Anti-interference control method for molecular weight distribution in styrene polymerization process based on dual event trigger mechanism

Through the dual event triggering mechanism and B-spline function approximation technology, the problems of poor stability and particle size distribution shape of molecular weight distribution during styrene polymerization in chemical production are solved, and effective control of non-Gaussian random systems are achieved, which improves the Communication efficiency and resource utilization.

CN114624998BActive Publication Date: 2024-12-06YANGZHOU UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202210021250.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-01-10
Publication Date
2024-12-06
Estimated Expiration
2042-01-10

AI Technical Summary

Technical Problem

The prior art is difficult to effectively control the molecular weight distribution in non-Gaussian random systems, especially in the process of styrene polymerization in chemical production, where there are problems of poor stability and difficult control of particle size distribution and shape.

Method used

Using an anti-interference control method based on the dual event trigger mechanism, the static molecular weight distribution model of styrene polymer is established, and the probability density function is used to approximate the probability density function, and the perturbation observer and PI controller are designed, combined with the Lyapunov stability analysis method to realize tracking control of the target probability density function.

Benefits of technology

Accurate control of the molecular weight distribution of the styrene polymerization process is achieved, channel communication efficiency is improved, data redundancy and waste of computing resources are reduced, and the impact of external disturbances on system control is reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114624998B_ABST
    Figure CN114624998B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for controlling the shape of molecular weight distribution in a styrene polymerization process based on a dual-event synchronous trigger mechanism in the field of process control, comprising the following steps: 1) establishing a styrene conservation equation, and modeling the molecular weight distribution of a final polymer product produced by a reactor; 2) using a B-spline function for approximation, and using a recursive least squares identification method to obtain a parameter matrix of a state space model; 3) respectively designing a disturbance observer and a dual-event trigger mechanism to estimate unknown interference and combine them with state feedback to form a PI controller; 4) using a Lyapunov stability analysis method to solve a linear matrix inequality equation to solve parameters such as each gain of a structured controller. The present invention can provide good interference suppression capability, save communication channels and computing resources, and significantly improve the stability of control in the styrene chemical polymerization process and good target tracking convergence.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of process control, and in particular to an anti-interference control method. Background Art

[0002] After years of development, the study of random control of random systems has become an important research direction in the control engineering community, and has accumulated a large number of research results and a complete theoretical system. Typical theoretical results of random system control include minimum variance control, random linear quadratic control, random neural network model control and other methods. The essence of these results is to track the statistical characteristics of the output random variables of the controlled system, that is, the main control research object is the mean and variance of the output, and the control target is achieved by controlling the mean and variance of the output variables of the random system to meet the specified target characteristics. However, most control systems are based on the assumption that their output random variables obey the normal distribution. In this way, when the system output and noise disturbances obey or are limited to Gaussian variables, their statistical characteristics mean and variance can control the system output probability density function.

[0003] With the extensive application of random system control theory in practical engineering, it is found that in actual production and life, such as grain particle size distribution in grain processing, molecular weight distribution in chemical production, flame distribution in furnace combustion, etc., the statistical characteristics of the contained or output variables are non-Gaussian random variables or do not obey the normal distribution, and the random disturbance noise in the system is often a non-Gaussian variable that is not normally distributed. This is because its probability density function is mostly asymmetric and multi-peaked, and the two statistical characteristics of mean and variance no longer accurately describe the probability characteristics of the output variable. Facing the industrial engineering control objects in the actual production and life process, with the popularization and rapid development of industry and information technology, the application of various production scenarios such as big data, artificial intelligence, image processing, and deep learning has greatly increased the requirements for control capabilities for the massive data generated. Therefore, the research on random system control of non-Gaussian variables has far-reaching practical engineering significance and application background, but the statistical characteristics of non-Gaussian variables can only be accurately described through a probability density function with a positive integral constraint. The application background of the present invention is to take the probability density function of the molecular weight distribution output of the styrene polymerization process in the chemical production background as an example. The systematic analysis is aimed at the modeling and control methods of non-Gaussian random systems, and a new research direction is explored, which is the random distribution control theory.

