Filter bag damage non-contact detection method and system based on acoustic array positioning

Abstract

The invention relates to the technical field of bag type dust collector filter bag abrasion detection, in particular to a filter bag damage non-contact detection method and system based on acoustic array positioning. The method comprises the following steps: arranging an acoustic sensor array in an air purification chamber of the dust remover, and collecting acoustic signals in real time; after band-pass filtering is carried out on the signals, feature vectors of the signals are extracted and matched with a pre-established acoustic fingerprint database, so that a filter bag damage event is recognized; after identification, based on the time difference of arrival of the signal to at least three sensors in the array, combining the coordinates of the sensors and the sound velocity, and utilizing a TDOA positioning algorithm to calculate the space coordinates of the damaged sound source; and finally, matching the coordinates with a pre-stored filter bag coordinate mapping table, determining the number of a specific damaged filter bag, and performing visual highlight marking on a pattern board plane layout of an application layer terminal. The online, real-time, automatic and accurate positioning of the damaged filter bag is realized, the problem that the prior art can only give an alarm but cannot position is solved, and the operation and maintenance efficiency and safety are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of filter bag wear detection of bag type dust collector, and particularly relates to a filter bag breakage non-contact detection method and system based on acoustic array positioning. BACKGROUND

[0002] The bag type dust collector is a core equipment in the field of industrial dust purification, and its dust removal efficiency depends on the filtering unit composed of thousands of filter bags. In the long-term operation, the filter bags will inevitably be damaged due to high temperature, corrosion, wear and repeated impact of pulse cleaning airflow. The damaged filter bags will cause the short circuit of high-concentration dust-containing airflow and direct discharge, which not only causes serious environmental pollution, but also makes the enterprise face environmental protection risks. Therefore, timely and accurate detection and positioning of damaged filter bags are the key to ensuring the stable and standard operation of the dust removal system. In the prior art, the detection methods for filter bag breakage mainly include the following types: Offline manual detection method: represented by fluorescent powder detection method, this method needs to stop the production line, manually throw powder and check each bag with the help of ultraviolet lamp, the whole process takes several hours to several days, causing huge production loss, and there are risks of personnel safety and chemical pollution. The manual auscultation method also depends on the experience of maintenance personnel, and has low accuracy in strong noise field and is invalid for micro breakage. Such methods are low in efficiency and cannot meet the needs of modern industrial continuous production.

[0003] Online macro alarm method: represented by differential pressure monitoring method and online dust concentration monitoring method, they can sense the abnormal state of the system (such as sudden change of system pressure drop and over-standard outlet dust concentration) and issue a macro alarm of fault. However, the core defect of these methods is that they can only judge whether breakage occurs, but cannot know which filter bag is broken, the key problem, and subsequent fault troubleshooting still needs to return to the offline manual detection, which fails to fundamentally solve the positioning problem.

[0004] In recent years, acoustic detection has shown potential as a non-contact method, but existing acoustic positioning technologies (such as CN119716736A and CN114046968B) are mostly for general spatial sound source positioning or for multiple equipment faults, and their algorithms and architectures lack specific adaptation to the specific acoustic characteristics (such as high-frequency airflow leakage sound) of filter bag breakage of the bag type dust collector and the complex industrial working conditions (high dust, high temperature and strong noise), making it difficult to achieve long-term stability and accurate positioning in this application scenario.

[0005] Therefore, there is a significant technical gap in the prior art: between macro online alarm and accurate offline positioning, there is a lack of a detection scheme that can online, real-time, automatically and accurately locate a single damaged filter bag. SUMMARY

[0006] The technical problems solved by the present application are: in order to solve the problems of lack of positioning ability, low efficiency, high cost, dependence on experience, poor accuracy, safety and environmental risk in the prior art in the background art, a filter bag breakage non-contact detection method based on acoustic array positioning is provided, which can detect and locate the specific filter bag that breaks in real time, online, automatically and accurately, so as to realize accurate positioning of faults, shorten fault troubleshooting time, reduce operation and maintenance cost and improve safety and environmental protection.

