Rapid identification and alarm method for gas-liquid two-phase leakage of alkylation device

Abstract

The invention discloses a method for quickly identifying and alarming gas-liquid two-phase leakage of an alkylation device. Gas-liquid two-phase leakage alarm is carried out based on an alkylation device gas-liquid two-phase leakage linkage alarm system installed on an alkylation device site, and the alkylation device gas-liquid two-phase leakage linkage alarm system comprises a combustible gas detector, a video sensing camera system and a leakage accident early warning processing center. The leakage accident early warning processing center is internally provided with a geographic information system module, an alkylation device field video identification module, an abnormal alarm processing module and a human-assisted research and judgment module, and an alkylation device field video identification module constructed based on a deep learning network is packaged in the alkylation device field video identification module. According to the invention, rapid identification and grading alarm of gas leakage and liquid leakage on site of the alkylation device are realized, the accuracy and efficiency of identification of leakage accidents on site of the alkylation device are effectively improved, and decision support is provided for rapid early warning of leakage risks.

Description

Technical Field

[0001] The invention relates to the technical field of safety engineering, and in particular to a method for quickly identifying and alarming gas-liquid two-phase leakage in an alkylation unit. Background Art

[0002] As an important synthetic method, alkylation reaction introduces alkyl groups into organic molecules through addition or displacement reactions and has been widely used in petrochemical production processes. With the construction and accelerated expansion of alkylation units in petrochemical enterprises and the increasing complexity of related processes, the corresponding safety issues have become increasingly prominent, especially the identification and monitoring of leaks in key dangerous areas has become a difficult problem that needs to be solved urgently. In order to quickly detect gas-liquid two-phase mixed leakage and issue timely warnings, and reduce the losses that may be caused by leakage accidents, it is particularly important to effectively carry out research on the rapid and accurate identification and monitoring and early warning of gas-liquid two-phase mixed leakage.

[0003] At present, the means and methods for monitoring the leakage of flammable or toxic gases in the petrochemical production process are relatively mature, while the identification and monitoring of liquid leaks are still under development and improvement. At this stage, there are various means for identifying and monitoring liquid leaks. Simply relying on manual inspections to monitor liquid leaks has high labor costs on the one hand, and is highly subjective on the other hand. It is easy to misjudge leakage accidents at different angles and has low efficiency; automatic monitoring and identification relying on sensors and other alarm equipment is more common in gas and smoke detection, but it is difficult to identify and monitor liquid leaks, and the accuracy is also low.

[0004] With the rapid development of big data and artificial intelligence technology, especially the widespread application of deep learning algorithms, it has become possible to monitor liquid leakage based on deep learning video recognition of liquid leakage targets. Therefore, it is urgent to propose a method for rapid identification and alarm of gas-liquid two-phase leakage in alkylation units, so as to realize rapid identification and alarm of gas-liquid two-phase mixed leakage in alkylation units, effectively improve the early warning and emergency response efficiency of petrochemical enterprises, and contribute to safe production. Summary of the invention

[0005] In view of the problem that liquid leakage identification and monitoring are difficult and of low precision at the current stage, the present invention proposes a method for rapid identification and alarm of gas-liquid two-phase leakage of an alkylation unit. Based on the construction method of an on-site video recognition module of an alkylation unit, a gas-liquid two-phase leakage linkage alarm system for an alkylation unit encapsulated with a rapid identification module for gas-liquid two-phase leakage of an alkylation unit is constructed. The gas-liquid two-phase leakage linkage alarm system for an alkylation unit is used to monitor the alkylation unit site in real time, thereby realizing rapid identification and alarm of gas and liquid leakages at the alkylation unit site, effectively improving the identification accuracy and efficiency of leakage accidents, and providing decision support for rapid early warning of leakage risks.

[0006] The present invention specifically adopts the following technical solutions:

[0007] A method for constructing an on-site video recognition module for an alkylation unit, which is applicable to an alkylation unit, specifically comprises the following steps:

[0008] Step 1, obtaining a gas-liquid two-phase leakage image, preprocessing each gas-liquid two-phase leakage image, obtaining multiple gas-liquid two-phase leakage image samples, and forming a gas-liquid two-phase leakage image sample data set, wherein the gas-liquid two-phase leakage image sample data set includes a training set and a test set;

[0009] Step 2: Building a rapid identification model for gas-liquid two-phase leakage based on a deep learning network;

[0010] Step 3, using the training set to train the gas-liquid two-phase leakage rapid identification model to obtain a trained gas-liquid two-phase leakage rapid identification model;

[0011] Step 4: Use the test set to verify the recognition effect of the gas-liquid two-phase leakage rapid identification model after training. If the recognition effect of the gas-liquid two-phase leakage rapid identification model after training has reached the best, proceed to step 5; otherwise, return to step 3 and continue to train the gas-liquid two-phase leakage rapid identification model using the training set;

[0012] Step 5: encapsulate the verified gas-liquid two-phase leakage rapid identification model and set it in the alkylation unit on-site video identification module.

[0013] Preferably, the gas-liquid two-phase leakage image is collected through two ways:

[0014] The first approach is to obtain gas-liquid two-phase leakage images from historical accident video files of gas-liquid two-phase leakage in alkylation units;

[0015] The second approach is to use a gas-liquid two-phase leakage simulation experimental device to simulate a gas-liquid two-phase leakage accident in the laboratory and obtain a gas-liquid two-phase leakage image.

[0016] Preferably, the preprocessing process of the gas-liquid two-phase leakage image is:

[0017] According to a preset pixel resolution, after filtering out gas-liquid two-phase leakage images with pixel resolutions lower than the preset pixel resolution, according to a preset noise range, filtering out gas-liquid two-phase leakage images with noise values ​​not meeting the preset noise range, and filtering out gas-liquid two-phase leakage images without gas-liquid two-phase leakage objects, to obtain filtered gas-liquid two-phase leakage images;

[0018] The screened gas-liquid two-phase leakage images are subjected to frame extraction processing to obtain multiple gas-liquid two-phase leakage images. On each gas-liquid two-phase leakage image, the position coordinates of the leakage liquid column, leakage droplets and leakage gas, as well as the type of leakage are marked on the gas-liquid two-phase leakage image using a labeling box, and a label document corresponding to the gas-liquid two-phase leakage image sample is generated. The gas-liquid two-phase leakage image with the label document is used as a gas-liquid two-phase leakage image sample.

[0019] Preferably, the preset pixel resolution is 240P.

[0020] Preferably, the number of gas-liquid two-phase leakage image samples is not less than 50,000;

[0021] The gas-liquid two-phase leakage image samples in the gas-liquid two-phase leakage image sample data set are randomly allocated to a training set and a test set, and the ratio of the number of gas-liquid two-phase leakage image samples in the training set to the number of gas-liquid two-phase leakage image samples in the test set is not less than 4:1.

[0022] Preferably, the deep learning network selects the target detection network YOLOv7, and uses the target detection network YOLOv7 to build a rapid identification model for gas-liquid two-phase leakage.

[0023] Preferably, the target detection network YOLOv7 takes into account both recognition speed and accuracy, and includes an input end, a backbone feature extraction network and a detection head;

[0024] The input end is used to perform adaptive image scaling processing, Mosaic data enhancement processing and adaptive anchor frame calculation on the input gas-liquid two-phase leakage image, wherein the adaptive image scaling processing is used to unify the input format of the gas-liquid two-phase leakage image, so that multiple input gas-liquid two-phase leakage images have the same size through scaling processing; the Mosaic data enhancement processing obtains the spliced ​​gas-liquid two-phase leakage image by splicing multiple gas-liquid two-phase leakage images in a random scaling, random cropping and random arrangement manner; the adaptive anchor frame calculation is used to automatically obtain the optimal anchor frame value during each training;

[0025] The backbone feature extraction network includes a plurality of CBS modules and a plurality of ELAN modules connected in sequence, and two adjacent ELAN modules are connected via an MP module;

[0026] The detection head includes a connected SPPCSPC structure, an ELAN module, an MP structure and a re-parameterized Rep structure, which are used as a classifier and a regressor of the network;

[0027] The SPPCSPC structure introduces the CSP structure into the SPP structure, and uses the SPP with the CSP structure to expand the receptive field for residual edge-assisted optimization and feature extraction;

[0028] The re-parameterized Rep structure introduces the RepConv structure and the residual structure to assist training, so as to reduce the complexity of the target detection network YOLOv7.

[0029] Preferably, the CBS module includes a convolution layer, a BN layer and a Silu layer; an activation function is set in the Silu layer, and the activation function is set to a sigmoid function.

[0030] Preferably, the ELAN module comprises a shallow feature extraction layer, a deep feature extraction layer and a reconstruction module;

[0031] The shallow feature extraction layer is set to a 3×3 convolutional layer;

[0032] The deep feature extraction layer consists of stacked ELABs and a residual connection;

[0033] The reconstruction module consists of a 3×3 convolution and a PixelShuffle operation layer.

[0034] Preferably, two branches are provided in the MP module for realizing super downsampling of parallel extraction and compression of features, wherein the first branch is provided with a maximum pooling layer and a 1×1 convolution layer connected in sequence; the other branch is provided with a 1×1 convolution layer and a convolution block with a 3×3 convolution kernel and a step size of 2 connected in sequence; the 1×1 convolution layers provided in the first branch and the second branch are both used to change the number of channels; a merging layer is provided after the first branch and the second branch, and the merging layer is used to add the processing results of the first branch and the second branch and output the result of super downsampling.

