Intelligent inspection and fire emergency linkage system for hazardous chemical storage

Abstract

This invention relates to the field of fire alarm technology, specifically to an intelligent inspection and fire emergency linkage system for hazardous chemical storage. The system includes: an inspection timing module, an airflow difference module, a diffusion determination module, a periodic locking module, and a linkage execution module. In this invention, inspection data is filtered and frame sequence integrity is verified according to a determination period to construct a concentration sequence with continuous constraints, reducing interference from abnormal data. The headwind and headwind concentration differences are calculated by combining detector coordinates with the main airflow direction, and the difference matrix is ​​rearranged using the vertical distance to the channel centerline. Gradient transitions are identified through change thresholds to extract abnormal points closer to the diffusion source. Consistency and displacement comparisons are performed on continuous periodic anomalies to retain stable diffusion records and suppress short-term disturbances. The system is graded based on the maximum difference and the linkage loop is encoded for verification and feedback, forming a closed loop of identification, location, and execution, improving early detection capability, location accuracy, and response reliability.

Description

Technical Field

[0001] The present invention relates to the technical field of fire alarm, and particularly to an intelligent inspection and fire-fighting emergency linkage system for hazardous chemical storage. Background Art

[0002] The technical field of fire alarm refers to a technical system that monitors the physical quantity changes in the precursor and initial stage of a fire and issues an alarm. It mainly includes core matters such as smoke concentration detection, temperature change detection, combustible gas concentration detection, flame spectrum identification, alarm signal transmission, and linkage control with sprinkler systems, smoke exhaust fans, fireproof rolling shutters, and audible and visual alarms. The method is to arrange smoke detectors, temperature detectors, combustible gas detectors, and flame detectors in the monitoring area to collect electrical signals. After the controller compares the thresholds, it triggers the relay to output a signal, and transmits the alarm information to the fire control host through a bus or wired communication line. Then, the host sends a start command to the relevant execution devices to complete the alarm and linkage control. Among them, the traditional intelligent inspection and fire-fighting emergency linkage system for hazardous chemical storage refers to a system applied to the scenario of warehouses storing inflammable, explosive, and toxic chemicals, which periodically checks the storage environment status and starts fire-fighting equipment when a fire or leakage occurs. For the problems of early identification of abnormal temperature, gas leakage, and fire in hazardous chemical warehouses and synchronous response of fire-fighting equipment, the traditional method is to manually inspect and record the temperature and humidity values at regular intervals, and combine the fixedly installed smoke detectors, temperature detectors, and combustible gas concentration detectors for on-site monitoring. When the detected value exceeds the set threshold, the fire control cabinet sends start signals to the solenoid valve, sprinkler pump, smoke exhaust fan, and audible and visual alarm, and uploads the alarm information to the duty room terminal through a wired network for manual confirmation and disposal.

[0003] Current methods primarily rely on manual, timed inspections to record temperature and humidity values, supplemented by monitoring with fixed detectors. This data exhibits discrete sampling characteristics, meaning that momentary leaks and short-term diffusions within inspection intervals may occur during recording gaps, potentially missing the optimal response window due to delayed information acquisition by the monitoring station. Alarm criteria largely depend on single-point threshold triggers, emphasizing absolute concentration values ​​while neglecting directional gradients and spatial correlations. Local disturbances such as airflow deflection during ventilation, open doors and windows, or vehicle entry and exit can cause short-term peaks at detection points, triggering false alarms. Conversely, dilution during diffusion can lead to prolonged periods without exceeding thresholds, resulting in missed alarms. For example, a leak located downstream and spreading along a channel may cause gradual increases in temperature at multiple points, but none of the individual points reach the threshold, delaying response initiation. Furthermore, the monitoring link lacks [further details needed]. Insufficient explicit verification of the integrity of reported sequences means that communication jitter or detector anomalies may be mistaken for real environmental changes, leading to unnecessary coordinated actions, downtime, and wasted resources. Anomaly localization often remains near a detector or a specific area, lacking quantitative clues about the source and direction of diffusion. On-site investigation relies on experience and manpower, and expanding the search scope increases response time. Coordinated execution is mostly based on unidirectional start signals from the control cabinet. When the feedback verification mechanism is insufficient, issues such as stuck solenoid valves, circuit failures, and actuator disconnections are difficult to detect in a timely manner. There may be risky scenarios where alarms are triggered but critical equipment does not actually operate, resulting in initial suppression failures, insufficient basis for personnel evacuation and isolation decisions, and ultimately amplifying the scope and scale of the accident's impact. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and propose an intelligent inspection and fire emergency linkage system for hazardous chemical storage.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a hazardous chemical storage intelligent inspection and fire emergency linkage system includes:

[0006] The inspection sequence module obtains inspection data from the combustible gas detectors in the hazardous chemical warehouse, reads combustible gas concentration data and time stamps, filters data with time stamps within a set period, sorts them by frame number, marks skipped data, and organizes them into an inspection cycle concentration dataset.

[0007] Based on the inspection cycle concentration dataset, the airflow difference module calculates the angle between the detector and the airflow direction, combined with the detector coordinates and the angle of the main airflow direction in the tank area. It then calculates the concentration difference between the headwind and the headwind according to the size of the angle, and obtains the concentration difference table of the storage area direction.

[0008] The diffusion determination module establishes a difference matrix based on the directional concentration difference table of the storage area, calculates the vertical distance between the detector and the center line in combination with the coordinates of the main channel center line, and reorders the matrix difference entries in the same row of the difference matrix according to the vertical distance between the detector and the center line of the main channel in the corresponding column from small to large, to obtain the rearranged directional difference sequence and vertical distance sequence. It calculates the difference change of adjacent difference entries after rearrangement, filters records with change exceeding the predetermined change threshold, extracts the maximum difference and the corresponding tank area coordinates, and generates a list of abnormal gas diffusion points in the storage area.

[0009] The periodic locking module compares the list of abnormal gas diffusion points in the storage area, filters records that are consistent within the period, calculates the displacement of points in adjacent periods, filters abnormal records whose displacement is less than a preset displacement threshold, and generates a continuous diffusion abnormal record table.

[0010] Based on the continuous diffusion anomaly record table, the linkage execution module determines the alarm level according to the maximum concentration difference, looks up the corresponding fire alarm controller output circuit address code in the fire control room, updates the relay status, and reads feedback information to obtain fire linkage execution status feedback.