[0007] The technical solution adopted by the present application to solve its technical problems is: a filter bag breakage non-contact detection method based on acoustic array positioning, comprising the following steps: S1, arranging a plurality of acoustic sensors in a geometric array in the clean gas chamber of the dust collector, and collecting acoustic signals in real time; S2, band-pass filtering the collected acoustic signals to extract ultrasonic signals in a preset frequency band, extracting a feature vector of the ultrasonic signals, matching with a pre-established acoustic fingerprint library, and determining whether a filter bag breakage event occurs according to the matching result; S3, after identifying the filter bag breakage event, based on the time difference of arrival of the ultrasonic signals corresponding to the breakage event to at least three sensors, the preset coordinates of each sensor and the speed of sound, the spatial coordinates of the sound source are calculated by TDOA positioning algorithm; S4, matching the calculated sound source coordinates with the pre-stored filter bag coordinate mapping table to determine the corresponding broken filter bag number; S5, on the human-computer interaction interface of the application layer terminal, in the virtual flower board plane layout diagram, the position corresponding to the broken filter bag number is highlighted and marked in a graphical manner.

[0008] Through acoustic fingerprint library matching and TDOA positioning, the fault is improved from macroscopic judgment to filter bag level microscopic positioning, realizing high recognition accuracy and high positioning precision.

[0009] Further, the establishment of the acoustic fingerprint library in S2 comprises the following steps: First, collect acoustic signal samples of the filter bag in normal state and various simulated breakage states; Second, at least one feature including short-time energy, zero-crossing rate, spectral centroid and mel-frequency cepstral coefficient is extracted from these acoustic signal samples to form a feature vector set; Finally, the feature vector set is normalized and stored as a template with working condition and breakage state labels to form an acoustic fingerprint library.

[0010] By collecting samples under different states and different working conditions, the established fingerprint library contains rich on-site feature information, making the subsequent recognition model have strong anti-interference ability and generalization ability, thereby ensuring the long-term stable operation and high recognition rate of the system in complex industrial environments.

[0011] Further, the S2 identifies the broken filter bag event, specifically: Real-time feature extraction from ultrasonic signals, pattern matching with templates in the acoustic fingerprint library; when the matching degree exceeds the preset threshold, it is determined that a breakage event has occurred, and an event trigger signal is generated; the arrival time stamp of the breakage event's acoustic signal starting wave front at each sensor in the acoustic sensor array is accurately recorded.

[0012] The introduction of event triggering and accurate recording of starting wave front timestamp realizes the real-time and automation of detection, ensures the accuracy and synchronization of the time difference data relied on by the subsequent TDOA positioning algorithm, and avoids the positioning failure caused by signal processing delay or false triggering.

[0013] Further, the pattern matching uses a support vector machine classifier, or a method combining energy threshold and frequency spectrum matching.

[0014] The machine learning method using support vector machine (SVM) classifier can handle high-dimensional and complex acoustic features, achieve recognition accuracy and reliability far beyond simple threshold judgment, and effectively distinguish between breakage signals and background noise; the method combining energy threshold and frequency spectrum matching provides a high computational efficiency and rapid response alternative solution, suitable for scenarios with extremely high real-time requirements.

[0015] Further, the TDOA positioning algorithm in S3 solves the following equation set to calculate the sound source coordinates P(x, y): , , , wherein, , , is the preset coordinates of the three sensors, , , is the time difference relative to the reference sensor, and v is the sound speed corrected in real time by temperature.

[0016] By solving this hyperbolic equation set, abstract time difference data can be converted into specific two-dimensional spatial coordinates, and the sound speed is required to be temperature corrected, improving the accuracy and robustness of the positioning algorithm in variable working condition environments, solving the problem of not considering industrial site environmental variables in general acoustic positioning technology.

[0017] Further, in S4, the calculated sound source coordinates are matched with the pre-stored filter bag coordinate mapping table, specifically: the Euclidean distance between the sound source coordinates P(x, y) and each filter bag center coordinate in the mapping table is calculated, the filter bag with the smallest Euclidean distance is determined as the damaged filter bag, and its number is taken as the matching result.

[0018] Using the smallest Euclidean distance as the determination criterion, a high-efficiency and clear result matching mechanism can be realized, which can accurately match the sound source coordinates calculated by the algorithm, which may have a small error, with the only filter bag closest to the physical location, thereby discretizing the continuous coordinate space and finally outputting the specific filter bag number that can be directly understood by the maintenance personnel.

[0019] Further, the algorithm used to calculate the spatial coordinates of the sound source is a positioning algorithm based on received signal strength indication or a positioning algorithm based on angle of arrival.

