Hazardous chemical substance safety production risk monitoring and dynamic early warning system

Abstract

The invention discloses a hazardous chemical substance safety production risk monitoring and dynamic early warning system, particularly relates to the technical field of industrial safety monitoring, and is used for solving the problem of monitoring failure caused by signal interference or offset in a mixed leakage scene of various hazardous chemical substances in the prior art. Multi-source monitoring data such as gas concentration and temperature are acquired in real time through a data acquisition module, and a conflict identification module detects reverse fluctuation characteristics of sensor data; the leakage verification module is used for reversely positioning a leakage source and verifying path shielding interference based on production pipe network topology and real-time flow data; after the physical shielding is eliminated, the physicochemical analysis module analyzes the signal counteracting characteristic of the hazardous chemical substance combination and verifies the matching between the diffusion path and the sensor layout; the dynamic compensation module dynamically corrects the monitoring data according to a physicochemical reaction mechanism; the risk assessment module generates a mixed leakage risk grade and triggers a graded emergency response instruction; accurate identification and dynamic early warning of mixed leakage risks are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial safety monitoring. More specifically, the present invention relates to a risk monitoring and dynamic early warning system for the safe production of hazardous chemicals. Background Art

[0002] In the production process of hazardous chemicals, existing safety monitoring systems usually deploy a variety of sensors to collect environmental parameters (such as gas concentration, temperature, pressure, etc.) in real time, and judge the leakage risks of single or a few hazardous chemicals based on preset thresholds; when it is detected that a certain parameter exceeds the safety range, the system triggers an alarm and links to emergency measures; such methods can effectively identify risks in the scenario of single hazardous chemical leakage, but in actual industrial scenarios, the coexistence and mixed leakage of multiple hazardous chemicals often occur.

[0003] Existing monitoring technologies have limitations in monitoring scenarios of mixed leakage of multiple hazardous chemicals. That is, when different hazardous chemicals leak simultaneously, their physical and chemical reactions may cause interference or cancellation of sensor signals (for example, the decrease in gas concentration caused by acid-base neutralization reaction), resulting in the system being unable to accurately identify the true composite leakage risk. This problem of false negatives or false positives caused by the signal annihilation effect seriously reduces the reliability of the monitoring system and increases potential accident hazards. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a risk monitoring and dynamic early warning system for the safe production of hazardous chemicals to solve the problems raised in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solutions: A risk monitoring and dynamic early warning system for the safe production of hazardous chemicals, including the following modules: A data acquisition module, configured to obtain multi-source monitoring data in the production environment of hazardous chemicals in real time, including gas concentration data and temperature data; A conflict identification module, configured to identify whether there is a reverse fluctuation exceeding a preset safety threshold in the sensor data corresponding to the leakage of at least two hazardous chemicals in the multi-source monitoring data; A leakage verification module, configured to reverse calculate the location of the leakage source based on the topological structure of the production pipeline network and the real-time flow data, and verify whether the reverse fluctuation is caused by the path occlusion from the leakage source to the sensor; A physical and chemical analysis module, when excluding the case caused by path occlusion, is configured to analyze whether the combination of hazardous chemicals corresponding to the reverse fluctuation has the signal cancellation characteristic of the sensor, and verify whether the diffusion path of the hazardous chemicals matches the sensor layout; A dynamic compensation module, which is used to generate a leakage signal correction model according to the physical and chemical reaction mechanism and perform dynamic compensation on multi-source monitoring data if the hazardous chemical combination has signal cancellation characteristics and the diffusion paths do not match; A risk assessment module, which is used to generate a mixed leakage risk level based on the compensated multi-source monitoring data and determine whether to trigger a hierarchical emergency response instruction matching the hazardous chemical combination.

[0006] In a preferred embodiment, multi-source monitoring data in the production environment of hazardous chemicals is obtained in real time, including gas concentration data and temperature data, including: Gas concentration sensors and temperature sensors are deployed at different process nodes in the production area of hazardous chemicals; The monitoring data of the gas concentration sensors and temperature sensors are aligned in real time through a timestamp synchronization protocol to generate time-synchronized multi-source monitoring data; Noise filtering processing and filtering processing are performed on the time-synchronized multi-source monitoring data; The calibrated and filtered multi-source monitoring data is stored in a real-time database.

[0007] In a preferred embodiment, it is identified whether the sensor data corresponding to the leakage of at least two hazardous chemicals in the multi-source monitoring data has a reverse fluctuation exceeding a preset safety threshold, including: The gas concentration data and temperature data within a continuous time window are extracted, and the differential fluctuation values of the gas concentration data and temperature data within the time window are calculated; If the differential fluctuation value exceeds the preset safety threshold of the corresponding hazardous chemical, it is determined that the gas concentration data and temperature data have a reverse fluctuation trend; According to whether the fluctuation directions of the gas concentration data and temperature data are opposite, it is judged whether the reverse fluctuation condition is met; If the reverse fluctuation condition is met, the hazardous chemical combination corresponding to the gas concentration data and temperature data is associated, and a reverse fluctuation event identifier is generated; Based on the equipment operation status data and environmental parameter data, pseudo-reverse fluctuation events caused by equipment start-stop or environmental mutation are excluded.

[0008] In a preferred embodiment, based on the topological structure of the production pipeline network and the real-time flow data, the leakage source location is calculated reversely to verify whether the reverse fluctuation is caused by the path occlusion from the leakage source to the sensor, including: According to the topological structure model of the production pipeline network and the real-time flow data, the flow change rate of each pipeline node is calculated, and candidate abnormal nodes with a flow change rate exceeding the dynamic threshold are screened; Combined with the historical leakage frequency of the candidate abnormal nodes and the pressure mutation directions of their upstream and downstream pipelines, leakage source probability weights are assigned to generate a leakage source candidate list sorted by weight; Based on real-time wind speed data and the diffusion rate of hazardous chemicals, correct the diffusion path direction of the leakage source candidate, and generate a theoretical diffusion trajectory that matches the sensor position. According to the physical layout data of the devices around the target sensor, detect whether there are obstacles on the path of the theoretical diffusion trajectory, and calculate the expected concentration attenuation range of the leaked substance reaching the sensor based on the geometric size of the obstacles and the diffusion characteristics of the hazardous chemicals. If the actual sensor signal fluctuation amplitude is less than the lower limit of the expected concentration attenuation range, determine that the reverse fluctuation is caused by path occlusion, and mark the corresponding leakage source candidate as low confidence.

[0009] In a preferred embodiment, when excluding the cause of path occlusion, analyze whether the hazardous chemical combination corresponding to the reverse fluctuation has the signal cancellation characteristic of the sensor, and verify whether the diffusion path of the hazardous chemicals matches the sensor layout, including: When excluding the cause of path occlusion, based on the reverse fluctuation event identifier, retrieve the dynamic physical and chemical reaction characteristics library of the hazardous chemical combination, and match whether the real-time phase difference between the gas concentration and the temperature change conforms to the time delay law of the synchronous neutralization reaction. If the real-time phase difference conforms to the time delay law, determine that the hazardous chemical combination has the signal cancellation characteristic of the sensor, and perform diffusion path verification to judge whether the diffusion path of the hazardous chemicals matches the sensor layout: If the real-time phase difference does not conform to the time delay law, verify whether the signal cancellation characteristic is caused by sensor sensitivity attenuation or hazardous chemical adsorption effect. Dynamically fuse the device maintenance log and the environmental mutation event library to exclude misjudgment of reverse fluctuations caused by equipment shutdown or environmental parameter jumps.

