Multi-stage pressure optimization control method and system for high-pressure hydrogenation reactor

Abstract

The invention relates to the technical field of high-pressure hydrogenation reactor control, and discloses a multi-stage pressure optimization control method and system for a high-pressure hydrogenation reactor, and the method comprises the steps: obtaining temperature, pressure and flow monitoring data, generating original parameter data, and deploying a distributed control network; collecting and analyzing multi-stage pressure sensor data, and generating a pressure change influence factor; sensing a valve state, and generating a valve change influence factor; collecting data of a flow sensor, and dividing a control area after calibration; and calculating a pressure optimization index based on the control area data, integrating multiple factors to generate comprehensive control data, and further generating a pressure control signal and outputting an instruction. The system comprises a control network deployment module, a pressure analysis module, a valve analysis module, a control area analysis module and a pressure control signal generation module. Through multi-dimensional data fusion, a distributed control network and dynamic optimization calculation, the accuracy and stability of pressure control of the high-pressure hydrogenation reactor are improved, and the method is suitable for efficient and safe production in the field of petrochemical engineering.

Description

Technical Field

[0001] The present invention relates to the technical field of high-pressure hydrogenation reactor control, and in particular to a multi-stage pressure optimization control method and system for a high-pressure hydrogenation reactor. Background Art

[0002] In the petrochemical industry, high-pressure hydrogenation reactors are core equipment, and the accuracy of their pressure control is directly related to reaction efficiency, product quality, and production safety. Traditional pressure control methods for high-pressure hydrogenation reactors typically use a single closed-loop control strategy, performing simple proportional-integral-derivative (PID) adjustments based solely on feedback signals from pressure sensors. This control approach has significant limitations.

[0003] In terms of control network architecture, traditional centralized control networks suffer from high single-point failure risks and poor scalability. A failure in a control node can lead to loss of control over the entire system. Furthermore, centralized control networks struggle to adapt to the increasing size and complexity of high-pressure hydrogenation reactors, and are unable to achieve refined control over each reactor area. For example, in large-scale high-pressure hydrogenation reactors, reaction processes may differ across different regions, requiring different pressure control strategies. However, centralized control networks lack the flexibility to divide regions and deploy control nodes.

[0004] Traditional methods provide a crude sense and control of valve status. Control is based solely on valve opening feedback, failing to comprehensively monitor and analyze the valve's operating status. For example, valves may experience wear, sticking, and other faults during long-term operation. Traditional methods are unable to promptly detect these abnormal conditions, which can easily lead to valve control failure and, in turn, affect the reactor's pressure stability. Furthermore, traditional methods fail to consider the impact of valve type and regional distribution on pressure control during valve control, making it impossible to achieve differentiated control and optimized scheduling of valves. Regarding pressure optimization calculations, traditional methods use fixed control parameters and simple mathematical models, making it impossible to dynamically adjust based on the reactor's real-time operating status. For example, the reactor's pressure optimization target may vary during different reaction stages. Traditional methods are unable to update the pressure optimization index in real time, resulting in suboptimal control. Furthermore, traditional methods lack a model training and optimization process when generating pressure control signals, resulting in low accuracy and reliability of the control signals. Summary of the Invention

[0005] The object of the present invention is to provide a multi-stage pressure optimization control method and system for a high-pressure hydrogenation reactor to solve the problems raised in the above background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a multi-stage pressure optimization control method for a high-pressure hydrogenation reactor, the method comprising: Step S1: acquiring temperature monitoring data, pressure monitoring data, and flow monitoring data of a high-pressure hydrogenation reactor; performing reactor parameter analysis on the temperature monitoring data, pressure monitoring data, and flow monitoring data to generate original parameter data; and deploying a control network based on the original parameter data to obtain a distributed control network; Step S2: collecting and analyzing multi-level pressure sensor data on the distributed control network to generate pressure change influencing factors; Step S3: sensing the status of the reactor valve according to the distributed control network, obtaining valve status data, and generating a valve change influencing factor based on the valve status data; Step S4: collecting flow sensor data from the reactor according to the distributed control network to obtain flow sensor collected data; sampling and calibrating the flow sensor collected data to generate flow calibration data; dividing the reactor control area based on the flow calibration data to generate control area data; Step S5: Based on the control zone data, the flow calibration data and the valve status data are subjected to reactor pressure optimization calculation to generate a pressure optimization index; the pressure optimization index, the pressure change influencing factor and the valve change influencing factor are integrated to generate comprehensive control data; a pressure control signal is generated for the comprehensive control data to generate pressure control signal data; and an instruction is output for the pressure control signal data to execute a multi-stage pressure optimization control operation.

[0007] Preferably, step S1 includes the following steps: Step S11: obtaining a high-pressure hydrogenation reactor temperature sample, a high-pressure hydrogenation reactor pressure sample, and a high-pressure hydrogenation reactor flow sample; Step S12: performing parameter analysis on the high-pressure hydrogenation reactor temperature sample, the high-pressure hydrogenation reactor pressure sample, and the high-pressure hydrogenation reactor flow rate sample to obtain temperature monitoring data, pressure monitoring data, and flow rate monitoring data of the high-pressure hydrogenation reactor; Step S13: performing reactor parameter analysis on the temperature monitoring data, pressure monitoring data, and flow monitoring data to generate original parameter data; Step S14: Deploy the control network according to the original parameter data to obtain a distributed control network.

[0008] Preferably, step S14 includes the following steps: Step S141: performing reactor structure analysis on the original parameter data to obtain structure data; Step S142: Divide the reactor area according to the structural data to generate area boundary data; Step S143: arranging control node positions based on the area boundary data to generate control node position data; Step S144: deploying distributed control nodes using the multi-dimensional control node position data, and calibrating control parameters of the deployed distributed control nodes, thereby generating distributed calibrated control nodes; Step S145: performing communication channel association on the distributed calibration control nodes to generate a distributed control network.

[0009] Preferably, step S2 includes the following steps: Step S21: performing multi-level pressure sensor data acquisition on the distributed control network to obtain multi-level pressure monitoring data, wherein the multi-level pressure monitoring data includes primary pressure monitoring data, intermediate pressure monitoring data, and advanced pressure monitoring data; Step S22: performing multi-level pressure gradient calculation on the primary pressure monitoring data and the advanced pressure monitoring data to obtain a pressure gradient index; performing reactor pressure-density difference impact analysis on the pressure gradient index based on the intermediate pressure monitoring data to generate pressure-density difference data; Step S23: performing pressure fluctuation analysis on the pressure gradient index based on the pressure density difference data to generate pressure fluctuation data; performing pressure fluctuation trajectory analysis on the pressure fluctuation data to generate pressure fluctuation trajectory data; Step S24: analyzing the degree of reactor pressure variation on the pressure fluctuation data using the pressure fluctuation trajectory data to generate a pressure variation influencing factor.

