Preprocessing device intelligent work starting method and system based on multi-model algorithm
By employing a multi-model algorithm-based collaborative control and real-time monitoring method for intelligent start-up of pretreatment units, the problems of low operational accuracy, low efficiency, and high safety risks during the start-up of pretreatment units in the oil refining and chemical industry have been solved, achieving efficient, safe, and stable automated start-up control.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-22
- Publication Date
- 2026-03-10
AI Technical Summary
The pretreatment units in the oil refining and chemical industry rely on manual operation during startup, which results in low operational accuracy, low efficiency, and high safety risks. Furthermore, existing automated control methods are unable to cope with complex operating conditions and multivariate coupling problems, especially lacking effective intelligent regulation in key stages such as cold circulation, dehydration and hot pressing, opening to atmospheric pressure, and opening to depressurization.
An intelligent start-up method for the pretreatment unit based on a multi-model algorithm is adopted. Through the coordinated work of the cold cycle sub-model, temperature control model, atmospheric pressure control algorithm and pressure reduction control algorithm, the heating furnace temperature and side line system are intelligently regulated in stages. Combined with real-time monitoring and early warning mechanisms, the automated and intelligent start-up of the pretreatment unit is achieved.
It improves the control precision and efficiency of the start-up process, reduces energy consumption, enhances safety, ensures stable product quality, avoids human error, and optimizes start-up sequence and operational standardization.
Smart Images
Figure CN121635221A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to an intelligent start-up method and system for a preprocessing device based on a multi-model algorithm, belonging to the field of industrial process control technology. Background Technology
[0002] As a capital-intensive, technology-intensive, and complex basic industry, the oil refining and chemical industry's start-up process is a key node connecting the static state of equipment with the dynamic operation of production. It involves the linkage of multiple units, the control of multiple parameters, and the prevention and control of high risks, which is crucial to the safe operation and efficiency improvement of enterprises. In the process of developing this invention, the inventors discovered at least the following problems in the existing technology: The start-up process of pretreatment units in the oil refining and chemical industry traditionally relies on manual operation, resulting in low operational precision, low efficiency, and high safety risks. Traditional start-up methods primarily depend on the operator's experience, which can easily lead to judgment errors due to differences in experience and fatigue, affecting equipment lifespan and product quality, and even triggering interlocking shutdowns. Furthermore, the traditional manual start-up process is time-consuming, consumes a large amount of energy and materials, and increases the company's operating costs.
[0003] While some automated control methods exist in existing technologies, most employ single models or simple control strategies, making it difficult to address the complex operating conditions and multivariate coupling issues during the start-up of pretreatment units. Particularly in critical stages such as cold cycling, dehydration and hot pressing, atmospheric pressure activation, and depressurization activation, existing technologies lack effective intelligent control methods, hindering the achievement of fully automated and intelligent start-up across the entire process. Summary of the Invention
[0004] To address the aforementioned issues, this invention proposes an intelligent start-up method and system for preprocessing devices based on a multi-model algorithm. This method enables intelligent start-up of the preprocessing device, reduces human intervention, improves start-up accuracy and safety, and avoids the impact of human error on the start-up process.
[0005] The technical solution adopted by this invention to solve its technical problem is as follows: In a first aspect, the present invention provides an intelligent start-up method for a preprocessing device based on a multi-model algorithm, comprising the following steps: Obtain relevant parameters of the cold cycle, and determine the corresponding target cold cycle sub-model and target regulating valve based on the cold cycle stage to which the relevant parameters belong; The relevant parameters of the cold cycle are input into the target cold cycle sub-model to obtain the first target adjustment parameter of the target regulating valve, and the target regulating valve is adjusted according to the first target adjustment parameter. When the cold cycle process meets the closed-loop cycle condition, the dehydration heat-related parameters and the set temperature curve are input into the temperature control model to obtain the second target adjustment parameters of the main fuel valve of the heater, and control the heater to gradually heat up and maintain the temperature. An atmospheric pressure tower side-stream delivery system is established by using an open atmospheric pressure control algorithm to intelligently regulate the open side-stream of the atmospheric pressure tower. A side-line power supply system for the pressure reducing tower is established by using an open / closed pressure reducing control algorithm, and the open side-line of the pressure reducing tower is intelligently regulated.
[0006] As one possible implementation of this embodiment, determining the corresponding target cold cycle sub-model and target regulating valve based on the cold cycle stage to which the cold cycle-related parameters belong includes: Based on the cold cycle related parameters, query the database to obtain the cold cycle sub-model and the regulating valve that have a mapping relationship with the cold cycle related parameters. Use the cold cycle sub-model as the target cold cycle sub-model for the cold cycle related parameters, and use the regulating valve as the target regulating valve for the cold cycle related parameters.
[0007] As one possible implementation of this embodiment, the set temperature curve includes a first set temperature curve, a second set temperature curve, and a third set temperature curve; the step of inputting the dehydration heat-related parameters and the set temperature curve into the temperature control model to obtain the second target adjustment parameter of the heating furnace fuel main valve includes: The dehydration heat-related parameters and the first set temperature curve are input into the temperature control model. The temperature control model processes the dehydration heat-related parameters and the first set temperature curve, and outputs a second target adjustment parameter for the main fuel valve of the heating furnace to control the temperature of the heating furnace to rise to the first target temperature corresponding to the first set temperature curve. When the furnace temperature is kept constant at the first target temperature for a first target time and the first condition is met, the dehydration thermally relevant parameters and the second set temperature curve are input into the temperature control model. The temperature control model processes the dehydration thermally relevant parameters and the second set temperature curve, and outputs a second target adjustment parameter for the furnace fuel main valve to control the furnace temperature to rise to the second target temperature corresponding to the second set temperature curve. The first condition includes the water content of the crude oil pipeline at the outlet of the electric desalting tank being less than or equal to the target water content. When the furnace temperature is kept constant at the second target temperature for the second target time and the second condition is met, the dehydration heat-related parameters and the third set temperature curve are input into the temperature control model. The temperature control model processes the dehydration heat-related parameters and the second set temperature curve to output a second target adjustment parameter for the furnace fuel main valve, so as to control the furnace temperature to rise to the third target temperature corresponding to the third set temperature curve and keep it constant at the third target temperature for the third target time. The second condition includes the top vapor dew point of the distillation column being less than or equal to the target dew point temperature.
