Safety management system and method based on chemical product production

Abstract

The invention discloses a safety management system and method based on chemical product production, and relates to the technical field of chemical product production, and the system comprises a reaction kettle sensor network deployment module, a chemical reaction process database construction module, and a reaction kettle real-time monitoring and management module. According to the method, the limitation that the reaction process is monitored only depending on basic parameters traditionally is broken through, high-precision sensor data and thermochromic coating color image information are creatively fused, and the reaction process can be more accurately grasped through multi-dimensional comparative analysis with a chemical reaction process database, abnormity can be found in time, corresponding measures are taken, and the method is suitable for large-scale popularization and application. Safety accidents caused by out-of-control reaction are effectively avoided; according to the method, the prediction data set of the reaction kettle in the current chemical reaction is predicted through the real-time monitoring data of the reaction kettle, the risk type of the reaction kettle is predicted in a targeted manner according to different chemical reactions, active prevention is achieved, and therefore the occurrence rate of dangerous accidents of the reaction kettle is reduced.

Description

Technical Field

[0001] This invention relates to the field of chemical product manufacturing technology, and more specifically to a safety management system and method for chemical product manufacturing. Background Technology

[0002] In modern chemical industries, the stable operation of reactors is crucial to the safety and efficiency of the entire production process. Chemical production is characterized by "high risk" and "accident amplification." Once a reactor fails, the consequences are often beyond imagination. Monitoring and managing reactors is a core task to ensure the smooth progress of chemical production and eliminate hidden dangers. Therefore, it is extremely necessary to monitor and manage reactors used in the production of chemical products.

[0003] Existing technologies, such as the invention application patent with announcement number CN118965145A, disclose a method and system for identifying dangerous areas based on downhole liquid level monitoring. By dynamically adjusting the sensor distribution density and acquisition frequency, multimodal data fusion, context awareness mechanism, and dangerous area prediction model, the system significantly improves accuracy and reliability, reduces maintenance costs and operational risks, and provides strong support for the safe and efficient operation of mines, oil extraction, and underground engineering through intelligent data processing and fault prediction.

[0004] Existing safety management systems for chemical production have the following shortcomings: First, the chemical reaction processes within reactors in chemical production are complex, with significant differences in reaction rates and heat release at different stages. Traditional monitoring only focuses on basic parameters such as temperature and pressure, making it difficult to accurately control the reaction process and easily leading to safety accidents due to uncontrolled reactions. Second, early warning systems for reactors in chemical production often rely on historical data thresholds. Real-time parameters are collected by sensors such as pressure, temperature, and liquid level, and compared with preset safety ranges to trigger alarms. This rule-based early warning method essentially confirms anomalies rather than predicts potential risks, resulting in a response process that lags behind the actual operating status of the equipment. It is mostly reactive, lacking predictability and failing to achieve proactive prevention. Summary of the Invention

[0005] The purpose of this invention is to provide a safety management system and method for chemical product production, which solves the problems existing in the background art.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The first aspect of the present invention provides a safety management system for chemical product production, including: a reactor sensor network deployment module, used to deploy a sensor network in the reactor during chemical product production, wherein the sensor network deployment includes thermocouple array deployment, pressure sensor deployment, thermochromic coating deployment and deployment of several other key sensors.

[0007] The chemical reaction process database construction module is used to collect historical data from the reactor and construct a chemical reaction process database. The chemical reaction process database includes data feature sets corresponding to each risk type of various chemical reactions at each stage, as well as regular datasets of various chemical reactions at each stage.

[0008] The real-time monitoring and management module for the reactor is used to collect monitoring data of the reactor in real time through the deployment of the reactor's sensor network, correct the temperature data of the reactor, import the temperature data and other monitoring data into the reactor's three-dimensional model, predict the risk type of the reactor, generate a risk control plan for the reactor, and execute the risk control plan at the reactor's management terminal.

[0009] A second aspect of the present invention provides a method for implementing the safety management system for chemical product production described herein, comprising: S1, deploying a sensor network in a reaction vessel during chemical product production, wherein the sensor network deployment includes a thermocouple array deployment, a pressure sensor deployment, a thermochromic coating deployment, and deployment of several other key sensors.