[0004] Nowadays, with the development of sensors and communication technology, information transmission in control systems mostly adopts network transmission, and periodic sampling, a highly conservative design method, is mostly used for information collection. However, high-frequency sampling will occupy or waste limited channel and bandwidth resources, and over-periodic sampling will bring a large amount of redundant information to occupy storage resources. The retention of redundant information affects the operation of the controller and leads to the occupation and waste of computing resources. In order to overcome the limitations of the periodic sampling mechanism, the event trigger mechanism was proposed in the 1990s. The core idea of ​​this mechanism is that when the system state or other objects meet a certain range of given functional relationship conditions, the event is not triggered and the data is not transmitted until the system object exceeds the set threshold, the event is triggered, and the transmission data is sent. The transmission of information is determined by the set "event", which effectively reduces the frequency of information data transmission, reduces channel load, and reduces resource waste.

[0005] Polymer is a chemical reaction process in which monomers are excited and combined with other molecular structures through polymerization reaction. In the chemical process, polymerization reaction also refers to the process in which low molecular weight monomers are combined and polymerized through a series of reactions such as chain initiation, chain growth, chain transfer, etc. to form the final high molecular weight polymer. The physicochemical properties of chemical products in production and life are affected by the microscopic molecular structure of the final polymer, thereby determining the quality and application of chemical polymer products. The molecular weight distribution of polymers is one of the important characteristic parameters that characterize the microscopic molecular structure of polymers, and is also the object of exploration and control of the present invention. The final polymer molecular weight distribution belongs to the category of distribution function, which is usually used to describe the proportion of molecules with different molecular weights in the polymerization product, and is in units of polymerization degree chain length j, and there are several different description methods: number-based molecular weight distribution, weight-based molecular weight distribution, etc. Finally, the molecular weight distribution control problem of chemical polymerization process needs to start with the following two modeling: one, a mechanism model derived from physicochemical mechanism; two, an empirical model based on a large amount of experimental data. Regardless of the model, there are many problems to be solved. The model based on the physicochemical reaction mechanism adopted in the present invention also needs to consider multiple aspects of the problem, such as the problem of external unknown disturbance, the accuracy of the spline function approximating its probability density function, etc. In the detailed introduction of the invention below, the design also provides corresponding solutions. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides an anti-interference control method for the molecular weight distribution in a styrene polymerization process based on a dual-event trigger mechanism, which solves the problems of poor stability and particle size distribution shape control in the presence of input interference in the existing reaction system.

[0007] The object of the present invention is achieved by: a method for anti-interference control of molecular weight distribution in styrene polymerization process based on a dual event trigger mechanism, comprising the following steps:

[0008] Step 1) obtaining a styrene conservation equation through a reaction process of a styrene polymer polymerization process, and establishing a static molecular weight distribution model of the polymer;

[0009] Step 2) using B-spline function for approximation, establishing a dynamic model of the spline function to calculate the corresponding weight vector, then establishing a state space model of the weight vector and the control input, collecting enough input and output data pairs, and using the recursive least squares identification method to obtain the parameter matrix of the state space model;

[0010] Step 3) For the state space model, combined with external disturbances, a disturbance observer and a dual event trigger mechanism are designed respectively to estimate the unknown disturbance and combine with state feedback to form a PI controller;

[0011] Step 4) Finally, the linear matrix inequality equation is solved by the Lyapunov stability analysis method to solve the various gains and other parameters of the structured controller. At the same time, it is proved that the lower limit of the minimum trigger interval avoids the Zeno phenomenon, realizes the tracking control of the target PDF, and completes the anti-interference control of the styrene polymer polymerization process.

