Chemical safety data acquisition control system and operation method

Abstract

The present application relates to the technical field of chemical process safety monitoring and automatic control, in particular to a kind of chemical safety data acquisition control system and operating method, comprising: obtaining the raw data of reaction kettle temperature and pressure and filtering verification to extract characteristic parameters;According to the characteristic parameters, calculate the dynamic heat difference value, calculate the cooling water flow and the feed cut-off ratio when the system is unbalanced to generate basic instructions;According to the unbalanced state and the fluid transmission delay time, determine the risk source and the cut-off node, generate the cut-off instruction;Combining the second derivative of temperature or pressure with the mechanism lag time to obtain the expected parameter increment and add to the basic instruction;Predict the temperature or pressure value after parameter backfall after execution, if it will be lower than the safety lower limit, generate reverse heat compensation instruction;Send each instruction to the field execution mechanism, realize the state inhibition and parameter recovery of exothermic reaction device.

Description

Technical Field

[0001] This invention relates to the field of chemical process safety monitoring and automatic control technology, specifically a chemical safety data acquisition and control system and its operation method. Background Technology

[0002] Currently, the safety monitoring and control mechanisms of chemical reaction units are usually based on threshold alarms or conventional interlocking strategies for single operating parameters such as temperature and pressure. When the reactor experiences accelerated exothermic reaction, enhanced temperature and pressure coupling, or abnormal material transport along pipelines, existing systems often cannot effectively filter and cross-validate the collected data, nor can they combine imbalance states, fluid transport delays, and actuator lags for coordinated control. In the process of handling abnormal operating conditions, existing technologies generally suffer from problems such as delayed judgment, unreasonable selection of cut-off nodes, or passive cooling adjustments that easily lead to overshoot and abnormal parameter drops, resulting in prolonged handling time for abnormal operating conditions and easily causing secondary fluctuations in system parameters. Summary of the Invention

[0003] The purpose of this invention is to provide a chemical safety data acquisition and control system and its operation method, which solves the following technical problems: the existing safety monitoring and control mechanisms for chemical reaction devices are insufficient in terms of effective filtering and cross-validation of acquired data, as well as linkage control based on imbalance state, fluid transmission delay time and actuator lag. There is a need for a chemical safety data acquisition and control system and its operation method that can filter and preprocess and cross-validate operating data, dynamically identify imbalance state, select risk cut-off nodes based on fluid transmission delay, and implement feedforward compensation and anti-overshoot control to achieve state suppression and parameter callback.

[0004] The objective of this invention can be achieved through the following technical solutions: On one hand, the present invention provides a chemical safety data acquisition and control system, comprising: a data acquisition and processing module, used to acquire raw operating condition data of temperature and pressure of the reactor in the chemical system, the lower limit of temperature safety, the lower limit of pressure safety, and the lag time of each field actuator; and to extract operating condition characteristic parameters based on the raw operating condition data; The state judgment module inputs the operating condition characteristic parameters into a pre-constructed state space model, determines the dynamic heat difference of the reactor based on the state space model, and determines the imbalance state of the chemical system based on the dynamic heat difference, and generates basic control commands. The risk cut-off module is used to determine the risk source based on the imbalance state, and based on the risk source, obtain the pipeline parameters and current material flow rate of the process pipeline network in the chemical system; retrieve the topology matrix of the process pipeline network and combine it with the pipeline parameters and the current material flow rate to determine the fluid transmission delay time, and generate a cut-off command; The feedforward control module is used to combine the lag time of each field actuator, the operating condition characteristic parameters and the original operating condition data to determine the expected parameter increment, and to correct the basic control command according to the expected parameter increment. The prediction and compensation module is used to continuously track and predict the lowest predicted value after the temperature or pressure inside the reactor drops after the basic control command or the cut-off command is executed, and to generate a reverse heat compensation command based on the lowest predicted value, the temperature safety lower limit and the pressure safety lower limit. The control output module is used to send the basic control command, the cut-off command, and the reverse heat compensation command to the corresponding field actuators.

[0005] On the other hand, the present invention provides a method for operating a chemical safety data acquisition and control system, comprising the following steps: S1. Obtain the original operating condition data of temperature and pressure of the reactor in the chemical system, the lower limit of temperature safety, the lower limit of pressure safety, and the lag time of each field actuator; extract the operating condition characteristic parameters based on the original operating condition data; S2. Calculate the dynamic heat difference of the reactor based on the operating condition characteristic parameters, determine the imbalance state of the chemical system based on the dynamic heat difference, and generate basic control commands. S3. Determine the risk source based on the imbalance state, obtain the pipeline parameters and current material flow rate of the process pipeline network in the chemical system; retrieve the topology matrix of the process pipeline network and determine the fluid transmission delay time by combining the pipeline parameters and the current material flow rate, and generate a cut-off command. S4. Read the lag time of the corresponding field actuator, determine the expected parameter increment based on the working condition characteristic parameters and the original working condition data combined with the lag time, and correct the basic control command according to the expected parameter increment; S5. After executing the basic control command or the cut-off command, predict the lowest predicted value after the temperature or pressure inside the reactor drops, generate a reverse heat compensation command based on the lowest predicted value, the lower limit of temperature safety, and the lower limit of pressure safety, and send the basic control command, the cut-off command, and the reverse heat compensation command to the corresponding field actuator.

[0006] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention uses a data acquisition and processing module to perform cross-validation on the temperature and pressure sequences in the original operating condition data, and extracts operating condition feature parameters after filtering and validation, thus solving the problem that existing systems cannot effectively filter and cross-validate the acquired data; 2. This invention uses a risk cut-off module to retrieve the topology matrix and calculates the fluid transmission delay time by combining it with the pre-acquired pipeline parameters of the process network. This overcomes the problems of actuator lag and overshoot and abnormal parameter drop caused by passive cooling adjustment in the background technology, improves the response speed of control commands when handling abnormal operating conditions, and reduces the oscillation amplitude during parameter adjustment. Attached Figure Description

[0007] 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.

[0008] Figure 1 This is a schematic diagram of a chemical safety data acquisition and control system provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating the operation method of a chemical safety data acquisition and control system provided in an embodiment of the present invention. Detailed Implementation

[0009] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0010] A chemical safety data acquisition and control system includes: a data acquisition and processing module for acquiring raw operating condition data of temperature and pressure of the reactor in the chemical system, the lower limit of temperature safety, the lower limit of pressure safety, and the lag time of each field actuator; and extracting operating condition characteristic parameters based on the raw operating condition data. The state judgment module inputs the operating condition characteristic parameters into a pre-built state space model, determines the dynamic heat difference of the reactor based on the state space model, and determines the imbalance state of the chemical system based on the dynamic heat difference, and generates basic control commands. The risk cut-off module is used to determine the risk source based on the imbalance state, and based on the risk source, obtain the pipeline parameters and current material flow rate of the process pipeline network in the chemical system; retrieve the topology matrix of the process pipeline network and combine the pipeline parameters and current material flow rate to determine the fluid transmission delay time, and generate a cut-off command. The feedforward control module is used to combine the lag time of each field actuator with the operating condition characteristic parameters and the original operating condition data to determine the expected parameter increment, and to correct the basic control command according to the expected parameter increment. The prediction and compensation module is used to continuously track and predict the lowest predicted value after the temperature or pressure inside the reactor drops after the basic control command or the cut-off command is executed, and to generate a reverse heat compensation command based on the lowest predicted value, the temperature safety lower limit and the pressure safety lower limit. The control output module is used to send basic control commands, cut-off commands, and reverse heat compensation commands to the corresponding field actuators.

