An acoustic imaging gas leak risk grading alarm method and system

Abstract

The application discloses an acoustic imaging gas leakage risk grading alarm method, which comprises the following steps: synchronously collecting acoustic signals, visual signals and environmental auxiliary parameters of a region to be monitored; performing background noise suppression and interference filtering on the collected acoustic signals; performing leakage feature extraction on the interference-suppressed acoustic signals, and judging whether to generate a pre-warning trigger signal or an alarm trigger signal according to the extracted features; after receiving the pre-warning trigger signal or the alarm trigger signal, calling a multi-source data to calculate a risk comprehensive score, and dividing a risk grade according to the risk comprehensive score; and outputting a differentiated alarm signal according to the divided risk grade, and matching a treatment requirement or a linkage control action corresponding to the risk grade. The application also discloses an acoustic imaging gas leakage risk grading alarm system for realizing the above method. The application realizes the grading quantitative evaluation of leakage risks, significantly reduces the false alarm rate of non-dangerous emissions, and realizes the accurate matching of alarms and emergency responses.

Description

Technical Field

[0001] This invention relates to the field of gas leak monitoring technology, and in particular to an acoustic imaging gas leak risk classification alarm method and system. Background Technology

[0002] In the fields of petrochemicals, energy pipelines, and natural gas storage and transportation, leaks of pressurized gases can easily lead to explosions, poisoning, and even environmental pollution accidents. Monitoring flammable, toxic, and hazardous gases is a core control objective for safe production. Traditional detection methods (such as combustible gas probes) are slow and struggle to quickly locate hidden leak points; manual inspections are inefficient and risky, failing to meet the needs of 24 / 7, large-scale monitoring. With the increasing safety and environmental protection demands for gas leak detection in the industrial sector, and the shortcomings of existing technologies in intelligent and refined management, these issues need to be addressed.

[0003] Acoustic imaging technology uses microphone arrays to capture ultrasonic / audible sound signals generated by the turbulence of leaking gas. It then generates an acoustic image using beamforming algorithms and overlays it with visible light images, enabling non-contact, visual, and rapid location of the leak source, and is gradually becoming a mainstream detection method. In recent years, the rise of acoustic imaging technology has provided a completely new solution for gas leak detection. By capturing the sound waves generated by the leak using microphone arrays and combining them with beamforming algorithms to convert the sound signals into visual images, it allows for "seeing sound with your eyes," enabling long-distance, non-contact location of the leak source. However, existing acoustic imaging equipment still has the following shortcomings in gas leak monitoring applications: First, the alarm mechanism is simplistic. Most mainstream acoustic imaging devices currently use a fixed sound pressure threshold to trigger alarms. When the detected sound pressure level exceeds the preset threshold, an alarm is issued. This binary "present / absent" alarm cannot distinguish the severity of the leak. Minor leaks and major leaks will both generate the same alarm signal, making it difficult for maintenance personnel to determine the priority of handling. Secondly, there is a lack of risk assessment capabilities. The risk of gas leaks depends not only on the size of the leak, but also on factors such as the properties of the leaked gas (flammable, toxic), the location of the leak, and environmental conditions. Existing equipment only provides approximate location of the leak and sound pressure level data, without integrating multi-dimensional information to form a comprehensive risk level assessment. Third, it cannot distinguish between hazardous and non-hazardous emissions. Industrial sites contain numerous non-hazardous gas emission sources, such as overpressure valve releases, flare discharges, and compressed air purging. These emissions also generate significant acoustic signals, leading to false alarms from traditional acoustic detectors. To avoid frequent false alarms, on-site personnel are often forced to raise the alarm threshold or move the detectors away, which actually reduces the sensitivity to actual hazardous leaks. Fourth, handheld devices are difficult to use for continuous monitoring. The current market is dominated by handheld acoustic imagers, which are suitable for periodic inspections, but cannot provide 24 / 7 uninterrupted monitoring of critical equipment. For intermittent leaks or sudden leaks at night, handheld inspection methods may miss the best time to deal with them.

[0004] Therefore, developing an online acoustic imaging method capable of risk-based alarm for gas leaks, enabling continuous monitoring, automatic risk assessment, and graded early warning of leaks, has significant theoretical and engineering application value.

