transmission and reception time with nanosecond precision. This eliminates clock offset and transmission delay differences between different acquisition nodes, achieving precise alignment of multiple sound signal data in the time dimension. For example, if four sound signal acquisition nodes are deployed around the transmission device, the master clock node will synchronously send PTP synchronization messages to these four slave nodes. Each slave node will timestamp its own acquired sound signal data with a precision within 100 nanoseconds at the hardware level, ensuring that the acquisition time base of the four sound signal data is completely unified.

[0029] After completing the nanosecond-level synchronization marking under the master-slave clock architecture, the synchronization signal data needs to be pre-amplified and anti-aliasing filtered to obtain standardized synchronization acoustic signal data. Pre-amplification is used for weak fault acoustic signals. An operational amplifier is used to increase the amplitude of the acoustic signal to a range recognizable by subsequent processing modules, avoiding feature loss due to excessively low signal amplitude. For example, a fault acoustic signal with an original amplitude of 0.1 millivolts is amplified to 1 volt to meet the input requirements of the analog-to-digital conversion module. Anti-aliasing filtering uses a low-pass filter to remove high-frequency noise components above the Nyquist frequency from the acoustic signal, preventing frequency aliasing during analog-to-digital conversion and ensuring the integrity of the acoustic signal characteristics. For example, when the sampling frequency is set to 48 kHz, the anti-aliasing filter will completely filter out high-frequency noise above 24 kHz, retaining only the effective fault acoustic signal frequency band. After undergoing pre-amplification and anti-aliasing filtering, the synchronized multi-channel acoustic signal data is transformed into standardized synchronized acoustic signal data with stable amplitude and controllable noise interference, providing a high-quality input foundation for subsequent noise reduction processing, fault feature extraction, and sound source localization calculation.

[0030] The noise reduction process performed on the synchronous acoustic signal data yields a fault feature enhancement signal, including: Short-time Fourier transform is performed on the synchronous acoustic signal data to extract the steady-state noise spectrum features; The target signal is obtained by filtering out the power frequency and fan steady-state interference signals in the steady-state noise spectrum characteristics based on adaptive notch filtering. The fault feature enhancement signal is obtained by calculating the target signal based on Kalman filtering.

[0031] First, a short-time Fourier transform (SFT) is performed on the synchronous acoustic signal data to extract the steady-state noise spectral characteristics. The SFT decomposes a continuous acoustic signal in the time domain into frequency domain signals within different time windows, enabling joint analysis of the time and frequency domains and thus accurately identifying the persistent steady-state noise components in the industrial environment. Specifically, the synchronous acoustic signal is divided into several overlapping time windows, and a Fourier transform is performed on the signal within each window to obtain the corresponding spectrum. Then, the frequency characteristics of fan operation and power frequency interference are extracted through the amplitude distribution of the spectrum. For example, when there is 50 Hz power frequency noise and 200 Hz fan operation noise in the workshop where the transmission device is located, the SFT can clearly locate the noise amplitude at these two frequency points, providing a clear target frequency band for subsequent filtering processing.

[0032] The target signal is obtained by filtering out steady-state interference signals such as power frequency and fan noise from the steady-state noise spectrum characteristics using adaptive notch filtering. The adaptive notch filter generates a notch filter for the corresponding frequency of the steady-state noise extracted in the first step, dynamically adjusting the filter parameters to accurately suppress these interference signals. This filtering method can minimize the impact of steady-state noise while preserving the effective fault sound signal. For example, for 50 Hz power frequency noise, the adaptive notch filter sets a deep attenuation stopband at that frequency, reducing the signal amplitude to less than 1% of its original value, while producing almost no attenuation for fault characteristic frequency signals outside 50 Hz, thus obtaining a target signal containing only the fault sound signal and a small amount of non-steady-state noise.

