A combustible and toxic gas alarm system for large oil depots

Abstract

The present invention relates to the field of gas alarm technology, and specifically to a large-scale oil depot combustible and toxic gas alarm system, comprising: an area layout module, a state acquisition module, a state judgment module and a coverage assessment module; generating a three-dimensional grid according to geographic information of the oil depot, and forming an initial detection area corresponding to the geographic information of the oil depot for any geographic information of the oil depot; detecting the gas concentration in the initial detection area, obtaining the diffusion path of each detection point on the initial detection area with a preset simulation route of the initial detection area, and outputting the gas concentration on the diffusion path as gas stratification data; establishing a concentration-time matrix according to the gas stratification data of each detection point on the initial detection area, and determining the candidate area corresponding to the gas leakage according to the time series correlation of the concentration of each detection point; reconstructing the concentration grid field of the gas leakage based on the effective coverage radius of the candidate area, and setting a leakage alarm strategy; achieving the response efficiency and accuracy of the gas alarm.

Description

Technical Field

[0001] The invention relates to the technical field of gas alarms, in particular to a combustible and toxic gas alarm system for a large oil depot. Background Art

[0002] Combustible gases are substances that can ignite and are in a gaseous state at normal temperature and pressure. When mixed in certain proportions, they can cause combustion or explosion. Toxic gases are toxic chemicals that are gaseous or highly volatile at normal temperature and pressure. In large oil depots, traditional point detectors have limited coverage and are difficult to capture widely diffused gases. This is especially true if there is a buildup of flammable light gases like hydrogen or stagnant heavy gases like hydrogen sulfide, which can lead to corresponding risks in oil depot management.

[0003] For example, Chinese patent publication number CN118247906A discloses a gas alarm system, which belongs to the field of gas safety supervision technology and includes a data acquisition module, an analysis and judgment module, a scheduling and processing module, a secondary judgment module, and an event summary module. The analysis and judgment module includes a leakage judgment unit, a hazard calculation unit, and a danger warning unit. The present invention can obtain monitoring data from all monitoring devices in the gas area, analyze the monitoring data to determine whether there are any anomalies, and further analyze the anomalies to determine the level and cause of the anomaly. Based on the level and cause of the anomaly, it can find the corresponding processing strategy to reasonably control the anomaly.

[0004] For example, Chinese patent publication No. CN114283552A discloses a combustible gas alarm system that drives a wireless solenoid valve. The combustible gas alarm system includes a combustible gas alarm module, a wireless solenoid valve module that can be connected to the combustible gas alarm module through wireless communication technology, and a first temperature sensor module that can also be connected to the combustible gas alarm module through wireless communication technology. The combustible gas alarm module also specifically includes a power supply module, a combustible gas sensor module, a second temperature sensor module, a data processing module, an alarm output module, a wireless communication module, and a key input module.

[0005] The existing technology describes how to generate an alarm based on anomalies identified by gas concentration, and how to correct the corresponding gas alarm by interpreting the temperature of the identified gas. However, the existing technology ignores the diffusion paths of different gases, and the location of the leak source relies on manual experience. It is impossible to quickly associate the leak diffusion path with the emergency strategy, resulting in the inability to accurately identify the leak source when a gas alarm is generated, which reduces the safety of oil depot management. Summary of the Invention

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a large oil depot combustible and toxic gas alarm system, including: an area layout module, used to generate a three-dimensional grid based on the oil depot geographic information, and for any oil depot geographic information, form an initial detection area corresponding to the oil depot geographic information.

[0007] The state acquisition module is used to detect the gas concentration in the initial detection area, obtain the diffusion path of each detection point in the initial detection area using a preset simulation route in the initial detection area, and output the gas concentration on the diffusion path as gas stratification data.

[0008] The state judgment module is used to establish a concentration-time matrix based on the gas stratification data of each detection point on the initial detection area, and determine the candidate area corresponding to the gas leakage based on the time series correlation of the concentration of each detection point.

[0009] The coverage assessment module is used to reconstruct the concentration grid field of gas leakage based on the effective coverage radius of the candidate area, and set the leakage alarm strategy according to the leakage characteristics of each grid in the concentration grid field.

