A control system and method for continuous production line of polymerization reaction

Abstract

The present application discloses a control system and method for a continuous production line of a polymerization reaction, which relates to the field of chemical technology. The control system and method are as follows: an exothermic equation for the polycondensation reaction is established according to a kinetic model of the reaction mechanism of the polyester polycondensation reaction and the reaction rate of the polycondensation product; a cooling heat equation is established according to a cooling system; a heat exchange equation between the polycondensation reactor and the environment is established, and a heat balance differential equation in the polycondensation reactor is obtained; the reaction temperature and component concentration in the polycondensation reactor are defined as the state of a reinforcement learning model; the cooling medium temperature is defined as the action of the reinforcement learning model; a state transfer function is established according to the heat balance differential equation; a hierarchical dynamic reward function is designed; and a reinforcement learning model is constructed; a DDPG algorithm is selected to train the reinforcement learning model; an optimal temperature control strategy is obtained using the trained reinforcement learning model to optimize the control of the polyester polycondensation reaction process; and the temperature stability in the polyester polycondensation reaction process is improved.

Description

Technical Field

[0001] The present application relates to the field of chemical technology, and in particular to a control system and method for a continuous production line of a polymerization reaction. Background Art

[0002] The polymerization production line is an industrialized continuous production system for producing high molecular polymers. It converts monomer raw materials into polymer products with specific properties through chemical reactions. In the production process of saturated polyester resins for powder coatings, the continuity and stability of the polymerization reaction are key factors to ensure product quality and production efficiency. Saturated polyester resins for powder coatings are an important industrial material, and its production process involves complex chemical reactions and physical changes.

[0003] The Chinese patent with publication number CN118146490A discloses a method for continuously and efficiently preparing high molecular weight polyester materials, comprising the following steps: (1) adding a polymerization monomer, an excess initiator and an excess catalyst into a reactor to carry out ring-opening polymerization until the reaction equilibrium is reached to obtain a high viscosity liquid; the molar ratio of the polymerization monomer to the initiator is 50:1 to 200:1, and the molar ratio of the polymerization monomer to the catalyst is 50:1 to 4000:1; (2) feeding the high viscosity liquid into a twin-screw extruder at a set speed through a melt metering pump, feeding the polymerization monomer melted in advance in the kettle from a feeding port on the twin-screw extruder through a metering pump, controlling the speed of the screw, and under the push of the screw, the material completes the polymerization reaction in the extruder, and is extruded through a die head and then cold cut into pellets. The invention first prepares a high-viscosity liquid feed containing excess catalyst and initiator, which can directly form a stable flow in an extruder, and then simply adjusts the feed speed of the high-viscosity liquid feed and the polymerization monomer and the material addition position to prepare polylactic acid with different molecular weights.

[0004] Polyester polycondensation reactions are highly sensitive to temperature. The existing production methods have large temperature fluctuations, resulting in uneven reactions. Too high a temperature can lead to darker product color and potential safety hazards, while too low a temperature can lead to a slow reaction rate, affecting production efficiency. Summary of the invention

[0005] The present application aims to solve one of the technical problems in the related art at least to a certain extent. To this end, one purpose of the present application is to propose a control system and method for a continuous production line of a polymerization reaction, which improves the stability of the reaction temperature during the polyester polycondensation reaction.

[0006] One aspect of the present application provides a method for controlling a continuous production line of a polymerization reaction, comprising:

[0007] Step S100: establishing a polycondensation reaction exothermic equation according to the reaction mechanism of the polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, establishing a cooling heat equation according to the cooling system, establishing a heat exchange equation between the polycondensation reactor and the environment, and obtaining a heat balance differential equation in the polycondensation reactor;

[0008] Step S200: defining the reaction temperature and component concentration in the polycondensation reactor as the state of the reinforcement learning model, defining the cooling medium temperature as the action of the reinforcement learning model, establishing a state transfer function according to the heat balance differential equation, designing a hierarchical dynamic reward function, and constructing a reinforcement learning model;

[0009] Step S300: Select the DDPG algorithm to train the reinforcement learning model to obtain a trained reinforcement learning model;

[0010] Step S400: Utilize the trained reinforcement learning model to obtain the optimal temperature control strategy and optimize the control of the polyester polycondensation reaction process.

[0011] Specifically, the method of establishing the heat balance differential equation of the polycondensation reaction according to the reaction mechanism of the polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, establishing the cooling heat equation according to the cooling system, and establishing the heat exchange equation between the polycondensation reactor and the environment is as follows:

[0012] Step S110: According to the reaction mechanism of polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, the reaction rate rP of the polycondensation product, the working volume V of the polycondensation reactor and the reaction enthalpy change Multiply them to get the reaction exothermic rate of the polyester polycondensation reaction The exothermic equation of the polycondensation reaction;

[0013] The reaction mechanism of the polyester polycondensation reaction is simplified to a reversible polycondensation reaction of a dibasic acid monomer and a diol monomer, and the reaction formula is: , wherein A is a dibasic acid monomer, B is a diol monomer, P is a polycondensation product, and W is water;

[0014] The kinetic model of the polyester polycondensation reaction is: ,in, is the forward reaction rate constant, represents the reverse reaction rate constant, rP represents the reaction rate of the polycondensation product, , represent the concentration of dibasic acid monomer, diol monomer, polycondensation product, and water, respectively;

[0015] The exothermic equation of the polycondensation reaction is: ,in, is the reaction rate of the polycondensation product, is the working volume of the polycondensation reactor, is the reaction enthalpy change;

[0016] Step S120: Calculate the cooling heat according to the difference between the reaction temperature of the cooling system and the temperature of the cooling medium, multiplied by the total heat transfer coefficient and the heat transfer area of ​​the polycondensation reactor. The cooling heat equation;

[0017] The cooling heat equation is: ,in, is the overall heat transfer coefficient, is the heat transfer area of ​​the polycondensation reactor, T is the reaction temperature, is the cooling medium temperature;

[0018] Step S130: According to Newton's cooling law, the temperature difference between the reaction temperature and the ambient temperature is calculated, and the temperature difference is multiplied by the convection heat transfer coefficient and the heat transfer area of ​​the polycondensation reactor to obtain the heat transfer coefficient between the polycondensation reactor and the environment. The heat exchange equation of

[0019] The heat exchange equation is: , where h is the convective heat transfer coefficient and Ta is the ambient temperature;

[0020] Step S140: establishing a heat balance differential equation in the polycondensation reactor according to the polycondensation reaction heat release equation, the cooling heat equation and the heat exchange equation;

[0021] The heat balance differential equation in the polycondensation reactor is: ,in, is the density of the reactants, is the specific heat capacity, is the change in reaction temperature within each time step t.