[0020] Further, in S5, the highlighting marking method includes: uniquely marking the damaged filter bag on the flower plate layout with a specific color, a flashing animation or an icon.

[0021] By highlighting the marking, the abstract filter bag number is converted into an intuitive and eye-catching graphical interface, so that the maintenance personnel do not need to interpret data or drawings, greatly shortening the fault response time and reducing the skill requirements for personnel.

[0022] Also provided is a filter bag damage non-contact detection system for implementing the filter bag damage non-contact detection method based on acoustic array positioning in the above scheme, comprising: An acoustic sensor array arranged in the clean gas chamber of the bag-type dust collector for collecting acoustic signals; An edge computing signal processor in communication connection with the acoustic sensor array; An application layer terminal in communication with the edge computing signal processor for receiving the damaged filter bag number, and comprising a display module for rendering and displaying a virtual flower plate layout and an alarm and rendering module for highlighting marking the damaged filter bag position on the layout and triggering sound and / or light alarms and / or remote alarms.

[0023] Further, the edge computing signal processor comprises: An acoustic fingerprint library module for storing and retrieving an acoustic fingerprint library; A TDOA positioning calculation module for calculating the time difference of arrival according to the signal arrival time stamps of each sensor channel and performing a TDOA algorithm to solve the sound source coordinates; The coordinate mapping module internally stores a filter bag coordinate mapping table and is configured to determine a filter bag with the smallest distance as a damaged filter bag by calculating the Euclidean distance between the sound source coordinate and the center coordinate of each filter bag.

[0024] The present application has the following advantages: (1) The present application completely solves the problem that the existing online monitoring technology can only perform macro alarm but cannot perform accurate positioning. Through the combination of the acoustic array and the TDOA positioning algorithm, the system can directly output the physical position or unique number of the specific damaged filter bag, and accurate operation of fault troubleshooting is realized. (2) The present application realizes online, real-time and automatic monitoring and positioning. The entire detection process does not require the dust remover to be shut down, nor does it require personnel to enter the inside of the equipment. Maintenance personnel can directly and quickly replace the target filter bag during the planned shutdown period according to the accurate position information provided by the system, so that the troubleshooting and processing time caused by filter bag damage is shortened to the minute level, and the operation and maintenance efficiency is improved by orders of magnitude. (3) Since the present application realizes rapid and accurate positioning without the need for special shutdown for troubleshooting, it can maximize the avoidance of unplanned production shutdowns caused by waiting for fault troubleshooting, thereby ensuring the continuity and stability of the production process and achieving significant economic benefits. (4) The present application uses a completely non-contact acoustic detection method, which fundamentally avoids the high-risk operation of maintenance personnel entering the high-temperature, high-dust and closed space inside the dust remover, thereby improving the operation safety. At the same time, the method completely eliminates the use of chemical reagents such as fluorescent powder, thereby eliminating the secondary pollution problem of the dust removal system and the environment caused thereby, and is more in line with the requirements of modern industry for green and environmental protection. (5) The present application integrates the whole process of automatic signal acquisition, intelligent feature recognition, automatic positioning calculation and result visualization alarm, realizes the complete automation of filter bag state monitoring, reduces the dependence on the personal experience of maintenance personnel, and provides data support for the state monitoring, fault prediction and health management of the equipment with the accurate position data output by the system. BRIEF DESCRIPTION OF DRAWINGS

[0025] The present application will be further described below in conjunction with the drawings and examples.

[0026] Figure 1 is a flowchart of the filter bag damage non-contact detection method of the present application.

[0027] Figure 2 is a planar arrangement diagram of the acoustic sensor array in the filter bag damage non-contact detection method of the present application.

[0028] Figure 3 is a TDOA positioning and coordinate mapping schematic diagram in the filter bag damage non-contact detection method of the present application.

[0029] Figure 4 is a structural block diagram of the filter bag breakage non-contact detection system of the present application. DETAILED DESCRIPTION

[0030] The present application will now be further described in detail with reference to the drawings. These drawings are simplified schematic diagrams and only show the basic structure of the present application in a schematic manner, and thus only show the components relevant to the present application.