[0008] As one possible implementation of this embodiment, the step of establishing an atmospheric pressure tower side-stream delivery system through an open atmospheric pressure control algorithm and intelligently regulating the open side-stream of the atmospheric pressure tower includes: Input the relevant parameters of atmospheric pressure and the atmospheric pressure set temperature curve into the atmospheric pressure temperature control model. The atmospheric pressure temperature control model outputs the third target adjustment parameters of the atmospheric pressure furnace fuel branch valve and the atmospheric pressure furnace flue gas damper so that the atmospheric pressure furnace temperature gradually increases to the atmospheric pressure target temperature at the set rate. The opening command of the steam regulating valve is determined based on the relevant parameters of steam boosting and the steam boosting control algorithm, and the steam regulating valve is adjusted based on the opening command. Real-time monitoring of parameters related to the establishment of atmospheric pressure tower sidelines; when the parameters related to the establishment of atmospheric pressure tower sidelines meet the establishment conditions, generating establishment prompt information for the corresponding atmospheric pressure tower sidelines.
[0009] As one possible implementation of this embodiment, the step of determining the opening command of the steam regulating valve based on steam rise parameters and the steam rise control algorithm includes: ,in The proportional control gain coefficient is the target flow rate. The target steam flow rate at any given time. Target traffic The target steam flow rate is the final value, and the rate of increase is the preset rate at which the steam flow rate increases. This refers to the cumulative time since the increase in volume began.
[0010] As one possible implementation of this embodiment, the step of establishing a side-line delivery system for the pressure reducing tower through an open / closed pressure reducing control algorithm, and intelligently regulating the open / closed side-line of the pressure reducing tower, includes: Input the relevant parameters for opening and closing the pressure reducing valve and the pressure reducing set temperature curve into the pressure reducing temperature control model. The pressure reducing temperature control model outputs the fourth target adjustment parameters for the pressure reducing furnace fuel branch valve and the pressure reducing furnace flue gas damper, so that the pressure reducing furnace temperature gradually increases to the pressure reducing target temperature according to the set rate. When the outlet temperature of the heating furnace reaches the target temperature threshold, a prompt will be made to perform a vacuuming operation on the pressure reducing tower. Real-time monitoring of parameters related to the establishment of the pressure relief tower sideline; when the parameters related to the establishment of the pressure relief tower sideline meet the establishment conditions, generating establishment prompt information for the corresponding pressure relief tower sideline.
[0011] As one possible implementation of this embodiment, the intelligent start-up method for the preprocessing device based on a multi-model algorithm includes the following steps: During the intelligent start-up process, relevant operating parameters are monitored in real time, and when the relevant operating parameters meet abnormal conditions, corresponding alarm information is generated.
[0012] Secondly, an intelligent start-up system for a preprocessing device based on a multi-model algorithm, provided by embodiments of the present invention, includes: The data acquisition module is used to acquire cold cycle related parameters and determine the corresponding target cold cycle sub-model and target regulating valve according to the cold cycle stage to which the cold cycle related parameters belong. The cold cycle control module is used to input the cold cycle related parameters into the target cold cycle sub-model, obtain the first target adjustment parameter of the target regulating valve, and adjust the target regulating valve according to the first target adjustment parameter; The temperature control module is used to input the dehydration heat-related parameters and the set temperature curve into the temperature control model when the cold cycle process meets the closed-loop cycle conditions, to obtain the second target adjustment parameter of the main fuel valve of the heating furnace, and to control the heating furnace to gradually heat up and maintain a constant temperature. The atmospheric pressure tower control module is used to establish an atmospheric pressure tower side-stream delivery system through an open atmospheric pressure control algorithm, and to intelligently regulate the open side-stream of the atmospheric pressure tower. The pressure reducing tower control module is used to establish a side-line external transmission system of the pressure reducing tower through the pressure reducing control algorithm, and to intelligently regulate the opening of the side-line of the pressure reducing tower.
[0013] Thirdly, an electronic device provided by an embodiment of the present invention includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of any of the above-described intelligent start-up methods for preprocessing devices based on multi-model algorithms.
[0014] Fourthly, embodiments of the present invention provide a storage medium storing a computer program, which, when run by a processor, executes the steps of any of the above-described intelligent start-up methods for preprocessing devices based on multi-model algorithms.
[0015] The beneficial effects of the technical solutions of the embodiments of the present invention are as follows: First, the above-mentioned technical solution overcomes the technical problem of insufficient control accuracy of traditional single model by adopting the technical means of multi-model collaborative control, specifically including the collaborative work of cold cycle sub-model, temperature control model, atmospheric pressure control algorithm and depressurization control algorithm. This achieves the technical effect of improving the control accuracy of the pretreatment device start-up process and realizes precise control of key stages such as cold cycle, dehydration and hot pressing, atmospheric pressure and depressurization. Second, the above technical solution adopts a phased intelligent control approach, specifically including determining the corresponding target cold cycle sub-model and target regulating valve based on cold cycle related parameters, and controlling the heating furnace to rise and maintain a constant temperature in stages according to the set temperature curve. This overcomes the technical problem of low efficiency in traditional manual operation, thereby achieving the technical effect of significantly improving start-up efficiency and reducing energy consumption, and realizing the automation and intelligence of the pretreatment device start-up process. Third, the above technical solution overcomes the technical problem of large safety hazards in traditional methods by adopting real-time monitoring and early warning technology, specifically including real-time monitoring of relevant operating parameters during intelligent start-up and generating corresponding alarm information when abnormal conditions are met, thereby achieving the technical effect of improving the safety of the start-up process and effectively avoiding safety accidents caused by operational errors. Fourth, the above technical solution overcomes the technical problem of unstable product quality in traditional methods by adopting the technical means of establishing a side-stream delivery system, specifically by establishing an atmospheric pressure tower side-stream delivery system through an open atmospheric pressure control algorithm and establishing a vacuum tower side-stream delivery system through an open vacuum control algorithm. This achieves the technical effect of improving product separation efficiency and product quality, ensuring the efficient separation and stable output of light fractions such as diesel and kerosene. Fifth, the above technical solution overcomes the technical problem of relying on human experience to make judgments in traditional methods by using database query mapping relationship technology, specifically including querying the database to obtain the corresponding target cold cycle sub-model and target regulating valve according to the cold cycle related parameters. This achieves the technical effect of improving the accuracy and consistency of start-up and avoids judgment deviations caused by differences in personnel experience. Sixth, the above technical solution adopts intelligent prompting technology, specifically including generating corresponding establishment prompt information when the relevant parameters for establishing the atmospheric pressure tower side line and the relevant parameters for establishing the vacuum tower side line meet the establishment conditions. This overcomes the technical problem of inaccurate timing of operation in traditional methods, thereby achieving the technical effects of optimizing the start-up sequence and improving the standardization of operation, ensuring the orderly establishment of each side line.