[0022] Specifically, the specific method of obtaining the reaction temperature and component concentration in the polycondensation reactor as the state, the cooling medium flow and cooling medium temperature of the cooling system as the action, and the heat balance differential equation as the environment is:

[0023] Step S210: defining a state variable set S of a reinforcement learning model, wherein the state s(t) in the state variable set S includes: a reaction temperature T in the polycondensation reactor, a component concentration CX, wherein the component concentration CX includes a dibasic acid monomer concentration CA, a diol monomer concentration CB, a polycondensation product concentration CP, and a water concentration CW, wherein t in s(t) represents a time step;

[0024] Step S220: defining an action variable set A of a reinforcement learning model, wherein the action a(t) in the action variable set A is the temperature of the cooling medium;

[0025] Step S230: Establishing a state transfer function based on the heat balance differential equation , represents the state at the current time step Next action After that, it transfers to the new state at the next time step ;

[0026] The state at the current time step Next action After that, it transfers to the new state at the next time step The specific method is:

[0027] Step S231: Substitute the reaction temperature T(t) and component concentration CX(t) of the current time step in state s(t) into the heat balance differential equation, and calculate the heat balance according to the action Adjust the cooling medium temperature and update the cooling heat ;

[0028] Step S232: Calculate the reaction temperature T(t+1) of the next time step according to the heat balance differential equation after the cooling heat is updated;

[0029] Step S233: Calculate the component concentration at the next time step according to the reaction temperature T(t+1) at the next time step and the kinetic model of the polyester polycondensation reaction;

[0030] Step S234: Output new status =[T(t+1),CX(t+1)].

[0031] Specifically, the specific method for designing the hierarchical dynamic reward function is:

[0032] Step S240: Preset target reaction temperature And the safe temperature range of polycondensation reactor , according to the absolute value of the difference between the target reaction temperature and the reaction temperature T(t) at the current time step t, the ratio is calculated with the safe temperature range to obtain the reaction temperature fluctuation amplitude δ(t);

[0033] The calculation formula of the reaction temperature fluctuation amplitude δ(t) is: ;

[0034] Step S250: Calculate the penalty function for the deviation of the temperature fluctuation amplitude from the target reaction temperature ;

[0035] The penalty function The calculation formula is: ,in, is the steepness parameter of the penalty function, is the threshold parameter of the penalty function;

[0036] Step S260: Count the number of time steps in which the reaction temperature continuously remains within the safe temperature range of the target reaction temperature in the time steps before the current time step t. , calculate the reward function for temperature stability ;

[0037] The reward function for temperature stability The calculation formula is: ,in, is the scaling factor of the reward function for temperature stability, is a logarithmic function;

[0038] Step S270: define the initial, middle and late stages of the polyester polycondensation reaction, wherein the initial stage corresponds to the esterification stage of the polyester polycondensation reaction, the middle stage corresponds to the transesterification stage of the polyester polycondensation reaction, and the late stage corresponds to the polycondensation stage of the polyester polycondensation reaction, and the weight coefficient of the penalty function is set according to different reaction stages. and the weight coefficient of the reward function , the weight coefficients of the penalty function and the reward function in the early, middle and late reaction stages are defined as , , and , , ;

[0039] Step S280: Weighted combination of penalty function and reward function to obtain hierarchical dynamic reward function ;

[0040] The calculation formula of the hierarchical dynamic reward function is: .

[0041] Specifically, the weight coefficient of the penalty function is set according to different reaction stages. and the weight coefficient of the reward function The specific method is:

[0042] Step S271: Calculate the concentration of the polycondensation product during the polyester polycondensation reaction and the final theoretical concentration The ratio between , taking the ratio as the reaction progress variable ;

[0043] Step S272: Setting the first reaction progress threshold As the critical point for the completion of the esterification stage, when the reaction progress variable reaches the first reaction progress threshold, the esterification stage is determined to be completed;

[0044] Step S273: Setting the second reaction progress threshold As the critical point for the completion of the transesterification stage, when the reaction progress variable reaches the second reaction progress threshold, the transesterification stage is determined to be completed;

[0045] Step S274: Setting the third reaction progress threshold As the critical point for the completion of the polycondensation stage, when the reaction progress variable reaches the third reaction progress threshold, the polycondensation stage is determined to be finished;

[0046] Step S275: Calculate the weight coefficient of the penalty function In each reaction stage, the linear interpolation method is used. For the esterification stage, the weight coefficient of the penalty function changes from Increase linearly to , for the transesterification stage, the weight coefficient of the penalty function changes from Increase linearly to , for the condensation stage, the weight coefficient of the penalty function is maintained at constant;

[0047] Step S276: Calculate the weight coefficient of the reward function , in each reaction stage, using the linear interpolation method, for the esterification stage, the weight coefficient of the reward function changes from Increase linearly to , for the transesterification stage, the weight coefficient of the reward function changes from Increase linearly to , for the condensation stage, the weight coefficient of the reward function is maintained at constant.

[0048] Specifically, the calculation method of the final theoretical concentration is:

[0049] Determine the stoichiometric ratio of dibasic acid monomers and diol monomers in polyester polycondensation reactions;

[0050] Calculate the molar numbers of the dibasic acid monomer and the diol monomer according to the feed amounts of the dibasic acid monomer and the diol monomer;

[0051] Calculate the theoretical molar number of the polycondensation product according to the molar number and stoichiometric ratio of the diol monomer and the dibasic acid monomer;

[0052] Calculate the theoretical mass of the polycondensation product according to the theoretical molar number and molar mass of the polycondensation product;

[0053] The final theoretical concentration of the polycondensation product was calculated based on the theoretical mass of the polycondensation product and the total mass of the reaction system.

