Petrochemical tank area multi-mode detection explosion-proof method and petrochemical tank area inspection robot

Abstract

The application provides a petrochemical tank area multi-modal detection explosion-proof method and a petrochemical tank area inspection robot, and belongs to the technical field of industrial safety detection. The method comprises the following steps: according to a preset inspection path, performing multi-modal detection on a target petrochemical tank area to obtain multi-modal detection data, analyzing the multi-modal detection data item by item, mapping corresponding analysis results to a time-space unified coordinate system constructed according to time stamps calibrated in advance based on the multi-modal detection data, performing cross-collaborative verification of the rechecked analysis results according to the positions of the analysis results in the time-space unified coordinate system, judging whether the cross-collaborative verification is valid, calculating an abnormal index correlation weight coefficient if the cross-collaborative verification is valid, determining a correlation degree according to the calculation result, and outputting a corresponding explosion-proof disposal plan according to the correlation degree. The method realizes early identification of small leakage of hazardous gas, improves the discrimination accuracy of gas-liquid two-phase leakage characteristics, reduces the false alarm rate and saves maintenance resources.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of industrial safety detection, and relates to a petrochemical tank area multi-modal detection explosion-proof method and a petrochemical tank area inspection robot. BACKGROUND

[0002] The petrochemical tank area has safety hazards due to containing explosive and other dangerous gases, and needs to be continuously monitored. However, the existing inspection devices applied to the petrochemical tank area, the traditional equipment mostly adopts a single sensor working mode, which cannot simultaneously obtain optical imaging, infrared temperature field distribution and ultrasonic spectrum characteristics of gas leakage forms, and cannot establish a dynamic correlation model of multi-physical field data. These problems lead to that a small leakage is difficult to be identified early, the gas-liquid two-phase leakage characteristic discrimination precision is insufficient, not only the manual re-inspection frequency of the high-risk area is greatly increased, but also the maintenance resources are wasted due to a high false alarm rate, and the progress of the whole area unmanned inspection of the petrochemical device is seriously restricted. SUMMARY

[0003] The purpose of the present application is to solve the above problems existing in the prior art, and a petrochemical tank area multi-modal detection explosion-proof method is provided.

[0004] The purpose of the present application can be realized by the following technical scheme: a petrochemical tank area multi-modal detection explosion-proof method, comprising:

[0005] According to a preset inspection path, multi-modal detection is performed on a target petrochemical tank area to obtain multi-modal detection data;

[0006] The multi-modal detection data are analyzed item by item, and the corresponding analysis results are mapped to a time-space unified coordinate system constructed according to a time stamp calibrated in advance according to the multi-modal detection data;

[0007] According to the position of the analysis result in the time-space unified coordinate system, cross-collaborative verification of the re-inspection analysis result is performed, and whether the cross-collaborative verification is established is judged.

[0008] If the cross-collaborative verification is established, an abnormal index correlation weight coefficient is calculated, the correlation degree is determined according to the calculation result, and a corresponding explosion-proof disposal plan is output according to the correlation degree.

[0009] As an optional embodiment of the present application, multi-modal detection is performed on a target petrochemical tank area to obtain multi-modal detection data, comprising:

[0010] The multi-modal detection data include visible light images, gas distribution data and preset frequency band sound wave signals, wherein the gas distribution data include spectral absorption data and plume distribution data;

[0011] The visible light images of the target petrochemical tank area are collected by using double visible light cameras;

[0012] Collecting spectral absorption data and plume distribution data of the target petrochemical tank area by using an optical gas imager;

[0013] Collecting sound wave signals in a preset frequency band by using an ultrasonic array sensor.

[0014] As an optional embodiment of the present application, cross-collaborative verification of the re-examination analysis result is performed according to the position of the analysis result in the unified space-time coordinate system, and it is determined whether the cross-collaborative verification is established, including:

[0015] If there is an abnormal analysis result, the multi-modal detection data is re-examined and analyzed based on the position of the abnormal analysis result in the unified space-time coordinate system, and a re-examination analysis result is obtained;

[0016] The correlation between the analysis result and the re-examination analysis result is calculated, and it is determined whether the correlation is successful according to whether the correlation meets a preset correlation standard, and a hierarchical corresponding measure is set according to the multi-modal detection data that is successfully correlated;

[0017] If the analysis result and the re-examination analysis result are both successfully correlated, it is determined that the cross-collaborative verification is established.

[0018] As an optional embodiment of the present application, the multi-modal detection data is analyzed item by item, including:

[0019] It is analyzed whether the visible light image contains an abnormal area, and if the visible light image contains an abnormal area, the visible light image is abnormal, and the edge profile of the abnormal area is extracted;

[0020] The position of the edge profile in the unified space-time coordinate system is marked, and a detection point position within a preset range is locked based on a pre-constructed three-dimensional digital model of the petrochemical tank area according to the mark.