[0010] The beneficial effects of the present invention are as follows: 1. The present invention forms an initial detection area by setting a three-dimensional grid according to different structures in the oil depot, and determines that the three-dimensional grid covers no blind spots according to the detection points associated with each position; at the same time, a preset simulation route is used to identify the gas concentration, and after preliminarily identifying its spatially related diffusion route, the gas concentration identified at each vertical level is output as gas stratification data, and the stratification weight of the diffusion path is used to improve the recognition accuracy of spatial positioning in different gas detection processes.

[0011] 2. The present invention further verifies the gas concentration output by the diffusion path with a time delay, identifies the corresponding time of the gas concentration change at each detection point, and then combines the propagation path related to the time delay with the diffusion path to divide the candidate area of ​​probabilistic leakage. According to the gas concentration identified in the candidate area, the various areas that may appear during gas diffusion are processed to improve the accuracy of leakage source positioning and the spatiotemporal correlation of gas diffusion identification.

[0012] 3. The present invention combines multiple groups of features identified in the diffusion path and propagation path, such as leakage characteristics, candidate area types and gas concentrations, to identify the weight corresponding to the current detection point, and sets the leakage alarm strategy based on the weight of each detection point to improve the current alarm system's response capability to various forms of alarms and improve the applicability of the current system. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The present invention will be further described below with reference to the accompanying drawings and examples.

[0014] Figure 1 This is a system framework diagram of a flammable and toxic gas alarm system for a large oil depot.

[0015] Figure 2 The diagram is a flow chart of the regional layout module of the combustible and toxic gas alarm system of a large oil depot.

[0016] Figure 3 The figure is a flow chart of the status acquisition module of the combustible and toxic gas alarm system of a large oil depot.

[0017] Figure 4 The present invention is a flow chart of a status judgment module of a combustible and toxic gas alarm system in a large oil depot.

[0018] Figure 5 The present invention is a flow chart of the coverage assessment module of the combustible and toxic gas alarm system of a large oil depot. DETAILED DESCRIPTION

[0019] The following embodiments of the present invention are described in detail. The embodiments described below are exemplary and are only used to explain the present invention, and are not to be construed as limiting the present invention. Where specific techniques or conditions are not specified in the embodiments, the techniques or conditions described in the literature in the art or in the product specifications shall be followed.

[0020] See Figure 1 A large oil depot combustible and toxic gas alarm system includes: an area layout module, a state acquisition module, a state judgment module and a coverage assessment module; wherein the output end of the area layout module is connected to the state acquisition module, the output end of the state acquisition module is connected to the state judgment module, and the output end of the state judgment module is connected to the coverage assessment module.

[0021] The area layout module is used to generate a three-dimensional grid based on the geographical information of the oil depot, and for any geographical information of the oil depot, form an initial detection area corresponding to the geographical information of the oil depot.

[0022] The state acquisition module is used to detect the gas concentration in the initial detection area, obtain the diffusion path of each detection point in the initial detection area using a preset simulation route in the initial detection area, and output the gas concentration on the diffusion path as gas stratification data.

[0023] The state judgment module is used to establish a concentration-time matrix based on the gas stratification data of each detection point on the initial detection area, and determine the candidate area corresponding to the gas leakage based on the time series correlation of the concentration of each detection point.

[0024] The coverage assessment module is used to reconstruct the concentration grid field of gas leakage based on the effective coverage radius of the candidate area, and set the leakage alarm strategy according to the leakage characteristics of each grid in the concentration grid field.

[0025] The oil depot's geographic information describes the area where the currently identified data resides. For example, this information can correspond to various locations, such as tank tops, firebreak boundaries, and pipeline flange connections. Devices, such as layered detectors, are then installed at the predicted leak points to collect toxic and flammable gases. This location is then considered the initial detection area.

[0026] like Figure 2 As shown, the implementation method of the regional layout module includes: taking the oil depot geographical information as a constraint condition, dynamically mapping the data in each three-dimensional grid, and generating a multi-location mapping constraint condition matrix after associating them according to the mapping table.

