Polyolefin production parameter optimization method, device and equipment and storage medium

Abstract

The invention relates to the technical field of automation, in particular to a polyolefin production parameter optimization method and device, equipment and a storage medium, and aims to improve the optimization efficiency of polyolefin production parameters. According to the method, for historical production parameter sets corresponding to a plurality of historical polyolefin categories, historical product prediction models are respectively constructed, and parameter optimization processing is iteratively executed on a to-be-optimized parameter set of a target polyolefin category through the historical product prediction models until the to-be-optimized parameter set meets a target category production condition, so that a target production parameter set is obtained. And in each iterative training, performing prediction comparison processing on the to-be-optimized parameter set through each historical product prediction model to obtain a similarity weight of each historical product prediction model, determining a recommended production parameter of the target polyolefin category according to each obtained similarity weight, updating the to-be-optimized parameter set according to the recommended production parameter, and obtaining the target polyolefin category. And executing the next parameter optimization processing.

Description

Technical Field

[0001] The present application relates to the field of automation technology, and in particular to a polyolefin production parameter optimization method, device, equipment and storage medium. Background Art

[0002] Polyolefin production parameters refer to the operating parameters of the production equipment during the polyolefin production process, such as PP molecular weight, crystallinity, the content of two types of ethylene-propylene copolymers, molecular weight, average ethylene content, copolymer sequence distribution, etc. By regulating the polyolefin production parameters, the micro-quality indicators and micro-distribution of polyolefin products will be directly affected.

[0003] However, in current industrial production, due to the limitations of production technology level and the complexity of production regulation, when a polyolefin production plant is newly put into production or a new grade of polyolefin products is produced, it is often necessary to rely on a large amount of trial operation and experimental data to conduct inverse optimization research on the polyolefin production parameters in order to find production parameters that meet the standards, which greatly reduces the optimization efficiency of polyolefin production parameters. Summary of the invention

[0004] The invention provides a polyolefin production parameter optimization method, system, equipment and storage medium, which are used to improve the optimization efficiency of polyolefin production parameters.

[0005] In a first aspect, the present application provides a method for optimizing polyolefin production parameters, the method comprising:

[0006] For each of the historical production parameter sets corresponding to the multiple historical polyolefin categories, corresponding historical product prediction models are constructed respectively; each historical production parameter set includes multiple historical production parameters of the corresponding historical polyolefin category and the molecular weight distribution of the polyolefin product;

[0007] Based on each historical product prediction model, parameter optimization processing is iteratively performed on the parameter set to be optimized of the target polyolefin category until the parameter set to be optimized meets the production conditions of the target category, thereby obtaining a target production parameter set.

[0008] In a second aspect, the present application provides a polyolefin production parameter optimization device, the device comprising:

[0009] A construction unit is used to construct corresponding historical product prediction models for historical production parameter sets corresponding to multiple historical polyolefin categories, respectively; each historical production parameter set includes multiple historical production parameters of the corresponding historical polyolefin category and molecular weight distribution of polyolefin products;

[0010] The optimization unit is used to iteratively perform parameter optimization processing on the parameter set to be optimized of the target polyolefin category based on each historical product prediction model until the parameter set to be optimized meets the production conditions of the target category, thereby obtaining a target production parameter set.

[0011] Optionally, the optimization unit is specifically used for:

[0012] Based on the historical product prediction models, the parameter set to be optimized is subjected to prediction comparison processing to obtain similarity weights of the historical product prediction models;

[0013] Determining recommended production parameters for the target polyolefin class based on the obtained similarity weights;

[0014] Based on the recommended production parameters, the parameter set to be optimized is updated, and based on the updated parameter set to be optimized, the next parameter optimization process is performed.

[0015] Optionally, the optimization unit is specifically used to:

[0016] Based on the parameter set to be optimized, construct a corresponding target product prediction model;

[0017] Based on the model similarity between each historical product prediction model and the target product prediction model, the similarity weight of each historical product prediction model is obtained.

[0018] Optionally, the optimization unit is specifically used to:

[0019] Based on the acquisition functions corresponding to the historical production prediction models, and in combination with the similarity weights, a recommendation function for the target polyolefin category is obtained;

[0020] Based on the recommendation function, the recommended production parameters are obtained.

[0021] Optionally, the target product prediction model includes local product prediction models in multiple local search spaces, and the optimization unit is further used to:

[0022] From the parameter set to be optimized, multiple parameter subsets to be optimized are obtained, and a local search space corresponding to each parameter subset to be optimized is determined;

[0023] The migration Bayesian processing is performed on each local search space respectively to obtain the local product prediction model corresponding to each local search space.

[0024] Optionally, the optimization unit is further used for:

[0025] For each local search space, the model similarity between each historical product prediction model and the corresponding local product prediction model is calculated respectively;

[0026] Based on the similarity of each model, the local similarity weight of each historical product prediction model is obtained; the local similarity weight represents the local similarity weight of the historical product prediction model in the corresponding local search space.