[0033] Kalman filtering is used to calculate the enhanced fault feature signal from the target signal. Kalman filtering establishes a state-space model of the acoustic signal, dynamically predicting and updating the optimal signal value at the current moment using the estimated value from the previous moment and the observed value from the current moment. This effectively suppresses interference from non-steady-state random noise while amplifying weak fault feature signals. The specific formula is: Current optimal estimate = Previous optimal estimate + Kalman gain × (Current observed value - Previous predicted observed value). After Kalman filtering, random noise in the target signal is further filtered out, and the amplitude and feature clarity of the fault acoustic signal are significantly improved, ultimately forming an enhanced fault feature signal that can be used for subsequent sound source localization analysis.

[0034] Based on the sound source localization reference point, the fault characteristic enhancement signal is judged to obtain the time delay deviation coefficient, spatial distribution deviation coefficient, time delay anomaly reference point, and spatial anomaly reference point, including: Based on the sound source localization reference point, the time delay of the fault feature enhancement signal is calculated to obtain the time delay deviation coefficient, and the time delay deviation coefficient is marked as the time delay anomaly reference point; The spatial distribution deviation coefficient is obtained by judging the distribution based on the spatial coordinates of the sound source localization reference point and the acquisition node, and the spatial abnormal reference point is marked based on the spatial distribution deviation coefficient.

[0035] First, the time delay of the fault feature enhancement signal is calculated based on the sound source localization reference point to obtain the time delay deviation coefficient, and this coefficient is marked as the time delay anomaly reference point. The sound source localization reference point is a pre-set theoretical sound source location under normal operating conditions of the transmission device. When calculating the time delay, the system calculates the theoretical propagation time delay based on the spatial distance between each acquisition node and the sound source localization reference point, combined with the sound speed. This theoretical time delay is then compared with the actual propagation time delay of the fault feature enhancement signal between the acquisition nodes to obtain the deviation value, and subsequently, the time delay deviation coefficient is calculated. The specific calculation formula is: Time delay deviation coefficient = |Actual propagation time delay - Theoretical propagation time delay| / Theoretical propagation time delay. For example, if the theoretical propagation delay between the acquisition node and the sound source positioning reference point is 2 milliseconds, and the actual detected fault sound signal propagation delay is 2.3 milliseconds, substituting into the formula, we can get the time delay deviation coefficient = |2.3-2| / 2 = 0.15. When this coefficient exceeds the preset threshold, the location corresponding to the node is marked as the time delay abnormal reference point. This reference point can intuitively reflect the degree of deviation of the fault sound source in the time propagation dimension.

[0036] Secondly, the spatial distribution is determined based on the spatial coordinates of the sound source localization reference point and the acquisition nodes to obtain the spatial distribution deviation coefficient. Based on this coefficient, spatial anomaly reference points are then marked. The system first acquires the three-dimensional spatial coordinates of each acquisition node, combines them with the three-dimensional coordinates of the sound source localization reference point, calculates the theoretical spatial distance from each acquisition node to the reference point, and then calculates the actual spatial distribution distance based on the amplitude distribution and propagation direction of the fault characteristic enhancement signal at each acquisition node, thus obtaining the spatial distribution deviation coefficient. The calculation formula is: Spatial distribution deviation coefficient = Σ|Actual spatial distribution distance - Theoretical spatial distribution distance| / Theoretical spatial distribution distance, where Σ represents the sum of the deviation values ​​over all acquisition nodes. For example, four acquisition nodes are deployed around the transmission device. The spatial distribution deviations of each node are calculated to be 0.1, 0.2, 0.05, and 0.15, respectively. After summing and dividing by 4, the spatial distribution deviation coefficient is 0.125. When this coefficient exceeds the preset threshold, the corresponding spatial location is marked as a spatial anomaly reference point. This reference point can reflect the degree of deviation of the fault sound source in the spatial distribution dimension, providing a core basis for subsequent positioning deviation calculation and three-dimensional calibration.

[0037] Based on the sound source localization reference point, the time delay of the fault characteristic enhancement signal is calculated to obtain the time delay deviation coefficient. The time delay deviation coefficient is marked as the time delay anomaly reference point, including: The time difference of the signal arriving at each acquisition node is calculated from the enhanced fault characteristic signal; The time difference is compared with the preset time delay threshold, and the time delay deviation coefficient is obtained based on the time difference and the time delay threshold. The corresponding sound source localization reference point is marked as the time delay abnormal reference point.