[0010] In a preferred embodiment, performing diffusion path verification to judge whether the diffusion path of the hazardous chemicals matches the sensor layout specifically is: Real-time monitor the temperature and pressure data of the leakage point. When the critical condition of the phase state conversion of the hazardous chemicals is reached, generate a gas-liquid two-phase diffusion path model, and calculate the proportion of the blind area volume not covered by the sensor in the gas-phase diffusion path. Retrieve the diffusion path data of the same hazardous chemical combination in the historical leakage event library, extract the average historical diffusion direction and the average rate, and calculate the current path direction deviation angle and the rate difference coefficient. If the proportion of the blind area volume exceeds the preset proportion threshold, or the current path direction deviation angle exceeds the preset angle threshold and the rate difference coefficient exceeds the preset difference coefficient, determine that the diffusion path of the hazardous chemicals does not match the sensor layout.

[0011] In a preferred embodiment, verifying whether the signal cancellation characteristic is caused by sensor sensitivity attenuation or hazardous chemical adsorption effect specifically is: According to the adsorption coefficient of the hazardous chemical combination and the sensor sensitivity attenuation model, calculate the attenuation amplitude of the effective concentration relative to the original leakage concentration; If the attenuation amplitude exceeds the lower threshold of the sensor detection dynamic range, it is determined that the signal cancellation characteristic is caused by adsorption or sensitivity attenuation.

[0012] In a preferred embodiment, if the hazardous chemical combination has a signal cancellation characteristic and the diffusion paths do not match, a leakage signal correction model is generated according to the physical and chemical reaction mechanism to dynamically compensate the multi-source monitoring data, including: Based on the synchronous neutralization reaction characteristics of the hazardous chemical combination and the verification result of the unmatched diffusion paths, retrieve the corresponding reaction rate equation and diffusion attenuation coefficient in the physical and chemical reaction mechanism library; Construct a leakage signal correction model according to the reaction rate equation and the diffusion attenuation coefficient. The leakage signal correction model includes the superposition relationship between the gas concentration compensation function and the temperature compensation function; Based on the real-time wind speed and temperature and humidity data, dynamically adjust the weight coefficient of the compensation function to generate the correction amounts of the gas concentration data and the temperature data; Superimpose the correction amounts on the gas concentration value and the temperature value of the original multi-source monitoring data, and output the compensated gas concentration data and temperature data.

[0013] In a preferred embodiment, generate a mixed leakage risk level based on the compensated multi-source monitoring data, and determine whether to trigger a hierarchical emergency response instruction matching the hazardous chemical combination, including: Extract the compensated gas concentration data and temperature data, combine the diffusion path verification result, and generate a mixed leakage risk level according to the diffusion range and real-time concentration gradient of the leaked substance; Based on the real-time environmental wind speed and diffusion range, generate the range boundary coordinates of the leakage impact area; Input the mixed leakage risk level and the range boundary coordinates of the impact area into the pre-stored hierarchical emergency response mapping table, and output the corresponding emergency response instruction type; If the mixed leakage risk level exceeds the preset safety threshold, trigger an emergency response instruction matching the type of hazardous chemical combination; Adjust the execution priority of the emergency response instruction according to the equipment operation status data.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. Based on the topology of the production pipeline network and real-time flow, reverse positioning of the leakage source is carried out. Combining path occlusion verification, it can accurately distinguish real leaks from physical interferences, avoiding misjudgment by a single sensor. Through the analysis of physical and chemical reaction characteristics, the causes of signal cancellation are identified and the matching of the diffusion path and sensor layout is verified, eliminating monitoring blind spots caused by chemical reactions of hazardous chemicals or equipment blind spots. Based on physical and chemical mechanisms, multi-source data is corrected in real time to restore the real leakage signal, solving the technical bottleneck of signal distortion in mixed leakage scenarios and significantly improving the reliability of leakage monitoring and the accuracy of risk identification in complex industrial environments.

[0015] 2. Through the synergistic effect of leakage source positioning, signal cancellation verification, dynamic compensation and hierarchical response, a full-process closed-loop management of mixed leakage risks is achieved. It can not only identify compound leakage events in scenarios where multiple hazardous chemicals coexist, but also dynamically adjust the risk assessment model according to the leakage diffusion path and real-time environmental parameters, generating hierarchical emergency instructions matching the characteristics of hazardous chemicals. This adaptive monitoring and response mechanism effectively reduces the rates of missed reports and false alarms, providing a more accurate and intelligent active defense ability for the safety of hazardous chemical production. Description of the Drawings

[0016] Figure 1 It is a schematic structural diagram of the hazardous chemical production safety risk monitoring and dynamic early warning system of the present invention; Figure 2 It is a flow chart for verifying the characteristics of signal cancellation of hazardous chemical combinations and judging the matching of diffusion paths of the present invention. Detailed Embodiments

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0018] Embodiment: Figure 1 A schematic structural diagram of the hazardous chemical production safety risk monitoring and dynamic early warning system of the present invention is given. The hazardous chemical production safety risk monitoring and dynamic early warning system includes the following modules: A data acquisition module, which is used to obtain multi-source monitoring data in the production environment of hazardous chemicals in real time, including gas concentration data and temperature data; A conflict identification module, which is used to identify whether the sensor data corresponding to the leakage of at least two hazardous chemicals in the multi-source monitoring data has a reverse fluctuation exceeding a preset safety threshold; A leakage verification module, which is used to inversely calculate the location of the leakage source based on the topological structure and real-time flow data of the production pipeline network, and verify whether the reverse fluctuation is caused by the path occlusion from the leakage source to the sensor; A physical and chemical analysis module, when the path occlusion is excluded, is used to analyze whether the hazardous chemical combination corresponding to the reverse fluctuation has the signal cancellation characteristic of the sensor, and verify whether the diffusion path of the hazardous chemical matches the sensor layout; A dynamic compensation module, if the hazardous chemical combination has the signal cancellation characteristic and the diffusion path does not match, is used to generate a leakage signal correction model according to the physical and chemical reaction mechanism, and perform dynamic compensation on the multi-source monitoring data; A risk assessment module, which is used to generate a mixed leakage risk level based on the compensated multi-source monitoring data, and determine whether to trigger a hierarchical emergency response instruction that matches the hazardous chemical combination.

[0019] Real-time obtain multi-source monitoring data in the production environment of hazardous chemicals, including gas concentration data and temperature data. The specific implementation is as follows: Deploy gas concentration sensors and temperature sensors at different process nodes in the production area of hazardous chemicals. The deployment positions of the gas concentration sensors and temperature sensors cover hazardous chemical storage tanks, pipelines, and reaction kettles. A set of gas concentration sensors is installed at the top and bottom of the hazardous chemical storage tank respectively to monitor the gas concentration gradient change inside and outside the tank; temperature sensors are arranged at the valve connections and bends of the pipeline to capture local temperature anomalies caused by leakage; gas concentration sensors and temperature sensors are evenly distributed on the outer wall of the reaction kettle to form a surrounding monitoring network. Through the above deployment method, ensure full coverage monitoring of the storage, transportation, and reaction links of hazardous chemicals.

[0020] Perform real-time alignment on the monitoring data of the gas concentration sensors and temperature sensors through the timestamp synchronization protocol to generate multi-source monitoring data with time synchronization. The timestamp synchronization protocol uses the Network Time Protocol. A clock chip is built into the data acquisition terminal of each sensor, and time calibration is performed with the main control server through the local area network. The calibration frequency is, for example, once per second, ensuring that the timestamp error of all sensor data is, for example, less than 10 milliseconds. The data sampling frequency of the gas concentration sensor is set to, for example, 2 times per second, and the temperature sensor is, for example, 1 time per second. The linear interpolation algorithm is used to align the timestamp of the temperature data to the acquisition time point of the gas concentration data to generate a multi-source monitoring data set with time synchronization.