[0027] In a third aspect, the present application provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, any one of the polyolefin production parameter optimization methods described in the first aspect is implemented.

[0028] In a fourth aspect, the present application provides a computer storage medium, wherein the computer-readable storage medium stores computer program instructions, and the computer program instructions are executed by a processor to implement any one of the polyolefin production parameter optimization methods described in the first aspect.

[0029] In a fifth aspect, an embodiment of the present application provides a computer program product, comprising computer program instructions, which, when executed by a processor, implement any one of the polyolefin production parameter optimization methods described in the first aspect.

[0030] The beneficial effects of the present invention are as follows:

[0031] The embodiment of the present application provides a method for optimizing polyolefin production parameters, which constructs corresponding historical product prediction models for each of the historical production parameter sets corresponding to multiple historical polyolefin categories, and iterates the parameter optimization process of the parameter set to be optimized of the target polyolefin category through each historical product prediction model, until the parameter set to be optimized meets the target category production conditions, and obtains the target production parameter set. And in each iterative training, the present application will predict and compare the parameter set to be optimized through each historical product prediction model, obtain the similarity weight of each historical product prediction model, and determine the recommended production parameters of the target polyolefin category according to each similarity weight obtained, and then update the parameter set to be optimized according to the recommended production parameters, and perform the next parameter optimization process. In this way, by fully mining historical production data, suitable polyolefin production parameters can be quickly found, the test debugging time can be reduced, the production cycle and optimization cost can be reduced, and the optimization efficiency of polyolefin production parameters can be greatly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0033] Figure 1 A schematic diagram of a polypropylene production process provided in an embodiment of the present application;

[0034] Figure 2A flow chart of a method for optimizing polyolefin production parameters provided in an embodiment of the present application;

[0035] Figure 3 A flow chart of a method for parameter optimization processing provided in an embodiment of the present application;

[0036] Figure 4 A schematic diagram of a polyolefin production parameter optimization process provided in an embodiment of the present application;

[0037] Figure 5 A schematic diagram of simulation optimization results of production parameters of a propylene multi-stage polymerization process provided in an embodiment of the present application;

[0038] Figure 6 A schematic diagram of the structure of a polyolefin production parameter optimization device provided in an embodiment of the present application;

[0039] Figure 7 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical scheme and advantages of the present application clearer, the technical scheme in the embodiment of the present application will be clearly and completely described below in conjunction with the drawings in the embodiment of the present application. Obviously, the described embodiment is only a part of the embodiment of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application. In the absence of conflict, the embodiments in the present application and the features in the embodiments can be arbitrarily combined with each other. In addition, although the logical order is shown in the flow chart, in some cases, the steps shown or described can be performed in an order different from that here.

[0041] The terms "first" and "second" in the specification and claims of the present application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the term "comprising" and any of their variations are intended to cover non-exclusive protection. For example, a process, method, system, product or device comprising a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or devices. "Multiple" in the present application can mean at least two, for example, can be two, three or more, and the embodiments of the present application are not limited.

[0042] The term "and / or" in the embodiments of the present application is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0043] It is understood that in the following specific implementations of the present application, data related to polyolefin production parameters and the like are involved. When the various embodiments of the present application are applied to specific products or technologies, relevant licenses or consents need to be obtained, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions. For example, relevant volunteers can be recruited and relevant agreements on volunteer authorization data can be signed, and then the data of these volunteers can be used for implementation; or, by implementing within the scope of an authorized organization, the following implementation methods are implemented by using the data of members within the organization to manage data; or, the relevant data used in the specific implementation are all simulated data, such as simulated data generated in a virtual scene.

[0044] The following is a brief introduction to the design concept of the embodiments of the present application.

[0045] Polyolefin production parameters refer to the operating parameters of the production equipment during the polyolefin production process, such as PP molecular weight, crystallinity, the content of two types of ethylene-propylene copolymers, molecular weight, average ethylene content, copolymer sequence distribution, etc. By regulating polyolefin production parameters, the micro-quality indicators and micro-distribution of polyolefin products will be directly affected. The micro-quality and distribution of polymer materials determine the different uses or characteristics of the materials, which in turn affects the market demand and competitiveness of polyolefin products. The controllable production of high-performance polyolefins is an important means to increase product added value and meet high-end needs. Therefore, how to stably control the micro-quality of polyolefin products by optimizing production parameters has become a key issue.

[0046] However, in current industrial production, due to the limitations of production technology level and the complexity of production regulation, when a polyolefin production plant is newly put into production or a new grade of polyolefin products is produced, it is often necessary to rely on a large amount of trial operation and experimental data to conduct inverse optimization research on the polyolefin production parameters in order to find production parameters that meet the standards. This not only leads to low optimization efficiency of polyolefin production parameters, long parameter optimization cycle, and serious waste of production capacity and resources, but also high cost of obtaining production test data, resulting in high economic and time costs of the optimization process.