[0054] Specifically, the DDPG algorithm is selected to train the reinforcement learning model, and the specific method of obtaining the trained reinforcement learning model is:

[0055] Step S310: Select the DDPG algorithm as the deep reinforcement learning algorithm for training the reinforcement learning model, and initialize the Actor network, the Critic network, and the corresponding target network;

[0056] Step S320: Initialize the experience playback buffer R;

[0057] Step S330: Set the number of training iterations M, the current number of iterations m=0;

[0058] Step S340: At the beginning of each training iteration, a state is randomly initialized and the agent interacts with the environment;

[0059] Step S350: using the quadruple sampled from the experience replay buffer, the target Q value is calculated. When the polyester polycondensation reaction is completed or the reaction temperature exceeds the safe temperature range of the target reaction temperature, the target Q value is equal to the hierarchical dynamic reward function value; when the polyester polycondensation reaction is not terminated and the reaction temperature does not exceed the safe temperature range of the target reaction temperature, the target Q value is composed of the hierarchical dynamic reward function and the Q value of the next state;

[0060] Step S360: Taking the mean square error between the target Q value and the state-action value function Q(s, a) output by the Critic network and maximizing the cumulative hierarchical dynamic reward function value of the action generated by the Actor network as the training goal, a trained reinforcement learning model is obtained.

[0061] Specifically, the specific method of using the trained reinforcement learning model to obtain the optimal temperature control strategy and optimizing the control of the polyester polycondensation reaction process is:

[0062] Step S410: obtaining the working volume and heat transfer area of ​​the polycondensation reactor of the polyester polycondensation reaction, recording the feeding amounts of the dibasic acid monomer and the diol monomer, obtaining the total heat transfer coefficient of the cooling system, and obtaining the ambient temperature and the reaction temperature during the reaction in real time;

[0063] Step S420: Utilize the trained reinforcement learning model to optimize the actual polyester polycondensation reaction process, and output the optimal cooling medium temperature to control the reaction temperature at each time step according to the real-time status in the polycondensation reactor.

[0064] One aspect of the present application provides a polymerization reaction continuous production line control system, comprising:

[0065] The heat balance building module is used to establish the heat release equation of the polycondensation reaction according to the reaction mechanism of the polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, establish the cooling heat equation according to the cooling system, establish the heat exchange equation between the polycondensation reactor and the environment, and obtain the heat balance differential equation in the polycondensation reactor;

[0066] The reinforcement learning building module is used to define the reaction temperature and component concentration in the polycondensation reactor as the state of the reinforcement learning model, define the cooling medium temperature as the action of the reinforcement learning model, establish the state transfer function according to the heat balance differential equation, design the hierarchical dynamic reward function, and build the reinforcement learning model;

[0067] The reinforcement learning training module is used to select the DDPG algorithm to train the reinforcement learning model and obtain a trained reinforcement learning model;

[0068] The temperature optimization control module is used to use the trained reinforcement learning model to obtain the optimal temperature control strategy and optimize the control of the polyester polycondensation reaction process.

[0069] One aspect of the present application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in a method for controlling a continuous production line of a polymerization reaction are implemented.

[0070] One aspect of the present application provides a readable storage medium storing a computer program, wherein the computer program is suitable for loading into a processor to execute steps in a method for controlling a continuous production line of a polymerization reaction.

[0071] The control system and method for a continuous polymerization production line proposed in this application have the following advantages over the prior art:

[0072] This application adopts a reinforcement learning method to perform real-time optimization control of the reaction temperature. The reaction temperature and component concentration in the polycondensation reactor are defined as states, and the cooling medium temperature is defined as an action. A state transfer function is established based on the heat balance differential equation. The optimal control strategy is obtained through training. The control strategy can be dynamically adjusted according to the real-time state to adapt to the complex dynamic characteristics of the polyester polycondensation reaction and achieve precise control of the reaction temperature.

[0073] This application designs a hierarchical dynamic reward function, which can perform targeted optimization according to the temperature control requirements of different stages. In the early stage of esterification, the focus is on penalizing temperature fluctuations to reach the target temperature as soon as possible, balancing penalties and rewards in the middle stage to maintain a constant temperature, and focusing on rewarding temperature stability in the later stage to ensure product quality. This hierarchical dynamic reward mechanism can better adapt to the specific requirements of temperature control at different stages of polyester condensation reactions.

[0074] This application considers both the temperature fluctuation amplitude and the temperature continuous stabilization time in the reward function. The temperature fluctuation is minimized by the penalty function, and the temperature continuous stabilization time is maximized by the reward function. This can suppress the temperature fluctuation while prompting the system to stabilize the temperature near the target reaction temperature for as long as possible, which is beneficial to improving the stability of product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 A method flow chart of a method for controlling a continuous production line of a polymerization reaction provided in this application;

[0076] Figure 2 A flow chart of a method for designing a hierarchical dynamic reward function in a polymerization reaction continuous production line control method provided in the present application;

[0077] Figure 3 A flow chart of a method for setting a weight coefficient of a penalty function and a weight coefficient of a reward function in a polymerization reaction continuous production line control method provided by the present application;

[0078] Figure 4 A functional module diagram of a polymerization reaction continuous production line control system provided in this application;

[0079] Figure 5 A schematic diagram of the structure of an electronic device provided in this application;

[0080] Figure 6 It is a schematic diagram of the structure of a readable storage medium provided by this application. DETAILED DESCRIPTION

[0081] In order to better understand the present application, various aspects of the present application will be described in more detail with reference to the accompanying drawings. It should be understood that these detailed descriptions are only descriptions of exemplary embodiments of the present application and do not limit the scope of the present application in any way. Throughout the specification, the same reference numerals refer to the same elements. The expression "and / or" includes any and all combinations of one or more of the associated listed items.

[0082] In the accompanying drawings, the size, dimensions and shapes of the elements have been slightly adjusted for ease of illustration. The drawings are for illustration only and are not strictly drawn to scale. As used herein, the terms "substantially", "approximately" and similar terms are used as terms of approximation, not as terms of degree, and are intended to illustrate the inherent deviations in measurements or calculations that will be recognized by those of ordinary skill in the art. In addition, in this application, the order in which the steps are described does not necessarily represent the order in which these processes occur in actual operation, unless otherwise specified or can be derived from the context.

[0083] It should also be understood that expressions such as "including", "comprising", "having", "containing" and / or "comprising" are open rather than closed expressions in this specification, which indicate the presence of the stated features, elements and / or components, but do not exclude the presence of one or more other features, elements, components and / or combinations thereof. In addition, when expressions such as "at least one of..." appear after a list of listed features, they modify the entire list of features rather than just the individual elements in the list. In addition, when describing embodiments of the present application, "may" is used to mean "one or more embodiments of the present application". And, the term "exemplary" is intended to refer to an example or illustration.

[0084] Unless otherwise defined, all words (including engineering terms and scientific and technological terms) used in this article have the same meaning as those commonly understood by ordinary technicians in the field to which this application belongs. It should also be understood that unless there is a clear explanation in this application, words defined in commonly used dictionaries should be interpreted as having the same meaning as their meaning in the context of the relevant technology, and should not be interpreted in an idealized or overly formal sense.