[0038] First, the time difference of the fault feature enhancement signal arriving at each acquisition node is calculated. After PTP hardware clock synchronization, each acquisition node records the arrival timestamp of the fault feature enhancement signal. The system selects one acquisition node as the reference node and calculates the difference between the arrival times of the signals from the other nodes and the reference node to obtain the signal arrival time difference between each pair of nodes. For example, if four acquisition nodes are deployed around the transmission device, and node 1 is selected as the reference node, with the signal arrival times recorded by nodes 2, 3, and 4 being 1.2 ms, 1.5 ms, and 1.1 ms respectively, and the arrival time of reference node 1 being 1.0 ms, then the time difference between node 2 and the reference node can be calculated as 0.2 ms, the time difference between node 3 and the reference node as 0.5 ms, and the time difference between node 4 and the reference node as 0.1 ms. This time difference directly reflects the difference in propagation delay of the fault sound signal between different acquisition nodes, providing basic data for subsequent deviation analysis.

[0039] The time difference is compared with a preset time delay threshold, and a time delay deviation coefficient is obtained based on the time difference and the time delay threshold. The corresponding sound source positioning reference point is marked as a time delay anomaly reference point. The system pre-sets a reasonable time delay threshold based on the sound signal propagation characteristics under normal operating conditions of the transmission device. This threshold represents the acceptable deviation range of the time difference. The formula for calculating the time delay deviation coefficient is: Time Delay Deviation Coefficient = |Actual Time Difference - Theoretical Time Difference| / Preset Time Delay Threshold, where the theoretical time difference is the standard propagation time difference calculated based on the spatial distance and sound speed between the sound source positioning reference point and each acquisition node. For example, if the theoretical value of a certain time difference is 0.2 milliseconds, the actual calculated time difference is 0.5 milliseconds, and the preset time delay threshold is 0.3 milliseconds, substituting into the formula, we get the time delay deviation coefficient = |0.5 - 0.2| / 0.3 = 1. When the coefficient is greater than 1, it indicates that the actual propagation delay has exceeded the normal acceptable range. The system will mark the corresponding sound source positioning reference point as a time delay abnormality reference point. This reference point can be directly used for subsequent positioning deviation calculation and three-dimensional calibration, providing an abnormal basis in the time dimension for the accurate positioning of faulty sound sources.

[0040] Based on the spatial coordinates of the sound source localization reference point and the acquisition node, a spatial distribution deviation coefficient is obtained. Based on this coefficient, spatially abnormal reference points are marked, including: Collect data from nodes designed for harsh working conditions to construct three-dimensional spatial calibration coordinates; Calculate the spatial distance deviation between the sound source localization reference point and each acquisition node; The spatial distance deviation is compared with the preset spatial threshold, and the spatial distribution deviation coefficient is obtained based on the spatial distance deviation and the spatial threshold. The sound source localization reference point corresponding to the spatial distribution deviation coefficient is marked as the spatial anomaly reference point.

[0041] First, the system collects data from the harsh operating conditions-resistant acquisition nodes to construct a three-dimensional spatial calibration coordinate system. These nodes are specifically designed for the harsh environments of industrial sites, characterized by high temperatures, high dust levels, and strong vibrations. After deployment, the system uses a laser rangefinder or a 3D laser scanner to collect the three-dimensional coordinates of each node within the space where the transmission device is located, thus constructing a three-dimensional spatial calibration coordinate system. For example, if four harsh operating conditions-resistant acquisition nodes are deployed around the transmission device, the coordinates of node 1 (0, 0, 0), node 2 (2, 0, 0), node 3 (1, 2, 0), and node 4 (1, 1, 3) will be obtained. These coordinates will serve as the basis for subsequent spatial distance calculations.

