An environmental protection and safety early warning system for production of alkyd resin

Abstract

The application discloses an environment-friendly and safe early warning system for production of alkyd resin, relates to the technical field of environment-friendly and safe technology, and constructs a multi-source real-time data pool, outputs a risk level based on the data pool, verifies a control strategy, issues a control instruction, executes hierarchical alarm, solidifies discharge and alarm records, and realizes risk assessment and verification of a second-level control instruction through multi-source real-time data fusion, cooperative prediction of a heat release rate and VOC, Bayesian threshold self-adaptation and digital twin simulation. The system integrates reflux regulation, heat load reduction and inert gas protection interlocking, and supports three-dimensional visualization and multi-terminal early warning interaction through fingerprint chaining, solidification of discharge and alarm records, breaks through the delay of traditional DCS monitoring and the limitation of static threshold, and takes into account environmental compliance and intrinsic safety.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of environmental protection and safety, and in particular to an environmentally friendly and safe early warning system for alcohol acid resin production. BACKGROUND

[0002] In the industrial production process of alcohol acid resin, high-temperature esterification reaction and solvent reflux process are generally used. This process is accompanied by the release of a large amount of volatile organic compounds and the accumulation of reaction heat. There are dual risks of unstable VOC emissions and explosive solvent atmosphere in the workshop. Traditional monitoring methods rely on single-point concentration detection and regular manual sampling. The monitoring delay is generally in minutes, making it difficult to dynamically predict the reaction heat release and volatile trends. At the same time, the existing system separates the environmental monitoring and safety interlocking operation, lacks a risk linkage judgment mechanism based on a unified data pool, and cannot generate a coordinated control strategy before the risk is formed. There is a lack of online modeling means for the correlation between acid value dynamic changes and VOC concentration, resulting in fixed risk thresholds that cannot be adaptively adjusted according to production batch fluctuations, and there is a risk of threshold misjudgment and interlocking action failure.

[0003] At present, the Chinese invention patent with application number CN202411882404.8 discloses a production process risk early warning system, method and equipment for a cushion plate. The system can accurately analyze the production compliance factors of the cushion plate production equipment and the environmental pollution index by collecting the production parameters of the cushion plate production equipment and the environmental parameters of the production equipment in real time. It not only helps to timely discover potential risks in the production process, so that the cushion plate production equipment can adapt to the cushion plate production process, but also effectively prevents production quality problems caused by cushion plate production equipment failure or environmental pollution. After the cushion plate production equipment completes the manufacturing of each sample cushion plate, the quality of each sample cushion plate is further evaluated by testing the parameters of each sample cushion plate, thereby realizing secondary risk early warning control of the cushion plate production process. This double early warning mechanism not only significantly improves the safety, stability and environmental protection of the cushion plate production, but also ensures the continuous reliability of the cushion plate quality.

[0004] The above-mentioned technology cannot timely integrate multi-source process data and realize risk prediction and second-level interlocking verification, and cannot simultaneously meet the environmental emission control and intrinsic safety linkage needs of alcohol acid resin production. SUMMARY

[0005] The technical problem solved by the present application is that the existing technology cannot timely integrate multi-source process data and realize risk prediction and second-level interlocking verification, and cannot simultaneously meet the environmental emission control and intrinsic safety linkage needs of alcohol acid resin production.

[0006] To solve the above technical problems, the present application provides the following technical solutions:

[0007] An environmentally friendly and safe early warning system for alkyd resin production, comprising a data fusion acquisition module, a risk prediction analysis module, a digital twin simulation module, a self-adaptive interlocking control module, a man-machine interaction early warning module and a compliance record tracing module;

[0008] The data fusion acquisition module is used to construct a multi-source real-time data pool;

[0009] The risk prediction analysis module is used to output a risk level based on the data pool;

[0010] The digital twin simulation module is used to verify the control strategy;

[0011] The self-adaptive interlocking control module is used to issue control instructions;

[0012] The man-machine interaction early warning module is used to perform hierarchical alarm;

[0013] The compliance record tracing module is used to solidify the emission and alarm records;

[0014] The data fusion acquisition module comprises a VOC spectrum acquisition unit, a heat flux density monitoring unit, an acid value micro-titration unit and a solvent reflux monitoring unit;

[0015] The VOC spectrum acquisition unit is used to obtain component mass concentration and write into the data pool;

[0016] The heat flux density monitoring unit is used to output a reaction kettle wall and condenser outer wall temperature gradient scanning signal and form a heat flux vector field;

[0017] The acid value micro-titration unit is used to collect measured acid value and moisture content of resin sample liquid, and output real-time acid value sequence;

[0018] The solvent reflux monitoring unit is used to obtain reflux mass flow and reflux temperature and synchronize to the data pool;

[0019] The risk prediction analysis module comprises a heat release rate inference unit, a VOC trend inference unit and a threshold adaptive unit;

[0020] After being used to receive real-time acid value sequence, the heat release rate inference unit, according to reaction order coefficients, resin formula parameters and reaction heat constants in a historical sample library, uses a spline fitting algorithm to calculate the current reaction rate constant, wherein the historical sample library is established by dynamic titration experiments corresponding to typical resin formulas;

[0021] Integrate reaction heat in a preset prediction window, output heat release rate sequence and predicted wall temperature sequence, and calculate local temperature rise amount of kettle top based on average value of kettle top heat flux density, and call a steam pressure analysis model to calculate instantaneous steam pressure of head space;

[0022] The vapor pressure analytical model fits the local temperature estimate with pre-stored component vapor-liquid equilibrium parameters including vapor pressure constant and fugacity coefficient according to Antoine equation, and outputs vapor pressure time series data;

[0023] The vapor pressure time series data is used to calculate corresponding VOC mass concentration according to preset vapor gas phase conversion coefficient, a VOC mass concentration change curve is generated by using a cubic spline interpolation method, and the VOC mass concentration change curve is written into a risk buffer area at a preset recording interval;