[0027] The oil depot geographic information is traversed using a multi-location mapping constraint matrix to determine the initial detection area corresponding to the oil depot geographic information.

[0028] The three-dimensional grid set at this time will set different grid sizes according to the different contents of the oil depot geographical information. For example, in the oil depot storage, the open area is larger than the The grid is laid out in the form of . If it is in a dense area such as a pipeline intersection or a storage tank group, The grid size is adjusted to include the various devices present therein to identify whether there is a leak at the corresponding location.

[0029] The initial inspection areas that are subsequently divided will be marked according to the content contained in the areas, such as initial inspection areas with different labels such as the perimeter of storage tanks, along pipelines, and loading and unloading areas.

[0030] The above mapping table contains the data corresponding to the geographical information of different oil depots, and then traverses the initial detection area that needs to be detected.

[0031] In one embodiment of the present invention, the preset simulation route is used to represent the gas diffusion path that may appear at the current position when there is gas leakage or flammable and toxic gas emission at some locations. For example, where there is a ventilation hole, the position close to the ventilation hole is mainly identified. For relatively closed spaces, the pipeline direction is arranged according to the direction as its simulation route. For open air, detectors are arranged around to identify possible simulation paths, and finally the multi-route gas distribution is completed.

[0032] At this time, the preset simulation route is based on the identification of the concentration at each position on the initial detection area. If a single point concentration anomaly is identified and the concentration gradient of the adjacent points is abnormal, a diffusion cloud map is generated based on the concentration gradient of the abnormal detection point. The routes in the diffusion cloud map are used for stratification judgment to screen out the gas stratification data on each simulation route.

[0033] The output gas layering data is used to illustrate the vertical positioning of the leakage source. After the gas concentrations collected by the detection points are arranged according to their locations, they are layered according to the relative vertical heights of each point to illustrate the specific problems of possible gas leakage. For example, a high concentration is detected at a height of 0.3m. The sudden drop in concentration identified at 1.5m indicates that the leakage source is near the ground, which may be a slight leak in the pipeline flange; or the abnormal methane concentration detected at 3m may be a leak in the breathing valve on the top of the tank.

[0034] At this time, a layered concentration measurement method is used to output all detection points that are identified as abnormal, and then these contents are used as subsequent detection to identify areas where leakage may occur and the direction of leakage in the area.

[0035] like Figure 3 As shown, the implementation method of the state acquisition module includes: extracting a preset simulation route based on the geographical information of the oil depot corresponding to the initial detection area.

[0036] According to the preset simulation route, the detection points on each initial detection area are judged for concentration anomalies, and the diffusion cloud map corresponding to each detection point is formed based on the concentration anomaly judgment results.

[0037] The paths on the diffusion cloud map are screened, starting from the abnormal detection point along the direction of decreasing concentration, and the paths that coincide with the preset simulation routes are screened out as the diffusion paths of the diffusion cloud map.

[0038] The gas concentration on the diffusion path of the diffusion cloud map is detected in layers, and the gas layered data is output based on the layered weight of the diffusion path after detection.

[0039] Preferably, the preset simulation route will select different paths according to the different representations of the oil depot geographical information. For example, the preset simulation route out of the vent will set a path that simulates the current gas leakage according to the real-time wind speed and wind direction deviation, and use this path as the preset simulation route for detection at this time; when the oil depot geographical information represents a closed area, a preset simulation route will be set according to its gas density and spatial geometric shape deviation. If it represents a pipeline, the pipeline will be used as its preset simulation route; if the oil depot geographical information represents an open-air location, a path that simulates gas leakage will be set according to its terrain elevation and weather as its preset simulation route.

[0040] When determining concentration anomalies at detection points on each initial detection area, single-point detection and adjacent-point gradient anomaly detection are performed on each detection point to identify whether there are abnormal conditions related to gas leakage in the initial detection area.