[0042] The system calculates the spatial distance deviation between the sound source localization reference point and each acquisition node. First, based on the transmission device structure under normal operating conditions, the system determines the three-dimensional coordinates of the sound source localization reference point. Then, using the three-dimensional spatial distance formula, it calculates the theoretical spatial distance from the reference point to each acquisition node. Simultaneously, considering the amplitude and propagation direction of the fault characteristic enhancement signal at each acquisition node, the actual spatial distance is calculated. Finally, the deviation between the two is obtained. The calculation formula is: Spatial distance deviation = |Actual spatial distance - Theoretical spatial distance|. For example, if the coordinates of the sound source localization reference point are 1, 1, 1, the calculated theoretical spatial distance to node 1 is √3 ≈ 1.732, and the actual spatial distance is 2.232. Therefore, the spatial distance deviation is |2.232 - 1.732| = 0.5.

[0043] Finally, the spatial distance deviation is compared with a preset spatial threshold. Based on this comparison, a spatial distribution deviation coefficient is obtained, and the sound source localization reference point corresponding to the spatial distribution deviation coefficient is marked as a spatial anomaly reference point. The system pre-sets a preset spatial threshold based on the structural accuracy and positioning requirements of the transmission device. The formula for calculating the spatial distribution deviation coefficient is: Spatial distribution deviation coefficient = Σ spatial distance deviation / (Number of acquisition nodes × Preset spatial threshold), where Σ represents the sum of the spatial distance deviations of all acquisition nodes. For example, if the spatial distance deviations of four acquisition nodes are 0.5, 0.3, 0.4, and 0.2 respectively, the sum is 1.4. With four acquisition nodes and a preset spatial threshold of 0.2, the spatial distribution deviation coefficient = 1.4 / (4 × 0.2) = 1.75. When this coefficient is greater than 1, it indicates that the spatial distribution of the faulty sound source has exceeded the normal acceptable range. The system will mark the corresponding sound source localization reference point as a spatial anomaly reference point, providing a spatial dimension basis for subsequent localization deviation calculation and three-dimensional calibration.

[0044] The signal transmission distance is obtained based on the abnormal reference point of the fault sound source and the preset positioning correction node, including: The first moment when the abnormal reference point of the fault sound source sends the positioning correction signal is obtained, and the second moment when the positioning correction node receives the positioning correction signal is obtained. The signal transmission duration is obtained based on the first and second time points; The signal transmission distance is obtained by combining the ambient temperature-compensated sound velocity, the signal transmission time, and the compensated sound velocity.

[0045] First, the system acquires the first moment when the faulty sound source abnormal reference point sends a positioning correction signal, and the second moment when the positioning correction node receives the signal. The faulty sound source abnormal reference point is a potential location of the faulty sound source determined through prior assessment. The system uses the signal transmitting module corresponding to this reference point as the trigger source and records the instant it sends the positioning correction signal as the first moment. Simultaneously, the preset positioning correction node deployed in the transmission device space records the time the signal is received; this time is the second moment. For example, if the transmitting module corresponding to the faulty sound source abnormal reference point sends a signal at 10 milliseconds, the first moment is recorded as 10 milliseconds; if the positioning correction node receives the signal at 10.002 seconds, the second moment is recorded as 10.002 seconds.

[0046] Secondly, the signal transmission duration is calculated based on the first and second time points. The signal transmission duration is the difference between the second and first time points. This duration directly reflects the total time it takes for the positioning correction signal to propagate from the fault sound source's abnormal reference point to the positioning correction node. The calculation formula is: Signal transmission duration = Second time point - First time point. For example, signal transmission duration = 10.002 seconds - 10 milliseconds = 0.002 seconds. This value provides the basic time data for subsequent distance calculations.

[0047] The signal transmission distance is obtained by combining the ambient temperature-compensated sound velocity, signal transmission duration, and the compensated sound velocity. Since the sound velocity changes with ambient temperature, temperature compensation is required to obtain an accurate compensated sound velocity. The sound velocity temperature compensation formula is: Compensated sound velocity = 331 m / s + 0.6 × Ambient temperature, where 331 m / s is the sound velocity at 0 degrees Celsius, and 0.6 is the correction factor for sound velocity changes with temperature. For example, when the ambient temperature is 25 degrees Celsius, the compensated sound velocity = 331 + 0.6 × 25 = 346 m / s. Based on this, the signal transmission distance is calculated as: Signal transmission distance = Compensated sound velocity × Signal transmission duration. Substituting into the example above, the signal transmission distance = 346 m / s × 0.002 s = 0.692 m. This allows for the accurate determination of the signal transmission distance between the fault sound source's abnormal reference point and the positioning correction node, providing reliable spatial parameters for subsequent positioning correction intensity calculations and 3D positioning calibration.