[0010] Preferably, step S24 includes the following steps: Step S241: performing a reactor pressure fluctuation time series analysis on the high-pressure hydrogenation reactor pressure fluctuation data using the high-pressure hydrogenation reactor pressure fluctuation trajectory data to generate high-pressure hydrogenation reactor pressure fluctuation time series data; performing a dynamic graph conversion on the high-pressure hydrogenation reactor pressure fluctuation time series data to generate a high-pressure hydrogenation reactor pressure fluctuation dynamic graph; Step S242: performing reactor pressure abnormality event analysis on the high-pressure hydrogenation reactor pressure fluctuation dynamic diagram to generate high-pressure hydrogenation reactor abnormal pressure event data; Step S243: performing a normal pressure change trend analysis on the pressure fluctuation dynamic diagram of the high-pressure hydrogenation reactor based on the abnormal pressure event data of the high-pressure hydrogenation reactor to generate normal pressure trend change data of the high-pressure hydrogenation reactor; Step S244: Using the abnormal pressure event data of the high-pressure hydrogenation reactor, the impact degree of the abnormal event is detected on the normal pressure trend change data of the high-pressure hydrogenation reactor to generate abnormal pressure impact degree data; based on the abnormal pressure impact degree data, the reactor pressure change analysis is performed on the pressure fluctuation data of the high-pressure hydrogenation reactor to generate a high-pressure hydrogenation reactor pressure change impact factor.

[0011] Preferably, step S3 includes the following steps: Step S31: sensing the status of the reactor valve according to the distributed control network and generating valve sensing data; Step S32: collecting valve parameters from the valve sensing data, and performing valve status identification on the collected valve parameters to obtain valve status data; Step S33: matching the valve status data with the preset valve database. If the valve status data does not match the preset valve database, the corresponding valve status data is marked as abnormal valve status data. If the valve status data does match the preset valve database, the corresponding valve status data is marked as normal valve status data. Step S34: performing valve opening evaluation on the abnormal valve status data to generate abnormal valve opening evaluation data; performing valve abnormality impact analysis on the normal valve status data based on the abnormal valve opening evaluation data to generate a valve change impact factor.

[0012] Preferably, step S31 includes the following steps: Step S311: collecting reactor valve sensor data on the distributed control network to obtain valve sensor data; Step S312: performing data filtering on the valve sensor data to generate valve filtered data; performing signal enhancement on the valve filtered data to generate valve enhanced data; Step S313: performing signal feature analysis on the valve enhancement data to generate valve signal feature data; performing regional reactor valve marking based on the valve signal feature data to generate regional valve marking data; Step S314: performing reactor valve type identification on the regional valve tag data to generate valve type identification data; performing reactor valve behavior perception based on the valve type identification data to generate valve perception data.

[0013] Preferably, step S4 includes the following steps: Step S41: collecting flow sensor data from the reactor according to the distributed control network to obtain flow sensor collected data, wherein the flow sensor includes a flow meter sensor, a pressure difference sensor, and a flow velocity sensor; Step S42: performing flow temporal and spatial distribution characteristic analysis on the flow sensor data to generate flow temporal and spatial distribution characteristic data; performing flow data sampling and calibration on the flow temporal and spatial distribution characteristic data to generate flow calibration data; Step S43: performing a reactor flow state evaluation on the flow calibration data to generate flow state evaluation data; performing a distribution characteristic evaluation on the flow calibration data to generate flow distribution characteristic evaluation data; Step S44: dividing the reactor control area according to the flow state evaluation data and the flow distribution characteristic evaluation data to generate control area data.

[0014] Preferably, step S5 includes the following steps: Step S51: performing reactor control chain analysis on the flow calibration data and valve status data based on the control zone data to generate control chain data; Step S52: performing reactor pressure optimization calculation based on the control chain data to generate a pressure optimization index; integrating the pressure optimization index, the pressure change influencing factor, and the valve change influencing factor to generate comprehensive control data; Step S53: Divide the comprehensive control data into data sets to generate a model training set and a model test set; use an optimization algorithm to perform model training on the model training set to generate a pressure control signal training model; use the model test set to perform model optimization iteration on the pressure control signal training model to generate a pressure control signal model; Step S54: importing the integrated control data into the pressure control signal model to generate a pressure control signal and generate pressure control signal data; outputting instructions to the pressure control signal data to perform a multi-level pressure optimization control operation.

[0015] Preferably, the present invention further includes a high-pressure hydrogenation reactor multi-stage pressure optimization control system for executing the above-mentioned high-pressure hydrogenation reactor multi-stage pressure optimization control method, the system comprising: The control network deployment module is used to obtain the temperature monitoring data, pressure monitoring data, and flow monitoring data of the high-pressure hydrogenation reactor; perform reactor parameter analysis on the temperature monitoring data, pressure monitoring data, and flow monitoring data to generate original parameter data; and deploy the control network based on the original parameter data to obtain a distributed control network; The reactor pressure analysis module is used to collect multi-level pressure sensor data from the distributed control network to obtain pressure monitoring data; perform multi-level pressure gradient calculation on the pressure monitoring data to obtain a pressure gradient index; perform pressure fluctuation analysis on the pressure gradient index to generate pressure fluctuation data; perform pressure fluctuation trajectory analysis on the pressure fluctuation data to generate pressure fluctuation trajectory data; and perform reactor pressure change analysis on the pressure fluctuation data using the pressure fluctuation trajectory data to generate a pressure change influencing factor. The reactor valve analysis module is used to sense the reactor valve status based on the distributed control network and generate valve sensing data; collect valve parameters from the valve sensing data and perform valve status identification on the collected valve parameters to obtain valve status data; analyze the impact of valve anomalies on the valve status data and generate valve change impact factors; The reactor control zone analysis module is used to collect data from the reactor flow sensor according to the distributed control network to obtain flow sensor collected data; perform flow data sampling and calibration on the flow sensor collected data to generate flow calibration data; divide the reactor control zone based on the flow calibration data to generate control zone data; The pressure control signal generation module is used to perform reactor pressure optimization calculations on flow calibration data and valve status data based on control zone data to generate a pressure optimization index; integrate the pressure optimization index, pressure change influencing factor, and valve change influencing factor to generate comprehensive control data; generate a pressure control signal based on the comprehensive control data to generate pressure control signal data; and output instructions for the pressure control signal data to execute multi-stage pressure optimization control operations.

[0016] Compared with the prior art, the present invention has the following beneficial effects: In terms of control network architecture, a distributed control network was constructed. Through a series of steps such as structural analysis of raw parameter data, regional division, control node location layout, node deployment and calibration, and communication channel association, a reasonable layout of control nodes and efficient communication were achieved. Distributed control networks have the advantages of decentralization, high reliability, and strong scalability. They can effectively reduce the risk of single-point failures and adapt to the development needs of large-scale and complex high-pressure hydrogenation reactors. For example, by using regional boundary data to layout control node locations, corresponding control nodes can be deployed in different areas according to the structural characteristics of the reactor and the reaction process, achieving refined control of each area. At the same time, the distributed control network can flexibly expand nodes and calibrate parameters, improving the adaptability and maintainability of the system.

[0017] By filtering, enhancing, analyzing features, identifying types, and sensing behavior of valve sensor data, the system can accurately perceive the operating status of valves and generate accurate valve perception data. Valve parameters are collected and their status is identified, and information is matched against a pre-set valve database. This allows for the timely detection of abnormal valve conditions, such as wear and jamming. Abnormal valve status data is then evaluated for opening and analyzed for impact, generating valve change influencing factors. This enables the system to monitor valve operating conditions in real time, promptly detect and address valve failures, and avoid pressure fluctuations and production accidents caused by valve control failure. Furthermore, differentiated control based on valve type identification data fully considers the impact of the characteristics and regional distribution of different valve types on pressure control, achieving optimized valve scheduling and precise control.