[0012] As a further limitation of the present invention, the reaction mechanism of the styrene polymer in the reactor in step 1) is as follows:

[0013] Chain Initiation:

[0014]

[0015]

[0016] Chain Growth:

[0017]

[0018] Chain transfer:

[0019]

[0020] Chain Termination:

[0021]

[0022] Where I represents the initiator, M is the monomer, i.e. styrene, P is the final polymer, I is the initiator, M 0 is the initial concentration of styrene monomer entering the reactor, I 0 represents the initial concentration of the initiator entering the reactor, the subscripts n, m, and j represent the chain length, and K trm ,K p ,K t ,K d ,K i is the reaction rate constant that conforms to the Arrhenius formula, such as Kd =A d e -Ed / RT , R is an active free radical, R + is the initial active free radical;

[0023] According to the polymerization reaction mechanism, the material conservation equation of styrene can be obtained:

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030] Where θ = V / F is the average residence time of the reactor, V is the volume of the continuous stirred tank reactor CSTR, and F is the total feed flow rate, which can be obtained Represents the total concentration of free radicals and final polymers. At the same time, the changes of free radicals and final polymers in the reaction process can be obtained from the above formula:

[0031]

[0032]

[0033] By setting the parameterized model equation to 0, we can obtain the steady-state solution of the equation:

[0034]

[0035] R j =α -1 R j-1 =α -j+1 R1(j≥2)

[0036]

[0037]

[0038] In the formula Substituting the above equation into the equation, we can get the molecular weight distribution of polymers with a chain length of j in static state, that is, the static molecular weight distribution model:

[0039]

[0040] As a further limitation of the present invention, step 2) is specifically: according to the static molecular weight distribution model, let Pj The distribution shape of a polymer is described by its PDF γ(y,u(t),d(t)):

[0041]

[0042] in represents the polymer concentration P of different chain lengths j j Falls within a pre-set known interval The probability of P j The PDF of the distribution shape is represented by γ, is the control input triggered by the system event, that is, the flow rate of monomer styrene entering the reactor, d(t) represents the external interference, and satisfies:

[0043]

[0044] Where α(t) is the disturbance state, W and V are the relevant parameters of the external disturbance;

[0045] Here we choose to use the square root B-spline function to approximate its function, as follows:

[0046]

[0047] in is the weight, D i (y)(i=1,2,...,n) is a pre-set basis function corresponding to the weight. The basis function is known, and the approximation of the entire spline function to the PDF can be controlled by controlling the weight, thereby controlling the output PDF.

[0048] According to the above square root spline function, it can be transformed into the following model:

[0049]

[0050] Where C0=[D1(y)D2(y)...D n-1 (y)], The weights corresponding to the PDF can be obtained by calculating the above formula. By collecting input and output data pairs, the system parameter matrices A0 and B0 are obtained by recursive least squares; thus, a dynamic model of control input and weights can be established:

[0051]

[0052] where V(t)∈R n-1 are independent weights, x(t) is the system state, H is the given system matrix, and (A0, B0) is observable.

[0053] As a further limitation of the present invention, step 3) is specifically:

[0054] Defining new augmented system variables The augmented system is:

[0055]

[0056] in And define the tracking error e v =V(t)-V g , V g is the target weight vector;

[0057] The dual-event trigger mechanism is adopted, and the corresponding event trigger scheme is designed for the system state and disturbance estimation. According to which of the two events exceeds the set threshold first, that is, the trigger condition is met, and the two events are triggered at the same time;

[0058] The state trigger scheme is designed as follows:

[0059]

[0060] in is the correlation coefficient, δ i ∈[0,1), i=1,2 is the given amount for setting the threshold, Φ i >0,i=1,2 is the pre-set event trigger matrix, a symmetric positive definite free weight matrix of appropriate dimension;

[0061] The interference trigger scheme is designed as follows:

[0062]

[0063] in is the estimated value of the external disturbance and and They are also the current sampling status and interference estimation information. and Is the latest information sent to the controller;

[0064] The next trigger moment is designed according to the following formula:

[0065]

[0066] where t∈[t k ,t k+1 ), is the minimum trigger time lower bound of the augmented system state event trigger at time t+1, is the minimum trigger time lower bound of the interference estimation event at time t+1, t k+1 is the event triggering time at time t+1.