[0048] The correction priority of the abnormal reference point of the fault sound source is obtained based on the positioning correction intensity and signal transmission distance, including: Preset correction strength weights and transmission distance weights; The positioning correction coefficient is obtained based on the positioning correction strength and the correction strength weight. The spatial distance coefficient is obtained based on the signal transmission distance and the transmission distance weight. A comprehensive correction coefficient is obtained based on the positioning correction coefficient and the spatial distance coefficient, and the correction priority is set from high to low based on the comprehensive correction coefficient.

[0049] First, preset the correction intensity weight and transmission distance weight. Based on the actual needs of transmission device fault detection, the system pre-sets the correction intensity weight and transmission distance weight. The correction intensity weight measures the impact of positioning deviation on calibration, while the transmission distance weight measures the impact of signal transmission distance on calibration response efficiency. The sum of the two weights is 1 to ensure the rationality of the weight allocation. For example, if the system prioritizes the effectiveness of deviation correction, the correction intensity weight can be set to 0.7, and the transmission distance weight to 0.3.

[0050] The positioning correction coefficient is obtained based on the positioning correction intensity and its weight. The positioning correction intensity is a quantified parameter calculated from the positioning deviation, reflecting the degree of positioning deviation of the abnormal reference point of the faulty sound source. The formula for calculating the positioning correction coefficient is: Positioning Correction Coefficient = Positioning Correction Intensity × Correction Intensity Weight. For example, if the positioning correction intensity of the abnormal reference point of the faulty sound source is 0.8 and the correction intensity weight is 0.7, then the positioning correction coefficient = 0.8 × 0.7 = 0.56. This coefficient reflects the proportion of contribution of the deviation degree to the overall correction.

[0051] The spatial distance coefficient is obtained based on the signal transmission distance and its weight. The signal transmission distance is the actual propagation distance between the fault sound source's abnormal reference point and the positioning correction node. The formula for calculating the spatial distance coefficient is: Spatial distance coefficient = Signal transmission distance × Transmission distance weight. For example, if the signal transmission distance is 0.692 meters and the transmission distance weight is 0.3, then the spatial distance coefficient = 0.692 × 0.3 = 0.2076. This coefficient reflects the contribution ratio of spatial distance in the comprehensive correction.

[0052] Finally, a comprehensive correction coefficient is obtained based on the positioning correction coefficient and the spatial distance coefficient. Correction priorities are then set from high to low based on this comprehensive correction coefficient. The formula for calculating the comprehensive correction coefficient is: Comprehensive Correction Coefficient = Positioning Correction Coefficient + Spatial Distance Coefficient. Substituting the values ​​into the formula, we get the comprehensive correction coefficient = 0.56 + 0.2076 = 0.7676. The system sorts the comprehensive correction coefficients of all abnormal reference points of the faulty sound sources. A higher coefficient indicates a more significant positioning deviation and a more critical calibration response efficiency for that reference point; therefore, it has a higher correction priority. Positioning correction nodes will sequentially perform three-dimensional positioning calibration on the abnormal reference points of the faulty sound sources according to this priority, ensuring that core fault points are addressed first and improving the overall efficiency of fault identification and positioning.

[0053] The 3D positioning calibration results are mapped to the 3D model of the transmission device to complete the visualization output of the fault sound source, including: Import the 3D model of the transmission device; Map the fault coordinates in the 3D positioning calibration results to the corresponding positions in the 3D model; Display the fault location and probability in the form of marked points or heat maps, and output fault location information.