[0047] In view of the above problems, the present application embodiment provides a method for optimizing polyolefin production parameters, which constructs corresponding historical product prediction models for each of the historical production parameter sets corresponding to multiple historical polyolefin categories, and iterates the parameter optimization process of the parameter set to be optimized of the target polyolefin category through each historical product prediction model, until the parameter set to be optimized meets the target category production conditions, and obtains the target production parameter set. And in each iterative training, the present application will predict and compare the parameter set to be optimized through each historical product prediction model, obtain the similarity weight of each historical product prediction model, and determine the recommended production parameters of the target polyolefin category according to each similarity weight obtained, and then update the parameter set to be optimized according to the recommended production parameters, and perform the next parameter optimization process. In this way, by fully mining historical production data, suitable polyolefin production parameters can be quickly found, the test debugging time can be reduced, the production cycle and optimization cost can be reduced, and the efficiency of production parameter optimization can be greatly improved.

[0048] The following briefly introduces the application scenarios to which the technical solutions of the embodiments of the present application can be applied. It should be noted that the application scenarios introduced below are only used to illustrate the embodiments of the present application and are not limited. In the specific implementation process, the technical solutions provided by the embodiments of the present application can be flexibly applied according to actual needs.

[0049] The solution provided in the embodiments of the present application can be applied to the application scenarios of process optimization in most polyolefin product production processes, and can be used to significantly improve the optimization efficiency of polyolefin production parameters. Figure 1 Shown is a schematic diagram of a polypropylene production process provided by an embodiment of the present application. The figure shows the entire production process from prepolymerization, cyclic polymerization to gas phase polymerization, reflecting how to achieve the molecular structure and quality control of the final product through reaction control at different stages in polypropylene production. As a thermoplastic, polypropylene can change its physical and chemical properties by adjusting the raw material ratio, reaction conditions, etc. in the production process to meet different application requirements. Different grades of polypropylene represent different categories of polypropylene products, and each grade of polypropylene has specific performance indicators, such as melt index, molecular weight distribution, rigidity, toughness, etc., for different application scenarios, such as films, fibers, injection molded products, etc. In order to prepare the target grade of polypropylene, the product molecular weight distribution is used as the optimization target, and 11 production parameters are selected as decision variables based on the actual operating conditions of the factory, such as the hydrogen feed amount of 4 reactors, the propylene feed amount of 4 reactors, the catalyst feed amount of the R-401 reactor, the co-catalyst feed amount and the ethylene feed amount. If the target production parameters suitable for the production conditions of the target polyolefin category can be accurately selected, the debugging time and waste generation will be reduced, and the polyolefin production efficiency will be improved. At the same time, the target production parameters can effectively control key properties such as molecular weight distribution and product purity, which is of great significance for subsequent processing and application in polyolefins.

[0050] Of course, the method provided in the embodiment of the present application is not limited to the above application scenarios, but can also be used in other possible application scenarios, which are not limited by the embodiment of the present application. The functions that can be realized by each device in the above application scenarios will be described in the subsequent method embodiments, and no further details will be given here.

[0051] Below, in combination with the application scenarios described above, the method provided by the exemplary embodiment of the present application is described with reference to the accompanying drawings. It should be noted that the above application scenarios are only shown to facilitate understanding of the spirit and principles of the present application, and the implementation methods of the present application are not limited in this regard.

[0052] See also Figure 2 As shown, it is a flow chart of a method for optimizing polyolefin production parameters provided in an embodiment of the present application, and the specific implementation process of the method is as follows:

[0053] Step 201: construct corresponding historical product prediction models for each of the historical production parameter sets corresponding to a plurality of historical polyolefin categories.

[0054] In the embodiment of the present application, each historical production parameter set includes multiple historical production parameters corresponding to the corresponding historical polyolefin category and the molecular weight distribution of the polyolefin product. The historical production parameter set includes specific production parameters collected from the production process of multiple historical polyolefin categories, which directly affects the production process of polyolefins and may include operating variables such as temperature, pressure, catalyst type and dosage, reaction time, type of reactor and operation mode. The molecular weight distribution of polyolefin products refers to the length distribution of molecular chains in the produced polyolefin products. Molecular weight distribution is an important factor in determining the physical properties of polyolefin materials. Therefore, the ultimate goal of optimizing polyolefin production parameters is to ensure that the molecular weight distribution of the obtained polyolefin product meets the expected material performance requirements. The historical product prediction model is a mathematical model constructed by machine learning, statistical modeling, etc. based on the historical production parameter set and the corresponding product molecular weight distribution. The model can learn known historical production parameters and predict the product molecular weight distribution of the target polyolefin category.