[0067] The disturbance observer design is as follows:

[0068]

[0069] Where z(t) is the designed auxiliary variable, L is the observer gain, and the disturbance estimation error is defined as: Combining the disturbance observation with the PI state feedback controller and considering the system state and disturbance estimation under the event triggering mechanism, the composite control design is:

[0070]

[0071] Where K is the gain of the composite controller.

[0072] As a further limitation of the present invention, step 4) is specifically:

[0073] Assume a parameter λ1>0 and there exists a diagonal matrix P2, R, Q1>0, satisfying the following linear matrix inequality:

[0074]

[0075] In the above linear inequality, R=KQ1, μ>0, It can be deduced that the augmented system is asymptotically stable under the designed composite controller and has good target weight tracking characteristics, where the controller and disturbance observer gains can be obtained by and Calculate and obtain;

[0076] For a dual-event trigger, the lower limit of the minimum trigger time interval is:

[0077]

[0078] in The minimum trigger interval for status event triggering. The minimum trigger interval for the interference event triggering can be verified to prove the existence of the lower limit of the minimum event trigger interval to avoid the Zeno phenomenon.

[0079] As a further limitation of the present invention, in step 1, styrene is used as a monomer, and azobisisobutyronitrile is added to initiate a chemical polymerization reaction in a reactor to obtain the desired styrene final polymer. The initial flow rates of the initiator and the monomer are fixed, and the initiator and the monomer are continuously added. The flow rate of the initiator entering the reactor is constant, and the monomer flow rate is controlled to enter the reactor in the form of a ratio of the unit and the total flow rate. At the same time, a heat solvent is introduced to ensure a constant temperature in the reactor. A flow monitor is added at the reactor inlet to monitor and control the ratio of the total logistics flowing in. Since the initiator flow rate is constant, only the feed flow rate of the monomer styrene needs to be controlled, and finally the final polymer concentration with a chain length of j is mainly controlled by the input.

[0080] Compared with the prior art, the present invention has the following beneficial effects:

[0081] (1) The square root B-spline function approximation method of the present invention is aimed at more accurately approximating the probability density function of the molecular weight distribution output of the styrene polymerization process;

[0082] (2) The present invention is based on the analysis of the reaction mechanism of styrene monomer in the reactor, and has better physical and chemical properties for the final polymer product;

[0083] (3) The present invention constructs a dual event trigger mechanism, introduces an event trigger mechanism into the state of the entire closed-loop system and the designed interference estimation, improves channel communication efficiency, and reduces data redundancy and waste of computing resources;

[0084] (4) The present invention constructs a disturbance observer to observe and estimate the external unknown disturbances that exist in large quantities in actual engineering and to offset and compensate them in the feedforward channel, thereby minimizing the impact of the disturbance on the system control. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0086] Figure 1 It is a flow chart of the present invention.

[0087] Figure 2 It is a schematic diagram of the molecular weight distribution control system of the styrene bulk polymerization process in the present invention.

[0088] Figure 3 This is a dual event trigger diagram of MATLAB / Simulink simulation verification results without using the present invention.

[0089] Figure 4 The 3D distribution diagram is output for the MATLAB / Simulink simulation verification results using the present invention. DETAILED DESCRIPTION

[0090] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0091] like Figure 1 The molecular weight distribution control method of styrene polymerization process based on the dual event synchronous trigger mechanism comprises the following steps:

[0092] Step 1) obtaining the conservation equation of styrene through the reaction process of the styrene polymer polymerization process, and establishing a static molecular weight distribution model of the polymer; the basic process is: taking styrene as a monomer, adding initiator azobisisobutylnitrile to carry out a chemical polymerization reaction in a reactor to obtain the required styrene final polymer, the initial flow of the initiator and the monomer is fixed, and the initiator and the monomer are continuously added, the flow of the initiator entering the reactor is constant, the monomer flow is controlled to enter the reactor in the form of a ratio of the unit and the total flow, and the heat solvent is introduced to ensure the constant temperature in the reactor, and a flow monitoring is added at the reactor inlet to monitor and control the inflow total logistics ratio. Since the initiator flow is constant, only the feed flow of the monomer styrene needs to be controlled, and finally the final polymer concentration with a chain length of j is mainly controlled input;