[0016] The above technical solution realizes the automation and intelligence of the pretreatment unit start-up process, reduces human intervention, improves start-up accuracy and safety, reduces energy and material consumption, avoids the impact of human operation errors on the start-up process, and is applicable to intelligent start-up control of pretreatment units in oil refining, chemical and other fields. Attached Figure Description
[0017] Figure 1 This is a flowchart illustrating an intelligent start-up method for a preprocessing device based on a multi-model algorithm, according to an exemplary embodiment. Figure 2 This is a schematic diagram illustrating the principle of an intelligent start-up system for a preprocessing device based on a multi-model algorithm, according to an exemplary embodiment. Detailed Implementation
[0018] To more clearly illustrate the technical features of the present invention, the invention will be described in detail below through specific embodiments and in conjunction with the accompanying drawings. The following description is merely exemplary and is not intended to limit the invention.
[0019] like Figure 1 As shown in the figure, an embodiment of the present invention provides an intelligent start-up method for a preprocessing device based on a multi-model algorithm. The multi-model algorithm includes a cold cycle model, a temperature control model, an atmospheric pressure control algorithm, and a depressurization control algorithm. The cold cycle model includes multiple cold cycle sub-models. The method includes the following steps: Step S1: Obtain the relevant parameters of the cold cycle, and determine the corresponding target cold cycle sub-model and target regulating valve from the cold cycle model according to the cold cycle stage to which the relevant parameters of the cold cycle belong.
[0020] In some embodiments, the start-up process of the pretreatment unit includes cold circulation, dehydration and hot tightening, atmospheric pressure start-up, and vacuum pressure start-up. Crude oil needs to be safely injected into the electrostatic desalting tank so that it can be desalted and partially dehydrated before entering the primary distillation tower. After preliminary fractionation in the primary distillation tower, it enters the atmospheric distillation tower. Diesel, kerosene, and other products are fractionated in the atmospheric distillation tower. Atmospheric residue is fractionated in the vacuum distillation tower to obtain lubricating oil components and catalytic feedstock. The cold circulation process circulates the crude oil through the electrostatic desalting tank, primary distillation tower, atmospheric distillation tower, and vacuum distillation tower. During the dehydration and hot tightening process, the temperature is gradually increased and maintained constant in the heating furnace to achieve dehydration and hot tightening. During the atmospheric pressure start-up process, an intelligent control system is established to establish a side-stream delivery system for the atmospheric distillation tower, achieving efficient separation and stable output of light fractions such as diesel and kerosene. During the vacuum pressure start-up process, atmospheric residue is separated under vacuum to obtain lubricating oil components and catalytic cracking feedstock. In some embodiments, the intelligent start-up method of this embodiment relies on a DCS control system.
[0021] In some embodiments, cold cycle parameters include, but are not limited to, the flow rates, pressures, and temperatures of the three pre-desalting feed lines, and the differential pressure of the mixer. The cold cycle stages include, but are not limited to, electric desalting oil loading, primary distillation column oil loading, atmospheric distillation column oil loading, and vacuum distillation column oil loading. For example, during the electric desalting oil loading stage, a stable crude oil feed is required; during the primary distillation column oil loading stage, a smooth rise in the column level is required to prevent vapor lock; during the atmospheric distillation column oil loading stage, stable multi-feed flow and furnace temperature are required; and during the vacuum distillation column oil loading stage, sudden changes in liquid level are required.
[0022] Step S2: Input the cold cycle related parameters into the target cold cycle sub-model, process the cold cycle related parameters through the target cold cycle sub-model to obtain the first target adjustment parameter of the target regulating valve, and adjust the target regulating valve according to the first target adjustment parameter.
[0023] In some embodiments, the cold cycle sub-model includes, but is not limited to, the desalting oil loading model, the primary distillation column oil loading model, the atmospheric distillation column oil loading model, and the vacuum distillation column oil loading model. In some embodiments, at different cold cycle stages, it is necessary to determine the corresponding adjustment parameters (e.g., target control valve opening) of the target control valve based on different cold cycle-related parameters in order to achieve intelligent control of the target control valve. In some embodiments, the system includes a database that records the target cold cycle sub-model and target control valve corresponding to each cold cycle-related parameter, so that the relevant cold cycle-related parameters are input into the corresponding target cold cycle sub-model and the first target adjustment parameter for the target control valve is output. For a detailed explanation of this part, please refer to the corresponding embodiments below, which will not be repeated here. In this embodiment, the cold cycle process is divided into multiple cold cycle stages, and each cold cycle stage corresponds to a cold cycle sub-model. Intelligent control of the cold cycle stage is achieved through the cold cycle sub-model.
[0024] Step S3: When the cold cycle process meets the closed-loop cycle condition, the dehydration heat-related parameters and the set temperature curve are input into the temperature control model. The temperature control model processes the dehydration heat-related parameters and the set temperature curve, and outputs the second target adjustment parameter for the main fuel valve of the heating furnace to control the heating furnace to gradually heat up and maintain a constant temperature.
[0025] In some embodiments, closed-loop conditions include, but are not limited to, acceptable crude oil water content. In some embodiments, the set temperature curve includes, but is not limited to, a temperature-time function curve (e.g., a temperature-time curve), and the set temperature curve includes a target temperature and a isothermal time. The temperature control model employs an LSTM+reinforcement learning model architecture. In some embodiments, the heating furnace includes an atmospheric pressure heating furnace and a vacuum heating furnace. Dehydration and hot-tightening related parameters include, but are not limited to, the outlet temperature of the atmospheric pressure heating furnace, the fuel gas pressure of the atmospheric pressure heating furnace, the fuel gas flow rate of the atmospheric pressure heating furnace, the flue gas oxygen content of the atmospheric pressure heating furnace, the furnace tube surface temperature of the atmospheric pressure heating furnace, the outlet temperature of the vacuum heating furnace, the fuel gas pressure of the vacuum heating furnace, the fuel gas flow rate of the vacuum heating furnace, the flue gas oxygen content of the vacuum heating furnace, and the furnace tube surface temperature of the vacuum heating furnace. In some embodiments, the second target adjustment parameter includes, but is not limited to, the opening degree of the fuel main valve of the atmospheric pressure heating furnace and the opening degree of the fuel main valve of the vacuum heating furnace. In some embodiments, the set temperature curve includes, but is not limited to, a first set temperature curve, a second set temperature curve, and a third set temperature curve. By inputting multiple set temperature curves into the temperature control model, the heating furnace is gradually heated (e.g., according to the heating rate of the set temperature curve) to different temperatures, thereby achieving intelligent control of the dehydration and hot-tightening process. For a detailed explanation of this part, please refer to the corresponding embodiments below, which will not be repeated here.