[0011] As a further embodiment of the present invention, the inspection cycle concentration dataset includes average concentration values, peak concentration points, and data integrity identifiers; the storage area directional concentration difference table includes a set of downwind difference values, a set of upwind difference values, and a directional classification identifier; the storage area gas diffusion anomaly point list includes anomaly point coordinates, maximum difference values, and anomaly level identifiers; the continuous diffusion anomaly record table includes continuous anomaly point numbers, periodic displacement parameters, and stability discrimination identifiers; and the fire-fighting linkage execution status feedback includes alarm level identifiers, loop action status, and feedback confirmation signals.

[0012] As a further aspect of the present invention, the change threshold is the upper limit of the threshold range determined based on the statistical characteristic value of the concentration difference between the windward and the headwind directions in the reservoir direction concentration difference table.

[0013] The records whose changes exceed a predetermined threshold are retained after sorting, where the difference between adjacent positions continuously exceeds the threshold.

[0014] As a further aspect of the present invention, the inspection timing module includes:

[0015] The data frame acquisition submodule obtains the data reported during the inspection from the combustible gas detectors in the hazardous chemical warehouse, reads the combustible gas concentration data and time stamps in the data frames, verifies the correspondence between the time stamps and the frame sequence number, removes abnormal marker entries, and obtains the concentration-time mapping.

[0016] The periodic screening submodule, based on the concentration-time mapping, filters entries whose time markers are within the determination period interval, determines the continuous state of the time series, reassembles the corresponding concentration series, and generates the number of sequences within the period.

[0017] The frame sequence verification submodule calls the sequence quantity within the period, arranges them in order according to the frame sequence number, and judges whether there is an offset state in the difference between adjacent frame sequence numbers. It marks the detector data with skipped number characteristics and collects the corresponding concentration entries to establish the inspection period concentration dataset.

[0018] As a further aspect of the present invention, the airflow difference module includes:

[0019] Based on the inspection cycle concentration dataset, the orientation parameter submodule collects the coordinates of the tank area detectors and the angle of the main airflow direction, unifies the coordinate reference, verifies the correspondence between the coordinate identifiers and the detector numbers, and generates coordinate direction mapping quantities.

[0020] The included angle calculation submodule constructs the detector position vector and the main airflow direction vector based on the coordinate direction mapping amount, calculates the vector projection relationship, determines the projection sign, divides the forward and reverse regions, and obtains the included angle sign sequence value.

[0021] The forward and reverse difference submodule calls the included angle symbol sequence value and associates it with the inspection cycle concentration dataset. It calculates the corresponding detector concentration difference according to the forward and reverse zone identifiers and performs intra-zone aggregation processing to establish a storage area directional concentration difference table.

[0022] As a further aspect of the present invention, the diffusion determination module includes:

[0023] The difference table building submodule, based on the storage area directional concentration difference table, collects the detector index and directional difference fields in the table and verifies the index alignment relationship, calculates the index cross-pairing difference and writes it into the matrix unit to generate the difference matrix quantity;

[0024] The vertical distance rearrangement submodule, based on the difference matrix, collects the coordinates of the main channel centerline and calculates the vertical distance between the detector coordinates and the centerline. It then reorders the matrix difference entries in the same row of the difference matrix according to the vertical distance between the detector and the main channel centerline in the corresponding column from smallest to largest, obtaining the rearranged direction difference sequence and vertical distance sequence. It also calculates the difference change of adjacent difference entries after rearrangement, compares the difference change with the change threshold, and marks records that exceed the threshold, thus establishing an over-threshold record identifier.

[0025] The anomaly list submodule calls the threshold record identifier and associates it with the storage area directional concentration difference table, filters the corresponding records, extracts the item with the largest directional difference and the corresponding storage tank area coordinates, and generates a list of abnormal gas diffusion points in the storage area.

[0026] As a further aspect of the present invention, the specific calculation formula for reordering the corresponding detectors according to their vertical distances from the center line of the main channel from smallest to largest, obtaining the rearranged direction difference sequence and vertical distance sequence, and calculating the change in the difference between adjacent difference entries after rearrangement is as follows:

[0027] ;

[0028] The change in difference is calculated, and the change in difference is compared with the change threshold. Records exceeding the threshold are marked, and an over-threshold record identifier is established.

[0029] in, Represents detector number The corresponding change in the difference is, Represents detector number In the vertical distance sorting sequence The difference in direction, Represents detector number In the vertical distance sorting sequence Vertical distance, Represents the difference in direction. Represents vertical distance. and Representing the detector number respectively The maximum and minimum terms of the corresponding direction difference sequence and Representing the detector number respectively The maximum and minimum terms of the corresponding vertical distance sequence, Represents detector number Number of items participating in the sorting Represents the numbering of the detector From item 2 to item 3 Items accumulated, Represents absolute value operation. This represents the square root operation.

[0030] As a further aspect of the present invention, the periodic locking module includes:

[0031] The periodic comparison submodule compares the list of abnormal gas diffusion points in the storage area for three consecutive inspection cycles, extracts the periodic point number and coordinate field, performs cross-matching of the same number, determines the consistency of the number within the three cycles and retains the consistent record, and obtains the number of consistent points in the three cycles.

[0032] The displacement calculation submodule collects the coordinates of corresponding points in adjacent periods based on the three-period consistent point quantities and calculates the planar displacement between adjacent periods. It forms a displacement sequence according to the period order and calculates the displacement value to generate the periodic displacement.

[0033] The locking and filtering submodule calls the periodic displacement amount and associates it with the three-period consistent point amount. It compares the displacement amount with the displacement threshold and retains the records that are less than the displacement threshold. It summarizes the corresponding number and coordinate field to generate a continuous diffusion anomaly record table.

[0034] As a further aspect of the present invention, the linkage execution module includes:

[0035] The difference level determination submodule collects and records concentration differences based on the continuous diffusion anomaly record table and extracts the item with the largest concentration difference. It then obtains the alarm level value by mapping the level threshold interval.

[0036] The loop addressing submodule collects the list of loop address codes output by the fire alarm controller and performs address matching and filtering based on the alarm level value to obtain the loop address code value;

[0037] The feedback acquisition submodule calls the circuit address code value, acquires the relay status and writes the status update instruction, acquires feedback information and performs a consistency judgment on the feedback code and status code, and obtains the fire linkage execution status feedback.

[0038] As a further aspect of the present invention, the preset displacement threshold is a displacement limit determined based on the historical statistical distribution of displacements of adjacent periodic points. The abnormal records with displacements less than the preset displacement threshold are selected by calculating the displacements of adjacent periodic points based on the difference in planar coordinates of the tank area coordinate points, and retaining abnormal records that meet the preset displacement threshold for multiple consecutive inspection cycles.