[0021] Noise filtering is performed on time-synchronized multi-source monitoring data. This includes baseline calibration of gas concentration data based on ambient background values and filtering of temperature data based on the equipment's operating status. The ambient background value is calculated by continuously collecting gas concentration data for, for example, 24 hours, in a leak-free state. Data points exceeding ±3 standard deviations of the mean are removed, and the hourly mean concentration is calculated to form a baseline background curve. When the difference between the real-time gas concentration data and the baseline background curve exceeds, for example, ±5% of the baseline value, it is considered a valid signal; otherwise, it is considered ambient noise and reset to zero.

[0022] The method for judging the operating status of the equipment is as follows: the current signal and vibration signal of the pipeline pump are collected. If the current value is within the rated range and there is no abnormal peak in the vibration spectrum, the equipment is determined to be in steady-state operation. At this time, the temperature data is filtered using a sliding average filter with a window length of, for example, 5 seconds; if the equipment is in a start-stop or load sudden change state, it is switched to weighted median filtering, and the weight is dynamically adjusted according to the current change rate. For example, when the current change rate exceeds 5 amperes per second, the weight of the current temperature data is 0.7, the weight of the data in the previous second is 0.2, and the weight of the data in the previous two seconds is 0.1.

[0023] The calibrated and filtered multi-source monitoring data is stored in a real-time database and transmitted to the conflict identification module. The real-time database uses a time series database structure, with storage fields including timestamp, sensor number, gas concentration value, temperature value, and data quality indicator field. The data quality indicator field is used to mark whether the data has passed the noise filtering process. The indicator value is 0 for raw data, 1 for calibrated data, and 2 for filtered data.

[0024] Data transmission uses a lightweight message queue protocol. The conflict identification module subscribes to the specified topic in the real-time database. When the real-time database writes new data, it is automatically pushed to the message queue of the conflict identification module. The push frequency is consistent with the data sampling frequency.

[0025] In the noise filtering process, the window length of the sliding average filter is dynamically adjusted according to the equipment type: for example, the window length of the reactor temperature data is 10 seconds, and the window length of the pipeline temperature data is 3 seconds.

[0026] The weight calculation rule for weighted median filtering is as follows: For example, if the current rate of change is less than 2 amperes per second, the current temperature data has a weight of 0.4, the data from the previous second has a weight of 0.3, and the data from the previous two seconds has a weight of 0.3. When the data quality indicator field is 2, the standard deviation of the sliding average filter result and the original data of the temperature data must be less than, for example, 1°C, or the relative error between the weighted median filter result and the original data must be less than, for example, 3%. If these conditions are not met, the conflict identification module refuses to process the data and triggers a sensor abnormality alarm.

[0027] Identify whether there is a reverse fluctuation in the sensor data corresponding to the leakage of at least two hazardous chemicals in the multi-source monitoring data, and the specific implementation is as follows: Extract the gas concentration data and temperature data within a continuous time window, and calculate the differential fluctuation value of the gas concentration data and temperature data within the time window. The length of the continuous time window is set according to the process characteristics of the hazardous chemical production link. For example, for a scenario where gas leakage diffuses quickly, the time window length is set to 30 seconds; for a scenario where diffusion is slow, the time window length is set to 5 minutes.

[0028] The calculation method of the differential fluctuation value is as follows: Calculate the difference between adjacent sampling points within the time window for the gas concentration data and temperature data respectively, and accumulate the absolute values to obtain the gas concentration differential fluctuation value and the temperature differential fluctuation value. For example, if there are 60 sampling points in the 30-second window for the gas concentration data, calculate the sum of the absolute values of the concentration differences between the 2nd to the 60th sampling points and the previous sampling point as the gas concentration differential fluctuation value.

[0029] If the differential fluctuation value exceeds the preset safety threshold of the corresponding hazardous chemical, it is determined that there is a reverse fluctuation trend in the gas concentration data and temperature data. The setting method of the preset safety threshold is as follows: For each hazardous chemical, collect historical data in the non-leakage state, and statistically calculate the maximum value of its gas concentration differential fluctuation value and temperature differential fluctuation value, and take 1.5 times the maximum value as the preset safety threshold of the hazardous chemical. For example, if the maximum gas concentration differential fluctuation value of a certain hazardous chemical without leakage is 200 ppm, then its preset safety threshold is set to 300 ppm. When the real-time gas concentration differential fluctuation value exceeds 300 ppm and the temperature differential fluctuation value exceeds the corresponding threshold, it is determined that there is a reverse fluctuation trend.

[0030] Judge whether the reverse fluctuation condition is met according to whether the fluctuation directions of the gas concentration data and temperature data are opposite. The judgment method of the fluctuation direction is as follows: Within the time window, if the gas concentration data rises with time while the temperature data drops with time, or the gas concentration data drops with time while the temperature data rises with time, it is determined that the fluctuation directions are opposite. For example, if the gas concentration data rises from 100 ppm to 150 ppm within 5 consecutive sampling points, and at the same time the temperature data drops from fifty °C to forty-five °C, it is determined as a reverse fluctuation; if both rise or fall synchronously, the reverse fluctuation condition is not met.

[0031] If the reverse fluctuation condition is met, associate the hazardous chemical combination corresponding to the gas concentration data and temperature data to generate a reverse fluctuation event identifier. The association method is as follows: According to the process parameters of the hazardous chemical storage tank or pipeline, match the hazardous chemical combinations that may leak simultaneously. For example, if the gas concentration sensor detects abnormal chlorine concentration and the temperature sensor detects abnormal temperature in the adjacent sulfuric acid storage tank area, then associate the hazardous chemical combination of chlorine and sulfuric acid to generate a reverse fluctuation event with the identifier "chloric acid leakage".

[0032] Based on the equipment operation status data and environmental parameter data, exclude the pseudo-reverse fluctuation events caused by equipment startup or shutdown or environmental mutation. The equipment operation status data includes the current signal and vibration signal of the pipeline pump. If the current value suddenly increases from 0 amperes to the rated value, such as 10 amperes, within the time window and the vibration signal shows an instantaneous peak, it is determined as the equipment startup stage. At this time, the reverse fluctuation event may be caused by the pump startup and shutdown disturbance, and this event needs to be excluded. The environmental parameter data includes environmental temperature and wind speed. If the environmental temperature mutates by more than, for example, 5 degrees Celsius within the time window, or the wind speed suddenly increases by more than, for example, 3 meters per second, it is determined as environmental mutation interference, and the associated reverse fluctuation event is excluded.

[0033] In the calculation of the differential fluctuation value, the sampling frequencies of the gas concentration data and temperature data are kept consistent. For example, both the gas concentration data and temperature data are collected at a frequency of 1 time per second to ensure the time alignment of the differential fluctuation value calculation. If the sampling frequencies of the two types of data are inconsistent, the low-frequency data is aligned to the time stamp of the high-frequency data through an interpolation algorithm. For example, the temperature data is collected once per second, and the gas concentration data is collected twice per second, then the temperature data is interpolated at the 0.5-second position to generate a sampling point sequence synchronized with the gas concentration data.

[0034] In the determination of the fluctuation direction, if the fluctuation directions of the gas concentration data and temperature data alternate multiple times within the time window, a majority voting mechanism is adopted. For example, within a 30-second time window, the gas concentration data first rises and then falls, and the temperature data first falls and then rises. Then, the proportion of the duration of the rising or falling trend is counted. If the total duration proportion of the gas concentration rising exceeds 60% and the total duration proportion of the temperature falling exceeds 60%, it is still determined as a reverse fluctuation.

