A method and system for detecting chlorine leakage in a reaction kettle production

Abstract

The present application relates to the technical field of chemical gas, and particularly relates to a method and system for detecting chlorine leakage in the production of a reaction kettle. The method comprises: obtaining chlorine concentration values and wind field vectors of multiple positions around the reaction kettle at each time; taking any time as a target time, determining a gas diffusion contribution coefficient of each position to the remaining positions at the target time; determining a real gas diffusion contribution coefficient of each position to the remaining positions at the target time; determining a gas flow correction density of each position at the target time; and based on the gas flow correction densities of all positions at the target time, locating a chlorine leakage source by using a density peak value clustering algorithm. The present application corrects the density by introducing the wind field vector to construct the gas diffusion contribution coefficient, overcomes the limitations of the traditional static diffusion, effectively distinguishes the real leakage source and the concentration collection point downwind, improves the accuracy and reliability of the chlorine leakage source positioning, and provides support for safety emergency disposal.

Description

Technical Field

[0001] This invention relates to the field of chemical gas technology, and in particular to a method and system for detecting chlorine leaks in a reaction vessel. Background Technology

[0002] In the production processes of many fine chemicals, such as the synthesis of 2-amino-4-methylbenzothiazole, the chlorination reaction by introducing chlorine gas into the reactor is an indispensable core process. However, the chemical production environment is complex, and during long-term operation, chlorine gas leaks may occur at weak points such as flanges, valves, and welds due to factors such as equipment aging, seal fatigue failure, or improper operation. Chlorine is a yellowish-green, highly toxic gas with a strong pungent odor and is also highly corrosive. Once chlorine gas leaks from the reactor into the workshop environment, it not only poses an immediate and serious threat to the lives of on-site operators but may also corrode precision equipment, triggering a chain reaction of accidents, causing production stoppages, and resulting in huge economic losses. Therefore, achieving rapid and accurate detection and source location of chlorine gas leaks is of paramount importance for ensuring personnel safety, protecting production facilities, and preventing environmental pollution.

[0003] Currently, a common approach to leak source localization is to deploy a monitoring network consisting of multiple gas sensors in potential leak areas, such as around a reactor. When a sensor in the network detects a target gas, a specific algorithm analyzes the concentration data from each sensor to infer the location of the leak source. The Density Peak Clustering (DPC) algorithm is used for leak source localization because it does not require a pre-defined number of clusters and can effectively identify clusters of arbitrary shapes.

[0004] However, traditional density peak clustering algorithms, when locating leak sources, only consider the concentration readings of gas sensors at each location and the Euclidean spatial distance between them. This inherently assumes that the gas diffuses uniformly in an ideal, static space with no wind and no obstructions. But in real chemical plant workshops, due to factors such as forced ventilation systems, equipment obstructions, and hot air convection, there is a complex and intense airflow field, i.e., a wind field. The wind field will greatly change the diffusion path of chlorine gas, causing the highest concentration point measured to often be the gas convergence point downwind of the chlorine leak source, rather than the actual physical leak source. Because traditional density peak clustering algorithms fail to incorporate the key environmental information of the wind field to correct their judgment of the correlation between data points, they will systematically misidentify the downwind concentration peak point as the leak source, leading to deviations in the location results, delaying emergency repairs, and increasing safety risks. Summary of the Invention

[0005] To address the problem of inaccurate chlorine leak location caused by the traditional density peak clustering algorithm ignoring the influence of wind field, this invention provides a chlorine leak detection method and system in reactor production.

[0006] In a first aspect, the present invention provides a method for detecting chlorine leakage in a reaction vessel production process, employing the following technical solution:

[0007] A method for detecting chlorine leaks in a reactor production process includes: acquiring chlorine concentration values ​​and wind field vectors at multiple locations around the reactor at various times; designating any location as a target location, any other location not including the target location as a reference location, and any time as a target time; acquiring the difference vector between the spatial coordinate vectors of the target location and the reference location; determining the gas diffusion contribution coefficient of the target location to the reference location at the target time based on the angle between the wind field vector of the target location at the target time and the difference vector, and the spatial coordinate vectors of the target location and the reference location; filtering the gas diffusion contribution coefficients of the target location to the reference location at multiple times including the target time to obtain the true gas diffusion contribution coefficient of the target location to the reference location at the target time; determining the corrected airflow density of the target location at the target time based on the chlorine concentration values ​​at all locations at the target time, and the gas diffusion contribution coefficients of the target location at the target time to all other locations not including the target location; and locating the chlorine leak source based on the corrected airflow density of all locations at the target time using a density peak clustering algorithm.