[0011] This embodiment provides a linkage control mechanism for a chemical safety data acquisition and control system. Please refer to [link / reference]. Figure 1 This embodiment is applied to a chemical reactor production line, which includes a reactor body, a jacketed cooling circuit, an upstream feed pipeline, a downstream buffer tank, an emergency shut-off valve, a material conveying pump, a heat tracing system, and a distributed control system. The field actuators in this embodiment mainly include an emergency shut-off valve and a material conveying pump for cutting off and conveying, as well as a jacketed cooling valve, a proportional valve of the heat tracing system, and a variable frequency cooling water pump for temperature regulation. The lag time of the field actuators is calculated by acquiring the mechanical dead time of each actuator and the time required for the transmission of the corresponding control medium, and adding the two together to form the total lag time. If the lag time data of the actuators is missing during the acquisition process, the maximum nominal lag value in the corresponding equipment file is directly retrieved for calculation. The data acquisition and processing module continuously receives process quantities such as reactor temperature, reactor top pressure, liquid level, and stirring speed. For example, it continuously samples multiple sets of gradually increasing temperature values ​​and corresponding pressure values. If the latest set of data experiences a sudden increase that does not conform to normal physical laws within a short period of time, the module performs filtering to remove transient abnormal signals that exceed the preset thermal inertia rate of change threshold of the equipment. Cross-validation is performed. For example, under the same material ratio and gas phase volume conditions, an increase in temperature is usually accompanied by an increase in pressure. If the rate of temperature rise exceeds a set threshold while the rate of pressure change approaches zero, or if the pressure rises sharply while the temperature remains stable, then at least one type of measurement deviation risk is identified. After filtering and validation, the module extracts characteristic parameters of the operating conditions, such as the rate of temperature rise, pressure fluctuation variance, and liquid level fluctuation index. Based on the temperature change and sampling interval within adjacent effective sampling periods, the current rate of temperature rise is extracted. If the rate of temperature rise in subsequent sampling periods is significantly greater than that in the previous period, it indicates that the temperature rise trend is accelerating. After receiving the above characteristic parameters, the status determination module calculates the heat difference between the current net heat release capacity of the reaction system and the maximum heat dissipation capacity of the cooling system; the specific formula is as follows: In the formula, This is the dynamic heat difference; The total generated heat output power of the reactor at the current moment is estimated by the state observer; This refers to the actual heat dissipation power provided by the cooling system under current operating conditions; when When it indicates that cooling can still cover the heat release; when And over time first derivative When this occurs, it indicates an increased risk of system imbalance.

[0012] Calculation of fluid transport delay time: Assuming that the calculated net heat release power at a certain moment is greater than the actual heat dissipation power of the current cooling system, a positive dynamic heat difference is generated; if this dynamic heat difference shows an increasing trend in subsequent consecutive sampling periods, it can be considered that the imbalance is intensifying; at this time, the module calculates the target cooling water flow rate and feed cutoff ratio, and generates basic control commands; specifically, the system uses the following physical quantification formula for calculation: in, Set the target cooling water flow rate. This is the current cooling water flow rate. The specific heat capacity of cooling water. The density of the cooling water, The outlet water temperature, The constant 3600 represents the inlet water temperature and is a unit conversion factor for the time dimension. in, This refers to the feed cutting ratio. The additional heat power under maximum feed conditions; in, For the time window of safe intervention To preset a safe temperature threshold, The current temperature. This represents the current rate of temperature rise.

[0013] Basic control commands can include combinations of increasing the cooling water valve opening from 60% to 90% and reducing the material conveying pump speed to 20% of its original speed or directly cutting it off; the risk cut-off module is used to perform path-level cut-off based on the location of the abnormal source and the process topology; this module can abstract the connection relationship between the reactor outlet, circulation branch, feed branch and downstream buffer tank into a topology matrix; Assume there are four nodes to ,in For reaction vessel, For discharge valve, For the inlet valve of the buffer tank, If it is a reflux pump, then in the connection matrix arrive , arrive , arrive The corresponding position can be set to 1, and other unconnected positions are set to 0; the risk disconnection module retrieves the topology matrix reflecting the fluid connectivity and uses the following formula to calculate the distance of abnormal fluid from the risk source to each downstream candidate node. Fluid transport delay time : In the formula, To reach the node Total delay time; From risk source to node The number of consecutive pipe segments contained between them; For the first The length of the pipe section; For the current material in the first The actual flow velocity within the pipe section; the system compares this delay time with the completion time of the actuators at each node. Only those that meet the conditions The nodes are included in the candidate process node sequence; Furthermore, the feedforward control module is used to compensate for the regulation lag caused by system imbalance and actuator action delay; by monitoring the current absolute value of temperature or pressure and its second-order change trend; if the first derivative of temperature shows a continuous upward trend, the second derivative can be considered positive, indicating that the temperature rise rate shows an accelerating increasing trend; then the actuator lag time is read, and combined with the mechanical action dead zone of the actuator and the time required for medium transmission, the total lag time is constituted. By combining this second-order change trend with the lag time, an expected parameter increment is formed, which is used to amplify the control action in advance. Specifically, at the engineering calculation level, to avoid control delays caused by complex integral calculations, the feedforward control module uses the following formula to calculate the expected parameter increment generated during the actuator's action lag period. : In the formula, The temperature or pressure value in the original operating condition data is substituted independently into the temperature or pressure value according to the control target for calculation. This is the first derivative of the temperature or pressure in the current operating condition data; It is the second derivative of temperature or pressure; The preset process equivalent gain coefficient is dimensionless. The total delay time for the implementing agency; By introducing first- and second-order terms of Taylor expansion, the nonlinear dynamic change trend of parameters is accurately characterized. The response curvature is weighted by combining the process equivalent gain coefficient. The process equivalent gain coefficient is pre-calibrated based on the heating response curve of similar historical reactors through parameter identification. The second-order rate of change is accurately converted into the expected parameter increment within the time delay window. The parameter identification is specifically performed offline using the least squares method. The convergence condition is that the root mean square of the estimation error is less than the preset threshold. The equivalent additional rate of change is multiplied by the total lag time to obtain the expected parameter increment. The module will then pre-increase the valve opening or trigger the feed cut-off command in advance on the basis of the basic control command to avoid the situation where heat continues to accumulate after the command has been issued. The prediction and compensation module is used to predict the temperature or pressure values ​​after the temperature drops. If the temperature drops after the system starts maximum cooling, and the lowest predicted temperature value will be lower than the set safety limit within a preset time window based on the current parameter drop rate, the module will generate a reverse heat compensation command in advance, such as partially opening the proportional valve of the heat tracing system to 15%, or reducing the variable frequency cooling water pump from the rated speed to 70% to offset the residual cooling inertia. The control output module sends the cut-off command to the corresponding emergency cut-off valve and material conveying pump emergency stop circuit of the safety instrument system, and sends the reverse heat compensation command to the proportional valve or variable frequency cooling water pump controller of the heat tracing system.

[0014] Preferably, cut-off commands and compensation commands are sent through different priority queues. The former uses a safety cascade link, while the latter uses a process regulation link to ensure that the propagation of abnormal states is blocked first and parameter compensation is implemented. If the data acquisition and processing module detects the loss of temperature and pressure signals and cannot replenish them through neighbor sampling, the system directly enters the fault-safe mode and issues a default control to maintain the current highest cooling level and suspend new material feeding.