[0035] When associating hazardous chemical combinations, if there are multiple groups of potentially associated hazardous chemical combinations in the same area, determine the priority according to the process flow diagram. For example, if chlorine and ammonia are stored in the same area and there is no cross-connection in their pipelines, then the hazardous chemical combinations in the same process link are preferably associated; if the relevance cannot be determined, multiple reverse fluctuation event identifiers are generated for manual review.

[0036] When excluding pseudo reverse fluctuation events, the determination of the equipment start-stop state needs to be combined with historical operation logs. For example, if the time point of the sudden change in the pump current signal matches the preset equipment maintenance plan, it is determined as a legal start-stop operation; otherwise, it is regarded as an abnormal disturbance. The determination of sudden changes in environmental temperature needs to refer to meteorological data. If the change in environmental temperature is consistent with the real-time meteorological station data, it is regarded as a natural environmental change, and related events are excluded.

[0037] Based on the topological structure of the production pipeline network and real-time flow data, reverse calculate the location of the leakage source to verify whether the reverse fluctuation is caused by the path occlusion from the leakage source to the sensor. The specific implementation is as follows: According to the topological structure model of the production pipeline network and real-time flow data, calculate the flow change rate of each pipeline node, and screen out candidate abnormal nodes whose flow change rate exceeds the dynamic threshold. The dynamic threshold is set according to the peak value of the flow change during the normal start-stop process in historical operating conditions data. For example, during the normal start-up process of a pump, the maximum value of the flow change rate is an increase of 100 liters per second, then the dynamic threshold is set to 120 liters / second; if the real-time flow change rate of a pipeline node exceeds this threshold, it is marked as a candidate abnormal node. Historical operating conditions data includes equipment start-stop records, valve operation logs, and corresponding flow change curves. By statistically analyzing the peak value distribution of flow changes under normal operations, determine the confidence interval of the dynamic threshold. For example, the start-up records of a pump in a chemical plant show that the maximum fluctuation of the flow change rate during normal start-stop is from 80 liters to 120 liters per second, then the dynamic threshold can be set to 144 liters / second (i.e., 1.2 times of 120 liters).

[0038] Combining the historical leakage frequency of candidate abnormal nodes with the pressure mutation directions of their upstream and downstream pipelines, assign leakage source probability weights, and generate a candidate list of leakage sources sorted by weight. The historical leakage frequency is extracted from the factory operation and maintenance database. For example, if a certain node has had 3 leakage events in the past year, its leakage frequency weight is 0.3; the pressure mutation directions of the upstream and downstream pipelines are judged through pressure sensor data. If the leakage point is upstream of the node, the upstream pressure drops and the downstream pressure rises, and the weight of the pressure mutation direction at this time is 0.5; if the leakage point is downstream of the node, the upstream pressure rises and the downstream pressure drops, and the weight is 0.2. Add the leakage frequency weight and the pressure mutation direction weight to generate the total weight of the candidate leakage source, and sort it from high to low according to the total weight. For example, if the leakage frequency weight of a candidate node is 0.3 and the pressure mutation direction weight is 0.5, the total weight is 0.8, and it is ranked higher in the list.

[0039] Based on real-time wind speed data and the diffusion rate of hazardous chemicals, correct the diffusion path direction of the leakage source candidate, and generate a theoretical diffusion trajectory that matches the sensor position. The real-time wind speed data is obtained through a weather station deployed in the factory area. For example, the current wind speed is 2 meters per second, and the wind direction is south-southeast. The diffusion rate of hazardous chemicals is determined by looking up the table according to its volatility and density. For example, the diffusion rate of chlorine gas is 0.5 meters per second. According to the vector superposition result of the wind speed and the diffusion rate, correct the diffusion path of the leakage source candidate. For example, if the leakage source is located in the northwest direction of the sensor and the wind speed will push the leaked substance to diffuse southeast, the theoretical diffusion trajectory will deflect southeast. For liquefied gas hazardous chemicals, the diffusion rate also needs to be adjusted according to the real-time temperature. For example, when the ambient temperature rises from 25°C to 35°C, the diffusion rate of liquefied petroleum gas increases from 0.6 meters per second to 0.8 meters per second.

[0040] According to the physical layout data of the equipment around the target sensor, detect whether there are obstacles on the path of the theoretical diffusion trajectory, and calculate the expected concentration attenuation range of the leaked substance reaching the sensor according to the geometric size of the obstacle and the diffusion characteristics of the hazardous chemical. The physical layout data is extracted from the factory's 3D modeling system. For example, there is a storage tank with a height of 3 meters, a width of 5 meters, and a carbon steel material in front of a certain sensor; the diffusion characteristics of hazardous chemicals include the bypass coefficient and the adsorption rate. For example, the adsorption rate of chlorine gas on the carbon steel surface is 5%. Calculate the cross-sectional area of the diffusion channel after bypass according to the height and width of the obstacle, and combine the adsorption rate and the diffusion rate. The expected concentration attenuation range is 30% to 50% of the initial concentration. If the obstacle is a porous structure (such as a grille), the attenuation range is further adjusted according to the porosity. For example, a grille with a porosity of 60% can expand the attenuation range to 40%-70%.

[0041] If the actual sensor signal fluctuation amplitude is less than the lower limit of the expected concentration attenuation range, determine that the reverse fluctuation is caused by path obstruction, and mark this leakage source candidate as low confidence. For example, the theoretical expected concentration attenuation lower limit is 30%, but the actual gas concentration fluctuation detected by the sensor is only 20% of the initial concentration. Then it is determined that the signal attenuation of the leaked substance is caused by path obstruction, and the credibility of this leakage source candidate is reduced and marked as low confidence. The low-confidence mark is used for priority exclusion or manual review in subsequent steps. For the candidate nodes marked as low confidence, the system automatically triggers a sensor calibration command, such as increasing the sampling frequency or starting cross-validation of redundant sensors.

[0042] Eliminate candidate abnormal nodes that match the valve switching log, where the valve switching log includes the operation time and the record of valve opening changes. For example, if the time point when the flow rate change rate of a candidate abnormal node exceeds the dynamic threshold is 10:00 am, and the valve switching log shows that a valve closing operation was performed between 9:58 am and 10:02 am, and the valve opening decreased from 100% to 0%, it is determined that the flow anomaly of this node is caused by a legal valve operation and is eliminated from the candidate list. The matching rule for the valve switching log is: If the overlap between the flow change time and the valve operation time exceeds 80%, and the flow change trend is consistent with the valve operation direction (such as valve closing resulting in a decrease in flow), it is regarded as a legal operation interference. For example, a valve decreased from 50% opening to 0% between 10:00:00 and 10:00:30, and the flow data showed a decrease from 200 liters per second to 0 liters per second between 10:00:10 and 10:00:40. The time overlap ratio is 66.7% (20 seconds / 30 seconds), which does not reach the 80% threshold, so this candidate node is not eliminated.

[0043] When calculating the leakage source probability weight, if a candidate abnormal node has no historical leakage record, an initial weight is assigned according to the average leakage frequency of adjacent nodes. For example, a newly commissioned node has no historical data, and the average leakage frequency of its adjacent nodes in the past year is 0.2 times, then the leakage frequency weight of this node is assigned 0.2. The range of adjacent nodes is defined as the directly connected upstream and downstream nodes in the topological structure model. For the pressure mutation direction weight, if the upstream and downstream pressure sensor data is missing, the pressure change trend is simulated according to the pipeline design pressure parameter. For example, the design pressure of a pipeline is 1 MPa. When leaking, the simulated value of the upstream pressure drops to 0.8 MPa, and the downstream pressure rises to 1.2 MPa, and the weight is assigned accordingly.