[0055] Specifically, taking the preparation of polypropylene of the target polyolefin category (target grade) as an example, the embodiment of the present application will analyze the relevant process variables of the propylene polymerization process, select the operating variables such as the hydrogen feed amount, propylene feed amount, catalyst feed amount, co-catalyst feed amount and ethylene feed amount of the reactor as production parameters, take the molecular weight distribution of the product as the quality indicator, use the factory system to collect the production parameter values ​​of different polyolefin grades, and calculate the molecular weight distribution curve based on the measured data as the historical production parameter set.

[0056] in, Represents n sets of data samples collected during the polypropylene production process. A large amount of historical data is collected through the factory system, covering the production data of multiple polyolefin grades, that is, the historically collected production parameter sets and the associated data of the corresponding product quality, which serve as the basic data for subsequent training and optimization of the model. k Represents the historical production parameters of the kth sample, which may include the hydrogen feed amount, propylene feed amount, catalyst feed amount, etc. k Represents the product molecular weight distribution of the kth sample, i.e., the target variable under the production parameter conditions. n: represents the total number of samples. A large amount of historical data collected through the factory system covers the production data of multiple polyolefin grades.

[0057] In a possible implementation, when data is scarce, the present application embodiment may use a Gaussian Process (GP) model to construct a historical product prediction model. The Gaussian process is a non-parametric Bayesian method for regression problems, which assumes that the relationship between data points can be represented by a kernel function, so that an unknown function can be predicted. The Gaussian process provides a mean estimate and uncertainty estimate of the function value, that is, for a given input point, not only the output value can be predicted, but also the uncertainty of its prediction can be estimated. The mean of the GP model is denoted as μ(x * ), the uncertainty is given by the variance σ 2 (x * ) indicates that the specific formula is as follows:

[0058]

[0059] Among them, σ 2 (x * ) represents the Gaussian process model at point x * The variance at , i.e. the credibility of the predicted value, reflects the uncertainty of the predicted product molecular weight distribution under specific production parameter conditions. If the uncertainty is large, it means that the Gaussian process model has a low prediction credibility for this point. In practice, a larger variance means that the model lacks sufficient training data support for this point, while a smaller variance indicates that the model's prediction for this point is more reliable.

[0060] k * Represents the evaluation point x * The covariance vector between them is used to measure the similarity between the evaluation point x and other historical data points; K represents a kernel matrix of size n×n, and each element K i,j Meet K i,j = k(x i ,x j), which is used to calculate the covariance matrix between each data point and measure the similarity between data points. Through the kernel function, the model can capture the similarity between the historical production parameter set and the target production parameter set. For example, if the polyolefin products under two different production conditions have similar molecular weight distributions, the kernel function will give them a higher similarity.

[0061] y contains the labels corresponding to all current data x, that is, the known product molecular weight distribution in the historical production parameter set.

[0062] represents the noise term, which indicates the measurement noise in the data, and I is the unit matrix. Noise is inevitable in the actual production process, such as sensor measurement errors or environmental factors. This term is used to represent the impact of these noises on the model prediction results to improve the robustness of the prediction.

[0063] In this way, by inputting the historical production parameter set and product quality data of the corresponding grade to the GP model corresponding to each historical grade, the product performance under the new production parameter combination is predicted, and support is provided for subsequent similarity evaluation. And by converting historical data into a prediction model, the embodiment of the present application effectively utilizes a large amount of data accumulated in the historical production process, avoids starting from scratch every time production optimization, and saves the cost of data collection and model training. Based on the knowledge of multiple historical product prediction models, it is possible to help infer the production parameters of the target polyolefin category, enhance the reliability of the model for predicting the production parameters of the new grade, and improve the generalization ability of the model.

[0064] Step 202: Based on each historical product prediction model, iteratively perform parameter optimization processing on the parameter set to be optimized of the target polyolefin category until the parameter set to be optimized meets the production conditions of the target category, thereby obtaining the target production parameter set.

[0065] In the embodiment of the present application, multiple rounds of optimization processing are performed on the parameters to be optimized of the target polyolefin category through the historical product prediction model to obtain the best production parameter combination of the target polyolefin category, that is, the target production parameter set, to maximize the product quality of the target polyolefin category.

[0066] Specifically, the ability of historical product prediction models built with data from different polyolefin grades to generalize to the target polyolefin grade is defined as Among them, f i (x) represents the model output of the i-th polyolefin grade. For each different polyolefin grade, there will be a corresponding historical product prediction model to predict its product molecular weight distribution. It is the historical production parameters and product molecular weight information related to polyolefin grade i. Different polyolefin grades in the polyolefin production process correspond to different production parameters such as catalyst feed amount, propylene or ethylene feed amount, etc. Therefore, by predicting different grades separately through multiple models, the production process of different grades can be better optimized.

[0067] In one possible implementation, reference Figure 3 As shown, when executing step 202, the embodiment of the present application will specifically perform the following operations in each parameter optimization process:

[0068] Step 221: Based on each historical product prediction model, a prediction comparison process is performed on the parameter set to be optimized to obtain the similarity weight of each historical product prediction model.