[0015] If the status judgment module detects If an abnormal jump occurs in a single sampling period but subsequently returns to normal, trend verification is performed through multiple preset consecutive sampling sub-periods to improve the robustness of the system's judgment and suppress false triggering by high-frequency interference. If the risk cutoff module calculates that multiple candidate nodes can act before arrival, the node with the smallest impact range is selected first; if none of them can meet the time limit, the upstream total cutoff node is directly triggered; if the second-order change trend obtained by the feedforward control module is negative, it means that the temperature and pressure have fallen back on their own, so the amplified feedforward amount is not superimposed to prevent over-adjustment. If the predicted value of the prediction compensation module is close to the safety lower limit, then graded compensation is used instead of one-time reverse heating; In an exemplary chemical exothermic reaction application scenario, when the reaction proceeds to the middle and later stages, the temperature and pressure of the reactor rise abnormally. After filtering and cross-validation, the system confirms the data is reliable and determines that the exothermic capacity has exceeded the cooling capacity for three consecutive cycles. The system calculates the basic control commands, increases the opening of the jacket cooling valve to the target opening, and sets the main feed cutoff ratio to 100%. At the same time, based on the pipeline length and material flow rate between the reactor and the buffer tank, the emergency shut-off valve closest to the reactor outlet is selected as the first execution point. Considering that there is still a delay of several seconds between the action of the cooling valve and the arrival of the cooling medium, the system adds feedforward compensation to issue the maximum output amplitude control command for the field actuator before the valve has fully reached its maximum opening. After the temperature drops from the peak, the system predicts that it will continue to drop and fall below the safety lower limit. The system generates a reverse heat compensation command to open the proportional valve of the heat tracing system to a low opening for a short time, so that the temperature stabilizes and returns to the safe range. This embodiment constructs a continuous closed-loop control chain by collecting data, determining the target, cutting off the control, feeding forward, preventing overshoot, and outputting the data. This enables the system to not only identify imbalances but also to make hierarchical responses to abnormal propagation paths, actuator hysteresis, and adjustment of residual inertia, thereby achieving rapid suppression and stable recovery of the exothermic reaction device. The data acquisition and processing module is used to perform cross-validation on the temperature and pressure sequences in the raw operating condition data.

[0016] The data acquisition and processing module deploys data acquisition cards and edge computing gateways between field instruments and the control system. The raw signals are processed at the edge computing layer before entering the control link. The data acquisition card can read signals from thermocouples, pressure transmitters, and level transmitters according to a fixed sampling period; under the fixed sampling period, it can continuously read multiple sets of temperature and pressure sequences; if the abrupt sample value exceeds the highest physical change rate threshold of the reaction system in the corresponding time window, the edge computing gateway can predict and correct the system state through extended Kalman filtering. For example, the predicted temperature of the system state at the previous moment is set as the first reference temperature, and the allowable natural variation range is within the preset fluctuation threshold. If the deviation between the current temperature sample value and the predicted value is greater than the preset allowable range, the data is determined to be transient high-frequency interference and suppressed, and the system outputs a corrected value that is closer to the predicted value. After filtering is completed, the edge computing gateway continues to perform cross-validation between temperature and pressure to overcome potential faults where a single variable is within a preset threshold but the coupling relationship between two variables is abnormal. In this reaction system, if the temperature rises, the pressure in the gas phase space of the reactor should usually rise synchronously; if the temperature remains around 92℃, but the pressure is consistently higher than the reasonable range calculated from the saturated vapor pressure curve, it is determined that the pressure sensor is experiencing zero-point drift. Conversely, if the pressure change is reasonable but the temperature deviates from the corresponding value of the gas-liquid phase equilibrium for a long time, it is determined that the temperature measuring point is attached with scale, causing the reading to lag. For example, the theoretical pressure range obtained through verification at a certain moment can be set to 0.17. Up to 0.20 The measured value remained at 0.24 for five consecutive cycles. If the above is true, the edge computing gateway will mark the pressure measurement point as suspicious and reduce its weight in the subsequent fusion calculation; based on the above, the gateway extracts the working condition characteristic parameters; the temperature rise rate is based on the stable trend of the sequence after removing transient high-frequency interference within a short time window, for example, after fitting [92.1, 92.3, 92.4, 92.6], it is obtained as 0.17℃ / 10s; Pressure fluctuation variance reflects whether there are abnormal fluctuations in the system with amplitudes greater than a preset amplitude threshold. For example, for the four most recent effective pressure values ​​[0.18, 0.18, 0.19, 0.19], the pressure fluctuation variance is lower than the set variance threshold, while for [0.18, 0.23, 0.17, 0.24], the pressure fluctuation variance is higher than the set variance threshold, indicating the presence of flash evaporation, pulse, or sensor status deviating from the preset normal state. The liquid level fluctuation index can be determined based on the ratio of the maximum difference to the average value of the liquid level within a short time window. That is, it is calculated by dividing the liquid level range within a preset time window by the average liquid level within that window. It is used to identify liquid level instability caused by feed disturbance, boiling, or agitation failure. After obtaining the above parameters, the gateway writes them into the shared memory area of ​​the distributed control system, so that subsequent modules can directly read them at a unified address, reducing cross-layer redundant calculations. If only one type of signal, temperature or pressure, is available, the gateway retains the univariate trend parameter but adds a confidence decline flag in the shared memory area, prompting the downstream to proceed with cautious judgment. If cross-validation finds that both types of signals deviate from the model simultaneously, and it is impossible to determine which measurement point is distorted, no forced correction is performed; instead, the original data is retained and an instrument maintenance flag is triggered. If the liquid level signal is temporarily interrupted, the liquid level fluctuation index uses the most recent valid value and maintains it for a preset time. After the maintenance time is exceeded, it is set to vacancy, and the downstream ignores this parameter without terminating the entire control process. During the operation of the reaction batch, if the temperature probe of the reactor is subjected to instantaneous interference and generates an abnormal transient extreme value signal, but the pressure value remains stable at the same time, the system will avoid misjudging it as thermal runaway. In another operation, the pressure transmitter experienced a persistently high reading due to liquid accumulation in the pressure guide tube, causing the temperature-pressure relationship to deviate from the gas-liquid equilibrium curve. The gateway identified this as a drift and reduced the weight of the pressure sequence. Ultimately, reliable characteristic parameters could still be provided based on the temperature rise rate and liquid level fluctuation. Through the above processing, the credibility and usability of the original operating condition data were improved, thereby enabling subsequent status judgments to be based on physically consistent, noise-controlled, and quantifiable inputs. The state judgment module runs within the programmable logic controller (PLC). Specifically, it inputs operating condition characteristic parameters into a pre-built state-space model, transforming the operating condition characteristics into a dynamic judgment of whether the current reaction's exothermic and heat dissipation capabilities exceed a preset safety threshold. The state-space model is a mathematical model characterizing the thermodynamic dynamic process, established using a system identification algorithm based on historical operating data of the reactor. It consists of state equations and observation equations. The state equation formula is: The observation equation is as follows: In the formula, For discrete time steps, the state vector This includes reactor temperature, pressure, material quantity, and average temperature of the jacket cooling water; control input vector. Including cooling water valve opening and feed rate; observation vector Including temperature rise rate, pressure fluctuation variance, and liquid level fluctuation index; system matrix Control input matrix and observation mapping matrix The dimensions are determined by the dimensions of the aforementioned vectors; This is the process noise vector. The observed noise vectors are all independent and identically distributed Gaussian white noise with zero mean and known covariance. The corresponding process noise covariance matrix and observed noise covariance matrix are pre-calibrated constant matrices. The process noise covariance matrix and observed noise covariance matrix are obtained by collecting historical background noise data of chemical reaction equipment under no-load operation and rated operating conditions, and calculating their statistical variances in advance. Matrix parameters of the state-space model , , The parameters were obtained offline using the least squares method by analyzing data samples from historically operating batches. Specifically, the parameter identification process involved: collecting multiple sets of historical time-series data sequences of the reactor under standard operating conditions; constructing a system identification equation with control input vector and state vector as regression variables; identifying the optimal element values ​​of the matrix parameters of the state-space model by solving the objective function that minimizes the sum of squared residuals; the identification convergence condition was that the root mean square of the fitting residuals of each observed variable after normalization was less than 0.05; where multi-source operating condition characteristic parameters were input into the state-space model to accurately estimate the current net heat release power. Specifically, through the state vector The total enthalpy of the reaction system is calculated by combining the reactor temperature and material quantity with preset specific heat capacity parameters, and then differentiated over time to obtain the current heat release power generated. Simultaneously, the actual heat dissipation power is calculated based on the average temperature of the jacket cooling water in the state vector. The difference between the two values ​​will output the dynamic heat difference value. .