[0085] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0086] Example 1

[0087] like Figure 1 As shown, a polymerization reaction continuous production line control method provided by the present application comprises:

[0088] Step S100: establishing a polycondensation reaction exothermic equation according to the reaction mechanism of the polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, establishing a cooling heat equation according to the cooling system, establishing a heat exchange equation between the polycondensation reactor and the environment, and obtaining a heat balance differential equation in the polycondensation reactor;

[0089] The specific method of establishing the heat balance differential equation of the polycondensation reaction according to the reaction mechanism of the polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, establishing the cooling heat equation according to the cooling system, and establishing the heat exchange equation between the polycondensation reactor and the environment is as follows:

[0090] Step S110: According to the reaction mechanism of polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, the reaction rate rP of the polycondensation product, the working volume V of the polycondensation reactor and the reaction enthalpy change Multiply them to get the reaction exothermic rate of the polyester polycondensation reaction The exothermic equation of the polycondensation reaction;

[0091] The reaction mechanism of the polyester polycondensation reaction is simplified to a reversible polycondensation reaction of a dibasic acid monomer and a diol monomer, and the reaction formula is: , wherein A is a dibasic acid monomer, B is a diol monomer, P is a polycondensation product, and W is water;

[0092] The kinetic model of the polyester polycondensation reaction is: ,in, is the forward reaction rate constant, represents the reverse reaction rate constant, rP represents the reaction rate of the polycondensation product, , represent the concentration of dibasic acid monomer, diol monomer, polycondensation product, and water, respectively;

[0093] The exothermic equation of the polycondensation reaction is: ,in, is the reaction rate of the polycondensation product, is the working volume of the polycondensation reactor, is the reaction enthalpy change.

[0094] Step S120: Calculate the cooling heat according to the difference between the reaction temperature of the cooling system and the temperature of the cooling medium, multiplied by the total heat transfer coefficient and the heat transfer area of ​​the polycondensation reactor. The cooling heat equation;

[0095] The cooling heat equation is: ,in, is the overall heat transfer coefficient, is the heat transfer area of ​​the polycondensation reactor, T is the reaction temperature, is the cooling medium temperature;

[0096] The total heat transfer coefficient refers to the overall heat transfer performance of the heat exchanger, which is related to the type, size, material, and fluid properties of the heat exchanger, and is determined by those skilled in the art based on the properties of the heat exchanger.

[0097] Step S130: According to Newton's cooling law, the temperature difference between the reaction temperature and the ambient temperature is calculated, and the temperature difference is multiplied by the convection heat transfer coefficient and the heat transfer area of ​​the polycondensation reactor to obtain the heat transfer coefficient between the polycondensation reactor and the environment. The heat exchange equation of

[0098] The heat exchange equation is: , where h is the convective heat transfer coefficient and Ta is the ambient temperature;

[0099] The convective heat transfer coefficient refers to the convective heat transfer coefficient between the polycondensation reactor and the environment, which is determined by those skilled in the art based on experience.

[0100] Step S140: establishing a heat balance differential equation in the polycondensation reactor according to the polycondensation reaction heat release equation, the cooling heat equation and the heat exchange equation;

[0101] The heat balance differential equation in the polycondensation reactor is: ,in, is the density of the reactants, is the specific heat capacity, is the change in reaction temperature within each time step t;

[0102] Reactant density refers to the density of the mixture of reactants; specific heat capacity refers to the specific heat capacity of the mixture of reactants; reaction temperature change refers to the temperature change of the mixture of reactants; reaction enthalpy change represents the total change in system enthalpy during the reaction, which is set by those skilled in the art. The reaction enthalpy change is a constant, indicating that the reaction always proceeds with the same enthalpy change; reaction temperature refers to the temperature of the mixture of reactants.

[0103] The heat balance differential equation of the polyester polycondensation reaction describes the quantitative relationship between the consumption of reactants, the generation of polycondensation products, the reaction temperature and the heat transfer during the polyester polycondensation reaction;

[0104] A heat balance differential equation is established as the core mathematical model of the entire control system, and the heat balance principle of polyester polycondensation reaction is used to provide theoretical guidance for subsequent reaction temperature control.

[0105] Step S200: defining the reaction temperature and component concentration in the polycondensation reactor as the state of the reinforcement learning model, defining the cooling medium temperature as the action of the reinforcement learning model, establishing a state transfer function according to the heat balance differential equation, designing a hierarchical dynamic reward function, and constructing a reinforcement learning model;

[0106] The specific method of defining the reaction temperature and component concentration in the polycondensation reactor as the state of the reinforcement learning model, defining the cooling medium temperature as the action of the reinforcement learning model, establishing the state transfer function according to the heat balance differential equation, and designing the hierarchical dynamic reward function to construct the reinforcement learning model is as follows:

[0107] Step S210: defining a state variable set S of a reinforcement learning model, wherein the state s(t) in the state variable set S includes: a reaction temperature T in the polycondensation reactor, a component concentration CX, wherein the component concentration CX includes a dibasic acid monomer concentration CA, a diol monomer concentration CB, a polycondensation product concentration CP, and a water concentration CW, wherein t in s(t) represents a time step;

[0108] X in the component concentration CX = {A, B, P, W}, where X represents the component, A is the dibasic acid monomer, B is the diol monomer, P is the polycondensation product, and W is water;

[0109] Step S220: defining an action variable set A of a reinforcement learning model, wherein the action a(t) in the action variable set A is the temperature of the cooling medium;

[0110] The adjustment range of the cooling medium temperature is set by technicians in this field according to equipment conditions and process requirements.

[0111] Step S230: Establishing a state transfer function based on the heat balance differential equation , represents the state at the current time step Next action After that, it transfers to the new state at the next time step ;

[0112] The state at the current time step Next action After that, it transfers to the new state at the next time step The specific method is:

[0113] Step S231: Substitute the reaction temperature T(t) and component concentration CX(t) of the current time step in state s(t) into the heat balance differential equation, and calculate the heat balance according to the action Adjust the cooling medium temperature and update the cooling heat ;

[0114] Step S232: Calculate the reaction temperature T(t+1) of the next time step according to the heat balance differential equation after the cooling heat is updated;

[0115] Step S233: Calculate the component concentration at the next time step according to the reaction temperature T(t+1) at the next time step and the kinetic model of the polyester polycondensation reaction;

[0116] Step S234: Output new status =[T(t+1),CX(t+1)].