[0021] As an optional embodiment of the present application, the multi-modal detection data is analyzed item by item, including:

[0022] The gas distribution data is analyzed, and if the gas distribution data contains a gas concentration higher than a historical normal concentration range, the gas distribution data is abnormal, and the corresponding gas is taken as a target gas;

[0023] The spectral absorption data of the target gas is analyzed, and it is identified whether the target gas is a leak-proof gas;

[0024] The plume distribution data of the target gas is analyzed, and a diffusion parameter of the target gas is calculated.

[0025] As an optional embodiment of the present application, if there is an abnormal analysis result, the following steps are further included:

[0026] extracting target data from the multi-modal detection data to construct a multi-modal correlation matrix;

[0027] correlation analysis on the target data, updating a data consistency index of the multi-modal correlation matrix, wherein updating the data consistency index comprises updating time synchronization, spatial consistency and intensity correlation.

[0028] As an optional embodiment of the present application, correlation analysis on the target data, updating a data consistency index of the multi-modal correlation matrix, comprises:

[0029] calculating a time deviation between the target data, and the time synchronization is correspondingly valued according to the time deviation;

[0030] calculating the Euclidean distance between the center point of the gas diffusion area and the positioning point of the gas source according to the target data, and the spatial consistency is correspondingly valued according to the Euclidean distance;

[0031] determining positive correlation or negative correlation by dynamically comparing the change trend of the frequency spectrum of the contrast gas distribution data and the sound wave signal, and the intensity correlation is correspondingly valued according to the determination result;

[0032] updating the data consistency index according to the time synchronization, the spatial consistency and the intensity correlation after corresponding valuation.

[0033] The present application also provides a petrochemical tank area inspection robot, which is applied to the petrochemical tank area multi-modal detection explosion-proof method as described above, comprising:

[0034] A vehicle body, two symmetrical rollers are arranged on the two sides of the vehicle body respectively, and the rollers are used to be connected with the bearing cable;

[0035] A detection unit, which comprises two visible light cameras and an ultrasonic array sensor fixedly arranged at the bottom of the vehicle body, and an optical gas imager fixedly arranged at the front end of the vehicle body along the advancing direction of the vehicle body;

[0036] A driving connection part, which is fixedly arranged at the top of the vehicle body and used to be connected with the traction cable.

[0037] As an optional embodiment of the present application, a groove is arranged on the outer circumferential surface of the roller, the groove comprises two symmetrical abutting surfaces, the abutting surfaces are used to abut against the bearing cable, and a plurality of magnets are respectively and intervally distributed on the abutting surfaces along the circumferential direction of the roller.

[0038] The present application also provides a petrochemical tank area multi-modal detection explosion-proof system, comprising:

[0039] An acquisition data module, which is used to perform multi-modal detection on a target petrochemical tank area according to a preset inspection path, and obtain multi-modal detection data.

[0040] a mapping module, configured to analyze the multi-modal detection data item by item, and map corresponding analysis results to a time-space unified coordinate system constructed according to time stamps calibrated in advance based on the multi-modal detection data;

[0041] a cross-collaborative verification module, configured to perform cross-collaborative verification of the analysis results according to positions of the analysis results in the time-space unified coordinate system, and determine whether the cross-collaborative verification is valid;

[0042] an output data module, configured to calculate an abnormal index correlation weight coefficient if the cross-collaborative verification is valid, determine a correlation degree according to a calculation result, and output a corresponding explosion-prevention disposal plan according to the correlation degree.

[0043] The present application also provides an electronic device, comprising:

[0044] a processor;

[0045] a memory for storing processor-executable instructions;

[0046] wherein the processor is configured to implement the petrochemical tank area multi-modal detection explosion-prevention method described above when executing the executable instructions.

[0047] Compared with the prior art, the present application realizes early identification of small leakage of hazardous gas by synchronously acquiring multi-modal detection data and dynamically correlating and analyzing, improves the discrimination accuracy of gas-liquid two-phase leakage characteristics, greatly reduces the frequency of manual re-inspection in high-risk areas, reduces the false alarm rate, saves maintenance resources, and promotes unmanned inspection of petrochemical tank areas. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 is a petrochemical tank area multi-modal detection explosion-prevention method flowchart of an embodiment of the present application;

[0049] Figure 2 is a three-dimensional structural schematic diagram of a petrochemical tank area inspection robot;

[0050] Figure 3 is a three-dimensional structural schematic diagram of a petrochemical tank area inspection robot; Figure 2

[0051] Figure 4 is a three-dimensional structural schematic diagram of a petrochemical tank area inspection robot;

[0052] Figure 5 is a petrochemical tank area multi-modal detection explosion-prevention system block diagram of an embodiment of the present application;

[0053] ​In the figure, the vehicle body 100; the roller 101; the anti-off plate 102; the magnet 103; the driving connection part 104; the traction cable 200; the bearing cable 201; the dual visible light camera 300; the ultrasonic array sensor 301; the optical gas imager 302; the radio frequency read-write sensor 303; the petrochemical tank area multi-modal detection explosion-proof system 10; the data acquisition module 11; the mapping module 12; the cross-collaborative verification module 13; the output data module 14. DETAILED DESCRIPTION

[0054] The following is a specific embodiment of the present application and further describes the technical solutions of the present application in conjunction with the drawings, but the present application is not limited to these embodiments.