[0017] For example, the estimated result at a certain moment is: the heat release power of the reaction is 410. Stirring and heat dissipation from the wall are negligible; the maximum heat dissipation capacity of the cooling system is 380. The dynamic heat difference is 30. The next sampling period is estimated to be 55. The next cycle is 85. If all three consecutive sampling periods show an upward trend, it indicates that the difference is diverging. To avoid misjudging a normal heating process with a temperature rise rate lower than the preset alarm value as an abnormal operating condition, this embodiment sets a dual threshold judgment condition: the temperature must exceed a preset safety threshold, and the temperature rise rate must exceed a preset alarm value. The safety threshold can be set to 95℃, and the alarm temperature rise rate can be set to 0.15℃ / s. When the temperature is 96.2℃ and the temperature rise rate is 0.22℃ / s, and the dynamic heat difference increases for three consecutive cycles, the system determines that it has entered the temperature runaway critical zone. After determining that an imbalance has been entered, the controller further reads the reaction activation energy, specific heat capacity and theoretical exothermic data of the batch of materials in the formula library in order to calculate the specific control adjustment amount; if the current material activation energy is lower than the preset activation energy threshold, it means that the reaction rate will be significantly accelerated after the temperature is further increased. At this time, the cooling water flow rate should be increased according to the preset high gain coefficient and the feed should be cut off at the maximum preset ratio. Assuming the calculation shows that an increase of 120 is needed Given the heat, cooling water inlet temperature, and allowable temperature rise, the controller calculates the required cooling water flow rate to be increased from 20... Increased to 34 Meanwhile, based on the additional heat that may result from the continued participation of the remaining feed in the reaction, the feed cut-off ratio is calculated to be 100% or 80%; the safety intervention time window represents the remaining time from the current judgment moment to before the loss of control and irreversibility, which can still be effectively intervened. For example, if it is estimated to be 12 seconds, then the subsequent cut-off and feedforward actions must be completed around this window. If the dynamic heat difference increases but then decreases in a certain period, the system will re-accumulate the number of divergences and will not use the previous count; if the temperature exceeds the threshold but the temperature rise rate is insufficient, the system will be judged to be in a high-level stable operating condition and will only enter the warning level and not the interlock level. If the material library lacks the activation energy or specific heat capacity of this batch, the controller will use a conservative upper limit value from similar historical materials. If the deviation between the input parameters is greater than the preset conflict threshold, for example, the temperature rise rate is greater than or equal to the preset extreme value but the cooling load estimate is lower than the preset lower limit, the status judgment module will mark the model result as unreliable and submit the anomaly to the maintenance diagnosis link, while maintaining a high level of manual prompts. In an exemplary chemical exothermic reaction application scenario, a batch of feed exhibits enhanced exothermic reaction in the later stages due to a high feed concentration. The edge computing layer extracts a temperature rise rate that increases from 0.11℃ / s to 0.22℃ / s, with the temperature reaching 96.2℃. The controller estimates a dynamic heat difference of 30 for three consecutive cycles. 55 and 85 Based on the activation energy parameters of the material, it was determined that further heating would significantly accelerate the reaction. Therefore, the target cooling water flow rate was calculated to be increased from 20... Increased to 34 Meanwhile, the feed cutoff ratio is set to 100%, and the remaining intervention time is estimated to be only 12 seconds. The aforementioned information is then sent to the subsequent cutoff and feedforward links. Through the above processing, the judgment of whether danger exists has been upgraded from a static over-limit judgment to a dynamic judgment of the degree of heat balance disruption, thereby enabling earlier identification of the true imbalance state and output of actionable intervention targets. The risk cutoff module is specifically used to: determine the candidate process nodes distributed on the fluid transmission path generated by the topology matrix of the process pipeline network, compare the fluid transmission delay time with the preset action completion time of the candidate process nodes, and select the candidate process node whose action completion time is less than the fluid transmission delay time and contains the fewest downstream process nodes as the cutoff node.

[0018] This embodiment provides a risk cut-off mechanism based on process topology and transmission delay. Specifically, although the aforementioned state judgment can identify that the reactor has entered an imbalance, if the cut-off action relies only on a pre-set single valve, it is easy to cause technical defects such as action lag or the cut-off range exceeding the preset isolation range. Therefore, this embodiment dynamically selects the most suitable cut-off node through the process topology information in the preset data acquisition and monitoring control system. In terms of specific control logic implementation, the system reads the topology matrix of the pre-built and stored pipeline and instrumentation flow diagram; assuming the anomaly source node is... Downstream, it connects sequentially to , , and These represent the outlet valve, buffer tank inlet valve, circulating pump, and bypass valve, respectively; in the topology matrix, → , → , → The connectivity position is set to 1; the system then reads the length and diameter of each pipe segment, and calculates the time it takes for the abnormal medium to reach each node based on the current flow velocity; for example... to The pipe is 2m long and the flow velocity is 1. The arrival time is approximately 2 seconds. to The main pipe is 6m long and takes approximately 6 seconds to arrive. At the same time, the system reads the completion time of each field actuator's action, such as It takes 0.8 seconds to close. The inlet valve requires 2.5 seconds to close. It takes 4 seconds for the flow rate to completely decrease after an emergency stop; therefore, it can be concluded that... and The preset action completion time is less than the fluid transfer delay time, thus enabling the action to be completed before abnormal materials arrive. Unable to achieve effective blocking; Furthermore, if multiple nodes can operate within the time limit, this embodiment introduces a selection strategy with the lowest process isolation level, that is, prioritizing the valve with the fewest process nodes based on the number of downstream process nodes contained in the fluid transmission path; the influence range of the process node is calibrated and stored in the distributed control system based on historical process flow data and preset shutdown level. For example, close Isolating only the current abnormal batch and closing the upstream main shut-off valve would shut down the entire unit; therefore, this is the preferred option. ;like If the system is malfunctioning or under maintenance, then select... If all local nodes fail to meet the arrival time limit, the higher-level main shut-off valve and material transfer pump will be triggered directly through the safety instrumented system to stop the emergency shutdown, and a conservative approach will be adopted to prevent the risk from spreading. If multiple paths run in parallel within the topology matrix, such as the main line and bypass simultaneously connecting to the downstream buffer tank, the system calculates the arrival time for each path and takes the earliest arrival time as the risk arrival time of that downstream node. If the execution status of a candidate valve is unknown, it is downgraded to a backup node. If the real-time flow velocity is unavailable, the maximum flow velocity estimate under the most recent stable operating condition is used to shorten the available action window. If the safety instrumented system reports that the preferred field actuator has failed to operate, the system automatically switches to the next node according to the priority sequence. In the aforementioned nitration production line, when the reactor is determined to be out of balance, the system retrieves the flowchart from the reactor to the downstream buffer tank and detects that there is a main discharge line and a low-flow return bypass for the abnormal medium. The calculation results show that the emergency shut-off valve near the reactor outlet of the main discharge line can be closed within 1 second, and the abnormal material will take about 3 seconds to reach the downstream side of the valve. Therefore, this valve is selected as the first shut-off node. Although the bypass valve can also be activated, its closure will affect the subsequent cleaning process and is not the shortest diffusion path, so it is used as an alternative. If the first shut-off node reports a fault during self-testing, the system immediately issues an emergency stop command for the delivery pump and a closing command for the downstream inlet valve to prevent the high-temperature material from diffusing into the buffer tank. Through the above processing, the cut-off action is transformed from a fixed interlock to a dynamic selection associated with real-time operating conditions and process topology, thereby achieving faster risk isolation and a smaller scope of downtime impact. The feedforward control module is specifically used to: compare the expected parameter increment with a preset parameter increment threshold; if it exceeds the increment threshold, then modify the basic control command to a maximum output amplitude control command for the field actuator.