[0010] S2. Collect historical data from the reactor and construct a chemical reaction process database. The chemical reaction process database includes data feature sets corresponding to each risk type of various chemical reactions at each stage, as well as regular datasets of various chemical reactions at each stage.

[0011] S3. By deploying a sensor network on the reactor, real-time monitoring data of the reactor is collected, the temperature data of the reactor is corrected, and the temperature data and other monitoring data of the reactor are imported into the three-dimensional model of the reactor to predict the risk type of the reactor and generate a risk control plan for the reactor. The management terminal of the reactor executes the risk control plan.

[0012] The beneficial effects of the present invention are as follows: (1) The present invention breaks through the limitations of traditional methods that rely solely on basic parameters to monitor the reaction process. It innovatively integrates high-precision sensor data (temperature, pressure) with thermochromic coating color image information (different from single parameter monitoring methods). By comparing and analyzing with a chemical reaction process database in multiple dimensions, it can more accurately grasp the reaction process, promptly detect abnormalities and take corresponding measures, and effectively avoid safety accidents caused by uncontrolled reaction.

[0013] (2) This invention uses real-time monitoring data of the reactor to predict the predicted dataset of the reactor in the current chemical reaction, and predicts the risk type of the reactor according to the different chemical reactions, anticipates potential risks, responds proactively, and achieves proactive prevention, thereby reducing the incidence of dangerous accidents in the reactor. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the system structure connection of the present invention.

[0016] Figure 2 This is a schematic diagram of the implementation steps of the method of the present invention. Detailed Implementation

[0017] 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 embodiments of the present invention, and not all embodiments. 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.

[0018] Reference Figure 1 As shown, the first aspect of the present invention provides a safety management system for chemical product production, including: a reactor sensor network deployment module, used to deploy a sensor network in the reactor during chemical product production, wherein the sensor network deployment includes thermocouple array deployment, pressure sensor deployment, thermochromic coating deployment, and deployment of several other key sensors.

[0019] In a specific embodiment of the present invention, the deployment of a sensor network in the reaction vessel during chemical product production includes: dividing the reaction vessel axially into a top, middle and bottom section, and further dividing the top, middle and bottom sections of the reaction vessel into a grid, resulting in several grid areas at the top, middle and bottom of the reaction vessel.

[0020] Thermocouple array deployment: Thermocouples are installed in several grid areas at the top, middle and bottom of the reactor, and the installation angle of the thermocouples is at a 45° angle with the agitator blades. The thermocouples are installed by flange insertion and equipped with protective sleeves.

[0021] For example, suppose three sets of thermocouples are arranged at the top, middle and bottom of the reactor, and each set contains three measuring points, for a total of 9 sets and 27 measuring points. Each set of thermocouples is distributed along the circumference of the reactor body with a spacing of 120°.

[0022] It should be noted that the installation angle is 45° to the stirring blades in order to avoid fluid interference.

[0023] Pressure sensor deployment: Pressure sensors are installed in several grid areas at the top, middle and bottom of the reactor. The pressure sensors are installed by threaded connection and equipped with condensation bends and isolation valves to prevent high-temperature media from directly contacting the sensor diaphragm.

[0024] Thermochromic coating deployment: High-pressure airless spraying of thermochromic coatings is used to coat several grid areas at the top, middle and bottom of the reactor, with each grid area coated with a different thermochromic coating formulation (to achieve precise temperature-color mapping), and an industrial camera is installed in the center of the top viewing window of the reactor.

[0025] Additional sensor deployment: Several additional sensors are installed in several grid areas at the top, middle and bottom of the reactor.

[0026] The chemical reaction process database construction module is used to collect historical data from the reactor and construct a chemical reaction process database. The chemical reaction process database includes data feature sets corresponding to each risk type of various chemical reactions at each stage, as well as regular datasets of various chemical reactions at each stage.

[0027] It should be noted that the historical data of the reactor is collected specifically through the reactor online monitoring system and the production planning system. The reactor online monitoring system is equipped with various sensors to monitor the reaction process data of the reactor, and the reactor production planning system is used to plan and schedule the production of the reactor.