[0039] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0040] In this invention, by filtering and verifying the integrity of the frame sequence of inspection data according to the judgment cycle, a concentration sequence with continuous constraints is constructed to reduce the interference of abnormal data. The difference between the headwind and tailwind concentrations is calculated by combining the detector coordinates and the direction of the main airflow, and the difference matrix is ​​rearranged by introducing the vertical distance to the center line of the channel. Gradient transitions are identified by the change threshold, and abnormal points closer to the diffusion source are extracted. Consistency and displacement comparisons are performed on continuous periodic anomalies to retain stable diffusion records and suppress short-term disturbances. The maximum difference is graded and linked to the loop coding verification feedback to form a closed loop of identification, positioning and execution, thereby improving the early detection capability, positioning accuracy and response reliability. Attached Figure Description

[0041] Figure 1 This is a system flowchart of the present invention;

[0042] Figure 2 This is a flowchart of the inspection timing module in this invention;

[0043] Figure 3 This is a flowchart of the airflow difference module in this invention;

[0044] Figure 4 This is a flowchart of the diffusion determination module in this invention;

[0045] Figure 5 This is a flowchart of the periodic locking module in this invention;

[0046] Figure 6 This is a flowchart of the linkage execution module in this invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0048] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0049] Please see Figure 1 A smart inspection and fire emergency response system for hazardous chemical storage includes:

[0050] The inspection sequence module obtains the data reported during the inspection period from the combustible gas detectors in the hazardous chemical warehouse, reads the combustible gas concentration data and time stamps in the data frames, filters the data whose time stamps are within the set judgment period, sorts them by frame number, marks the detector data with skipped numbers, and organizes them into an inspection cycle concentration dataset.

[0051] The airflow difference module is based on the concentration dataset of the inspection cycle. It combines the detector coordinates of the tank area and the angle of the main airflow direction to calculate the angle between the detector position and the airflow direction. Based on the size of the angle, it calculates the concentration difference between the downwind and upwind directions respectively, and obtains the concentration difference table of the storage area direction.

[0052] The diffusion determination module establishes a difference matrix based on the directional concentration difference table of the storage area, and calculates the vertical distance between the detector and the center line in combination with the coordinates of the main channel center line. The difference entries in the same row of the difference matrix are reordered according to the vertical distance between the detector and the center line of the main channel in the corresponding column from smallest to largest, so as to obtain the rearranged directional difference sequence and vertical distance sequence. The change in the difference between adjacent difference entries after the rearrangement is calculated and compared with the predetermined change threshold. Records with changes exceeding the threshold are filtered out, the maximum difference and the corresponding storage tank area coordinates are extracted, and a list of abnormal gas diffusion points in the storage area is generated.

[0053] The cycle lock control module compares the list of abnormal gas diffusion points in the storage area for three consecutive inspection cycles, filters the records that remain consistent within the cycle, calculates the displacement of points in adjacent cycles, compares it with the preset displacement threshold, retains abnormal records with displacement less than the displacement threshold, and generates a continuous diffusion abnormal record table.

[0054] The linkage execution module determines the alarm level based on the continuous diffusion anomaly record table and the maximum concentration difference. It then looks up the corresponding fire alarm controller output circuit address code in the fire control room, updates the relay status, and reads feedback information to obtain the fire linkage execution status feedback.

[0055] The inspection cycle concentration dataset includes average concentration values, peak concentration points, and data integrity indicators; the storage area directional concentration difference table includes sets of downwind and upwind differences, and directional classification indicators; the storage area gas diffusion anomaly point list includes anomaly point coordinates, maximum difference values, and anomaly level indicators; the continuous diffusion anomaly record table includes continuous anomaly point numbers, periodic displacement parameters, and stability judgment indicators; the fire linkage execution status feedback includes alarm level indicators, loop action status, and feedback confirmation signals.

[0056] Please see Figure 2 The inspection sequence module includes:

[0057] The data frame acquisition submodule obtains the data reported during the inspection from the combustible gas detectors in the hazardous chemical warehouse, reads the combustible gas concentration data and time stamps in the data frames, verifies the correspondence between the time stamps and the frame sequence number, removes abnormal marker entries, and obtains the concentration-time mapping.

[0058] The standard current loop output of the fixed combustible gas detector in the hazardous chemical warehouse is connected to the acquisition channel. The acquisition channel samples the 4-20mA current value of each channel at a fixed sampling interval of 2 seconds and simultaneously reads the frame number and timestamp field reported by the detector. First, the current value is mapped to 0-100% LEL concentration value according to the range and the original sampling millisecond timestamp is retained. The range mapping adopts linear conversion, that is, 4mA corresponds to 0% LEL, 20mA corresponds to 100% LEL, and for current samples exceeding 3.8-20.5mA, it is directly converted. The following is marked as an out-of-bounds anomaly: A double-key verification is established for the frame sequence number and time stamp, forming an increasing sequence with the frame sequence number as the primary key and the time stamp as the primary key. Each record from the same detector is compared to check if the time difference and frame difference between two adjacent records satisfy the sampling interval constraint. The constraint is fixed as follows: if the adjacent time difference falls within the 1.6-2.4s interval and the adjacent frame difference is 1, it is considered a consistent record; if the time difference falls within the 1.6-2.4s interval but the frame difference is not 1, it is considered a candidate for frame skipping; if the frame difference is 1 but the time difference exceeds 1.6-2.4s... If a record is deemed to be a candidate for time-stamp drift, and if both conditions are not met, it is deemed a strong anomaly. Time-stamp drift candidates are backfilled using the timestamp of the most recent consistent record as a baseline, multiplied by 2 seconds by the frame difference, and the backfilled value is written to the quality identifier field. Strong anomaly records are directly removed, and their original frame number, original timestamp, and original current value are recorded before removal for subsequent quality statistics. To ensure reproducibility of subsequent periodic screening, the judgment period interval is determined by the start and end times in the inspection plan and extended by 10 seconds at both ends. Records within the buffer window are only used for continuity judgment and are not included in the periodic sequence. For each detector, a concentration-time mapping is generated within one inspection period, with the structured fields fixed as detector number, frame number, corrected timestamp, concentration value, quality identifier, and original current value. For each detector, the proportion of the longest consecutive segment with a frame difference of 1 is calculated, and 0.9 is used as the consistency lower limit. If the value is below 0.9, the entire data for that detector in this period is marked as low confidence and retained for missing rate statistics but not included in subsequent difference calculations.