[0024] The peak value of the VOC mass concentration curve, the time when the peak value appears, and the concentration rising slope are written into a risk assessment table;

[0025] The threshold adaptive unit is used to write the latest batch data including heat release rate deviation, VOC concentration deviation and response control amount deviation after obtaining a batch production completion signal;

[0026] After receiving the latest batch data, an incremental Bayesian updating algorithm is called to update the posterior distribution of the risk threshold, the historical batch deviation data is updated, and the threshold solidification process is triggered when the confidence interval shrinks, the prediction error converges or the sample capacity is saturated, and a threshold table with a confidence interval is output, the threshold table includes a lower confidence limit, an upper confidence limit and an expected value;

[0027] The threshold table is published to the risk buffer area and the adaptive interlocking control module through an internal message queue;

[0028] The threshold adaptive unit uses an incremental Bayesian learning method to update the posterior distribution of the historical batch deviation data and correct the posterior distribution based on the control response error vector, and optimizes the parameters in the threshold table;

[0029] The risk prediction analysis module further receives control response data from the adaptive interlocking control module, and is used to correct the prediction parameters and update the next round of risk assessment input.

[0030] Preferably, the digital twin simulation module includes a virtual reaction kettle instance unit, a vapor cloud diffusion simulation unit and an interlocking strategy pre-play unit;

[0031] The virtual reaction kettle instance unit is used to construct a three-dimensional finite element model, and the model parameters of the three-dimensional finite element model include shell thickness distribution, mixer blade type, heat conduction oil jacket structure and pressure support layout;

[0032] The shell thickness measurement, the stirring torque measurement and the feed temperature measurement are received, and the shell thickness measurement, the stirring torque measurement and the feed temperature measurement are mapped to the corresponding finite element boundary conditions;

[0033] The explicit coupled solver is invoked to perform transient iteration on the thermal stress field, and outputs the shell nodal temperature matrix, the shell principal stress matrix, and the blade shaft torque distribution vector.

[0034] Write the node number list, temperature matrix, principal stress matrix and timestamp into the simulation buffer;

[0035] The vapor cloud diffusion simulation unit is used to obtain the heat flow vector field and VOC concentration curve, and generate an unstructured three-dimensional CFD mesh in combination with the workshop geometric boundary conditions.

[0036] Within each preset time step, the incompressible Navier-Stokes solver based on pressure-velocity coupling is invoked and the component transport equations are superimposed to calculate the mass fraction field of volatile components.

[0037] Execute the volume fraction threshold truncation algorithm to extract the isosurface corresponding to the lower explosion limit and output the isosurface vertex coordinate set, grid index mapping table and time identifier;

[0038] Write the isosurface vertex coordinate set, mesh index mapping table and time identifier into the simulation buffer via IPC shared memory;

[0039] The interlocking strategy pre-simulation unit is used to construct an equivalent discrete event-driven model based on the hardware PLC logic table after obtaining the control instruction sequence proposed by the adaptive interlocking control module.

[0040] In each preset discrete event cycle, control commands are injected as event triggering conditions into the virtual reactor instance unit and the steam cloud diffusion simulation unit, and the simulation clock is advanced synchronously.

[0041] When obtaining the simulation end signal, the risk constraint verification function is called to determine the principal stress matrix of the shell, the nodal temperature matrix, and the volume fraction of the explosive vapor cloud:

[0042] If the judgment result is within the threshold table limit, an instruction approval identifier is generated and the approval status, control instruction sequence, and simulation annotation are written to the control instruction queue.

[0043] If any judgment result exceeds the threshold, an instruction rejection flag is generated and the rejection reason code and maximum deviation parameter are recorded to the rejection log file.

[0044] Preferably, the adaptive interlocking control module includes a reflux regulation unit, a heat load reduction unit, and an inert gas protection unit;

[0045] The reflux regulation unit is used to subscribe to the real-time value of VOC mass concentration in the data pool and call the lower limit explosion limit (LEL) data in the configuration file to generate a target threshold. When the real-time value of VOC mass concentration reaches the preset first concentration threshold, the following control is executed:

[0046] reading the real-time value of the reflux mass flow rate, calculating a target reflux increment, and sending an analog output signal to the frequency converter, the analog output signal being used to increase the rotational speed of the reflux pump;

[0047] In each sampling period, the deviation between the read value of the flow meter and the target value is calculated, and if the deviation exceeds a preset percentage, a correction pulse is sent out;

[0048] When the real-time value of the VOC mass concentration is lower than a preset second concentration threshold value within a preset time period, the rotational speed of the reflux pump is adjusted in the reverse direction to a steady-state value at the same slope, and the adjustment curve is archived;

[0049] The heat load reduction unit is used to obtain a wall temperature sequence and calculate a temperature rise rate;

[0050] When the temperature rise rate exceeds a preset first temperature rise rate threshold value, a bypass valve opening adjustment logic is triggered, and the bypass valve opening adjustment logic is:

[0051] An opening control signal is sent to the electrical converter, and the outlet temperature of the jacket heat transfer oil is sampled;

[0052] If the outlet temperature of the jacket heat transfer oil is higher than the target temperature, the opening of the bypass valve is increased to a first opening value;

[0053] If the temperature rise rate threshold value is lower than a preset second temperature rise rate threshold value and the outlet temperature of the jacket heat transfer oil is less than or equal to the target temperature, the opening of the bypass valve is decreased to a second opening value;

[0054] The valve position change, the temperature rise rate correction amount, and the observation time scale are written into a control log;

[0055] The inert gas protection unit is used to receive the oxygen content in the kettle and the pressure in the kettle in real time through a serial bus, and when the VOC concentration risk index and the oxygen content deviation index both meet the triggering conditions, a nitrogen flow control signal is output;

[0056] The adaptive interlock control module is triggered based on the heat release rate sequence, the predicted wall temperature sequence, and the VOC mass concentration curve output by the risk prediction analysis module.