[0035] Preferably, the step 3 specifically includes the following steps:

[0036] Step 3.1, set the total number of training times, loss value, accuracy, recall rate and mAP@0.5 value;

[0037] The mAP@0.5 is the average precision of the image samples when the intersection over union (IoU) is 0.5;

[0038] Step 3.2, inputting the gas-liquid two-phase leakage image samples in the training set into the gas-liquid two-phase leakage rapid identification model constructed in step 2, using the gas-liquid two-phase leakage rapid identification model to identify the leakage type in the gas-liquid two-phase leakage image, and obtaining the leakage type identification result of the gas-liquid two-phase leakage image;

[0039] Step 3.2, optimize the gas-liquid two-phase leakage rapid identification model based on the Adam algorithm, compare the leakage type identification result of the gas-liquid two-phase leakage rapid identification model with the label document corresponding to the gas-liquid two-phase leakage image sample, calculate the loss value, accuracy, recall rate and mAP@0.5 value of the gas-liquid two-phase leakage rapid identification model, and compare them with the preset loss value, accuracy, recall rate and mAP@0.5 value. If the loss value is less than the preset loss value and the accuracy, recall rate and mAP@0.5 value are all greater than the preset accuracy, recall rate and mAP@0.5 value, proceed to step 3.4; otherwise, proceed to step 3.3;

[0040] Step 3.3, update the confidence threshold and IOU threshold of the gas-liquid two-phase leakage rapid identification model to obtain an updated gas-liquid two-phase leakage rapid identification model, continue to randomly select gas-liquid two-phase leakage image samples in the training set and input them into the updated gas-liquid two-phase leakage rapid identification model, use the updated gas-liquid two-phase leakage rapid identification model to identify the gas-liquid two-phase leakage image samples, obtain the gas-liquid two-phase leakage image sample leakage type identification result, and then return to step 3.2;

[0041] Step 3.4, complete the training of the gas-liquid two-phase leakage rapid identification model to obtain the trained gas-liquid two-phase leakage rapid identification model.

[0042] Preferably, in the gas-liquid two-phase leakage rapid identification model, the initial value of the confidence threshold is set to 0.25, and the maximum value of the confidence threshold is not higher than 0.5; the initial value of the IOU threshold is set to 0.45.

[0043] Preferably, the step 4 specifically includes the following steps:

[0044] Step 4.1, input the gas-liquid two-phase leakage image samples in the test set into the trained gas-liquid two-phase leakage rapid identification model, use the gas-liquid two-phase leakage rapid identification model to identify the leakage type in the gas-liquid two-phase leakage image samples, and obtain the leakage type identification result of the gas-liquid two-phase leakage image samples;

[0045] Step 4.2, compare the recognition results of the gas-liquid two-phase leakage rapid identification model of the gas-liquid two-phase leakage image sample with the label document corresponding to the gas-liquid two-phase leakage image sample, calculate the loss value, accuracy, recall rate and mAP@0.5 value of the recognition results of the gas-liquid two-phase leakage rapid identification model, and compare them with the preset loss value, accuracy, recall rate and mAP@0.5 value. If the loss value is less than the preset loss value and the accuracy, recall rate and mAP@0.5 value are all greater than the preset accuracy, recall rate and mAP@0.5 value, it proves that the recognition effect of the gas-liquid two-phase leakage rapid identification model after training has reached the optimal value, and proceed to step 5; otherwise, return to step 3 and continue to use the training set to train the gas-liquid two-phase leakage rapid identification model.

[0046] Preferably, the alkylation unit on-site video recognition module is externally connected to a Web terminal, for providing an API service interface for the external Web terminal;

[0047] The request parameters of the video data source in the alkylation unit on-site video recognition module include the recognition type detectType, image name name, image number number, source address sourceAddress, source type sourceType, and recognition status status;

[0048] The return parameters in the on-site video recognition module of the alkylation unit include image name name, image number number, source address sourceAddress, source type sourceType, recognition status status, request status code, current recognition result currentResults, whether the parameter has been deleted deleted, recognition type detectType, last alarm recognition result lastAlarmResults, prompt message msg, and timestamp timestamp;

[0049] When the identification type detectType is 0, it indicates that the identification type is liquid phase leakage; when the identification type detectType is 1, it indicates that the identification type is gas phase leakage; when the source type sourceType is 0, it indicates that the source type is video stream; when the source type sourceType is 1, it indicates that the source type is mp4 standard video file; when the identification status status is 0, it indicates that the on-site video identification module of the alkylation unit is not turned on; when the identification status status is 1, it indicates that the on-site video identification module of the alkylation unit is turned on for identification; when the identification status status is 2, it indicates that the on-site video identification module of the alkylation unit has ended its identification.

[0050] A gas-liquid two-phase leakage linkage alarm system for an alkylation unit, wherein an alkylation unit on-site video recognition module constructed by the construction method of the alkylation unit on-site video recognition module as described above is internally arranged, including a combustible gas detector, a video sensing camera system, and a leakage accident early warning processing center, wherein the combustible gas detector and the video sensing camera system are respectively connected to the leakage accident early warning processing center;

[0051] The combustible gas detector is used to monitor the alkylation gas concentration at the alkylation unit site in real time;

[0052] The video sensing camera system is a camera device used to obtain on-site images of the alkylation unit and transmit them to the leakage accident early warning processing center in real time;

[0053] The leakage accident early warning and processing center is provided with a geographic information system module, an alkylation unit on-site video recognition module, an abnormal alarm processing module and a manual auxiliary judgment module; the geographic information system module is used to display the leakage risk video in the on-site monitoring area of ​​the alkylation unit, is connected with the combustible gas detector and the camera equipment in the video perception camera system, and is provided with an alarm; the alkylation unit on-site video recognition module is used to quickly identify the gas phase leakage and liquid phase leakage of the alkylation unit, and determine whether to trigger a first-level alarm or a second-level alarm; the abnormal alarm processing module is connected with the alarm, and is used to receive the leakage early warning signal issued by the alkylation unit on-site video recognition module and the combustible gas detector and use the alarm to alarm, alerting the on-site staff of the alkylation unit to the leakage; the manual auxiliary judgment module is used to manually verify the second-level alarm determined by the alkylation unit on-site video recognition module. When the manual auxiliary judgment module verifies that the leakage at the alkylation unit is true, the emergency response is immediately initiated to enter the leakage emergency disposal process. If the manual auxiliary judgment module verifies that the leakage at the alkylation unit is not true, the situation at the alkylation unit site continues to be monitored.

[0054] Preferably, the alarm is configured as a combustible gas alarm, which is connected to the leakage accident warning and processing center via intranet wifi or 5G signal, and includes a fixed combustible gas alarm and a portable combustible gas alarm, wherein the fixed combustible gas alarm is arranged at the pipe rack of the pipe gallery near the material reactor in the alkylation unit area, and the portable combustible gas alarm is carried by on-site inspection personnel.

[0055] Preferably, the resolution of the camera device in the video perception camera system is not less than 240P, and the camera device in the video perception camera system is connected to the leakage accident warning processing center via wifi or 5G signal.

[0056] Preferably, the on-site video recognition module of the alkylation unit is arranged on a GPU server of a leakage accident early warning processing center.

[0057] A method for quickly identifying and alarming a gas-liquid two-phase leakage in an alkylation unit, using the gas-liquid two-phase leakage linkage alarm system for an alkylation unit as described above, specifically comprises the following steps:

[0058] Step 1: a combustible gas detector is installed on site of the alkylation unit. The combustible gas detector has a preset alarm threshold for excessive alkyl gas concentration. The combustible gas detector compares the real-time measured alkyl gas concentration with the preset alarm threshold for excessive alkyl gas concentration to give a gas leakage warning.

[0059] When the alkane gas concentration measured by the combustible gas detector in real time exceeds the preset alkane gas concentration over-limit alarm threshold, the combustible gas detector sends a gas leakage warning signal to the leakage accident warning processing center; when the alkane gas concentration measured by the combustible gas detector in real time does not exceed the preset alkane gas concentration over-limit alarm threshold, the combustible gas detector does not send a gas leakage warning signal to the leakage accident warning processing center;

[0060] Step 2: an alkylation unit on-site video recognition module is set up at the alkylation unit site, and the alkylation unit on-site video recognition module is used to monitor in real time whether gas leakage and liquid leakage occur, and when a gas leakage is detected, a gas leakage warning signal is sent to a leakage accident warning processing center, and when a liquid leakage is detected, a liquid leakage warning signal is sent to the leakage accident warning processing center;

[0061] Step 3: After installing the on-site alarm of the alkylation unit, the combustible gas detector and the on-site video recognition module of the alkylation unit are used to monitor the situation of the alkylation unit in real time. The leakage accident warning and processing center triggers the alarm to issue a first-level alarm or a second-level alarm based on the leakage warning signal issued by the on-site combustible gas detector and the on-site video recognition module of the alkylation unit, so as to remind the on-site staff of the alkylation unit to handle it in time.