[0059] Table 1. Example of data frame cleaning during inspection cycle:

[0060]

[0061] As shown in Table 1, the frame difference between frames 1204 and 1202 is 2, and the time difference is 4.0 seconds. Therefore, it is identified as a candidate for frame skipping and the time point 1203 is filled back. The current value of 21.0mA exceeds the upper limit of 20.5mA and is therefore rejected. The proportion of the longest consecutive segment with a frame difference of 1 is 0.92, which is higher than the consistency threshold of 0.90. Therefore, the data for this period meets the requirements and enters the subsequent processing.

[0062] The periodic screening submodule filters entries whose time markers are within the determination period range based on the concentration-time mapping, determines the continuous state of the time series and reassembles the corresponding concentration series to generate the number of sequences within the period.

[0063] Based on the concentration-time mapping, closed-interval filtering is performed on entries whose time markers fall within the judgment period range, with both start and end times participating. After filtering, the concentration sequence is recombined in ascending order of time markers. When two records with the same time marker are found, the record with the original valid quality identifier is retained first, and the remaining duplicate entries are deleted. Quantifiable discrimination is used for the continuity of the time series. The time difference between adjacent records is calculated to see if it falls within the 1.6-2.4s range, and the length of continuous abnormal segments is counted. If the length of continuous abnormal segments exceeds 3, the data of this detector in this period is marked as discontinuous and removed. For data that passes the continuity discrimination, the sequence quantity within the period is output. The sequence quantity within the period includes at least the detector number, the time series within the period, the concentration series within the period, the continuity ratio, the number of removed entries, and the number of backfilled entries, so that subsequent frame sequence verification can directly reuse the cleaning results of the same batch and maintain the same time axis caliber as the inspection period.

[0064] The frame sequence verification submodule calls the sequence quantity within the cycle, arranges them in order according to the frame sequence number, and judges whether there is an offset state in the difference between adjacent frame sequence numbers. It marks the detector data with skip number characteristics and collects the corresponding concentration entries to establish the inspection cycle concentration dataset.

[0065] The sequence quantities within the call period are arranged sequentially with the frame number as the primary key. For each detector, the difference between adjacent frame numbers is calculated pairwise. A difference of 1 indicates normal continuity, while a difference greater than 1 indicates offset and is marked as having skipped number characteristics. The first record of the skipped number segment is written into the skipped number characteristic list. At the same time, the corresponding concentration entries within the skipped number segment are collected and the skipped number starting frame and skipped number span fields are added. For detectors with more than 3 skipped numbers in a single period, a high-risk label is written, and the number of skipped numbers and the maximum span are recorded in the quality statistics field. Normal frame entries and skipped number collection entries are uniformly written into the inspection period concentration dataset. The inspection period concentration dataset is grouped by detector number, and each group outputs the concentration sequence, time sequence, frame sequence, and quality statistics field within the period. The quality statistics field includes at least the number of removed entries, the number of backfilled entries, the proportion of the main continuous segment, and the number of skipped numbers, thereby ensuring that subsequent directional difference calculations can trace back the source, cleaning method, and frame continuity of each concentration within the same period.

[0066] Please see Figure 3 The airflow difference module includes:

[0067] The orientation parameter submodule is based on the inspection cycle concentration dataset. It collects the coordinates of the detectors in the tank area and the angle of the main airflow direction, unifies the coordinate reference, verifies the correspondence between the coordinate identifier and the detector number, and generates the coordinate direction mapping quantity.

[0068] Based on the concentration dataset from the inspection cycle, the set of detector numbers participating in the calculation for this cycle is read. The planar coordinates and installation height of each detector are collected from the tank area deployment log and uniformly converted to the same plant area coordinate benchmark. The coordinate unit is fixed in meters and three decimal places are retained. The main airflow direction angle is read from the output of the on-site meteorological instrument or wind vane and aligned to the same time axis at 2-second intervals consistent with the inspection sampling. Two rounds of verification are performed on the correspondence between coordinate identifiers and detector numbers. The first round is performed by matching the detector number in the logbook with the installation point number and the loop address in the controller loop address table with the detector number. Consistency matching is performed, and records that fail to match are directly removed. In the second round, a self-check is triggered for successfully matched detectors, and the reported number is verified to be consistent with the ledger number before being included in the calculation. Noise reduction is performed on the angle of the main airflow direction. The median angle is calculated using 10 sliding windows, and outliers deviating from the median by more than 25° within the window are replaced. The outlier threshold of 25° is calibrated and fixed by the 95th percentile value of wind direction fluctuation statistics for 7 consecutive days under leak-free conditions. The output coordinate direction mapping is fixed with the following fields: detector number, coordinate X, coordinate Y, angle of the main airflow direction, angle time stamp, and coordinate reference identifier.

[0069] The included angle calculation submodule constructs the detector position vector and the main airflow direction vector based on the coordinate direction mapping amount, calculates the vector projection relationship, determines the projection sign, and divides the forward and reverse regions to obtain the included angle sign sequence value;

[0070] Based on the coordinate direction mapping, a detector position vector and a main airflow direction vector are constructed. The position vector takes the coordinates of the center point of the main channel of the storage area as the origin, and the vector is a two-dimensional vector pointing from the origin to the detector coordinates. The main airflow direction vector is obtained by angle transformation and normalized. The angle is defined as true north 0° and clockwise as positive. When calculating the vector projection relationship, the dot product of each detector position vector and the main airflow direction vector is performed. If the dot product result is greater than 0, it is judged as a forward zone; if it is less than 0, it is judged as a reverse zone; if it is equal to 0, it is judged as a boundary zone and treated as a reverse zone to avoid misjudgment of forward zone. The boundary zone discrimination introduces a threshold of absolute dot product value less than 0.05. 0.05 is determined by the dot product distribution under the typical point density of 8-15m spacing after the position vector is normalized, and is used to cover points close to the vertical direction. The partition identifier obtained by each detector over time in this cycle is written into the angle symbol sequence value, so that the subsequent difference calculation can directly read the forward and reverse zone member sets at the same time at each time mark t and maintain a consistent partition aperture.

[0071] The forward and reverse difference submodule calls the included angle symbol sequence value and associates it with the inspection cycle concentration dataset. It calculates the corresponding detector concentration difference according to the forward and reverse zone identifiers and performs intra-zone aggregation processing to establish a storage area directional concentration difference table.