[0057] Preferably, the human-computer interaction early warning module includes a three-dimensional visualization terminal unit, a hierarchical alarm pushing unit, and a semantic analysis and broadcast unit:

[0058] The three-dimensional visualization terminal unit is used to display a three-dimensional model of the plant in the central control room and on the mobile terminal, and to superimpose in real time an isosurface layer generated by the VOC mass concentration curve output by the risk prediction analysis module;

[0059] The hierarchical alarm pushing unit is configured to compare the current prediction result and the actual monitoring data based on the confidence interval in the threshold table generated by the threshold self-adaptive unit, make a level judgment, and push the alarm information corresponding to the risk level to the DCS system, the mobile terminal and the large screen in the duty room respectively.

[0060] The semantic analysis and broadcast unit is configured to convert the key control instructions and the alarm information generated in the control module into semantic prompts.

[0061] Preferably, the compliance record tracing module comprises an environmental emission report generation unit, a safety event fingerprint on-chain unit and an external interface unit.

[0062] The environmental emission report generation unit is configured to generate the unit emission rate and the total emission amount based on the VOC emission data collected by the data fusion collection module and the actual VOC concentration change trend after the execution of the control module, and form a structured emission record.

[0063] The safety event fingerprint on-chain unit is configured to combine and encode the predicted risk record generated by the risk prediction and analysis module, the control response result generated by the control module and the alarm sequence triggered by the human-computer interaction module, generate a hash fingerprint and write it into a private chain.

[0064] The external interface unit is configured to provide a standardized environmental compliance data access interface to an ecological environment supervision platform through an OAuth2.0 protocol.

[0065] Preferably, the digital twin simulation module further comprises a parameter mapping extraction mechanism and a fast proxy model calling mechanism, which are configured to call a reduced-order proxy model to perform second-level fast simulation based on the stored working condition mapping data during simulation running, and generate the main stress, temperature distribution and steam cloud volume distribution results corresponding to the control strategy.

[0066] Preferably, the steam cloud diffusion simulation unit continuously calculates the steam cloud boundary and predicts the farthest diffusion distance based on the real-time updated three-dimensional grid and multi-step time advancing algorithm.

[0067] Preferably, the safety event fingerprint on-chain unit records the block timestamp and node signature immediately after generating the hash fingerprint.

[0068] The beneficial effects of the present application: the present application discloses that through multi-source real-time data fusion, heat release rate and VOC collaborative prediction, Bayesian threshold self-adaptation and digital twin simulation, risk assessment and second-level control instruction verification are realized. The system integrates reflux regulation, heat load reduction and inert gas protection interlocking, and through fingerprint chaining solidification emission and alarm record, supports three-dimensional visualization and multi-terminal early warning interaction, breaks through the delay of traditional DCS monitoring and static threshold limit, and considers environmental compliance and intrinsic safety. BRIEF DESCRIPTION OF DRAWINGS

[0069] Figure 1 A basic flowchart of an environmental protection and safety early warning system for alkyd resin production is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0070] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments.

[0071] Embodiments, with reference to Figure 1 , an environmental protection and safety early warning system for alkyd resin production is provided, comprising a data fusion acquisition module, a risk prediction analysis module, a digital twin simulation module, a self-adaptive interlocking control module, a man-machine interaction early warning module and a compliance record tracing module.

[0072] The data fusion acquisition module is used to construct a multi-source real-time data pool.

[0073] The risk prediction analysis module outputs a risk level based on the data pool.

[0074] The digital twin simulation module is used to verify the control strategy.

[0075] The self-adaptive interlocking control module is used to issue control instructions.

[0076] The man-machine interaction early warning module is used to execute graded alarm.

[0077] The compliance record tracing module is used to solidify emission and alarm records.

[0078] The present application discloses that through multi-source real-time data fusion, heat release rate and VOC collaborative prediction, Bayesian threshold self-adaptation and digital twin simulation, risk assessment and second-level control instruction verification are realized. The system integrates reflux regulation, heat load reduction and inert gas protection interlocking, and through fingerprint chaining solidification emission and alarm record, supports three-dimensional visualization and multi-terminal early warning interaction, breaks through the delay of traditional DCS monitoring and static threshold limit, and considers environmental compliance and intrinsic safety.

[0079] The data fusion acquisition module includes a VOC spectrum acquisition unit, a heat flux density monitoring unit, an acid value microtitering unit, and a solvent reflux monitoring unit.

[0080] The VOC spectral acquisition unit is used to obtain the component mass concentration and write it into the data pool.

[0081] The VOC spectral acquisition unit enables the acquisition and structured recording of the mass concentration of volatile organic components at the top of the reactor within seconds, providing continuous concentration baseline data for VOC trend analysis and explosive vapor cloud determination.

[0082] The heat flux density monitoring unit is used to output the temperature gradient scanning signal between the reactor wall and the outer wall of the condenser and to form a heat flux vector field.

[0083] The heat flux density monitoring unit generates a temperature gradient and heat flux vector field covering the outer wall of the reactor and condenser, which are used to establish a dynamic relationship between the heat release rate and the thermal stress of the shell.

[0084] The acid value microtiter unit is used to collect resin sample solutions to measure acid value and moisture content, and output a real-time acid value sequence.

[0085] The acid value microtitering unit outputs real-time acid value and moisture content sequences, enabling online quantification of reaction progress and exothermic characteristics, and providing key stoichiometric parameters for rate inference and energy balance analysis.

[0086] The solvent reflux monitoring unit is used to acquire reflux mass flow rate and reflux temperature and synchronize them to the data pool.

[0087] The solvent reflux monitoring unit continuously records the reflux mass flow rate and reflux temperature, forming a time series of solvent phase equilibrium state, which provides real-time feedback for VOC concentration control and reflux adjustment logic.

[0088] The data fusion acquisition module synchronously collects and standardizes multi-source real-time parameters in the alkyd resin production process, forming a complete data pool that includes volatile organic compound concentration, heat flux density distribution, acid value changes, and solvent reflux status. This provides continuous and accurate basic data support for subsequent risk prediction, digital twin simulation, and interlocking control, ensuring that reaction process variables and environmental emission parameters are calculated and transmitted in a unified manner at the same time.