[0093] The reaction mechanism of styrene polymerization in the reactor is as follows:

[0094] Chain Initiation:

[0095]

[0096]

[0097] Chain Growth:

[0098]

[0099] Chain transfer:

[0100]

[0101] Chain Termination:

[0102]

[0103] Among them, I represents the initiator, M is the monomer, i.e., styrene, and R is the active free radical. + is the initial active free radical. The subscripts n, m, and j represent the chain length, K trm ,K p ,K t ,K d ,K i is the reaction rate constant. The reaction rate constant conforms to the Arrhenius formula, such as K d =A d e -Ed / RT . Starting from the chain initiation, the initiator is K d The reaction decomposes into free active radicals R+ The free active radical combines with the monomer styrene to produce a free radical with a chain length of 1. During the chain growth process, as its fate suggests, the free radical continues to grow as K p The reaction rate is combined with monomer styrene to produce macromolecular free radicals. Chain termination is the polymerization of two free radicals with different chain lengths to produce a final stable polymer. In this process, the free radical activity disappears and the free radical is consumed in the combination. Chain transfer is the reaction between free radicals and monomers with K trm In the reaction rate combination, free radicals are transferred from monomers to destroy and absorb into a part of the final polymer, while the transferred molecules do not have a stable structure, forming new free radicals.

[0104] Further, according to the polymerization reaction mechanism, the material conservation equation of styrene can be obtained:

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111] Where θ = V / F is the average residence time of the reactor. V is the volume of the continuous stirred tank reactor (CSTR), and F is the total feed flow rate. Represents the total concentration of free radicals and final polymers. At the same time, the changes of free radicals and final polymers in the reaction process can be obtained from the above formula:

[0112]

[0113]

[0114] By setting the parameterized model equation to 0, we can obtain the steady-state solution of the equation:

[0115]

[0116] R j =α -1 R j-1 =α -j+1 R1(j≥2)

[0117]

[0118]

[0119] In the formula Substituting the above equation into the equation, we can get the molecular weight distribution of polymer with chain length j at static state:

[0120]

[0121] From the above formula we can see that Therefore, the molecular weight distribution (MWD) control problem under static conditions is transformed into the PDF control problem of chain length j.

[0122] Step 2) Then, since its static model and probability density function (PDF) are normalized, the output static molecular weight distribution control problem can be converted into the output PDF control problem, and then the B-spline function is selected for approximation, and the dynamic model of the spline function is established to calculate the corresponding weight vector, and then the state space model of its weight vector and control input is established. After collecting enough input and output data pairs, the recursive least squares identification method is used to obtain the parameter matrix of the state space model; specifically as follows:

[0123] According to the static molecular weight distribution model, let P j The distribution shape of a polymer is described by its PDF γ(y,u(t),d(t)):

[0124]

[0125] in represents the polymer concentration P of different chain lengths j j Falls within a pre-set known interval The probability of is the control input triggered by the system event, that is, the flow rate of monomer styrene entering the reactor, d(t) represents the external interference, and satisfies:

[0126]

[0127] Here we choose to use the square root B-spline function to approximate its function, as follows:

[0128]

[0129] in is the weight, D i (y)(i=1,2,...,n) is a pre-set basis function corresponding to the weight. The basis function is known. It is only necessary to control the weight to control the approximation of the entire spline function to the PDF, thereby controlling the output PDF. According to the above formula, the square root spline function can be transformed into the following model:

[0130]

[0131] Where C0=[D1(y)D2(y)...D n-1 (y)], Since the probability density function is normalized, must be satisfied, so only n-1 weights are independent, so the weight constraint problem needs to be considered in the modeling process.