[0005] Therefore, through beneficial exploration and research, the applicant has found a solution to the above problems, and the technical solution to be introduced below is the result of this research. Summary of the Invention

[0006] One of the technical problems to be solved by this invention is to provide an acoustic imaging gas leak risk classification alarm method to address the shortcomings of existing technologies. By automatically adjusting the time for stable imaging in the target area and setting the sound pressure value to be greater than the national standard, risk classification early warning and quantitative detection can be achieved. This solves the problems of conventional methods failing to detect potential problems in a timely manner, low detection accuracy, and incomplete functions, thereby improving alarm accuracy and emergency response efficiency when gas leaks occur, and promoting refined safety management in industrial scenarios.

[0007] The second technical problem to be solved by the present invention is to provide an acoustic imaging gas leak risk classification alarm system that implements the above-mentioned acoustic imaging gas leak risk classification alarm method.

[0008] As a first aspect of the present invention, an acoustic imaging gas leak risk classification alarm method includes: Simultaneously collect acoustic signals, visual signals, and environmental auxiliary parameters of the area to be monitored; Background noise suppression and interference filtering are performed on the collected acoustic signals to enhance the identification of the leakage signals; Leakage features are extracted from the acoustic signal after interference suppression, and a warning trigger signal or an alarm trigger signal is generated based on the extracted features. Upon receiving the warning trigger signal or the alarm trigger signal, multi-source data is used to calculate a comprehensive risk score, and the risk level is classified based on the comprehensive risk score; and Based on the risk level classification, output differentiated alarm signals and match the corresponding handling requirements or linkage control actions.

[0009] In a preferred embodiment of the present invention, the step of extracting leakage features from the acoustic signal after interference suppression and determining whether to generate a warning trigger signal or an alarm trigger signal based on the extracted features includes: Extract the sound pressure characteristics, frequency characteristics, and time sequence characteristics of the leakage sound signal; The extracted sound pressure features, frequency features, and temporal features are compared with the pre-stored regional gas leakage acoustic features. A leakage trend factor is generated based on the comparison results; As the leakage trend factor gradually increases, the leakage mass flow rate and cumulative leakage amount are calculated; and The warning trigger signal or the alarm trigger signal is generated based on the calculation results.

[0010] In a preferred embodiment of the present invention, the extracted sound pressure features, frequency features, and temporal features are compared with pre-stored regional gas leakage acoustic features, including: The extracted features are compared with a preset reference acoustic profile database to identify hazardous gas leaks and eliminate false alarms of non-hazardous emissions.

[0011] In a preferred embodiment of the present invention, the reference acoustic profile database pre-stores acoustic features of hazardous gas leaks and acoustic features of non-hazardous emissions; the acoustic features of hazardous gas leaks include acoustic features of hydrogen, methane, and ammonia; the acoustic features of non-hazardous emissions include acoustic features of compressed air purging and acoustic features of overpressure valve venting.

[0012] In a preferred embodiment of the present invention, the step of calculating a comprehensive risk score by calling multi-source data after receiving the early warning trigger signal or the alarm trigger signal, and classifying the risk level according to the comprehensive risk score, includes: Upon receiving the warning trigger signal or the alarm trigger signal, multi-source data is invoked; Read real-time leakage trend factors; As the real-time leakage trend factor gradually increases, it becomes crucial to identify the type of leaking gas and the equipment where the leak occurred. Assess the area and extent of the impact as the leak spreads; A comprehensive risk score is calculated based on the identification and assessment results; Risk levels are determined based on the comprehensive risk score.

[0013] In a preferred embodiment of the present invention, the calculation of the comprehensive risk score includes: Multiple factors, including leakage amount, gas hazard parameters, environmental diffusion factor, leakage trend factor, and equipment importance factor, are collected and input into a preset risk classification model to calculate the comprehensive risk score.

[0014] In a preferred embodiment of the present invention, the risk classification model is calculated using a multi-level weighted scoring and fuzzy logic hybrid reasoning method; the comprehensive risk score is calculated as follows: Risk = Leakage factor × Gas hazard factor × Environmental diffusion factor × Equipment sensitivity factor, or by weighted summation.

[0015] In a preferred embodiment of the present invention, it further includes: Based on the risk level, the system pushes alarm information and handling requirements corresponding to the risk level to the terminal and automatically triggers at least one of the following linkage control actions: starting the ventilation system, starting the sprinkler system, shutting off the upstream valve, and triggering the emergency shutdown system.

[0016] In a preferred embodiment of the present invention, it further includes: Record the actual handling results before and after the alarm, optimize the weight and classification threshold of the risk classification model periodically based on the actual handling results, and supplement the newly emerging non-hazardous emission acoustic features into the reference acoustic profile database.

