Polyolefin polymerization reaction kinetic parameter correction method, device, equipment and medium
By constructing a polyolefin production model and using intelligent optimization algorithms to automatically correct polymerization reaction kinetic parameters, the inefficiency and poor accuracy caused by relying on human experience in existing technologies have been solved, achieving efficient and robust parameter correction and model prediction.
Patent Information
- Application Number
- CN202511705591.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-20
- Publication Date
- 2026-02-13
AI Technical Summary
In existing technologies, the correction of kinetic parameters for polyolefin polymerization reactions relies on human experience, which leads to low efficiency, easy getting trapped in local optima, poor model prediction accuracy and poor applicability, and difficulty in maintaining robustness under different operating conditions.
By constructing a polyolefin production model based on steady-state actual production data and product quality information from a pre-set polyolefin production unit, and utilizing intelligent optimization algorithms such as autocovariance matrix adaptive evolution strategy, differential evolution algorithm, and particle swarm optimization algorithm, the polymerization reaction kinetic parameters are automatically corrected, and an objective function and iterative optimization module are constructed to achieve efficient parameter correction.
It improves the model's prediction accuracy and applicability, reduces reliance on human experience, enables rapid and automated parameter correction, and enhances the model's robustness and industrial applicability.
Smart Images
Figure CN121528340A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of computer, in particular to a polyolefin polymerization reaction kinetics parameter correction method, device, equipment and medium. BACKGROUND
[0002] Polyolefin is an important synthetic resin, and its production process involves complex polymerization reaction mechanism, such as chain initiation, chain growth, chain transfer and chain termination, and multiple elementary reaction steps. However, due to the complexity, multiphase and inherent difficulty of experimental measurement of the polyolefin polymerization reaction system, these key kinetic parameters are often difficult to accurately obtain by direct experimental means, making it difficult to establish an accurate production model.
[0003] At present, the existing related parameter correction scheme mainly relies on the manual experience of engineers for 'trial and error' adjustment, or uses some simple numerical optimization methods, which makes: 1) low efficiency, when facing complex optimization problems with multiple parameters and multiple objectives, manual adjustment is difficult to meet the requirements; 2) easy to fall into local optimal solution, and cannot guarantee to find the global optimal parameter combination, resulting in poor prediction accuracy of the model; 3) poor applicability, and it is difficult to ensure the prediction robustness of the model under different working conditions. SUMMARY
[0004] Therefore, the purpose of the present application is to provide a polyolefin polymerization reaction kinetics parameter correction method, device, equipment and medium, which can realize efficient and automatic parameter correction under the condition of greatly reducing the dependence on manual experience, and improve the prediction accuracy, applicability and robustness of the model. The specific scheme is as follows:
[0005] In the first aspect, the present application provides a polyolefin polymerization reaction kinetics parameter correction method, comprising:
[0006] Based on a preset polyolefin production device, the target attribute values of the polyolefin products corresponding to a plurality of different grades are determined based on the steady-state actual production data and the actual product quality information when producing the polyolefin products corresponding to the plurality of different grades; wherein the grade is used to represent the production condition, the polyolefin product includes homopolymer, random copolymer and block copolymer, and the attribute includes several of weight average molecular weight, number average molecular weight, molecular weight distribution, melt index and product density;
[0007] After building a computing environment, a polyolefin production model describing the polyolefin production process corresponding to a plurality of grades is constructed, and the polymerization reaction kinetics parameters in the polyolefin production model are initialized; the polymerization reaction kinetics parameters include pre-exponential factor, activation energy and reaction order;
[0008] construct a target function based on the preset parameter optimization target information, the steady-state actual production data, the actual product quality information, and the target attribute value;
[0009] determine a target polyolefin production model based on the preset parameter correction framework, the target function, the polyolefin production model, and a preset iterative optimization termination condition;
[0010] based on the target polyolefin production model, trigger a target polyolefin product attribute value prediction operation when the verification passes.
[0011] Optionally, the steady-state actual production data and the actual product quality information of the preset polyolefin production device when producing a plurality of different grades of polyolefin products are used to determine the target attribute value of the polyolefin product corresponding to each grade, including:
[0012] Obtain the steady-state actual production data of the preset polyolefin production device when producing a plurality of different grades of polyolefin products; wherein the steady-state actual production data includes catalyst feed flow, cocatalyst feed flow, propylene and comonomer feed flow, hydrogen feed flow, reactor temperature and reactor pressure;
[0013] Sample the polyolefin powder corresponding to each grade to determine the sampling result;
[0014] Determine the actual product quality information of the polyolefin product corresponding to each grade based on the sampling result and gel permeation chromatography; the actual product quality information includes weight average molecular weight, number average molecular weight, molecular weight distribution and product density.
[0015] Optionally, before constructing the polyolefin production model describing the polyolefin production process corresponding to a plurality of grades, it further includes:
[0016] For any type of catalyst, collect product powder corresponding to a plurality of grades, and perform gel permeation chromatography analysis based on the product powder to determine the corresponding number of active sites of the catalyst.
[0017] Optionally, after building a computing environment, constructing a polyolefin production model describing the polyolefin production process corresponding to a plurality of grades, and initializing the polyolefin production model, the polyolefin production model includes:
[0018] After building a target database required for a preset parameter correction framework and installing a preset chemical simulation software, a polyolefin production model is constructed based on the preset chemical simulation software or a preset programming language and the number of active sites of the catalyst; the polyolefin production model is a model including a plurality of polymerization reaction mechanisms in a polyolefin production process corresponding to the grades;
[0019] An initial value corresponding to a polymerization reaction kinetics parameter in the polyolefin production model is configured to complete a parameter initialization operation;
[0020] Based on the initial value, a corresponding parameter search range is determined.