[0132] Based on this, the weight vector can be calculated. Further considering the dynamics of the weight vector, the parameter matrix A0, B0 of the system can be obtained by using the subspace identification method; that is, the parameter matrix of the system can be obtained by solving the input and output data using the least squares method; based on the calculated system matrix, the following state space model is established:

[0133]

[0134] where V(t)∈R n-1 are independent weights, x(t) is the system state, H is the given system matrix, and (A0, B0) is observable.

[0135] Step 3) For the state space model, considering external disturbances, effectively reducing the waste of resources caused by data transmission redundancy and other problems in the system control process, designing disturbance observers and dual event triggering mechanisms respectively, realizing the estimation of unknown disturbances and combining them with state feedback to form a PI controller, and adding a dual event triggering mechanism to further reduce the number of system state and disturbance data transmissions, improve transmission efficiency, and save communication channel resources; specifically: define a new augmented variable The augmented system is:

[0136]

[0137] in Define the expected error e v =V(t)-V g , V g is the target weight vector.

[0138] First, the event trigger mechanism used is a non-periodic control mode. The system data transmission is determined by the "event" of the set threshold condition. When the threshold is met, the event is not triggered and the system data will not be transmitted. When the threshold is not met, the event is triggered and data transmission begins. The communication channel is not occupied at other times, which reduces data redundancy and saves computing and channel resources. The present invention adopts a dual-event synchronous trigger mechanism, designs corresponding event trigger schemes for system states and disturbance estimates, and follows the principle of "first come, first served, synchronous triggering", that is, whichever of the two events exceeds the set threshold first, that is, the trigger condition is met, and the two events are triggered simultaneously.

[0139] The state trigger scheme is designed as follows:

[0140]

[0141] in is the correlation coefficient, δ i ∈[0,1), i=1,2 is the given amount for setting the threshold, Φ i >0,i=1,2 is the pre-set event trigger matrix, A symmetric positive definite free weights matrix of appropriate dimension.

[0142] The interference trigger scheme is designed as follows:

[0143]

[0144] in is the estimated value of the external disturbance and and They are also the current sampling status and interference estimation information. and Is the latest information sent to the controller.

[0145] From the above, we can see that the two event trigger design mechanisms are triggered simultaneously. No matter which one exceeds the set threshold first and reaches the trigger condition, the other one will be triggered at the same time. The next trigger moment is designed according to the following formula:

[0146]

[0147] where t∈[t k ,t k+1 ), is the minimum trigger time lower bound of the augmented system state event trigger at time t+1, is the minimum trigger time lower bound of the interference estimation event at time t+1, t k+1 is the event triggering time at time t+1; the disturbance observer design scheme is as follows:

[0148]

[0149] Where z(t) is the designed auxiliary variable and L is the observer gain. The disturbance estimation error is defined as: Combining the disturbance observation with the PI state feedback controller and considering the system state and disturbance estimation under the event triggering mechanism, the composite control design is:

[0150]

[0151] Where K is the gain of the composite controller.

[0152] Step 4) Finally, the Lyapunov stability analysis method and the linear matrix inequality equation are used to solve the various gains and other parameters of the structured controller, and at the same time, it is proved that the minimum trigger interval lower limit avoids the Zeno phenomenon, realizes the tracking control of the target probability density function, completes the anti-interference control of the styrene polymer polymerization process, and avoids resource waste;

[0153] The linear matrix inequalities for solving the controller and observer gains are:

[0154] Theorem 1: Assume a parameter λ1>0, and there exists a diagonal matrix P2, R, Q1>0, satisfying the following matrix inequality:

[0155]

[0156] In the above linear inequality, R=KQ1, μ>0, It can be deduced that the augmented system is asymptotically stable under the designed composite controller and has good target weight tracking characteristics, where the controller and disturbance observer gains can be obtained by and Calculated.