[0069] In the embodiment of the present application, after the historical product prediction model is established through the historical data of different polyolefin grades, its ability to be generalized to the target polyolefin category will be evaluated. The parameter set to be optimized of the target polyolefin category is predicted by the established historical product prediction model, and the similarity between them is determined according to the correctness of the observed results, so as to obtain the similarity weight of each historical product prediction model. By calculating the similarity weight, the similarity between the historical production parameters and the target polyolefin category can be more accurately evaluated, so as to dynamically select the most relevant historical model for optimization, improve the accuracy of optimization, and reduce errors.

[0070] Specifically, the embodiment of the present application will construct a corresponding loss function for each historical product prediction model to measure the parameter set to be optimized for the target polyolefin category of each historical product prediction model. The correctness of the sorting is to measure whether the predictions of each historical product prediction model for the target polyolefin category can correctly sort the similarities between its production parameters and historical parameters. Through this loss function, the embodiment of the present application can continuously adjust the model parameters so that it can better capture the similarities between the target polyolefin category and the historical production data. In this way, the loss function based on sorting can improve the prediction accuracy of the model for the target production parameter set, reduce the trial and error cost in the production process, and improve production efficiency. Given n t For target task evaluation >1, we define the loss as the number of incorrectly ordered observation pairs, and the loss function is formulated as follows:

[0071]

[0072] Wherein, i represents the label of the i-th polyolefin category (brand), and when i=t, it represents the target polyolefin brand.

[0073] Represents the loss function of the historical product prediction model corresponding to the i-th historical polyolefin category, which is used to measure the accuracy of the corresponding historical product prediction model in predicting the target polyolefin category. The loss function is used to evaluate the performance of different historical prediction models in the sorting task. If the value of the loss function is lower, it means that the prediction result of the model is closer to the actual value.

[0074] The set of production parameters representing the tth sample, such as operating variables such as raw material feed rate, temperature, pressure, etc. in polyolefin production.

[0075] Indicates that after removing the kth sample, the model produces a parameter set This step is similar to cross-validation and aims to prevent the model from overfitting. Represents the data of the lth historical polyolefin sample through the historical product prediction model corresponding to the i-th polyolefin category The results of the prediction. These two functions represent the prediction results of the target production parameters and historical production parameters respectively. The target production parameters are to be optimized, while the historical production parameters are used to guide model training.

[0076] and Represents the true observation value of the kth sample of the target polyolefin category and the true observation value of the lth sample of the historical polyolefin category. These two parameters are used to compare the difference between the model prediction value and the actual observation value, helping the model to optimize the prediction results.

[0077] 1() represents the indicator function, which is used to evaluate whether there is an error between the predicted result and the actual observed value. If the order between the predicted value and the true value is as expected, the value of the indicator function is 1, indicating that the prediction result is correct; otherwise, the value is 0.

[0078] ⊕ represents the XOR operator, which is used to compare the relationship between the predicted value and the actual value. The XOR operator is used to determine the difference between the predicted ranking and the actual ranking. In the ranking task, a correct prediction will result in the XOR operation being zero, while an incorrect prediction will result in a non-zero value.

[0079] In a possible implementation, the embodiment of the present application will construct a corresponding target product prediction model based on the parameter set to be optimized, and obtain the similarity weight of each historical product prediction model based on the model similarity between each historical product prediction model and the target product prediction model.

[0080] Specifically, for a small amount of known data of the target polyolefin category (target grade), a data sample of some target grades is selected to construct a GP model as the target product prediction model, and the accuracy of the GP model predicting the unselected data is used to determine the similarity of the GP model constructed by the selected target grade data to the target task. For the target product prediction model, the embodiment of the application can use leave-one-out cross-validation to evaluate the prediction of the target product prediction model. The loss function of the target product prediction model is as follows:

[0081]

[0082] in, Represents that after removing the kth sample, the model Comparison between the predicted value f and its true observed value. Used to determine whether the model's predicted value for the target polyolefin class is smaller than the true observed value. This term is true if the model prediction is lower than the true value.

[0083] Represents the comparison between the true observed value of the kth sample of the target polyolefin class and the lth sample of the historical polyolefin class. Used to evaluate the relative relationship between the target polyolefin class data and the historical polyolefin class data. This term is true if the predicted value of the target polyolefin class is less than the historical polyolefin class.

[0084] This formula is a supplement to the previous loss function and is used to further determine whether the target product prediction model of the target polyolefin category is accurately ranked among different historical data sets. By comparing the size relationship between the predicted value and the actual observed value, the formula can effectively optimize the ranking accuracy of the model, making the model more reliable in distinguishing between different polyolefin samples, and ultimately improving the effect of the model in optimizing production parameters. By introducing this loss function, this scheme can ensure that the prediction results of the target production parameters are not only close to the actual values, but also can accurately reflect the relative order of product parameters in the production process. This helps to better control the quality of the production process while further improving the accuracy and reliability of the prediction model.