[0062] Preferably, the triggering condition of the first level alarm is:

[0063] The combustible gas detector and the alkylation unit on-site video recognition module simultaneously detect a leak at the alkylation unit site within 1 second, and when the combustible gas detector sends out a gas leak warning signal and the alkylation unit on-site video recognition module simultaneously sends out a gas leak warning signal and a liquid leak warning signal, the alarm triggers a first-level alarm;

[0064] The triggering conditions for the alarm to trigger the secondary alarm are:

[0065] The combustible gas detector sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds;

[0066] Or the on-site video recognition module of the alkylation unit sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds;

[0067] or the on-site video recognition module of the alkylation unit sends out a liquid leakage warning signal and the liquid leakage warning signal is not interrupted for 5 consecutive seconds;

[0068] or the combustible gas detector sends out a gas leakage warning signal and at the same time the on-site video recognition module of the alkylation unit sends out a gas leakage warning signal;

[0069] Or the combustible gas detector sends out a gas leakage warning signal and at the same time the on-site video recognition module of the alkylation unit sends out a liquid leakage warning signal;

[0070] Or the on-site video recognition module of the alkylation unit simultaneously sends out a gas leakage warning signal and a liquid leakage warning signal.

[0071] Preferably, the first-level alarm triggering and the second-level alarm triggering conditions are that the combustible gas detector sends out a gas leakage warning signal and the alkylation unit on-site video recognition module sends out a gas leakage warning signal at the same time, or the combustible gas detector sends out a gas leakage warning signal and the alkylation unit on-site video recognition module sends out a liquid leakage warning signal at the same time, or the alkylation unit on-site video recognition module sends out a gas leakage warning signal and a liquid leakage warning signal at the same time, and the alarm at the alkylation unit site immediately alarms;

[0072] The secondary alarm triggering condition is that the combustible gas detector sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds, or the on-site video recognition module of the alkylation unit sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds, or the on-site video recognition module of the alkylation unit sends out a liquid leakage warning signal and the liquid leakage warning signal is not interrupted for 5 consecutive seconds, and the alarm on the alkylation unit site is delayed to alarm.

[0073] The beneficial effects of the present invention are:

[0074] This patent provides a gas-liquid two-phase leakage linkage alarm system for an alkylation unit. By setting a leakage accident early warning processing center encapsulated with an alkylation unit on-site video recognition module in the gas-liquid two-phase leakage linkage alarm system of the alkylation unit, a gas-liquid two-phase leakage rapid recognition model trained by a deep network is used to quickly identify gas-liquid two-phase leakage in the on-site image of the alkylation unit. The deep learning gas-liquid two-phase rapid recognition is combined with the leakage gas concentration monitoring, which effectively improves the recognition accuracy and recognition efficiency of leakage accidents, solves the problems of low efficiency, high false alarm and missed alarm rate, and long response time caused by relying solely on manual video monitoring and recognition, meets the needs of petrochemical enterprises for rapid and accurate recognition of gas-liquid two-phase leakage at key risk sites of the alkylation unit, and provides technical support and decision support for rapid early warning of leakage risks.

[0075] The present invention also proposes a method for quickly identifying and alarming gas-liquid two-phase leakage in an alkylation unit. A gas-liquid two-phase leakage linkage alarm system for an alkylation unit is used to realize real-time monitoring of key areas of gas-liquid two-phase leakage risk at the alkylation unit site. By acquiring real-time video streams of key risk points at the alkylation unit site in real time, the gas-liquid two-phase leakage linkage alarm system for the alkylation unit is used to quickly identify the leakage type at the alkylation unit site and perform graded alarms, which is beneficial for assisting emergency duty personnel in conducting comprehensive monitoring, early warning and analysis of leakage accidents in complex petrochemical scenarios, and effectively improving the level of intelligent monitoring of leakage accidents in complex scenario conditions of alkylation units of petrochemical enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 The present invention is a structural schematic diagram of a method for rapid identification and alarm of gas-liquid two-phase leakage in an alkylation unit.

[0077] Figure 2 The present invention is a flow chart of alarm triggering in the method for rapid identification and alarm of gas-liquid two-phase leakage in an alkylation unit. DETAILED DESCRIPTION

[0078] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0079] Example 1

[0080] This embodiment proposes a method for constructing an on-site video recognition module of an alkylation unit, which is applicable to an alkylation unit and specifically includes the following steps:

[0081] Step 1: obtain a gas-liquid two-phase leakage image, preprocess each gas-liquid two-phase leakage image to obtain multiple gas-liquid two-phase leakage image samples, form a gas-liquid two-phase leakage image sample data set, and randomly allocate the gas-liquid two-phase leakage image samples in the gas-liquid two-phase leakage image sample data set to a training set and a test set in a ratio of 4:1 to form a training set and a test set.

[0082] In this embodiment, the collected gas-liquid two-phase leakage images are no less than 50,000, and the gas-liquid two-phase leakage images are collected through two ways:

[0083] The first approach is to obtain gas-liquid two-phase leakage images from historical accident video files of gas-liquid two-phase leakage in alkylation units;

[0084] The second approach is to use a gas-liquid two-phase leakage simulation experimental device to simulate a gas-liquid two-phase leakage accident in the laboratory and obtain a gas-liquid two-phase leakage image.

[0085] The liquid phase leakage in the gas-liquid two-phase leakage image is the acidic corrosive liquid in the reaction part of the alkylation unit. The liquid phase leakage has obvious identifiable characteristics such as dark black or dark yellow like soy sauce and slightly viscous, which can corrode the outer walls of valves, pipe galleries and other facilities. The main components of the gas phase leakage include methane gas and some hydrocarbon substances. The appearance is colorless and odorless, but when accompanied by the vaporization of some leaked liquid, it presents a "yellow-white mist" appearance.

[0086] The preprocessing process of gas-liquid two-phase leakage image is as follows:

[0087] According to the preset pixel resolution, in order to ensure the clarity of the gas-liquid two-phase leakage image and facilitate the accurate identification of the leakage type in the subsequent gas-liquid two-phase leakage image, the preset pixel resolution in this embodiment is 240P. After screening out the gas-liquid two-phase leakage images with pixel resolution lower than 240P, the gas-liquid two-phase leakage images with noise values ​​not meeting the preset noise range are screened out according to the preset noise range, and the gas-liquid two-phase leakage images without gas-liquid two-phase leakage are screened out to obtain the screened gas-liquid two-phase leakage images.

[0088] The screened gas-liquid two-phase leakage images are subjected to frame extraction processing to obtain multiple gas-liquid two-phase leakage images. On each gas-liquid two-phase leakage image, the position coordinates of the leakage liquid column, leakage droplets and leakage gas, as well as the type of leakage are marked on the gas-liquid two-phase leakage image using a labeling box, and a label document corresponding to the gas-liquid two-phase leakage image sample is generated. The gas-liquid two-phase leakage image with the label document is used as a gas-liquid two-phase leakage image sample.

[0089] Step 2: Build a rapid identification model for gas-liquid two-phase leakage based on the deep learning network.

[0090] The deep learning network used in this embodiment is the target detection network YOLOv7.

[0091] The target detection network YOLOv7 takes into account both recognition speed and accuracy, and includes an input end, a backbone feature extraction network and a detection head.

[0092] The input end is used to perform adaptive image scaling processing, Mosaic data enhancement processing and adaptive anchor frame calculation on the input gas-liquid two-phase leakage image, wherein the adaptive image scaling processing is used to unify the input format of the gas-liquid two-phase leakage image, so that multiple input gas-liquid two-phase leakage images have the same size after scaling processing; the Mosaic data enhancement processing obtains the spliced ​​gas-liquid two-phase leakage image by splicing multiple gas-liquid two-phase leakage images by random scaling, random cropping and random arrangement; the adaptive anchor frame calculation is used to automatically obtain the optimal anchor frame value during each training.

[0093] The backbone feature extraction network includes a plurality of CBS modules and a plurality of ELAN modules connected in sequence, and two adjacent ELAN modules are connected via an MP module.

[0094] The detection head includes a connected SPPCSPC structure, an ELAN module, an MP structure and a re-parameterized Rep structure, which are used as a classifier and a regressor of the network.

[0095] The SPPCSPC structure introduces the CSP structure into the SPP structure, and uses the SPP with the CSP mechanism to expand the receptive field for residual edge-assisted optimization and feature extraction.

[0096] The re-parameterized Rep structure introduces the RepConv structure and the residual structure to assist training, so as to reduce the complexity of the target detection network YOLOv7.

[0097] In this embodiment, 4 CBS modules, 4 ELAN modules and 3 MP modules are arranged in the backbone feature extraction network, wherein the CBS module includes a convolutional layer, a BN layer and a Silu layer, an activation function is arranged in the Silu layer, and the activation function is set to a sigmoid function.

[0098] The ELAN module includes a shallow feature extraction layer, a deep feature extraction layer and a reconstruction module, wherein the shallow feature extraction layer is set to a 3×3 convolution layer, the deep feature extraction layer is composed of stacked ELABs and a residual connection; the reconstruction module is composed of a 3×3 convolution and a PixelShuffle operation layer.