[0028] It should also be noted that the conventional dataset includes, but is not limited to, temperature datasets over time.

[0029] It should be noted again that by combining component analysis data from LIMS (Laboratory Management System) with operation logs from MES (Manufacturing Execution System), standardized ETL tools can achieve automatic integration of multi-source heterogeneous data. Based on mature time-series databases (such as InfluxDB) and distributed computing frameworks (such as Spark), massive amounts of historical batch data can be processed efficiently. By employing pattern recognition, statistical process control (SPC), and machine learning algorithms (such as clustering and anomaly detection), the baseline range of process parameters for each reaction stage (regular dataset) can be extracted from normal operation data. At the same time, by comparing deviation / accident batch data, the characteristic fingerprints of risks such as thermal runaway and overpressure at specific stages (such as the slope of temperature rise + cooling valve saturation linkage) can be mined. Finally, a structured and queryable risk feature library and process baseline library are formed. The existing technology is relatively mature and will not be elaborated here.

[0030] The real-time monitoring and management module for the reactor is used to collect monitoring data of the reactor in real time through the deployment of the reactor's sensor network, correct the temperature data of the reactor, import the temperature data and other monitoring data into the reactor's three-dimensional model, predict the risk type of the reactor, generate a risk control plan for the reactor, and execute the risk control plan at the reactor's management terminal.

[0031] It should be noted that the construction of the three-dimensional model of the reactor adopts multiphysics coupling simulation technology to build a high-precision three-dimensional model of the reactor, and incorporates fluid dynamics, heat and mass transfer and chemical reaction dynamics. The existing technology is relatively mature, so it will not be elaborated here.

[0032] In a specific embodiment of the present invention, the monitoring data of the reactor includes temperature, heating rate, temperature fluctuation coefficient, pressure, pressure increase rate, pressure fluctuation coefficient, thermochromic coating image, and several other key parameters of several grid areas at the top, middle and bottom. The temperature fluctuation coefficient is specifically the ratio of the standard deviation to the mean of the temperature data, which reflects the stability of the temperature data. The larger the temperature fluctuation coefficient, the more severe the temperature fluctuation.

[0033] In a specific embodiment of the present invention, the temperature data of the reactor is corrected, and the specific correction process is as follows: ST1, the temperature of several grid areas at the top, middle and bottom of the reactor is corrected using a first-order inertial model, wherein the first-order inertial model is specifically: Where T0(t) is the temperature collected by the thermocouple. It represents the rate of change of temperature with time, where τ is a time constant, specifically calibrated through a step response experiment (e.g., when a type K thermocouple steps at 100 degrees Celsius, τ is 0.8 s), and outputs the first temperature of several grid regions at the top, middle, and bottom of the reactor.

[0034] ST2. Based on the thermochromic coating images of several grid areas at the top, middle, and bottom of the reactor, and combined with the quantification relationship between color change and temperature error, the second temperature of several grid areas at the top, middle, and bottom of the reactor is corrected. The hue change rate and brightness change of several grid areas at the top, middle, and bottom of the reactor are obtained. After applying the hue change rate correction model H_correction = H*(1+k*ΔL), where ΔL is the brightness change, H is the hue change rate before correction, and k is the correction coefficient, the corrected hue change rate of several grid areas at the top, middle, and bottom of the reactor is output.

[0035] It should be noted that the correction coefficients are specifically determined in advance through linear regression fitting of experimental data, and the specific principle is as follows: ΔH_observed represents the actual observed rate of hue change, ΔL represents the change in brightness, and H_original represents the uncorrected rate of hue change.

[0036] ST3. Calculate the weighting factors of thermocouples and thermochromic coatings for several grid regions at the top, middle and bottom of the reactor, respectively.

[0037] ST4. Determining the model through temperature. Where T1 and T2 represent the first temperature and the second temperature, respectively, and the corrected temperatures of several grid areas at the top, middle and bottom of the reactor are obtained by fusion, which are used as the temperature data of the reactor. ω1' and ω2' are the weighting factors of the thermocouple and the thermochromic coating after homogenization, respectively.