[0008] The beneficial effects are as follows: By introducing the angular relationship between the spatial difference vector between the target location and the reference location and the wind field vector, an evaluation model for the gas diffusion contribution coefficient is constructed, which can effectively reflect the influence of the wind field on the chlorine diffusion path, fully consider the dominant role of the complex airflow field in actual chemical workshops on gas diffusion, and overcome the limitations of traditional methods that assume static uniform diffusion; by eliminating noise interference through filtering, the true gas diffusion contribution coefficient is obtained, improving the accuracy and stability of the contribution coefficient calculation; the density peak clustering algorithm based on airflow-corrected density achieves more accurate location of leakage sources in wind field environments; it effectively distinguishes between the true leakage source point and the downwind concentration convergence point, avoiding the systematic error of traditional algorithms misjudging high concentration points as leakage sources; by integrating wind field information to correct the correlation judgment between data points, the accuracy and reliability of leakage source location are improved, providing support for rapid emergency response and safety risk control of chlorine leaks.

[0009] Furthermore, the chlorine concentration value is collected using a chlorine sensor based on the principle of ultraviolet absorption spectroscopy.

[0010] Furthermore, the wind field vector is acquired using a three-dimensional ultrasonic anemometer.

[0011] Furthermore, the gas diffusion contribution coefficient satisfies:

[0012] In the formula, For position At any moment For location The gas diffusion contribution coefficient, For position At any moment Wind field vector and difference vector The angle between them and Positions With position spatial coordinate vector, For position At any moment The wind field vector, For hyperparameters, For the standard normalized function, It is a natural exponential function. This is the modulus symbol.

[0013] The beneficial effects are as follows: by constructing a composite function that includes the wind direction angle factor and the spatial distance attenuation factor, a scientific evaluation of the gas diffusion contribution coefficient is achieved. The angle factor can effectively reflect the promoting effect of wind direction on diffusion. The contribution is the largest when the wind direction is consistent with the position difference vector, the contribution is moderate when it is perpendicular, and the contribution is the smallest when they are opposite. The distance attenuation factor reflects the law that gas diffusion attenuates with increasing distance. At the same time, it takes into account the influence of wind speed on diffusion distance. The greater the wind speed, the slower the attenuation and the wider the diffusion influence range.

[0014] Furthermore, the filtering process includes: sorting the data within a sliding time window, which contains the gas diffusion contribution coefficients of the target location to the reference location at the target time and several times prior to the target time, according to their numerical values; and selecting the median of the sorted sliding time window as the true gas diffusion contribution coefficient of the target location to the reference location at the target time.

[0015] The beneficial effects are as follows: By processing the gas diffusion contribution coefficient within the sliding time window through median filtering, the interference of instantaneous noise and outliers can be effectively eliminated, improving the stability and reliability of contribution coefficient calculation. Median filtering has good denoising ability while maintaining edge characteristics, avoiding the data smoothing over-smoothing problem that may occur with mean filtering. The sliding time window mechanism makes full use of historical data information. Through statistical analysis of data at multiple time points, it can more accurately reflect the true trend of gas diffusion contribution. The selection of the median ensures that the filtering result is not affected by extreme values, improving the robustness of the algorithm.

[0016] Furthermore, the length of the sliding time window is 9.

[0017] Furthermore, the airflow correction density satisfies:

[0018] In the formula, For position At any moment Airflow correction density, For position At any moment The chlorine concentration value, For position At any moment The chlorine concentration value, For position At any moment For location The real gas diffusion contribution coefficient.

[0019] The beneficial effects are as follows: by combining the current location concentration value with the weighted sum of the concentration contributions from other locations, a comprehensive evaluation model for airflow correction density is constructed, which fully considers the directionality and transmissibility of gas diffusion under the influence of the wind field and effectively reflects the true correlation between each monitoring point; the current location concentration value reflects the direct measurement information of that point, while the concentration contributions from other locations reflect the influence of upstream points on downstream points under the influence of the wind field. By weighting through the real gas diffusion contribution coefficient, the accuracy of the contribution calculation is ensured.