[0099] The MP module is provided with two branches for realizing super downsampling of parallel extraction and compression of features, wherein the first branch is provided with a maximum pooling layer and a 1×1 convolution layer connected in sequence; the other branch is provided with a 1×1 convolution layer and a convolution block with a 3×3 convolution kernel and a step size of 2 connected in sequence; the 1×1 convolution layers provided in the first branch and the second branch are both used to change the number of channels; a merging layer is provided after the first branch and the second branch, and the merging layer is used to add the processing results of the first branch and the second branch and output the result of super downsampling.

[0100] Step 3, using the training set to train the gas-liquid two-phase leakage rapid identification model to obtain the trained gas-liquid two-phase leakage rapid identification model, specifically includes the following steps:

[0101] Step 3.1, set the total number of training times, loss value, accuracy, recall rate and mAP@0.5 (average precision of image samples when the intersection over union (IoU) is 0.5) values.

[0102] In this embodiment, the loss rate is set to 0.1, the accuracy is set to 0.9, the recall rate is set to 0.9, and the mAP@0.5 value is set to 0.9.

[0103] Step 3.2, input the gas-liquid two-phase leakage image samples in the training set into the gas-liquid two-phase leakage rapid identification model constructed in step 2, and use the gas-liquid two-phase leakage rapid identification model to identify the leakage type in the gas-liquid two-phase leakage image to obtain the leakage type identification result of the gas-liquid two-phase leakage image.

[0104] Step 3.2: Optimize the rapid identification model for gas-liquid two-phase leakage based on the Adam algorithm, compare the leakage type identification result of the rapid identification model for gas-liquid two-phase leakage with the label document corresponding to the gas-liquid two-phase leakage image sample, calculate the loss value, accuracy, recall rate and mAP@0.5 value of the identification result of the rapid identification model for gas-liquid two-phase leakage, and compare them with the preset loss value, accuracy, recall rate and mAP@0.5 value. If the loss value is less than the preset loss value and the accuracy, recall rate and mAP@0.5 value are all greater than the preset accuracy, recall rate and mAP@0.5 value, proceed to step 3.4; otherwise, proceed to step 3.3.

[0105] Step 3.3, update the confidence threshold and IOU threshold of the gas-liquid two-phase leakage rapid identification model to obtain the updated gas-liquid two-phase leakage rapid identification model, continue to randomly select gas-liquid two-phase leakage image samples in the training set and input them into the updated gas-liquid two-phase leakage rapid identification model, use the updated gas-liquid two-phase leakage rapid identification model to identify the gas-liquid two-phase leakage image samples, obtain the identification result of the leakage type of the gas-liquid two-phase leakage image samples, and then return to step 3.2.

[0106] In this embodiment, the initial value of the confidence threshold in the gas-liquid two-phase leakage rapid identification model is set to 0.25, and the initial value of the IOU threshold is set to 0.45.

[0107] Step 3.4, complete the training of the gas-liquid two-phase leakage rapid identification model to obtain the trained gas-liquid two-phase leakage rapid identification model.

[0108] Step 4, use the test set to verify the recognition effect of the gas-liquid two-phase leakage rapid identification model after training. If the recognition effect of the gas-liquid two-phase leakage rapid identification model after training has reached the best, proceed to step 5; otherwise, return to step 3 and continue to train the gas-liquid two-phase leakage rapid identification model using the training set.

[0109] In this embodiment, the verification process of the gas-liquid two-phase leakage rapid identification model specifically includes the following steps:

[0110] Step 4.1, input the gas-liquid two-phase leakage image samples in the test set into the trained gas-liquid two-phase leakage rapid identification model, use the gas-liquid two-phase leakage rapid identification model to identify the leakage type in the gas-liquid two-phase leakage image samples, and obtain the leakage type identification result of the gas-liquid two-phase leakage image samples.

[0111] Step 4.2, compare the recognition results of the gas-liquid two-phase leakage rapid identification model of the gas-liquid two-phase leakage image sample with the label document corresponding to the gas-liquid two-phase leakage image sample, calculate the loss value, accuracy, recall rate and mAP@0.5 value of the recognition results of the gas-liquid two-phase leakage rapid identification model, and compare them with the preset loss value, accuracy, recall rate and mAP@0.5 value. If the loss value is less than the preset loss value and the accuracy, recall rate and mAP@0.5 value are all greater than the preset accuracy, recall rate and mAP@0.5 value, it proves that the recognition effect of the gas-liquid two-phase leakage rapid identification model after training has reached the optimal value, and proceed to step 5; otherwise, return to step 3 and continue to use the training set to train the gas-liquid two-phase leakage rapid identification model.

[0112] Step 5: encapsulate the verified gas-liquid two-phase leakage rapid identification model and set it in the alkylation unit on-site video identification module.

[0113] In this embodiment, the on-site video recognition module of the alkylation unit is externally connected to a Web terminal, which is used to provide an API service interface for the external Web terminal.

[0114] The request parameters of the video data source in the alkylation unit on-site video recognition module include the recognition type detectType, image name name, image number number, source address sourceAddress, source type sourceType, and recognition status status;

[0115] The return parameters in the on-site video recognition module of the alkylation unit include image name name, image number number, source address sourceAddress, source type sourceType, recognition status status, request status code, current recognition result currentResults, whether the parameter deleted is deleted, recognition type detectType, last alarm recognition result lastAlarmResults, prompt message msg, and timestamp timestamp.

[0116] When the identification type detectType is 0, it indicates that the identification type is liquid phase leakage; when the identification type detectType is 1, it indicates that the identification type is gas phase leakage; when the source type sourceType is 0, it indicates that the source type is video stream; when the source type sourceType is 1, it indicates that the source type is mp4 standard video file; when the identification status status is 0, it indicates that the on-site video identification module of the alkylation unit is not turned on; when the identification status status is 1, it indicates that the on-site video identification module of the alkylation unit is turned on for identification; when the identification status status is 2, it indicates that the on-site video identification module of the alkylation unit has ended its identification.

[0117] It can be seen that this embodiment builds a rapid identification model for gas-liquid two-phase leakage based on the target detection network YOLOv7. By training the rapid identification model for gas-liquid two-phase leakage, it realizes the rapid identification of gas-phase leakage and liquid-phase leakage in the gas-liquid two-phase leakage image at the alkylation unit site, laying a foundation for the monitoring and early warning of gas-liquid two-phase leakage.

[0118] Example 2

[0119] This embodiment proposes a gas-liquid two-phase leakage linkage alarm system for an alkylation unit, including a combustible gas detector, a video sensing camera system, and a leakage accident early warning processing center. The combustible gas detector and the video sensing camera system are respectively connected to the leakage accident early warning processing center.

[0120] The combustible gas detector is used to monitor the alkylation gas concentration at the site of the alkylation unit in real time.

[0121] The video perception camera system is a camera device used to obtain on-site images of the alkylation unit. In order to ensure the clarity of the obtained on-site gas-liquid two-phase leakage images of the alkylation unit, the resolution of the camera device used in the video perception camera system in this embodiment is not less than 240P. It is connected to the leakage accident early warning processing center via wifi or 5G signals, and is used to transmit the on-site gas-liquid two-phase leakage images of the alkylation unit to the leakage accident early warning processing center in real time.

[0122] The leakage accident early warning processing center is provided with a geographic information system module, an alkylation unit on-site video recognition module constructed by the construction method of the alkylation unit on-site video recognition module proposed in Example 1, an abnormal alarm processing module and a manual auxiliary judgment module; the geographic information system module is used to display the leakage risk video in the on-site monitoring area of ​​the alkylation unit, is connected to the combustible gas detector and the camera equipment in the video perception camera system, and is provided with an alarm; the alkylation unit on-site video recognition module is arranged on the GPU server of the leakage accident early warning processing center, and is used to quickly identify the gas phase leakage and liquid phase leakage of the alkylation unit and trigger the first level alarm or Level 2 alarm; the abnormal alarm processing module is connected to the alarm, and is used to receive the leakage warning signal issued by the on-site video recognition module of the alkylation unit and the combustible gas detector and use the alarm to alarm, alerting the on-site staff of the alkylation unit to the leakage; the manual auxiliary analysis and judgment module is used to manually verify the level 2 alarm determined by the on-site video recognition module of the alkylation unit. When the manual auxiliary analysis and judgment module verifies that the leakage at the alkylation unit is true, the emergency response is immediately initiated to enter the leakage emergency disposal process. If the manual auxiliary analysis and judgment module verifies that the leakage at the alkylation unit is not true, the situation at the alkylation unit site continues to be monitored.

[0123] In this embodiment, the alarm is configured as a combustible gas alarm installed at the alkylation unit site, which is connected to the leakage accident warning and processing center via intranet wifi or 5G signal, including a fixed combustible gas alarm and a portable combustible gas alarm, wherein the fixed combustible gas alarm is installed at the pipe rack of the pipe gallery near the material reactor in the alkylation unit area, and the portable combustible gas alarm is carried by on-site inspection personnel.

[0124] Example 3

[0125] This embodiment proposes a method for quickly identifying and alarming a gas-liquid two-phase leakage in an alkylation unit, using the gas-liquid two-phase leakage linkage alarm system for an alkylation unit described in Example 2. Figure 1 As shown, the specific steps include:

[0126] Step 1: Install a combustible gas detector at the alkylation unit. The combustible gas detector has a preset alarm threshold for excessive alkyl gas concentration. The combustible gas detector compares the real-time measured alkyl gas concentration with the preset alarm threshold for excessive alkyl gas concentration to issue a gas leakage warning.