[0038] In a specific embodiment of the present invention, the method for correcting the second temperature of several grid areas at the top, middle and bottom of the reactor is as follows: based on the image recognition color information of the thermochromic coating of several grid areas at the top, middle and bottom of the reactor, and combined with the coating color blocks of several grid areas, the initial temperature of several grid areas at the top, middle and bottom of the reactor is found based on the color information of each coating color block of the thermochromic coating at each temperature preset in the data warehouse.

[0039] It should be noted that the color information of each coated color block of the thermochromic coating at each temperature was calibrated in advance through experiments.

[0040] The relationship between the color space distance difference of each coated color block in a thermochromic coating and the mapping temperature difference is constructed based on the principle of ΔT=p0+p1*ΔE+p2*(ΔE). 2 , where p0, p1, and p2 are model coefficients, respectively, and ΔE is the color space distance.

[0041] It should be noted that the relationship between the color space distance difference between each coated color block and the mapping temperature difference in the thermochromic coating and the color space distance between each display color block was specifically calibrated beforehand through experiments. The specific details are as follows:

[0042] In this embodiment, the CIELAB color space is selected (for uniform perception, the Euclidean distance between two points in the color space is proportional to the color difference perceived by the human eye).

[0043] Using the X-Rite ColorChecker standard color chart, which contains 24 known color patches (such as pure white, pure black, etc.), the Lab value of each color patch is extracted (for camera color calibration). The camera transformation matrix M is calculated, and the ΔE error of the transformed color patch is verified. When the error meets the requirements, the following experiment is performed:

[0044] A coating sample with a certain color patch was placed in a constant temperature chamber. An industrial camera was used to take an image of the coating, and the average Lab values ​​were extracted and denoted as L1, a1, and b1. The temperature was gradually increased from 50°C to 200°C, and the data was recorded every 10 seconds. At each temperature point, an image of the coating and an image of each color patch were taken using an industrial camera. The color chart image was transformed using M-transformation and the illumination shift was compensated. The same transformation and compensation were applied to the coating image, and the average Lab values ​​were extracted and denoted as L2, a2, and b2.

[0045] Through calculation

[0046] The difference between the actual temperature of the constant temperature chamber and the temperature reflected by the coating is ΔT = T_measured - T_reflected.

[0047] Establish a dataset showing the correspondence between ΔE and ΔT (e.g., ΔE = 2.3 corresponds to ΔT = +0.5℃), and use quadratic polynomial fitting to capture nonlinear relationships, i.e., ΔT = p0 + p1 * ΔE + p2 * (ΔE). 2 The model coefficients p0, p1, and p2 are solved using the least squares method. For example, for a red coating, the fitting result might be ΔT = -0.12*(ΔE). 2 +0.85*ΔE-1.2.

[0048] Based on the thermochromic coating images of several grid regions at the top, middle, and bottom of the reactor, the color space distance difference was extracted and calculated.

[0049] Based on the color blocks of the thermochromic coating in several grid areas at the top, middle and bottom of the reactor, the color space distance difference of several grid areas at the top, middle and bottom of the reactor is imported into the relationship between the color space distance difference of the corresponding color block and the mapped temperature difference, and the mapped temperature difference is calculated.

[0050] The initial temperature of several grid regions at the top, middle, and bottom of the reactor is added to the difference between the initial temperature and the mapped temperature to obtain the second temperature of several grid regions at the top, middle, and bottom of the reactor.

[0051] In a specific embodiment of the present invention, the weighting factors of the thermocouples in several grid regions at the top, middle, and bottom of the reactor are specifically determined through a thermocouple weighting model. Sure.

[0052] It should be noted that the thermocouple weighting model combines the temperature fluctuation coefficient and the heating rate, which affect the weights through a product. When the TFC is high or the heating rate exceeds a threshold, the weight decreases; conversely, the weight increases. The weights are dynamically adjusted through an exponential decay function. The faster the heating rate, the more likely the thermocouple will experience a delay due to thermal inertia, resulting in a lower measured value and more severe decay.