[0085] According to the loss function constructed in the above steps, the loss value of each historical product prediction model and the target product prediction model generalized to the target task can be calculated, thereby obtaining the model similarity between each historical product prediction model and the target product prediction model. Specifically, the smaller the loss value of the model, the higher its similarity with the target task. Therefore, the model similarity can be measured by the inverse relationship of the loss value, that is, the model with a small loss value has a higher similarity.

[0086] In a possible implementation, the present application embodiment can estimate the similarity probability weight by sampling, that is, each time from According to a certain proportion, some data are randomly selected to obtain a sample data set and replace It is used to measure the loss function of the historical product prediction model and the target product prediction model, so as to calculate the similarity weight w of each historical product prediction model. i , the sampling estimation process will be repeated S times, and the calculation formula of the similarity weight of each model is as follows:

[0087]

[0088] in, Represents the loss value of the i-th historical product prediction model, which is used to measure the performance of the model on a given data set. i represents the i-th model from 1 to t.

[0089] w i Represents the similarity weight of the i-th historical product prediction model. The higher the weight, the more similar the models are.

[0090] S represents the number of repeated sampling, which means that part of the data Conduct multiple sampling;

[0091] Represents the model set with the smallest loss value, that is, the model with the smallest loss value is selected from all models. Indicates that if the i-th model belongs to the model set with the smallest loss value, the value is 1;

[0092] Represents all satisfaction The models are summed up as a normalization factor.

[0093] Step 2022: Determine the recommended production parameters for the target polyolefin category based on the obtained similarity weights.

[0094] In the embodiment of the present application, after obtaining the similarity weights of each historical product prediction model, the similarity weights are used to calculate the recommended sampling points, thereby obtaining the recommended production parameters of the target polyolefin category to optimize the production process of the target polyolefin category.

[0095] In a possible implementation, the embodiment of the present application obtains a recommendation function for the target polyolefin category based on the acquisition function corresponding to each historical production prediction model and in combination with each similarity weight.

[0096] Specifically, using the mean and uncertainty estimates provided by the GP model, a corresponding acquisition function can be designed for each historical product prediction model. The formula of the acquisition function is as follows:

[0097]

[0098] α EI Measure at a given data point The possibility of expected improvement at parameter x under the production parameter conditions helps determine which new production parameters are most likely to bring significant production increases. That is, considering the current data points The corresponding optimal value f(x * ) and the expected improvement of the new recommended sampling point x. By maximizing α EI , the most promising next recommended sampling point can be found, thus gradually approaching the global optimal solution and obtaining the target production parameter set for the target polyolefin category.

[0099] f(x * ) represents product quality indicators such as molecular weight distribution of polyolefins under the best production parameter conditions known so far.

[0100] x represents the newly selected sampling point x from the production parameter set to be optimized, and f(x) represents the function value of the newly selected sampling point x, that is, the prediction performance of the model at the new sampling point.

[0101] The above acquisition function is based on the Expected Improvement (EI) criterion. Exploration and exploitation are balanced by favoring uncertain points with higher improvement potential. By avoiding points that are too close to the current optimal solution, promising areas can be effectively mined and converge to the global optimal solution rather than just the local optimal solution. In this way, by calculating the acquisition function, the embodiment of the present application can estimate the potential benefits of exploration at new sampling points. If f(x) of the sampling point x is equal to the historical optimal value f(x * ) is large, it means there may be more room for improvement at this sampling point.

[0102] In a possible implementation, the embodiment of the present application may also weight the acquisition functions of multiple historical product prediction models to form a final recommendation function, ensuring that the model with better performance occupies a larger proportion in the final sampling decision. The embodiment of the present application takes into account the different data scales involved in the historical production tasks and the target production tasks, and provides a probability weighting strategy to combine f 1 ,…,f t The acquisition functions of multiple historical product prediction models are combined using probability weights. The weighted combination formula is as follows:

[0103]

[0104] Among them, α(x) represents the weighted recommendation function, which integrates the acquisition functions from different models and obtains the final recommendation result in a weighted manner;

[0105] t represents the number of participating models, which is the total number of historical product prediction models and target product prediction models;

[0106] w i Represents the similarity weight of the i-th historical product prediction model;

[0107] EI i (x) represents the acquisition function of the mixed historical product prediction model and the target product prediction model, and then calculates the expected improvement of the integration, which is used to measure the expected improvement that the model can bring under a given input x.

[0108] In a possible implementation, the present application embodiment can further simplify the acquisition function based on the combination of the cumulative distribution function and the probability density function of the standard normal distribution. To make the calculation more convenient, the specific formula is as follows:

[0109]

[0110] in, represents the standardized expectation cue, which is the standardized value of the difference between the mean and the optimal value.

[0111] Φ() represents the cumulative probability density function, which is used to measure the possible benefits of exploring new sampling points.