[0089] The risk prediction and analysis module includes a heat release rate inference unit, a VOC trend inference unit, and a threshold adaptation unit.

[0090] After receiving the real-time acid value sequence, the exothermic rate inference unit calculates the current reaction rate constant based on the reaction order coefficients, resin formulation parameters, and reaction heat constants in the historical sample library, which is established by dynamic titration experiments corresponding to typical resin formulations.

[0091] Integrate the reaction heat within the preset prediction window, output the exothermic rate sequence and the predicted wall temperature sequence.

[0092] The predicted wall temperature sequence is extracted by the longest peak search algorithm to obtain the highest wall temperature and bind it with the time identifier, and then written to the risk buffer zone.

[0093] The exothermic rate inference unit realizes the calculation of the reaction rate constant based on the real-time acid value sequence and the integration of the heat release, outputs the continuous exothermic rate sequence and the predicted wall temperature sequence, and provides energy balance data for the thermal runaway trend judgment.

[0094] The VOC trend inference unit is used to obtain the heat flow vector field and the exothermic rate sequence, and calculate the local temperature rise at the top of the kettle based on the average value of the heat flux density at the top of the kettle, and call the steam pressure analysis model to calculate the instantaneous steam pressure in the head space.

[0095] The steam pressure analysis model fits the local temperature estimate with the pre-stored component vapor-liquid equilibrium parameters according to the Antoine equation, and outputs the steam pressure time series data.

[0096] The steam pressure time series data is calculated according to the preset steam gas phase conversion coefficient to obtain the corresponding VOC mass concentration, and the VOC mass concentration change curve is generated by the cubic spline interpolation method, and written to the risk buffer zone at a preset recording interval.

[0097] The peak value of the VOC mass concentration curve, the time when the peak value appears, and the concentration rising slope are written into the risk assessment table.

[0098] The parameters required by the steam pressure analysis model include the parameter set of the target component, which includes the steam pressure constant, the fugacity coefficient and the boiling point. The parameter set is a pre-defined static data set for component level, which is obtained by extracting the material safety data table and component data in the system deployment stage. The steam pressure model of different formulations is defined by engineers in the configuration file, and the current batch formula is bound at runtime through the parameter calling mechanism, to ensure the consistency of the component and the validity of the parameters of the steam pressure prediction.

[0099] The VOC trend inference unit constructs a spatial steam pressure dynamic model by the heat flow vector field and the exothermic rate sequence, generates the VOC mass concentration curve and the concentration change gradient, and provides the concentration evolution basis for the steam cloud formation risk identification.

[0100] The threshold adaptive unit is used to write the latest batch data after obtaining the batch production completion signal, including the exothermic rate deviation, the VOC concentration deviation and the response control amount deviation.

[0101] Upon receiving the latest batch data, the incremental Bayesian updating algorithm is called to update the posterior distribution of the risk threshold, with the aim of minimizing the risk prediction bias. The historical batch bias data is updated a posteriori, and the threshold solidification process is triggered when the confidence interval shrinks, the prediction error converges, or the sample capacity saturates. The threshold table with confidence interval is output, including the lower confidence limit, the upper confidence limit and the expected value.

[0102] The threshold table is published to the risk buffer area and the adaptive interlocking control module through the internal message queue.

[0103] The threshold adaptive unit uses the incremental Bayesian learning method to update the historical batch bias data a posteriori and incrementally correct the posterior distribution based on the control response error vector, optimizing the parameters in the threshold table.

[0104] The threshold adaptive unit realizes dynamic adjustment and a posteriori optimization of the risk threshold through incremental Bayesian updating of historical batch bias data, forms a threshold table with confidence interval and drives the adaptive interlocking logic, improving the real-time and accuracy of risk assessment.

[0105] In addition, the threshold adaptive unit also receives control response data packets from the adaptive interlocking control module, and extracts the error amount between the actual response and the expected response according to the preset deviation evaluation rules, forming a deviation vector.

[0106] After the deviation vector and the historical batch bias data are fused, they are input into the Bayesian updating function as incremental samples, so as to continuously update the posterior distribution and output the latest threshold table containing the expected value, confidence interval and adjustment amount.

[0107] This mechanism realizes continuous closed-loop optimization between prediction and control, improving the dynamic adaptability of threshold setting and the real-time correction ability of control strategy.

[0108] The risk prediction analysis module further receives control response data from the adaptive interlocking control module and is used to correct prediction parameters and update the input of the next round of risk assessment

[0109] The risk prediction analysis module realizes dynamic assessment of reaction risk and real-time adjustment of threshold by jointly modeling the reaction heat release, volatile organic compound concentration change and historical deviation data in the alkyd resin production process. This module combines stoichiometric parameters and thermodynamic state variables to form a continuous risk level output, providing a pre-judgment basis for digital twin simulation and interlocking control.

[0110] The digital twin simulation module includes a virtual reaction kettle instance unit, a steam cloud diffusion simulation unit and an interlocking strategy pre-play unit.

[0111] The virtual reactor instance unit is used to build a three-dimensional finite element model, and model parameters of the three-dimensional finite element model include shell thickness distribution, agitator blade type, heat conduction oil jacket structure and pressure support layout.

[0112] The above parameters are static offline input data, which are obtained from process design drawings, structure detection reports and manufacturing materials in the device modeling stage, loaded at system initialization, and kept constant during the running process.

[0113] The shell thickness measurement value, stirring torque measurement value and feed temperature measurement value are received, and the shell thickness measurement value, stirring torque measurement value and feed temperature measurement value are mapped to the corresponding finite element boundary conditions.

[0114] The explicit coupling solver is called to perform transient iteration on the thermal stress field, and the shell node temperature matrix, shell principal stress matrix and blade shaft torque distribution vector are output.

[0115] The node number list, temperature matrix, principal stress matrix and time stamp are written into the simulation cache area.