[0019] This embodiment provides a feedforward compensation and overshoot prevention control mechanism that takes into account execution lag and residual inertia. Specifically, although the basic control and risk cut-off mentioned above can respond to imbalances, in exothermic reaction scenarios, there is a dual mechanical and fluid lag between the issuance of control commands and their actual application to the reactor temperature and pressure. Furthermore, after applying the maximum cooling load, the system will continue to decline due to thermal inertia, easily falling below the safety lower limit. Therefore, this embodiment sets up feedforward control and overshoot prevention compensation control respectively. Specifically, the feedforward control module continuously calculates the first and second derivatives of temperature or pressure. The first derivative reflects the current rate of change, while the second derivative reflects whether the rate of change is still accelerating. If the temperature is 94.0℃, 95.0℃, and 96.5℃ for three consecutive cycles, with a sampling interval of 5 seconds, the first derivatives for the first two intervals are approximately 0.20℃ / s and 0.30℃ / s, respectively, indicating that the temperature rise is accelerating, and the corresponding second derivatives are greater than zero. The module then reads the actuator lag time. For example, if the dead zone of the cooling valve's mechanical action is 1 second, and it takes 4 seconds for the cooling medium to pass through the pipeline and enter the jacket before affecting the reactor temperature, the total lag time is 5 seconds. The module discretizes the second derivative of temperature or pressure, that is, it extracts the difference between the first derivatives of adjacent segments and converts it into an equivalent additional rate of change through the process gain coefficient. For example, the above... and The difference divided by the sampling interval Converted to second derivative parameters ; The module combines the current first derivative. Second derivative With lag time Substituting into the feedforward formula, the expected parameter increment is calculated as follows: If, based on the current second-order trend and the product calculation, the temperature will continue to increase by 1.75°C beyond the expected range within these 5 seconds, then the system will immediately add feedforward compensation to the basic control command. When the combination of the second derivative and the lag time exceeds the preset threshold, it indicates that the trend of acceleration combined with the execution lag has reached a dangerous level. At this time, the maximum output amplitude control command for the field actuator is directly issued. The preset incremental threshold is the maximum transient parameter deviation value that the equipment can withstand during the control lag period, which is calibrated based on the historical safe operation data of the reactor. For example, the target opening of the cooling valve is directly adjusted to 100%; in terms of communication, this control task is placed with higher priority to avoid ordinary monitoring data occupying the bus transmission time slot; if trend recording frames and interlock control frames exist simultaneously within the industrial bus cycle, the system increases the communication priority of the interlock control frames in the industrial bus to shorten the actual issuance delay of control commands. After executing the maximum cooling load, the prediction and compensation module predicts the lowest future value based on the parameter fallback curve and estimates the compensation enthalpy value required to prevent overshoot. During the prediction and compensation phase, the system prioritizes calculating the compensation enthalpy based on the lower limit of temperature safety. When the predicted pressure is synchronously lower than the lower limit of pressure safety, the system calls a preset temperature-pressure coupling coefficient to perform weighted correction on the calculated compensation enthalpy value. The specific calculation of the compensation enthalpy value uses the following formula: In the formula, The required compensated enthalpy value, in units of ; The equivalent comprehensive heat capacity of the reactor system and the current batch of materials is determined by looking up and weighting the data from a pre-stored table of basic thermodynamic properties of the materials. The unit is [unit missing]. ; The lower limit of the system temperature setting, in units of ; This represents the lowest predicted future temperature based on the current rate of decline, in units of... .

[0020] When predicted At that time, the system generates a value including the compensated enthalpy. The system receives the reverse heat compensation command and establishes a direct control function mapping based on the principle of energy conservation: the target opening degree of the proportional valve of the heat tracing system. The formula is: In the formula, This refers to the rated mass flow rate of steam at full scale of the proportional valve. Latent heat of vapor, The preset compensation time; or synchronously combining the characteristics of the actuator to generate inverter control commands, thereby adjusting the operating frequency of the variable frequency cooling water pump. The following formula will be used to adjust downwards: in, This is the current steady-state operating frequency. The pre-calibrated frequency derating factor, in units of This establishes an absolute mapping from enthalpy to mechanical action.

[0021] Heat compensation can be achieved in two ways: one is to output a small opening command to the proportional valve of the heat tracing system to compensate for heat for a short time; the other is to reduce the speed of the variable frequency cooling water pump to reduce the heat dissipation rate. The two can be used independently or in combination in stages. For example, the speed of the variable frequency cooling water pump can be reduced first and then the proportional valve of the heat tracing system can be opened at a low opening. If the first derivative is positive but the second derivative is close to zero, it means that the parameters have not been further accelerated. At this time, the feedforward module only performs small compensation and does not output the maximum amplitude control command. If the actuator lag time data is missing, the maximum nominal lag value in the equipment file is used for calculation; if the bus priority adjustment fails, the system sends the most critical control frame through the security link; for the prediction compensation module, if the predicted drop value has not yet fallen below the safety lower limit, slow cooling is used instead of heat tracing; if the heat tracing system is unavailable, compensation is only performed by reducing the speed of the variable frequency cooling water pump; if the temperature has returned to the stable zone and the parameter drop rate is close to zero, compensation is automatically terminated. In the aforementioned production batch, the temperature curve continued to rise after reaching 95°C, with the heating rate increasing from 1°C every 5 seconds to 1.5°C every 5 seconds. The jacketed cooling valve required approximately 5 seconds from activation to actual effectiveness. The system calculated that the acceleration trend combined with the lag exceeded the threshold, so it directly adjusted the cooling valve to its maximum output amplitude and elevated the control task to high priority on the bus. After 20 seconds, the system predicted that the temperature would further decrease from the safe lower limit of 90°C to 88.5°C, so it reduced the variable frequency cooling water pump from 100% speed to 72% in advance and briefly opened the proportional valve of the heat tracing system to 10%, ultimately stabilizing the temperature above 90°C. Through the above processing, the response delay caused by the execution lag is offset in advance, and the overshoot caused by the fallback is suppressed after strong intervention, thereby achieving a fast and stable safe adjustment effect.