[0021] Optionally, the target function is constructed based on the preset parameter optimization target information, the steady-state actual production data, the actual product quality information and the target attribute value, including:
[0022] Preset parameter optimization target information is obtained; wherein the preset parameter optimization target information includes attribute prediction error thresholds corresponding to the weight average molecular weight, the number average molecular weight and the product density, respectively, and the attribute prediction error threshold is an attribute prediction error threshold of the polyolefin production model;
[0023] Weight coefficients corresponding to the weight average molecular weight, the number average molecular weight and the product density are obtained;
[0024] Based on the weight coefficients, the target attribute values, the steady-state actual production data, the actual product quality information and the polyolefin production model, a target function is constructed; the target function is used to determine the total deviation between the predicted attribute values output by the polyolefin production model and the corresponding target attribute values.
[0025] Optionally, the target polymerization reaction kinetics parameter combination is determined based on the preset parameter correction framework, the target function, the polyolefin production model and a preset iteration optimization termination condition, including:
[0026] A preset parameter correction framework is obtained; the preset parameter correction framework includes one or more combinations of a self-covariance matrix adaptive evolution strategy, a differential evolution algorithm and a particle swarm optimization algorithm;
[0027] Based on the preset parameter correction framework and the parameter search range, population initialization is performed to obtain an initialization result; the initialization result includes a crossover probability and an initial individual number, and the individual is a polymerization reaction kinetics parameter group;
[0028] Based on the corresponding polyolefin production model, each individual is subjected to parallel model simulation to determine a simulation result of the current iteration round;
[0029] Based on the simulation results and the objective function, determine the target total error corresponding to each grade in the current iteration round;
[0030] Based on the total target error and the preset parameter optimization target information, individual screening and population update are performed to obtain the updated population, and the set of polymerization reaction kinetic parameters that do not exceed any of the attribute prediction error thresholds are recorded.
[0031] Based on the updated population, the process jumps back to the step of performing parallel model simulations on each individual based on the corresponding polyolefin production model until the preset iterative optimization termination condition is met. Then, the target polymerization reaction kinetic parameter combination is selected from the recorded sets of polymerization reaction kinetic parameters.
[0032] Optionally, the step of parameter calibration of the polyolefin production model based on the target polymerization reaction kinetic parameter combination, and model validation using the corresponding calibrated model, includes:
[0033] The polyolefin production model is calibrated based on the target polymerization reaction kinetic parameter combination to determine the calibrated model.
[0034] Based on the corrected model and the steady-state actual production data of other grades that did not participate in the iterative optimization, the attribute prediction results are determined.
[0035] Based on the attribute prediction results and the preset parameter optimization target information, it is determined whether the corrected model has passed the verification, so as to determine the model verification result.
[0036] Secondly, this application provides a device for correcting the kinetic parameters of a polyolefin polymerization reaction, comprising:
[0037] The target attribute value determination module is used to determine the target attribute value of each polyolefin product corresponding to a certain grade based on the steady-state actual production data and actual product quality information of a preset polyolefin production device when producing multiple polyolefin products corresponding to different grades; wherein, the grade is used to represent the production condition, the polyolefin product includes homopolymer, random copolymer and block copolymer, and the attribute includes several of the following: weight average molecular weight, number average molecular weight, molecular weight distribution, melt index and product density;
[0038] The production model construction module is used to construct a polyolefin production model describing the production process of multiple polyolefin grades after setting up the computing environment, and to initialize the polymerization reaction kinetic parameters in the polyolefin production model; the polymerization reaction kinetic parameters include pre-exponential factor, activation energy, and reaction order;
[0039] The target function construction module is configured to construct a target function based on preset parameter optimization target information, the steady-state actual production data, the actual product quality information, and the target attribute value;
[0040] The iterative optimization module is configured to determine a target polyolefin polymerization kinetics parameter combination based on a preset parameter correction framework, the target function, the polyolefin production model, and a preset iterative optimization termination condition.
[0041] The parameter correction module is configured to perform parameter correction on the polyolefin production model based on the target polyolefin polymerization kinetics parameter combination, and perform model verification using the corresponding corrected model, so as to trigger an attribute value prediction operation of a target polyolefin product based on the corrected model when the verification is passed.
[0042] In a third aspect, the present application provides an electronic device, comprising:
[0043] A memory configured to save a computer program;
[0044] A processor configured to execute the computer program to implement the steps of the polyolefin polymerization kinetics parameter correction method described above.
[0045] In a fourth aspect, the present application provides a computer readable storage medium configured to save a computer program, which is executed by a processor to implement the steps of the polyolefin polymerization kinetics parameter correction method described above.