[0054] First, the 3D model of the transmission device is imported. The system pre-acquires a CAD 3D model of the transmission device or a point cloud model generated through 3D laser scanning, completely restoring the actual structural dimensions and spatial layout of the transmission device, including precise 3D coordinate information of core components such as gears, bearings, and shafts, providing a unified spatial reference for subsequent fault coordinate mapping. For example, for a large gear transmission device in a metallurgical workshop, the system imports its complete 3D model, including the gearbox, bearing housing, and drive shaft. The model's coordinate system is perfectly aligned with the 3D spatial calibration coordinate system of the field acquisition nodes, ensuring spatial consistency in subsequent mapping.

[0055] The system maps the fault coordinates from the 3D positioning calibration results to the corresponding positions in the 3D model. The 3D positioning calibration results contain the accurate 3D coordinates of the fault sound source's abnormal reference point after calibration. Based on a unified spatial coordinate system, the system matches these fault coordinates with the structural coordinates of the transmission device's 3D model, calculating the corresponding spatial position of the fault coordinates in the 3D model, thus completing the accurate mapping from actual detection data to the virtual model. For example, if the fault coordinates obtained after 3D positioning calibration are X=1.2 meters, Y=0.8 meters, and Z=0.5 meters, the system substitutes these coordinates into the coordinate system of the 3D model, matching them to the corresponding installation position of a bearing inside the transmission device's gearbox, completing the model mapping of the fault location.

[0056] Finally, the system displays the fault location and probability in the form of markers or heatmaps, outputting fault location information. The system generates visual markers based on the fault coordinates at the corresponding locations in the 3D model. Marker points directly highlight the fault location, visually indicating the specific component where the fault occurred. Heatmaps, on the other hand, use different colored heat areas to cover the corresponding locations based on the fault probability; areas with higher probability are more prominent. Simultaneously, the system outputs a fault location report containing information such as fault coordinates, faulty components, and fault probability. For example, for the aforementioned bearing fault, the system generates a red highlighted marker at the corresponding bearing location in the 3D model, along with a heatmap area indicating a 95% fault probability at that location, and outputs a location report containing fault coordinates and fault type. Maintenance personnel can quickly pinpoint the fault location directly through the visual interface, significantly improving the efficiency of fault diagnosis and repair.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A sound source localization method for fault identification in transmission devices, characterized in that, The method includes the following steps: Acquire multi-channel acoustic signal data from the transmission device, and perform PTP hardware clock synchronization processing on the multi-channel acoustic signal data to obtain synchronized acoustic signal data; The synchronous acoustic signal data is denoised to obtain the fault feature enhancement signal, and the sound source localization reference point is set based on the fault feature enhancement signal. Based on the sound source localization reference point, the fault feature enhancement signal is judged to obtain the time delay deviation coefficient, spatial distribution deviation coefficient, time delay anomaly reference point, and spatial anomaly reference point; The fault sound source abnormal reference point is constructed based on the time delay abnormal reference point and the spatial abnormal reference point; the positioning deviation of the fault sound source abnormal reference point is obtained based on the preset time delay deviation weight and time delay deviation coefficient, and the preset spatial deviation weight and spatial distribution deviation coefficient. Based on the positioning deviation, a positioning correction intensity of the abnormal reference point of the fault sound source is generated; based on the positioning correction intensity, a positioning correction signal of the sound source positioning reference point is generated; and the signal transmission distance is obtained based on the abnormal reference point of the fault sound source and the preset positioning correction node. The correction priority of the abnormal reference point of the fault sound source is obtained based on the positioning correction intensity and signal transmission distance; the positioning correction node performs three-dimensional positioning calibration on the abnormal reference point of the fault sound source according to the correction priority and the positioning correction signal. The three-dimensional positioning calibration results are mapped to the three-dimensional model of the transmission device to complete the visualization output of the fault sound source.

2. The sound source localization method for fault identification of a transmission device according to claim 1, characterized in that, Acquire multi-channel acoustic signal data from the transmission device, and perform PTP hardware clock synchronization processing on the multi-channel acoustic signal data to obtain synchronized acoustic signal data, including: Synchronization signal data is obtained by synchronizing and marking the multi-channel audio signal data based on the hardware timestamp in the master-slave clock architecture of PTP hardware. The synchronization signal data is pre-amplified and anti-aliasing filtered to obtain standardized synchronization sound signal data.