[0041] That is, the realization method of the concentration abnormality judgment result includes: performing concentration detection on each detection point in the initial detection area along the direction of the preset simulation route, identifying the gas item at each detection point, and the gas item represents the current detection gas, such as detecting benzene, toluene and Toxic gases such as flammable gases produced by diesel leaks are detected according to the different contents of the current detection. Then, according to the concentration of each detection point, the candidate abnormal points are identified first. The candidate abnormal points represent the threshold value exceeding the normal gas concentration. Then, the concentration gradient around the candidate abnormal points is detected to indicate the presence of abnormal gas concentration, and finally the abnormal gas concentration judgment is completed.

[0042] Import the gas items of each detection point and determine whether the gas concentration corresponding to the gas item exceeds the theoretical value. If it exceeds, mark it as a candidate abnormal point.

[0043] With the candidate outlier point as the center, the concentration gradient between it and its adjacent points is calculated. When the concentration gradient value is greater than the gradient threshold, the corresponding candidate outlier point is marked as a leakage feature. The detection point marked as a leakage feature is used as the concentration anomaly judgment result.

[0044] At this time, a basic alarm threshold is set according to the gas type and application scenario. This basic alarm threshold represents the theoretical value of a certain gas that needs to be alarmed, for example The candidate anomaly mark is triggered, and the adjacent concentration detection is performed on the corresponding detection point according to the gas concentration of its adjacent detection points. At the same time, the standard deviation of the candidate anomaly point marked as a continuous time is determined. This time can be set within 10S or other time periods to prevent the currently identified candidate anomaly point from being an invalid signal; then its concentration gradient is verified, and the gradient threshold will select different thresholds according to the location of the current data, such as the gradient in a confined space. ,open air The gradient threshold can be set using the average value of the concentration gradient when the gas leak is identified in the historical data. At the same time, the gas concentration value of a single detection point can also use the average value of the historical data as its theoretical value to identify whether there is a leakage at a single point.

[0045] The identified leakage features will be expressed as leakage features in various descriptive forms, such as strong diffusion features, possible leakage diffusion, etc., according to the part of the leakage features that is greater than the threshold, indicating that the current detection point exceeds the normal value.

[0046] It should be noted that the currently output concentration anomaly determination result will include multiple abnormal detection points. When generating the diffusion cloud map, the detection points in the corresponding areas will be combined into a diffusion cloud map according to these abnormal detection points.

[0047] Preferably, when the diffusion cloud map is formed, the concentration contour lines of multiple detection points output in the concentration anomaly identification are combined with their coordinates to form a diffusion cloud map to illustrate the current diffusion form. Then, the time when the concentration gradient and other numerical values ​​in the diffusion cloud map appear and the concentration gradient value are reversed to identify whether the diffusion path in the diffusion cloud map coincides with the preset simulation route.

[0048] Specifically, the diffusion path of the diffusion cloud map is implemented by generating the shortest path corresponding to the abnormal detection point in the diffusion cloud map along the direction of decreasing concentration. This shortest path represents a method for calculating the shortest path, such as the Dijkstra algorithm, by connecting multiple detection points in the direction of decreasing concentration where the abnormal detection point is located to form a shortest path. This shortest path is determined by the relative distance between each detection point. The distance between detection points can be calculated based on the Euclidean distance.

[0049] The overlap between the shortest path corresponding to the anomaly detection point and the preset simulation route is determined, and when the overlap exceeds the overlap threshold, the corresponding shortest path is regarded as the diffusion path of the current diffusion cloud map.

[0050] At this time, the overlap between the output shortest path and the preset simulation route will be calculated based on the ratio of the intersection of the two routes to their union.

[0051] Preferably, the paths on the diffusion cloud map are screened. In addition to determining the degree of overlap between the paths and the preset simulation routes, it is also necessary to identify the concentration deviation between the concentration distribution on the paths and the preset simulation routes to set the stratification weights and ultimately obtain the output gas stratification data.

[0052] That is, the implementation method of gas stratification data also includes: performing stratified detection at the detection points in the vertical direction of the diffusion path, identifying the concentration gradient difference of each detection point in the vertical direction as the stratification weight of each detection point. At this time, stratified detection is to identify whether there is an obvious difference in the concentration gradient in the vertical direction, and the concentration difference value identified in the vertical direction of each detection point is used as its weight. This weight is set by dividing the concentration difference value by the sum of the concentration difference values ​​identified at all detection points to indicate whether the gas leakage and diffusion at each detection point will cause differences in the vertical direction after the diffusion path is identified.