[0046] It can be seen that, in the present application, based on the preset polyolefin production device, the steady-state actual production data and the actual product quality information of the polyolefin products corresponding to a plurality of different brands in the production process are determined, and the target attribute value of the polyolefin product corresponding to each brand is determined; wherein, the brand is used to represent the production condition, the polyolefin product includes homopolymer, random copolymer and block copolymer, and the attribute includes several of weight average molecular weight, number average molecular weight, molecular weight distribution, melt index and product density; after building the computing environment, the polyolefin production model describing the polyolefin production process corresponding to a plurality of brands is constructed, and the polymerization reaction kinetics parameters in the polyolefin production model are initialized; the polymerization reaction kinetics parameters include pre-exponential factor, activation energy and reaction order; based on the preset parameter optimization target information, the steady-state actual production data, the actual product quality information and the target attribute value, the objective function is constructed; based on the preset parameter correction framework, the objective function, the polyolefin production model and the preset iterative optimization termination condition, the target polymerization reaction kinetics parameter combination is determined; based on the target polymerization reaction kinetics parameter combination, the parameter correction of the polyolefin production model is carried out, and the corresponding corrected model is used for model verification, so that when the verification is passed, based on the corrected model, the attribute value prediction operation corresponding to the target polyolefin product is triggered. That is, in the present application, based on the steady-state actual production data and the actual product quality information of the polyolefin products corresponding to a plurality of different brands in the actual production process, the target attribute value is determined, then the polyolefin production model describing the polyolefin production process corresponding to a plurality of brands is constructed, and the polymerization reaction kinetics parameters in the model are initialized, then the objective function is constructed based on the preset parameter optimization target information and the target attribute value, the target polymerization reaction kinetics parameter combination is determined based on the preset parameter correction framework, the objective function, the polyolefin production model and the preset iterative optimization termination condition, and then the parameter correction of the polyolefin production model is carried out based on the target polymerization reaction kinetics parameter combination, and when the corrected model is verified, the attribute value prediction operation corresponding to the target polyolefin product is triggered based on the model. In this way, under the condition of greatly reducing the dependence on artificial experience, efficient and automatic parameter correction can be realized, and the prediction accuracy, applicability and robustness of the model can be improved. BRIEF DESCRIPTION OF DRAWINGS
[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only embodiments of the present application, and those skilled in the art can obtain other drawings according to the provided drawings without creating any inventive labor.
[0048] Figure 1A polyolefin polymerization reaction kinetics parameter correction method flow chart is provided for the present application;
[0049] Figure 2 A specific polyolefin polymerization reaction kinetics parameter correction method flow chart is provided for the present application;
[0050] Figure 3 An iteration optimization process flow diagram based on a preset parameter correction framework is provided for the present application;
[0051] Figure 4 A convergence diagram of the iteration optimization process based on the preset parameter correction framework is provided for the present application;
[0052] Figure 5 A polyolefin polymerization reaction kinetics parameter correction device structure diagram is provided for the present application;
[0053] Figure 6 An electronic device structure diagram is provided for the present application. DETAILED DESCRIPTION
[0054] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.
[0055] At present, the existing related parameter correction schemes mostly rely on the manual experience of engineers for 'trial and error' adjustment or adopt some simple numerical optimization methods, which makes: 1) low efficiency, when facing complex optimization problems with multiple parameters and multiple objectives, manual adjustment is difficult to meet the requirements; 2) easy to fall into local optimal solution, unable to guarantee to find the global optimal parameter combination, resulting in poor accuracy of model prediction; 3) poor applicability, difficult to guarantee the prediction robustness of the model under different working conditions.
[0056] Therefore, the present application provides a polyolefin polymerization reaction kinetics parameter correction scheme, which can realize efficient and automatic parameter correction under the condition of greatly reducing the dependence on manual experience, and improve the prediction accuracy, applicability and robustness of the model.
[0057] Referring to Figure 1 The embodiments of the present application disclose a polyolefin polymerization reaction kinetics parameter correction method, which comprises:
[0058] In step S11, based on the preset polyolefin production device, the steady-state actual production data when producing polyolefin products corresponding to a plurality of different brands, and the actual product quality information, the target attribute value of the polyolefin product corresponding to each brand is determined; wherein the brand is used to represent the production condition, the polyolefin product includes homopolymer, random copolymer and block copolymer, and the attribute includes several of weight average molecular weight, number average molecular weight, molecular weight distribution, melt index and product density.
[0059] In combination Figure 2 As shown in the embodiment, taking the production process of polypropylene (a typical polyolefin) as an example, first, data needs to be collected, that is, the steady-state actual production data of the preset polyolefin production device when producing polyolefin products corresponding to a plurality of different brands is obtained; wherein the steady-state actual production data includes catalyst feed flow, catalyst promoter feed flow, propylene and comonomer feed flow, hydrogen feed flow, reactor temperature and reactor pressure; the polyolefin powder corresponding to each brand is sampled to determine the sampling result; based on the sampling result and gel permeation chromatography (GPC), the actual product quality information of the polyolefin product corresponding to each brand is determined; the actual product quality information includes weight average molecular weight (Weight-average Molecular Weight, abbreviated as Mw), number average molecular weight (Number-average Molecular Weight, abbreviated as Mn), molecular weight distribution (Molecular Weight Distribution, abbreviated as MWD) and product density.
[0060] Further, for the working conditions from the same catalyst, by GPC analysis of the product powder of the corresponding different brands, the active site number of the catalyst can be obtained, that is, for any type of catalyst, a plurality of product powders corresponding to the brand are collected, and gel permeation chromatography analysis is performed based on the product powder to determine the corresponding catalyst active site number.
[0061] In step S12, after building the computing environment, a polyolefin production model describing the polyolefin production process corresponding to a plurality of brands is constructed, and the polymerization reaction kinetics parameters in the polyolefin production model are initialized; the polymerization reaction kinetics parameters include pre-exponential factor, activation energy and reaction order.