[0018] By dividing the comprehensive control data into model training and test sets, and using optimization algorithms for model training and optimization iterations to generate a pressure control signal model, the accuracy and reliability of the control signal can be improved. By importing the comprehensive control data into the model to generate pressure control signal data and output instructions, multi-stage pressure optimization control is achieved. The pressure control strategy can be dynamically adjusted according to different reaction stages and regional requirements, improving reaction efficiency and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a working principle diagram of the multi-stage pressure optimization control method for a high-pressure hydrogenation reactor according to the present invention; Figure 2 Design drawings for deploying modules for the control network; Figure 3 Design diagram for multi-stage pressure sensor data analysis. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] See also Figure 1-Figure 3 The multi-stage pressure optimization control method of the high-pressure hydrogenation reactor involved in the present invention is specifically implemented as follows: Step S1: Acquire temperature monitoring data, pressure monitoring data, and flow monitoring data of a high-pressure hydrogenation reactor; perform reactor parameter analysis on the temperature monitoring data, pressure monitoring data, and flow monitoring data to generate original parameter data; and deploy a control network based on the original parameter data to obtain a distributed control network.

[0022] Step S2: Collect and analyze multi-level pressure sensor data on the distributed control network to generate pressure change influencing factors.

[0023] Step S3: sensing the status of the reactor valve according to the distributed control network, obtaining valve status data, and generating a valve change influencing factor based on the valve status data.

[0024] Step S4: collecting flow sensor data from the reactor according to the distributed control network to obtain flow sensor collected data; sampling and calibrating the flow sensor collected data to generate flow calibration data; dividing the reactor control area based on the flow calibration data to generate control area data.

[0025] Step S5: Based on the control zone data, the flow calibration data and the valve status data are subjected to reactor pressure optimization calculation to generate a pressure optimization index; the pressure optimization index, the pressure change influencing factor and the valve change influencing factor are integrated to generate comprehensive control data; a pressure control signal is generated for the comprehensive control data to generate pressure control signal data; and an instruction is output for the pressure control signal data to execute a multi-stage pressure optimization control operation.

[0026] Example 1: In step S1, it is specifically achieved in the following ways: obtaining a temperature sample of a high-pressure hydrogenation reactor, a pressure sample of a high-pressure hydrogenation reactor and a flow sample of a high-pressure hydrogenation reactor. The temperature samples here cover the temperature values ​​of different parts of the reactor in different time periods, such as the reactor inlet, the middle of the catalyst bed, the outlet and other positions, and the temperature data collected at certain time intervals. The time interval can be set according to actual production needs, such as every minute, every five minutes, etc. The pressure sample includes the overall pressure inside the reactor and the local pressure data of each key area. The collection of these pressure data also follows certain time laws and spatial distribution principles to ensure the comprehensiveness and representativeness of the sample. The flow sample mainly involves the flow information of reaction raw gas, hydrogen and products, etc. Factors such as the size of the flow and the change trend need to be considered during collection.

[0027] Parameter analysis is performed on these acquired temperature, pressure, and flow samples. During parameter analysis, specific data processing algorithms and models are used to screen, organize, and calculate the various parameters in the samples. For example, for temperature samples, the average, maximum, and minimum values, as well as the rate and trend of temperature change, are analyzed; for pressure samples, the pressure fluctuation range and duration of stable states are studied; and for flow samples, the total flow rate, instantaneous flow rate, and flow rate change patterns are calculated, thereby obtaining temperature, pressure, and flow monitoring data for the high-pressure hydrogenation reactor. This monitoring data accurately reflects the reactor's current operating status and the specific conditions of each parameter.

[0028] The temperature, pressure, and flow monitoring data are analyzed for their respective reactor parameters to generate raw parameter data. This step requires combining the various monitoring data with the reactor's structural and process parameters for comprehensive analysis. For example, structural parameters such as the reactor's volume, catalyst loading, and properties, as well as process parameters such as the setpoints and allowable ranges for reaction temperature, pressure, and flow, are considered. Through in-depth analysis of this data, the correlation between the various monitoring data and the overall operating status of the reactor is determined, generating raw parameter data that comprehensively describes the reactor's operating status.

[0029] Based on the original parameter data, a control network is deployed to create a distributed control network. This control network deployment begins with a reactor structural analysis of the original parameter data to generate structural data. This structural analysis includes a detailed study of the reactor's geometry, the distribution of internal components, and the functional division of each area, thereby clarifying the reactor's structural characteristics and the functions of each component.

[0030] Based on the obtained structural data, the reactor is divided into zones, generating zone boundary data. The principle of zone division is to divide the reactor into several relatively independent zones based on the differences in function, temperature, pressure, flow rate, and other parameters within the reactor. For example, zones can be divided into feed preheating zones, reaction zones, product cooling zones, and so on, each with clearly defined boundaries.

[0031] Control node locations are placed based on the region boundary data, generating control node location data. Control node placement must ensure accurate monitoring and control of parameters in each region, so representative locations should be selected. For example, control nodes should be placed in the center of each region or in locations sensitive to parameter changes to ensure timely acquisition of accurate data and effective control.

[0032] Distributed control nodes are deployed using multi-dimensional control node location data. Control parameters are calibrated for these deployed distributed control nodes to generate distributed calibrated control nodes. When deploying distributed control nodes, factors such as inter-node communication capabilities, data transmission stability, and reliability must be considered to ensure that each node can work together. After deployment, the control parameters of each control node are calibrated to accurately reflect and control the parameters of the corresponding area.

[0033] The distributed calibration control nodes are linked through communication channels to generate a distributed control network. This channel association ensures fast and accurate data transmission between the control nodes. Appropriate communication protocols and methods are selected to establish stable communication links, enabling efficient operation of the entire distributed control network and comprehensive monitoring and control of the high-pressure hydrogenation reactor. These steps complete step S1, laying a solid foundation for subsequent pressure optimization control.

[0034] Example 2: The specific implementation method of step S2 is: multi-level pressure sensor data acquisition is performed on the distributed control network to obtain multi-level pressure monitoring data, wherein the multi-level pressure monitoring data includes primary pressure monitoring data, intermediate pressure monitoring data and advanced pressure monitoring data. During the acquisition process, the primary pressure sensor is arranged at the inlet section and buffer area of ​​the reactor to obtain the initial pressure state of the raw gas when it enters the reactor in real time. The sampling frequency is set to 1 time per second to capture the instantaneous fluctuation of the inlet pressure; the intermediate pressure sensor is distributed at different height positions of the catalyst bed, such as 1 meter below the top of the bed, 0.5 meters above the middle and bottom, and data is collected every 5 seconds, focusing on monitoring the pressure changes in the active area of ​​the catalyst during the reaction; the advanced pressure sensor is installed at the outlet section of the reactor and the pipeline node before product separation, with a sampling frequency of 0.5 times per second, to record the pressure decay after the reaction and the pressure stability during product transportation.

[0035] Perform multi-level pressure gradient calculations on the primary pressure monitoring data and the advanced pressure monitoring data to obtain the pressure gradient index. Specifically, perform a difference operation on the instantaneous pressure value in the primary pressure monitoring data and the pressure value at the corresponding moment in the advanced pressure monitoring data, and then divide it by the axial distance from the reactor inlet to the outlet to obtain the pressure change gradient per unit length. For example, at a certain moment, the primary pressure is 10.5MPa, the advanced pressure is 9.8MPa, and the axial length of the reactor is 8 meters. The pressure gradient is (10.5-9.8) / 8=0.0875MPa / m, which is the basic data of the pressure gradient index. At the same time, combined with the pressure gradient data of different time periods, analyze its changing trend throughout the reaction cycle, such as the gradient difference between the start-up and shutdown stages and the stable operation stage.