[0053] It should be noted that the preset simulation route focuses on gas leakage in the horizontal direction, and the stratified weight indicates whether there is diffusion from bottom to top, etc., to prevent inaccurate positioning of the leakage source when identifying it.

[0054] In one embodiment of the present invention, when judging whether there is a problem with the gas concentration in each direction under gas stratification, they are combined in time sequence to identify candidate areas that may cause abnormal gas concentration at the current location, and classified according to the possible leakage causes in the candidate areas to identify each candidate area.

[0055] For example, the values ​​of each gas layer data at each detection point within ±30S are traversed, the time series correlation coefficient is set, the propagation judgment direction of multiple detection point combinations with strong time series correlation is determined, and the candidate area is set based on the position of the propagation judgment direction.

[0056] The time series correlation coefficient obtained at this time is used to determine whether there is a relative time delay when a gas leak occurs, to assist in determining the time series characteristics of the current presence of flammable and toxic gases, etc. It is suitable for identifying instantaneous leaks or dynamic leaks in other situations. As for the diffusion path identified in the gas stratification data, it is used to identify situations where there is a continuous leak or a relatively stable leak state, to capture the location of the leak point in space.

[0057] Preferably, the concentration-time matrix is ​​composed of concentration-time matrices corresponding to multiple detection points based on the values ​​of the detection points contained in the gas stratification data in the corresponding time period, so as to illustrate the temporal correlation of each detection point.

[0058] like Figure 4 As shown, the implementation of the state judgment module includes extracting multiple detection point pairs from the gas stratification data, setting a time series correlation coefficient based on the similarity of the detection point pairs at multiple time points. The time series correlation coefficient is calculated based on the gas concentrations of the two detection points in the detection point pair, and using the Pearson correlation coefficient to determine whether there is a correlation between the gas concentrations at multiple time points. The time series correlation coefficient is then output as the time interval when the absolute value of the time series correlation coefficient reaches its maximum. The output time interval represents the period of time when the absolute value of the similarity calculated using the Pearson correlation coefficient is the maximum. This period can represent the sequence of concentration changes at the two points in the detection point pair and the time delay corresponding to this sequence.

[0059] For example, if the detection point pair obtained is i and j, the time series correlation coefficient is It can be expressed as follows.

[0060] ;in, Indicates different time offsets The time series correlation coefficient of the next detection point i and j, at this time Indicates the number of time points whose time offset is less than the number of currently set time points. ; Indicates the gas concentration value at detection point i at time t; Indicates the mean gas concentration at detection point i, which is expressed as the mean value of the time period corresponding to the currently selected time point; Indicates time The gas concentration value at detection point j at time ; represents the mean gas concentration at detection point j; Indicates the standard deviation of gas concentration at detection point i; It represents the standard deviation of the gas concentration at detection point j. When calculating the correlation between two detection point pairs, the main focus is on determining whether there is an offset within the realization period of the previous and subsequent offsets. If the time series correlation coefficient is greater than 0, it means that the concentration change at point i is earlier than that at point j, and the gas propagates from i to j. Otherwise, the gas propagates from j to i. The essence of this time series correlation coefficient is to find the most matching time offset between the two time series to identify which point's gas concentration changes first and to determine its propagation situation in relative time.

[0061] After connecting each detection point pair according to their time series correlation coefficient, the detection point pairs with the same time interval are combined to obtain the propagation path corresponding to the gas stratification data.

[0062] The connection method at this time is to first connect the two detection points in the detection point pair according to the cases where their time series correlation coefficient is greater than 0 and less than 0, and then connect the detection points in the area where the current gas stratification data is located according to the part with the same time point interval. Then, a rough propagation path is obtained. This propagation path will include its direction after connection to explain the path along which the gas leakage propagates in the scenario of time offset.

[0063] The propagation path is then combined with the existing diffusion path. By integrating the propagation path and the existing diffusion path, the data loss caused by time delay in the gas space diffusion process is compensated, thereby constructing a candidate area with both scene adaptability and path spatiotemporal correlation.