[0127] When the alkane gas concentration measured in real time by the combustible gas detector exceeds the preset alkane gas concentration over-limit alarm threshold, the combustible gas detector sends a gas leakage warning signal to the leakage accident warning processing center; when the alkane gas concentration measured in real time by the combustible gas detector does not exceed the preset alkane gas concentration over-limit alarm threshold, the combustible gas detector does not send a gas leakage warning signal to the leakage accident warning processing center.

[0128] Step 2: an alkylation unit on-site video recognition module is set up at the alkylation unit site, and the alkylation unit on-site video recognition module is used to monitor in real time whether gas leakage and liquid leakage occur, and when a gas leakage is detected, a gas leakage warning signal is sent to the leakage accident warning processing center, and when a liquid leakage is detected, a liquid leakage warning signal is sent to the leakage accident warning processing center.

[0129] Step 3: After installing the on-site alarm of the alkylation unit, the combustible gas detector and the on-site video recognition module of the alkylation unit are used to monitor the situation of the alkylation unit in real time. The leakage accident warning and processing center triggers the alarm to issue a first-level alarm or a second-level alarm based on the leakage warning signal issued by the on-site combustible gas detector and the on-site video recognition module of the alkylation unit, so as to remind the on-site staff of the alkylation unit to handle it in time.

[0130] In this embodiment, the alarm triggering conditions are as follows: Figure 2 shown.

[0131] The triggering conditions of the first level alarm are:

[0132] When the combustible gas detector and the alkylation unit on-site video recognition module simultaneously detect a leak at the alkylation unit site within 1 second, and the combustible gas detector sends out a gas leak warning signal and the alkylation unit on-site video recognition module simultaneously sends out a gas leak warning signal and a liquid leak warning signal, the alarm triggers a first-level alarm.

[0133] The triggering conditions for the alarm to trigger the secondary alarm are:

[0134] The combustible gas detector sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds;

[0135] Or the on-site video recognition module of the alkylation unit sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds;

[0136] or the on-site video recognition module of the alkylation unit sends out a liquid leakage warning signal and the liquid leakage warning signal is not interrupted for 5 consecutive seconds;

[0137] or the combustible gas detector sends out a gas leakage warning signal and at the same time the on-site video recognition module of the alkylation unit sends out a gas leakage warning signal;

[0138] Or the combustible gas detector sends out a gas leakage warning signal and at the same time the on-site video recognition module of the alkylation unit sends out a liquid leakage warning signal;

[0139] Or the on-site video recognition module of the alkylation unit simultaneously sends out a gas leakage warning signal and a liquid leakage warning signal.

[0140] The first-level alarm triggering and the second-level alarm triggering conditions are that the combustible gas detector sends out a gas leakage warning signal and the alkylation unit on-site video recognition module sends out a gas leakage warning signal, or the combustible gas detector sends out a gas leakage warning signal and the alkylation unit on-site video recognition module sends out a liquid leakage warning signal, or the alkylation unit on-site video recognition module sends out a gas leakage warning signal and a liquid leakage warning signal at the same time, and the alkylation unit on-site alarm immediately alarms;

[0141] The secondary alarm triggering condition is that the combustible gas detector sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds, or the on-site video recognition module of the alkylation unit sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds, or the on-site video recognition module of the alkylation unit sends out a liquid leakage warning signal and the liquid leakage warning signal is not interrupted for 5 consecutive seconds, and the alarm on the alkylation unit site is delayed to alarm.

[0142] It can be seen that the method for rapid identification and alarm of gas-liquid two-phase leakage in the alkylation unit proposed in this embodiment is a multi-level alarm mode based on multiple abnormal leakage forms. By combining the combustible gas detector with the on-site video recognition module of the alkylation unit, rapid identification and alarm of gas-liquid two-phase leakage in the alkylation unit is achieved, which effectively improves the identification accuracy and efficiency of leakage accidents and solves the problems of low efficiency, high false alarm and missed alarm rates and long response time caused by relying solely on manual video monitoring and identification.

[0143] Example 4

[0144] In this embodiment, the gas-liquid two-phase leakage linkage alarm system of the alkylation unit described in Example 2 and the gas-liquid two-phase leakage rapid identification and alarm method of the alkylation unit described in Example 3 are applied to the alkylation unit site.

[0145] In this embodiment, a leakage simulation experimental device for gas-liquid two-phase leakage is placed near the alkylation unit reactor E01 to simulate a leakage accident, ensuring that the simulated accident location is within the sensing range of the explosion-proof camera and the online fixed combustible gas detector Liye LY800.

[0146] An alkylation unit gas-liquid two-phase leakage linkage alarm system is installed near the alkylation unit reactor E01. The alkylation unit gas-liquid two-phase leakage linkage alarm system is used to monitor whether a gas-liquid two-phase leakage occurs near the alkylation unit reactor E01. After the alkylation unit gas-liquid two-phase leakage linkage alarm system is turned on, a gas-liquid two-phase leakage is simulated near the alkylation unit reactor E01. At the beginning of the actual measurement phase, the online fixed combustible gas detector Liye LY800 sends a gas leakage warning signal and the alkylation unit on-site video recognition module simultaneously sends a gas leakage warning signal and a liquid leakage warning signal. The on-site alarm point is highlighted in red on the online map in the geographic information system module of the alkylation unit gas-liquid two-phase leakage linkage alarm system, and the alarm is triggered to immediately send a first-level alarm prompt sound, immediately start the emergency procedure, send an alarm message to the safety management personnel, and push the leakage accident information, such as "a leakage accident occurred in the alkylation unit reactor E01, please confirm and take emergency measures immediately."

[0147] Example 5

[0148] In this embodiment, the gas-liquid two-phase leakage linkage alarm system of the alkylation unit described in Embodiment 2 and the gas-liquid two-phase leakage rapid identification and alarm method of the alkylation unit described in Embodiment 3 are applied to the alkylation unit site.

[0149] In this embodiment, a leakage simulation experimental device for gas-liquid two-phase leakage is placed near the alkylation unit reactor E05 to simulate a leakage accident. Multiple explosion-proof cameras and multiple online fixed combustible gas detectors Liye LY800 are set near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system of the alkylation unit can obtain the camera visible light video and the combustible gas alarm signal in real time.

[0150] An alkylation unit gas-liquid two-phase leakage linkage alarm system was installed near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system was used to monitor whether there was a gas-liquid two-phase leakage near the alkylation unit reactor E05. After the gas-liquid two-phase leakage linkage alarm system was turned on, a gas-liquid two-phase leakage was simulated near the alkylation unit reactor E05. At the beginning of the actual measurement phase, the online fixed combustible gas detector Liye LY800 issued a gas leakage warning signal and the gas leakage warning signal was not interrupted for 5 consecutive seconds. During this period, the on-site video identification module of the alkylation unit did not issue a gas leakage warning signal or a liquid leakage warning signal. The on-site alarm point is highlighted in red on the online map in the geographic information system module of the alkylation unit gas-liquid two-phase leakage linkage alarm system, and the alarm is triggered to immediately sound a secondary alarm. After the abnormal alarm processing module confirms the triggering alarm condition, the manual auxiliary judgment module of the alkylation unit gas-liquid two-phase leakage linkage alarm system sends a manual verification signal to the video monitoring center or the emergency command center to determine whether it is a misjudgment based on the results of manual observation. If manual observation determines that a leakage accident has occurred, the emergency procedure is immediately initiated, and an alarm message is sent to the safety management personnel to push the leakage accident information, such as "Alkylation unit reactor E05 has a leakage accident, please confirm immediately and take emergency measures." If manual observation does not determine that a leakage accident has occurred, the verification personnel eliminate and clear the alarm prompt in the manual auxiliary judgment module.

[0151] Example 6

[0152] In this embodiment, the gas-liquid two-phase leakage linkage alarm system of the alkylation unit described in Embodiment 2 and the gas-liquid two-phase leakage rapid identification and alarm method of the alkylation unit described in Embodiment 3 are applied to the alkylation unit site.

[0153] In this embodiment, a leakage simulation experimental device for gas-liquid two-phase leakage is placed near the alkylation unit reactor E05 to simulate a leakage accident. Multiple explosion-proof cameras and multiple online fixed combustible gas detectors Liye LY800 are set near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system of the alkylation unit can obtain the camera visible light video and the combustible gas alarm signal in real time.