[0036] The impact of reactor pressure-density differences on the pressure gradient index is analyzed based on intermediate pressure monitoring data to generate pressure-density difference data. Because the intermediate pressure monitoring point is located in the catalyst bed, the gas flow state in this area is complex due to the filling of catalyst particles, and the pressure density changes significantly. By comparing the intermediate pressure data at different bed heights at the same time, the pressure difference between adjacent monitoring points is calculated. Then, combined with the gas state equation (ignoring the formula and only describing the logic), the pressure difference is converted into a corresponding density change. For example, if the pressure at the top of the bed is 10.2MPa, the pressure in the middle is 10.0MPa, and the gas temperature is 300°C, the state equation estimates that the gas density difference between the top and middle is 0.5kg / m³. This data is used to construct a pressure-density difference matrix to reflect the density distribution characteristics at different locations within the bed.

[0037] Based on the pressure-density difference data, the pressure gradient index is subjected to pressure fluctuation analysis to generate pressure fluctuation data. Here, it is necessary to correlate the pressure gradient index with the pressure-density difference data in time and space, and analyze the fluctuation amplitude and frequency of the pressure gradient when the density difference changes. For example, when the density of a certain area of ​​the catalyst bed increases by 0.3 kg / m³ due to coking, the corresponding pressure gradient fluctuates from 0.08 MPa / m to 0.12 MPa / m in the next 30 minutes. The start time, duration and fluctuation range of the fluctuation are recorded to form a pressure fluctuation data sequence. At the same time, the characteristic differences of pressure fluctuations under different operating conditions (such as changes in feed gas composition and reaction temperature adjustments) are analyzed. For example, when the load increases by 20%, the pressure fluctuation frequency increases from 2 times per minute to 5 times per minute.

[0038] The pressure fluctuation data is subjected to pressure fluctuation trajectory analysis to generate pressure fluctuation trajectory data. Using the time series analysis method, the curve of the pressure value changing with time in the pressure fluctuation data is digitized to extract key feature points, such as the time and value of the peak and trough. For example, a certain section of pressure fluctuation data shows that the pressure reaches a peak of 10.8MPa at 8:00, drops to a trough of 10.2MPa at 8:15, and rises back to 10.6MPa at 8:30. These time points are matched with the pressure values ​​to form trajectory coordinates, and a continuous pressure fluctuation trajectory curve is drawn by interpolation. At the same time, combined with the spatial position of each monitoring point in the distributed control network, the pressure fluctuation trajectory is visualized in the three-dimensional model of the reactor. For example, different colored lines are used to represent the pressure fluctuation trajectory at different positions in the catalyst bed area.

[0039] The pressure fluctuation trajectory data is used to analyze the degree of reactor pressure variation and generate a pressure variation influencing factor. Specifically, characteristic parameters are first extracted from the pressure fluctuation trajectory data, including fluctuation amplitude (the difference between peak and valley values), fluctuation period (the time interval between adjacent peaks), and fluctuation symmetry (the ratio of rising and falling edges). For example, a trajectory has a fluctuation amplitude of 0.6 MPa, a period of 15 minutes, a rising edge time of 8 minutes, and a falling edge time of 7 minutes. A pressure variation degree assessment model is then established (ignoring formulas and descriptive logic). These characteristic parameters are compared with preset normal operating thresholds to calculate the degree of deviation. For example, if the normal fluctuation amplitude threshold is 0.5 MPa, the deviation of the current amplitude of 0.6 MPa is 20%. Combining the deviations of the fluctuation period and symmetry, a weighted calculation is performed to obtain a comprehensive deviation coefficient, which is the core parameter of the pressure variation influencing factor. In addition, the mutual influence of pressure changes in different regions must also be considered, such as the conduction time and degree of influence of inlet pressure fluctuations on bed pressure changes. Through statistical analysis of historical data, an influence correlation matrix between regions is established, and finally a pressure change influencing factor is generated that can comprehensively reflect the degree of overall pressure change in the reactor.

[0040] Furthermore, the pressure fluctuation trajectory data of the high-pressure hydrogenation reactor is used to perform a reactor pressure fluctuation time series analysis on the pressure fluctuation data of the high-pressure hydrogenation reactor to generate the pressure fluctuation time series data of the high-pressure hydrogenation reactor. During the time series analysis, the pressure fluctuation trajectory data of each monitoring point is arranged in chronological order based on the time axis to form a multi-dimensional time series data set. For example, the pressure fluctuation data of the inlet section, the middle of the bed, and the outlet section are aligned according to the second-level timestamp to construct a three-dimensional time series matrix, in which each row represents the pressure value of each monitoring point at a certain moment. The pressure fluctuation time series data of the high-pressure hydrogenation reactor is converted into a dynamic graph to generate a high-pressure hydrogenation reactor pressure fluctuation dynamic graph. Using computer graphics processing technology, the pressure changes in the time series data are displayed in the form of color gradients or dynamic curve changes, such as using red to represent high-pressure areas and blue to represent low-pressure areas. The colors change dynamically over time, intuitively presenting the spatiotemporal evolution of pressure fluctuations.

[0041] The pressure fluctuation dynamic graph of the high-pressure hydrogenation reactor is analyzed for abnormal reactor pressure events to generate abnormal pressure event data for the high-pressure hydrogenation reactor. By setting abnormality identification rules, if the pressure value exceeds the upper limit of the normal range by 10% or falls below the lower limit by 5%, or the pressure fluctuation rate exceeds 0.2MPa / min, it is determined to be an abnormal event. For example, at a certain moment in the dynamic graph, the pressure in the middle of the bed suddenly rises from 10.0MPa to 10.9MPa, exceeding the normal upper limit of 10.5MPa by 4%, and the rising rate is 0.3MPa / min, the abnormal event mark is triggered, and data such as the time, location, pressure change amplitude and rate of the event are recorded.

[0042] Based on the abnormal pressure event data from the high-pressure hydrogenation reactor, the pressure fluctuation dynamic diagram of the high-pressure hydrogenation reactor is analyzed for normal pressure change trends, generating normal pressure trend change data for the high-pressure hydrogenation reactor. After excluding the abnormal event data, the remaining normal pressure fluctuation data is smoothed using a moving average or exponential smoothing method (ignoring the formula, describing the method) to eliminate short-term fluctuation noise and extract long-term change trends. For example, a 72-hour moving average of the pressure data during the stable operation phase produces a trend line showing a slow decrease in pressure over time with a slope of -0.01 MPa / h, indicating the natural decay rate of pressure under normal operating conditions.

[0043] The abnormal pressure event data of the high-pressure hydrogenation reactor is used to detect the impact of abnormal events on the normal pressure trend change data of the high-pressure hydrogenation reactor, and abnormal pressure impact data is generated. The normal pressure trend data before and after the abnormal event are compared, and the deviation of the abnormal event from the trend line is calculated. For example, before an abnormal event occurs, the normal pressure trend line decreases by 0.01MPa per hour. After the event occurs, the trend line decreases by 0.02MPa per hour, and the deviation is 0.01MPa / h. This is used to quantify the impact of the abnormal event on the pressure trend. Based on the abnormal pressure impact data, the pressure fluctuation data of the high-pressure hydrogenation reactor is analyzed for reactor pressure changes to generate the high-pressure hydrogenation reactor pressure change impact factor. The impact data is combined with the characteristic parameters of the pressure fluctuation, and the weight of each factor is determined by the hierarchical analysis method. Finally, a quantitative indicator that can characterize the overall impact of the abnormal event on the reactor pressure is synthesized, namely the high-pressure hydrogenation reactor pressure change impact factor.