[0026] Step S4: Establish an atmospheric pressure tower side-stream delivery system through an open atmospheric pressure control algorithm, and intelligently regulate the open side-stream of the atmospheric pressure tower.
[0027] In some embodiments, opening to atmospheric pressure includes, but is not limited to, temperature control of the atmospheric pressure furnace, control of steam output from the atmospheric pressure tower, and alerts for opening the side feed line from the atmospheric pressure tower. In this embodiment, intelligent regulation of the atmospheric pressure tower side feed line delivery system is achieved by setting an atmospheric pressure control algorithm. For a detailed explanation of this part, please refer to the corresponding embodiments below, which will not be repeated here.
[0028] Step S5: Establish a side-line power supply system for the pressure reducing tower using an open / closed pressure control algorithm, and intelligently regulate the open side-line power supply of the pressure reducing tower.
[0029] In some embodiments, pressure reduction and opening include, but are not limited to, temperature control of the pressure reducing furnace and indication of side-line opening of the pressure reducing tower. In this embodiment, intelligent regulation of the pressure reducing tower side-line delivery system is achieved by setting a pressure reduction and opening control algorithm. For a detailed explanation of this part, please refer to the corresponding embodiments below, which will not be repeated here.
[0030] In some embodiments, determining the target cold cycle sub-model and target regulating valve corresponding to the cold cycle related parameters from the cold cycle model according to the cold cycle stage to which the cold cycle related parameters belong includes: querying a database to obtain a cold cycle sub-model and regulating valve that have a mapping relationship with the cold cycle related parameters, using the cold cycle sub-model as the target cold cycle sub-model for the cold cycle related parameters, and using the regulating valve as the target regulating valve for the cold cycle related parameters. For example, the system presets a database as shown in Table 1. After activating the start button of the preprocessing device, each cold cycle related parameter is acquired through the corresponding sensor, and the cold cycle related parameters are input into the corresponding target cold cycle sub-model based on different cold cycle related parameters, so that the first target regulating parameter of the target regulating valve is output through the target cold cycle sub-model. In some embodiments, the first target regulating parameter typically includes the target regulating valve opening degree. In this embodiment, during the cold cycle process, corresponding intelligent control is performed on different cold cycle stages through different cold cycle sub-models, which can ensure the safe start-up and operation of each cold cycle stage. In some embodiments, each cold cycle sub-model adopts a machine learning model architecture. For example, the electrostatic desalting oil filling model employs an LSTM (Long Short-Term Memory) + attention mechanism model architecture. The LSTM layer captures temporal dependencies, the attention mechanism layer dynamically weights key features, the fully connected layer performs feature fusion to generate control strategies, and the output layer outputs multi-valve coordinated commands. The model is trained using a large amount of historical training data corresponding to the electrostatic desalting oil filling model (e.g., historical cold cycle related parameters and corresponding optimal control valve openings) to obtain the final electrostatic desalting oil filling model. Specifically, the electrostatic desalting oil filling model includes an input layer, an LSTM layer, an attention layer, a fully connected layer, and an output layer, employing a mean squared error loss function and the Adam optimizer. In some embodiments, the cold cycle related parameters input to the electrostatic desalting oil filling model and the control valve opening information output are normalized. The primary distillation column oil filling model uses a gradient boosting tree model architecture to smoothly control the bottom liquid level of the primary distillation column, preventing cavitation or overflow. The crude distillation tower loading model is obtained by training a large amount of historical training data (e.g., historical cold cycle parameters and corresponding optimal control valve openings) on the crude distillation tower model. For example, the model structure specifically includes the number of trees (e.g., 500-1000 trees), tree depth (e.g., maximum depth of 6-8 to prevent overfitting and maintain model simplicity), learning rate (e.g., 0.05-0.1), and subsampling ratio (e.g., 0.8, each tree uses 80% of the samples and features for training to increase robustness). The squared error is used as the objective function, and optimization is performed using built-in algorithms such as greedy algorithms. In some embodiments, the input cold cycle parameters and output control valve opening information of the crude distillation tower loading model can also be normalized separately.The atmospheric pressure tower oil filling model employs a multi-task learning neural network (MTL). The system controls four feed valves to ensure balanced flow in each path while maintaining a stable total load, preventing flow deviation and localized overheating in the furnace tubes. During model training, a large amount of historical training data corresponding to the atmospheric pressure tower oil filling model (e.g., historical cold cycle parameters and corresponding optimal control valve openings) is used to train the model. For example, the model architecture includes a shared bottom layer (e.g., 3-4 fully connected layers (e.g., 128 -> 256 -> 128), using the ReLU activation function. All input features first enter this shared network), task-specific heads (e.g., smaller task-specific networks (e.g., 64 -> 32 -> 1) are built for each of the four valves. Each task head takes the output of the shared layer as input), and an output layer (e.g., four output neurons, typically using the Sigmoid function to ensure the output is between 0 and 1). The weighted sum of the MSEs of each task is used as the loss function, and the Adam optimizer is employed. In some embodiments, the input cold cycle parameters and output control valve opening information of the atmospheric pressure tower oil filling model can also be normalized separately. The vacuum tower oil filling model employs reinforcement learning (TD3). The core idea of reinforcement learning is to learn a policy to maximize accumulated future rewards. The TD3 algorithm can handle control problems with delayed rewards well. Similarly, the model is trained using a large amount of historical training data from the vacuum tower oil filling model to obtain the vacuum tower oil filling model. For example, the model architecture of the vacuum tower oil filling model includes an Actor network and a Critic network, using MSE and policy gradient as loss functions, and employing the Adam optimizer. In some embodiments, the input cold cycle parameters and output control valve opening information of the vacuum tower oil filling model can also be normalized separately.
[0031] Table 1 Database
[0032] Of course, those skilled in the art will understand that the database in Table 1 described above is only an example. Other existing or future cold cycle-related parameters and their correspondence with the regulating valve, if applicable to this embodiment, are also within the scope of protection of this application and are incorporated herein by reference.