[0154] An alkylation unit gas-liquid two-phase leakage linkage alarm system was installed near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system was used to monitor whether there was a gas-liquid two-phase leakage near the alkylation unit reactor E05. After the gas-liquid two-phase leakage linkage alarm system was turned on, a gas-liquid two-phase leakage was simulated near the alkylation unit reactor E05. At the beginning of the actual measurement phase, the online fixed combustible gas detector Liye LY800 did not issue a gas leakage warning signal, but the on-site video identification module of the alkylation unit issued a gas leakage warning signal and the gas leakage warning signal was not interrupted for 5 consecutive seconds. The on-site alarm point is highlighted in red on the online map in the geographic information system module of the alkylation unit gas-liquid two-phase leakage linkage alarm system, and the alarm is triggered to immediately sound a secondary alarm. After the abnormal alarm processing module confirms the triggering alarm condition, the manual auxiliary judgment module of the alkylation unit gas-liquid two-phase leakage linkage alarm system sends a manual verification signal to the video monitoring center or the emergency command center to determine whether it is a misjudgment based on the results of manual observation. If manual observation determines that a leakage accident has occurred, the emergency procedure is immediately initiated, and an alarm message is sent to the safety management personnel to push the leakage accident information, such as "Alkylation unit reactor E05 has a leakage accident, please confirm immediately and take emergency measures." If manual observation does not determine that a leakage accident has occurred, the verification personnel eliminate and clear the alarm prompt in the manual auxiliary judgment module.

[0155] Example 7

[0156] In this embodiment, the gas-liquid two-phase leakage linkage alarm system of the alkylation unit described in Embodiment 2 and the gas-liquid two-phase leakage rapid identification and alarm method of the alkylation unit described in Embodiment 3 are applied to the alkylation unit site.

[0157] In this embodiment, a leakage simulation experimental device for gas-liquid two-phase leakage is placed near the alkylation unit reactor E05 to simulate a leakage accident. Multiple explosion-proof cameras and multiple online fixed combustible gas detectors Liye LY800 are set near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system of the alkylation unit can obtain the camera visible light video and the combustible gas alarm signal in real time.

[0158] An alkylation unit gas-liquid two-phase leakage linkage alarm system is installed near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system is used to monitor whether there is a gas-liquid two-phase leakage near the alkylation unit reactor E05. After the gas-liquid two-phase leakage linkage alarm system is turned on, a gas-liquid two-phase leakage is simulated near the alkylation unit reactor E05. At the beginning of the actual measurement phase, the online fixed combustible gas detector Liye LY800 did not issue a gas leakage warning signal, but the on-site video identification module of the alkylation unit issued a liquid leakage warning signal and the liquid leakage warning signal was not interrupted for 5 consecutive seconds. The on-site alarm point is highlighted in red on the online map in the geographic information system module of the alkylation unit gas-liquid two-phase leakage linkage alarm system, and the alarm is triggered to immediately sound a secondary alarm. After the abnormal alarm processing module confirms the triggering alarm condition, the manual auxiliary judgment module of the alkylation unit gas-liquid two-phase leakage linkage alarm system sends a manual verification signal to the video monitoring center or the emergency command center to determine whether it is a misjudgment based on the results of manual observation. If manual observation determines that a leakage accident has occurred, the emergency procedure is immediately initiated, and an alarm message is sent to the safety management personnel to push the leakage accident information, such as "Alkylation unit reactor E05 has a leakage accident, please confirm immediately and take emergency measures." If manual observation does not determine that a leakage accident has occurred, the verification personnel eliminate and clear the alarm prompt in the manual auxiliary judgment module.

[0159] Example 8

[0160] In this embodiment, the gas-liquid two-phase leakage linkage alarm system of the alkylation unit described in Embodiment 2 and the gas-liquid two-phase leakage rapid identification and alarm method of the alkylation unit described in Embodiment 3 are applied to the alkylation unit site.

[0161] In this embodiment, a leakage simulation experimental device for gas-liquid two-phase leakage is placed near the alkylation unit reactor E05 to simulate a leakage accident. Multiple explosion-proof cameras and multiple online fixed combustible gas detectors Liye LY800 are set near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system of the alkylation unit can obtain the camera visible light video and the combustible gas alarm signal in real time.

[0162] An alkylation unit gas-liquid two-phase leakage linkage alarm system is installed near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system is used to monitor whether there is a gas-liquid two-phase leakage near the alkylation unit reactor E05. After the gas-liquid two-phase leakage linkage alarm system is turned on, a gas-liquid two-phase leakage is simulated near the alkylation unit reactor E05. At the beginning of the actual measurement phase, the online fixed combustible gas detector Liye LY800 sends a gas leakage warning signal and the alkylation unit on-site video identification module sends a gas leakage warning signal. The on-site alarm point is highlighted in red on the online map in the geographic information system module of the alkylation unit gas-liquid two-phase leakage linkage alarm system, and the alarm is triggered to immediately sound a secondary alarm. After the abnormal alarm processing module confirms the triggering alarm condition, the manual auxiliary judgment module of the alkylation unit gas-liquid two-phase leakage linkage alarm system sends a manual verification signal to the video monitoring center or the emergency command center to determine whether it is a misjudgment based on the results of manual observation. If manual observation determines that a leakage accident has occurred, the emergency procedure is immediately initiated, and an alarm message is sent to the safety management personnel to push the leakage accident information, such as "Alkylation unit reactor E05 has a leakage accident, please confirm immediately and take emergency measures." If manual observation does not determine that a leakage accident has occurred, the verification personnel eliminate and clear the alarm prompt in the manual auxiliary judgment module.

[0163] Example 9

[0164] In this embodiment, the gas-liquid two-phase leakage linkage alarm system of the alkylation unit described in Embodiment 2 and the gas-liquid two-phase leakage rapid identification and alarm method of the alkylation unit described in Embodiment 3 are applied to the alkylation unit site.

[0165] In this embodiment, a leakage simulation experimental device for gas-liquid two-phase leakage is placed near the alkylation unit reactor E05 to simulate a leakage accident. Multiple explosion-proof cameras and multiple online fixed combustible gas detectors Liye LY800 are set near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system of the alkylation unit can obtain the camera visible light video and the combustible gas alarm signal in real time.

[0166] An alkylation unit gas-liquid two-phase leakage linkage alarm system is installed near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system is used to monitor whether there is a gas-liquid two-phase leakage near the alkylation unit reactor E05. After the gas-liquid two-phase leakage linkage alarm system is turned on, a gas-liquid two-phase leakage is simulated near the alkylation unit reactor E05. At the beginning of the actual measurement phase, the online fixed combustible gas detector Liye LY800 sends out a gas leakage warning signal, and the alkylation unit on-site video identification module sends out a liquid leakage warning signal. The on-site alarm point is highlighted in red on the online map in the geographic information system module of the alkylation unit gas-liquid two-phase leakage linkage alarm system, and the alarm is triggered to immediately sound a secondary alarm. After the abnormal alarm processing module confirms the triggering alarm condition, the manual auxiliary judgment module of the alkylation unit gas-liquid two-phase leakage linkage alarm system sends a manual verification signal to the video monitoring center or the emergency command center to determine whether it is a misjudgment based on the results of manual observation. If manual observation determines that a leakage accident has occurred, the emergency procedure is immediately initiated, and an alarm message is sent to the safety management personnel to push the leakage accident information, such as "Alkylation unit reactor E05 has a leakage accident, please confirm immediately and take emergency measures." If manual observation does not determine that a leakage accident has occurred, the verification personnel eliminate and clear the alarm prompt in the manual auxiliary judgment module.

[0167] Example 10

[0168] In this embodiment, the gas-liquid two-phase leakage linkage alarm system of the alkylation unit described in Embodiment 2 and the gas-liquid two-phase leakage rapid identification and alarm method of the alkylation unit described in Embodiment 3 are applied to the alkylation unit site.

[0169] In this embodiment, a leakage simulation experimental device for gas-liquid two-phase leakage is placed near the alkylation unit reactor E05 to simulate a leakage accident. Multiple explosion-proof cameras and multiple online fixed combustible gas detectors Liye LY800 are set near the alkylation unit reactor E05. The gas-liquid two-phase leakage linkage alarm system of the alkylation unit can obtain the camera visible light video and the combustible gas alarm signal in real time.

[0170] An alkylation unit gas-liquid two-phase leakage linkage alarm system is installed near the alkylation unit reactor E05. The alkylation unit gas-liquid two-phase leakage linkage alarm system is used to monitor whether there is a gas-liquid two-phase leakage near the alkylation unit reactor E05. After the alkylation unit gas-liquid two-phase leakage linkage alarm system is turned on, a gas-liquid two-phase leakage is simulated near the alkylation unit reactor E05. At the beginning of the actual measurement phase, the online fixed combustible gas detector Liye LY800 did not issue a gas leakage warning signal, but the alkylation unit on-site video identification module simultaneously issued a gas leakage warning signal and a liquid leakage warning signal. The on-site alarm point is highlighted in red on the online map in the geographic information system module of the alkylation unit gas-liquid two-phase leakage linkage alarm system, and the alarm is triggered to immediately sound a secondary alarm. After the abnormal alarm processing module confirms the triggering alarm condition, the manual auxiliary judgment module of the alkylation unit gas-liquid two-phase leakage linkage alarm system sends a manual verification signal to the video monitoring center or the emergency command center to determine whether it is a misjudgment based on the results of manual observation. If manual observation determines that a leakage accident has occurred, the emergency procedure is immediately initiated, and an alarm message is sent to the safety management personnel to push the leakage accident information, such as "Alkylation unit reactor E05 has a leakage accident, please confirm immediately and take emergency measures." If manual observation does not determine that a leakage accident has occurred, the verification personnel eliminate and clear the alarm prompt in the manual auxiliary judgment module.