[0044] Example 3: Step S3 is implemented as follows: Reactor valve status is sensed based on the distributed control network to generate valve sensing data. During sensing, data is first collected from the reactor valve sensors in the distributed control network to obtain valve sensor data. The valve sensors include displacement sensors, pressure sensors, and temperature sensors installed on the valves. The displacement sensors collect position signals of the valve opening, the pressure sensors collect pressure signals before and after the valve, and the temperature sensors collect temperature signals of the valve body. These sensors are distributed across various locations on the valve to comprehensively capture valve operating status information. The acquisition frequency is determined based on the valve's importance and operating conditions. For critical valves such as feed valves and hydrogen circulation valves, the acquisition frequency is twice per second. For auxiliary valves such as drain valves, the acquisition frequency is 0.5 times per second.

[0045] Valve sensor data is filtered to generate valve filtered data. Because sensor data may be affected by external factors such as electromagnetic interference and mechanical vibration, generating noise signals, data filtering is necessary. Digital filtering methods, such as Kalman filtering and median filtering (ignoring the specific formulas and describing the logic), are used to remove high-frequency noise and random interference from the data. For example, when valve opening data collected by a displacement sensor experiences a sudden jump in value, a median filter algorithm is used to take the median value of several data points before and after the moment as the valid data at the current moment, eliminating the influence of the outlier. Signal enhancement is performed on the valve filtered data to generate valve enhanced data. Through signal amplification and noise reduction, the signal-to-noise ratio is improved, making the valve status characteristics more distinct. For example, the weak pressure change signal collected by the pressure sensor is amplified by an amplifier, and wavelet noise reduction technology is used to further remove residual noise, making the pressure signal clearer.

[0046] Signal feature analysis is performed on valve enhancement data to generate valve signal feature data. Time and frequency domain analysis is performed on the enhanced signal to extract characteristic parameters that characterize the valve status. In time domain analysis, parameters such as the mean, variance, peak value, rise time, and fall time are calculated. In frequency domain analysis, the time domain signal is converted to a frequency domain signal via Fourier transform, and the frequency components and energy distribution of the signal are analyzed. For example, the time domain curves of the valve opening displacement signal during normal opening and closing processes exhibit specific shapes and characteristic parameters, such as opening time, closing time, and maximum speed. Extracting these parameters can be used to determine whether the valve is operating normally. Regional reactor valve labeling is performed using valve signal feature data to generate regional valve labeling data. Valves are categorized and labeled based on their installation location and the region they belong to within the reactor. For example, valves in the reactor feed area can be labeled as "feed zone valves," while valves in the catalyst bed area can be labeled as "bed zone valves," etc., to facilitate subsequent management and analysis of valves in different regions.

[0047] Reactor valve type identification is performed on regional valve tag data to generate valve type identification data. Valve types, such as gate valves, globe valves, ball valves, and check valves, are identified based on factors such as their structural characteristics and functional purpose. Different valve types have different operating characteristics and failure modes. Valve type identification provides a basis for subsequent valve status assessment and anomaly analysis. For example, the primary failure mode of gate valves is sealing surface leakage, while the primary failure mode of ball valves is ball wear. Valve type identification enables the establishment of targeted fault warning indicators. Reactor valve behavior perception is performed based on valve type identification data to generate valve perception data. This data combines information such as valve type, installation location, and operating parameters to detect and record valve opening and closing movements, opening adjustment, and other behaviors. For example, for a feed valve, the system detects its opening and closing times, the opening adjustment process, and its operating status at different openings, forming a complete record of valve behavior.

[0048] The valve sensor data is then used to collect valve parameters and perform valve status assessment on the collected parameters to obtain valve status data. Valve parameter collection includes real-time collection of parameters such as valve opening, pressure, temperature, flow rate, and vibration. These parameters directly reflect the valve's operating status. For example, collecting valve opening parameters can determine whether the valve is in the normal operating position, while collecting valve vibration parameters can determine whether the valve has a mechanical failure. When assessing the status of the collected valve parameters, a parameter threshold range for normal valve operation is established, and the collected real-time parameters are compared with the threshold range. If the parameters are within the threshold range, the valve is considered normal; if the parameters are outside the threshold range, the valve is considered abnormal. For example, if the normal operating temperature range of a globe valve is set to 50°C to 80°C, a recorded temperature of 90°C indicates that the valve temperature is abnormal.

[0049] The valve status data is matched against the preset valve database. If the match between the valve status data and the preset valve database fails, the corresponding valve status data is marked as abnormal valve status data. If the match between the valve status data and the preset valve database succeeds, the corresponding valve status data is marked as normal valve status data. The preset valve database stores information such as parameter characteristics and failure modes of various types of valves under normal operating conditions. By matching the real-time collected valve status data with the information in the database, it is possible to quickly determine whether the valve is abnormal. For example, the database records that the leakage of a certain type of gate valve in the normally closed state should be less than 0.1L / min. If the leakage of the gate valve is detected to be 0.3L / min, the match fails and the valve is marked as abnormal valve status data.

[0050] Valve opening evaluation is performed on abnormal valve status data to generate abnormal valve opening evaluation data. When a valve status is abnormal, the valve opening is evaluated and the discrepancy between the valve opening and actual process requirements is analyzed. For example, if a feed valve's opening cannot reach the set value, the valve's mechanical structure and actuator are inspected and analyzed to assess the cause and extent of the abnormal valve opening. This abnormal valve opening evaluation data, including the opening deviation and possible fault causes, is generated. Based on the abnormal valve opening evaluation data, the impact of the abnormality is analyzed on normal valve status data to generate a valve variation impact factor. The impact of abnormal valve opening changes on the entire reactor pressure system is analyzed. For example, abnormal feed valve opening reduces feed flow, thereby affecting the pressure distribution and reaction rate within the reactor. By establishing a correlation model between valve anomalies and reactor pressure changes (ignoring formulas and descriptive logic), the abnormal valve opening evaluation data is comprehensively analyzed with other normal valve status data to calculate the extent of the valve anomaly's impact on reactor pressure changes. This factor quantifies the impact of valve status changes on reactor pressure control, providing a basis for subsequent pressure optimization control.

[0051] Example 4: The specific operation of step S4 is as follows: The reactor flow sensor data is collected according to the distributed control network to obtain flow sensor data. The flow sensors include flowmeter sensors, differential pressure sensors, and flow velocity sensors. For example, an electromagnetic flowmeter sensor is installed on the reactor's feed gas inlet pipeline to collect real-time feed gas volume flow rate, with a measurement range of 0-1000 m³ / h and an accuracy of 0.5. Differential pressure sensors are installed above and below the catalyst bed to monitor the pressure difference across the bed to indicate catalyst bed blockage, with a measurement range of 0-10 kPa and an accuracy of 0.1 kPa. An ultrasonic flow velocity sensor is installed on the reactor's product outlet pipeline to measure the product gas flow rate and subsequently calculate the product flow rate, with a measurement range of 0-30 m / s and an accuracy of 1%. These sensors are connected to the data acquisition system via the distributed control network and collect flow-related data in real time at a set sampling frequency (e.g., once per second).