[0022] Please see Figure 2 A method for operating a chemical safety data acquisition and control system includes the following steps: S1. Obtain the original operating condition data of temperature and pressure of the reactor in the chemical system, the lower limit of temperature safety, the lower limit of pressure safety, and the lag time of each field actuator; extract the operating condition characteristic parameters based on the original operating condition data; S2. Calculate the dynamic heat difference of the reactor based on the operating condition characteristic parameters, determine the imbalance state of the chemical system based on the dynamic heat difference, and generate basic control commands. S3. Determine the risk source based on the imbalance state, obtain the pipeline parameters and current material flow rate of the process pipeline network in the chemical system; retrieve the topology matrix of the process pipeline network and determine the fluid transmission delay time by combining the pipeline parameters and the current material flow rate, and generate a cut-off command. S4. Read the lag time of the corresponding field actuator, determine the expected parameter increment based on the working condition characteristic parameters and the original working condition data combined with the lag time, and correct the basic control command according to the expected parameter increment; S5. After executing the basic control command or the cut-off command, predict the lowest predicted value after the temperature or pressure inside the reactor drops, generate a reverse heat compensation command based on the lowest predicted value, the lower limit of temperature safety, and the lower limit of pressure safety, and send the basic control command, the cut-off command, and the reverse heat compensation command to the corresponding field actuator.

[0023] This embodiment provides an operational step chain for a chemical safety data acquisition and control system; specifically, in the aforementioned nitration reaction production line, the entire method revolves around the entire process of a single batch from stable operation, initial appearance of abnormal conditions, risk diffusion to recovery to stability, forming a continuous execution process; Specifically, S1 is used to ensure the reliability of data input; after the system collects raw data such as temperature and pressure, it first cleans up transient high-frequency interference and then performs physical consistency verification to obtain characteristic parameters that can be analyzed; S2 is used to realize imbalance identification and basic intervention calculation, that is, to determine whether the heat release of the reaction exceeds the heat dissipation capacity based on the heat difference, and output the basic actions of cooling and cutting off the feed after confirming that it has entered the danger zone, while giving the remaining intervention time. S3 is used to block the spread of risk, that is, to select the most suitable shut-off valve or pump based on the abnormal source and pipeline parameters; S4 is used to perform feedforward compensation before the control takes effect, to estimate the future parameter overshoot by combining the second-order change trend of the parameter and the execution lag, and to convert the parameter overshoot into a higher intensity control compensation; S5 is used to prevent the system from continuing to decline and falling below the safety lower limit after strong intervention, to generate a reverse heat compensation command by predicting the temperature or pressure drop curve. For example, if the characteristic parameters output by S1 are respectively the temperature rise rate as ℃ / s, pressure fluctuation variance is And the liquid level fluctuation index is S2 will increase the temperature rise rate. Pressure fluctuation variance and liquid level fluctuation index The dynamic heat difference is obtained after inputting the state model. Furthermore, the temperature increased for three consecutive cycles, thus indicating an imbalance and generating basic control commands, which included a cooling flow target of 34m³ / h. 3 / h and feed cut-off ratio 100%; after S3 reads the topology matrix of the process pipeline network, it determines that the preferred cut-off node is the outlet valve and generates a cut-off command; S4 further reads the second derivative of the temperature and the lag time of the actuator. For example, when the second derivative of the temperature reflects that the heating rate is still increasing and the lag time is 5 seconds, the product result corresponds to an expected temperature increment of 1.5℃. The system converts this increment into an additional compensation amount and adds it to the basic control command to improve the control output corresponding to the cooling command. S5 continuously predicts the temperature or pressure value after the action is executed. If it is expected that the temperature will be lower than the safety limit within a preset time, a reverse heat compensation command is generated, such as reducing the speed of the variable frequency cooling water pump or opening the proportional valve of the heat tracing system. The cut-off command and compensation command are sent to the corresponding actuators to complete one closed-loop process. If the reliable characteristic parameters are not obtained in the S1 stage, the S2 stage will not enter the normal model judgment, but will directly adopt the fail-safe mode. If the S2 judgment result does not meet the imbalance condition, the subsequent steps will not initiate emergency cut-off, but will only maintain normal adjustment. If no local node capable of timely action is matched in S3, the upstream total cutoff strategy is switched; if the expected parameter increment calculated in S4 is negative, it indicates that the trend is weakening, so no large-scale feedforward compensation is superimposed; if the predicted value in S5 is always higher than the safety lower limit, no reverse compensation is generated, and only the cooling gradual recovery logic is retained. During the production batch operation, S1 corrects the collected transient abnormal data into a reasonable sequence, S2 confirms that the heat difference is continuously diverging and gives the basic action of increasing the cooling water flow and completely cutting off the feed, S3 selects the emergency shut-off valve near the reactor for isolation, S4 directly adjusts the opening of the cooling valve to the full range because the temperature rise is detected to be accelerated and the execution lag is long, S5 after the danger is cleared, the cooling intensity is moderately recovered and a small amount of heat is applied, so that the batch can safely return to the control range without interrupting the utilities of the entire unit. S1 specifically includes: performing cross-validation on the temperature sequence and pressure sequence in the original operating condition data.

[0024] Step S1 specifically includes: In response to the differences in execution time granularity caused by different sensor hardware and calculation windows, the system introduces a timing synchronization and resampling coordination mechanism between the data acquisition and control links: generating an original sequence with a unified timestamp using a 1-second reference clock beat; for feature data extraction and calculation for 10-second and 5-second calculation windows, the module uniformly adopts a sliding window algorithm to extract timestamp-aligned data columns, and uses a zero-order hold to map back to a 1-second reference control cycle for collaborative scheduling, ensuring that multi-source signals have a strict consistent time reference when superimposing basic instructions, feedforward instructions and compensation instructions; Obtain the original sequence of temperature and pressure in the reactor; assume that the four consecutive sampled values ​​are temperature [91.9, 92.2, 116.8, 92.4] and pressure [0.18, 0.18, 0.19, 0.19]; among them, 116.8℃ obviously does not conform to thermal inertia, and the extended Kalman filter will correct this point according to the previous state; The system utilizes the thermodynamic equation of state and phase equilibrium curves for cross-validation; under specific gas phase space and material composition conditions, the theoretical pressure should be between 0.17 and 0.20 at a temperature of 92℃. The actual measurement showed a continuous deviation of up to 0.24. If the pressure measurement point is found to be drifting, the system extracts feature parameters after obtaining a reliable sequence; the temperature rise rate reflects a stable heating trend. Pressure fluctuation variance reflects whether there are violent pressure oscillations; liquid level fluctuation index reflects whether the liquid phase is unstable due to bubbling, feed pulses, or abnormal stirring; if the temperature has smoothly risen from 92.0℃ to 93.5℃ in the last 10 seconds, the temperature rise rate can be taken as 0.15℃ / s; if the pressure value is around 0.18... If the fluctuations are small, the pressure fluctuation variance will be within the preset fluctuation threshold; if the liquid level fluctuates up and down within a short time window, accounting for more than 5% of the average liquid level, the liquid level fluctuation index will increase. If both temperature and pressure exhibit unreliable fluctuations within the same time period, the system prioritizes retaining the original time sequence and marks it as low confidence. If the level sensor is not installed, the level fluctuation index is empty, and the subsequent model operates solely based on temperature and pressure characteristics. If cross-validation detects a slight deviation in the data but does not meet the duration threshold, the drift characteristic is recorded without triggering a fault determination. In another batch on the same production line, the temperature probe experienced a transient abnormal change, which was corrected after filtering. Subsequently, the pressure signal deviated from the equilibrium curve for a long time and was determined to be drift caused by liquid accumulation in the pressure guide tube. The system finally extracted a high temperature rise rate and a small pressure variance, indicating that attention should be paid to the enhanced exothermic reaction at this time. Through the above processing, the original input is transformed into a stable, reliable and easy-to-model feature set, thereby enabling accurate initiation of subsequent judgments; S2 specifically includes: inputting the operating condition characteristic parameters into a pre-constructed state-space model, and determining the dynamic heat difference of the reactor based on the state-space model.