[0055] Embodiment 1

[0056] The crude oil, finished oil, chemicals and the like stored in the petrochemical tank area are mostly flammable and explosive substances, and their vapors mixed with air may form an explosive environment, directly threatening personal safety, so the petrochemical tank area explosion-proof is the core link of the multiple protection system. The petrochemical tank area multi-modal detection explosion-proof method as shown in the embodiment includes: Figure 1

[0057] S1, according to a preset inspection path, performing multi-modal detection on a target petrochemical tank area to obtain multi-modal detection data;

[0058] S2, analyzing the multi-modal detection data item by item, and mapping the corresponding analysis results to a time-space unified coordinate system constructed according to a time stamp calibrated in advance according to the multi-modal detection data;

[0059] S3, performing cross-collaborative verification of the rechecked analysis results according to the positions of the analysis results in the time-space unified coordinate system, and determining whether the cross-collaborative verification is valid;

[0060] S4, if the cross-collaborative verification is valid, calculating an abnormal index correlation weight coefficient, determining a correlation degree according to the calculation result, and outputting a corresponding explosion-proof disposal plan according to the correlation degree.

[0061] ​A patrol path is formulated in advance according to the distribution of various facilities in the petrochemical tank area, a patrol device is used and a control signal is issued to the patrol device, so that the petrochemical tank area is detected in multiple modes according to the patrol path. The patrol device can be an explosion-proof robot, which can smoothly complete the patrol according to the patrol path and transmit the detected multi-modal detection data for analysis. Multi-modal data is information composed of multiple different types of data modalities, which can be text, image, audio, video, sensor data, time series data, etc. The detection data of the petrochemical tank area obtained by the patrol device in this embodiment is multi-modal, and all the obtained detection data is taken as multi-modal detection data, which is then transmitted to the control center of the petrochemical tank area for subsequent analysis.

[0062] According to the time stamp of each modal detection data, the spatial dimension and the time dimension are integrated to construct a unified spatio-temporal coordinate system, and the time axis of the multi-modal data is aligned. Each modal detection data is analyzed separately to determine whether each modal detection data is normal or abnormal, and the analysis results are mapped to the unified spatio-temporal coordinate system according to the time stamp.

[0063] According to the analysis results of each modal detection data, cross-collaborative verification is performed to check whether each modal detection data of the same petrochemical tank area location is all normal, all abnormal or partially abnormal, on the one hand to ensure the real-time nature of the detection data, and on the other hand based on the all normal, all abnormal or partially abnormal condition, to calculate whether the correlation between each modal detection data exceeds a preset correlation standard, when the cross-collaborative verification is established, further calculate the abnormal index correlation weight coefficient, determine the correlation degree according to the specific value calculated, trigger remote alarm and push emergency disposal plan.

[0064] Preferably, the target petrochemical tank area is detected in multiple modes to obtain multi-modal detection data, including:

[0065] The multi-modal detection data includes visible light images, gas distribution data and preset frequency band acoustic wave signals, wherein the gas distribution data includes spectral absorption data and plume distribution data;

[0066] A dual visible light camera is used to collect visible light images of the target petrochemical tank area;

[0067] An optical gas imaging instrument is used to collect spectral absorption data and plume distribution data of the target petrochemical tank area;

[0068] An ultrasonic array sensor is used to collect preset frequency band acoustic wave signals.

[0069] The multi-modal detection data collected in this embodiment includes visible light images, spectral absorption data about gas, plume distribution data and acoustic wave signal data.

[0070] By setting double visible cameras on the inspection device, the double visible cameras scan the pipelines, flanges and other facilities of the target petrochemical tank area and the ground of the target petrochemical tank area in real time according to the inspection path. The double visible cameras are equipped with adaptive lighting lamps, the brightness of which is dynamically adjusted according to the environmental illumination, so that clear images with a resolution of 1920x1080 can still be obtained in the night or low-illumination petrochemical tank area environment. By setting an optical gas imager on the inspection device to collect the spectral absorption data of the target petrochemical tank area, the optical gas imager is provided with a plurality of spectral filter adjustment grooves, and the spectral filter corresponding to different wavelengths is arranged in each groove. The spectral filter corresponding to the wavelength can focus on the characteristic absorption spectrum of the target gas according to the control signal. The non-cooled infrared detector in the optical gas imager is also used to scan the gas plume distribution, analyze the gas plume distribution data to obtain the gas concentration data and diffusion velocity parameters, etc. The ultrasonic array sensor composed of a rectangular array structure of multiple sensing units is arranged on the inspection device to capture the sound signals in the frequency band of 20-40 kHz.