[0052] The flow sensor data is used to analyze the temporal and spatial distribution characteristics of the reactor flow rate, generating flow rate temporal and spatial distribution data. For example, during the 8:00-9:00 AM timeframe, the feed gas inlet flow rate peaks at 850 m³ / h at 8:15 AM, then gradually decreases, stabilizing at around 780 m³ / h at 8:45 AM. The catalyst bed pressure differential is 2.5 kPa at 8:00 AM, rising to 3.2 kPa at 8:30 AM as the reaction proceeds, and then remaining relatively stable. The product outlet flow rate averages 15 m / s from 8:00-8:30 AM and 16 m / s from 8:30-9:00 AM. By analyzing the temporal trends of these data and combining them with the spatial location of the sensors within the reactor (e.g., inlet, bed, outlet), a flow rate distribution diagram is plotted as it changes over time and space. For example, with time as the horizontal axis and the axial position of the reactor as the vertical axis, a spatiotemporal distribution curve of the raw gas flow rate is drawn to intuitively display the flow changes at different positions at different times.

[0053] Flow data sampling and calibration are performed on the spatiotemporal distribution characteristic data of the flow to generate flow calibration data. Since the data collected by the sensor may contain errors, sampling and calibration processing is required. When sampling, the appropriate sampling interval is determined based on the actual process requirements and data processing capabilities, such as sampling once per minute, and the continuous flow data is converted into discrete sampling points. During the calibration process, a standard flow source is used to calibrate the sensor. For example, a standard gas with a known flow rate is passed through the electromagnetic flowmeter, and the sensor's measured value is compared with the standard value. The error is calculated and the sensor's measured data is corrected based on the error. Assuming the standard flow rate is 500m³ / h and the electromagnetic flowmeter's measured value is 505m³ / h, the error is +1%. The subsequent measurement data is corrected using the calibration coefficient to obtain accurate flow calibration data.

[0054] The flow calibration data is used to evaluate the reactor flow state and generate flow state assessment data. A flow state assessment indicator system is established, including flow stability, volatility, and the degree of deviation from the set value. For example, if the target flow rate of the feed gas is set at 800 m³ / h, over a period of time (e.g., 2 hours), the flow calibration data has an average value of 795 m³ / h, a standard deviation of 10 m³ / h, a maximum deviation of +20 m³ / h (i.e., 820 m³ / h), and a minimum deviation of -15 m³ / h (i.e., 785 m³ / h). This data is used to assess the flow state and determine whether the flow is stable within a reasonable range and whether there are significant fluctuations or deviations from the target value. If the flow fluctuates significantly or frequently deviates from the target value, it may affect the reactor's pressure stability and reaction performance, requiring further analysis.

[0055] The flow calibration data is evaluated for distribution characteristics to generate flow distribution characteristic evaluation data. The flow distribution in different areas of the reactor is analyzed. For example, in the catalyst bed area, the flow distribution uniformity within the bed is calculated using differential pressure sensor data and flow velocity sensor data. Assuming that the flow velocities at different locations in the bed vary significantly, with the flow velocity near the reactor wall being 12 m / s and the flow velocity in the center area being 18 m / s, the velocity difference reaches 50%, indicating an uneven flow distribution, which may lead to inconsistent catalyst utilization efficiency and affect the reaction effect. By evaluating the flow distribution characteristics, the uniformity and rationality of the flow distribution within the reactor are determined, providing a basis for subsequent control zone division.

[0056] The reactor control area is divided based on the flow state evaluation data and the flow distribution characteristic evaluation data to generate control area data. Based on the spatiotemporal distribution characteristics of the flow, the state evaluation results, and the distribution characteristic evaluation results, the reactor is divided into different control areas. For example, the reactor is divided into the following control areas: The feed control area, encompassing the feed gas inlet pipeline and surrounding areas, primarily controls the feed gas flow rate to ensure it meets process requirements. The flow status assessment in this area focuses on the stability of the feed gas flow rate and its deviation from the target value.

[0057] The reaction control zone, primarily the catalyst bed, focuses on flow distribution uniformity and pressure differential changes within the bed to ensure uniform reaction and proper catalyst operation. The flow distribution characteristics assessment in this area focuses on analyzing flow velocity differences and pressure differential trends within the bed.

[0058] The product control zone, encompassing the product outlet pipeline and surrounding areas, controls the product flow rate and velocity, ensuring smooth discharge of the product from the reactor while maintaining pressure balance within the reactor. The flow status assessment in this area focuses on the stability of the product flow and its compatibility with the reaction rate.

[0059] When dividing control zones, specific process parameters and equipment structures can be combined to further subdivide the control areas. For example, within the reaction control zone, the catalyst bed height can be divided into three sub-control zones: upper, middle, and lower. Each sub-control zone has independent flow monitoring and control nodes to more accurately control the flow distribution within the bed.

[0060] In order to more clearly demonstrate the basis and results of the control zone division, the following table can be listed as an example:

[0061] Through the above steps, the reactor control zones are divided and detailed control zone data is generated, including the location range, monitoring sensor layout, evaluation indicators, and control targets of each control zone. This control zone data will provide the basis for subsequent pressure optimization calculations and control, enabling precise adjustment of pressure control based on the flow characteristics and process requirements of different zones, thereby improving the reactor's operational stability and reaction efficiency.

[0062] Example 5: The implementation method of step S5 is: based on the control zone data, the flow calibration data and the valve status data are analyzed for the reactor control chain to generate control chain data. For example, assuming that the reactor is divided into three main control zones: the feed control zone, the reaction control zone, and the product control zone, the flow calibration data of the feed control zone shows that the raw gas flow rate fluctuates from 800m³ / h to 750m³ / h within a certain period of time. At the same time, the valve status data of the reaction control zone shows that the opening of the feed valve slowly decreases from 50% to 45%. At this time, it is necessary to analyze the correlation between the flow fluctuation and the change in valve opening, and establish the control chain logic: the feed valve opening decreases → the raw gas flow rate decreases → the raw material supply amount in the reaction control zone decreases → the product generation amount changes accordingly → the product flow rate in the product control zone changes accordingly. Through this causal correlation analysis, the parameter changes in each control zone are connected in series to form a complete control chain, the path and influence relationship of the parameter transmission are clarified, and control chain data containing the parameter correlation information of each link is generated.

[0063] Based on control chain data, reactor pressure optimization calculations are performed to generate a pressure optimization index. For example, considering the relationship between feed flow and valve opening in the control chain, adjusting the feed valve opening from 50% to 48% can result in a corresponding decrease in feed gas flow of approximately 3%, based on historical data. This is combined with the dynamic relationship between pressure and flow within the reactor (e.g., a decrease in flow leads to a decrease in reactor inlet pressure). A pressure optimization calculation model (ignoring formulas and descriptive logic) comprehensively considers the impact of parameter changes in each control zone on pressure. For example, a reduction in feed flow rate reduces the inlet pressure by 0.2 MPa. At the same time, the reaction rate in the reaction control zone decreases due to insufficient raw materials, which reduces the bed pressure by 0.1 MPa. The reduction in product flow rate in the product control zone reduces the outlet pressure by 0.15 MPa. These pressure changes are calculated by weight (such as inlet pressure impact weight 0.4, bed pressure impact weight 0.3, outlet pressure impact weight 0.3), and the comprehensive pressure change value is 0.2×0.4+0.1×0.3+0.15×0.3=0.155 MPa. This is used as the basic data of the pressure optimization index, reflecting the comprehensive impact of the current control chain parameter changes on the pressure.