[0064] Taking the detection point on the propagation path as the starting point, the propagation path and the diffusion path are connected to obtain the emphasis directions of the propagation path and the diffusion path, and the candidate areas are divided according to the combined emphasis directions.

[0065] At this time, the detection points on the propagation path and the diffusion path are integrated, and the leakage source can be accurately located by using directional consistency and other content, thereby improving the response efficiency and accuracy of the oil depot gas alarm system.

[0066] The implementation method of connecting the propagation path and the diffusion path includes: mapping each detection point to the propagation path and the diffusion path, forming a propagation vector and a diffusion vector in sequence; when forming the diffusion vector, the detection point is mapped to the diffusion path and the nearest diffusion vector is found; for example, let the detection point i find the nearest diffusion vector on the diffusion path Expressed as, ;in, 、 Respectively represent the coordinates of the detection point i, which can represent the position of the detection point relative to the plane map; the detection point selected at this time needs to be a point that can be mapped to the propagation path and the diffusion path. For the propagation vector, it is represented by finding the adjacent point j of the detection point i to form the propagation vector of the detection point i and the detection point j. , ;in, 、 Respectively represent the coordinates of detection point j; if the detection point j is propagated to the detection point i at this time, the calculated value in the corresponding vector will be negative.

[0067] The propagation vector and the diffusion vector are weighted and combined into the focus direction, and the candidate regions are divided according to the directional discreteness of the focus direction.

[0068] At this time, the two vectors are weighted and summed to obtain the vector value represented by the emphasis direction. At this time, the weight will be biased towards the diffusion situation in different spaces. The weight of the diffusion vector is set to 0.7, and the weight of the propagation direction is set to 0.3.

[0069] That is, focus on direction Expressed as, ;in, represents the weight of the propagation vector, It represents the weight of the diffusion vector. When calculating the emphasis direction, the input propagation vector and diffusion vector will be normalized to prevent the problem of excessively large values.

[0070] The directional dispersion is then calculated using the circular variance calculation method to verify the uniformity of these focused directions and measure the degree of directional dispersion of the detection points where gas diffusion currently exists.

[0071] Directional dispersion Expressed as, ;in, Indicates the number of vectors corresponding to the emphasis direction, where k ranges from 1 to N; Indicates the angle corresponding to the focus direction. This angle is based on the angle of the vector corresponding to the focus direction in the current plane. If the focus directions are highly consistent, the calculated direction dispersion is small; otherwise, it is too large.

[0072] For example, when the directional dispersion is less than 0.2, it indicates a high degree of consistency in the focus direction. The area encompassing the corresponding focus direction is designated as a candidate region. This candidate region represents a core leak candidate, indicating the path to the primary leak source. When the directional dispersion is between 0.2 and 0.5, it indicates a potential diffusion edge or interference region. The direction of this region exhibits a certain degree of dispersion, requiring verification based on gas concentration to determine if it represents a diffusion edge. The region represented by this value is designated as a candidate region. When the directional dispersion is greater than 0.5, it indicates a significant amount of noise, with significant variations in gas diffusion direction and concentration. This region is considered a noise region. During normal diffusion identification, this region has a low correlation with the leak source location and requires direct processing, such as a correlation alarm or gas concentration identification. The corresponding region is designated as a candidate region. Furthermore, when candidate regions are divided based on directional dispersion, each region is labeled based on its directional dispersion to indicate its candidate type, such as a core leak candidate, a diffusion edge candidate, or a noise region candidate.

[0073] In one embodiment of the present invention, the effective coverage radius is used to describe the range that can be inferred based on the concentration of the leakage when a corresponding leakage point exists in the candidate area, and the data obtained from its short-time instantaneous measurement and long-time cumulative measurement are used as its leakage characteristics to describe how to adopt the corresponding linkage alarm method after the leakage characteristics are identified.

[0074] The leakage characteristics include short-term high-concentration pulse signals such as valve rupture, which trigger emergency interlocks, close valves, start spraying and other measures to sound an alarm; they may also include low-concentration continuous accumulation time such as flange micro-leakage, which triggers operation and maintenance work orders and marks high-frequency leakage points. After these high-frequency leakage points are marked, alarm processing is carried out to timely control leaks at multiple locations and prevent large-scale leakage incidents.