[0033] In some embodiments, the set temperature curve includes a first set temperature curve, a second set temperature curve, and a third set temperature curve. Inputting the dehydration thermally relevant parameters and the set temperature curve into a temperature control model to obtain a second target adjustment parameter for the furnace fuel main valve includes: inputting the dehydration thermally relevant parameters and the first set temperature curve into the temperature control model; the temperature control model, by processing the dehydration thermally relevant parameters and the first set temperature curve, outputting a second target adjustment parameter for the furnace fuel main valve to control the furnace temperature to rise to a first target temperature corresponding to the first set temperature curve; when the furnace temperature remains constant at the first target temperature for a first target time and meets a first condition, the dehydration thermally relevant parameters and the second set temperature curve are... A temperature control model is input to the heating furnace. This model processes the dehydration thermally relevant parameters and the second set temperature curve to output a second target adjustment parameter for the furnace fuel main valve. This parameter controls the furnace temperature to rise to the second target temperature corresponding to the second set temperature curve. When the furnace temperature remains constant at the second target temperature for a second target time and meets a second condition, the dehydration thermally relevant parameters and the third set temperature curve are input to the temperature control model. This model processes these parameters and outputs a second target adjustment parameter for the furnace fuel main valve to control the furnace temperature to rise to the third target temperature corresponding to the third set temperature curve and remain constant at the third target temperature for a third target time. In some embodiments, the first set temperature curve includes a first target temperature (e.g., 90°C), the second set temperature curve includes a second target temperature (e.g., 150°C), and the third set temperature curve includes a third target temperature (e.g., 250°C). In this embodiment, by inputting different set temperature curves and current dehydration heat-related parameters into the temperature control model, and outputting a second target adjustment parameter (e.g., fuel valve opening) for the furnace fuel valve through the temperature control model, a phased, gradual temperature rise and constant temperature control of the furnace is achieved. In some embodiments, the first condition includes, but is not limited to, the water content of the crude oil pipeline at the outlet of the electric desalting tank being less than or equal to the target water content, and the second condition includes, but is not limited to, the top vapor dew point of the distillate pipeline at the top of the primary distillation column being less than or equal to the target dew point temperature. Of course, those skilled in the art will understand that the above-mentioned first and second conditions are merely examples, and other existing or future first and second conditions that are applicable to this application are also within the scope of protection of this application and are incorporated herein by reference. For example, the first and second conditions can be adjusted based on different specific processes.
[0034] In some embodiments, an atmospheric pressure control algorithm is used to establish a side-stream delivery system to achieve intelligent control of the atmospheric pressure tower's side-stream. This includes: inputting atmospheric pressure-related parameters and the atmospheric pressure set temperature curve into the atmospheric pressure temperature control model; outputting third target adjustment parameters for the four fuel branch valves and flue gas dampers of the atmospheric pressure furnace through the atmospheric pressure temperature control model, so that the atmospheric pressure furnace temperature gradually increases to the second target temperature at a set rate; determining the opening command for the steam regulating valve based on the steam feed rate related parameters and the steam feed rate control algorithm, and adjusting the steam regulating valve based on the opening command; and monitoring the atmospheric pressure tower side-stream establishment-related parameters in real time. When the atmospheric pressure tower side-stream establishment-related parameters meet the establishment conditions, generating establishment prompt information for the corresponding atmospheric pressure tower side-stream. In some embodiments, before the atmospheric pressure control algorithm runs, parameters such as the heating rate of the heating furnace (e.g., the atmospheric pressure set temperature curve), the target heating value (e.g., the second target temperature included in the atmospheric pressure set temperature curve), the steam start rate, the target value of stripping steam, the opening degree of the atmospheric top circulation regulating valve, the opening degree of the atmospheric first-stage regulating valve, the opening degree of the atmospheric second-stage regulating valve, the opening degree of the atmospheric first-stage regulating valve, the opening degree of the atmospheric second-stage regulating valve, the opening degree of the atmospheric third-stage regulating valve, and the opening degree of the atmospheric fourth-stage regulating valve are set in advance. In some embodiments, the atmospheric pressure related parameters include, but are not limited to, the temperatures of the four branches of the atmospheric pressure furnace, the flow balance coefficient of the four branches, and the temperature of the atmospheric pressure furnace outlet main pipe. The third target regulation parameters include, but are not limited to, the opening degree of the four fuel branch valves of the atmospheric pressure furnace and the opening degree of the flue gas damper of the atmospheric pressure furnace. In some embodiments, the atmospheric pressure furnace is also called an atmospheric pressure heating furnace. When the outlet temperature of the atmospheric pressure furnace reaches 320°C, a switch to open circuit prompt message is generated to prompt the user to switch to open circuit. In some embodiments, the steam boosting related parameters include, but are not limited to, the target steam flow rate and the current steam flow rate. For relevant explanations of the steam boosting control algorithm, please refer to the corresponding embodiments below, which will not be repeated here. In some embodiments, parameters related to the establishment of the atmospheric distillation column sideline are monitored in real time so that when the relevant parameters meet the establishment conditions, a prompt message for the establishment of the corresponding atmospheric distillation column sideline is generated. For example, the atmospheric distillation column sideline opening monitoring and control system establishes the atmospheric distillation column circulation reflux, atmospheric distillation column intermediate reflux, atmospheric distillation column intermediate reflux, atmospheric distillation column external discharge, atmospheric distillation column external discharge, atmospheric distillation column external discharge, and atmospheric distillation column external discharge in sequence after the conditions such as the heater temperature, atmospheric distillation column top liquid level, atmospheric distillation column 1 liquid level, atmospheric distillation column 2 liquid level, and atmospheric distillation column 3 liquid level are met. After the atmospheric distillation column sideline opening monitoring and control system meets the conditions for establishing circulation or opening a sideline, it promptly reminds the activation of the atmospheric distillation column top, atmospheric distillation column 1 intermediate, atmospheric distillation column 2 intermediate, atmospheric distillation column 1, atmospheric distillation column 2, atmospheric distillation column 3, and atmospheric distillation column 4 pumps to ensure that the atmospheric distillation column establishes the intermediate circuit and sideline discharge in an orderly manner. In some embodiments, the establishment conditions can be set differently according to different processes. For example, if the liquid level of the first line of the normal flow line is higher than 80%, the valve for the first line of the normal flow line is opened; if the liquid level of the second line of the normal flow line is higher than 80%, the valve for the second line of the normal flow line is opened; if the liquid level of the third line of the normal flow line is higher than 80%, the valve for the third line of the normal flow line is opened; if the liquid level of the second line of the normal flow line is higher than 60%, the valve for the first line of the normal flow line is opened; if the liquid level of the third line of the normal flow line is higher than 60%, the valve for the second line of the normal flow line is opened, etc.