[0017] As a second aspect of the present invention, an acoustic imaging gas leak risk classification alarm system includes: A multi-source data acquisition module, which is used to simultaneously acquire acoustic signals, visual signals and environmental auxiliary parameters of the area to be monitored; An interference suppression module is used to suppress background noise and filter out interference from the acquired acoustic signal. The leakage feature extraction and validity determination module is used to extract leakage features from the acoustic signal after interference suppression, and determine whether to generate a warning trigger signal or an alarm trigger signal based on the extracted features. A multi-dimensional risk grading module, which, upon receiving the early warning trigger signal or the alarm trigger signal, calls multi-source data to calculate a comprehensive risk score and classifies the risk level based on the comprehensive risk score; and The differentiated alarm and visualization module is used to output differentiated alarm signals according to the risk level classification, and overlay and display audio-visual images and visible light images.

[0018] Compared with the prior art, the present invention has the following beneficial effects: 1. Achieve graded and quantitative assessment of leakage risk. This invention overcomes the limitation of existing acoustic imaging equipment, which can only determine whether a leak has occurred using binary methods. By integrating multiple factors such as leakage amount, gas hazard, environmental diffusion conditions, equipment importance, and leakage trend, a comprehensive risk index is calculated and risk levels are classified, enabling maintenance personnel to accurately determine the severity of the leak and the priority of its handling.

[0019] 2. Significantly reduces the false alarm rate of non-hazardous emissions. This invention uses a preset reference acoustic profile database to compare the real-time extracted acoustic features with the pre-stored hazardous gas leak features and non-hazardous emission features in the database, effectively distinguishing between hazardous leaks and non-hazardous emissions, and avoiding frequent false alarms caused by common non-hazardous emissions in industrial sites.

[0020] 3. Supports continuous online monitoring and adaptive optimization. This invention adopts an online system architecture, supporting uninterrupted monitoring around the clock, avoiding the omission of intermittent or nighttime leaks due to the inspection cycle of handheld devices; at the same time, it regularly optimizes the weights and classification thresholds of the risk classification model through a closed-loop feedback mechanism, and dynamically updates the reference acoustic profile database to achieve adaptive improvement of system performance.

[0021] 4. Achieve precise matching between alarms and emergency responses. This invention outputs differentiated alarm signals based on the classified risk levels, pushes alarm information and handling requirements corresponding to the risk level to the terminal, and automatically triggers at least one linkage control action, such as starting the ventilation system, starting the sprinkler system, shutting off the upstream valve, or triggering the emergency shutdown system. This achieves strong matching between alarms and risk levels and handling priorities, thereby improving emergency response efficiency. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of the acoustic imaging gas leak risk classification alarm method of the present invention.

[0024] Figure 2 This is a flowchart of the leakage detection process of the present invention.

[0025] Figure 3 This is a flowchart of the multi-dimensional risk classification process of the present invention.

[0026] Figure 4 This is a schematic diagram of the acoustic imaging gas leak risk classification alarm system of the present invention. Detailed Implementation

[0027] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below with reference to specific illustrations.

[0028] This embodiment uses the natural gas pipeline monitoring area of ​​a chemical plant as an example to provide a detailed description of the acoustic imaging gas leak risk classification alarm method and system of the present invention.

[0029] See Figure 1 The acoustic imaging gas leak risk classification alarm method of this embodiment includes the following steps: Step S10: Synchronously acquire acoustic signals, visual signals, and environmental auxiliary parameters of the area to be monitored. In this embodiment, acoustic signals, visual signals, and environmental auxiliary parameters of the area to be monitored are acquired synchronously at a rate of 10 frames per second. Specifically, a 64-channel microphone array is used to acquire acoustic signals generated by leakage, a high-definition visible light camera is used to acquire on-site video images, and temperature, humidity, and wind speed data are acquired using temperature and humidity sensors and wind speed sensors.

[0030] Step S20 involves background noise suppression and interference filtering of the acquired acoustic signal to enhance the discernibility of the leakage signal. In this embodiment, a generalized sidelobe canceller adaptive beamforming algorithm is used to suppress background noise and filter interference from the acquired acoustic signal, thereby enhancing the discernibility of the leakage signal. After interference suppression processing, the signal-to-noise ratio of the leakage signal is improved by approximately 12 dB.

[0031] Step S30: Extract leakage features from the acoustic signal after interference suppression, and determine whether to generate a warning trigger signal or an alarm trigger signal based on the extracted features.