[0072] The angle symbol sequence value is called and associated with the inspection cycle concentration dataset. At the same time mark t, the concentration value sets of all detectors in the forward zone and the concentration value sets of all detectors in the reverse zone are extracted respectively, and the representative values ​​of the zone are calculated respectively. The representative values ​​of the zone are obtained by removing extreme values ​​and averaging. First, the highest and lowest values ​​are removed by sorting by concentration, and then the average value is calculated. When the number of detectors in the zone is less than 3, the removal is not performed and the average value is directly taken. The forward and reverse difference value is obtained by subtracting the representative value of the reverse zone from the representative value of the forward zone, and the difference value is filled back into the direction difference value field of each detector at that time to form the storage area direction concentration difference value table. Consistency check is performed on the time dimension of the difference value table. If the number of valid data entries in the forward zone or the reverse zone is 0 at a certain time, the difference value at that time is recorded as missing and does not participate in the subsequent matrix filling. Inspection cycles with a missing rate of more than 0.1 are marked as low quality cycles and are not included in the three-cycle comparison, thus ensuring that diffusion determination is based only on the complete and aligned difference value data of the partition.

[0073] Table 2. Example of reservoir coordinate orientation and direction difference:

[0074]

[0075] As shown in Table 2, under a wind direction of 90.0°, the average value of the forward wind zone is 17.2% LEL, the average value of the reverse wind zone is 10.8% LEL, and the direction difference is 6.4% LEL.

[0076] Please see Figure 4 The diffusion determination module includes:

[0077] The difference table building submodule is based on the reservoir directional concentration difference table. It collects the detector index and directional difference fields in the table and verifies the index alignment relationship. It calculates the index cross-pairing difference and writes it into the matrix cell to generate the difference matrix.

[0078] Based on the reservoir directional concentration difference table, detector index and directional difference fields are collected. The detector index is represented by a continuous integer code obtained by mapping the controller loop address and is guaranteed to be unique within the same reservoir. The index alignment relationship is verified by checking whether each detector index in the difference table exists in the coordinate direction mapping; if not, it is deleted. For existing but duplicate indexes, only the record with the highest quality identifier is retained, and the quality identifier is sorted in order of original validity over backfill correction over low confidence. The index cross-pairing difference is calculated and written into the matrix cell, with the matrix row corresponding to detector index i, the matrix column corresponding to detector index j, and the cell value written as the directional difference i. The absolute value of the directional difference j is subtracted and the sign bit is retained for subsequent anomaly point location. When i=j, the cell is set to 0. To avoid abnormal amplification of the matrix due to occasional noise, each directional difference is first normalized to the 0-1 interval and linearly scaled. The upper and lower limits of scaling are taken as the minimum and maximum values ​​of the directional difference sequence of the detector in the current period. If the absolute value of the difference between the maximum and minimum values ​​is less than 0.5%LEL, the entire difference sequence is set to 0 and marked as a low fluctuation detector to avoid numerical instability caused by an excessively small denominator. The difference matrix is ​​output and the matrix index is kept to correspond one-to-one with the detector index to ensure that the same index system can be reused in subsequent vertical distance rearrangement.

[0079] The vertical distance rearrangement submodule, based on the difference matrix, collects the coordinates of the main channel centerline and calculates the perpendicular distance between the detector coordinates and the centerline. It then reorders the matrix difference entries in the same row of the difference matrix according to the perpendicular distance between the detector and the main channel centerline in the corresponding column, from smallest to largest, to obtain the rearranged direction difference sequence and vertical distance sequence. The specific formula for calculating the change in difference between adjacent difference entries after rearrangement is as follows:

[0080] ;

[0081] The change in difference is calculated, and the change in difference is compared with the change threshold. Records exceeding the threshold are marked, and an over-threshold record identifier is established.

[0082] in, Represents detector number The corresponding change in difference is, Represents detector number In the vertical distance sorting sequence The difference in direction, Represents detector number In the vertical distance sorting sequence Vertical distance, Represents the difference in direction. Represents vertical distance. and Representing the detector number respectively The maximum and minimum terms of the corresponding direction difference sequence and Representing the detector number respectively The maximum and minimum terms of the corresponding vertical distance sequence, Represents detector number Number of items participating in the sorting Represents the numbering of the detector From item 2 to item 3 Items accumulated, Represents absolute value operation. Represents the square root operation;

[0083] The coordinates of the main channel centerline are collected based on the difference matrix, and the vertical distance between the detector coordinates and the centerline is calculated. The centerline of the main channel is determined by the coordinates of two endpoints from the factory CAD or survey points, and a straight line equation is fitted. The vertical distance is expressed in meters and retained to three decimal places. The matrix entries in the same row i are reordered in ascending order of the vertical distance of the detectors in the corresponding columns, resulting in a rearranged direction difference sequence and a vertical distance sequence. The distance values ​​in the sequence are deduplicated. If the vertical distance difference between two adjacent detectors is less than 0.2m, they are considered to be in the same bandwidth. Within the bandwidth, the detectors are sorted in ascending order by index, and their difference entries are treated as consecutive adjacent items. The 0.2m limit is fixed by the statistical upper limit of the re-measurement of the layout construction error. When the difference between the maximum and minimum values ​​of the vertical distance sequence is less than 0.2m, the vertical distance is considered to be within the same bandwidth. When the coverage is insufficient at 1.0m, the difference change is not calculated. For sequences that meet the conditions, the difference change is calculated according to a given formula and compared with the change threshold. The change threshold is obtained through field experiment calibration. In the leak-free condition A, 3,600 inspection cycles are collected to obtain the difference change distribution, and the 99th percentile value is taken to form the initial threshold value. When the initial threshold value falls between 0.24 and 0.27, it is fixed to 0.26. Then, in the controlled release condition B, the threshold is adjusted down by 0.01 steps with a recall rate of 0.95 as a constraint, and the false alarm rate of condition A is controlled to not exceed 0.02. The threshold is determined to be 0.25. Records exceeding the threshold are written into the threshold record identifier and the detector number, change value and corresponding rearranged index position are retained for subsequent list extraction.

[0084] The anomaly list submodule calls the threshold record identification quantity and associates it with the storage area directional concentration difference table, filters the corresponding records, extracts the item with the largest directional difference and the corresponding storage tank area coordinate point, and generates a list of abnormal gas diffusion points in the storage area.

[0085] The system calls the threshold record identifier and associates it with the storage area directional concentration difference table. It filters all records of threshold times by the same detector index and extracts the item with the largest directional difference and its occurrence time. It then checks the coordinate directional mapping at that time to obtain the corresponding storage tank area coordinate point and writes it into the abnormal point list. When the same detector exceeds the threshold multiple times in one cycle, it only retains the coordinate point corresponding to the item with the largest directional difference to form a unique abnormal point record for a single cycle. The abnormal point list fields are fixed as detector number, detector index, abnormal time, maximum directional difference, coordinate X, coordinate Y, difference change, and quality statistics field. This allows subsequent cycle locking to directly perform cross-cycle consistency filtering at the number and coordinate level and maintain traceable association with the upstream difference matrix and vertical distance sorting.