[0025] Step S2 specifically includes: the system inputs the operating condition characteristic parameters into the state space model to obtain the dynamic heat difference; for example, at a certain moment, the model estimates the net heat release power to be 390. The current cooling system has a maximum heat dissipation capacity of 370. The difference is 20. The next cycle will be 50. The next cycle will be 90. Since the difference increases for three consecutive cycles, it is determined to be divergent; combined with the preset temperature and temperature rise rate thresholds, for example, if the current temperature is 96℃ and the temperature rise rate is 0.22℃ / s, exceeding the set 95℃ and 0.15℃ / s, the system confirms that it has entered the temperature runaway critical zone. After confirmation, the system calculates the intervention amount based on the current material parameters. If the activation energy of the material reaction is lower than the preset activation energy threshold, further heating will exacerbate heat release more quickly, thus the feed cut-off ratio tends to be higher. Specific heat capacity affects the amount of heat corresponding to a unit temperature rise, thereby affecting the calculation of cooling flow rate. If an additional 120 is required, it is estimated that... Based on the heat output and the cooling water temperature difference and heat exchange efficiency, the target cooling water flow rate needs to be increased to 34 m³ / h. 3 / h; if the remaining feed continues to enter, approximately 60 will be introduced again. If there is potential for heat release, the cutoff ratio should be at least 80%, and a control command for complete cutoff should be generated if necessary. If the dynamic heat difference is divergent but the temperature has not exceeded the safety threshold, the system enters the preparatory control state, first increasing cooling but not cutting off the feed; if the temperature exceeds the threshold but the temperature rise rate is very low, it is determined that the operating condition is in a high-level stable state, and the runaway interlock is not directly triggered; if the parameters of a certain batch of materials are missing, the conservative upper limit value of similar materials is used for estimation; if the calculated cooling flow exceeds the equipment capacity, the system directly outputs a preset safety control command that includes full-scale cooling and maximum proportion feed cut-off. In this nitration batch, the model obtained an increasing dynamic heat difference value three times in a row, indicating that the net heat release power of the system was continuously greater than the maximum heat dissipation capacity; and since the material is sensitive to activation energy in the current temperature range, the system calculated the control amount to increase the cooling water flow rate target to near the upper limit and cut off the feed. Through the above processing, the risk assessment is combined with the thermal properties of materials and the heat dissipation capacity of equipment, thereby enabling more targeted calculation of basic intervention quantities; S3 specifically includes: retrieving the topology matrix of the process pipeline network within the chemical system; taking the risk source as the starting point; determining the fluid transmission delay time based on the pipeline parameters and the current material flow rate; generating a fluid transmission path; and determining a cut-off node from the candidate process nodes distributed along the fluid transmission path based on the fluid transmission delay time and the preset action completion time of the candidate process nodes, and generating a cut-off command.

[0026] Step S3 specifically includes: the system first generates a topology matrix from the flowchart; assuming the risk source is the reactor node. There are three downstream connections: one leading to the outlet valve. Then connect to the buffer tank inlet valve Right now Connected to the bypass valve Right now Connected to the circulating pump Then connect to the external cooler inlet valve Right now ; The corresponding connected terms in the matrix are assigned a value of 1; the parameters of each pipe segment are read, for example... arrive 2m in length arrive 5m in length through arrive The total length is 8m; the current material flow rates are 1m / s, 0.8m / s, and 1.2m / s, respectively. Therefore, the calculated arrival times for each path to the first node are 2 seconds, 6.25 seconds, and 6.67 seconds, respectively. Combined with the completion times of candidate process nodes, such as... It takes 0.8 seconds. It takes 2 seconds. An emergency stop is effective within 4 seconds. Therefore, The action completion time is less than the fluid transport delay time; the system compares the downtime impact levels; if shut down... This only affects the current batch; close. If shutting down the upstream material conveying pump will affect the receiving of materials in the downstream storage tanks, and shutting down the upstream material conveying pump will affect the entire unit, then the principle of minimizing the scope of downtime while meeting the time limit should be prioritized. And generate cut-off commands, such as first priority. Second priority The third priority material conveying pump was shut down immediately. If a path has parallel branches, the system calculates the arrival time of each parallel branch and uses the shortest arrival time as the condition for constraining the cutting sequence. If multiple nodes have the same time limit and shutdown level, the component with a higher success rate and better running status shall be selected first. If the current flow rate is missing, the nearest maximum flow rate is used as a substitute; if none of the local nodes meet the pre-arrival action, the priority sequence starts directly from the total cut-off. In this nitration system, the reactor outlet valve The component closest to the risk source and whose closure only affects the current production batch is therefore prioritized first; while the downstream buffer tank inlet valve, although it can also be closed promptly, will affect subsequent material receiving and switching, and is therefore ranked second; if If the status feedback is abnormal, the system will immediately execute the second priority option in sequence; Through the above processing, a dynamic cutoff sequence is generated based on both propagation delay and the impact of work stoppage, thereby achieving more precise risk isolation; S4 specifically includes: comparing the expected parameter increment with a preset parameter increment threshold; if it exceeds the increment threshold, then modifying the basic control command to a maximum output amplitude control command for the field actuator.