[0031] Example 1 As shown in Figure 1 , a filter bag breakage non-contact detection method based on acoustic array positioning includes the following steps: Step 1: In the clean gas chamber of the dust collector, a plurality of acoustic sensors are arranged in a geometric array, and acoustic signals are collected in real time; Step 2: The collected acoustic signals are band-pass filtered to extract ultrasonic signals in a predetermined frequency band, and the feature vectors of the ultrasonic signals are extracted, matched with a pre-established acoustic fingerprint library, and the occurrence of a filter bag breakage event is determined according to the matching result; Step 3: After identifying the filter bag breakage event, based on the time difference of arrival of the ultrasonic signal corresponding to the breakage event at at least three sensors, the predetermined coordinates of each sensor, and the speed of sound, the spatial coordinates of the sound source are calculated by TDOA positioning algorithm; Step 4: The calculated sound source coordinates are matched with the pre-stored filter bag coordinate mapping table to determine the corresponding broken filter bag number; Step 5: On the human-computer interaction interface of the application layer terminal, the position corresponding to the broken filter bag number is highlighted and marked in a graphical manner in the virtual flower board plane layout.

[0032] In step 1, at least three high-sensitivity ultrasonic microphone probes are arranged at a safe distance above the outlet plane of the filter bag on the top or side wall of the clean gas chamber of the dust collector. In this embodiment, four high-sensitivity ultrasonic microphone probes are used, as shown in Figure 2 , forming a rectangular shape; the four high-sensitivity ultrasonic microphone probes are all selected from MEMS or piezoelectric ultrasonic sensors with a frequency response range of 20 kHz to 60 kHz, which have the advantages of high sensitivity, strong anti-interference ability, small size, etc.; each ultrasonic microphone probe is packaged in an IP67 level high-temperature resistant (such as 150℃) and corrosion resistant (such as 316L stainless steel) shell to ensure long-term stable operation in harsh working conditions. During the system installation and debugging stage, a two-dimensional Cartesian coordinate system is established with one corner point of the flower board as the coordinate origin (0, 0), and the coordinates of the four ultrasonic microphone probes are accurately measured and recorded: , , , These coordinates are the basis for subsequent positioning calculations.

[0033] In step two, the establishment of the acoustic fingerprint library is specifically as follows: First, collect working condition data: collect at least 30 minutes of background sound in the new filter bag and normal operation stage, covering different air volume and temperature ranges; collect acoustic signals under different damage sizes, positions and air flow velocities by simulating damage in different positions of the filter bag (such as 1 mm, 3 mm, 5 mm, etc.) in a controlled experiment; collect mixed noise under different fan frequencies and dust concentration conditions to improve the model's anti-interference ability.

[0034] Second, extract acoustic features: extract features from each collected acoustic signal, first extract time domain features, the short-time energy extraction formula is: , and the zero-crossing rate is: ; then extract frequency domain features, extract spectral centroid, spectral width and spectral roll-off point; then extract voiceprint features, including: Mel-frequency cepstral coefficient (MFCC), which can simulate human auditory perception and adjust the Mel filter distribution in the ultrasonic frequency band; Chroma feature (pitch contour), and STFT time-frequency diagram, which can use convolutional neural network (CNN) to extract deep features from the obtained time-frequency diagram.

[0035] Next, generate acoustic fingerprint templates: first, combine the above time domain, frequency domain and time-frequency features into a unified feature vector F: , where d is the feature dimension; then use zero mean unit variance (Z-score) or Min-Max normalization to eliminate the differences between collection devices and environmental influences; then use PCA or t-SNE to reduce the dimension of the features to generate compact fingerprint templates for fast database retrieval; then store and label each template, including working condition type, collection working condition parameters and timestamp, where the working condition type includes normal, slight damage and severe damage, and the collection working condition parameters include air volume, temperature and filter bag number; finally, use a vector database (such as Faiss, Annoy) for storage, which supports fast similarity retrieval of high-dimensional features, and the similarity retrieval formula is: , where is the current detection feature vector, is the template vector, and the similarity range is 0-1.

[0036] In step two, the filter bag damage event is identified, specifically: real-time extraction of features from ultrasonic signals and pattern matching with templates in the acoustic fingerprint library; when the matching degree exceeds the preset threshold, it is determined that a damage event has occurred, and an event trigger signal is generated; accurately record the arrival time stamp of the damage event acoustic signal starting wave front to each sensor in the acoustic sensor array. Among them, the pattern matching uses a support vector machine classifier, or a method combining energy threshold and spectral matching.