[0086] Please see Figure 5 The periodic locking module includes:

[0087] The periodic comparison submodule compares the list of abnormal gas diffusion points in the storage area for three consecutive inspection cycles, extracts the periodic point number and coordinate field, performs cross-matching of the same number, determines the consistency of the number within the three cycles and retains the consistent record, and obtains the number of consistent points in the three cycles.

[0088] The list of abnormal gas diffusion points in the storage area for three consecutive inspection cycles is compared and recorded as P1, P2, and P3 in chronological order, ensuring that the interval between adjacent cycles does not exceed 30 minutes. If the interval exceeds 30 minutes, the on-site operating conditions are deemed incomparable, and the three-cycle window is abandoned, sliding forward one cycle to regroup the window. The cycle point number and coordinate fields are extracted, and cross-matching with the same number is performed. First, an index is created for P1 by point number, then the point number is searched in P2 and written into the candidate consistency set, and finally, the candidate consistency set is filtered again using P3. Only numbers that exist in all three cycles are retained. To avoid false consistency caused by identical numbers but incorrect coordinate entry, coordinate consistency constraints are added to the candidate consistent set. The coordinate difference between P1 and P2 and the coordinate difference between P2 and P3 are calculated. Any difference greater than 5.0m is discarded. 5.0m is taken from the minimum span limit of a single zone in the tank area to ensure the spatial interpretability of stable diffusion of the same leakage source. Consistent records are retained to form consistent point quantities for three cycles, and the maximum directional difference and the change of difference for each of the three cycles are carried in the fields for subsequent displacement calculation and stability screening.

[0089] The displacement calculation submodule is based on the consistent point quantities of three periods. It collects the coordinates of corresponding points in adjacent periods and calculates the planar displacement between adjacent periods. It forms a displacement sequence according to the period order and calculates the displacement value to generate the periodic displacement.

[0090] The displacement is calculated using Euclidean distance and rounded to three decimal places in meters. For each point, two adjacent displacement values ​​are generated, corresponding to P1 to P2 and P2 to P3 respectively, forming a displacement sequence in cyclical order. The cyclic displacement is taken as the maximum value in the displacement sequence to ensure that any significant drift in any adjacent cycle triggers rejection. The displacement threshold is calibrated through field experiments. In controlled release conditions, with the release point fixed, the displacement distribution over three cyclic windows is statistically analyzed, with the 95th percentile falling within 1.2-1.6m. In mobile interference conditions, a mobile source is used to statistically analyze the displacement, with the 95th percentile falling within 3.5-6.0m. To balance stable leak source locking and mobile interference rejection, the displacement threshold is set to 2.0m, and the false locking rate is verified to be no more than 0.01% after 30 days of trial operation. The cyclic displacement is written into the structure and kept consistent with the point number and coordinate fields, enabling threshold comparison for locking screening within the same field system.

[0091] The locking and filtering submodule calls the periodic displacement and associates it with the consistent point values ​​of the three periods. It compares the displacement with the displacement threshold and retains records that are less than the displacement threshold. It summarizes the corresponding number and coordinate fields to generate a continuous diffusion anomaly record table.

[0092] The system retrieves the periodic displacement and associates it with the consistent point values ​​across three periods. It compares the displacement with a threshold of 2.0m and retains records less than the threshold. Simultaneously, it reads the maximum directional difference for that point over three periods and calculates the three-value range. This range is compared with a range threshold of 3.0%LEL. If the range exceeds 3.0%LEL, it is considered unstable diffusion and removed from the list. 3.0%LEL is taken from the 99th percentile of the maximum directional difference range across periods for the same point in leak-free condition A to suppress occasional locking caused by sudden wind direction changes. The system summarizes the retained records, including the number and coordinate fields, and writes the mean and difference of the maximum directional differences over three periods. The mean change, periodic displacement, and stability fields are used to generate a continuous diffusion anomaly record table, which serves as the sole input for the linked execution. After a 30-day trial run, the difference in false alarms between single-cycle threshold judgment and the introduction of three-cycle locking was compared. The single-cycle judgment output 1,184 anomalies, of which 212 were actual leaks or continuous anomalies verified by manual review, resulting in a false alarm rate of 0.821%. After introducing three-cycle locking, the output was 164 anomalies, of which 151 were actual leaks or continuous anomalies, resulting in a false alarm rate of 0.079. The false alarm rate decreased by 0.742 while maintaining usable detection capability, demonstrating that locking constraints have a suppressive effect on random fluctuation triggering.

[0093] Please see Figure 6 The linkage execution module includes:

[0094] The difference grading submodule, based on the continuous diffusion anomaly record table, collects and records concentration differences and extracts the item with the largest concentration difference using the following specific calculation formula:

[0095] ;

[0096] The alarm level value is obtained by mapping the level threshold range.

[0097] in, Represents the alarm level value. Representing the The concentration difference of the abnormal records Represents the number of abnormal records. This represents the concentration value corresponding to the highest concentration in the set of abnormal records. Represents the concentration benchmark value and is consistent with , Same unit, Represents the relay response delay and is Same unit, Represents the sampling time interval value. Represents the relay target status code value. This represents the initial status code value of the relay. Represents the baseline value of the status code. Represents the output circuit impedance and is related to Same unit, Represents the impedance reference value;