[0071] Preferably, cross-collaborative verification of the analysis results is performed according to the positions of the analysis results in the spatiotemporal unified coordinate system, and it is determined whether the cross-collaborative verification is valid, including:

[0072] If there is an abnormal analysis result, the multi-modal detection data is rechecked based on the position of the abnormal analysis result in the spatiotemporal unified coordinate system, and rechecked analysis results are obtained by analysis;

[0073] The correlation between the analysis results and the rechecked analysis results is calculated, and it is determined whether the correlation is successful according to whether the correlation meets a preset correlation standard;

[0074] If the analysis results and the rechecked analysis results are both successfully correlated, it is determined that the cross-collaborative verification is valid.

[0075] After analyzing each modality detection data separately, if the analysis result of one of the modalities is abnormal, the remaining modalities detection data at the coordinate position of the abnormal modality in the spatiotemporal unified coordinate system is rechecked. For example, when the analysis result of the visible light image is that it contains an abnormal area, a hierarchical response mechanism is triggered, the real position of the abnormality in the petrochemical tank area is determined according to the position of the abnormality in the spatiotemporal unified coordinate system, and the gas distribution data and the preset frequency band acoustic signal at the real position are detected again in real time to analyze whether they are all normal, all abnormal or partially abnormal. Then, the correlation between the visible light image and the gas distribution data is calculated according to the all normal, all abnormal or partially abnormal gas distribution data, and if the correlation meets the standard, it indicates that the suspected gas leakage is highly likely, and at this time, the measure of marking the leakage point coordinate is taken. Based on the correlation between the visible light image and the rechecked gas distribution data, the correlation with the rechecked preset frequency band acoustic signal is calculated, and if the correlation meets the standard, it indicates that the suspected gas leakage is more likely. When the correlation between the visible light image, the rechecked gas distribution data and the preset frequency band acoustic signal meets the standard, the cross-verification is established, and the correlation weight coefficient W of the three abnormal indicators is calculated. The correlation weight coefficient W of the abnormal indicators calculated is compared with the preset threshold. In this embodiment, the threshold T is set to 0.85, and if W ≥ T, it is determined as a strong correlation leakage event, a remote alarm is triggered, and an emergency disposal plan is pushed. The hierarchical response mechanism is triggered according to the analysis result containing an abnormality, the rechecking of the remaining modalities detection data is started, and related measures are taken according to whether the correlation meets the standard. In this embodiment, when the analysis result of one of the modalities contains an abnormality, the related measure taken at this time is to start the rechecking of the remaining modalities detection data. According to the analysis result of one of the rechecking results, i.e., the analysis result of the gas distribution data or the preset frequency band acoustic signal, the correlation with the analysis result containing an abnormality is calculated, and if the correlation meets the standard, the related measure taken at this time is to automatically mark the leakage point coordinate. Then, based on the modalities detection data whose correlation meets the standard, the correlation with the analysis result of the preset frequency band acoustic signal or the gas distribution data is calculated, and if all the correlations meet the standard, the related measure taken at this time is to update the abnormal indicator correlation weight coefficient according to the updated data consistency indicators, i.e., the time synchronization, the spatial consistency and the intensity correlation, to determine whether it is a strong correlation leakage event, and if it is, a remote alarm is triggered and an emergency disposal plan is pushed. It should be noted that the measures taken in the hierarchical response mechanism can be set according to the types of the actual detection data and whether the correlation meets the standard.

[0076] Preferably, the multi-modality detection data is analyzed item by item, including:

[0077] analyzing whether the visible light image contains an abnormal area, if the visible light image contains an abnormal area, the visible light image is abnormal, and an edge profile of the abnormal area is extracted;

[0078] The position of the edge profile in the space-time unified coordinate system is marked, and based on the pre-constructed petrochemical tank area three-dimensional digital model, the detection point position within the preset range is locked according to the mark.

[0079] The double visible light cameras continuously capture the image of the tank body and the pipeline surface of the target petrochemical tank area. When an abnormal area appears in the visible light image, such as abnormal reflection spots or wet marks in dry areas, it is determined as a suspected leakage area. An edge enhancement process is automatically triggered to extract the edge profile features of the abnormal area, mark the real position coordinates of the suspected leakage area, and lock the detection point position within the preset range based on the pre-constructed petrochemical tank area three-dimensional digital model, such as locking the associated storage tank flange, pipeline weld or valve connection within the range directly above the real position coordinates.