[0020] Furthermore, the method for locating the chlorine leak source includes: obtaining the Euclidean distance between the target location and all locations with a gas flow correction density greater than the target location, and recording the minimum value among the Euclidean distances as the minimum distance of the target location at the target time; based on the gas flow correction density of all locations at the target time and the minimum location, searching for outliers in the decision map generated by the density peak clustering algorithm, and identifying the location corresponding to the outlier as the chlorine leak source.

[0021] The beneficial effects are as follows: By calculating the minimum Euclidean distance between the target location and all high-density locations, an evaluation index for local density distribution characteristics is constructed, which can effectively identify density peak points. It fully utilizes the core idea of ​​the density peak clustering algorithm and uses distance information to assist density information in anomaly detection. Through the anomaly identification mechanism in the decision graph, the precise location of chlorine leak sources is achieved. Combining airflow-corrected density and spatial distribution characteristics, the accuracy and robustness of leak source identification in complex wind field environments are improved. It avoids the problem of traditional methods misjudging downwind high-concentration points as leak sources, and provides reliable support for the safe emergency response to chlorine leaks in chemical plants.

[0022] Furthermore, the method of finding outliers includes: in response to any data point in the decision map having an airflow correction density greater than a preset density threshold at the target time, and the minimum distance of the data point at the target time being greater than a preset distance threshold, identifying the data point as an outlier.

[0023] Secondly, this invention provides a chlorine leak detection system for reactor production, employing the following technical solution:

[0024] A chlorine leak detection system for reactor production includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned chlorine leak detection method for reactor production is implemented.

[0025] By adopting the above technical solution, a computer program for detecting chlorine leakage in a reaction vessel is generated and stored in a memory so that it can be loaded and executed by a processor. A terminal device can then be made based on the memory and processor for convenient use.

[0026] The present invention has the following technical effects:

[0027] (1) Abandoning the ideal assumption of windless and unobstructed static space in the traditional density peak clustering algorithm, the algorithm integrates the key environmental information of wind field vector. By combining the spatial difference vector between the target position and the reference position with the angle between the wind field vector at the target time, the gas diffusion contribution coefficient is calculated, thereby more accurately assessing the influence of the wind field on the chlorine diffusion path. Then, the airflow correction density at the target position is corrected by the real gas diffusion contribution coefficient, which effectively avoids the systematic bias of the traditional algorithm in misjudging the downwind concentration accumulation point as the leakage source. This makes the density peak identified by the density peak clustering algorithm correspond to the real physical leakage source, improving the accuracy of leakage source location in complex scenarios.

[0028] (2) To address the issue of noise generated by industrial field sensor data due to equipment vibration, electromagnetic interference, etc., the gas diffusion contribution coefficient is filtered to eliminate abnormal fluctuation interference, ensuring that the calculated real gas diffusion contribution coefficient is more in line with the actual diffusion law. This allows the airflow correction density obtained based on the real gas diffusion contribution coefficient to more accurately reflect the possibility of each location as a leakage source, avoiding false density peaks caused by noise.

[0029] (3) Fully consider the complex airflow scenarios such as forced ventilation system, equipment obstruction, and hot air convection in chemical plant workshops. Dynamically capture the real-time airflow direction and intensity through wind field vector. Correct the physical obstruction of diffusion path by combining spatial difference vector with equipment layout. So that the algorithm is no longer limited to ideal environment and can stably adapt to the complex airflow field of real industrial scenarios. Solve the problem of traditional algorithm positioning failure in non-ideal environment and improve the industrial applicability of the method.

[0030] (4) By more accurately locating the real leak source, the deviation in the repair direction caused by the misjudgment of traditional algorithms can be avoided, ensuring that maintenance personnel can lock the leak location earlier and quickly start interlocking control measures such as valve closure, leak sealing, and local ventilation. This can effectively shorten the repair response time, reduce the spread of chlorine leaks, and reduce safety hazards such as personnel poisoning, equipment corrosion, and explosion risks, thus providing support for the safety of chemical production.

[0031] (5) The airflow-corrected density replaces the traditional static density calculation based solely on concentration and Euclidean distance, making the density judgment of the density peak clustering algorithm more in line with the physical laws of chlorine diffusion. The real leak source will continuously diffuse to the surrounding area, forming a high density and contributing significantly to the diffusion downstream. In contrast, the false downwind concentration peak is characterized by a high concentration but a low contribution to the diffusion upstream. Therefore, by correcting the density, the real leak source is amplified, making it easier to pinpoint the real leak source through the clustering results and improving the overall accuracy of safety detection. Attached Figure Description

[0032] Figure 1 This is a flowchart of a method for detecting chlorine leakage in a reactor production process according to an embodiment of the present invention.