[0037] The TDOA positioning algorithm in step three calculates the sound source coordinate P(x, y) as follows: a sensor is selected as a reference, for example, the sensor S1(0, 0) is selected as the reference, and the time difference of arrival of sound waves at the other three sensors relative to the reference sensor is calculated as follows: Figure 3 The distance difference equations are established as follows: the distance difference between the sound source P and any two sensors is equal to the sound speed v multiplied by the time difference of arrival where v is the sound speed that is corrected in real time according to the temperature; then the above equations are solved, the above equations are nonlinear hyperbolic equations, and direct solving is complex; in actual engineering, efficient iterative algorithms such as Chan algorithm or Taylor series expansion least square method are used to solve the optimal estimation coordinate P(x, y) of the sound source; Finally, the Euclidean distance between the sound source coordinate P(x, y) and the center coordinate of each filter bag in the mapping table is calculated, the filter bag with the smallest Euclidean distance is determined as the damaged filter bag, and the number of the damaged filter bag is taken as the matching result, for example, the filter bag number R5-C12.

[0038] The highlighted marking method in step five includes uniquely marking the damaged filter bag on the flower plate layout with a specific color, a flashing animation or an icon.

[0039] In order to verify the effectiveness of the embodiment, the following simulation experiment is performed: The experimental environment is as follows: filter bags are arranged on a 5mx4m simulated clean room flower plate, four ultrasonic sensors are installed 1.5m above the flower plate, the coordinates of the four ultrasonic sensors are S1(0, 0), S2(5, 0), S3(5, 4) and S4(0, 4), the environmental temperature is 25℃, the sound speed v≈346m / s, compressed air is used to simulate the leakage of the damaged filter bag at the specified position through a Ø2mm nozzle, and the positioning accuracy test data are shown in Table 1.

[0040] Table 1: Positioning accuracy test data

[0041] ​​​​​​​​​According to the data comparison in Table 1, the peak deviation of the system positioning data from the actual damage location is 4.5 cm; combined with the actual working conditions of industrial dust removal equipment, the filter bag center spacing generally adopts an engineering layout of ≥20 cm; this deviation range is only 22.5% of the spacing value, meaning that in a three-dimensional coordinate system, the shortest distance between adjacent filter bags is 4.4 times the system error, which has met the engineering demand for accurately locking a single damaged filter bag; through actual working condition simulation tests, the experimental results have proved the accuracy and feasibility of the technical scheme of the embodiment.

[0042] The basic principle of the embodiment is that when the filter bag is damaged, under the action of the pressure difference inside and outside the dust collector, the high-pressure airflow carries the dust from the damage and sprays it out in the form of a turbulent jet flow at high speed, the jet flow rubs and shears with the surrounding static or low-speed air, and periodic vortex shedding occurs at the edge of the damage, this physical process will excite high-frequency sound waves with a specific frequency band (mainly 20 kHz-60 kHz), the sound wave signal has concentrated energy, good directivity, and a frequency much higher than common mechanical vibration and fan noise (usually <1 kHz), which provides a physical premise for effectively extracting leakage characteristics from strong background noise. It is very difficult to judge damage directly from the original acoustic signal, the embodiment introduces an acoustic fingerprint library as intelligent recognition, through a large number of experiments in advance, the physical characteristics (time domain energy, frequency domain spectrum structure, time-frequency distribution) of the leakage sound waves under different damage states (such as hole size, shape) and different working conditions (such as pressure difference, air volume) are digitally modeled to form a high-dimensional feature vector template. In real-time detection, through pattern matching or similarity calculation, the real-time signal characteristics are compared with the fingerprint library template, this process converts a complex physical discrimination problem into a quantifiable classification and identification problem based on data driving, so as to realize reliable and automatic detection of damage events. Once the damage sound source is identified, accurate positioning relies on acoustic array positioning technology, i.e. the TDOA (Time Difference of Arrival) principle, such as Figure 3As shown, the sound wave propagates outward from the break point P in the form of a spherical wave. Due to the different spatial positions of the sound source and the sensors in the array, there is a slight difference in the time when the sound wave reaches each sensor. Under the condition that the sound speed v is known, each time difference corresponds to a rotating hyperboloid with a constant distance difference between the sound source and the two sensors. A spatial array composed of at least three sensors can obtain multiple such hyperboloids, and the intersection of the hyperboloids is the spatial coordinates of the sound source P. By solving the hyperbolic equation set and converting the time measurement value into the spatial coordinate value, the location of the sound source is found from the sound heard. The calculated sound source coordinates are a continuous point in the physical space. In order to correspond them to the discrete filter bag numbers, the nearest neighbor algorithm is used in this embodiment. During system initialization, the coordinates of the center of each filter bag in the same coordinate system are accurately measured and stored in advance. After positioning, the Euclidean distance between the sound source coordinates and all filter bag center coordinates is calculated, and the filter bag with the smallest distance is determined as the break source. As long as the positioning accuracy of the system is much smaller than the distance between the filter bags, the uniqueness and accuracy of the mapping result can be guaranteed.