[0032] See Figure 2 Step S30 specifically includes the following sub-steps: Step S31: Extract the sound pressure level, frequency characteristics, and timing characteristics of the leakage sound signal. In this embodiment, the extracted sound pressure level is 85 dB, the center frequency is 35 kHz, and the timing characteristics show that the sound pressure level increases for five consecutive frames.

[0033] Step S32 involves comparing the extracted sound pressure, frequency, and timing features with pre-stored regional gas leak acoustic features. Specifically, the extracted features are compared with a preset reference acoustic profile database to identify hazardous gas leaks and eliminate false alarms of non-hazardous emissions. The reference acoustic profile database pre-stores acoustic features of hazardous gas leaks and non-hazardous emissions. Hazardous gas leak acoustic features include those of hydrogen, methane, and ammonia, while non-hazardous emission acoustic features include those of compressed air purging and overpressure valve venting. In this embodiment, the extracted features match the acoustic features of a methane leak but do not match any non-hazardous emission acoustic features, thus identifying it as a hazardous gas leak.

[0034] Step S33: Generate a leakage trend factor based on the comparison results. In this embodiment, the leakage trend factor is 0.7, showing an increasing trend.

[0035] Step S34: As the leakage trend factor gradually increases, calculate the leakage mass flow rate and cumulative leakage amount. Based on the theoretical model of sound pressure level and leakage amount, the instantaneous leakage mass flow rate is calculated to be 0.5 g / s, and the cumulative leakage amount reaches 10 kg.

[0036] Step S35: Generate a warning trigger signal or an alarm trigger signal based on the calculation results. In this embodiment, an alarm trigger signal is generated.

[0037] Step S40: After receiving a warning trigger signal or an alarm trigger signal, the system calls up multi-source data to calculate the comprehensive risk score and classifies the risk level based on the comprehensive risk score.

[0038] See Figure 3 Upon receiving an alarm trigger signal, follow these sub-steps: Step S41: Call multi-source data to obtain data such as gas type, leakage amount, and environmental parameters.

[0039] Step S42: Read the real-time leakage trend factor. In this embodiment, the real-time leakage trend factor is 0.7 and shows a gradually increasing trend.

[0040] Step S43: As the real-time leakage trend factor gradually increases, identify the type of leaking gas and the criticality of the equipment where the leak occurred. In this embodiment, the leaking gas type is methane, and the equipment where the leak occurred is a natural gas main pipeline, which is a highly critical piece of equipment.

[0041] Step S44: Assess the area and extent of the impact of the leak diffusion. Based on the wind speed (0.5 m / s) and temperature (25°C) data collected by environmental sensors, and combined with the gas diffusion model, the estimated impact range of the leak diffusion is approximately 50 square meters.

[0042] Step S45: Calculate the comprehensive risk score based on the identification and assessment results. The calculation of the comprehensive risk score includes: aggregating multiple factors from leakage amount, gas hazard parameters, environmental diffusion factors, leakage trend factors, and equipment importance factors, and inputting these aggregated factors into a preset risk classification model.

[0043] In this embodiment, the values ​​of each factor are as follows: Leakage factor: 0.7; Gas hazard factor: Methane, lower explosive limit is 5%, toxicity level is low toxicity, factor value is 0.6; Environmental diffusion factor: indoor environment, poor ventilation, factor value is 0.8; Equipment sensitivity factor: natural gas main pipeline, factor value is 0.9.

[0044] The risk grading model uses a multi-level weighted scoring method combined with fuzzy logic reasoning for calculation. In this embodiment, the comprehensive risk score is calculated using a multiplicative method: Risk = 0.7 × 0.6 × 0.8 × 0.9 = 0.3024.

[0045] In addition, the risk grading model can also calculate the comprehensive risk score using a weighted summation method. The calculation formula is: Risk = w1 × Leakage Factor + w2 × Gas Hazard Factor + w3 × Environmental Diffusion Factor + w4 × Equipment Sensitivity Factor, where w1, w2, w3, and w4 are preset weight coefficients, and the sum of each coefficient is 1.

[0046] The preset weighting coefficients are: w1=0.4, w2=0.3, w3=0.2, w4=0.1. The values ​​of each factor are the same as in Example 1, so the overall risk score is 0.4×0.7+0.3×0.6+0.2×0.8+0.1×0.9=0.71. This score is classified as a high-risk level, and the system outputs a high-risk alarm signal and matches the corresponding handling requirements.