[0157] Proof: Construct the Lyapunov function: According to step 3, we can get:

[0158]

[0159] The inequality According to Schur's complement theorem: in is a normal number, when exist Get Both have: Where P(0) is the initial value of P(t). It can be concluded that the designed closed-loop system can converge to the designed compact set. For the augmented system in the closed-loop system Included The integral term, according to Babalat's theorem, exists as t tends to infinity, then

[0160] Theorem 1: For a two-event trigger, the lower limit of its minimum trigger time interval is:

[0161]

[0162] in The minimum trigger interval for status event triggering. is the minimum trigger interval for interference events. The existence of the minimum event trigger interval lower limit can be verified to avoid the Zeno phenomenon.

[0163] Proof: Order where t∈[t k ,t k+1 ) to:

[0164]

[0165]

[0166] or

[0167] The solution to the above inequality is The upper bounds of To obtain, among which Taking zero on the right side, we can get and Finally, the lower limits of the minimum trigger time of the state trigger mechanism and the interference trigger mechanism are obtained respectively:

[0168] In summary, it can be concluded that no matter which of the two triggers is triggered first, the lower limit of the minimum interval between their triggering times is always greater than zero, so the Zeno phenomenon in event triggering can be avoided.

[0169] In addition, if Figure 3-4 As shown, the above control method is simulated and verified in MATLAB / Simulink, and the results prove that the output distribution obtained by adopting the control method of the present invention is obviously better than the distribution result of adopting the present invention.

[0170] The above embodiments are only used to help understand the method and core idea of ​​the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made to the present invention without departing from the principles of the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