[0117] like Figure 2 As shown, it is a flow chart of a method for designing a hierarchical dynamic reward function in a polymerization reaction continuous production line control method provided by the present application;

[0118] The specific method for designing the hierarchical dynamic reward function is:

[0119] Step S240: Preset target reaction temperature And the safe temperature range of polycondensation reactor , according to the absolute value of the difference between the target reaction temperature and the reaction temperature T(t) at the current time step t, the ratio is calculated with the safe temperature range to obtain the reaction temperature fluctuation amplitude δ(t);

[0120] The calculation formula of the reaction temperature fluctuation amplitude δ(t) is: ;

[0121] The safe temperature range of the polycondensation reactor is usually set to a certain range above and below the target reaction temperature; preferably, =20°C; the safe temperature range refers to the range of fluctuations of the target reaction temperature;

[0122] The reaction temperature fluctuation amplitude δ(t) is used to measure the degree of deviation of the reaction temperature from the target temperature in the current state.

[0123] The target reaction temperature refers to the optimal reaction temperature determined to achieve the expected product quality and yield during the polyester polycondensation reaction, and is set by those skilled in the art based on experience and actual needs.

[0124] Step S250: Calculate the penalty function for the deviation of the temperature fluctuation amplitude from the target reaction temperature ;

[0125] The penalty function The calculation formula is: ,in, is the steepness parameter of the penalty function, is the threshold parameter of the penalty function;

[0126] The steepness parameter of the penalty function and the threshold parameter of the penalty function are used to control the change rate and triggering timing of the penalty function, and are set by those skilled in the art according to actual needs.

[0127] Step S260: Count the number of time steps in which the reaction temperature continuously remains within the safe temperature range of the target reaction temperature in the time steps before the current time step t. , calculate the reward function for temperature stability ;

[0128] The reward function for temperature stability The calculation formula is: ,in, is the scaling factor of the reward function for temperature stability, is a logarithmic function;

[0129] Step S270: define the initial, middle and late stages of the polyester polycondensation reaction, wherein the initial stage corresponds to the esterification stage of the polyester polycondensation reaction, the middle stage corresponds to the transesterification stage of the polyester polycondensation reaction, and the late stage corresponds to the polycondensation stage of the polyester polycondensation reaction, and the weight coefficient of the penalty function is set according to different reaction stages. and the weight coefficient of the reward function , the weight coefficients of the penalty function and the reward function in the early, middle and late reaction stages are defined as , , and , , ;

[0130] like Figure 3 As shown, it is a flow chart of a method for setting a weight coefficient of a penalty function and a weight coefficient of a reward function in a polymerization reaction continuous production line control method provided by the present application;

[0131] The weight coefficient of the penalty function is set according to different reaction stages. and the weight coefficient of the reward function The specific method is:

[0132] Step S271: Calculate the concentration of the polycondensation product during the polyester polycondensation reaction and the final theoretical concentration The ratio between , taking the ratio as the reaction progress variable ;

[0133] The calculation method of the final theoretical concentration is:

[0134] Determine the stoichiometric ratio of dibasic acid monomers and diol monomers in polyester polycondensation reactions;

[0135] The stoichiometric ratio is determined based on the molar ratio of the dibasic acid monomer to the diol monomer in the polyester polycondensation reaction;

[0136] Calculate the molar numbers of the dibasic acid monomer and the diol monomer according to the feed amounts of the dibasic acid monomer and the diol monomer;

[0137] Calculate the theoretical molar number of the polycondensation product according to the molar number and stoichiometric ratio of the diol monomer and the dibasic acid monomer;

[0138] Calculate the theoretical mass of the polycondensation product according to the theoretical molar number and molar mass of the polycondensation product;

[0139] The final theoretical concentration of the polycondensation product was calculated based on the theoretical mass of the polycondensation product and the total mass of the reaction system.

[0140] Exemplarily, the method for calculating the final theoretical concentration is based on the polyester polycondensation reaction equation nHO-R1-OH + nHOOC-R2-COOH → HO-(R1-OOC-R2-COO)nH + (2n-1)H2O, where HO-R1-OH represents a diol monomer, HOOC-R2-COOH represents a dibasic acid monomer, HO-(R1-OOC-R2-COO)nH represents a generated polycondensation product, n represents the number of repetitions of each monomer in the polyester reaction chain, and H2O represents water;

[0141] According to the polyester polycondensation reaction equation, the molar ratio of diol monomer to dibasic acid monomer is ;

[0142] According to the feeding amount of diol monomer and dibasic acid monomer and , calculate the molar numbers of diol monomer and dibasic acid monomer respectively and ,in, and are the molar masses of the diol monomer and the dibasic acid monomer, respectively;

[0143] According to the stoichiometric ratio, the theoretical molar number of the polycondensation product is calculated. Since the molar ratio of the diol monomer to the dibasic acid monomer is 1:1, the theoretical molar number of the polycondensation product is is equal to the molar number of diol monomer or dibasic acid monomer, that is ;

[0144] Calculate the theoretical mass of the polycondensation product. Calculate the theoretical mass of the polycondensation product based on the theoretical molar number and molar mass of the polycondensation product. ,Right now ,in, is the molar mass of the polycondensation product, which is calculated based on the structural formula of the monomer, namely ,in, is the molar mass of water;

[0145] Calculate the final theoretical concentration based on the theoretical mass of the polycondensation product and the total mass of the reaction system. , the calculation formula is: ,in, It is the total mass of the reaction system, including the sum of the masses of diol monomer, dibasic acid monomer and other additives.