[0035] In some embodiments, determining the opening command of the steam regulating valve based on steam rise parameters and a steam rise control algorithm includes: ,in The proportional control gain coefficient is the target flow rate. The target steam flow rate at any given time. Target traffic The target steam flow rate is the final value, and the rate of increase is the preset rate at which the steam flow rate increases. The cumulative time, in seconds, is the time from the start of the steam ramp. In some embodiments, the steam regulating valve is located in the atmospheric pressure tower. In some embodiments, a steam ramp curve (e.g., a steam ramp change curve over time) can be preset to determine the target steam flow rate at time t. Based on the steam ramp curve, the steam ramp is gradually increased to the target value at a set rate. In some embodiments, the proportional control gain coefficient includes, but is not limited to, 0.05. In this embodiment, a soft-start function can reduce the water hammer effect. In some embodiments, r is set to 1.5 to achieve nonlinear smooth acceleration, avoiding the sluggishness of linear growth while preventing the recklessness of squared growth. In some embodiments, the effect of the soft-start function on suppressing the water hammer effect can be tested in the field.
[0036] In some embodiments, a pressure reducing tower side-line delivery system is established through an opening and closing pressure reducing control algorithm to achieve intelligent control of the pressure reducing tower side-line. This includes: inputting relevant opening and closing pressure reducing parameters and the pressure reducing set temperature curve into the pressure reducing temperature control model; outputting fourth target adjustment parameters for the four fuel branch valves and flue gas dampers of the pressure reducing furnace through the pressure reducing temperature control model, so that the pressure reducing furnace temperature gradually increases to the third target temperature at a set rate; prompting a vacuuming operation for the pressure reducing tower when the heater outlet temperature reaches the target temperature threshold; and monitoring relevant parameters for establishing the pressure reducing tower side-line in real time, generating a prompt message for establishing the corresponding pressure reducing tower side-line when the relevant parameters for establishing the pressure reducing tower side-line meet the establishment conditions. In some embodiments, before the opening and closing pressure reducing operation, parameters such as the heater heating rate (e.g., the pressure reducing set temperature curve), the heating target value (e.g., the third target temperature), the steam opening rate, the opening degree of the top circulation regulating valve, the opening degree of the first regulating valve, the opening degree of the second regulating valve, the opening degree of the first regulating valve, the opening degree of the second regulating valve, the opening degree of the third regulating valve, the opening degree of the fourth regulating valve, and the opening degree of the fifth regulating valve are set in advance. In some embodiments, the parameters related to pressure reduction and depressurization include, but are not limited to, the temperatures of the four branches of the pressure reducing furnace, the flow balance coefficient of the four branches, and the temperature of the main outlet pipe of the pressure reducing furnace. The fourth target adjustment parameters include, but are not limited to, the opening degree of the four fuel branch valves of the pressure reducing furnace and the opening degree of the flue gas damper of the pressure reducing furnace. In some embodiments, the parameters related to the establishment of the pressure reducing tower side line are monitored in real time so that when the parameters related to the establishment of the pressure reducing tower side line meet the establishment conditions, a prompt message for the establishment of the corresponding pressure reducing tower side line is generated. For example, the atmospheric and vacuum distillation tower establishes the middle section reflux, and the side line opening monitoring and control sequentially establishes the top reflux, top external supply, opens the first pressure reducing line, establishes the middle section of the first pressure reducing line, opens the second pressure reducing line, establishes the middle section of the second pressure reducing line, opens the third pressure reducing line, opens the fourth pressure reducing line, and opens the fifth pressure reducing line. After the pressure reducing tower side line opening monitoring and control meets the conditions for establishing the circulation or opening the side line, a timely reminder is given to start the pumps of the top pressure reducing line, the middle section of the first pressure reducing line, the middle section of the second pressure reducing line, the external supply of the top pressure reducing line, the second pressure reducing line, the third pressure reducing line, the fourth pressure reducing line, and the fifth pressure reducing line, ensuring that the pressure reducing tower establishes the middle section loop and the side line external supply in an orderly manner. In some embodiments, the conditions for establishing the pressure control system can be set differently depending on the process. For example, if the level of the second reducing line is above 85%, the valve for supplying fuel to the second reducing line is opened; if the level of the third reducing line is above 75%, the valve for supplying fuel to the third reducing line is opened; if the level of the fourth reducing line is above 65%, the valve for supplying fuel to the fourth reducing line is opened; if the level of the second reducing line is above 20%, the valve for the first reducing line is opened; if the level of the third reducing line is above 30%, the valve for the second reducing line is opened, and so on. In some embodiments, both the atmospheric pressure temperature control model and the depressurization temperature control model adopt an LSTM+reinforcement learning model architecture. For example, LSTM complex state perception and dynamic prediction learns the thermal inertia of the heating furnace (e.g., the delay from valve action to temperature response), and predicts the temperature trend in the short term based on the current state and historical sequences. Reinforcement learning acts as the decision-maker, receiving the state (e.g., current temperature, predicted temperature, system inertial characteristics, etc.) enhanced by LSTM, and outputs the optimal fuel valve opening command.In some embodiments, the total reward = W1 * tracking reward + W2 * stability reward + W3 * safety reward - penalty. For example, the tracking reward encourages the actual temperature to closely follow the set temperature rise curve, the stability reward penalizes drastic fluctuations in temperature or valves to pursue smooth operation, and the safety reward includes imposing a large penalty when current key parameters (e.g., furnace tube branch temperature difference, over-temperature) approach the safety limit to ensure absolute safety. The penalty is used to avoid excessive valve operation and reduce equipment wear.
[0037] In some embodiments, the method further includes step S15 (not shown), in which relevant operating parameters are monitored in real time during the intelligent start-up process, and corresponding alarm information is generated when the relevant operating parameters meet abnormal conditions. For example, various relevant operating parameters are monitored in real time throughout the entire start-up process. In some embodiments, relevant operating parameters include, but are not limited to, the flow rates of the atmospheric and pressure furnaces, the four feed flow rates of the atmospheric furnace, and the two feed flow rates of the vacuum furnace. When the relevant operating parameters are not within the corresponding parameter range, corresponding alarm information is generated to promptly remind the user and ensure safe start-up.