[0116] When the simulation cache area is simulated and run, the thermal stress response model based on principal component simplification is called to output the shell principal stress matrix and temperature distribution estimation value within the preset time window.

[0117] The steam cloud diffusion simulation unit is used to obtain the heat flow vector field and VOC concentration curve, and generate an unstructured three-dimensional CFD grid combined with the geometric boundary conditions of the workshop.

[0118] In each preset time step, the incompressible Navier-Stokes solver based on pressure-velocity coupling is called and the component transport equation is superimposed to calculate the mass fraction field of the volatile component.

[0119] The volume fraction threshold interception algorithm is executed to extract the lower explosive limit corresponding to the isosurface and output the isosurface vertex coordinate set, grid index mapping table and time identifier.

[0120] The isosurface vertex coordinate set, grid index mapping table and time identifier are written into the simulation cache area through the IPC shared memory.

[0121] The steam cloud diffusion simulation unit generates an unstructured three-dimensional CFD grid using the heat flow vector field and VOC concentration curve, dynamically calculates the volatile component mass fraction field and extracts the lower explosive limit isosurface, and provides spatial distribution data for the determination of potential steam cloud formation area and diffusion boundary.

[0122] The interlocking strategy pre-play unit is used to build an equivalent discrete event driven model according to the hardware PLC logic table after obtaining the control instruction sequence simulated by the adaptive interlocking control module.

[0123] Inject control instructions as event trigger conditions into the virtual reactor instance unit and the vapor cloud diffusion simulation unit at every preset discrete event cycle and advance the simulation clock synchronously.

[0124] When the simulation end signal is acquired, call the risk constraint checking function to determine the shell principal stress matrix, node temperature matrix, and explosive vapor cloud volume fraction:

[0125] If the determination result is within the threshold value table limit range, generate an instruction approval identifier and write the approval state, control instruction sequence, and simulation annotation into the control instruction queue.

[0126] If any determination result exceeds the threshold value, generate an instruction rejection identifier and record the rejection reason code and maximum deviation parameter to the rejection log file.

[0127] The interlocking strategy pre-play unit performs discrete event-driven simulation on the control instructions before they are issued, verifies whether the shell principal stress, node temperature, and vapor cloud volume fraction meet the threshold requirements after execution, realizes pre-determination of instruction approval or rejection, and provides real-time safety checking for interlocking execution.

[0128] The digital twin simulation module is used to perform second-level fast verification of the control strategy during the generation stage to ensure that its execution will not cause equipment overload or explosive environmental risk. To achieve fast response, this module uses a multi-layer precision model design, pre-constructs a parameter mapping table through fine finite element simulation and CFD modeling in the early stage, and calls a reduced-order surrogate model for approximate deduction during operation.

[0129] The adaptive interlocking control module includes a reflux adjustment unit, a heat load reduction unit, and an inert gas protection unit.

[0130] The reflux adjustment unit is used to subscribe to real-time VOC mass concentration values in the data pool and generate target threshold values by calling lower explosive limit (LEL) data in the configuration file. When the real-time VOC mass concentration value reaches a preset first concentration threshold, the following control is performed:

[0131] Read the real-time reflux mass flow value, calculate the target reflux increment, and send an analog output signal to the frequency converter, which is used to increase the reflux pump speed.

[0132] In each sampling period, the deviation of the flowmeter readback value from the target value is checked. If the deviation exceeds a preset percentage, a correction pulse is sent.

[0133] When the real-time VOC mass concentration value is below a preset second concentration threshold for a preset time period, the reflux pump speed is adjusted in the opposite direction to the steady-state value at the same slope, and the adjustment curve is archived.

[0134] The reflux adjustment unit dynamically calculates the target reflux increment and adjusts the reflux pump speed based on the comparison of the real-time VOC mass concentration value and the lower explosion limit threshold, realizing closed-loop control of the solvent reflux rate and providing real-time flow adjustment capability for inhibiting VOC accumulation in the vapor space.

[0135] The heat load reduction unit is used to obtain a wall temperature sequence and calculate a temperature rise rate.

[0136] When the temperature rise rate exceeds a preset first temperature rise rate threshold, a bypass valve opening degree adjustment logic is triggered, and the bypass valve opening degree adjustment logic is:

[0137] An opening degree control signal is sent to the electrical converter, and the jacket heat transfer oil outlet temperature is sampled.

[0138] If the jacket heat transfer oil outlet temperature is higher than the target temperature, the bypass valve opening degree is increased to a first opening degree value.

[0139] If the temperature rise rate threshold is lower than a preset second temperature rise rate threshold and the jacket heat transfer oil outlet temperature is less than or equal to the target temperature, the bypass valve opening degree is decreased to a second opening degree value.

[0140] The valve position change, temperature rise rate correction amount, and observation time scale are written into a control log.

[0141] The heat load reduction unit calculates the temperature rise rate through the wall temperature sequence and drives the bypass valve opening degree adjustment, proportionally distributes the heat load of the heat transfer oil, realizes dynamic control of the reactor wall temperature, and provides temperature adjustment basis for preventing local overheating and thermal runaway.

[0142] The inert gas protection unit is used to receive the oxygen content in the reactor and the pressure in the reactor in real time through a serial bus, and when the VOC concentration risk index and the oxygen content deviation index meet the triggering conditions at the same time, output a nitrogen flow control signal.

[0143] The inert gas protection unit generates a nitrogen flow control signal by real-time monitoring of the oxygen content and pressure in the reactor, reduces the oxygen content in the reactor under the premise of maintaining the upper limit of the pressure safety margin, and ensures that the reaction environment is maintained under low-oxygen conditions, thereby providing a continuous protection mechanism for preventing the formation of explosive mixtures.

[0144] After the control instruction is executed, the system automatically collects the change data of the key response parameters in the current control process, including but not limited to: the actual concentration change after the reflux mass flow adjustment, the wall temperature response trend after the temperature rise control, and the oxygen content response curve in the reactor after the nitrogen injection.