[0064] The pressure optimization index, pressure variation impact factor, and valve variation impact factor are integrated to generate comprehensive control data. Assuming the pressure optimization index is 0.155 MPa, the pressure variation impact factor is determined through analysis in Example 2 as a quantitative value reflecting the pressure fluctuation amplitude and the impact of abnormal events (e.g., 0.8, with a larger value indicating a greater degree of pressure variation). The valve variation impact factor is determined through analysis in Example 3 as a quantitative value representing the impact of valve abnormalities on pressure (e.g., 0.6, with a larger value indicating a greater impact of valve abnormalities). These data are normalized (e.g., converted to a numerical range of 0-1) and weighted according to a preset integration rule (e.g., a weight of 0.5 for the pressure optimization index, a weight of 0.3 for the pressure variation impact factor, and a weight of 0.2 for the valve variation impact factor). The resulting comprehensive control data value is 0.155 × 0.5 + 0.8 × 0.3 + 0.6 × 0.2 = 0.4475, which comprehensively reflects the degree of influence of various factors related to the current reactor pressure control.

[0065] The comprehensive control data is divided into data sets to generate model training sets and model test sets. For example, 1,000 sets of comprehensive control data from the past three months are collected and divided into training sets and test sets in a ratio of 7:3. The training set contains 700 sets of data for model training, and the test set contains 300 sets of data for model optimization and iteration. When dividing, ensure that the data set covers comprehensive control data under different operating conditions (such as normal operation, start-up and shutdown, and load adjustment) to ensure the generalization ability of the model. For example, the training set contains 400 sets of normal operation data, 200 sets of load adjustment data, and 100 sets of start-up and shutdown data, and the test set contains 150 sets of normal operation data, 90 sets of load adjustment data, and 60 sets of start-up and shutdown data, so that the distribution of the data set is basically consistent with the proportion of actual operating conditions.

[0066] An optimization algorithm is used to train the model training set to generate a pressure control signal training model. An appropriate optimization algorithm (such as a neural network algorithm, support vector machine algorithm, etc., ignoring specific formulas and using description logic) is selected, using the comprehensive control data in the training set as input and the corresponding pressure control signal as output (such as a valve opening control signal or a compressor speed control signal). Using a neural network algorithm as an example, a network structure is constructed consisting of an input layer, a hidden layer, and an output layer. The number of input layer nodes is determined by the characteristic dimensions of the comprehensive control data (for example, if the comprehensive control data includes three features: pressure optimization index, pressure variation influencing factor, and valve variation influencing factor, the number of input layer nodes is 3). The number of output layer nodes is determined by the type of pressure control signal (for example, if the control signal includes two signals: feed valve opening adjustment and hydrogen replenishment amount adjustment, the number of output layer nodes is 2). The network parameters are iteratively optimized using the training set data, adjusting the weights and biases to gradually reduce the error between the model output and the actual pressure control signal until the preset training accuracy requirements are met, thus generating a pressure control signal training model.

[0067] The pressure control signal training model is optimized and iterated using the model test set to generate a pressure control signal model. The 300 sets of comprehensive control data from the test set are input into the trained model to obtain the model-predicted pressure control signal. This is then compared with the actual pressure control signal, and the prediction error (such as mean squared error, mean absolute error, etc.) is calculated. If the error exceeds a preset threshold (e.g., mean squared error greater than 0.05), the cause of the error is analyzed, the model structure is adjusted, or the algorithm parameters are optimized, and training and testing are repeated until the model's prediction error on the test set meets the requirements. For example, if the mean squared error is 0.08 during the first test, exceeding the threshold of 0.05, by increasing the number of hidden layer nodes or adjusting the learning rate, the mean squared error is reduced to 0.04 after retraining, meeting the requirements, and finally generating the pressure control signal model.

[0068] The integrated control data is imported into the pressure control signal model to generate the pressure control signal data. Taking the current integrated control data of 0.4475 as an example, the pressure control signal model is input. The model then outputs the corresponding pressure control signal based on the parameter relationships obtained through training and optimization. Assuming the model outputs a 2% increase in the feed valve opening and a 5% increase in the hydrogen replenishment volume, the generated pressure control signal data contains these specific control instructions and related parameters (such as the control signal's execution priority and time window).

[0069] The pressure control signal data is output as a command to perform multi-stage pressure optimization control. The generated pressure control signal data is transmitted via a distributed control network to the corresponding actuators (such as the feed valve actuator and the hydrogen compressor controller). These actuators adjust valve opening or equipment operating parameters based on the command to achieve optimized control of reactor pressure. For example, upon receiving a command to increase the opening by 2%, the feed valve actuator actuates the valve, increasing the opening from 48% to 50%. This increases the feed gas flow rate and the reactor inlet pressure, thereby affecting the pressure distribution throughout the reactor through the control chain, achieving optimized multi-stage pressure control. During the command output process, information such as the execution time and status of the control signal is recorded to facilitate subsequent evaluation and traceability of the control effect. This complete process, from data integration and model training to control signal generation and execution, ensures precise and optimized multi-stage pressure control of the high-pressure hydrogenation reactor to meet process operating requirements.

[0070] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0071] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A multi-stage pressure optimization control method for a high-pressure hydrogenation reactor, characterized in that: The following steps are involved: Step S1: acquiring temperature monitoring data, pressure monitoring data, and flow monitoring data of a high-pressure hydrogenation reactor; performing reactor parameter analysis on the temperature monitoring data, pressure monitoring data, and flow monitoring data to generate original parameter data; and deploying a control network based on the original parameter data to obtain a distributed control network; Step S2: collecting and analyzing multi-level pressure sensor data on the distributed control network to generate pressure change influencing factors; Step S3: sensing the status of the reactor valve according to the distributed control network, obtaining valve status data, and generating a valve change influencing factor based on the valve status data; Step S4: collecting data from the reactor flow sensor according to the distributed control network to obtain flow sensor collected data; Perform flow data sampling and calibration on the data collected by the flow sensor to generate flow calibration data; Divide the reactor control area based on the flow calibration data and generate control area data; Step S5: performing reactor pressure optimization calculation on the flow calibration data and valve status data based on the control zone data to generate a pressure optimization index; integrating the pressure optimization index, the pressure change influencing factor, and the valve change influencing factor to generate comprehensive control data; generating a pressure control signal based on the comprehensive control data to generate pressure control signal data; The pressure control signal data is commanded to output to perform multi-level pressure optimization control operations.

2. The multi-stage pressure optimization control method for a high-pressure hydrogenation reactor according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: obtaining a high-pressure hydrogenation reactor temperature sample, a high-pressure hydrogenation reactor pressure sample, and a high-pressure hydrogenation reactor flow sample; Step S12: performing parameter analysis on the high-pressure hydrogenation reactor temperature sample, the high-pressure hydrogenation reactor pressure sample, and the high-pressure hydrogenation reactor flow rate sample to obtain temperature monitoring data, pressure monitoring data, and flow rate monitoring data of the high-pressure hydrogenation reactor; Step S13: performing reactor parameter analysis on the temperature monitoring data, pressure monitoring data, and flow monitoring data to generate original parameter data; Step S14: Deploy the control network according to the original parameter data to obtain a distributed control network.

3. The multi-stage pressure optimization control method for a high-pressure hydrogenation reactor according to claim 2, characterized in that: Step S14 includes the following steps: Step S141: performing reactor structure analysis on the original parameter data to obtain structure data; Step S142: Divide the reactor area according to the structural data to generate area boundary data; Step S143: arranging control node positions based on the area boundary data to generate control node position data; Step S144: deploying distributed control nodes using the multi-dimensional control node position data, and calibrating control parameters of the deployed distributed control nodes, thereby generating distributed calibrated control nodes; Step S145: performing communication channel association on the distributed calibration control nodes to generate a distributed control network.