[0044] When correcting the diffusion path direction, if the diffusion rate of the hazardous chemical is significantly affected by temperature, for example, the evaporation rate of liquefied petroleum gas increases at high temperatures, the diffusion rate is adjusted according to the real-time temperature data. For example, the current temperature is 30 °C, and the diffusion rate of liquefied petroleum gas is corrected to 0.6 meters per second by looking up the table; if the temperature rises to 40 °C, the rate is adjusted to 0.8 meters per second. The real-time temperature data is provided by temperature sensors deployed near the candidate leakage source nodes. For the superposition effect of the diffusion direction and the wind speed, the vector synthesis method is used to calculate the synthesis direction. For example, the wind speed is 2 meters per second in the southeast direction, and the diffusion rate is 0.5 meters per second in the due east direction, then the synthesis direction is east by south, and the synthesis rate is 2.06 meters per second.

[0045] When detecting obstacles on the path of the theoretical diffusion trajectory, if the obstacle is a dynamic device (such as a movable hoisting machine), the occlusion state is updated according to the real-time position data of the device. For example, a certain crane stays on the theoretical diffusion path from 10:00 to 10:15. The system automatically adds the crane to the obstacle list during this period and calculates the impact of its size (length 10 meters, width 4 meters, height 6 meters) on the leaked substance. If the crane moves out of the path at 10:15, it is removed from the list and the concentration decay range is recalculated.

[0046] It should be noted that based on the topological structure of the production pipeline network and the real-time flow data, the leakage source location is inversely calculated, the reverse fluctuation caused by the path occlusion is verified, the leakage diffusion path is accurately simulated through the pipeline network topology, and the real-time flow is combined to dynamically correct the model, breaking through the limitation of the traditional method relying on a single sensor signal. For example, the real-time flow data can locate the flow anomaly node caused by the leakage, and the topological structure determines the leakage diffusion direction. The combination of the two can quickly narrow down the range of the leakage source. Compared with the existing technology, this method solves the problem of fuzzy leakage source location in complex pipeline networks, improves the location accuracy, reduces the false alarm rate, and realizes the collaborative application of structural data and dynamic parameters.

[0047] Figure 2 The flow chart for verifying the combined signal cancellation characteristics of hazardous chemicals and judging the matching of the diffusion path of the present invention is given; when the path occlusion is excluded, it is analyzed whether the combined hazardous chemicals corresponding to the reverse fluctuation have the signal cancellation characteristics of the sensor, and it is verified whether the diffusion path of the hazardous chemicals matches the sensor layout. The specific implementation is as follows: Based on the reverse fluctuation event identifier, the dynamic characteristic library of the physical and chemical reactions of the combined hazardous chemicals is retrieved, and it is matched whether the real-time phase difference between the gas concentration and the temperature change conforms to the time delay law of the synchronous neutralization reaction. The dynamic characteristic library of physical and chemical reactions is a pre-built database in the factory, which stores the neutralization reaction characteristic data of different combinations of hazardous chemicals, including the corresponding relationship between the gas concentration decrease rate and the temperature increase rate, the reaction delay time, etc. For example, the time difference between the temperature change and the gas concentration decrease in the reaction of hydrochloric acid and sodium hydroxide is 5 seconds. The time difference is obtained by simulating the neutralization reaction process in the laboratory, collecting the time series data of the concentration and temperature sensors, and calculating the peak interval. The matching method is: after normalizing the real-time gas concentration data and temperature data to the same dimension, the sliding window cross-correlation algorithm is used to calculate the position of the maximum correlation of the two sets of data. If the time difference of the maximum correlation is within the error range of, for example, ±1 second from the record in the library, it is considered that the phase difference conforms.

[0048] If the real-time phase difference conforms to the time delay law, it is determined that the hazardous chemical combination has the signal cancellation characteristic of the sensor, and the diffusion path verification is executed. Specifically, it includes: real-time monitoring of the temperature and pressure data at the leakage point, and generating a gas-liquid two-phase diffusion path model when the critical condition of the phase state conversion of the hazardous chemical is reached. The critical condition of the phase state conversion is extracted from the pre-set physical and chemical parameter library. For example, when the temperature of liquid ammonia is higher than -33°C or the pressure is lower than 0.8 MPa, gasification occurs, and a gaseous diffusion path is generated. The generation rule of the gas-phase diffusion path is: according to the real-time wind speed data (such as 2 meters per second) and the gaseous diffusion rate (such as 0.5 meters per second), combined with the terrain and obstacle data, a three-dimensional diffusion fan-shaped area model is generated through an interpolation algorithm, with a coverage angle of, for example, 120 degrees and a radius of, for example, 50 meters. Based on this model, the three-dimensional coordinate data of the sensor layout is imported into the spatial analysis system, and the volume ratio of the high-concentration (such as >50 ppm) area not covered by the sensor in the diffusion area is calculated using Boolean operations.

[0049] Retrieve the diffusion path data of the same hazardous chemical combination from the historical leakage event library, extract the historical average diffusion direction and average rate, and calculate the deviation angle of the current path direction and the rate difference coefficient. The historical leakage event library is stored classified by hazardous chemical type, environmental wind speed, and temperature range. For example, 20 historical path records are stored under the category of "chlorine leakage - wind speed 2 - 4 m / s - temperature 25 - 30°C". Extract the historical average diffusion direction (such as 5 degrees east-south) and average rate (such as 0.6 meters per second) under the category. The current path direction is calculated as follows: a vector direction is generated based on the coordinate difference between the leakage source location and the highest concentration point at the diffusion front. For example, if the leakage source coordinates are (100, 200, 5) and the front-end point is (150, 250, 5), then the current direction is the east-north angle arctan[(250 - 200) / (150 - 100)] = 26.6 degrees. The deviation angle is the vector angle between the current direction and the historical average, such as the angle between 26.6 degrees and 5 degrees is 21.6 degrees. The formula for the rate difference coefficient is (historical average rate - current rate) / historical average rate. For example, (0.6 - 0.2) / 0.6 = 0.67.

[0050] If the proportion of the blind area volume exceeds the preset proportion threshold, or the deviation angle of the current path direction exceeds the preset angle threshold and the rate difference coefficient exceeds the preset difference coefficient, it is determined that the diffusion path of the hazardous chemical does not match the sensor layout. The setting rule of the preset proportion threshold is as follows: according to the safety level of the factory area, the maximum allowable blind area volume ratio in the first-class explosion-proof area is, for example, 5%, and in the second-class area is, for example, 10%. The preset angle threshold is determined by statistically analyzing the maximum direction deviation in historical normal leakage events. For example, it is initially set to 15 degrees. If a historical record with a new deviation of 25 degrees is added and determined to be a legal leakage, the threshold is dynamically updated to 25 degrees. The initial value of the preset difference coefficient is, for example, 0.3, indicating that when the current rate deviates from the historical average by more than 30%, it is regarded as abnormal. For example, when the proportion of the blind area volume is 12% or the direction deviation angle is 21.6 degrees and the difference coefficient is 0.67, both trigger the mismatch determination.