[0043] Embodiment Two The difference from Embodiment One is that instead of TDOA positioning algorithm, based on Received Signal Strength Indication (RSSI), the signal strength received by different probes is compared to locate the break point. The closer the probe is to the break point, the stronger the signal. This method is simple to implement, but the positioning accuracy is lower than TDOA, and is easily affected by sound wave reflection and attenuation.

[0044] Embodiment Three The difference from Embodiment One is that instead of TDOA positioning algorithm, based on Angle of Arrival (AOA), a directional microphone array (such as a phased array) is used. The angle of arrival of the signal to the array can be measured to locate the break point. This method requires more complex sensors and algorithms, and the cost is higher.

[0045] Embodiment Four The difference from Embodiment One is that all the raw data of the probes are transmitted to a central server for processing, instead of at the edge. This approach is convenient for unified management and algorithm upgrading, but requires high network bandwidth and has data transmission delay, which may affect real-time performance.

[0046] Embodiment Five The difference from Embodiment One is that in certain working conditions, high-frequency vibration sensors are used. When air leakage occurs, it may cause specific high-frequency vibrations of the screen or nearby structures. By analyzing the vibration signals, it is also possible to attempt to locate the break point, but this is a contact-type measurement, which is more complex to install and has a more complex signal transmission path, making it more difficult to locate.

[0047] Embodiments two to five can be used as supplements or variants of embodiment one in different application scenarios, but the combination of the ultrasonic array, TDOA algorithm and edge computing in embodiment one achieves the best balance in terms of accuracy, real-time performance, cost and reliability.

[0048] Embodiment six As shown in Figure 4 A filter bag breakage non-contact detection system for implementing the filter bag breakage non-contact detection method based on acoustic array positioning in embodiment one, comprising an acoustic sensor array, an edge computing signal processor and an application layer terminal, wherein the acoustic sensor array is arranged in the clean gas chamber of the bag-type dust collector for collecting acoustic signals; the edge computing signal processor is in communication connection with the acoustic sensor array; the application layer terminal is in communication with the edge computing signal processor for receiving the number of the broken filter bag, and comprises a display module for rendering and displaying a virtual flower board layout diagram, and an alarm and rendering module for highlighting the position of the broken filter bag on the layout diagram and triggering sound and / or light alarms and / or remote alarms.

[0049] Specifically, the edge computing signal processor comprises an acoustic fingerprint library module, a TDOA positioning solution module and a coordinate mapping module, the acoustic fingerprint library module is used for storing and retrieving the acoustic fingerprint library; the TDOA positioning solution module is used for calculating the time difference of arrival according to the time stamps of the signals of each sensor channel, and performing TDOA algorithm to solve the coordinates of the sound source; the coordinate mapping module has a filter bag coordinate mapping table stored therein, and is configured to determine the filter bag with the smallest distance as the broken filter bag by calculating the Euclidean distance between the coordinates of the sound source and the coordinates of the center of each filter bag.

[0050] Based on the above ideal embodiments according to the present application, the relevant personnel can make various changes and modifications without deviating from the technical idea of the present application. The technical scope of the present application is not limited to the contents of the specification, and must be determined according to the scope of the claims.