[0033] Figure 2 This is a decision graph generated by the density peak clustering algorithm in a chlorine leak detection method in a reactor production process according to an embodiment of the present invention. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] This invention discloses a method for detecting chlorine gas leakage in a reactor during production, referring to... Figure 1 This includes steps S1-S5:

[0036] S1: Obtain the chlorine concentration and wind field vector at multiple locations around the reactor at various times.

[0037] First, at key locations around the reactor, such as flange connections, valve groups, welds, sampling ports, and vents, a chlorine sensor based on ultraviolet absorption spectroscopy is deployed at each location. For example, chlorine concentration data is acquired at a sampling frequency of ten times per second. For each sensor, the chlorine concentration value detected at each moment and the three-dimensional spatial coordinate vector are acquired as the data for the corresponding location. Simultaneously, the environmental wind field information of the reactor area is acquired, and a discrete wind field vector model covering the entire monitoring area is constructed. Specifically, a three-dimensional ultrasonic anemometer is pre-installed at each location in all main ventilation ducts, open areas, and around the equipment. It is synchronously measured in real time at the same sampling frequency as the chlorine sensor, and interpolation is used to obtain the wind field vector covering the entire area. The wind field vector at each sensor location at each moment can be queried from the wind field model.

[0038] S2: Record any time as the target time, and determine the gas diffusion contribution coefficient of each position to the other positions at the target time.

[0039] It should be noted that the purpose of this step is to assess the likelihood of leaked gas at any location contributing to the chlorine concentration at another location under the current wind field conditions, in order to eliminate the negative impact of instantaneous environmental disturbances. Therefore, this step will obtain the gas diffusion contribution coefficient of each location to the other locations at each time point based on the numerical performance of the wind field vector data corresponding to different locations. The logic behind this is as follows: if a location is located directly above another location in the wind direction, it indicates that this location is likely to contribute more to the chlorine concentration at the other location, and the corresponding gas diffusion contribution coefficient is larger; if the distance between two locations is smaller, the confidence level of the likelihood of this location contributing more to the chlorine concentration at the other location is higher, and the corresponding gas diffusion contribution coefficient is larger; if the wind speed at a location is higher, it indicates that the strong wind is more likely to transport the gas to a farther distance, and the confidence level of the likelihood of this location contributing more to the chlorine concentration at the other location is higher, and the corresponding gas diffusion contribution coefficient is larger.

[0040] Any position is designated as the target position, any other position not including the target position is designated as the reference position, and any time is designated as the target time. The difference vector between the spatial coordinate vectors of the target position and the reference position is obtained. Based on the angle between the wind field vector of the target position at the target time and the difference vector, as well as the spatial coordinate vectors of the target position and the reference position, the gas diffusion contribution coefficient of the target position to the reference position at the target time is determined.

[0041] Specifically, the gas diffusion contribution coefficient satisfies:

[0042] ;

[0043] In the formula, For position At any moment For location The gas diffusion contribution coefficient, For position At any moment Wind field vector and difference vector The angle between them and Positions With position spatial coordinate vector, For position At any moment The wind field vector, For hyperparameters, For the standard normalized function, It is a natural exponential function. This is the modulus symbol.

[0044] Implementers can set hyperparameters according to the specific implementation situation, for example, 0.001. The existence of hyperparameters is to prevent... and When the value is 0, the formula becomes meaningless.

[0045] in, The larger, the more it means The closer the value is to 1, The closer the value is to 0°, the better the... The position is located at the 1st The greater the probability of being directly downwind from the location, the more likely the location is to be... The position for the th The greater the contribution of chlorine concentration at each location, the higher the corresponding position. The position for the th The larger the gas diffusion contribution coefficient at a given location, the greater the contribution; conversely, the smaller the location, the greater the contribution. The smaller, the better. The closer the value is to 0, The closer the value is to 180°, the more significant the change. The position is located at the 1st The greater the likelihood of the location being directly upwind, the more likely it is that the position is... The position for the th The contribution of chlorine concentration at each location may be smaller, corresponding to the [missing information - likely a specific location or position]. The position for the th The smaller the gas diffusion contribution coefficient at each location. The smaller the value, the better. The position for the th The greater the contribution of chlorine concentration at a given location, the more credible the correlation. The position for the th The larger the gas diffusion contribution coefficient at a given location, the better. The larger the wind, the stronger the gas will be, indicating that the gas will be carried from the first... The greater the likelihood that the item will be transported to a more distant location, the higher the probability of the corresponding item being transported. The position for the th The greater the contribution of chlorine concentration at a given location, the more credible the correlation. The position for the th The larger the gas diffusion contribution coefficient at a given location, the better.