[0075] like Figure 5 As shown, the implementation method of the coverage assessment module includes: based on the effective coverage radius of the candidate area, each grid in the three-dimensional grid is recombined according to the detection points covered by the candidate area, the leakage characteristics, candidate area type and gas concentration in each candidate area are extracted, and each grid in the concentration grid field represents any set of leakage characteristics, candidate area type and gas concentration to form a concentration grid field.

[0076] The above leakage features are obtained by extracting the leakage features marked at each detection point during gas diffusion, and then describing the gas concentration value detected at the detection point. When dividing the candidate areas, the label content corresponding to the candidate areas is introduced; these data are then used as a set of data pairs to describe the relevant situation of each detection point in the concentration grid field.

[0077] Calculate the weight of each grid in the concentration grid field at any time and obtain the average weight of each grid in the concentration grid field.

[0078] A strategy search is performed based on the average weight and maximum weight of each grid in the concentration grid field, and a leakage alarm strategy is set for each grid in the concentration grid field.

[0079] The weight of each grid in the concentration grid field at any given moment is calculated by setting scores for the leakage feature and candidate area type, then multiplying the ratio of the gas concentration at the current detection point to the gas concentration of all detection points in the corresponding concentration grid field by the leak feature and candidate area type scores. The sum of these sums is the weight of each grid. The average of these weights at multiple moments is the average weight of each grid. The score for the leakage feature is set based on whether there is strong diffusion and, if so, the form of diffusion. This score can be set based on industry experience or the ratio of the frequency of the same leakage feature at the detection point to the total frequency of leakage features. The score for the candidate area type is set in the same way as the leak feature.

[0080] When performing strategy retrieval later, the average weight and maximum weight in each grid are used to represent the instantaneous and cumulative amount of a single grid, and the similarity between these values ​​and the preset strategies in the database is used for extraction. For example, the part with the greatest similarity after retrieval is used as the leakage alarm strategy adopted by the corresponding grid at this time by calculation methods such as the Pearson correlation coefficient.

[0081] Preferably, when identifying its effective coverage radius, the area can be represented by the minimum inscribed circle corresponding to its candidate area. When the area is an irregular type such as a long strip, the extreme points in the candidate area can also be used to set the minimum circumscribed circle containing the maximum value and the smaller value, and divide the current candidate area into multiple areas included in the range to express the coverage part of the current candidate area that is within the effective coverage radius.

[0082] Preferably, the implementation method of setting the leakage alarm strategy further includes: receiving the leakage alarm strategy of each grid in the concentration grid field, and executing the leakage alarm strategy in sequence according to the maximum weight of each grid in the concentration grid field.

[0083] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention, which are still covered by the scope of protection of the present invention.

Claims

1. A large oil depot combustible and toxic gas alarm system, characterized by: include: The area layout module is used to generate a three-dimensional grid based on the geographical information of the oil depot, and for any geographical information of the oil depot, form an initial detection area corresponding to the geographical information of the oil depot; The state acquisition module is used to detect the gas concentration in the initial detection area, obtain the diffusion path of each detection point in the initial detection area using a preset simulation route in the initial detection area, and output the gas concentration on the diffusion path as gas stratification data; The state judgment module is used to establish a concentration-time matrix based on the gas stratification data of each detection point in the initial detection area, and determine the candidate area corresponding to the gas leakage based on the time series correlation of the concentration of each detection point; The implementation methods of the status judgment module include: Extract multiple groups of detection point pairs from the gas stratification data to determine the similarity of the detection point pairs at multiple time points and set the time series correlation coefficient; After connecting each detection point pair according to their time series correlation coefficient, the detection point pairs with the same time interval are combined to obtain the propagation path corresponding to the gas stratification data; Taking the detection point on the propagation path as the starting point, the propagation path and the diffusion path are connected to obtain the focus directions of the propagation path and the diffusion path, and the candidate areas are divided according to the combined focus directions; Ways to connect the propagation path with the diffusion path include: Map each detection point to the propagation path and the diffusion path, and form the propagation vector and the diffusion vector in turn; The propagation vector and the diffusion vector are weighted and combined into a focus direction, and the candidate area is divided according to the directional discreteness of the focus direction; The coverage assessment module is used to reconstruct the concentration grid field of gas leakage based on the effective coverage radius of the candidate area, and set the leakage alarm strategy according to the leakage characteristics of each grid in the concentration grid field.