[0062] In this embodiment, after collecting the data, a computing environment supporting the scheme of this embodiment is built, and a polyolefin production model is established according to the collected data, that is, after building a target database required for a preset parameter correction framework and installing a preset chemical simulation software, a polyolefin production model is constructed based on the preset chemical simulation software or a preset programming language and the number of active sites of the catalyst; the polyolefin production model is a model containing a plurality of polymerization reaction mechanisms in a polyolefin production process corresponding to the grades; initial values corresponding to the kinetic parameters of the polymerization reaction in the polyolefin production model are configured to complete a parameter initialization operation; and a corresponding parameter search range is determined based on the initial values.
[0063] It can be understood that the search range of the parameter is set relative to the initial value thereof: the search range of the pre-exponential factor can be , the search range of the activation energy can be , and the search range of the reaction order can be .
[0064] It should be understood that, regarding the building of the computing environment, in a specific implementation, a computer or server installed with Python (version 3.x) or MATLAB (version R2013a or later) can be selected to configure the required computing libraries, including specific libraries for data processing, numerical calculation, parallel processing, and intelligent optimization algorithms.
[0065] It can be understood that, if the model is to be established in a commercial chemical simulation software such as Aspen Plus, it is necessary to ensure that the corresponding genuine simulation software suite is installed in the running environment and the interface is configured. The Aspen Plus commercial simulation software contains the polymer kinetics equation of the system, and in this embodiment, by building an optimization algorithm and a COM interface through Python, automatic and rapid calling of the Aspen Plus software can be realized.
[0066] Further, regarding the establishment of a polyolefin production model, taking polypropylene as an example, a polypropylene polymerization reaction mechanism model containing multiple grades can be established in a chemical engineering simulation software or through a mathematical programming language. The core of the model is the polymerization reaction kinetics, which generally includes chain initiation, chain growth, chain transfer to monomer / hydrogen, and chain termination, etc. According to the catalyst characteristics and reaction mechanism, the model can identify the key kinetic parameters that need to be corrected, mainly the pre-exponential factor and activation energy of each elementary reaction. The initial values of these parameters can be preliminarily set according to laboratory data or engineering experience. Among them, the grades are HP648T, MT50, MT60, RP340R, which include both homopolymerization and copolymerization production conditions. The catalyst characteristics are obtained by GPC analysis of the product powders of different grades in step S11, and the HR catalyst has three active sites.
[0067] In step S13, a target function is constructed based on the preset parameter optimization target information, the steady-state actual production data, the actual product quality information, and the target attribute value.
[0068] In this embodiment, after the model is constructed, an optimization problem is constructed, taking the reaction kinetic parameters as optimization variables, and taking the minimum total deviation between the model predicted attribute value and the target attribute value as the optimization target, that is, obtaining preset parameter optimization target information; wherein the preset parameter optimization target information includes attribute prediction error thresholds corresponding to the weight average molecular weight, the number average molecular weight, and the product density, respectively, and the attribute prediction error threshold is the attribute prediction error threshold of the polyolefin production model; obtaining weight coefficients corresponding to the weight average molecular weight, the number average molecular weight, and the product density, respectively; based on the weight coefficients, the target attribute value, the steady-state actual production data, the actual product quality information, and the polyolefin production model, a target function is constructed; the target function is used to determine the total deviation between the predicted attribute value output by the polyolefin production model and the corresponding target attribute value.
[0069] Specifically, in this embodiment, according to the industrial requirements, the optimization target is set, that is, the preset parameter optimization target information. For example, it is required that the prediction error of the corrected model for the key product attributes satisfies: the relative error of the density <2%, and the relative error of the molecular weight <5%. Based on this, a target function (total deviation F) of the following form is constructed:
[0070] .
[0071] In the formula, denotes the product density, MW denotes the molecular weight (including the weight average molecular weight and the number average molecular weight), and the subscript sim denotes the model simulation value; the subscript act denotes the actual value collected in step S11, that is, the target attribute value. and This is a weighting coefficient that can be adjusted according to the importance of different attributes. When multiple grades are involved in optimization, the total deviation F is the sum of the deviations of all grades.
[0072] Step S14: Based on the preset parameter correction framework, the objective function, the polyolefin production model, and the preset iterative optimization termination condition, determine the target polymerization reaction kinetic parameter combination.
[0073] In this embodiment, an intelligent optimization algorithm is selected and configured to obtain a preset parameter correction framework. The intelligent optimization algorithm is selected from one or more combinations of covariance matrix adaptive evolution strategy, differential evolution algorithm, particle swarm optimization algorithm, or other optimization algorithms, or other types of intelligent optimization algorithms. Multi-grade collaborative optimization is performed, incorporating data from multiple different grades (operating conditions) into the optimization process to find a unified parameter set that can simultaneously meet the accuracy requirements of multiple scenarios, thereby enhancing the robustness and versatility of the model. Combined with... Figure 3 As shown, the process involves: obtaining a preset parameter correction framework; the preset parameter correction framework includes one or more combinations of autocovariance matrix adaptive evolution strategy, differential evolution algorithm, and particle swarm optimization algorithm; initializing the population based on the preset parameter correction framework and the parameter search range to obtain initialization results; the initialization results include crossover probability and initial number of individuals, where each individual is a set of polymerization reaction kinetic parameters; performing parallel model simulations on each individual based on the corresponding polyolefin production model to determine the simulation results of the current iteration round; determining the target total error corresponding to each grade in the current iteration round based on the simulation results and the objective function; performing individual screening and population update based on the target total error and the preset parameter optimization target information to obtain an updated population, and recording the polymerization reaction kinetic parameter sets that do not exceed any attribute prediction error threshold; based on the updated population, returning to the step of performing parallel model simulations on each individual based on the corresponding polyolefin production model until the preset iteration optimization termination condition is met, and then selecting the target polymerization reaction kinetic parameter combination from the recorded sets of polymerization reaction kinetic parameters. It is understood that the initialization results may also include optimization algorithm hyperparameters. The types of hyperparameters will vary depending on the selected intelligent optimization algorithm. For example, evolutionary algorithms may have mutation factors in their hyperparameters, while non-evolutionary algorithms may not have mutation factors.