[0046] S3: Determine the contribution coefficient of each location to the real gas diffusion at the target time for the other locations.

[0047] It is important to note that in real industrial environments, wind field vector data measured by anemometers are easily affected by instantaneous, non-continuous local airflow disturbances. This can lead to spurious, dramatic fluctuations in the wind field vector data. For example, rapid movement of people or brief opening and closing of nearby doors and windows can generate a short-lived but high-intensity local airflow near the anemometer. This disturbance is not a stable airflow generated by the overall ventilation system of the plant and cannot truly represent the macroscopic path of chlorine transport over long distances. If wind field vectors containing such spurious fluctuations are directly used to calculate the gas diffusion contribution coefficient, the value of the gas diffusion contribution coefficient will also exhibit drastic jumps that do not conform to physical reality. For example, a brief gust of wind pointing in a certain direction can cause the algorithm to momentarily misinterpret it as a fundamental change in the main gas diffusion direction, resulting in the subsequently calculated leak source location fluctuating erratically between the true source point and the spurious disturbance direction, severely affecting the stability and reliability of the location results. Therefore, this step uses a median filtering algorithm to optimize the gas diffusion contribution coefficients of different sensors at each location at each time step, obtaining the true gas diffusion contribution coefficient at each location at each time step. The specific implementation logic is as follows: at each time step, the median filtering algorithm will refer to the gas diffusion contribution coefficients within the sliding time window of the most recent multiple time steps, including the current time step, sort them, and then select the value in the middle of the sorted values ​​as the final output of the current time step.

[0048] The gas diffusion contribution coefficients of the target location to the reference location at multiple time points including the target time are filtered to obtain the true gas diffusion contribution coefficients of the target location to the reference location at the target time.

[0049] Specifically, the filtering includes:

[0050] The data within a sliding time window, which includes the contribution coefficients of the target location to the gas diffusion at the target time and several times before the target time, are sorted according to their numerical values.

[0051] The median within the sorted sliding time window is selected as the real gas diffusion contribution coefficient of the target location to the reference location at the target time.

[0052] The implementer can set the length of the sliding window according to the specific implementation situation, for example, 9.

[0053] S4: Determine the corrected airflow density at each location at the target time.

[0054] It's important to note that a real chlorine leak source not only has a detectable concentration at its own location, but more importantly, it contributes to the chlorine concentration of a vast downwind area. Therefore, the corrected airflow density at a given location at a given time should be determined by both the chlorine concentration at that location and its contribution to the chlorine concentration at all downstream locations. Thus, this step calculates the corrected airflow density at each location at each time based on the actual gas diffusion contribution coefficient and the original chlorine concentration. The logic is as follows: the higher the original chlorine concentration at a location at a given time, the higher its corresponding corrected airflow density. A higher actual gas diffusion contribution coefficient at a location at a given time indicates a greater contribution to the chlorine concentration at other locations. Therefore, even if the original concentration at a certain upwind source is not the highest, its contribution to the downstream will result in a higher corrected airflow density, surpassing all other locations and becoming the true density peak.

[0055] Based on the chlorine concentration values ​​at all locations at the target time, and the gas diffusion contribution coefficient of the target location at the target time to all other locations excluding the target location, determine the gas flow correction density at the target location at the target time.

[0056] Specifically, the airflow correction density satisfies:

[0057] ;

[0058] In the formula, For position At any moment Airflow correction density, For position At any moment The chlorine concentration value, For position At any moment The chlorine concentration value, For position At any moment For location The real gas diffusion contribution coefficient.

[0059] S5: Based on the airflow correction density at all locations at the target time, the density peak clustering algorithm is used to locate the chlorine leak source.

[0060] Specifically, locating the chlorine leak source includes:

[0061] Obtain the Euclidean distances between the target location and all locations with airflow correction densities greater than the target location, and record the minimum value among the Euclidean distances as the minimum distance of the target location at the target time.

[0062] Based on the corrected density of airflow at all locations at the target time and the minimum location, outliers are found in the decision map generated by the density peak clustering algorithm, and the locations corresponding to the outliers are identified as chlorine leakage sources.