2. A large oil depot combustible and toxic gas alarm system according to claim 1, characterized in that: The implementation methods of the regional layout module include: Taking the oil depot geographic information as a constraint, the data in each three-dimensional grid is dynamically mapped and associated according to the mapping table to generate a multi-location mapping constraint matrix; The oil depot geographic information is traversed using a multi-location mapping constraint matrix to determine the initial detection area corresponding to the oil depot geographic information.

3. A large oil depot combustible and toxic gas alarm system according to claim 1, characterized in that: The implementation methods of the status acquisition module include: Extract the preset simulation route based on the geographical information of the oil depot corresponding to the initial detection area; According to the preset simulation route, the detection points on each initial detection area are judged for concentration anomalies, and the diffusion cloud corresponding to each detection point is formed according to the concentration anomaly judgment results; Screen the paths on the diffusion cloud map, starting from the abnormal detection point along the direction of decreasing concentration, and select the path that coincides with the preset simulation route as the diffusion path of the diffusion cloud map; The gas concentration on the diffusion path of the diffusion cloud map is detected in layers, and the gas layered data is output based on the layered weight of the diffusion path after detection.

4. A large oil depot combustible and toxic gas alarm system according to claim 3, characterized in that: The methods for realizing abnormal concentration determination results include: Conduct concentration tests on each detection point in the initial detection area in sequence along the direction of the preset simulation route, and identify the gas item at each detection point; Import the gas items at each detection point and determine whether the gas concentration corresponding to the gas item exceeds the theoretical value. If it exceeds, mark it as a candidate abnormal point; Taking the candidate outlier point as the center, the concentration gradient between it and the adjacent points is calculated. When the concentration gradient value is greater than the gradient threshold, the corresponding candidate outlier point is marked as a leakage feature; the detection point marked as a leakage feature is used as the concentration anomaly judgment result.

5. A large oil depot combustible and toxic gas alarm system according to claim 3, characterized in that: The implementation of the diffusion path of the diffusion cloud map also includes: Starting from the abnormal detection point in the diffusion cloud map, the shortest path corresponding to the abnormal detection point is generated along the direction of decreasing concentration; The overlap between the shortest path corresponding to the anomaly detection point and the preset simulation route is determined, and when the overlap exceeds the overlap threshold, the corresponding shortest path is regarded as the diffusion path of the current diffusion cloud map.

6. A large oil depot combustible and toxic gas alarm system according to claim 3, characterized in that: Gas stratification data can also be implemented by: The detection points along the vertical direction of the diffusion path are used for stratified detection, and the concentration gradient difference of each detection point in the vertical direction is identified as the stratified weight of each detection point.

7. A large oil depot combustible and toxic gas alarm system according to claim 1, characterized in that: The implementation of the coverage assessment module includes: Based on the effective coverage radius of the candidate area, each grid in the three-dimensional grid is reassembled according to the detection points covered by the candidate area, and the leakage characteristics, candidate area type and gas concentration in each candidate area are extracted. Each grid is made to represent any set of leakage characteristics, candidate area type and gas concentration to form a concentration grid field; Calculate the weight of each grid in the concentration grid field at any time and obtain the average weight of each grid in the concentration grid field; A strategy search is performed based on the average weight and maximum weight of each grid in the concentration grid field, and a leakage alarm strategy is set for each grid in the concentration grid field.

8. A large oil depot combustible and toxic gas alarm system according to claim 7, characterized in that: The implementation of setting leakage alarm strategy also includes: The leakage alarm strategy of each grid in the concentration grid field is received, and the leakage alarm strategy is executed sequentially according to the maximum weight of each grid in the concentration grid field.

Images