Claims

1. A method for anti-interference control of molecular weight distribution in styrene polymerization process based on a dual event trigger mechanism, characterized in that: The following steps are involved: Step 1) obtains the conservation equation of styrene through the reaction process of the styrene polymer polymerization process, and establishes a static molecular weight distribution model of the polymer. The reaction mechanism of the styrene polymer in the reactor is as follows: Chain Initiation: Chain Growth: Chain transfer: Chain Termination: Wherein, M is the monomer, i.e., styrene, P is the final polymer, I is the initiator, and M 0 is the initial concentration of styrene monomer entering the reactor, I 0 represents the initial concentration of the initiator entering the reactor, the subscripts n, m, and j represent the chain length, and K trm ,K p ,K t ,K d ,K i is the reaction rate constant that conforms to the Arrhenius formula, K d =A d e -EdRT , R is an active free radical, R + is the initial active free radical; According to the polymerization reaction mechanism, the material conservation equation of styrene can be obtained: Where θ = V / F is the average residence time of the reactor, V is the volume of the continuous stirred tank reactor CSTR, and F is the total feed flow rate, which can be obtained Represents the total concentration of free radicals and final polymers. At the same time, the changes of free radicals and final polymers in the reaction process can be obtained from the above formula: By setting the parameterized model equation to 0, we can obtain the steady-state solution of the equation: R j =α -1 R j-1 =α -j+1 R1(j≥2) P2=θ[K trm MR2+K t R1 2 ] In the formula Substituting the above equation into the equation, we can get the molecular weight distribution of polymers with a chain length of j in static state, that is, the static molecular weight distribution model: Step 2) Use B-spline function for approximation, establish a dynamic model of spline function to calculate the corresponding weight vector, then establish the state space model of its weight vector and control input, collect enough input and output data pairs, and use recursive least squares identification method to obtain the parameter matrix of the state space model, specifically: according to the static molecular weight distribution model, let P j The distribution shape of a polymer is described by its PDF γ(y,u(t),d(t)): in represents the polymer concentration P of different chain lengths j j Falls within a pre-set known interval The probability of j The PDF of the distribution shape is represented by γ, is the control input triggered by the system event, that is, the flow rate of monomer styrene entering the reactor, d(t) represents the external interference, and satisfies: Where α(t) is the disturbance state, W and V are the relevant parameters of the external disturbance; Here we choose to use the square root B-spline function to approximate its function, as follows: in is the weight, D i (y)(i=1,2,...,n) is a pre-set basis function corresponding to the weight. The basis function is known, and the approximation of the entire spline function to the PDF can be controlled by controlling the weight, thereby controlling the output PDF. According to the above square root spline function, it can be transformed into the following model: Where C0(y)=[D1(y)D2(y)...D n-1 (y)], The weights corresponding to the PDF can be obtained by calculating the above formula. By collecting input and output data pairs, the system parameter matrices A0 and B0 are obtained by recursive least squares; thus, a dynamic model of control input and weights can be established: where V(t)∈R n-1 are independent weights, x(t) is the system state, H is the given system matrix, (A0, B0) is observable; Step 3) For the state space model, combined with external disturbances, the disturbance observer and the dual event trigger mechanism are designed respectively to estimate the unknown disturbance and combine with the state feedback to form a PI controller, specifically: Defining new augmented system variables Its augmented system is: in And define the tracking error e v =V(t)-V g , V g is the target weight vector; The dual-event trigger mechanism is adopted, and the corresponding event trigger scheme is designed for the system state and disturbance estimation. According to which of the two events exceeds the set threshold first, that is, the trigger condition is met, and the two events are triggered at the same time; The state trigger scheme is designed as follows: in is the correlation coefficient, δ i ∈[0,1), i=1,2 is the given amount for setting the threshold, Φ i >0,i=1,2 is the pre-set event trigger matrix, a symmetric positive definite free weight matrix of appropriate dimension; The interference trigger scheme is designed as follows: in is the estimated value of the external disturbance and and They are also the current sampling status and interference estimation information. and Is the latest information sent to the controller; The next trigger moment is designed according to the following formula: where t∈[t k ,t k+1 ), is the minimum trigger time lower bound of the augmented system state event trigger at time t+1, is the minimum trigger time lower bound of the interference estimation event at time t+1, t k+1 is the event triggering time at time t+1; The disturbance observer design is as follows: Where z(t) is the designed auxiliary variable, L is the observer gain, and the disturbance estimation error is defined as: Combining the disturbance observation with the PI state feedback controller and considering the system state and disturbance estimation under the event triggering mechanism, the composite control design is: Where K is the gain of the composite controller; Step 4) Finally, the linear matrix inequality equation is solved by the Lyapunov stability analysis method to solve the various gains and other parameters of the structured controller. At the same time, it is proved that the minimum trigger interval lower limit avoids the Zeno phenomenon, realizes the tracking control of the target PDF, and completes the anti-interference control of the styrene polymer polymerization process. Specifically: Assume a parameter λ1>0 and there exists a diagonal matrix The following linear matrix inequality is satisfied: In the above linear inequality, R=KQ1, μ>0, It can be deduced that the augmented system is asymptotically stable under the designed composite controller and has good target weight tracking characteristics, where the controller and disturbance observer gains can be expressed by K = RQ1 -1 and L = P2 -1 X is calculated; For a dual-event trigger, the lower limit of the minimum trigger time interval is: in The minimum trigger interval for status event triggering. The minimum trigger interval for the interference event triggering can be verified to prove the existence of the lower limit of the minimum event trigger interval to avoid the Zeno phenomenon.

2. The anti-interference control method for molecular weight distribution of styrene polymerization process based on dual event trigger mechanism according to claim 1, characterized in that: In step 1), styrene is used as a monomer, and azobisisobutylnitrile is added to initiate a chemical polymerization reaction in a reactor to obtain the desired styrene final polymer. The initial flow rates of the initiator and the monomer are fixed, and the initiator and the monomer are continuously added. The flow rate of the initiator entering the reactor is constant, and the monomer flow rate is controlled to enter the reactor in the form of a ratio of a unit flow rate to a total flow rate. At the same time, a heat solvent is introduced to ensure that the temperature in the reactor is constant. A flow monitor is added at the reactor inlet to monitor and control the ratio of the inflowing total logistics. Since the initiator flow rate is constant, only the feed flow rate of the monomer styrene needs to be controlled, and finally the final polymer concentration with a chain length of j is mainly controlled by the input.

Citation Information

Patent Citations

  • Predictive functional control-based molecular weight output PDF control method

    CN105676637A

  • Styrene bulk polymerization anti-interference distribution shape control method based on interference observer

    CN112684707A