[0063] Specifically, the process of finding anomalies includes:

[0064] If the airflow correction density corresponding to any data point in the decision map at the target time is greater than a preset density threshold, and the minimum distance corresponding to that data point at the target time is greater than a preset distance threshold, then that data point is identified as an outlier. See [link to relevant documentation]. Figure 2 Where D4 is the identified anomaly.

[0065] Implementers can set density and distance thresholds according to the specific implementation situation. For example, the density threshold is 45 and the distance threshold is 18.

[0066] Based on the location results, the system highlights the suspected leak location on the central monitoring interface and associates it with the specific equipment component, providing on-site personnel with emergency repair guidance.

[0067] This invention also discloses a chlorine leak detection system in reactor production, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a chlorine leak detection method in reactor production according to the present invention.

[0068] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0069] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for detecting a chlorine gas leak in a production of a reaction vessel, characterized by, The method comprises the following steps: obtaining the chlorine concentration values and wind field vectors of multiple positions around the reactor at each time point; Any location is designated as the target location, and any other location not including the target location is designated as the reference location. Any time is designated as the target time. The difference vector between the spatial coordinate vectors of the target location and the reference location is obtained. Based on the angle between the wind field vector of the target location at the target time and the difference vector, and the spatial coordinate vectors of the target location and the reference location, the gas diffusion contribution coefficient of the target location to the reference location at the target time is determined, including: In the formula, For position At any moment For location The gas diffusion contribution coefficient, For position At any moment Wind field vector and difference vector The angle between them and Positions With position spatial coordinate vector, For position At any moment The wind field vector, For hyperparameters, For the standard normalized function, It is a natural exponential function. The modulus symbol; filtering the gas diffusion contribution coefficients of the target position to the reference position at multiple time points including the target time point to obtain the real gas diffusion contribution coefficient of the target position to the reference position at the target time point; According to the chlorine concentration values of all positions at the target moment and the gas diffusion contribution coefficients of the target position at the target moment to all positions except the target position, a gas flow correction density of the target position at the target moment is determined, comprising: ; wherein is the position at time the air flow correction density, is the position at time the chlorine concentration value, is the position at time the chlorine concentration value, is the position at time the real gas diffusion contribution coefficient for the position ; locating the chlorine leakage source based on the air flow corrected density of all positions at the target time point and using the density peak value clustering algorithm.

2. A method for detecting a chlorine gas leak in the production of a reaction vessel according to claim 1, characterized in that, The chlorine concentration values are collected by a chlorine sensor based on the principle of ultraviolet absorption spectrum.

3. The method for detecting chlorine leakage in the production of a reaction kettle according to claim 1, characterized in that, The wind field vectors are collected using a three-dimensional ultrasonic anemometer.

4. The method for detecting chlorine leakage in the production of a reaction kettle according to claim 1, characterized in that, The filtering comprises the following steps: sorting the data in the sliding time window containing the gas diffusion contribution coefficients of the target position to the reference position at the target time point and several time points before the target time point according to the numerical values; selecting the median of the sorted sliding time window as the real gas diffusion contribution coefficient of the target position to the reference position at the target time point.

5. A method for detecting a chlorine gas leak in the production of a reaction vessel according to claim 4, characterized in that, The length of the sliding time window is 9.

6. A method for detecting a chlorine gas leak in the production of a reaction vessel according to claim 1, characterized by, The locating of the chlorine leakage source comprises the following steps: obtaining the Euclidean distances between the target position and all positions with air flow corrected density greater than that of the target position, and recording the minimum value in the Euclidean distances as the minimum distance of the target position at the target time point; locating the chlorine leakage source by finding the abnormal points in the decision graph generated by the density peak value clustering algorithm based on the air flow corrected density of all positions at the target time point and the minimum position, and identifying the positions corresponding to the abnormal points as the chlorine leakage source.

7. A method for detecting a chlorine gas leak in the production of a reaction vessel according to claim 6, characterized in that, The finding of the abnormal points comprises the following steps: determining that a data point in the decision graph is an abnormal point when the air flow corrected density corresponding to the data point at the target time point is greater than a preset density threshold value and the minimum distance corresponding to the data point at the target time point is greater than a preset distance threshold value.

8. A system for detecting a chlorine gas leak in a production of a reaction vessel, characterized by, The method comprises the following steps: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for detecting chlorine leakage in reactor production according to any one of claims 1-7 is realized.

Images