[0074] It should be noted that in the embodiment, parallel computing technology is used to simultaneously process model calculations of multiple polyolefin grades or multiple operating conditions to improve optimization efficiency. Moreover, optimization calculation (solving the optimization problem of model parameters using intelligent optimization algorithm) and model verification (verifying the prediction accuracy of the optimized model) work together to realize automatic optimization of model parameters and accuracy verification. In a specific embodiment, the convergence process of the above optimization can be as shown in Figure 4 .
[0075] In a specific embodiment, the process of selecting differential evolution algorithm to optimize and correct the kinetic parameters can be as follows:
[0076] (1) Initialization: Set the population size (e.g. 50), mutation factor f (e.g. 0.7), crossover probability CR (e.g. 0.9) and other hyperparameters. Randomly generate an initial population (i.e. multiple sets of kinetic parameter combinations) within the reasonable search range of parameters set in step S12;
[0077] (2) Iterative optimization: Enter the main loop. For each individual (i.e. each set of parameters) in the population, assign it to an independent model calculation process through parallel computing. The process updates the set of parameters to the polypropylene model, runs the simulation under the corresponding operating condition, and obtains the predicted product property values;
[0078] (3) Evaluation and selection: Calculate the objective function value (total deviation F) of the individual according to the simulation results. The algorithm generates a new generation of population according to the mutation, crossover and selection rules, and retains the parameter combination with smaller total deviation. This process is repeated continuously, and the population as a whole evolves towards the direction of smaller total deviation. At the same time, record the parameter combination that meets all the single error threshold values.
[0079] (4) Result acquisition and verification: Set the iteration stopping conditions, such as reaching the maximum evolution generation (e.g. 500 generations) or the objective function value does not decrease significantly for consecutive generations. After optimization, select the set of parameters with the smallest total error from the recorded parameter combinations that meet the convergence conditions as the final optimal parameter combination.
[0080] Step S15, based on the target polyolefin production model, the parameter correction is performed based on the target polyolefin production model, and the corresponding corrected model is used for model verification, so that when the verification is passed, the target polyolefin product corresponding to the attribute value prediction operation is triggered based on the corrected model.
[0081] In the embodiment, after obtaining the target polymerization kinetics parameter combination, the target polymerization kinetics parameter combination is applied to the production model, and verification is performed to determine whether the correction is successful, that is, the polyolefin production model is parameter-corrected based on the target polymerization kinetics parameter combination to determine a corrected model; based on the corrected model and the steady-state actual production data corresponding to other grades that do not participate in the iterative optimization, attribute prediction results are determined; based on the attribute prediction results and the preset parameter optimization target information, it is judged whether the corrected model passes the verification to determine a model verification result.
[0082] It should be understood that, in the embodiment, after the target polymerization kinetics parameter combination is applied to the model, another set of production data that does not participate in the foregoing optimization process is required for model verification. In the embodiment, after optimization and correction, the average error of product density prediction of the model for multiple grades can be reduced to 1.5%, and the average error of molecular weight prediction can be reduced to 1%, as shown in FIG. 2, both of which meet the preset industrial application requirements, proving that the scheme in the embodiment is effective and feasible. Figure 4
[0083] In summary, the scheme in the embodiment has the beneficial effects shown in the following:
[0084] 1) Significantly improve the model precision: through the global search capability of the intelligent optimization algorithm, the local optimal solution can be effectively avoided, and the optimal kinetics parameter combination can be found, so that the model prediction error is reduced to an acceptable level for the polyolefin industry;
[0085] 2) Enhance the model versatility: through the collaborative optimization of multiple grades (working conditions), the parameter set obtained can be applied to multiple polyolefin grades and production working conditions, greatly improving the industrial applicability and robustness of the model;
[0086] 3) Improve the correction efficiency: compared with the traditional inefficient manual parameter adjustment, the automatic and parallel optimization process of the scheme can shorten the model correction time from several weeks to several hours, thereby greatly improving the research and engineering efficiency;
[0087] 4) Support process optimization and innovation: a high-fidelity model that has been accurately corrected is a reliable digital basis for production process optimization, product quality control, fault diagnosis, and development of new products and new grades.
[0088] It can be seen that, in the present application, the target attribute value is determined based on the steady-state actual production data and actual product quality information of the polyolefin products corresponding to a plurality of different brands in the actual production process, then a polyolefin production model describing the polyolefin production process corresponding to a plurality of brands is constructed, and the polymerization reaction kinetics parameters in the model are initialized, then a target function is constructed based on the preset parameter optimization target information and the target attribute value, the target polymerization reaction kinetics parameter combination is determined based on the preset parameter correction framework, the target function, the polyolefin production model and the preset iterative optimization termination condition, after that, the polyolefin production model is parameter corrected based on the target polymerization reaction kinetics parameter combination, and when the corrected model is verified, the attribute value prediction operation corresponding to the target polyolefin product is triggered based on the model. In this way, efficient and automatic parameter correction can be realized under the condition of greatly reducing the dependence on artificial experience, and the prediction accuracy, applicability and robustness of the model are improved.