[0053] The weighting factors of the thermochromic coating in several grid areas at the top, middle, and bottom of the reactor are specifically determined through a thermochromic coating weighting model. Sure.

[0054] In the above, e is the natural constant, TFC is the temperature fluctuation coefficient, T_yz is the heating rate threshold, ΔE is the color space distance, ΔE' is the color space distance threshold, and H' is the hue change rate threshold (the maximum hue change rate during the experiment).

[0055] It should be noted that the heating rate threshold is specifically extracted from conventional datasets of various chemical reactions at each stage, taking into account the current reaction stage of the reactor.

[0056] This invention breaks through the limitations of traditional methods that rely solely on basic parameters to monitor reaction processes. It innovatively integrates high-precision sensor data (temperature, pressure) with color image information of thermochromic coatings (different from single-parameter monitoring methods). By comparing and analyzing the data with a chemical reaction process database in multiple dimensions, it can more accurately grasp the reaction process, promptly detect anomalies, and take corresponding measures to effectively avoid safety accidents caused by uncontrolled reactions.

[0057] In a specific embodiment of the present invention, the prediction process for predicting the risk type of the reactor is as follows: obtaining the prediction dataset of the reactor in the current chemical reaction from the three-dimensional model of the reactor, extracting the data feature sets corresponding to each risk type and each stage of each chemical reaction from the chemical reaction process database, identifying the prediction stage of the reactor according to the prediction cycle of the reactor, obtaining the data feature sets of the reactor in the prediction stage of each risk type of the current chemical reaction, thereby identifying the characteristic risk type of the reactor.

[0058] For example, if the current chemical reaction in the reactor is a polymerization reaction, the prediction dataset consists of the reactor's heat release rate during the prediction period, the cutoff temperature and temperature rise rate of several grid regions at the top, middle and bottom. If the cutoff temperature of a certain grid region is greater than or equal to the safe cutoff temperature of the current chemical reaction in the prediction period, or the temperature rise rate is greater than or equal to the safe temperature rise rate of the current chemical reaction in the prediction period, then local overheating is determined to be a characteristic risk type of the reactor. If the heat release rate of the reactor during the prediction period is greater than or equal to the safe heat release rate of the current chemical reaction in the prediction period, then reaction runaway is determined to be a characteristic risk type of the reactor.

[0059] If the current chemical reaction in the reactor is esterification, the prediction dataset consists of the conversion rate and the proportion of side reactions in the reactor during the prediction period. When the conversion rate of the reactor during the prediction period is less than the expected conversion rate of the current chemical reaction in the reactor during the prediction period, or when the proportion of side reactions is greater than the expected proportion of side reactions, the reaction stagnation is determined to be the characteristic risk type of the reactor.

[0060] If the current chemical reaction in the reactor is an oxidation reaction, the prediction dataset consists of the reactor's cutoff pressure, pressure surge rate, and gas volume fraction during the prediction period. If the reactor's pressure surge rate during the prediction period is greater than or equal to the reactor's allowable pressure surge rate, gas volume fraction during the prediction period, or cutoff pressure during the prediction period, then the pressure surge is determined to be a characteristic risk type of the reactor.

[0061] In the above, the safe cutoff temperature and the safe temperature rise rate are specifically the data feature set of the polymerization reaction when it is locally overheated, and the safe exothermic rate is the data feature set of the polymerization reaction when it is runaway.

[0062] For example, if the risk type of the reactor is local overheating, the local overheating grid area of ​​the reactor is summarized, and the medium is sprayed onto the local overheating grid area by a rotating nozzle in the top area of ​​the reactor.

[0063] If the risk type of the reactor is runaway, then stop the monomer feed and start the inhibitor pulse injection.

[0064] If the risk type of the reactor is reaction stagnation, the catalyst flow rate will be automatically increased or the stirring speed will be increased.

[0065] If the risk type of the reactor is pressure surge, the pressure balancing valve will automatically open.

[0066] This invention uses real-time monitoring data from the reactor to predict the predicted dataset of the reactor in the current chemical reaction. Based on different chemical reactions, it specifically predicts the risk type of the reactor, anticipates potential risks, and proactively responds to achieve proactive prevention, thereby reducing the incidence of dangerous reactor accidents.