[0027] Step S4 specifically includes: In S4, the system calculates the first and second derivatives of temperature or pressure in real time; if the temperature sequence is 95.0℃, 96.0℃, and 97.8℃, and the sampling interval is 5 seconds, then the first derivative is first 0.20℃ / s, and then 0.36℃ / s, indicating that the temperature rise is accelerating; at the same time, the system reads the dead time of the cooling valve's mechanical action of 1 second and the pure lag of the cooling medium delivery of 4 seconds, for a total lag of 5 seconds; When performing incremental estimation, the system strictly distinguishes between the first and second rates of change of temperature. Taking a sampling interval of 5 seconds as an example, if the first derivative of the previous window is 0.20℃ / s and the current first derivative is 0.36℃ / s, then the system calculates the true second derivative as (0.36-0.20) / 5 = 0.032℃ / s. 2 The system uses the current first derivative of 0.36℃ / s and the second derivative of 0.032℃ / s. 2 Substituting the gain coefficient K and the lag time of 5 seconds into the feedforward formula that includes Taylor expansion, the expected parameter increment is calculated. ℃; If the product estimated based on the current second-order trend is 2.2℃, the temperature will still have a margin of 2.2℃ within these 5 seconds. When the increment of the expected parameter is greater than or equal to the preset increment threshold, the system issues a maximum output amplitude control command for the field actuator and simultaneously modifies the priority label of the preset bus communication data frame, so that the control frame is sent with priority over the trend recording frame and the ordinary monitoring frame. In S5, after the temperature begins to drop due to the maximum cooling load and feed cut-off, the system estimates the compensation enthalpy based on the current rate of temperature drop, equipment heat capacity, and remaining heat of reaction. If the predicted temperature will drop below the safety lower limit by 2°C, and the equivalent comprehensive heat capacity of the reactor is 50 kW per 1°C, the system will compensate for the temperature drop. The required enthalpy compensation is approximately 100. The system can output a small opening command to the heat tracing proportional valve, such as 10% to 15%, or reduce the cooling water pump speed from 100% to a preset safe speed to reduce the intensity of subsequent heat dissipation. Either action can be selected, or the cooling water pump speed can be reduced first and then the heat tracing can be slightly reduced. If the second derivative is close to zero or negative, it indicates that the trend of change is below the preset acceleration threshold, and basic control is maintained. If the bus communication priority label modification fails, the system will still send the maximum output amplitude control command for the field actuator through the existing safety control channel. If the estimated value of the compensated enthalpy is lower than the preset minimum action threshold, the heat tracing will not be activated, and the pump speed will be slowly restored. If the heat tracing system itself is in a fault state, only the cooling water pump speed will be reduced. If the predicted drop amplitude exceeds the preset oscillation judgment threshold, multiple small-amplitude heat compensations will be used. During the dangerous temperature rise phase of this nitration batch, the system detected that the rate of temperature increase continued to accelerate, while the cooling circuit had a total lag of 5 seconds. Therefore, the cooling valve was directly adjusted to the maximum output amplitude, and the frame was sent in the high priority channel of the industrial bus. After the reaction was suppressed, the system predicted that the temperature might drop to 88.5℃, which is lower than the 90℃ safety limit. Therefore, the variable frequency cooling water pump was first slowed down, and then the heating proportional valve was briefly opened to make the reactor temperature stabilize in the range of 90℃ to 91℃. Through the above processing, two measures, namely communication priority enhancement and reverse heat compensation, are implemented to simultaneously solve the problems of control response delay and regulation overshoot, thereby achieving rapid and stable convergence after hazard disposal.

[0028] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A chemical safety data acquisition and control system, characterized in that, include: The data acquisition and processing module is used to acquire raw operating condition data of temperature and pressure of the reactor in the chemical system, the lower limit of temperature safety, the lower limit of pressure safety, and the lag time of each field actuator; and to extract operating condition characteristic parameters based on the raw operating condition data. The state judgment module inputs the operating condition characteristic parameters into a pre-constructed state space model, determines the dynamic heat difference of the reactor based on the state space model, and determines the imbalance state of the chemical system based on the dynamic heat difference, and generates basic control commands. The risk cut-off module is used to determine the risk source based on the imbalance state, and based on the risk source, obtain the pipeline parameters and current material flow rate of the process pipeline network in the chemical system; The topology matrix of the process pipeline network is retrieved and the fluid transmission delay time is determined by combining the pipeline parameters and the current material flow rate, and a cut-off command is generated. The feedforward control module is used to combine the lag time of each field actuator, the operating condition characteristic parameters and the original operating condition data to determine the expected parameter increment, and to correct the basic control command according to the expected parameter increment. The prediction and compensation module is used to continuously track and predict the lowest predicted value after the temperature or pressure inside the reactor drops after the basic control command or the cut-off command is executed, and to generate a reverse heat compensation command based on the lowest predicted value, the temperature safety lower limit and the pressure safety lower limit. The control output module is used to send the basic control command, the cut-off command, and the reverse heat compensation command to the corresponding field actuators.

2. The chemical safety data acquisition and control system according to claim 1, characterized in that, The data acquisition and processing module is used to perform cross-validation on the temperature and pressure sequences in the original operating condition data.

3. The chemical safety data acquisition and control system according to claim 1, characterized in that, The risk cutoff module is specifically used to: generate a fluid transmission path based on the topology matrix of the process pipeline network, determine the candidate process nodes distributed on the fluid transmission path, compare the fluid transmission delay time with the preset action completion time of the candidate process node, and select the candidate process node whose action completion time is less than the fluid transmission delay time and contains the fewest downstream process nodes as the cutoff node.

4. The chemical safety data acquisition and control system according to claim 1, characterized in that, The feedforward control module is specifically used to: compare the expected parameter increment with a preset parameter increment threshold; if it exceeds the increment threshold, then modify the basic control command to a maximum output amplitude control command for the field actuator.

5. A method for operating a chemical safety data acquisition and control system, characterized in that, Includes the following steps: S1. Obtain the original operating condition data of temperature and pressure of the reactor in the chemical system, the lower limit of temperature safety, the lower limit of pressure safety, and the lag time of each field actuator; extract the operating condition characteristic parameters based on the original operating condition data; S2. Calculate the dynamic heat difference of the reactor based on the operating condition characteristic parameters, determine the imbalance state of the chemical system based on the dynamic heat difference, and generate basic control commands. S3. Determine the risk source based on the imbalance state, and obtain the pipeline parameters and current material flow rate of the process pipeline network in the chemical system; The topology matrix of the process pipeline network is retrieved and the fluid transmission delay time is determined by combining the pipeline parameters and the current material flow rate, and a cut-off command is generated. S4. Read the lag time of the corresponding field actuator, determine the expected parameter increment based on the working condition characteristic parameters and the original working condition data combined with the lag time, and correct the basic control command according to the expected parameter increment; S5. After executing the basic control command or the cut-off command, predict the lowest predicted value after the temperature or pressure inside the reactor drops, generate a reverse heat compensation command based on the lowest predicted value, the lower limit of temperature safety, and the lower limit of pressure safety, and send the basic control command, the cut-off command, and the reverse heat compensation command to the corresponding field actuator.

6. The operation method of a chemical safety data acquisition and control system according to claim 5, characterized in that, S1 specifically includes: performing cross-validation on the temperature sequence and pressure sequence in the original operating condition data.

7. The operation method of a chemical safety data acquisition and control system according to claim 5, characterized in that, The step S2, which calculates the dynamic heat difference of the reactor based on the operating condition characteristic parameters, includes: inputting the operating condition characteristic parameters into a pre-constructed state-space model, and determining the dynamic heat difference of the reactor based on the state-space model.

8. The operation method of a chemical safety data acquisition and control system according to claim 5, characterized in that, S3 specifically includes: retrieving the topology matrix of the process pipeline network within the chemical system; taking the risk source as the starting point; determining the fluid transmission delay time based on the pipeline parameters and the current material flow rate; generating a fluid transmission path; and determining a cut-off node from the candidate process nodes distributed along the fluid transmission path based on the fluid transmission delay time and the preset action completion time of the candidate process nodes, and generating a cut-off command.

9. The operation method of a chemical safety data acquisition and control system according to claim 5, characterized in that, S4 includes: comparing the expected parameter increment with a preset parameter increment threshold; if the increment exceeds the threshold, then modifying the basic control command to a maximum output amplitude control command for the field actuator.

Images