[0080] Preferably, the multi-modal detection data is analyzed item by item, and further includes:

[0081] Analyzing the gas distribution data, if the gas concentration is higher than the historical normal concentration range, the gas distribution data is abnormal, and the corresponding gas is taken as the target gas;

[0082] Analyzing the spectral absorption data of the target gas, and identifying whether the target gas is a leak-proof gas;

[0083] Analyzing the plume distribution data of the target gas, and calculating the diffusion parameters of the target gas.

[0084] The optical gas imager continuously captures infrared images of the surface of the tank and pipeline. If the gas concentration captured exceeds the historical normal concentration range, it indicates that the gas distribution data is abnormal. The gas exceeding the historical normal concentration range is regarded as the target gas for further identification. The active spectral enhancement technology is used to improve the identification ability of the molecular absorption spectrum characteristics of the target gas, effectively distinguish the environmental interference signal from the real gas leakage characteristics, and confirm whether the target gas is a leak-proof gas, including flammable gas, such as methane. If it is a leak-proof gas, there is a leak-proof gas leakage in the target petrochemical tank area. The detected leakage area is marked with pseudo-color, and timely intervention is required. After confirming that it is a leak-proof gas, the plume distribution data scanned by the non-cooled infrared detector is used to extract the plume shape of the leak-proof gas and track the leak source of the leak-proof gas. The concentration data and diffusion speed parameters of the leak-proof gas are analyzed and output. The concentration data of the leak-proof gas output by the optical gas imager is subjected to Kalman filter denoising, and the filter window is set to 5 seconds sliding average. The concentration trend curve is generated and the abnormal period when the fluctuation rate exceeds 15% is marked.

[0085] In addition, when analyzing the abnormal area contained in the visible light image, the real position of the abnormal area is focused and scanned to obtain rechecked gas distribution data for subsequent cross-verification.

[0086] Preferably, the multi-modal detection data is analyzed item by item, and further comprises:

[0087] The preset frequency band sound wave signal is subjected to fast Fourier transform to extract a frequency spectrum graph of the target frequency band.

[0088] The frequency spectrum graph is identified to determine whether it contains a peak characteristic of the leak-proof gas. If it contains, the preset frequency band sound wave signal is abnormal. The phase difference of the sound wave signal corresponding to the peak characteristic received by the sensor is calculated by a beam forming algorithm to locate the sound source direction of the sound wave signal.

[0089] In this embodiment, the preset frequency band sound wave signal is a sound wave signal in the frequency band of 20-40 kHz, and the sampling frequency of the ultrasonic array sensor is 1 MHz. The fast Fourier transform is performed on the sound wave signal in the frequency band of 20-40 kHz to extract a frequency spectrum graph in the frequency band of 25 kHz. Whether the frequency spectrum graph contains a peak characteristic of the leak-proof gas, such as a peak characteristic of methane gas, is identified. The phase difference of the signal received by each sensing unit is calculated by a beam forming algorithm to preliminarily locate the sound source direction of the sound wave signal corresponding to the leak-proof gas. It should be noted that the mentioned original state of the leak-proof gas can be gas or liquid evaporation.

[0090] It also needs to be explained that the item-by-item analysis of the multi-modal detection data means that the detection data of each modality is analyzed separately, the analysis process is parallel, and the analysis results can be transmitted to each other, so that the analysis results of each modality detection data can be fused and analyzed, and cross-verification is performed. Specifically, when the abnormal area is contained in the visible light image, the coordinates of the true position of the abnormal area are automatically synchronized to the optical gas imager, guiding it to focus scanning, and whether the abnormal area has a leak-proof gas leakage is verified again through the secondary verification of the molecular absorption spectrum characteristics. The ultrasonic array sensor collects the sound wave signal of a specific frequency band of the abnormal area in real time, and further verifies whether there is a leak-proof gas leakage through peak feature recognition. Or when the optical gas imager captures a gas concentration that exceeds the historical normal concentration range, the corresponding visible light image, sound wave signal and molecular absorption spectrum characteristics at the position of the unified coordinate system in space-time are linked to retrieve and cross-verify, and a three-dimensional heat map containing the leakage position, leakage material type and diffusion trend is generated.