[0146] Step S272: Setting the first reaction progress threshold As the critical point for the completion of the esterification stage, when the reaction progress variable reaches the first reaction progress threshold, the esterification stage is determined to be completed;

[0147] The first reaction progress threshold is set by those skilled in the art based on experience and actual needs; preferably, =0.2;

[0148] When the reaction progress variable reaches 0.2, the esterification stage is considered to be over and the ester exchange stage begins;

[0149] During the reaction, near infrared spectroscopy was used to measure the real-time changing concentration of the polycondensation product;

[0150] Step S273: Setting the second reaction progress threshold As the critical point for the completion of the transesterification stage, when the reaction progress variable reaches the second reaction progress threshold, the transesterification stage is determined to be completed;

[0151] The second reaction progress threshold is set by those skilled in the art based on experience and actual needs; preferably, =0.6;

[0152] When the reaction progress variable reaches 0.6, the transesterification stage is considered to be over and the polycondensation stage begins;

[0153] Step S274: Setting the third reaction progress threshold As the critical point for the completion of the polycondensation stage, when the reaction progress variable reaches the third reaction progress threshold, the polycondensation stage is determined to be finished;

[0154] The third reaction progress threshold is set by those skilled in the art based on experience and actual needs; preferably, =0.8;

[0155] When the reaction progress variable reaches 0.8, the polycondensation stage is considered to be over and the reaction is complete;

[0156] Step S275: Calculate the weight coefficient of the penalty function In each reaction stage, the linear interpolation method is used. For the esterification stage, the weight coefficient of the penalty function changes from Increase linearly to , for the transesterification stage, the weight coefficient of the penalty function changes from Increase linearly to , for the condensation stage, the weight coefficient of the penalty function is maintained at constant;

[0157] In the esterification stage, the calculation formula of the weight coefficient of the penalty function is: ;

[0158] In the transesterification stage, the calculation formula of the weight coefficient of the penalty function is: ;

[0159] In the condensation stage, the calculation formula of the weight coefficient of the penalty function is: ;

[0160] Step S276: Calculate the weight coefficient of the reward function , in each reaction stage, using the linear interpolation method, for the esterification stage, the weight coefficient of the reward function changes from Increase linearly to , for the transesterification stage, the weight coefficient of the reward function changes from Increase linearly to , for the condensation stage, the weight coefficient of the reward function is maintained at constant;

[0161] In the esterification stage, the calculation formula of the weight coefficient of the reward function is: ;

[0162] In the transesterification stage, the calculation formula of the weight coefficient of the reward function is: ;

[0163] In the condensation stage, the calculation formula of the weight coefficient of the reward function is: ;

[0164] The weight coefficients of the penalty function and reward function in the early, middle and late reaction stages , , and , , It is set by those skilled in the art according to the temperature control requirements and expert experience of the polyester polycondensation reaction.

[0165] Step S280: Weighted combination of penalty function and reward function to obtain hierarchical dynamic reward function ;

[0166] The calculation formula of the hierarchical dynamic reward function is: .

[0167] Introducing dynamic weights at different reaction stages can adapt the reward function to the characteristics of the reaction process; for example, in the early stage, it focuses on penalizing temperature fluctuations to reach the target reaction temperature as quickly as possible, in the middle stage, it balances fluctuation penalties and stability rewards to maintain a constant temperature, and in the later stage, it focuses on rewarding temperature stability to ensure product quality.

[0168] The above steps design a hierarchical dynamic reward function, which comprehensively considers the penalty function of temperature fluctuation and the reward function of temperature stability, and dynamically adjusts the weights of the two according to the reaction stage to adapt to the temperature control requirements of the entire reaction process.

[0169] Step S300: Select the DDPG algorithm to train the reinforcement learning model to obtain a trained reinforcement learning model;

[0170] The specific method of selecting the DDPG algorithm to train the reinforcement learning model to obtain the trained reinforcement learning model is:

[0171] Step S310: Select the DDPG algorithm as the deep reinforcement learning algorithm for training the reinforcement learning model, and initialize the Actor network, the Critic network, and the corresponding target network;

[0172] The Actor network receives the state as input, outputs a deterministic action, and initializes the Actor network parameters using random initialization.

[0173] The Critic network receives state and action as input, outputs the state-action value function Q(s,a), and initializes the Critic network parameters by random initialization.

[0174] Initialize the target Actor network and target Critic network. Their network structures are the same as the original networks, and their parameters are initialized as a copy of the original network parameters.

[0175] Step S320: Initialize the experience playback buffer R;

[0176] The experience replay buffer is used to store state transition data generated by the interaction between the agent and the environment;

[0177] Step S330: Set the number of training iterations M, the current number of iterations m=0;

[0178] Step S340: At the beginning of each training iteration, a state is randomly initialized and the agent interacts with the environment;

[0179] The specific process of the agent interacting with the environment is as follows: inputting the state s into the Actor network, obtaining the action a, applying the action a to the current time step, and obtaining the state s' of the next time step; forming a quaternion consisting of the state s of the current time step, the action a, the hierarchical dynamic reward function and the state s' of the next time step, and storing it in the experience playback buffer;

[0180] Step S350: using the quadruple sampled from the experience replay buffer, the target Q value is calculated. When the polyester polycondensation reaction is completed or the reaction temperature exceeds the safe temperature range of the target reaction temperature, the target Q value is equal to the hierarchical dynamic reward function value; when the polyester polycondensation reaction is not terminated and the reaction temperature does not exceed the safe temperature range of the target reaction temperature, the target Q value is composed of the hierarchical dynamic reward function and the Q value of the next state;

[0181] When the polyester polycondensation reaction is not terminated and the reaction temperature does not exceed the safe temperature range of the target reaction temperature, the target Q value is calculated as follows: , where y represents the target Q value, γ is the discount factor, Q' represents the target Critic network, Represents the target Actor network;

[0182] The discount factor is usually between 0.9 and 0.99, indicating the importance of future rewards.

[0183] Step S360: Taking the mean square error between the target Q value and the state-action value function Q(s, a) output by the Critic network and maximizing the cumulative hierarchical dynamic reward function value of the action generated by the Actor network as the training goal, a trained reinforcement learning model is obtained.

[0184] Step S400: using the trained reinforcement learning model to obtain the optimal temperature control strategy to optimize the polyester polycondensation reaction process;

[0185] The specific method of using the trained reinforcement learning model to obtain the optimal temperature control strategy and optimize the control of the polyester polycondensation reaction process is:

[0186] Step S410: obtaining the working volume and heat transfer area of ​​the polycondensation reactor of the polyester polycondensation reaction, recording the feeding amounts of the dibasic acid monomer and the diol monomer, obtaining the total heat transfer coefficient of the cooling system, and obtaining the ambient temperature and the reaction temperature during the reaction in real time;

[0187] Step S420: Utilize the trained reinforcement learning model to optimize the actual polyester polycondensation reaction process, and output the optimal cooling medium temperature to control the reaction temperature at each time step according to the real-time status in the polycondensation reactor.