[0047] Among them, the leakage trend factor is obtained by the rate of change of leakage amount over time, and is used to determine whether the leakage is in a stable state, an increasing state, or a decreasing state; the gas hazard parameter includes at least one of the following: lower explosive limit, concentration that immediately endangers life or health, and toxicity level; the environmental diffusion factor is used to characterize the area and range of the leakage diffusion effect.

[0048] Step S46: Classify the risk level based on the comprehensive risk score. Preset thresholds: low risk (<0.2), medium risk (0.2-0.5), high risk (0.5-0.8), and significant risk (≥0.8). In this embodiment, a comprehensive risk score of 0.3024 is classified as medium risk.

[0049] Step S50: Based on the classified risk level, output differentiated alarm signals and match the corresponding handling requirements or linkage control actions. Specifically, based on the medium risk level, push alarm information and handling requirements corresponding to the medium risk level to the terminal: the central control room screen displays a yellow alarm pop-up window with the content "Natural gas medium risk leak (0.5g / s), it is recommended to handle within 2 hours"; at the same time, push the same information to the duty mobile terminal. In addition, automatically trigger linkage control actions: start the local exhaust ventilation system.

[0050] The acoustic imaging gas leak risk classification and alarm method in this embodiment also includes a closed-loop feedback optimization step. This step records the actual handling results before and after the alarm, periodically optimizes the weights and classification thresholds of the risk classification model based on the actual handling results, and supplements the reference acoustic profile database with newly emerging non-hazardous emission acoustic features. The system ran continuously for 30 days, recording three suspected alarms caused by compressed air purging. After manual confirmation that these were non-hazardous emissions, the acoustic features of the compressed air purging were supplemented into the reference acoustic profile database. Simultaneously, based on historical handling data, the lower limit threshold for the risk level of methane was adjusted from 0.20 to 0.18, and the upper limit threshold was adjusted from 0.50 to 0.48. After optimization, no false alarms occurred in the subsequent 15 days, and the false alarm rate decreased by approximately 40%.

[0051] See Figure 4 This embodiment employs an acoustic imaging gas leak risk classification alarm system to implement the above method. The system includes a multi-source data acquisition module 100, an interference suppression module 200, a leak feature extraction and validity determination module 300, a multi-dimensional risk classification module 400, and a differentiated classification alarm and visualization display module 500.

[0052] The multi-source data acquisition module 100 is used to simultaneously acquire acoustic signals, visual signals, and environmental auxiliary parameters of the area to be monitored. In this embodiment, the module employs a 64-channel microphone array, a high-definition visible light camera, and temperature, humidity, and wind speed sensors.

[0053] The interference suppression module 200 is connected to the multi-source data acquisition module 100 and is used to suppress background noise and filter interference from the acquired acoustic signals.

[0054] The leakage feature extraction and validity determination module 300 is connected to the interference suppression module 200. It is used to extract leakage features from the acoustic signal after interference suppression and determine whether to generate a warning trigger signal or an alarm trigger signal based on the extracted features.

[0055] The multi-dimensional risk classification module 400 is connected to the leakage feature extraction and validity determination module 300. It is used to calculate the comprehensive risk score by calling multi-source data after receiving a warning trigger signal or an alarm trigger signal, and to classify the risk level according to the comprehensive risk score.

[0056] The differentiated alarm and visualization module 500 is connected to the multi-dimensional risk classification module 400. It outputs differentiated alarm signals based on the classified risk levels and overlays audio-visual images and visible light images. In this embodiment, the module includes a display screen used to mark leakage areas with different colors according to different risk levels; for example, medium-risk areas are marked in yellow, and high-risk areas are marked in red.

[0057] The acoustic imaging gas leak risk classification alarm method and system of the present invention can be widely used in industrial sites with pressurized gas leak risks such as petrochemical, natural gas storage and transportation, energy pipeline corridors, and mines. It can adapt to harsh working conditions such as complex background noise, multi-source leak superposition, and extreme environments.

[0058] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. An acoustic imaging gas leak risk classification alarm method, characterized in that, include: Simultaneously collect acoustic signals, visual signals, and environmental auxiliary parameters of the area to be monitored; Background noise suppression and interference filtering are performed on the collected acoustic signals to enhance the identification of the leakage signals; Leakage features are extracted from the acoustic signal after interference suppression, and a warning trigger signal or an alarm trigger signal is generated based on the extracted features. Upon receiving the warning trigger signal or the alarm trigger signal, multi-source data is used to calculate the comprehensive risk score, and the risk level is classified according to the comprehensive risk score. as well as Based on the risk level classification, output differentiated alarm signals and match the corresponding handling requirements or linkage control actions.