[0091] Based on the unified coordinate system in space-time, the target data corresponds to the time point of the sudden increase of the gas concentration in the multi-modal detection data, the outline position of the target gas diffusion area in the visible light image at the same time, and the intensity change value of the ultrasonic signal in the corresponding period. Align the time axis of these multi-modal detection data to construct a multi-modal correlation matrix. It should be noted that the relationship between the extracted target data and the analysis results of the anomaly can be included or not included, for example, the data of the sudden increase of the gas concentration may or may not belong to the gas concentration data higher than the historical normal concentration range. The matrix adopts a sliding time window mechanism, the window width is 30 seconds, and the data consistency index: time synchronization α, spatial consistency β and intensity correlation γ are updated to calculate the abnormal index correlation weight coefficient W, W= (0.4α+0.3β+0.3γ). The value range of the abnormal index correlation weight coefficient W is [0, 1], and the closer the value is to 1, the stronger the collaborative correlation between the multi-modal data.

[0092] Preferably, for the correlation analysis of the target data, the data consistency index of the multi-modal correlation matrix is updated, comprising:

[0093] Calculate the time deviation between the target data, and the time synchronization is correspondingly valued according to the time deviation;

[0094] Calculate the Euclidean distance between the center point of the gas diffusion area and the positioning point of the gas source according to the target data, and the spatial consistency is correspondingly valued according to the Euclidean distance;

[0095] Dynamically compare the change trend of the contrast gas distribution data and the frequency spectrum of the sound wave signal, determine the positive correlation or negative correlation, and the intensity correlation is correspondingly valued according to the determination result;

[0096] The data consistency index is updated according to the time synchronization, spatial consistency and intensity correlation after corresponding value.

[0097] For the value of time synchronization a, compared with the moment of sudden increase of contrast gas concentration, plume generation time and ultrasonic amplitude mutation time, if the time deviation of the three is within 2 seconds, it is determined that the time is highly synchronized, a=1; if the time deviation is 2 seconds<|D|≤6 seconds, linear deduction is carried out according to the amount of time deviation, and a=1-(|D|-2) / 4.

[0098] For the value of spatial consistency b, the Euclidean distance between the center point of the gas diffusion area in the visible light image obtained at the same time and the sound source positioning point of the ultrasonic wave to the gas is calculated to determine the spatial consistency; when the distance between the two is less than 0.5 meters, it is considered that the spatial position is completely matched, b=1; when the distance is 0.5 meters

[0099] For the value of intensity correlation g, the trend comparison method is used to dynamically compare the gas concentration change rate and the ultrasonic signal intensity increase, and a positive correlation score is given when the two show the same trend, and if the concentration increases but the signal intensity decreases, it is determined as negative correlation, and g=ACxAS, wherein AC is the gas concentration change rate, and AS is the ultrasonic signal intensity increase percentage, and when the calculation result is greater than 1.0, g=1, and when AC and AS change in opposite directions, g=0.

[0100] By applying the above method to the explosion-proof detection of petrochemical tank area, multi-modal detection data is synchronously obtained, and dynamic correlation analysis is carried out, early identification of small leakage of hazardous gas is realized, discrimination accuracy of gas-liquid two-phase leakage characteristics is improved, frequency of manual re-inspection in high-risk area is greatly reduced, false positive rate is reduced, maintenance resources are saved, and unmanned inspection of petrochemical tank area is promoted.

[0101] Embodiment 2

[0102] The embodiment also provides a petrochemical tank area inspection robot, which is applied to the explosion-proof method of the petrochemical tank area multi-modal detection of embodiment 1, as shown in Figures 2-4 The petrochemical tank area inspection robot comprises:

[0103] The vehicle body 100 is provided with two rollers 101 symmetrically arranged on both sides of the vehicle body 100, the rollers 101 are used to be connected with the bearing cable 201, and the vehicle body 100 is provided with support and rolls along the extension direction of the bearing cable 201.

[0104] The detection unit comprises two visible light cameras 300 and an ultrasonic array sensor 301 fixedly arranged at the bottom of the vehicle body 100, and an optical gas imager 302 is fixedly arranged at the front end of the vehicle body 100 in the advancing direction of the vehicle body 100.

[0105] A driving connection part 104 is fixedly arranged on the top of the vehicle body 100, and the driving connection part 104 is fixed with the traction cable 200.

[0106] It should be noted that the dual visible light camera 300 refers to two visible light imaging units integrated on the explosion-proof inspection robot, and the optical gas imager 302 includes a non-cooled infrared detector.

[0107] Further limited, the outer circumferential surface of the roller 101 is provided with a groove, the groove includes two abutting surfaces arranged symmetrically, the abutting surfaces are used for abutting with the bearing cable 201, along the circumferential direction of the roller 101, a plurality of mounting grooves are respectively and intermittently distributed on the abutting surfaces, and a magnet 103 is arranged in each mounting groove. The magnet 103 does not protrude from the abutting surface, and the magnet 103 can generate a magnetic attraction force with the bearing cable 201. Along the circumferential direction of the roller 101, the gap distance between the adjacent two magnets 103 is less than the length of the magnet 103, so as to realize continuous coverage of the magnetic field and improve the adsorption stability.