3. The sound source localization method for fault identification of a transmission device according to claim 2, characterized in that, The noise reduction process performed on the synchronous acoustic signal data yields a fault feature enhancement signal, including: Short-time Fourier transform is performed on the synchronous acoustic signal data to extract the steady-state noise spectrum features; The target signal is obtained by filtering out the power frequency and fan steady-state interference signals in the steady-state noise spectrum characteristics based on adaptive notch filtering. The fault feature enhancement signal is obtained by calculating the target signal based on Kalman filtering.

4. The sound source localization method for fault identification of a transmission device according to claim 3, characterized in that, Based on the sound source localization reference point, the fault characteristic enhancement signal is judged to obtain the time delay deviation coefficient, spatial distribution deviation coefficient, time delay anomaly reference point, and spatial anomaly reference point, including: Based on the sound source localization reference point, the time delay of the fault feature enhancement signal is calculated to obtain the time delay deviation coefficient, and the time delay deviation coefficient is marked as the time delay anomaly reference point; The spatial distribution deviation coefficient is obtained by judging the distribution based on the spatial coordinates of the sound source localization reference point and the acquisition node, and the spatial abnormal reference point is marked based on the spatial distribution deviation coefficient.

5. The sound source localization method for fault identification of a transmission device according to claim 4, characterized in that, Based on the sound source localization reference point, the time delay of the fault characteristic enhancement signal is calculated to obtain the time delay deviation coefficient. The time delay deviation coefficient is marked as the time delay anomaly reference point, including: The time difference of the signal arriving at each acquisition node is calculated from the enhanced fault characteristic signal; The time difference is compared with the preset time delay threshold, and the time delay deviation coefficient is obtained based on the time difference and the time delay threshold. The corresponding sound source localization reference point is marked as the time delay abnormal reference point.

6. The sound source localization method for fault identification of a transmission device according to claim 5, characterized in that, Based on the spatial coordinates of the sound source localization reference point and the acquisition node, a spatial distribution deviation coefficient is obtained. Based on this coefficient, spatially abnormal reference points are marked, including: Collect data from nodes designed for harsh working conditions to construct three-dimensional spatial calibration coordinates; Calculate the spatial distance deviation between the sound source localization reference point and each acquisition node; The spatial distance deviation is compared with the preset spatial threshold, and the spatial distribution deviation coefficient is obtained based on the spatial distance deviation and the spatial threshold. The sound source localization reference point corresponding to the spatial distribution deviation coefficient is marked as the spatial anomaly reference point.

7. The sound source localization method for fault identification of a transmission device according to claim 6, characterized in that, The signal transmission distance is obtained based on the abnormal reference point of the fault sound source and the preset positioning correction node, including: The first moment when the abnormal reference point of the fault sound source sends the positioning correction signal is obtained, and the second moment when the positioning correction node receives the positioning correction signal is obtained. The signal transmission duration is obtained based on the first and second time points; The signal transmission distance is obtained by combining the ambient temperature-compensated sound velocity, the signal transmission time, and the compensated sound velocity.

8. The sound source localization method for fault identification of a transmission device according to claim 7, characterized in that, The correction priority of the abnormal reference point of the fault sound source is obtained based on the positioning correction intensity and signal transmission distance, including: Preset correction strength weights and transmission distance weights; The positioning correction coefficient is obtained based on the positioning correction strength and the correction strength weight. The spatial distance coefficient is obtained based on the signal transmission distance and the transmission distance weight. A comprehensive correction coefficient is obtained based on the positioning correction coefficient and the spatial distance coefficient, and the correction priority is set from high to low based on the comprehensive correction coefficient.

9. The sound source localization method for fault identification of a transmission device according to claim 8, characterized in that, The 3D positioning calibration results are mapped to the 3D model of the transmission device to complete the visualization output of the fault sound source, including: Import the 3D model of the transmission device; Map the fault coordinates in the 3D positioning calibration results to the corresponding positions in the 3D model; Display the fault location and probability in the form of marked points or heat maps, and output fault location information.

Images