[0067] Reference Figure 2 As shown, a second aspect of the present invention provides a method for implementing the safety management system for chemical product production described in the present invention, comprising: S1, deploying a sensor network in a reaction vessel during chemical product production, wherein the sensor network deployment includes a thermocouple array deployment, a pressure sensor deployment, a thermochromic coating deployment, and deployment of several other key sensors.

[0068] S2. Collect historical data from the reactor and construct a chemical reaction process database. The chemical reaction process database includes data feature sets corresponding to each risk type of various chemical reactions at each stage, as well as regular datasets of various chemical reactions at each stage.

[0069] S3. By deploying a sensor network on the reactor, real-time monitoring data of the reactor is collected, the temperature data of the reactor is corrected, and the temperature data and other monitoring data of the reactor are imported into the three-dimensional model of the reactor to predict the risk type of the reactor and generate a risk control plan for the reactor. The management terminal of the reactor executes the risk control plan.

[0070] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A safety management system for chemical product production, characterized in that, include: The reactor sensor network deployment module is used to deploy a sensor network in a reactor during chemical product production. The sensor network deployment includes thermocouple array deployment, pressure sensor deployment, thermochromic coating deployment, and deployment of several other key sensors. A chemical reaction process database construction module is used to collect historical data from reaction vessels and construct a chemical reaction process database. The chemical reaction process database includes data feature sets corresponding to each risk type of various chemical reactions at each stage, as well as regular datasets of various chemical reactions at each stage. The real-time monitoring and management module for the reactor is used to collect monitoring data of the reactor in real time through the deployment of the reactor's sensor network, correct the temperature data of the reactor, import the temperature data and other monitoring data into the reactor's three-dimensional model, predict the risk type of the reactor, generate a risk control plan for the reactor, and execute the risk control plan at the reactor's management terminal.

2. The safety management system for chemical product production according to claim 1, characterized in that, The specific content of deploying a sensor network in the reaction vessel during chemical product production includes: The reactor is axially divided into top, middle and bottom, and the reactor body at the top, middle and bottom is gridded to obtain several grid regions at the top, middle and bottom of the reactor. Thermocouple array deployment: Thermocouples are installed in several grid areas at the top, middle and bottom of the reactor, and the installation angle of the thermocouples is at a 45° angle with the agitator blades. The thermocouples are installed by flange insertion and equipped with protective sleeves. Pressure sensor deployment: Pressure sensors are installed in several grid areas at the top, middle and bottom of the reactor. The pressure sensors are installed by threaded connection and equipped with condensation bends and isolation valves. Thermochromic coating deployment: High-pressure airless spraying of thermochromic coatings is used to coat several grid areas at the top, middle and bottom of the reactor, and each grid area is coated with a thermochromic coating of a different formulation. An industrial camera is installed in the center of the viewing window at the top of the reactor. Additional sensor deployment: Several additional sensors are installed in several grid areas at the top, middle and bottom of the reactor.

3. The safety management system for chemical product production according to claim 1, characterized in that, The monitoring data of the reactor includes temperature, heating rate, temperature fluctuation coefficient, pressure, pressure increase rate, pressure fluctuation coefficient, thermochromic coating image, and several other key parameters for several grid areas at the top, middle, and bottom.

4. The safety management system for chemical product production according to claim 3, characterized in that, The temperature data of the reactor obtained by the correction is specifically corrected as follows: ST1. A first-order inertial model is used to correct the temperature of several grid regions at the top, middle, and bottom of the reactor. Specifically, the first-order inertial model is: T_correction. Where T0(t) is the temperature collected by the thermocouple. It represents the rate of change of temperature with time, where τ is a time constant. Specifically, it is calibrated through a step response experiment and outputs the first temperature of several grid regions at the top, middle, and bottom of the reactor. ST2. Based on the thermochromic coating images of several grid areas at the top, middle, and bottom of the reactor, and combined with the quantization relationship between color change and temperature error, the second temperature of several grid areas at the top, middle, and bottom of the reactor is corrected. The hue change rate and brightness change of several grid areas at the top, middle, and bottom of the reactor are obtained. After applying the hue change rate correction model H_correction = H*(1+k*ΔL), where ΔL is the brightness change, H is the hue change rate before correction, and k is the correction coefficient, the corrected hue change rate of several grid areas at the top, middle, and bottom of the reactor is output. ST3. Calculate the weighting factors of thermocouples and thermochromic coatings for several grid regions at the top, middle and bottom of the reactor, respectively. ST4. Determining the model through temperature. Where T1 and T2 represent the first temperature and the second temperature, respectively, and the corrected temperatures of several grid areas at the top, middle and bottom of the reactor are obtained by fusion, which are used as the temperature data of the reactor. ω1' and ω2' are the weighting factors of the thermocouple and the thermochromic coating after homogenization, respectively.