[0108] A plurality of U-shaped anti-falling plates 102 are also fixedly arranged on the vehicle body 100, and a cable space is formed between the anti-falling plate 102 and the corresponding roller 101. The bearing cable 201 extends out after passing through the cable space, and the bearing cable 201 does not contact the anti-falling plate 102.

[0109] Further preferably, a radio frequency position tag is arranged on the bearing cable 201, and a radio frequency read-write sensor 303 is arranged on the vehicle body 100, and the radio frequency read-write sensor 303 can read the position information on the radio frequency position tag.

[0110] A wireless charging module, an industrial switch and a wireless communication module are also arranged in the vehicle body, and an explosion-proof antenna is arranged on the vehicle body. The wireless charging module is axially aligned and electromagnetically coupled with an external charging base station.

[0111] Using the robot, the optical imaging of the gas leakage form, the infrared temperature field distribution and the ultrasonic spectrum characteristics can be synchronously obtained, a dynamic correlation model of multi-physical field data can be established, and the discrimination accuracy of micro leakage and gas-liquid two-phase leakage characteristics can be improved.

[0112] Embodiment 3

[0113] Based on the same principle as the foregoing method, a petrochemical tank area multi-modal detection explosion-proof system 10 is also proposed, as shown in Figure 5 The explosion-proof system 10 comprises:

[0114] The data acquisition module 11 is used for performing multi-modal detection on the target petrochemical tank area according to a preset inspection path, and obtaining multi-modal detection data.

[0115] A mapping module 12 is configured to analyze the multi-modal detection data item by item and map corresponding analysis results to a time-space unified coordinate system constructed according to time stamps calibrated in advance based on the multi-modal detection data;

[0116] A cross-collaborative verification module 13 is configured to perform cross-collaborative verification of the analysis results according to their positions in the time-space unified coordinate system, and determine whether the cross-collaborative verification is valid.

[0117] An output data module 14 is configured to calculate an abnormal index correlation weight coefficient if the cross-collaborative verification is valid, determine a correlation degree according to the calculation result, and output a corresponding explosion-prevention disposal plan according to the correlation degree.

[0118] Embodiment 4

[0119] Further, an electronic device is provided, comprising:

[0120] a processor;

[0121] a memory for storing processor-executable instructions;

[0122] The processor is configured to implement the petrochemical tank farm multi-modal detection explosion-prevention method of embodiment 1 when executing the executable instructions.

[0123] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present application are only used to explain the relative positional relationship, movement condition, etc. between components in a certain posture (as shown in the drawings), and if the certain posture changes, the directional indications will also change accordingly.

[0124] In addition, it should be noted that the descriptions such as “first”, “second”, “one” and the like in the present application are only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features limited by “first” and “second” can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of “a plurality of” is at least two, such as two, three, etc., unless otherwise specifically limited. The terms “connection”, “fixation” and the like should be understood broadly, for example, “fixation” can be fixed connection, or detachable connection, or integral; can be mechanical connection, or electrical connection; can be direct connection, or indirect connection through an intermediate medium; can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise specifically limited. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0125] In addition, the technical solutions among various embodiments of the present application can be combined with each other, but it must be based on that a person skilled in the art can realize, when the combination of the technical solutions appears contradictory or cannot be realized, it should be considered that the combination of the technical solutions does not exist, and is not within the protection scope required by the present application.

[0126] The specific embodiments described herein are merely illustrative of the spirit of the present application. Those skilled in the art can make various modifications or supplements to the described specific embodiments or replace them with similar ways, but will not deviate from the spirit of the present application or exceed the scope defined by the appended claims.

Claims

1. A petrochemical tank farm multi-modal detection explosion proof method, characterized in that, The method comprises the following steps: According to the preset inspection path, the target petrochemical tank area is detected by multi-modal detection to obtain multi-modal detection data; The multi-modal detection data is analyzed item by item, and the corresponding analysis result is mapped to the time-space unified coordinate system constructed by the time stamp calibrated in advance according to the multi-modal detection data; According to the position of the analysis result in the time-space unified coordinate system, the cross-verification of the re-inspection analysis result is carried out, and whether the cross-verification is established is judged; If there is an abnormal analysis result, the multi-modal detection data is re-inspected and analyzed based on the position of the abnormality in the time-space unified coordinate system, and the re-inspection analysis result is obtained, wherein the re-inspection analysis refers to the analysis of each modal detection data separately, and if the analysis result of one of the modal detection data is abnormal, the remaining modal detection data of the coordinate position is re-inspected according to the coordinate position of the modal detection data in the time-space unified coordinate system; If there is an abnormal analysis result, it further comprises: Extracting target data from the multi-modal detection data to construct a multi-modal correlation matrix; Correlation analysis of the target data, updating the data consistency index of the multi-modal correlation matrix for calculating the abnormal index correlation weight coefficient, wherein the data consistency index includes time synchronization, spatial consistency and intensity correlation; If the cross-verification is established, the abnormal index correlation weight coefficient is calculated according to the data consistency index, the correlation degree is determined according to the calculation result, and the corresponding explosion-proof disposal plan is output according to the correlation degree.