[0051] If the real-time phase difference does not conform to the time delay law, verify whether the signal cancellation characteristic is caused by the attenuation of sensor sensitivity or the adsorption effect of hazardous chemicals. The adsorption coefficient is obtained from the physical and chemical parameter library and determined by testing the adsorption rate of hazardous chemicals on the surface of the sensor sensitive material in the laboratory. For example, the adsorption rate of chlorine on the surface of the activated carbon sensor is 10%. The test method is as follows: inject chlorine into a closed cavity until the concentration reaches 100 ppm, continuously monitor the sensor reading until it stabilizes at 90 ppm, and calculate the adsorption rate as (100 - 90) / 100 = 10%. The sensor sensitivity attenuation model is constructed based on the device usage duration. For example, calibrate the sensor reference value every 6 months and calculate the percentage of the current sensitivity to the initial value. If the sensitivity of a sensor used for two years decays to 80%, the detection lower limit rises from the initial 50 ppm to 62.5 ppm. If the current effective concentration decays to 90 ppm after adsorption and the effective detection value after superimposing the sensitivity attenuation is 90×0.8 = 72 ppm, which is lower than the adjusted detection lower limit of 62.5 ppm, it is determined that the signal cancellation is caused by the combined effect of adsorption and sensitivity attenuation.

[0052] Dynamically integrate the device maintenance log and the environmental mutation event library to exclude misjudgments of reverse fluctuations caused by device shutdown or environmental parameter jumps. The device maintenance log includes the device start and stop times, valve operation records, and maintenance personnel marking information. For example, the shutdown time of a certain pump is from 10:00 to 10:10. If a reverse fluctuation event occurs from 10:05 to 10:15, the time overlap ratio is 50% (5 minutes / 10 minutes), which does not reach the 80% exclusion threshold, so the event is retained. The environmental mutation event library stores meteorological data (such as the instantaneous wind speed jump value) and device abnormal alarm records. For example, during a certain fluctuation period, it is detected that the wind speed suddenly increases from 2 m / s to 8 m / s. The system automatically associates the meteorological records of this period, marks it as "wind speed mutation interference", and excludes the associated events.

[0053] When calculating the proportion of the blind area volume in the gas-phase diffusion path, if there is a multi-phase mixed diffusion scenario (such as partial vaporization of liquefied petroleum gas and partial liquid retention), the blind area volume is calculated by weighting according to the proportion of the phase change mass. For example, if the gaseous mass proportion is 70%, the diffusion blind area proportion is 70% of the total blind area; the liquid residue area may cover the sensor blind spot (such as the trench area), but since it does not involve gaseous monitoring, it is not included in the diffusion path verification scope of this patent.

[0054] When verifying the deviation angle of the direction, the processing method for the path bifurcation scenario is: perform a weighted average calculation on the directions of the primary and secondary diffusion branches. For example, the weight of the diffusion direction of the primary branch is 0.7, and the weight of the secondary branch is 0.3. Combining with the mean direction of the historical path, ensure that the deviation angle calculation reflects the overall diffusion trend. The weight is determined according to the concentration proportion of the diffusion branch. For example, the concentration proportion of the primary branch is 70%, and the secondary branch is 30%.

[0055] It should be noted that when analyzing the signal cancellation characteristics and verifying the diffusion path matching after excluding path occlusion, multiple factors (chemical reaction, sensor layout) are comprehensively considered to judge the authenticity of the leakage. For example, if the leakage of chlorine and sulfuric acid causes the concentration and temperature to fluctuate in the opposite direction, by verifying whether the diffusion path bypasses the sensor blind spot, the real leakage and interference signals can be distinguished. Compared with the prior art that only relies on threshold alarm, through the double verification of path matching and signal cancellation, the risk of misjudgment is reduced, and the leakage identification accuracy is improved.

[0056] If the hazardous chemical combination has the signal cancellation characteristic and the diffusion path does not match, a leakage signal correction model is generated according to the physical and chemical reaction mechanism to dynamically compensate the multi-source monitoring data. The specific implementation is as follows: Based on the synchronous neutralization reaction characteristics of the hazardous chemical combination and the verification result of the unmatched diffusion path, the corresponding reaction rate equation and diffusion attenuation coefficient in the physical and chemical reaction mechanism library are retrieved. The physical and chemical reaction mechanism library is a pre-set database that stores the chemical reaction rate equation parameters and diffusion attenuation coefficients of different hazardous chemical combinations. For example, the rate equation form of the reaction between hydrochloric acid and sodium hydroxide is a first-order reaction equation, and the rate constant is determined by laboratory measurement to be, for example, 0.05 per second, and the diffusion attenuation coefficient is fitted according to the leakage experiment data to be, for example, 0.2. The synchronous neutralization reaction characteristics include the time sequence relationship between the gas concentration decrease and the temperature increase. For example, for every 10 ppm decrease in concentration, the temperature increases by 0.5 °C. This characteristic is obtained through the data regression analysis of historical leakage events.

[0057] A leakage signal correction model is constructed based on the reaction rate equation and the diffusion attenuation coefficient. The leakage signal correction model includes the superposition relationship between the gas concentration compensation function and the temperature compensation function. The form of the gas concentration compensation function is that the concentration correction amount is equal to the original concentration multiplied by the quantity in the brackets which is one plus the diffusion attenuation coefficient multiplied by the reaction rate. For example, if the original concentration is 100 ppm, the diffusion attenuation coefficient is 0.2, and the reaction rate is 0.05, then the correction amount is 100 multiplied by the quantity in the brackets which is one plus 0.2 multiplied by 0.05, equal to 101 ppm. The form of the temperature compensation function is that the temperature correction amount is equal to the original temperature plus the quantity in the brackets which is the heat of reaction effect multiplied by the reaction rate. For example, if the heat of reaction effect is 50 kJ released per mole and the reaction rate is 0.05, then the temperature correction amount increases by 2.5 °C. The superposition relationship means that the gas concentration correction amount and the temperature correction amount are output in a linear addition manner. For example, the gas correction amount of 101 ppm and the temperature correction amount of 2.5 °C are superposed to generate the final compensation value.

[0058] Based on the real-time wind speed and temperature and humidity data, the weight coefficients of the compensation function are dynamically adjusted to generate the correction amounts of the gas concentration data and the temperature data. The weight coefficients are allocated according to the influence degree of the environmental parameters on diffusion and reaction. For example, when the wind speed exceeds 5 m per second, the weight of the diffusion attenuation coefficient is increased to, for example, 0.8, and the weight of the reaction rate is decreased to, for example, 0.2; when the humidity exceeds 70%, the temperature compensation weight is increased to, for example, 0.6. The adjustment rule of the weight coefficients is determined by a multiple linear regression model. For example, a mapping table of wind speed, humidity and weight coefficients is established, and the weight value is obtained by looking up the table according to the real-time parameters.

[0059] The correction amounts are superposed on the gas concentration values and temperature values of the original multi-source monitoring data to output the compensated gas concentration data and temperature data. For example, if the original gas concentration is 100 ppm and the correction amount is plus 1 ppm, then the compensated concentration is 101 ppm; if the original temperature is 25 °C and the correction amount is plus 2.5 °C, the compensated temperature is 27.5 °C. The superposition operation is performed in real time in the data processor to ensure that the compensated data is synchronized with the sensor sampling frequency. For example, it is compensated once per second.

[0060] When constructing the gas concentration compensation function, if the hazardous chemical combination is a multi-component mixed reaction, such as the neutralization reaction when chlorine and ammonia coexist, the reaction rate weights are allocated according to the concentration ratios of each component. For example, if the chlorine concentration accounts for 60% and the ammonia accounts for 40%, then the reaction rate weights are 0.6 and 0.4 respectively, and the total reaction rate is the weighted sum of the rates of each component. The ratio is calculated in real time from the gas concentration values in the multi-source monitoring data.

[0061] When dynamically adjusting the weight coefficient, if the environmental parameters change suddenly, for example, the wind speed suddenly increases from 2 meters per second to 8 meters per second, the weight is adjusted according to the preset priority rules. For example, when the wind speed changes suddenly, the weight adjustment of the diffusion attenuation coefficient is prioritized, and the reaction rate weight remains constant to avoid drastic fluctuations in the compensation amount. The priority rules are set according to the factory safety level. For example, in the first-level explosion-proof area, the diffusion compensation is prioritized, and in the second-level area, the reaction compensation is prioritized.