[0188] Example 2

[0189] like Figure 4 As shown, a polymerization reaction continuous production line control system provided by the present application includes:

[0190] The heat balance building module is used to establish the heat release equation of the polycondensation reaction according to the reaction mechanism of the polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, establish the cooling heat equation according to the cooling system, establish the heat exchange equation between the polycondensation reactor and the environment, and obtain the heat balance differential equation in the polycondensation reactor;

[0191] The reinforcement learning building module is used to define the reaction temperature and component concentration in the polycondensation reactor as the state of the reinforcement learning model, define the cooling medium temperature as the action of the reinforcement learning model, establish the state transfer function according to the heat balance differential equation, design the hierarchical dynamic reward function, and build the reinforcement learning model;

[0192] The reinforcement learning training module is used to select the DDPG algorithm to train the reinforcement learning model and obtain a trained reinforcement learning model;

[0193] The temperature optimization control module is used to use the trained reinforcement learning model to obtain the optimal temperature control strategy and optimize the control of the polyester polycondensation reaction process.

[0194] Example 3

[0195] Figure 5 1 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Figure 5 As shown, according to another aspect of the present application, an electronic device is also provided. The electronic device may include one or more processors and one or more memories. The memories store computer readable codes, and when the computer readable codes are run by one or more processors, a polymerization reaction continuous production line control method as described above may be executed.

[0196] The method or system according to the embodiment of the present application can also be used by Figure 5 The electronic device architecture shown in FIG. Figure 5 As shown, the electronic device may include a bus, one or more CPUs, a read-only memory (ROM), a random access memory (RAM), a communication port connected to a network, an input / output component, a hard disk, and the like. A storage device in an electronic device, such as a ROM or a hard disk, may store a method for controlling a continuous production line of a polymerization reaction provided in the present application. A method for controlling a continuous production line of a polymerization reaction may, for example, include: establishing an exothermic equation for the polycondensation reaction according to a kinetic model of the reaction mechanism of the polyester polycondensation reaction and the reaction rate of the polycondensation product, establishing a cooling heat equation according to the cooling system, establishing a heat exchange equation between the polycondensation reactor and the environment, and obtaining a heat balance differential equation in the polycondensation reactor; defining the reaction temperature and component concentration in the polycondensation reactor as the state of the reinforcement learning model, defining the cooling medium temperature as the action of the reinforcement learning model, establishing a state transfer function according to the heat balance differential equation, designing a hierarchical dynamic reward function, and constructing a reinforcement learning model; selecting a DDPG algorithm to train the reinforcement learning model to obtain a trained reinforcement learning model; using the trained reinforcement learning model to obtain the optimal temperature control strategy to optimize the control of the polyester polycondensation reaction process. Furthermore, the electronic device may also include a user interface. Of course, Figure 5The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 5 One or more components of an electronic device are shown.

[0197] Example 4

[0198] Figure 6 Schematic diagram of a readable storage medium structure provided by an embodiment of the present application. Figure 6 As shown, it is a readable storage medium according to an embodiment of the present application. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are executed by the processor, a polymerization reaction continuous production line control method according to an embodiment of the present application described with reference to the above figures can be executed. The storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache). Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0199] In addition, according to the embodiments of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, for example: according to the reaction mechanism of the polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, establish the polycondensation reaction exothermic equation, establish the cooling heat equation according to the cooling system, establish the heat exchange equation between the polycondensation reactor and the environment, and obtain the heat balance differential equation in the polycondensation reactor; define the reaction temperature and component concentration in the polycondensation reactor as the state of the reinforcement learning model, define the cooling medium temperature as the action of the reinforcement learning model, establish the state transfer function according to the heat balance differential equation, design the hierarchical dynamic reward function, and construct the reinforcement learning model; select the DDPG algorithm to train the reinforcement learning model to obtain a trained reinforcement learning model; use the trained reinforcement learning model to obtain the optimal temperature control strategy to optimize the polyester polycondensation reaction process. When the computer program is executed by the central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0200] The methods, apparatuses, and devices of the present application may be implemented in many ways. For example, the methods, apparatuses, and devices of the present application may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present application are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present application may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the method according to the present application. Therefore, the present application also covers recording media storing programs for executing the method according to the present application.

[0201] In addition, the parts of the above-mentioned technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive redundancy.

[0202] The specific implementation modes as described above further describe the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation mode of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for controlling a continuous production line of a polymerization reaction, characterized in that: include: According to the reaction mechanism of polyester polycondensation reaction and the kinetic model of the reaction rate of polycondensation product, the polycondensation reaction exothermic equation is established, the cooling heat equation is established according to the cooling system, the heat exchange equation between the polycondensation reactor and the environment is established, and the heat balance differential equation in the polycondensation reactor is obtained; The reaction temperature and component concentration in the polycondensation reactor are defined as the state of the reinforcement learning model, the cooling medium temperature is defined as the action of the reinforcement learning model, the state transfer function is established according to the heat balance differential equation, the hierarchical dynamic reward function is designed, and the reinforcement learning model is constructed; Select the DDPG algorithm to train the reinforcement learning model to obtain a trained reinforcement learning model; The trained reinforcement learning model is used to obtain the optimal temperature control strategy and optimize the control of the polyester polycondensation reaction process; The specific method for designing the hierarchical dynamic reward function is: Preset target reaction temperature And the safe temperature range of polycondensation reactor , according to the absolute value of the difference between the target reaction temperature and the reaction temperature T(t) at the current time step t, the ratio is calculated with the safe temperature range to obtain the reaction temperature fluctuation amplitude δ(t); Calculate the penalty function for the temperature fluctuation deviation from the target reaction temperature ; Count the number of time steps before the current time step t in which the reaction temperature continuously remains within the safe temperature range of the target reaction temperature , calculate the reward function for temperature stability ; Define the initial, middle and late stages of polyester polycondensation reaction. The initial stage corresponds to the esterification stage of polyester polycondensation reaction, the middle stage corresponds to the transesterification stage of polyester polycondensation reaction, and the late stage corresponds to the polycondensation stage of polyester polycondensation reaction. Set the weight coefficient of penalty function according to different reaction stages. and the weight coefficient of the reward function , the weight coefficients of the penalty function and the reward function in the early, middle and late reaction stages are defined as , , and , , ; The penalty function and reward function are weighted together to obtain a hierarchical dynamic reward function .