[0145] The above response data is packaged into a control response data packet in real time and returned to the risk prediction analysis module and the threshold adaptive unit through an internal message queue.

[0146] The risk prediction analysis module compares the prediction results with the actual response, identifies the existing deviation indicators, and fine-tunes the prediction model input weights or fitting parameters.

[0147] The threshold adaptive unit updates the risk threshold distribution based on the error amount of the current control response, response delay, and maximum deviation, and further optimizes the control strategy generation basis for the next period.

[0148] The adaptive interlocking control module is triggered based on the heat release rate sequence, predicted wall temperature sequence, and VOC mass concentration curve output by the risk prediction analysis module.

[0149] The adaptive interlocking control module realizes dynamic control of solvent reflux rate, heat load distribution, and nitrogen protection through real-time monitoring and closed-loop adjustment of key parameters such as VOC concentration, wall temperature sequence, and in-pot oxygen content, forming an interlocking system linked with the risk prediction model. This module can automatically correct parameters when the reaction state changes rapidly, ensuring that the reaction kettle operates within the safe process boundary.

[0150] The human-computer interaction warning module includes a three-dimensional visualization terminal unit, a hierarchical warning push unit, and a semantic analysis broadcast unit.

[0151] The three-dimensional visualization terminal unit is used to display the three-dimensional model of the workshop in the central control room and mobile terminal, and to superimpose the isosurface layer generated by the VOC mass concentration curve output by the risk prediction analysis module in real time.

[0152] The three-dimensional visualization terminal unit superimposes the VOC concentration isosurface and spatial distribution on the three-dimensional model of the workshop, realizing the visualization of multiple parameter states and providing spatial positioning information of the reaction zone environment and risk area for operators.

[0153] The hierarchical warning push unit compares the current prediction results and actual monitoring data based on the confidence interval in the threshold table generated by the threshold adaptive unit, makes a level judgment, and pushes the warning information corresponding to the risk level to the DCS system, mobile terminal, and duty room large screen.

[0154] The hierarchical warning push unit synchronizes the warning information to the DCS system, mobile terminal, and duty room large screen according to the risk level, realizes multi-channel parallel notification, and ensures that personnel at different operation levels and positions can receive risk prompts in the first time.

[0155] The semantic analysis broadcast unit is used to convert the key control instructions and warning information generated in the control module into semantic prompts.

[0156] The semantic analysis broadcast unit converts the key control password into a short format voice prompt, and reduces the ambiguity of understanding through semantic structure, realizes the instant and accurate transmission of operation instruction, and provides an audio instruction channel for the rapid execution of on-site personnel.

[0157] The man-machine interaction early warning module realizes the rapid transmission and on-site execution of monitoring information and control decisions through multi-dimensional visual display, hierarchical alarm push and semantic instruction broadcast. The module establishes a real-time information channel between operators, control systems and on-site terminals, ensures that risk information is output in an intuitive, hierarchical and operable manner, and improves the speed of instruction response and transmission accuracy.

[0158] The compliance record tracing module includes an environmental emission report generation unit, a safety event fingerprint on-chain unit and an external interface unit:

[0159] The environmental emission report generation unit is used to generate unit production emission rate and total emission based on VOC emission data collected by the data fusion collection module and actual VOC concentration change trend after execution of the control module, and form structured emission records.

[0160] The environmental emission report generation unit compiles the total VOC emission and unit production emission rate collected and generates a standardized report, providing structured emission data archives for internal audit and environmental protection supervision of enterprises.

[0161] The safety event fingerprint on-chain unit is used to combine and encode the predicted risk records generated by the risk prediction analysis module, the control response results generated by the control module, and the alarm sequence triggered by the man-machine interaction module, generate a hash fingerprint and write it into a private chain.

[0162] The safety event fingerprint on-chain unit generates a hash fingerprint for the alarm sequence and writes it into a private chain, ensuring that the safety event record has a timestamp and uniqueness, and realizing the non-tamperable traceability of the whole life cycle.

[0163] The external interface unit is used to provide a standardized environmental compliance data access interface to the ecological environment supervision platform through the OAuth2.0 protocol.

[0164] The external interface unit outputs a standardized data interface through the OAuth2.0 protocol, so that the compliance emission information and safety records can be safely connected and accessed online with the ecological environment supervision platform.

[0165] The compliance record tracing module realizes the emission compliance archive of the production process and the traceable record of safety events by uniformly managing VOC emission data, alarm events and access interfaces. The module establishes a data closed loop from on-site collection to external supervision, ensuring that all environmental and safety information has completeness, non-tamperability and standardized connection capability.

[0166] The digital twin simulation module further comprises a parameter mapping extraction mechanism and a fast proxy model calling mechanism, which are used to call the reduced-order proxy model to perform second-level fast simulation based on the stored working condition mapping data during simulation running, and generate the main stress, temperature distribution and steam cloud volume distribution results corresponding to the control strategy.

[0167] The steam cloud diffusion simulation unit continuously calculates the steam cloud boundary and predicts the farthest diffusion distance based on the real-time updated three-dimensional grid and multi-step time marching algorithm.

[0168] The security event fingerprint chaining unit records the block timestamp and node signature immediately after generating the hash fingerprint.

[0169] The present application has obvious technical progress in safety control and environmental protection compliance, realizes second-level multi-source data fusion and process variable collaborative modeling, constructs a multi-channel data pool of VOC spectrum, heat flux density, acid value titration and reflux monitoring, etc., ensures the timeliness and parameter integrity of monitoring, the collaborative prediction ability of VOC concentration, realizes the reaction trend prediction and risk level dynamic adjustment through piecewise linear fitting, cubic interpolation and Bayesian adaptive mechanism, introduces a digital twin simulation module to verify the potential stress overload and steam cloud diffusion in parallel before control execution, ensures that the control instruction has safety margin, the control instruction issuing closed-loop mechanism supports the instruction approval and rejection based on the result feedback, and retains the complete simulation annotation and deviation parameters, opens up the data chain from the field to the supervision, ensures that the alarm and emission records are tamper-proof and traceable through the fingerprint chaining and OAuth interface mechanism, meets the multi-terminal and multi-dimensional interaction demand, supports 3D visual model and semantic broadcast linkage, improves the operator response efficiency and understanding accuracy, and supplements the structural defects of traditional DCS system in risk pre-play, dynamic threshold adaptation and compliance evidence, and has good industrial application prospect.