[0098] Based on the continuous diffusion anomaly record table, concentration differences are collected and used to form a grading batch. The batch window is fixed as the set of valid anomaly records from the most recent inspection cycle. When the number of anomaly records n is less than 3, the batch is marked as observation and the grading is delayed by one cycle to reduce sparse triggering false alarms. For each anomaly record, the concentration difference ΔCi is read and its absolute ratio to the concentration benchmark Cref is calculated. At the same time, the concentration value Cmax corresponding to the highest concentration in the anomaly record set is read and its absolute ratio to Cref is calculated. The concentration benchmark Cref is fixed at 10.0% LEL and is compared with ΔCi. Cmax is kept in the same unit to ensure consistent normalization; the relay response delay Rk is obtained through on-site action testing. The contact voltage reversal time is recorded simultaneously with the issuance of the status update command, repeated 50 times, and the median is fixed at 0.20s. The sampling time interval Ts is fixed at 2.0s, consistent with the inspection sampling; the relay target status code Sr is set to 1, the initial status code S0 to 0, and the status code reference value Sref to 1 to normalize the status difference; the output circuit impedance Zl is obtained by injecting a 1mA test current and measuring the voltage during circuit self-testing, repeated 10 times, and the average value is 38.0Ω. The anti-base value Zref is set to 40.0Ω to match the design nominal value. Substituting the above parameters into the given formula, the alarm level value G is calculated, and the alarm level is mapped according to the level threshold range. The threshold range is fixed as follows: G < 0.80 maps to Level 1 warning, 0.80 ≤ G ≤ 1.60 maps to Level 2 alarm, and G > 1.60 maps to Level 3 linkage. For example, in a batch with n=5, ΔCi is 6.4%LEL, 7.1%LEL, 5.8%LEL, 8.0%LEL, and 6.9%LEL, and Cmax is 18.2%LEL. Substituting these parameters, G=1 is calculated. The result, mapped to Level 3 linkage, shows that the difference level and maximum concentration level of the abnormal batch are significantly higher than the benchmark, and the loop executability meets the linkage requirements. The advantage of the formula is that by involving the average relative amplitude of the concentration difference of the abnormal records, the relative amplitude of the maximum concentration, the relay response, and the loop state factor in the calculation, the classification level simultaneously covers diffusion intensity and executability. The experimental comparison shows that when only the concentration difference threshold is used for judgment, the false alarm rate of Level 3 is 14.6%. After introducing the G classification level, the false alarm rate of Level 3 decreases to 6.1%, and the actual leakage Level 3 trigger rate increases from 78.0% to 86.0%.

[0099] Table 3 Comparison of Judgment Threshold and Linkage Effect:

[0100]

[0101] As shown in Table 3, the threshold combination A02 achieves a balance between the detection rate and the false alarm rate, so 0.80 and 1.60 are selected as the level boundaries.

[0102] The loop addressing submodule collects the list of loop address codes output by the fire alarm controller based on the alarm level value and performs address matching and filtering to obtain the loop address code value.

[0103] The fire alarm controller output loop address code list is collected based on the alarm level value, and address matching and filtering are performed. The address list is stored in three fields: alarm level, loop address code, and controlled device type. The loop address code corresponding to the level 3 linkage is fixed to the exhaust linkage relay and the audible and visual alarm loop. The level 2 alarm corresponds to the on-site audible and visual alarm and the duty room prompt. The level 1 warning corresponds to only recording and push. When performing address matching, the candidate address set is first filtered by alarm level, and then filtered by storage area number and detector loop partition field to ensure that the linkage is not triggered across areas. After obtaining the loop address code value, a status update command frame is constructed and written to the transmission buffer. The command frame contains the target address code, the target status code Sr, the command sequence number, and the check field. The check field adopts 16-bit cyclic redundancy check and is implemented consistently at the sending and receiving ends to prevent bit errors, thereby ensuring that the controller-side addressing and action commands can be reproduced and transmitted within the same encoding system.

[0104] The feedback acquisition submodule calls the loop address code value, acquires the relay status and writes the status update instruction, acquires feedback information and performs a consistency judgment on the feedback code and status code, and obtains the fire linkage execution status feedback.

[0105] After calling the loop address code value, the relay status is collected and a status update command is written. After the command is sent, a timeout window of 2 sampling periods is used. Starting 0.5s after sending, the relay status input point is polled and collected once every 0.2s until 4.0s. The collected feedback code and the target status code are compared for consistency. If they match for 3 consecutive times, the execution is considered successful and a successful execution feedback is output. If they do not match for 3 consecutive times, the execution is considered unsuccessful and a resend is triggered. If the resend still fails, a linkage failure alarm is written and the batch of G, the measured value of the loop impedance and the measured value of the response delay are recorded in the event database for subsequent threshold recalibration. Taking the example, the above G=2.14 is mapped to level 3 linkage. After matching the loop address code A07, a closing command is sent. Starting at 1.1s, the feedback code is collected for 3 consecutive times and is consistent with Sr. The execution is considered successful and the fire linkage execution status feedback is output as closed. At the same time, the coordinates of the abnormal point and the execution result are pushed to the site for location and investigation.

[0106] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A smart inspection and fire emergency response system for hazardous chemical storage, characterized in that, The system includes: The inspection sequence module obtains inspection data from the combustible gas detectors in the hazardous chemical warehouse, reads combustible gas concentration data and time stamps, filters data with time stamps within a set period, sorts them by frame number, marks skipped data, and organizes them into an inspection cycle concentration dataset. Based on the inspection cycle concentration dataset, the airflow difference module calculates the angle between the detector and the airflow direction, combined with the detector coordinates and the angle of the main airflow direction in the tank area. It then calculates the concentration difference between the headwind and the headwind according to the size of the angle, and obtains the concentration difference table of the storage area direction. The diffusion determination module establishes a difference matrix based on the directional concentration difference table of the storage area, calculates the vertical distance between the detector and the center line in combination with the coordinates of the main channel center line, and reorders the matrix difference entries in the same row of the difference matrix according to the vertical distance between the detector and the center line of the main channel in the corresponding column from small to large, to obtain the rearranged directional difference sequence and vertical distance sequence. It calculates the difference change of adjacent difference entries after rearrangement, filters records with change exceeding the predetermined change threshold, extracts the maximum difference and the corresponding tank area coordinates, and generates a list of abnormal gas diffusion points in the storage area. The periodic locking module compares the list of abnormal gas diffusion points in the storage area, filters records that are consistent within the period, calculates the displacement of points in adjacent periods, filters abnormal records whose displacement is less than a preset displacement threshold, and generates a continuous diffusion abnormal record table. Based on the continuous diffusion anomaly record table, the linkage execution module determines the alarm level according to the maximum concentration difference, looks up the corresponding fire alarm controller output circuit address code in the fire control room, updates the relay status, and reads feedback information to obtain fire linkage execution status feedback.

2. The intelligent inspection and fire emergency response system for hazardous chemical storage as described in claim 1, characterized in that, The inspection cycle concentration dataset includes average concentration values, peak concentration points, and data integrity identifiers; the storage area directional concentration difference table includes a set of downwind difference values, a set of upwind difference values, and a directional classification identifier; the storage area gas diffusion anomaly point list includes anomaly point coordinates, maximum difference values, and anomaly level identifiers; the continuous diffusion anomaly record table includes continuous anomaly point numbers, periodic displacement parameters, and stability discrimination identifiers; the fire-fighting linkage execution status feedback includes alarm level identifiers, loop action status, and feedback confirmation signals.