2. A method for controlling a continuous polymerization production line according to claim 1, characterized in that: The specific method of establishing the heat balance differential equation of the polycondensation reaction according to the reaction mechanism of the polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, establishing the cooling heat equation according to the cooling system, and establishing the heat exchange equation between the polycondensation reactor and the environment is as follows: According to the reaction mechanism of polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, the reaction rate rP of the polycondensation product, the working volume V of the polycondensation reactor and the reaction enthalpy change Multiply them to get the reaction exothermic rate of the polyester polycondensation reaction The exothermic equation of the polycondensation reaction; The cooling heat is calculated based on the difference between the reaction temperature of the cooling system and the temperature of the cooling medium, and the product of the total heat transfer coefficient and the heat transfer area of ​​the polycondensation reactor. The cooling heat equation; According to Newton's cooling law, the temperature difference between the reaction temperature and the ambient temperature is calculated, and the temperature difference is multiplied by the convection heat transfer coefficient and the heat transfer area of ​​the polycondensation reactor to obtain the heat exchange coefficient between the polycondensation reactor and the environment. The heat exchange equation of According to the polycondensation reaction exothermic equation, cooling heat equation and heat exchange equation, the heat balance differential equation in the polycondensation reactor is established.

3. A method for controlling a continuous polymerization production line according to claim 2, characterized in that: The specific method of defining the reaction temperature and component concentration in the polycondensation reactor as the state of the reinforcement learning model, defining the cooling medium temperature as the action of the reinforcement learning model, and establishing the state transfer function according to the heat balance differential equation is: A state variable set S of the reinforcement learning model is defined, wherein the state s(t) in the state variable set S includes: a reaction temperature T in the polycondensation reactor, a component concentration CX, wherein the component concentration CX includes a dibasic acid monomer concentration CA, a diol monomer concentration CB, a polycondensation product concentration CP, and a water concentration CW, wherein t in s(t) represents a time step; An action variable set A of a reinforcement learning model is defined, wherein an action a(t) in the action variable set A is a cooling medium temperature; Establishing the state transfer function based on the heat balance differential equation , represents the state at the current time step Next action After that, it transfers to the new state at the next time step .

4. A method for controlling a continuous polymerization production line according to claim 3, characterized in that: The weight coefficient of the penalty function is set according to different reaction stages. and the weight coefficient of the reward function The specific method is: Calculate the concentration of polycondensation products during polyester polycondensation and the final theoretical concentration The ratio between , taking the ratio as the reaction progress variable ; Setting the first response progress threshold As the critical point for the completion of the esterification stage, when the reaction progress variable reaches the first reaction progress threshold, the esterification stage is determined to be completed; Set the second reaction progress threshold As the critical point for the completion of the transesterification stage, when the reaction progress variable reaches the second reaction progress threshold, the transesterification stage is determined to be completed; Set the third reaction progress threshold As the critical point for the completion of the polycondensation stage, when the reaction progress variable reaches the third reaction progress threshold, the polycondensation stage is determined to be finished; Calculate the weight coefficient of the penalty function In each reaction stage, the linear interpolation method is used. For the esterification stage, the weight coefficient of the penalty function changes from Increase linearly to , for the transesterification stage, the weight coefficient of the penalty function changes from Increase linearly to , for the condensation stage, the weight coefficient of the penalty function is maintained at constant; Calculate the weight coefficient of the reward function , in each reaction stage, using the linear interpolation method, for the esterification stage, the weight coefficient of the reward function changes from Increase linearly to , for the transesterification stage, the weight coefficient of the reward function changes from Increase linearly to , for the condensation stage, the weight coefficient of the reward function is maintained at constant.

5. A method for controlling a continuous polymerization production line according to claim 4, characterized in that: The calculation method of the final theoretical concentration is: Determine the stoichiometric ratio of dibasic acid monomers and diol monomers in polyester polycondensation reactions; Calculate the molar numbers of the dibasic acid monomer and the diol monomer according to the feed amounts of the dibasic acid monomer and the diol monomer; Calculate the theoretical molar number of the polycondensation product according to the molar number and stoichiometric ratio of the diol monomer and the dibasic acid monomer; Calculate the theoretical mass of the polycondensation product according to the theoretical molar number and molar mass of the polycondensation product; The final theoretical concentration of the polycondensation product was calculated based on the theoretical mass of the polycondensation product and the total mass of the reaction system.

6. A method for controlling a continuous polymerization production line according to claim 5, characterized in that: The specific method of selecting the DDPG algorithm to train the reinforcement learning model to obtain the trained reinforcement learning model is: Select the DDPG algorithm as the deep reinforcement learning algorithm for training the reinforcement learning model, initialize the Actor network, Critic network and the corresponding target network; Initialize the experience replay buffer R; Set the number of training iterations M, the current number of iterations m=0; At the beginning of each training iteration, a state is randomly initialized and the agent interacts with the environment; The target Q value is calculated using the quadruple sampled from the experience replay buffer. When the polyester polycondensation reaction is completed or the reaction temperature exceeds the safe temperature range of the target reaction temperature, the target Q value is equal to the hierarchical dynamic reward function value; when the polyester polycondensation reaction is not terminated and the reaction temperature does not exceed the safe temperature range of the target reaction temperature, the target Q value is composed of the hierarchical dynamic reward function and the Q value of the next state; The training objectives are to minimize the mean square error between the target Q value and the state-action value function Q(s, a) output by the Critic network and to maximize the cumulative hierarchical dynamic reward function value of the action generated by the Actor network to obtain a trained reinforcement learning model.

7. A polymerization reaction continuous production line control system, which is used to implement a polymerization reaction continuous production line control method according to any one of claims 1 to 6, characterized in that: include: The heat balance building module is used to establish the heat release equation of the polycondensation reaction according to the reaction mechanism of the polyester polycondensation reaction and the kinetic model of the reaction rate of the polycondensation product, establish the cooling heat equation according to the cooling system, establish the heat exchange equation between the polycondensation reactor and the environment, and obtain the heat balance differential equation in the polycondensation reactor; The reinforcement learning building module is used to define the reaction temperature and component concentration in the polycondensation reactor as the state of the reinforcement learning model, define the cooling medium temperature as the action of the reinforcement learning model, establish the state transfer function according to the heat balance differential equation, design the hierarchical dynamic reward function, and build the reinforcement learning model; The reinforcement learning training module is used to select the DDPG algorithm to train the reinforcement learning model and obtain a trained reinforcement learning model; The temperature optimization control module is used to use the trained reinforcement learning model to obtain the optimal temperature control strategy and optimize the control of the polyester polycondensation reaction process.

8. An electronic device, characterized in that: The invention comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the method for controlling a continuous production line of a polymerization reaction as described in any one of claims 1 to 6 are implemented.

9. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor to execute the steps in the method for controlling a continuous production line for a polymerization reaction as described in any one of claims 1 to 6.

Images