[0112] φ() is the probability function of the standard normal distribution, which is used to measure the potential of new sampling points in unexplored areas.

[0113] σ(x) represents the standard deviation of the observation point x, indicating the uncertainty of the point. The larger the standard deviation, the greater the uncertainty of the model's prediction for the point.

[0114] In a possible implementation, after obtaining the recommendation function, the embodiment of the present application can obtain the next recommended sampling point with the greatest improvement expectation by maximizing the recommendation function, and then reversely solve it as the production parameter of the next test, that is, the recommended production parameter of the target polyolefin category. The specific formula is as follows:

[0115]

[0116] Among them, x best Represents the optimal sampling point, i.e. the final recommended production parameters, which can bring the best improvement in the production process after the similarity calculation between the historical production forecast model and the target production forecast model. The selection of this point means that it can maximize the expected improvement of the target polyolefin category in the production process.

[0117] X represents all possible production parameter sets. During the production process, different production parameter combinations (such as temperature, pressure, raw material ratio, etc.) will affect the quality and production efficiency of polyolefins. Set X includes different combinations of various production parameters, and the present application embodiment needs to find the optimal value of the target polyolefin category from them.

[0118] Thus, from all possible sampling points x∈X, select the one that can make the recommendation function The point x where the maximum value is reached best , which is the next recommended sampling point that can bring the greatest expected improvement in the production parameters, i.e., the recommended production parameters for the target polyolefin category.

[0119] In a possible implementation, in order to further improve the optimization efficiency and accuracy of production parameters, the embodiment of the present application will adopt a parallel strategy to obtain multiple subsets of parameters to be optimized from the set of parameters to be optimized, and construct local search spaces corresponding to each subset of parameters to be optimized, and perform migration Bayesian processing on each local search space to obtain the local product prediction model corresponding to each local search space, so as to measure the similarity of each model in the local range. In this way, each local search space is only responsible for processing a part of the parameters, reducing the parameter dimension of each subset, reducing the computational complexity, and performing parameter optimization in different local areas at the same time, increasing the possibility of finding the global optimal solution. The risk of falling into the local optimum can be effectively avoided, so as to get closer to or reach the global optimum.

[0120] Specifically, it is necessary to first determine the scope of the local search space. If the local search space is limited, the GP model may not be able to explore a wide enough range of solutions, resulting in an incomplete understanding of the target task behavior and a decrease in model performance. Moreover, if the local search space is too large, the GP model may explore irrelevant or data-inefficient areas, resulting in inaccurate similarity comparisons. Therefore, the embodiment of the present application will use the principle of sampling without replacement to extract data subsets. Among them, the hyperparameter m represents the proportion of the selected data ratio, and then use The GP agent model is constructed based on the data, and the recommended next sampling point x is obtained by maximizing the EI acquisition function. best,m . And use the standard Euclidean distance method to measure x best,m arrive The distance is

[0121] The formula is as follows:

[0122]

[0123] Among them, d represents the dimension of the parameter space, that is, the number of features considered in the parameter optimization process, and each dimension in the parameter space corresponds to;

[0124] Represents the production parameter set of the current sampling point, including the values ​​of the parameter set in each dimension, and is the sampling point used for local optimization;

[0125] s j Represents the standard deviation in the jth dimension, indicating the range of variation of the production parameters of this dimension in historical data.

[0126] x best,m,j Represents the value of the optimal point of the mth historical model in the jth dimension. It is the optimal solution inferred from the historical model and reflects the optimal parameters of a historical model in this dimension, close to x best,m The data are selected to determine the local search space.

[0127] In this way, through this distance, the closeness of the current sampling point to the historical optimal point can be evaluated to guide the subsequent search direction. If the distance is small, it means that the current sampling point is close to the local optimal point and may be a better candidate solution. If the distance is large, it means that the current sampling point deviates from the local optimal point and needs further optimization. The standardization process ensures that the values ​​of different dimensions can be compared on the same scale to avoid the influence of some dimensions being too large or too small. The closeness between the sampling point and the optimal point is determined by the standardized distance, so as to reasonably define the scope of the local search space, help the model quickly find the area closest to the optimal point in the high-dimensional feature space, and thus guide the further optimization of production parameters.

[0128] In one possible implementation, to obtain the most compact local search space containing data The present application embodiment also constructs a constrained optimization problem, and its formula is as follows:

[0129]

[0130] Among them, two constraints define the boundaries of the local search space, that is, all known sampling points x and the current optimal point xbest must fall within the upper and lower bounds of the local search space, ensuring that the local search space can contain all currently valuable information. In the polyolefin synthesis scenario, each component of x represents a different production parameter, such as reaction temperature, pressure, feed rate, etc.