[0062] If there is signal cancellation and path mismatch in the hazardous chemical combination, the monitoring data is corrected through the physical and chemical reaction mechanism. For example, when the concentration decays due to a neutralization reaction, the compensation amount is dynamically adjusted according to the reaction rate and diffusion coefficient to restore the true leakage concentration. Compared with the static correction of the existing technology, combining real-time environmental parameters (wind speed, temperature, and humidity) with reaction characteristics improves the data reliability.

[0063] Generate a mixed leakage risk level based on the compensated multi-source monitoring data, and determine whether to trigger a hierarchical emergency response instruction matching the hazardous chemical combination. The specific implementation is as follows: Extract the compensated gas concentration data and temperature data, combine the diffusion path verification results, and generate a mixed leakage risk level according to the diffusion range of the leaked substance and the real-time concentration gradient. The diffusion path verification result is the conclusion of whether the verified diffusion path matches the sensor layout. For example, when the diffusion path is determined to be mismatched, the risk level calculation needs to consider the cumulative effect of potential leakage concentration in the uncovered area. The mixed leakage risk level is generated by multiplying the compensated gas concentration value by the toxicity weight coefficient of the hazardous chemical combination and adding the product of the temperature value and the reaction activity coefficient. The toxicity weight coefficient and reaction activity coefficient are extracted from the pre-set hazardous chemical property library, which stores the toxicity levels, reaction activity parameters, and emergency treatment priority data of different hazardous chemicals. For example, the toxicity weight coefficient of chlorine is 0.8, the reaction activity coefficient is 0.5, the compensated gas concentration is 120 ppm, and the temperature is 30 °C, then the risk level calculation is 120 multiplied by 0.8 plus 30 multiplied by 0.5, and the result is 111.

[0064] Generate the range boundary coordinates of the leakage impact area based on the real-time environmental wind speed and diffusion range. The diffusion range is determined according to the boundary of the uncovered area calculated in the diffusion path verification result. For example, if the fan-shaped coverage angle of the diffusion path is 120 degrees and the radius is 50 meters, combined with the real-time wind speed of 5 meters per second, the three-dimensional coordinate boundary of the diffusion area is generated through the interpolation algorithm. The interpolation algorithm is a conventional method for spatial data processing. Specifically, based on the vector superposition result of the diffusion rate and wind speed, a position coordinate sequence at the front end of the diffusion is generated, and the coordinate sequence is connected to form a closed area boundary. For example, from 116.3 degrees to 116.5 degrees east longitude, from 39.9 degrees to 40.1 degrees north latitude, and from 0 to 10 meters above sea level.

[0065] Input the mixed leakage risk level and the boundary coordinates of the affected area range into the pre-stored hierarchical emergency response mapping table, and output the corresponding emergency response instruction type. The hierarchical emergency response mapping table is a database pre-stored in the factory, storing the corresponding relationships between different hazardous chemical combinations, risk levels, affected area ranges, and emergency response instructions. For example, when the chlorine leakage risk level exceeds 100 and the affected area includes the production workshop, the mapping instruction is to immediately isolate the leakage source and evacuate the workshop personnel. The construction rule of the mapping table is generated based on the handling experience of historical leakage events and safety specifications. For example, the emergency measures for each type of hazardous chemical combination are formulated by the factory safety department based on accident review data.

[0066] If the mixed leakage risk level exceeds the preset safety threshold, trigger the emergency response instruction matching the hazardous chemical combination type. The preset safety threshold is dynamically set according to the safety level of the factory area. For example, the safety threshold for the first-class explosion-proof area is 50, and for the second-class area is 80. If the calculated mixed leakage risk level is 111 and the leakage area is the first-class area, it is determined to exceed the threshold. The triggering logic of the emergency response instruction is to call the pre-stored instruction template according to the hazardous chemical combination type. For example, when chlorine and ammonia are combined and leaked, the instruction template includes closing the connecting valves, injecting neutralizing agents, and evacuating the personnel within a radius of 200 meters, and the execution order is adjusted according to the real-time risk level. For example, when the risk level is higher than 100, the valve closing is given priority.

[0067] Adjust the execution priority of the emergency response instruction according to the equipment operation status data. The equipment operation status data includes the real-time operation status of pumps, valves, and ventilation equipment. For example, if the pump associated with the leakage source is in the running state, the pump shutdown instruction is given priority. The priority adjustment rule is to allocate execution weights according to the impact degree of the equipment state on the leakage diffusion. For example, the operation of the pump will accelerate the leakage diffusion, and the weight coefficient of its shutdown operation is set to 0.9, and the weight coefficient of the ventilation equipment startup is set to 0.5. The system executes the instructions in descending order of weights.

[0068] When generating the boundary coordinates of the affected area of the leakage, if there are multiple branches in the diffusion path, such as diffusion bifurcation due to obstacles, calculate the affected areas of each branch separately and merge them into the overall range. For example, the main branch diffusion direction is due east, and the affected area is from 116.3 degrees to 116.4 degrees east longitude, and the secondary branch direction is northeast, and the affected area is from 116.35 degrees to 116.45 degrees east longitude. The merged range coordinates are from 116.3 degrees to 116.45 degrees east longitude. The merging method is to take the minimum bounding rectangle of the coordinates of each branch. For example, calculate the maximum and minimum values of the longitude and latitude of all branch coordinates to generate a rectangular area containing all branches.

[0069] The calculations involved in the embodiments are all dimensionless numerical calculations, and the preset parameters and threshold selections in the calculations are set by those skilled in the art according to the actual situation.

[0070] It should be noted that the present invention can be deployed on the device itself to implement embedded applications, or can also run on a PC or other terminal with a user interface, so as to meet various hardware environments and usage requirements.

[0071] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from a website, computer, server, or data center to another website, computer, server, or data center by wire (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains a set of one or more available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0072] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices, and modules described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0073] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there can be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings, direct couplings, or communication connections shown or discussed with each other can be through some interfaces, and the indirect couplings or communication connections of the devices or modules can be in electrical, mechanical, or other forms.

[0074] The module described as a separation component may or may not be physically separated. The component shown as a module may or may not be a physical module. It may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0075] In addition, in each embodiment of this application, each functional module can be integrated into a processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0076] If the described function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of this application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0077] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0078] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A safety production risk monitoring and dynamic early warning system for hazardous chemicals, characterized in that, It includes the following modules: The data acquisition module is used to obtain multi-source monitoring data in the production environment of hazardous chemicals in real time, including gas concentration data and temperature data; The conflict identification module is used to identify whether the sensor data corresponding to the leakage of at least two hazardous chemicals in the multi-source monitoring data has a reverse fluctuation exceeding the preset safety threshold; The leakage verification module is used to reverse calculate the leakage source location based on the topological structure of the production pipeline network and the real-time flow data, and verify whether the reverse fluctuation is caused by the path occlusion from the leakage source to the sensor; The physical and chemical analysis module is used to analyze whether the hazardous chemical combination corresponding to the reverse fluctuation has the signal cancellation characteristic of the sensor and verify whether the diffusion path of the hazardous chemical matches the sensor layout when the path occlusion is excluded; The dynamic compensation module is used to generate a leakage signal correction model according to the physical and chemical reaction mechanism and perform dynamic compensation on the multi-source monitoring data if the hazardous chemical combination has the signal cancellation characteristic and the diffusion path does not match; The risk assessment module is used to generate a mixed leakage risk level based on the compensated multi-source monitoring data and determine whether to trigger a hierarchical emergency response instruction matching the hazardous chemical combination.