2. The acoustic imaging gas leak risk classification alarm method as described in claim 1, characterized in that, The step of extracting leakage features from the acoustic signal after interference suppression and determining whether to generate a warning trigger signal or an alarm trigger signal based on the extracted features includes: Extract the sound pressure characteristics, frequency characteristics, and time sequence characteristics of the leakage sound signal; The extracted sound pressure features, frequency features, and temporal features are compared with the pre-stored regional gas leakage acoustic features. A leakage trend factor is generated based on the comparison results; As the leakage trend factor gradually increases, the leakage mass flow rate and cumulative leakage amount are calculated; and The warning trigger signal or the alarm trigger signal is generated based on the calculation results.

3. The acoustic imaging gas leak risk classification alarm method as described in claim 2, characterized in that, The extracted sound pressure features, frequency features, and temporal features are compared with pre-stored regional gas leakage acoustic features, including: The extracted features are compared with a preset reference acoustic profile database to identify hazardous gas leaks and eliminate false alarms of non-hazardous emissions.

4. The acoustic imaging gas leak risk classification alarm method as described in claim 3, characterized in that, The reference acoustic profile database pre-stores acoustic features of hazardous gas leaks and acoustic features of non-hazardous emissions; the acoustic features of hazardous gas leaks include the acoustic features of hydrogen, methane, and ammonia; the acoustic features of non-hazardous emissions include the acoustic features of compressed air purging and the acoustic features of overpressure valve venting.

5. The acoustic imaging gas leak risk classification alarm method as described in claim 1, characterized in that, Upon receiving the warning trigger signal or the alarm trigger signal, the process of calling multi-source data to calculate a comprehensive risk score and classifying risk levels based on the comprehensive risk score includes: Upon receiving the warning trigger signal or the alarm trigger signal, multi-source data is invoked; Read real-time leakage trend factors; As the real-time leakage trend factor gradually increases, it becomes crucial to identify the type of leaking gas and the equipment where the leak occurred. Assess the area and extent of the impact as the leak spreads; A comprehensive risk score is calculated based on the identification and assessment results; Risk levels are determined based on the comprehensive risk score.

6. The acoustic imaging gas leak risk classification alarm method as described in claim 5, characterized in that, The calculated risk score includes: Multiple factors, including leakage amount, gas hazard parameters, environmental diffusion factor, leakage trend factor, and equipment importance factor, are collected and input into a preset risk classification model to calculate the comprehensive risk score.

7. The acoustic imaging gas leak risk classification alarm method as described in claim 6, characterized in that, The risk grading model is calculated using a multi-level weighted scoring and fuzzy logic hybrid reasoning method; the comprehensive risk score is calculated as follows: Risk = Leakage factor × Gas hazard factor × Environmental diffusion factor × Equipment sensitivity factor, or by weighted summation.

8. The acoustic imaging gas leak risk classification alarm method as described in claim 1, characterized in that, Also includes: Based on the risk level, the system pushes alarm information and handling requirements corresponding to the risk level to the terminal and automatically triggers at least one of the following linkage control actions: starting the ventilation system, starting the sprinkler system, shutting off the upstream valve, and triggering the emergency shutdown system.

9. The acoustic imaging gas leak risk classification alarm method according to any one of claims 1 to 8, characterized in that, Also includes: Record the actual handling results before and after the alarm, optimize the weight and classification threshold of the risk classification model periodically based on the actual handling results, and supplement the newly emerging non-hazardous emission acoustic features into the reference acoustic profile database.

10. An acoustic imaging gas leak risk classification alarm system that implements the acoustic imaging gas leak risk classification alarm method as described in any one of claims 1 to 9, characterized in that, include: A multi-source data acquisition module, which is used to simultaneously acquire acoustic signals, visual signals and environmental auxiliary parameters of the area to be monitored; An interference suppression module is used to suppress background noise and filter out interference from the acquired acoustic signal. The leakage feature extraction and validity determination module is used to extract leakage features from the acoustic signal after interference suppression, and determine whether to generate a warning trigger signal or an alarm trigger signal based on the extracted features. A multi-dimensional risk grading module is used to calculate a comprehensive risk score by calling multi-source data after receiving the early warning trigger signal or the alarm trigger signal, and to classify the risk level according to the comprehensive risk score. as well as The differentiated alarm and visualization module is used to output differentiated alarm signals according to the risk level classification, and overlay and display audio-visual images and visible light images.

Images