[0038] In the cold cycle process, the above technical solution determines the target cold cycle sub-model and target regulating valve corresponding to the relevant cold cycle parameters from the cold cycle model. The relevant cold cycle parameters are input into the corresponding target cold cycle sub-model, and the first target regulating parameter of the target regulating valve is output through the target cold cycle sub-model, thus achieving intelligent control of the cold cycle process. In the dehydration and hot tightening process, the relevant dehydration and hot tightening parameters and the first set temperature curve are input into the temperature control model. The temperature control model processes the relevant dehydration and hot tightening parameters and the first set temperature curve, and outputs a second target regulating parameter for the main fuel valve of the heating furnace, causing the heating furnace temperature to rise and remain constant at a set temperature rate, thus achieving intelligent control of the dehydration and hot tightening process. In the atmospheric pressure and depressurization processes, atmospheric pressure tower side-line delivery systems and depressurization tower side-line delivery systems are established through atmospheric pressure control algorithms and depressurization control algorithms, respectively, achieving intelligent control of the atmospheric pressure tower side-line and depressurization tower side-line. The above technical solution achieves intelligent start-up of the pretreatment unit, reducing manual intervention; it also improves start-up accuracy and avoids the impact of human error on the start-up process.
[0039] like Figure 2 As shown in the figure, an intelligent start-up system for a preprocessing device based on a multi-model algorithm provided by an embodiment of the present invention includes: The data acquisition module is used to acquire cold cycle related parameters and determine the corresponding target cold cycle sub-model and target regulating valve according to the cold cycle stage to which the cold cycle related parameters belong. The cold cycle control module is used to input the cold cycle related parameters into the target cold cycle sub-model, obtain the first target adjustment parameter of the target regulating valve, and adjust the target regulating valve according to the first target adjustment parameter; The temperature control module is used to input the dehydration heat-related parameters and the set temperature curve into the temperature control model when the cold cycle process meets the closed-loop cycle conditions, to obtain the second target adjustment parameter of the main fuel valve of the heating furnace, and to control the heating furnace to gradually heat up and maintain a constant temperature. The atmospheric pressure tower control module is used to establish an atmospheric pressure tower side-stream delivery system through an open atmospheric pressure control algorithm, and to intelligently regulate the open side-stream of the atmospheric pressure tower. The pressure reducing tower control module is used to establish a side-line external transmission system of the pressure reducing tower through the pressure reducing control algorithm, and to intelligently regulate the opening of the side-line of the pressure reducing tower.
[0040] An electronic device provided by an embodiment of the present invention includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of any of the above-described intelligent start-up methods for preprocessing devices based on multi-model algorithms.
[0041] Specifically, the aforementioned memory and processor can be general-purpose memory and processor, without any specific limitations. When the processor runs the computer program stored in the memory, it can execute the aforementioned intelligent start-up method for the preprocessing device based on the multi-model algorithm.
[0042] Those skilled in the art will understand that the structure of the electronic device does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or combine some components, or split some components, or have different component arrangements.
[0043] In some embodiments, the electronic device may further include a touchscreen for displaying a graphical user interface (e.g., an application launch screen) and receiving user actions on the graphical user interface (e.g., launching an application). Specifically, the touchscreen may include a display panel and a touch panel. The display panel may be configured as an LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode), or similar type. The touch panel can collect user touch or non-touch operations on or near it and generate pre-set operation commands, such as user actions using fingers, styluses, or any suitable object or accessory on or near the touch panel. Additionally, the touch panel may include a touch detection device and a touch controller. The touch detection device detects the user's touch position and posture, and detects the signals generated by the touch operation, transmitting the signals to the touch controller. The touch controller receives touch information from the touch detection device, converts it into information that the processor can process, sends it to the processor, and can also receive and execute commands from the processor. Furthermore, touch panels can be implemented using various types of sensors, including resistive, capacitive, infrared, and surface acoustic wave sensors, as well as any future technologies. Moreover, the touch panel can cover the display panel. Users can operate on or near the touch panel, which is covered by the graphical user interface displayed on the display panel. After detecting the operation on or near the touch panel, the touch panel transmits it to the processor to determine the user input. The processor then responds to the user input by providing corresponding visual output on the display panel. Additionally, the touch panel and display panel can be implemented as two separate components or integrated together.
[0044] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it includes the steps of the methods described in the embodiments. The storage medium includes, but is not limited to, ROM, RAM, magnetic disks, optical disks, etc.
[0045] Corresponding to the above application startup method, this embodiment of the invention also provides a storage medium storing a computer program, which, when run by a processor, executes the steps of any of the above-described intelligent start-up methods for preprocessing devices based on multi-model algorithms.
[0046] The application launch device provided in this application embodiment can be specific hardware on the device or software or firmware installed on the device. The device provided in this application embodiment has the same implementation principle and technical effects as the foregoing method embodiments. For the sake of brevity, any parts not mentioned in the device embodiment can be referred to the corresponding content in the foregoing method embodiments. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can all be referred to the corresponding processes in the above method embodiments, and will not be repeated here.
[0047] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0048] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and there may be other division methods in actual implementation. Furthermore, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interface, and the indirect coupling or communication connection of the apparatus or modules may be electrical, mechanical, or other forms.
[0049] The modules described as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.
[0050] In addition, the functional modules in the embodiments provided in this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.
[0051] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0052] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0053] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.
Claims
1. A method for intelligent start-up of a pre-treatment device based on a multi-model algorithm, characterized in that, The method comprises the following steps: obtaining a cold cycle related parameter, determining a corresponding target cold cycle sub-model and a target regulating valve according to a cold cycle stage to which the cold cycle related parameter belongs; inputting the cold cycle related parameter into the target cold cycle sub-model to obtain a first target regulating parameter of the target regulating valve, and adjusting the target regulating valve according to the first target regulating parameter; when a closed cycle condition is met during a cold cycle process, inputting a dehydration heat related parameter and a set temperature curve into a temperature control model to obtain a second target regulating parameter of a heating furnace fuel total valve, and controlling the heating furnace to gradually increase and maintain temperature according to the second target regulating parameter; establishing an atmospheric tower side line external delivery system through an atmospheric pressure control algorithm to intelligently control the opening of the side line of the atmospheric tower; establishing a vacuum tower side line external delivery system through a pressure reduction control algorithm to intelligently control the opening of the side line of the vacuum tower.