5. The safety management system for chemical product production according to claim 4, characterized in that, The correction obtains the second temperature of several grid regions at the top, middle, and bottom of the reactor. The specific correction method is as follows: Based on the image recognition color information of the thermochromic coating in several grid areas at the top, middle and bottom of the reactor, and combined with the coating color blocks in several grid areas, the initial temperature of several grid areas at the top, middle and bottom of the reactor is found based on the color information of each coating color block of the thermochromic coating at various temperatures in the data warehouse. The relationship between the color space distance difference of each coated color block in a thermochromic coating and the mapping temperature difference is constructed based on the principle of ΔT=p0+p1*ΔE+p2*(ΔE). 2 , where p0, p1, and p2 are model coefficients, respectively, and ΔE is the color space distance; Based on the thermochromic coating images of several grid regions at the top, middle and bottom of the reactor, the color space distance difference was extracted and calculated. Based on the color blocks of the thermochromic coating in several grid areas at the top, middle and bottom of the reactor, the color space distance difference of several grid areas at the top, middle and bottom of the reactor is imported into the relationship between the color space distance difference of the corresponding color block and the mapped temperature difference, and the mapped temperature difference is calculated. The initial temperature of several grid regions at the top, middle, and bottom of the reactor is added to the difference between the initial temperature and the mapped temperature to obtain the second temperature of several grid regions at the top, middle, and bottom of the reactor.

6. The safety management system for chemical product production according to claim 5, characterized in that, The weighting factors of the thermocouples in several grid regions at the top, middle, and bottom of the reactor are specifically determined through a thermocouple weighting model. Sure; The weighting factors of the thermochromic coating in several grid areas at the top, middle, and bottom of the reactor are specifically determined through a thermochromic coating weighting model. Sure; In the above, e is the natural constant, TFC is the temperature fluctuation coefficient, T_yz is the heating rate threshold, ΔE is the color space distance, ΔE' is the color space distance threshold, and H' is the hue change rate threshold.

7. The safety management system for chemical product production according to claim 1, characterized in that, The specific prediction process for the risk type of the reactor is as follows: The prediction dataset of the reactor in the current chemical reaction is obtained from the 3D model of the reactor, and the data feature sets corresponding to each risk type and each stage of each chemical reaction are extracted from the chemical reaction process database. Based on the prediction cycle of the reactor, the prediction stage of the reactor is identified, and the data feature sets of the reactor in the prediction stage of each risk type in the current chemical reaction are obtained, thereby identifying the characteristic risk type of the reactor.

8. A method for implementing the safety management system for chemical product production as described in any one of claims 1-7, characterized in that, include: S1. Deploy a sensor network in the reaction vessel during chemical product production. The sensor network deployment includes thermocouple array deployment, pressure sensor deployment, thermochromic coating deployment, and deployment of several other key sensors. S2. Collect historical data of the reactor and construct a chemical reaction process database. The chemical reaction process database includes data feature sets corresponding to each risk type of various chemical reactions at each stage and regular datasets of various chemical reactions at each stage. S3. By deploying a sensor network on the reactor, real-time monitoring data of the reactor is collected, the temperature data of the reactor is corrected, and the temperature data and other monitoring data of the reactor are imported into the three-dimensional model of the reactor to predict the risk type of the reactor and generate a risk control plan for the reactor. The management terminal of the reactor executes the risk control plan.

Images