[0089] Referring to Figure 5 The present application also discloses a polyolefin polymerization reaction kinetics parameter correction device, as shown in the accompanying drawings, comprising:
[0090] The target attribute value determination module 11 is configured to determine the target attribute value of the polyolefin product corresponding to each brand based on the steady-state actual production data and actual product quality information when producing a plurality of polyolefin products corresponding to different brands in a preset polyolefin production device; wherein the brand represents the production condition, the polyolefin product includes homopolymer, random copolymer and block copolymer, and the attribute includes several of weight average molecular weight, number average molecular weight, molecular weight distribution, melt index and product density;
[0091] The production model construction module 12 is configured to construct a polyolefin production model describing the polyolefin production process corresponding to a plurality of brands after building a computing environment, and initialize the polymerization reaction kinetics parameters in the polyolefin production model; the polymerization reaction kinetics parameters include pre-exponential factor, activation energy and reaction order;
[0092] The target function construction module 13 is configured to construct a target function based on the preset parameter optimization target information, the steady-state actual production data, the actual product quality information and the target attribute value;
[0093] The iterative optimization module 14 is configured to determine a target polymerization reaction kinetics parameter combination based on the preset parameter correction framework, the target function, the polyolefin production model and the preset iterative optimization termination condition;
[0094] The parameter correction module 15 is configured to perform parameter correction on the polyolefin production model based on the target set of polyolefin reaction kinetics parameters, and perform model verification using the corresponding corrected model, so as to trigger a target polyolefin product attribute value prediction operation based on the corrected model when the verification is passed.
[0095] In some embodiments, the target attribute value determination module 11 can be specifically configured to: obtain steady-state actual production data of a preset polyolefin production device in production of a plurality of polyolefin products corresponding to different grades; wherein the steady-state actual production data includes catalyst feed flow, cocatalyst feed flow, propylene and comonomer feed flow, hydrogen feed flow, reactor temperature and reactor pressure; sample polyolefin powders corresponding to each of the grades to determine sampling results; determine actual product quality information of the polyolefin products corresponding to each of the grades based on the sampling results and gel permeation chromatography; and the actual product quality information includes weight average molecular weight, number average molecular weight, molecular weight distribution and product density.
[0096] In some embodiments, the polyolefin polymerization reaction kinetics parameter correction device can be further configured to: collect a plurality of product powders corresponding to each of the grades for any type of catalyst, and perform gel permeation chromatography analysis based on the product powders to determine the corresponding number of catalyst active sites.
[0097] In some embodiments, the production model construction module 12 can be specifically configured to: after building a target database required for a preset parameter correction framework and installing a preset chemical simulation software, construct a polyolefin production model based on the preset chemical simulation software or a preset programming language and the number of catalyst active sites; the polyolefin production model is a model including polymerization reaction mechanisms in polyolefin production processes corresponding to each of the grades; configure initial values of polymerization kinetics parameters in the polyolefin production model to complete a parameter initialization operation; and determine a corresponding parameter search range based on the initial values.
[0098] In some embodiments, the target function construction module 13 can be specifically configured to: obtain preset parameter optimization target information, wherein the preset parameter optimization target information comprises attribute prediction error thresholds corresponding to the weight average molecular weight, the number average molecular weight and the product density respectively, and the attribute prediction error thresholds are attribute prediction error thresholds of the polyolefin production model; obtain weight coefficients corresponding to the weight average molecular weight, the number average molecular weight and the product density respectively; and construct a target function based on the weight coefficients, the target attribute values, the steady-state actual production data, the actual product quality information and the polyolefin production model, so as to determine total deviations between predicted attribute values output by the polyolefin production model and the corresponding target attribute values.
[0099] In some embodiments, the iterative optimization module 14 can be specifically configured to: obtain a preset parameter correction framework, wherein the preset parameter correction framework comprises a combination of one or more of a self-covariance matrix adaptive evolution strategy, a differential evolution algorithm and a particle swarm optimization algorithm; perform population initialization based on the preset parameter correction framework and the parameter search range to obtain an initialization result, wherein the initialization result comprises a crossover probability and an initial number of individuals, and the individuals are polycondensation reaction kinetics parameter groups; perform parallel model simulation on each of the individuals based on the corresponding polyolefin production model to determine a simulation result of a current iteration round; determine target total errors corresponding to each of the grades in the current iteration round based on the simulation result and the target function; perform individual screening and population updating based on the target total errors and the preset parameter optimization target information to obtain an updated population, and record the polycondensation reaction kinetics parameter groups that do not exceed any of the attribute prediction error thresholds; and based on the updated population, jump back to the step of performing parallel model simulation on each of the individuals based on the corresponding polyolefin production model until a preset iterative optimization termination condition is met, and then select a target polycondensation reaction kinetics parameter combination from the recorded several polycondensation reaction kinetics parameter groups.
[0100] In some embodiments, the parameter correction module 15 can be specifically configured to: perform parameter correction on the polyolefin production model based on the target polycondensation reaction kinetics parameter combination to determine a corrected model; determine attribute prediction results based on the corrected model and steady-state actual production data corresponding to other grades that do not participate in iterative optimization; and determine a model verification result by judging whether the corrected model passes verification based on the attribute prediction results and the preset parameter optimization target information.