[0170] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (or computer- readable storage media) having computer-usable program code embodied in the medium. The medium can be any available storage media that can be accessed by a computer. By way of example, and not limitation, such computer-usable storage media can include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other storage medium(s) that can be used to carry or store desired computer program code in the form of instructions or data structures and that can be accessed by a computer. Also, the present application can be embodied in a computer program product that can be traded as goods or merchandise, through a computer-based platform or Figure 1 one or more functions specified in the flow or flows and / or blocks Figure 1 one or more functions specified in the flow or flows and / or blocks

[0171] It should be noted that the above-mentioned embodiments are only used to illustrate but not to limit the technical solutions of the present application, and although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application, and they should be covered in the scope of the claims of the present application.

Claims

1. An environmental safety early warning system for alkyd resin production, characterized in that, It includes a data fusion and acquisition module, a risk prediction and analysis module, a digital twin simulation module, an adaptive interlocking control module, a human-machine interaction early warning module, and a compliance record traceability module; The data fusion and acquisition module is used to construct a multi-source real-time data pool; The risk prediction and analysis module is used to output a risk level based on the data pool; The digital twin simulation module is used to verify the control strategy; The adaptive interlocking control module is used to issue control commands; The human-computer interaction early warning module is used to execute tiered alarms; The compliance record traceability module is used to solidify emission and alarm records; The data fusion and acquisition module includes a VOC spectrum acquisition unit, a heat flux density monitoring unit, an acid value microtitering unit, and a solvent reflux monitoring unit; The VOC spectral acquisition unit is used to acquire the component mass concentration and write it into the data pool; The heat flux density monitoring unit is used to output the temperature gradient scanning signal between the reactor wall and the condenser outer wall and to form a heat flux vector field. The acid value microtiter unit is used to collect the acid value and moisture content of the resin sample solution and output the real-time acid value sequence. The solvent reflux monitoring unit is used to acquire reflux mass flow rate and reflux temperature and synchronize them to the data pool. The risk prediction and analysis module includes a heat release rate inference unit, a VOC trend inference unit, and a threshold adaptive unit. After receiving the real-time acid value sequence, the exothermic rate inference unit calculates the current reaction rate constant based on the reaction order coefficient, resin formulation parameters and reaction heat constant in the historical sample library, which is established by dynamic titration experiments corresponding to typical resin formulations. Integrate the reaction heat within a preset prediction window and output the heat release rate sequence and the predicted wall temperature sequence; The predicted wall temperature sequence is extracted using the longest peak search algorithm, the highest wall temperature is bound to a time identifier, and then written into the risk buffer. The VOC trend inference unit is used to obtain the heat flux vector field and heat release rate sequence, calculate the local temperature rise at the top of the vessel based on the average heat flux density at the top of the vessel, and call the vapor pressure analytical model to calculate the instantaneous vapor pressure in the head space. The vapor pressure analytical model calculates the local temperature estimate by fitting it with the pre-stored component vapor-liquid equilibrium parameters based on the Antoine equation. The component vapor-liquid equilibrium parameters include the vapor pressure constant and the fugacity coefficient, and outputs vapor pressure time series data. The vapor pressure time series data is used to calculate the corresponding VOC mass concentration based on the preset vapor phase conversion coefficient. The VOC mass concentration change curve is generated by cubic spline interpolation and written to the risk buffer at preset recording intervals. Write the peak value, the time of peak occurrence, and the slope of concentration increase of the VOC mass concentration curve into the risk assessment table; The threshold adaptive unit is used to write the latest batch data after acquiring the batch production completion signal. The latest batch data includes heat release rate deviation, VOC concentration deviation, and response control quantity deviation. After receiving the latest batch of data, the incremental Bayesian update algorithm is called to update the posterior distribution of the risk threshold. With the goal of minimizing the risk prediction bias, the historical batch bias data is updated posteriorly. The threshold solidification process is triggered when the confidence interval shrinks, the prediction error converges, or the sample size is saturated. The threshold table with confidence intervals is output, which includes the lower confidence limit, the upper confidence limit, and the expected value. The threshold table is published to the risk buffer and adaptive interlocking control module via an internal message queue. The threshold adaptive unit uses the incremental Bayesian learning method to update the posterior data of historical batch deviation data and incrementally correct the posterior distribution based on the control response error vector, thereby optimizing the parameters in the threshold table. The risk prediction and analysis module further receives control response data from the adaptive interlocking control module and uses it to correct prediction parameters and update the risk assessment input for the next round.