[0171] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms used herein are only for describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0172] In the present invention, the directions or positional relationships indicated by terms such as "upper", "lower", "bottom", "top", etc. are based on the directions or positional relationships shown in the accompanying drawings. They are relational words determined only for the convenience of describing the structural relationships of the various parts or elements of the present invention. They do not specifically refer to any part or element in the present invention and cannot be understood as limitations on the present invention.

[0173] In the present invention, terms such as "connected" and "connection" should be understood in a broad sense, indicating that the connection can be fixed, integral or detachable; it can be directly connected or indirectly connected through an intermediate medium. Relevant scientific research or technical personnel in this field can determine the specific meaning of the above terms in the present invention according to specific circumstances, and they should not be understood as limiting the present invention.

[0174] Of course, the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by technicians in this technical field within the essential scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A method for constructing an on-site video recognition module for an alkylation unit, characterized in that: Applicable to alkylation unit, specifically comprising the following steps: Step 1, obtaining a gas-liquid two-phase leakage image, preprocessing each gas-liquid two-phase leakage image, obtaining multiple gas-liquid two-phase leakage image samples, and forming a gas-liquid two-phase leakage image sample data set, wherein the gas-liquid two-phase leakage image sample data set includes a training set and a test set; Step 2: Building a rapid identification model for gas-liquid two-phase leakage based on a deep learning network; Step 3, using the training set to train the gas-liquid two-phase leakage rapid identification model to obtain a trained gas-liquid two-phase leakage rapid identification model; Step 4: Use the test set to verify the recognition effect of the gas-liquid two-phase leakage rapid identification model after training. If the recognition effect of the gas-liquid two-phase leakage rapid identification model after training has reached the best, proceed to step 5; otherwise, return to step 3 and continue to train the gas-liquid two-phase leakage rapid identification model using the training set; Step 5: encapsulate the verified gas-liquid two-phase leakage rapid identification model and set it in the alkylation unit on-site video identification module.

2. The method for constructing an on-site video recognition module for an alkylation unit according to claim 1, characterized in that: The gas-liquid two-phase leakage image is collected in two ways: The first approach is to obtain gas-liquid two-phase leakage images from historical accident video files of gas-liquid two-phase leakage in alkylation units; The second approach is to use a gas-liquid two-phase leakage simulation experimental device to simulate a gas-liquid two-phase leakage accident in the laboratory and obtain a gas-liquid two-phase leakage image.

3. The method for constructing an on-site video recognition module for an alkylation unit according to claim 1, characterized in that: The preprocessing process of the gas-liquid two-phase leakage image is as follows: According to a preset pixel resolution, after filtering out gas-liquid two-phase leakage images with pixel resolutions lower than the preset pixel resolution, according to a preset noise range, filtering out gas-liquid two-phase leakage images with noise values ​​not meeting the preset noise range, and filtering out gas-liquid two-phase leakage images without gas-liquid two-phase leakage objects, to obtain filtered gas-liquid two-phase leakage images; The screened gas-liquid two-phase leakage images are subjected to frame extraction processing to obtain multiple gas-liquid two-phase leakage images. On each gas-liquid two-phase leakage image, the position coordinates of the leakage liquid column, leakage droplets and leakage gas, as well as the type of leakage are marked on the gas-liquid two-phase leakage image using a labeling box, and a label document corresponding to the gas-liquid two-phase leakage image sample is generated. The gas-liquid two-phase leakage image with the label document is used as a gas-liquid two-phase leakage image sample.

4. The method for constructing an on-site video recognition module for an alkylation unit according to claim 3, characterized in that: The preset pixel resolution is 240P.

5. The method for constructing an on-site video recognition module of an alkylation unit according to claim 1, characterized in that: The number of gas-liquid two-phase leakage image samples is not less than 50,000; The gas-liquid two-phase leakage image samples in the gas-liquid two-phase leakage image sample data set are randomly allocated to a training set and a test set, and the ratio of the number of gas-liquid two-phase leakage image samples in the training set to the number of gas-liquid two-phase leakage image samples in the test set is not less than 4:

1.

6. The method for constructing an on-site video recognition module of an alkylation unit according to claim 1, characterized in that: The deep learning network selects the target detection network YOLOv7, and uses the target detection network YOLOv7 to build a rapid recognition model for gas-liquid two-phase leakage.

7. The method for constructing an on-site video recognition module of an alkylation unit according to claim 6, characterized in that: The target detection network YOLOv7 takes into account both recognition speed and accuracy, including an input end, a backbone feature extraction network and a detection head; The input end is used to perform adaptive image scaling processing, Mosaic data enhancement processing and adaptive anchor frame calculation on the input gas-liquid two-phase leakage image, wherein the adaptive image scaling processing is used to unify the input format of the gas-liquid two-phase leakage image, so that multiple input gas-liquid two-phase leakage images have the same size through scaling processing; the Mosaic data enhancement processing obtains the spliced ​​gas-liquid two-phase leakage image by splicing multiple gas-liquid two-phase leakage images in a random scaling, random cropping and random arrangement manner; the adaptive anchor frame calculation is used to automatically obtain the optimal anchor frame value during each training; The backbone feature extraction network includes a plurality of CBS modules and a plurality of ELAN modules connected in sequence, and two adjacent ELAN modules are connected via an MP module; The detection head includes a connected SPPCSPC structure, an ELAN module, an MP structure and a re-parameterized Rep structure, which are used as a classifier and a regressor of the network; The SPPCSPC structure introduces the CSP structure into the SPP structure, and uses the SPP with the CSP structure to expand the receptive field for residual edge-assisted optimization and feature extraction; The re-parameterized Rep structure introduces the RepConv structure and the residual structure to assist training, so as to reduce the complexity of the target detection network YOLOv7.

8. The method for constructing an on-site video recognition module for an alkylation unit according to claim 7, characterized in that: The CBS module includes a convolution layer, a BN layer and a Silu layer; an activation function is set in the Silu layer, and the activation function is set to a sigmoid function.

9. The method for constructing an on-site video recognition module for an alkylation unit according to claim 7, characterized in that: The ELAN module includes a shallow feature extraction layer, a deep feature extraction layer and a reconstruction module; The shallow feature extraction layer is set to a 3×3 convolutional layer; The deep feature extraction layer consists of stacked ELABs and a residual connection; The reconstruction module consists of a 3×3 convolution and a PixelShuffle operation layer.

10. The method for constructing an on-site video recognition module of an alkylation unit according to claim 7, characterized in that: The MP module is provided with two branches for realizing super downsampling of parallel extraction and compression of features, wherein the first branch is provided with a maximum pooling layer and a 1×1 convolution layer connected in sequence; the other branch is provided with a 1×1 convolution layer and a convolution block with a 3×3 convolution kernel and a step size of 2 connected in sequence; the 1×1 convolution layers provided in the first branch and the second branch are both used to change the number of channels; a merging layer is provided after the first branch and the second branch, and the merging layer is used to add the processing results of the first branch and the second branch and output the result of super downsampling.

11. The method for constructing an on-site video recognition module of an alkylation unit according to claim 1, characterized in that: The step 3 specifically includes the following steps: Step 3.1, set the total number of training times, loss value, accuracy, recall rate and mAP@0.5 value; The mAP@0.5 is the average precision of the image samples when the intersection over union (IoU) is 0.5; Step 3.2, inputting the gas-liquid two-phase leakage image samples in the training set into the gas-liquid two-phase leakage rapid identification model constructed in step 2, using the gas-liquid two-phase leakage rapid identification model to identify the leakage type in the gas-liquid two-phase leakage image, and obtaining the leakage type identification result of the gas-liquid two-phase leakage image; Step 3.2, optimize the gas-liquid two-phase leakage rapid identification model based on the Adam algorithm, compare the leakage type identification result of the gas-liquid two-phase leakage rapid identification model with the label document corresponding to the gas-liquid two-phase leakage image sample, calculate the loss value, accuracy, recall rate and mAP@0.5 value of the gas-liquid two-phase leakage rapid identification model, and compare them with the preset loss value, accuracy, recall rate and mAP@0.5 value. If the loss value is less than the preset loss value and the accuracy, recall rate and mAP@0.5 value are all greater than the preset accuracy, recall rate and mAP@0.5 value, proceed to step 3.4; otherwise, proceed to step 3.3; Step 3.3, update the confidence threshold and IOU threshold of the gas-liquid two-phase leakage rapid identification model to obtain an updated gas-liquid two-phase leakage rapid identification model, continue to randomly select gas-liquid two-phase leakage image samples in the training set and input them into the updated gas-liquid two-phase leakage rapid identification model, use the updated gas-liquid two-phase leakage rapid identification model to identify the gas-liquid two-phase leakage image samples, obtain the gas-liquid two-phase leakage image sample leakage type identification result, and then return to step 3.2; Step 3.4, complete the training of the gas-liquid two-phase leakage rapid identification model to obtain the trained gas-liquid two-phase leakage rapid identification model.

12. The method for constructing an on-site video recognition module of an alkylation unit according to claim 11, characterized in that: In the gas-liquid two-phase leakage rapid identification model, the initial value of the confidence threshold is set to 0.25, and the maximum value of the confidence threshold is not higher than 0.5; the initial value of the IOU threshold is set to 0.

45.