[0101] Further, the embodiments of the present application also disclose an electronic device, Figure 6is a structural diagram of an electronic device 20 according to an exemplary embodiment, and the contents in the figure should not be considered as any limitation on the use scope of the present application.
[0102] Figure 6 A structural diagram of an electronic device 20 is provided in the present embodiment. The electronic device 20 can specifically include at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 is configured to store a computer program, which is loaded and executed by the processor 21 to implement the related steps in the polyolefin polymerization kinetics parameter correction method disclosed in any of the preceding embodiments. In addition, the electronic device 20 in the present embodiment can be specifically an electronic computer.
[0103] In the present embodiment, the power supply 23 is configured to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 is capable of creating a data transmission channel between the electronic device 20 and external devices, and the communication protocol followed by the communication interface 24 can be any communication protocol applicable to the technical solution of the present application, which is not specifically limited herein; the input / output interface 25 is configured to obtain external input data or output data to the outside, and the specific interface type can be selected according to the specific application needs, which is not specifically limited herein.
[0104] In addition, the memory 22 as a carrier for resource storage can be a read-only memory, a random access memory, a magnetic disk, or an optical disk, etc., and the resources stored thereon can include an operating system 221, a computer program 222, etc., and the storage mode can be temporary storage or permanent storage.
[0105] The operating system 221 is configured to manage and control each hardware device on the electronic device 20 and the computer program 222, and can be Windows Server, Netware, Unix, Linux, etc. In addition to the computer program capable of completing the polyolefin polymerization kinetics parameter correction method executed by the electronic device 20 disclosed in any of the preceding embodiments, the computer program 222 can further include computer programs capable of completing other specific work.
[0106] Further, the present application further discloses a computer readable storage medium for storing a computer program; wherein the computer program is executed by a processor to implement the polyolefin polymerization kinetics parameter correction method disclosed above. The specific steps of the method can refer to the corresponding contents disclosed in the preceding embodiments, which will not be repeated here.
[0107] The various embodiments described in the specification are progressive in nature, and each embodiment highlights the differences from other embodiments. The same or similar parts among the various embodiments can be mutually referred to. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method.
[0108] Those skilled in the art will further appreciate that the individual steps of the examples described in connection with the embodiments disclosed herein can be embodied in electronic hardware, computer software, or combinations of both. The various examples have been described in relation to the described embodiments, as a means of generalizing the interchangeability of hardware and software. Whether employing hardware or software, the described functionality is implemented as desired by the particular application and design constraints. Skilled artisans appreciate that the replacement of one part by a different part, the addition of new parts, or the omission of existing parts, can be made without departing from the scope of the present application.
[0109] The steps of a method or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in random access memory (RAM), flash memory, read-only memory (ROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
[0110] Finally, it needs to be pointed out that, in this document, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or sequence between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or device including the element.
[0111] The above describes the technical solutions provided by the present application in detail, and the principles and implementation manners of the present application are described by using specific examples. The above description of the embodiments is only for helping to understand the method of the present application and its core idea; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manner and application range can be changed, and the above description of the specification should not be understood as limiting the present application.
Claims
1. A method for correcting kinetic parameters of polyolefin polymerization reaction, characterized in that, include: Based on a pre-designed polyolefin production facility, steady-state actual production data and actual product quality information are used to determine the target attribute values of the polyolefin products corresponding to each grade when producing multiple different grades of polyolefin products. The grade is used to indicate the production condition, and the polyolefin products include homopolymers, random copolymers, and block copolymers. The attributes include several of the following: weight-average molecular weight, number-average molecular weight, molecular weight distribution, melt index, and product density. After setting up the computing environment, a polyolefin production model describing the production process of multiple polyolefin grades is constructed, and the polymerization reaction kinetic parameters in the polyolefin production model are initialized; the polymerization reaction kinetic parameters include pre-exponential factor, activation energy, and reaction order; An objective function is constructed based on the preset parameter optimization target information, the steady-state actual production data, the actual product quality information, and the target attribute value; Based on the preset parameter correction framework, the objective function, the polyolefin production model, and the preset iterative optimization termination condition, the combination of target polymerization reaction kinetic parameters is determined; The polyolefin production model is calibrated based on the target polymerization reaction kinetic parameter combination, and the model is validated using the calibrated model. When the validation is successful, the property value prediction operation corresponding to the target polyolefin product is triggered based on the calibrated model.
2. The method for correcting the kinetic parameters of polyolefin polymerization reaction according to claim 1, characterized in that, The method, based on a pre-set polyolefin production apparatus, uses steady-state actual production data and actual product quality information during the production of multiple polyolefin products corresponding to different grades to determine the target attribute values of the polyolefin products corresponding to each grade, including: Obtain steady-state actual production data of a preset polyolefin production unit when producing multiple polyolefin products corresponding to different grades; wherein, the steady-state actual production data includes catalyst feed flow rate, cocatalyst feed flow rate, propylene and comonomer feed flow rate, hydrogen feed flow rate, reactor temperature and reactor pressure. Sampling was conducted on the polyolefin powder corresponding to each of the aforementioned grades to determine the sampling results; Based on the sampling results and gel permeation chromatography, the actual product quality information of the polyolefin products corresponding to each grade is determined; the actual product quality information includes weight-average molecular weight, number-average molecular weight, molecular weight distribution, and product density.