4. The multi-stage pressure optimization control method for a high-pressure hydrogenation reactor according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: performing multi-level pressure sensor data acquisition on the distributed control network to obtain multi-level pressure monitoring data, wherein the multi-level pressure monitoring data includes primary pressure monitoring data, intermediate pressure monitoring data, and advanced pressure monitoring data; Step S22: performing multi-level pressure gradient calculation on the primary pressure monitoring data and the advanced pressure monitoring data to obtain a pressure gradient index; performing reactor pressure-density difference impact analysis on the pressure gradient index based on the intermediate pressure monitoring data to generate pressure-density difference data; Step S23: performing pressure fluctuation analysis on the pressure gradient index based on the pressure density difference data to generate pressure fluctuation data; performing pressure fluctuation trajectory analysis on the pressure fluctuation data to generate pressure fluctuation trajectory data; Step S24: analyzing the degree of reactor pressure variation on the pressure fluctuation data using the pressure fluctuation trajectory data to generate a pressure variation influencing factor.

5. The multi-stage pressure optimization control method for a high-pressure hydrogenation reactor according to claim 4, characterized in that: Step S24 includes the following steps: Step S241: performing a reactor pressure fluctuation time series analysis on the high-pressure hydrogenation reactor pressure fluctuation data using the high-pressure hydrogenation reactor pressure fluctuation trajectory data to generate high-pressure hydrogenation reactor pressure fluctuation time series data; performing a dynamic graph conversion on the high-pressure hydrogenation reactor pressure fluctuation time series data to generate a high-pressure hydrogenation reactor pressure fluctuation dynamic graph; Step S242: performing reactor pressure abnormality event analysis on the high-pressure hydrogenation reactor pressure fluctuation dynamic diagram to generate high-pressure hydrogenation reactor abnormal pressure event data; Step S243: performing a normal pressure change trend analysis on the pressure fluctuation dynamic diagram of the high-pressure hydrogenation reactor based on the abnormal pressure event data of the high-pressure hydrogenation reactor to generate normal pressure trend change data of the high-pressure hydrogenation reactor; Step S244: Using the abnormal pressure event data of the high-pressure hydrogenation reactor, the impact degree of the abnormal event is detected on the normal pressure trend change data of the high-pressure hydrogenation reactor to generate abnormal pressure impact degree data; based on the abnormal pressure impact degree data, the reactor pressure change analysis is performed on the pressure fluctuation data of the high-pressure hydrogenation reactor to generate a high-pressure hydrogenation reactor pressure change impact factor.

6. The multi-stage pressure optimization control method for a high-pressure hydrogenation reactor according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: sensing the status of the reactor valve according to the distributed control network and generating valve sensing data; Step S32: collecting valve parameters from the valve sensing data, and performing valve status identification on the collected valve parameters to obtain valve status data; Step S33: matching the valve status data with the preset valve database. If the valve status data does not match the preset valve database, the corresponding valve status data is marked as abnormal valve status data. If the valve status data does match the preset valve database, the corresponding valve status data is marked as normal valve status data. Step S34: performing valve opening evaluation on the abnormal valve status data to generate abnormal valve opening evaluation data; performing valve abnormality impact analysis on the normal valve status data based on the abnormal valve opening evaluation data to generate a valve change impact factor.

7. The multi-stage pressure optimization control method for a high-pressure hydrogenation reactor according to claim 6, characterized in that: Step S31 includes the following steps: Step S311: collecting reactor valve sensor data on the distributed control network to obtain valve sensor data; Step S312: performing data filtering on the valve sensor data to generate valve filtered data; performing signal enhancement on the valve filtered data to generate valve enhanced data; Step S313: performing signal feature analysis on the valve enhancement data to generate valve signal feature data; performing regional reactor valve marking based on the valve signal feature data to generate regional valve marking data; Step S314: performing reactor valve type identification on the regional valve tag data to generate valve type identification data; performing reactor valve behavior perception based on the valve type identification data to generate valve perception data.

8. The multi-stage pressure optimization control method for a high-pressure hydrogenation reactor according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: collecting flow sensor data from the reactor according to the distributed control network to obtain flow sensor collected data, wherein the flow sensor includes a flow meter sensor, a pressure difference sensor, and a flow velocity sensor; Step S42: performing flow temporal and spatial distribution characteristic analysis on the flow sensor data to generate flow temporal and spatial distribution characteristic data; performing flow data sampling and calibration on the flow temporal and spatial distribution characteristic data to generate flow calibration data; Step S43: performing a reactor flow state evaluation on the flow calibration data to generate flow state evaluation data; performing a distribution characteristic evaluation on the flow calibration data to generate flow distribution characteristic evaluation data; Step S44: dividing the reactor control area according to the flow state evaluation data and the flow distribution characteristic evaluation data to generate control area data.

9. The multi-stage pressure optimization control method for a high-pressure hydrogenation reactor according to claim 1, characterized in that: Step S5 includes the following steps: Step S51: performing reactor control chain analysis on the flow calibration data and valve status data based on the control zone data to generate control chain data; Step S52: performing reactor pressure optimization calculation based on the control chain data to generate a pressure optimization index; integrating the pressure optimization index, the pressure change influencing factor, and the valve change influencing factor to generate comprehensive control data; Step S53: Divide the comprehensive control data into data sets to generate a model training set and a model test set; use an optimization algorithm to perform model training on the model training set to generate a pressure control signal training model; use the model test set to perform model optimization iteration on the pressure control signal training model to generate a pressure control signal model; Step S54: importing the integrated control data into the pressure control signal model to generate a pressure control signal and generate pressure control signal data; outputting instructions to the pressure control signal data to perform a multi-level pressure optimization control operation.

10. A multi-stage pressure optimization control system for a high-pressure hydrogenation reactor, characterized in that: A system for executing the multi-stage pressure optimization control method for a high-pressure hydrogenation reactor according to any one of claims 1 to 9, comprising: The control network deployment module is used to obtain the temperature monitoring data, pressure monitoring data, and flow monitoring data of the high-pressure hydrogenation reactor; perform reactor parameter analysis on the temperature monitoring data, pressure monitoring data, and flow monitoring data to generate original parameter data; and deploy the control network based on the original parameter data to obtain a distributed control network; The reactor pressure analysis module is used to collect multi-level pressure sensor data from the distributed control network to obtain pressure monitoring data; perform multi-level pressure gradient calculation on the pressure monitoring data to obtain a pressure gradient index; perform pressure fluctuation analysis on the pressure gradient index to generate pressure fluctuation data; perform pressure fluctuation trajectory analysis on the pressure fluctuation data to generate pressure fluctuation trajectory data; and perform reactor pressure change analysis on the pressure fluctuation data using the pressure fluctuation trajectory data to generate a pressure change influencing factor. The reactor valve analysis module is used to sense the reactor valve status based on the distributed control network and generate valve sensing data; collect valve parameters from the valve sensing data and perform valve status identification on the collected valve parameters to obtain valve status data; analyze the impact of valve anomalies on the valve status data and generate valve change impact factors; The reactor control zone analysis module is used to collect data from the reactor flow sensor according to the distributed control network to obtain flow sensor collected data; perform flow data sampling and calibration on the flow sensor collected data to generate flow calibration data; divide the reactor control zone based on the flow calibration data to generate control zone data; The pressure control signal generation module is used to perform reactor pressure optimization calculations on flow calibration data and valve status data based on control zone data to generate a pressure optimization index; integrate the pressure optimization index, pressure change influencing factor, and valve change influencing factor to generate comprehensive control data; generate a pressure control signal based on the comprehensive control data to generate pressure control signal data; and output instructions for the pressure control signal data to execute multi-stage pressure optimization control operations.

Images