13. The method for constructing an on-site video recognition module of an alkylation unit according to claim 11, characterized in that: The step 4 specifically includes the following steps: Step 4.1, input the gas-liquid two-phase leakage image samples in the test set into the trained gas-liquid two-phase leakage rapid identification model, use the gas-liquid two-phase leakage rapid identification model to identify the leakage type in the gas-liquid two-phase leakage image samples, and obtain the leakage type identification result of the gas-liquid two-phase leakage image samples; Step 4.2, compare the recognition results of the gas-liquid two-phase leakage rapid identification model of the gas-liquid two-phase leakage image sample with the label document corresponding to the gas-liquid two-phase leakage image sample, calculate the loss value, accuracy, recall rate and mAP@0.5 value of the recognition results of the gas-liquid two-phase leakage rapid identification model, and compare them with the preset loss value, accuracy, recall rate and mAP@0.5 value. If the loss value is less than the preset loss value and the accuracy, recall rate and mAP@0.5 value are all greater than the preset accuracy, recall rate and mAP@0.5 value, it proves that the recognition effect of the gas-liquid two-phase leakage rapid identification model after training has reached the optimal value, and proceed to step 5; otherwise, return to step 3 and continue to use the training set to train the gas-liquid two-phase leakage rapid identification model.

14. The method for constructing an on-site video recognition module of an alkylation unit according to claim 1, characterized in that: The alkylation unit on-site video recognition module is externally connected to a Web terminal, and is used to provide an API service interface for the external Web terminal; The request parameters of the video data source in the alkylation unit on-site video recognition module include the recognition type detectType, image name name, image number number, source address sourceAddress, source type sourceType, and recognition status status; The return parameters in the on-site video recognition module of the alkylation unit include image name name, image number number, source address sourceAddress, source type sourceType, recognition status status, request status code, current recognition result currentResults, whether the parameter has been deleted deleted, recognition type detectType, last alarm recognition result lastAlarmResults, prompt message msg, and timestamp timestamp; When the identification type detectType is 0, it indicates that the identification type is liquid phase leakage; when the identification type detectType is 1, it indicates that the identification type is gas phase leakage; when the source type sourceType is 0, it indicates that the source type is video stream; when the source type sourceType is 1, it indicates that the source type is mp4 standard video file; when the identification status status is 0, it indicates that the on-site video identification module of the alkylation unit is not turned on; when the identification status status is 1, it indicates that the on-site video identification module of the alkylation unit is turned on for identification; when the identification status status is 2, it indicates that the on-site video identification module of the alkylation unit has ended its identification.

15. A gas-liquid two-phase leakage linkage alarm system for an alkylation unit, characterized in that: The internal arrangement is an alkylation unit on-site video recognition module constructed by the construction method of the alkylation unit on-site video recognition module according to any one of claims 1 to 14, comprising a combustible gas detector, a video sensing camera system, and a leakage accident early warning processing center, wherein the combustible gas detector and the video sensing camera system are respectively connected to the leakage accident early warning processing center; The combustible gas detector is used to monitor the alkylation gas concentration at the alkylation unit site in real time; The video sensing camera system is a camera device used to obtain on-site images of the alkylation unit and transmit them to the leakage accident early warning processing center in real time; The leakage accident early warning and processing center is provided with a geographic information system module, an alkylation unit on-site video recognition module, an abnormal alarm processing module and a manual auxiliary judgment module; the geographic information system module is used to display the leakage risk video in the on-site monitoring area of ​​the alkylation unit, is connected with the combustible gas detector and the camera equipment in the video perception camera system, and is provided with an alarm; the alkylation unit on-site video recognition module is used to quickly identify the gas phase leakage and liquid phase leakage of the alkylation unit, and determine whether to trigger a first-level alarm or a second-level alarm; the abnormal alarm processing module is connected with the alarm, and is used to receive the leakage early warning signal issued by the alkylation unit on-site video recognition module and the combustible gas detector and use the alarm to alarm, alerting the on-site staff of the alkylation unit to the leakage; the manual auxiliary judgment module is used to manually verify the second-level alarm determined by the alkylation unit on-site video recognition module. When the manual auxiliary judgment module verifies that the leakage at the alkylation unit is true, the emergency response is immediately initiated to enter the leakage emergency disposal process. If the manual auxiliary judgment module verifies that the leakage at the alkylation unit is not true, the situation at the alkylation unit site continues to be monitored.

16. The gas-liquid two-phase leakage linkage alarm system of the alkylation unit according to claim 15, characterized in that: The alarm is configured as a combustible gas alarm, which is connected to the leakage accident warning and processing center via intranet wifi or 5G signal, and includes a fixed combustible gas alarm and a portable combustible gas alarm. The fixed combustible gas alarm is arranged at the pipe rack of the pipe gallery near the material reactor in the alkylation unit area, and the portable combustible gas alarm is carried by on-site inspection personnel.

17. The gas-liquid two-phase leakage linkage alarm system of the alkylation unit according to claim 15, characterized in that: The resolution of the camera equipment in the video perception camera system is not less than 240P, and the camera equipment in the video perception camera system is connected to the leakage accident warning processing center via wifi or 5G signal.

18. The gas-liquid two-phase leakage linkage alarm system of the alkylation unit according to claim 15, characterized in that: The alkylation unit on-site video recognition module is arranged on the GPU server of the leakage accident early warning processing center.

19. A method for rapid identification and alarm of gas-liquid two-phase leakage in an alkylation unit, characterized in that: The gas-liquid two-phase leakage linkage alarm system for an alkylation unit according to any one of claims 15 to 18 specifically comprises the following steps: Step 1: a combustible gas detector is installed on site of the alkylation unit. The combustible gas detector has a preset alarm threshold for excessive alkyl gas concentration. The combustible gas detector compares the real-time measured alkyl gas concentration with the preset alarm threshold for excessive alkyl gas concentration to give a gas leakage warning. When the alkane gas concentration measured by the combustible gas detector in real time exceeds the preset alkane gas concentration over-limit alarm threshold, the combustible gas detector sends a gas leakage warning signal to the leakage accident warning processing center; when the alkane gas concentration measured by the combustible gas detector in real time does not exceed the preset alkane gas concentration over-limit alarm threshold, the combustible gas detector does not send a gas leakage warning signal to the leakage accident warning processing center; Step 2: an alkylation unit on-site video recognition module is set up at the alkylation unit site, and the alkylation unit on-site video recognition module is used to monitor in real time whether gas leakage and liquid leakage occur, and when a gas leakage is detected, a gas leakage warning signal is sent to a leakage accident warning processing center, and when a liquid leakage is detected, a liquid leakage warning signal is sent to the leakage accident warning processing center; Step 3: After installing the on-site alarm of the alkylation unit, the combustible gas detector and the on-site video recognition module of the alkylation unit are used to monitor the situation of the alkylation unit in real time. The leakage accident warning and processing center triggers the alarm to issue a first-level alarm or a second-level alarm based on the leakage warning signal issued by the on-site combustible gas detector and the on-site video recognition module of the alkylation unit, so as to remind the on-site staff of the alkylation unit to handle it in time.

20. The method for rapid identification and alarm of gas-liquid two-phase leakage in an alkylation unit according to claim 19, characterized in that: The triggering conditions of the first level alarm are: The combustible gas detector and the alkylation unit on-site video recognition module simultaneously detect a leak at the alkylation unit site within 1 second, and when the combustible gas detector sends out a gas leak warning signal and the alkylation unit on-site video recognition module simultaneously sends out a gas leak warning signal and a liquid leak warning signal, the alarm triggers a first-level alarm; The triggering conditions for the alarm to trigger the secondary alarm are: The combustible gas detector sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds; Or the on-site video recognition module of the alkylation unit sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds; or the on-site video recognition module of the alkylation unit sends out a liquid leakage warning signal and the liquid leakage warning signal is not interrupted for 5 consecutive seconds; or the combustible gas detector sends out a gas leakage warning signal and at the same time the on-site video recognition module of the alkylation unit sends out a gas leakage warning signal; Or the combustible gas detector sends out a gas leakage warning signal and at the same time the on-site video recognition module of the alkylation unit sends out a liquid leakage warning signal; Or the on-site video recognition module of the alkylation unit simultaneously sends out a gas leakage warning signal and a liquid leakage warning signal.

21. The method for rapid identification and alarm of gas-liquid two-phase leakage in an alkylation unit according to claim 20, characterized in that: The first-level alarm triggering and the second-level alarm triggering conditions are that the combustible gas detector sends out a gas leakage warning signal and the alkylation unit on-site video recognition module sends out a gas leakage warning signal, or the combustible gas detector sends out a gas leakage warning signal and the alkylation unit on-site video recognition module sends out a liquid leakage warning signal, or the alkylation unit on-site video recognition module sends out a gas leakage warning signal and a liquid leakage warning signal at the same time, and the alkylation unit on-site alarm immediately alarms; The secondary alarm triggering condition is that the combustible gas detector sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds, or the on-site video recognition module of the alkylation unit sends out a gas leakage warning signal and the gas leakage warning signal is not interrupted for 5 consecutive seconds, or the on-site video recognition module of the alkylation unit sends out a liquid leakage warning signal and the liquid leakage warning signal is not interrupted for 5 consecutive seconds, and the alarm on the alkylation unit site is delayed to alarm.

Images