2. The petrochemical tank farm multi-modal detection blast prevention method of claim 1, wherein, Multi-modal detection is carried out on the target petrochemical tank area to obtain multi-modal detection data, which comprises: The multi-modal detection data includes visible light image, gas distribution data and preset frequency band sound signal, wherein the gas distribution data includes spectral absorption data and plume distribution data; A dual visible light camera is used to collect the visible light image of the target petrochemical tank area; An optical gas imager is used to collect the spectral absorption data and plume distribution data of the target petrochemical tank area; An ultrasonic array sensor is used to collect the preset frequency band sound signal.

3. The petrochemical tank farm multi-modal detection blast prevention method of claim 2, wherein, According to the position of the analysis result in the time-space unified coordinate system, the cross-verification of the re-inspection analysis result is carried out, and whether the cross-verification is established is judged, which comprises: Calculate the correlation between the analysis result and the re-inspection analysis result, determine whether the correlation is successful according to whether the correlation meets the preset correlation standard, and set hierarchical corresponding measures according to the multi-modal detection data with successful correlation; If the analysis result and the re-inspection analysis result are both successfully correlated, it is determined that the cross-verification is established.

4. The petrochemical tank farm multi-modal detection blast prevention method of claim 3, wherein, The multi-modal detection data is analyzed item by item, which comprises: Analyze whether the visible light image contains an abnormal area, if it contains an abnormal area, the visible light image is abnormal, and the edge profile of the abnormal area is extracted; The position of the edge profile in the time-space unified coordinate system is marked, and based on the pre-constructed three-dimensional digital model of the petrochemical tank area, the detection point position within the preset range is locked according to the mark.

5. The petrochemical tank farm multi-modal detection blast prevention method of claim 3, wherein, The multi-modal detection data is analyzed item by item, which comprises: The gas distribution data is analyzed, if the gas concentration is higher than the historical normal concentration range, the gas distribution data is abnormal, and the corresponding gas is taken as a target gas; The spectral absorption data of the target gas is analyzed to identify whether the target gas is a leak-proof gas; The plume distribution data of the target gas is analyzed to calculate the diffusion parameters of the target gas.

6. The petrochemical tank farm multi-modal detection blast prevention method of claim 3, wherein, The multi-modal detection data is analyzed item by item, and further comprising: The preset frequency band acoustic wave signal is subjected to fast Fourier transform to extract a frequency spectrum of a target frequency band; It is identified whether the frequency spectrum contains a peak characteristic of a leak-proof gas, if it contains, the preset frequency band acoustic wave signal is abnormal, and a phase difference of the acoustic wave signal corresponding to the peak characteristic received by the sensor is calculated by a beam forming algorithm to locate a sound source direction of the acoustic wave signal.

7. The petrochemical tank farm multi-modal detection blast prevention method of claim 3, wherein, The target data is correlated and analyzed to update a data consistency index of the multi-modal correlation matrix, comprising: A time deviation between the target data is calculated, and the time synchronization is correspondingly valued according to the time deviation; A Euclidean distance between a center point of a gas diffusion area and a gas source positioning point is calculated according to the target data, and the spatial consistency is correspondingly valued according to the Euclidean distance; The change trend of the contrast gas distribution data and the frequency spectrum of the acoustic wave signal is dynamically determined to be positively correlated or negatively correlated, and the intensity correlation is correspondingly valued according to the determination result; The data consistency index is updated according to the time synchronization, the spatial consistency and the intensity correlation after corresponding valuation.

8. A petrochemical tank farm inspection robot applied to the petrochemical tank farm multi-modal detection explosion-proof method according to any one of claims 1-7, characterized in that, Comprising: A vehicle body, two symmetrical rollers are arranged on both sides of the vehicle body, and the rollers are used to be connected with a bearing cable; A detection unit comprising two visible light cameras and an ultrasonic array sensor fixedly arranged at the bottom of the vehicle body, and an optical gas imager fixedly arranged at the front end of the vehicle body along the advancing direction of the vehicle body; A driving connection part fixedly arranged at the top of the vehicle body and used to be connected with a traction cable.

9. The petrochemical tank farm inspection robot of claim 8, wherein, The outer periphery of the roller is provided with a groove, the groove comprises two symmetrical abutting surfaces, the abutting surfaces are used to abut against the bearing cable, and a plurality of magnets are respectively and intermittently distributed on the abutting surfaces along the circumferential direction of the roller.

Images