2. The hazardous chemical production safety risk monitoring and dynamic early warning system according to claim 1, wherein Obtain multi-source monitoring data in the production environment of hazardous chemicals in real time, including gas concentration data and temperature data, including: Deploy gas concentration sensors and temperature sensors at different process nodes in the production area of hazardous chemicals; Align the monitoring data of the gas concentration sensors and temperature sensors in real time through the timestamp synchronization protocol to generate time-synchronized multi-source monitoring data; Perform noise filtering processing and filtering processing on the time-synchronized multi-source monitoring data; Store the calibrated and filtered multi-source monitoring data in the real-time database.

3. The hazardous chemical production safety risk monitoring and dynamic early warning system according to claim 1, characterized in that, Identify whether the sensor data corresponding to the leakage of at least two hazardous chemicals in the multi-source monitoring data has a reverse fluctuation exceeding the preset safety threshold, including: Extract the gas concentration data and temperature data within a continuous time window, and calculate the differential fluctuation value of the gas concentration data and temperature data within the time window; If the differential fluctuation value exceeds the preset safety threshold of the corresponding hazardous chemical, it is determined that the gas concentration data and temperature data have a reverse fluctuation trend; Judge whether the reverse fluctuation condition is met according to whether the fluctuation directions of the gas concentration data and temperature data are opposite; If the reverse fluctuation condition is met, associate the hazardous chemical combination corresponding to the gas concentration data and temperature data, and generate a reverse fluctuation event identifier; Exclude pseudo-reverse fluctuation events caused by equipment startup and shutdown or environmental mutations based on equipment operation status data and environmental parameter data.

4. The hazardous chemical production safety risk monitoring and dynamic early warning system according to claim 1, wherein Reverse calculate the leakage source location based on the topological structure of the production pipeline network and the real-time flow data, and verify whether the reverse fluctuation is caused by the path occlusion from the leakage source to the sensor, including: According to the topological structure model of the production pipeline network and the real-time flow data, calculate the flow change rate of each pipeline node, and screen the candidate abnormal nodes whose flow change rate exceeds the dynamic threshold; Combine the historical leakage frequency of the candidate abnormal nodes and the pressure mutation direction of the upstream and downstream pipelines, assign leakage source probability weights, and generate a leakage source candidate list sorted by weight; Based on real-time wind speed data and the diffusion rate of hazardous chemicals, correct the diffusion path direction of leakage source candidates, and generate a theoretical diffusion trajectory that matches the sensor location; According to the physical layout data of the devices around the target sensor, detect whether there are obstacles on the path of the theoretical diffusion trajectory, and calculate the expected concentration attenuation range of the leaked substance reaching the sensor based on the geometric size of the obstacles and the diffusion characteristics of the hazardous chemicals; If the actual sensor signal fluctuation amplitude is less than the lower limit of the expected concentration attenuation range, determine that the reverse fluctuation is caused by path occlusion, and mark the corresponding leakage source candidate as low confidence.

5. The hazardous chemical production safety risk monitoring and dynamic early warning system according to claim 1, characterized in that, When excluding the cause of path occlusion, analyze whether the hazardous chemical combination corresponding to the reverse fluctuation has the signal cancellation characteristic of the sensor, and verify whether the diffusion path of the hazardous chemicals matches the sensor layout, including: When excluding the cause of path occlusion, based on the reverse fluctuation event identifier, retrieve the dynamic physical and chemical reaction characteristic library of the hazardous chemical combination, and match whether the real-time phase difference between the gas concentration and the temperature change conforms to the time delay law of the synchronous neutralization reaction; If the real-time phase difference conforms to the time delay law, determine that the hazardous chemical combination has the signal cancellation characteristic of the sensor, and perform diffusion path verification to determine whether the diffusion path of the hazardous chemicals matches the sensor layout: If the real-time phase difference does not conform to the time delay law, verify whether the signal cancellation characteristic is caused by sensor sensitivity attenuation or hazardous chemical adsorption effect; Dynamically fuse the device maintenance log and the environmental mutation event library to exclude misjudgment of reverse fluctuations caused by device shutdown or environmental parameter jumps.

6. The hazardous chemical production safety risk monitoring and dynamic early warning system according to claim 5, wherein, Performing diffusion path verification to determine whether the diffusion path of the hazardous chemicals matches the sensor layout specifically means: Real-time monitor the temperature and pressure data of the leakage point. When the critical condition of hazardous chemical phase transition is reached, generate a gas-liquid two-phase diffusion path model, and calculate the proportion of the blind area volume not covered by the sensor in the gas-phase diffusion path; Retrieve the diffusion path data of the same hazardous chemical combination in the historical leakage event library, extract the average historical diffusion direction and the average rate, and calculate the deviation angle of the current path direction and the rate difference coefficient; If the proportion of the blind area volume exceeds the preset proportion threshold, or the deviation angle of the current path direction exceeds the preset angle threshold and the rate difference coefficient exceeds the preset difference coefficient, determine that the diffusion path of the hazardous chemicals does not match the sensor layout.

7. The hazardous chemical production safety risk monitoring and dynamic early warning system according to claim 5, characterized in that, Verifying whether the signal cancellation characteristic is caused by sensor sensitivity attenuation or hazardous chemical adsorption effect specifically means: According to the adsorption coefficient of the hazardous chemical combination and the sensor sensitivity attenuation model, calculate the attenuation amplitude of the effective concentration relative to the original leakage concentration; If the attenuation amplitude exceeds the lower limit threshold of the sensor detection dynamic range, determine that the signal cancellation characteristic is caused by adsorption or sensitivity attenuation.

8. The hazardous chemical production safety risk monitoring and dynamic early warning system according to claim 1, characterized in that, If the hazardous chemical combination has the signal cancellation characteristic and the diffusion path does not match, generate a leakage signal correction model according to the physical and chemical reaction mechanism, and perform dynamic compensation on the multi-source monitoring data, including: Based on the synchronous neutralization reaction characteristics of the hazardous chemical combination and the verification result of the unmatched diffusion path, retrieve the corresponding reaction rate equation and diffusion attenuation coefficient in the physical and chemical reaction mechanism library; Construct a leakage signal correction model based on the reaction rate equation and the diffusion attenuation coefficient. The leakage signal correction model includes the superposition relationship between the gas concentration compensation function and the temperature compensation function; Based on the real-time wind speed, temperature and humidity data, dynamically adjust the weight coefficients of the compensation functions to generate the correction amounts of the gas concentration data and the temperature data; Superimpose the correction amounts on the gas concentration value and the temperature value of the original multi-source monitoring data, and output the compensated gas concentration data and temperature data.

9. The hazardous chemical production safety risk monitoring and dynamic early warning system according to claim 1, wherein, Generate a mixed leakage risk level based on the compensated multi-source monitoring data, and determine whether to trigger a hierarchical emergency response instruction matching the hazardous chemical combination, including: Extract the compensated gas concentration data and temperature data, combine the diffusion path verification results, and generate a mixed leakage risk level according to the diffusion range and the real-time concentration gradient of the leaked substance; Generate the range boundary coordinates of the leakage impact area based on the real-time environmental wind speed and the diffusion range; Input the mixed leakage risk level and the range boundary coordinates of the impact area into the pre-stored hierarchical emergency response mapping table, and output the corresponding emergency response instruction type; If the mixed leakage risk level exceeds the preset safety threshold, trigger an emergency response instruction matching the hazardous chemical combination type; Adjust the execution priority of the emergency response instruction according to the equipment operation status data.

Images