2. The pre-treatment plant intelligent start-up method based on a multi-model algorithm according to claim 1, characterized in that, The method comprises the following steps: According to the cold cycle related parameter, the cold cycle sub-model and the regulating valve which have a mapping relationship with the cold cycle related parameter are obtained from the database, and the cold cycle sub-model is used as the target cold cycle sub-model of the cold cycle related parameter, and the regulating valve is used as the target regulating valve of the cold cycle related parameter.
3. The pre-treatment plant intelligent start-up method based on a multi-model algorithm according to claim 1, characterized in that, The set temperature curve comprises a first set temperature curve, a second set temperature curve and a third set temperature curve; the dehydration heat related parameter and the set temperature curve are inputted into the temperature control model to obtain the second target regulating parameter of the heating furnace fuel total valve, which comprises the following steps: The dehydration heat related parameter and the first set temperature curve are inputted into the temperature control model, and the temperature control model processes the dehydration heat related parameter and the first set temperature curve to output the second target regulating parameter of the heating furnace fuel total valve, so as to control the temperature of the heating furnace to rise to a first target temperature corresponding to the first set temperature curve; When the temperature of the heating furnace is maintained at the first target temperature for a first target time and a first condition is met, the dehydration heat related parameter and the second set temperature curve are inputted into the temperature control model, and the temperature control model processes the dehydration heat related parameter and the second set temperature curve to output the second target regulating parameter of the heating furnace fuel total valve, so as to control the temperature of the heating furnace to rise to a second target temperature corresponding to the second set temperature curve, and the first condition comprises that the water content of the crude oil pipeline at the outlet of the electric desalting tank is less than or equal to a target water content. When the heating furnace temperature is constant at the second target temperature for a second target time and a second condition is met, the dehydration thermal correlation parameter and the third set temperature curve are input into the temperature control model, the temperature control model outputs a second target adjustment parameter of the heating furnace fuel total valve by processing the dehydration thermal correlation parameter and the third set temperature curve, so as to control the heating furnace temperature to rise to a third target temperature corresponding to the third set temperature curve and be constant at the third target temperature for a third target time, and the second condition includes that the overhead vapor dew point of the initial distillation column overhead pipeline is less than or equal to a target dew point temperature.
4. The pre-treatment plant intelligent start-up method based on a multi-model algorithm according to claim 1, characterized in that, The open atmospheric pressure control algorithm is used to establish the atmospheric pressure column side line external delivery system, and the atmospheric pressure column side line is intelligently regulated and controlled, including: The open atmospheric pressure correlation parameter and the atmospheric pressure set temperature curve are input into the atmospheric pressure temperature control model, and the third target adjustment parameter of the atmospheric pressure furnace fuel branch valve and the atmospheric pressure furnace flue gas damper is output by the atmospheric pressure temperature control model, so that the atmospheric pressure furnace temperature gradually rises to the atmospheric pressure target temperature at a set rate; The opening degree instruction of the steam adjusting valve is determined according to the steam extraction amount correlation parameter and the steam extraction amount control algorithm, and the steam adjusting valve is adjusted based on the opening degree instruction; The atmospheric pressure column side line establishment correlation parameter is monitored in real time, and when the atmospheric pressure column side line establishment correlation parameter meets the establishment condition, an establishment prompt information about the corresponding atmospheric pressure column side line is generated.
5. The pre-treatment plant intelligent start-up method based on a multi-model algorithm according to claim 4, characterized in that, The opening degree instruction of the steam adjusting valve is determined according to the steam extraction amount correlation parameter and the steam extraction amount control algorithm, including: wherein is a proportional control gain coefficient, target flow is the target steam flow at time target flow rate is the final target value of the steam flow rate, and the raising rate is a preset rate of the steam flow rate increase, is the time accumulated from the start of the raising of the steam flow rate.
6. The pre-treatment plant intelligent start-up method based on a multi-model algorithm according to claim 1, characterized in that, The open pressure reduction control algorithm is used to establish the vacuum column side line external delivery system, and the vacuum column side line is intelligently regulated and controlled, including: The open pressure reduction correlation parameter and the pressure reduction set temperature curve are input into the pressure reduction temperature control model, and the fourth target adjustment parameter of the pressure reduction furnace fuel branch valve and the pressure reduction furnace flue gas damper is output by the pressure reduction temperature control model, so that the pressure reduction furnace temperature gradually rises to the pressure reduction target temperature at a set rate; When the heating furnace outlet temperature reaches a target temperature threshold, a vacuum operation on the pressure reduction column is prompted; The pressure reduction column side line establishment correlation parameter is monitored in real time, and when the pressure reduction column side line establishment correlation parameter meets the establishment condition, an establishment prompt information about the corresponding pressure reduction column side line is generated.
7. The pre-treatment plant intelligent start-up method based on a multi-model algorithm according to any one of claims 1 to 6, characterized in that, Including the following steps: In the intelligent start-up process, the related operating parameters are monitored in real time, and when the related operating parameters meet the abnormal condition, corresponding alarm information is generated.
8. A pre-treatment plant intelligent start-up system based on a multiple model algorithm, characterized in that, Including: A data acquisition module is configured to acquire cold cycle correlation parameters, determine a target cold cycle sub-model and a target adjusting valve according to a cold cycle phase to which the cold cycle correlation parameters belong; and A cold cycle control module is configured to input the cold cycle correlation parameters into the target cold cycle sub-model, obtain a first target adjustment parameter of the target adjusting valve, and adjust the target adjusting valve according to the first target adjustment parameter. The temperature control module is used for inputting the dehydration heat closely related parameters and the set temperature curve into the temperature control model when the cold cycle process meets the closed cycle condition, obtaining the second target adjustment parameter of the heating furnace fuel total valve, and controlling the heating furnace to gradually increase and keep constant temperature according to the second target adjustment parameter; The atmospheric tower control module is used for establishing the atmospheric tower side line external delivery system through the atmospheric tower control algorithm, and intelligently controls the atmospheric tower side line. The vacuum tower control module is used for establishing the vacuum tower side line external delivery system through the vacuum tower control algorithm, and intelligently controls the vacuum tower side line.
9. An electronic device, comprising: The electronic device comprises a processor, a memory and a bus, the memory stores machine readable instructions executable by the processor, the processor and the memory communicate through the bus when the electronic device runs, and the processor executes the machine readable instructions to execute the steps of the intelligent starting method of the pretreatment device based on the multi-model algorithm.
10. A storage medium, characterized by The storage medium stores a computer program, and the computer program is executed by the processor to execute the steps of the intelligent starting method of the pretreatment device based on the multi-model algorithm.