Claims

1. A non-contact detection method for filter bag damage based on acoustic array positioning, characterized in that, Includes the following steps: S1. In the clean air chamber of the dust collector, multiple acoustic sensors are arranged in a geometric array and acoustic signals are collected in real time. S2. Bandpass filtering is performed on the collected acoustic signal to extract the ultrasonic signal in the preset frequency band, and the feature vector of the ultrasonic signal is extracted and matched with the pre-established acoustic fingerprint database. Based on the matching result, it is determined whether a filter bag breakage event has occurred. S3. After identifying the filter bag damage event, based on the arrival time difference of the ultrasonic signal corresponding to the damage event to at least three sensors, the preset coordinates of each sensor and the sound velocity, the spatial coordinates of the sound source are calculated by the TDOA positioning algorithm. S4. Match the calculated sound source coordinates with the pre-stored filter bag coordinate mapping table to determine the corresponding damaged filter bag number. S5. On the human-computer interaction interface of the application layer terminal, the location corresponding to the number of the damaged filter bag is highlighted in a graphical manner in the virtual flower board layout diagram.

2. The non-contact detection method for filter bag damage based on acoustic array positioning according to claim 1, characterized in that, The establishment of the acoustic fingerprint database in S2 includes the following steps: First, acoustic signal samples of the filter bags were collected under normal conditions and various simulated damage conditions; Secondly, at least one feature among short-time energy, zero-crossing rate, spectral centroid, and Mel frequency cepstral coefficients is extracted from these acoustic signal samples to form a feature vector set. Finally, the feature vector set is normalized and stored as a template with labels for operating conditions and damage status, forming an acoustic fingerprint database.

3. The non-contact detection method for filter bag damage based on acoustic array positioning according to claim 1, characterized in that, S2 identifies filter bag damage events specifically as follows: Features are extracted from ultrasonic signals in real time and pattern matched with templates in the acoustic fingerprint database. When the matching degree exceeds a preset threshold, a damage event is determined to have occurred, and an event trigger signal is generated. The arrival timestamps of the initial wavefront of the acoustic signal of the damage event to each sensor in the acoustic sensor array are accurately recorded.

4. The non-contact detection method for filter bag damage based on acoustic array positioning according to claim 3, characterized in that, The pattern matching employs a support vector machine classifier, or a combination of energy threshold and spectrum matching.

5. The non-contact detection method for filter bag damage based on acoustic array positioning according to claim 1, characterized in that, In S3, the TDOA localization algorithm calculates the sound source coordinates P(x, y) by solving the following system of equations: ， ， ， in, , , The preset coordinates for the three sensors, , , v is the arrival time difference relative to the reference sensor, and v is the speed of sound corrected for temperature in real time.

6. The non-contact detection method for filter bag damage based on acoustic array positioning according to claim 5, characterized in that, In S4, the calculated sound source coordinates are matched with the pre-stored filter bag coordinate mapping table. Specifically, the Euclidean distance between the sound source coordinates P(x, y) and the center coordinates of each filter bag in the mapping table is calculated. The filter bag with the smallest Euclidean distance is identified as the damaged filter bag, and its number is used as the matching result.

7. The non-contact detection method for filter bag damage based on acoustic array positioning according to claim 1, characterized in that, The algorithm used to calculate the spatial coordinates of the sound source is either a positioning algorithm based on the received signal strength indication or a positioning algorithm based on the angle of arrival.

8. The non-contact detection method for filter bag damage based on acoustic array positioning according to claim 1, characterized in that, The S5 highlights the damaged filter bag by using a specific color, flashing animation, or icon to uniquely mark it on the ceiling plan.

9. A non-contact filter bag damage detection system, used to implement the non-contact filter bag damage detection method based on acoustic array positioning as described in any one of claims 1 to 8, characterized in that, include: An acoustic sensor array is arranged in the clean air chamber of a bag filter to collect acoustic signals; An edge computing signal processor is communicatively connected to the acoustic sensor array; The application layer terminal communicates with the edge computing signal processor to receive the damaged filter bag number, and includes a display module for rendering and displaying a virtual flower board layout diagram, and an alarm and rendering module for highlighting the location of the damaged filter bag on the layout diagram and triggering audible and visual and / or remote alarms.

10. A non-contact detection system for filter bag damage according to claim 9, characterized in that, The edge computing signal processor includes: The acoustic fingerprint database module is used to store and retrieve acoustic fingerprints. The TDOA positioning and solving module is used to calculate the time difference of arrival based on the signal arrival timestamps of each sensor channel and to execute the TDOA algorithm to solve for the sound source coordinates. The coordinate mapping module stores a filter bag coordinate mapping table and is configured to determine the filter bag with the smallest Euclidean distance from the sound source coordinates to the center coordinates of each filter bag by calculating the Euclidean distance between the sound source coordinates and the center coordinates of each filter bag.

Images