2. The environmental safety early warning system for alkyd resin production as described in claim 1, characterized in that, The digital twin simulation module includes a virtual reactor instance unit, a steam cloud diffusion simulation unit, and an interlocking strategy pre-simulation unit. The virtual reactor instance unit is used to construct a three-dimensional finite element model. The model parameters of the three-dimensional finite element model include shell thickness distribution, stirrer blade shape, heat transfer oil jacket structure and pressure support layout. Receive shell thickness measurement, stirring torque measurement, and feed temperature measurement, and map the shell thickness measurement, stirring torque measurement, and feed temperature measurement to the corresponding finite element boundary conditions; The explicit coupled solver is invoked to perform transient iteration on the thermal stress field, and outputs the shell nodal temperature matrix, the shell principal stress matrix, and the blade shaft torque distribution vector. Write the node number list, temperature matrix, principal stress matrix and timestamp into the simulation buffer; The vapor cloud diffusion simulation unit is used to obtain the heat flow vector field and VOC concentration curve, and generate an unstructured three-dimensional CFD mesh in combination with the workshop geometric boundary conditions. Within each preset time step, the incompressible Navier-Stokes solver based on pressure-velocity coupling is invoked and the component transport equations are superimposed to calculate the mass fraction field of volatile components. Execute the volume fraction threshold truncation algorithm to extract the isosurface corresponding to the lower explosion limit and output the isosurface vertex coordinate set, grid index mapping table and time identifier; Write the isosurface vertex coordinate set, mesh index mapping table and time identifier into the simulation buffer via IPC shared memory; The interlocking strategy pre-simulation unit is used to construct an equivalent discrete event-driven model based on the hardware PLC logic table after obtaining the control instruction sequence proposed by the adaptive interlocking control module. In each preset discrete event cycle, control commands are injected as event triggering conditions into the virtual reactor instance unit and the steam cloud diffusion simulation unit, and the simulation clock is advanced synchronously. When obtaining the simulation end signal, the risk constraint verification function is called to determine the principal stress matrix of the shell, the nodal temperature matrix, and the volume fraction of the explosive vapor cloud: If the judgment result is within the threshold table limit, an instruction approval identifier is generated and the approval status, control instruction sequence, and simulation annotation are written to the control instruction queue. If any judgment result exceeds the threshold, an instruction rejection flag is generated and the rejection reason code and maximum deviation parameter are recorded to the rejection log file.

3. The environmental safety early warning system for alkyd resin production as described in claim 2, characterized in that, The adaptive interlocking control module includes a reflux regulation unit, a heat load reduction unit, and an inert gas protection unit; The reflux regulation unit is used to subscribe to the real-time value of VOC mass concentration in the data pool and call the lower limit explosion limit (LEL) data in the configuration file to generate a target threshold. When the real-time value of VOC mass concentration reaches the preset first concentration threshold, the following control is executed: Read the real-time value of the reflux mass flow rate, calculate the target reflux increment, and send an analog output signal to the variable frequency drive. The analog output signal is used to increase the speed of the reflux pump. In each sampling cycle, the deviation between the flow meter readout value and the target value is checked. If the deviation exceeds a preset percentage, a correction pulse is issued. When the real-time value of VOC mass concentration is lower than the preset second concentration threshold within a preset time period, the reflux pump speed is adjusted in the opposite direction with the same slope to the steady-state value and the adjustment curve is archived. The heat load reduction unit is used to acquire the wall temperature sequence and calculate the temperature rise rate. When the temperature rise rate exceeds a preset first temperature rise rate threshold, the bypass valve opening adjustment logic is triggered. The bypass valve opening adjustment logic is as follows: Send opening control signals to the electrical converter and sample the jacket heat transfer oil outlet temperature; If the outlet temperature of the jacket heat transfer oil is higher than the target temperature, the opening degree of the bypass valve will be increased to the first opening degree value. If the temperature rise rate threshold is lower than the preset second temperature rise rate threshold and the jacket heat transfer oil outlet temperature is less than or equal to the target temperature, then the bypass valve opening will be reduced to the second opening value. Write the valve position change, temperature rise rate correction, and observation time scale into the control log; The inert gas protection unit is used to receive the oxygen content and pressure inside the reactor in real time via a serial bus, and outputs a nitrogen flow control signal when the VOC concentration risk index and the oxygen content deviation index simultaneously meet the trigger conditions. The adaptive interlocking control module is triggered based on the heat release rate sequence, predicted wall temperature sequence, and VOC mass concentration curve output by the risk prediction and analysis module.

4. The environmental safety early warning system for alkyd resin production as described in claim 3, characterized in that, The human-computer interaction early warning module includes a three-dimensional visualization terminal unit, a hierarchical alarm push unit, and a semantic parsing and broadcasting unit. The three-dimensional visualization terminal unit is used to display the three-dimensional model of the workshop in the central control room and on mobile devices, and to overlay the isosurface layer generated by the VOC mass concentration curve output by the risk prediction and analysis module in real time. The hierarchical alarm push unit is used to compare the confidence interval in the threshold table generated by the threshold adaptive unit with the current prediction results and actual monitoring data to determine the level, and push the alarm information of the corresponding risk level to the DCS system, mobile terminal and duty room screen respectively. The semantic parsing and broadcasting unit is used to convert key control commands and alarm information generated in the control module into semantic prompts.

5. The environmental safety early warning system for alkyd resin production as described in claim 4, characterized in that, The compliance record traceability module includes an environmental emission report generation unit, a security event fingerprint uploading unit, and an external interface unit: The environmental emission report generation unit is used to generate the unit output emission rate and total emission based on the VOC emission data collected by the data fusion acquisition module and the actual VOC concentration change trend after the control module is executed, and to form a structured emission record. The security event fingerprinting unit is used to combine and encode the predicted risk records generated by the risk prediction and analysis module, the control response results generated by the control module, and the alarm sequences triggered by the human-computer interaction module to generate a hash fingerprint and write it into the private chain. The external interface unit is used to provide a standardized environmental compliance data access interface to the ecological environment supervision platform through the OAuth2.0 protocol.

6. The environmental safety early warning system for alkyd resin production as described in claim 5, characterized in that, The digital twin simulation module also includes a parameter mapping extraction mechanism and a fast proxy model invocation mechanism. The parameter mapping extraction mechanism and the fast proxy model invocation mechanism are used to call the reduced-order proxy model based on the stored working condition mapping data to perform second-level fast simulation during simulation, and generate principal stress, temperature distribution and steam cloud volume distribution results corresponding to the control strategy.

7. The environmental safety early warning system for alkyd resin production as described in claim 6, characterized in that, The vapor cloud diffusion simulation unit continuously calculates the vapor cloud boundary and predicts the farthest diffusion distance based on a real-time updated three-dimensional mesh and a multi-step time progression algorithm.

8. The environmental safety early warning system for alkyd resin production as described in claim 7, characterized in that, The security event fingerprinting unit immediately records the block timestamp and node signature after generating the hash fingerprint.

Images