3. The method for correcting the kinetic parameters of polyolefin polymerization reaction according to claim 2, characterized in that, Before constructing the polyolefin production model describing the polyolefin production process corresponding to multiple of the aforementioned grades, the method further includes: For any type of catalyst, multiple product powders corresponding to the stated grades are collected, and gel permeation chromatography is performed on the product powders to determine the number of corresponding catalyst active sites.
4. The method for correcting the kinetic parameters of polyolefin polymerization reaction according to claim 3, characterized in that, After setting up the computing environment, a polyolefin production model is constructed to describe the production process of multiple polyolefin grades, and the polymerization reaction kinetic parameters in the polyolefin production model are initialized, including: After building the target database required for the preset parameter calibration framework and installing the preset chemical simulation software, a polyolefin production model is constructed based on the preset chemical simulation software or preset programming language and the number of catalyst active sites; the polyolefin production model is a model that includes the polymerization reaction mechanism in the polyolefin production process corresponding to multiple grades. Configure the initial values of the polymerization reaction kinetic parameters in the polyolefin production model to complete the parameter initialization operation; Based on the initial value, the corresponding parameter search range is determined.
5. The method for correcting the kinetic parameters of polyolefin polymerization reaction according to claim 4, characterized in that, The objective function is constructed based on the preset parameter optimization target information, the steady-state actual production data, the actual product quality information, and the target attribute value, including: Obtain preset parameter optimization target information; wherein, the preset parameter optimization target information includes attribute prediction error thresholds corresponding to the weight-average molecular weight, the number-average molecular weight and the product density respectively, and the attribute prediction error thresholds are attribute prediction error thresholds of the polyolefin production model; Obtain the weighting coefficients corresponding to the weight-average molecular weight, the number-average molecular weight, and the product density, respectively; Based on the weighting coefficients, the target attribute values, the steady-state actual production data, the actual product quality information, and the polyolefin production model, an objective function is constructed; the objective function is used to determine the total deviation between the predicted attribute values output by the polyolefin production model and the corresponding target attribute values.
6. The method for correcting the kinetic parameters of polyolefin polymerization reaction according to claim 5, characterized in that, The determination of the target polymerization reaction kinetic parameter combination based on the preset parameter correction framework, the objective function, the polyolefin production model, and the preset iterative optimization termination condition includes: Obtain a preset parameter correction framework; the preset parameter correction framework includes one or more combinations of autocovariance matrix adaptive evolution strategy, differential evolution algorithm and particle swarm optimization algorithm; Population initialization is performed based on the preset parameter correction framework and the parameter search range to obtain initialization results; the initialization results include crossover probability and initial number of individuals, where the individuals are a set of polymerization reaction kinetic parameters; Based on the corresponding polyolefin production model, parallel model simulations are performed on each individual to determine the simulation results of the current iteration. Based on the simulation results and the objective function, determine the target total error corresponding to each grade in the current iteration round; Based on the total target error and the preset parameter optimization target information, individual screening and population update are performed to obtain the updated population, and the set of polymerization reaction kinetic parameters that do not exceed any of the attribute prediction error thresholds are recorded. Based on the updated population, the process jumps back to the step of performing parallel model simulations on each individual based on the corresponding polyolefin production model until the preset iterative optimization termination condition is met. Then, the target polymerization reaction kinetic parameter combination is selected from the recorded sets of polymerization reaction kinetic parameters.
7. The method for correcting the kinetic parameters of polyolefin polymerization reaction according to any one of claims 1 to 6, characterized in that, The process of parameter calibration of the polyolefin production model based on the target polymerization reaction kinetic parameter combination, and model validation using the corresponding calibrated model, includes: The polyolefin production model is calibrated based on the target polymerization reaction kinetic parameter combination to determine the calibrated model. Based on the corrected model and the steady-state actual production data of other grades that did not participate in the iterative optimization, the attribute prediction results are determined. Based on the attribute prediction results and the preset parameter optimization target information, it is determined whether the corrected model has passed the verification, so as to determine the model verification result.
8. A device for correcting kinetic parameters of polyolefin polymerization reaction, characterized in that, include: The target attribute value determination module is used to determine the target attribute value of each polyolefin product corresponding to a certain grade based on the steady-state actual production data and actual product quality information of a preset polyolefin production device when producing multiple polyolefin products corresponding to different grades; wherein, the grade is used to represent the production condition, the polyolefin product includes homopolymer, random copolymer and block copolymer, and the attribute includes several of the following: weight average molecular weight, number average molecular weight, molecular weight distribution, melt index and product density; The production model construction module is used to construct a polyolefin production model describing the production process of multiple polyolefin grades after setting up the computing environment, and to initialize the polymerization reaction kinetic parameters in the polyolefin production model; the polymerization reaction kinetic parameters include pre-exponential factor, activation energy, and reaction order; The objective function construction module is used to construct an objective function based on preset parameters, optimized objective information, the steady-state actual production data, the actual product quality information, and the objective attribute value. The iterative optimization module is used to determine the combination of target polymerization reaction kinetic parameters based on a preset parameter correction framework, the objective function, the polyolefin production model, and a preset iterative optimization termination condition. The parameter calibration module is used to calibrate the polyolefin production model based on the target polymerization reaction kinetic parameter combination, and to perform model verification using the corresponding calibrated model. When the verification is successful, the module triggers the prediction operation of the attribute value corresponding to the target polyolefin product based on the calibrated model.
9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the method for correcting kinetic parameters of polyolefin polymerization reaction as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, Used to store a computer program, which, when executed by a processor, implements the method for correcting kinetic parameters of polyolefin polymerization reaction as described in any one of claims 1 to 7.