3. The intelligent inspection and fire emergency response system for hazardous chemical storage as described in claim 1, characterized in that: The threshold for the change is the upper limit of the threshold range determined based on the statistical characteristic value of the concentration difference between the windward and headwind directions in the reservoir direction concentration difference table; The records whose changes exceed a predetermined threshold are retained after sorting, where the difference between adjacent positions continuously exceeds the threshold.

4. The intelligent inspection and fire emergency response system for hazardous chemical storage as described in claim 1, characterized in that, The inspection sequence module includes: The data frame acquisition submodule obtains the data reported during the inspection from the combustible gas detectors in the hazardous chemical warehouse, reads the combustible gas concentration data and time stamps in the data frames, verifies the correspondence between the time stamps and the frame sequence number, removes abnormal marker entries, and obtains the concentration-time mapping. The periodic screening submodule, based on the concentration-time mapping, filters entries whose time markers are within the determination period interval, determines the continuous state of the time series, reassembles the corresponding concentration series, and generates the number of sequences within the period. The frame sequence verification submodule calls the sequence quantity within the period, arranges them in order according to the frame sequence number, and judges whether there is an offset state in the difference between adjacent frame sequence numbers. It marks the detector data with skipped number characteristics and collects the corresponding concentration entries to establish the inspection period concentration dataset.

5. The intelligent inspection and fire emergency response system for hazardous chemical storage as described in claim 1, characterized in that, The airflow differential module includes: Based on the inspection cycle concentration dataset, the orientation parameter submodule collects the coordinates of the tank area detectors and the angle of the main airflow direction, unifies the coordinate reference, verifies the correspondence between the coordinate identifiers and the detector numbers, and generates coordinate direction mapping quantities. The included angle calculation submodule constructs the detector position vector and the main airflow direction vector based on the coordinate direction mapping amount, calculates the vector projection relationship, determines the projection sign, divides the forward and reverse regions, and obtains the included angle sign sequence value. The forward and reverse difference submodule calls the included angle symbol sequence value and associates it with the inspection cycle concentration dataset. It calculates the corresponding detector concentration difference according to the forward and reverse zone identifiers and performs intra-zone aggregation processing to establish a storage area directional concentration difference table.

6. The intelligent inspection and fire emergency response system for hazardous chemical storage as described in claim 1, characterized in that, The diffusion determination module includes: The difference table building submodule, based on the storage area directional concentration difference table, collects the detector index and directional difference fields in the table and verifies the index alignment relationship, calculates the index cross-pairing difference and writes it into the matrix unit to generate the difference matrix quantity; The vertical distance rearrangement submodule, based on the difference matrix, collects the coordinates of the main channel centerline and calculates the vertical distance between the detector coordinates and the centerline. It then reorders the matrix difference entries in the same row of the difference matrix according to the vertical distance between the detector and the main channel centerline in the corresponding column from smallest to largest, obtaining the rearranged direction difference sequence and vertical distance sequence. It also calculates the difference change of adjacent difference entries after rearrangement, compares the difference change with the change threshold, and marks records that exceed the threshold, thus establishing an over-threshold record identifier. The anomaly list submodule calls the threshold record identifier and associates it with the storage area directional concentration difference table, filters the corresponding records, extracts the item with the largest directional difference and the corresponding storage tank area coordinates, and generates a list of abnormal gas diffusion points in the storage area.

7. The intelligent inspection and fire emergency response system for hazardous chemical storage as described in claim 6, characterized in that, The process of reordering the detectors according to their vertical distances from the center line of the main channel from smallest to largest yields a rearranged sequence of directional differences and a sequence of vertical distances. The specific formula for calculating the change in difference between adjacent difference entries after rearrangement is as follows: ； The change in difference is calculated, and the change in difference is compared with the change threshold. Records exceeding the threshold are marked, and an over-threshold record identifier is established. in, Represents detector number The corresponding change in difference is, Represents detector number In the vertical distance sorting sequence The difference in direction, Represents detector number In the vertical distance sorting sequence Vertical distance, Represents the difference in direction. Represents vertical distance. and Representing the detector number respectively The maximum and minimum terms of the corresponding direction difference sequence and Representing the detector number respectively The maximum and minimum terms of the corresponding vertical distance sequence, Represents detector number Number of items participating in the sorting Represents the numbering of the detector From item 2 to item 3 Items accumulated, Represents absolute value operation. This represents the square root operation.

8. The intelligent inspection and fire emergency response system for hazardous chemical storage as described in claim 1, characterized in that, The periodic locking module includes: The periodic comparison submodule compares the list of abnormal gas diffusion points in the storage area for three consecutive inspection cycles, extracts the periodic point number and coordinate field, performs cross-matching of the same number, determines the consistency of the number within the three cycles and retains the consistent record, and obtains the number of consistent points in the three cycles. The displacement calculation submodule collects the coordinates of corresponding points in adjacent periods based on the three-period consistent point quantities and calculates the planar displacement between adjacent periods. It forms a displacement sequence according to the period order and calculates the displacement value to generate the periodic displacement. The locking and filtering submodule calls the periodic displacement amount and associates it with the three-period consistent point amount. It compares the displacement amount with the displacement threshold and retains the records that are less than the displacement threshold. It summarizes the corresponding number and coordinate field to generate a continuous diffusion anomaly record table.

9. The intelligent inspection and fire emergency response system for hazardous chemical storage as described in claim 1, characterized in that, The linkage execution module includes: The difference level determination submodule collects and records concentration differences based on the continuous diffusion anomaly record table and extracts the item with the largest concentration difference. It then obtains the alarm level value by mapping the level threshold interval. The loop addressing submodule collects the list of loop address codes output by the fire alarm controller and performs address matching and filtering based on the alarm level value to obtain the loop address code value; The feedback acquisition submodule calls the circuit address code value, acquires the relay status and writes the status update instruction, acquires feedback information and performs a consistency judgment on the feedback code and status code, and obtains the fire linkage execution status feedback.

10. The intelligent inspection and fire emergency response system for hazardous chemical storage as described in claim 1, characterized in that: The preset displacement threshold is a displacement limit determined based on the historical statistical distribution of displacements of adjacent periodic points. The screening of abnormal records with displacements less than the preset displacement threshold is done by calculating the displacements of adjacent periodic points based on the difference in plane coordinates of the tank area coordinate points, and retaining abnormal records that meet the preset displacement threshold for multiple consecutive inspection cycles.

Images