[0131] l, u represent the lower and upper bounds of each production parameter, respectively, and define the desirable range of each production parameter, i.e., the boundary of the local search space. In polyolefin synthesis, l and u may represent the minimum and maximum allowable values ​​of process parameters such as temperature and pressure. These values ​​can be set according to actual production experience or theoretical ranges to ensure that the search space contains a reasonable combination of production parameters, and the present application embodiment does not specifically limit this.

[0132] θ represents the parameter vector, including the upper and lower bounds l and u, and defines the boundary conditions of the local search space. In practical applications, θ is used to control the local search range of the parameter optimization algorithm. By adjusting θ, the size of the search space can be changed, guiding the algorithm to optimize within the effective production parameter range.

[0133] In this way, by minimizing the scope of the local search space and reducing the distance between the upper and lower bounds as much as possible, it can ensure that the production parameters are optimized within a more precise range and avoid invalid searches.

[0134] Furthermore, the compactness of the local search space is achieved by This is achieved by squaring the term, which will penalize the maximum range in each dimension. Therefore, it is easy to get the optimal solution θ given * =(l * ,u * ):

[0135]

[0136] Thus, the upper bound u * and the lower bound l * It is determined by maximizing and minimizing the distance between the sampling point and the optimal point, respectively. These boundary values ​​define the scope of the search space, so that the optimization algorithm only searches in the most relevant area, avoiding unnecessary calculations and invalid sampling.

[0137] In one possible implementation, after constructing the local search space, the embodiment of the present application will, for each local search space, calculate the model similarity between each historical product prediction model and the corresponding local product prediction model, and obtain the local similarity weight of each historical product prediction model in the corresponding local search space based on the model similarities, thereby calculating the next recommended sampling point and recommended production parameters of each local search space through the local similarity weight.

[0138] Specifically, in each local search space, the embodiment of the present application can The Bayesian Optimization (BO) process of transfer learning is performed in the process. The BO algorithm is an iterative algorithm for global optimization. In the embodiment of the present application, the BO algorithm can be used to determine at which parameter point the experiment should be performed each time a sample is taken. The GP model estimates the potential return of the sampling point, and the BO algorithm will continuously update the model based on these estimates to find the global optimal solution. For example, if a historical production prediction model m1 has a high similarity weight w1 with the current task, then the BO optimization process will rely more on the information provided by this model and perform more frequent exploration in the local search space. In this way, by calculating The local similarity weight w between each proxy model and the target task t helps the model share information between similar tasks, thereby accelerating the optimization process. Substitute the local similarity weight w into the recommendation function to calculate the most potential next recommended sampling point And record its recommended value α(x t+1,m′ ). If a sampling point x in the current local search space t+1,m′ The BO optimization algorithm recommends the point as a candidate for the next test and records the recommended value of the point. In the polyolefin synthesis process, it means selecting the sampling point that can maximize the production efficiency or quality. In this way, by optimizing parameters in the local search space, the model explores in the subspace close to the target task, reducing invalid data sampling, making each sampling more likely to be close to the optimal solution, and improving the accuracy and efficiency of the optimization process.

[0139] Step 2023: Based on the recommended production parameters, the parameter set to be optimized is updated, and based on the updated parameter set to be optimized, the next parameter optimization process is performed.

[0140] In the embodiments of the present application, after each parameter optimization process is performed to obtain the recommended production parameters for the target polyolefin category, the parameter set to be optimized for the target polyolefin category will be updated, and the next parameter optimization process will be iteratively performed, thereby improving the posterior probability density of the model and improving the model's ability to describe the polyolefin polymerization process, until the parameter set to be optimized meets the production conditions of the target category and the target production parameter set is obtained.

[0141] Specifically, the embodiment of the present application can select the sampling point with the largest recommendation value from all local search spaces as the best recommended sampling point for actual experiment. Its contribution to the performance of the target grade of polyolefin is verified by experiment, and the experimental results are updated to the target grade of data set, and the parameter optimization process is repeated until the ideal production parameters that meet the target grade of polyolefin are found.

[0142] Please refer to Figure 4 FIG. 1 is a schematic diagram of a polyolefin production parameter optimization process provided by an embodiment of the present application. In the parameter optimization process, the present application collects data samples of multiple historical brands, such as the above Figure 4The data in the "Historical Brand 1 Data", "Historical Brand 2 Data" to "Historical Brand m Data" contain historical data related to different historical production parameters and product quality indicators. The data of each historical brand are used to build the corresponding historical GP model. Each historical GP model reflects the relationship between the historical production parameters of the brand and the product molecular weight distribution, in preparation for subsequent transfer learning. At the same time, the known sample data of the target brand is collected. The initial target brand data may be relatively limited, and only part of the sample data is used for modeling. Similar to the historical brand, based on part of the sample data of the target brand, a target GP model of the target brand is constructed to predict the performance of new data points. For each GP model of the historical brand, a local search space is constructed separately, such as Figure 4 The "local search space 1", "local search space 2", etc. shown in the figure. The local search space is the scope for evaluating and exploring the local similarity of each historical brand model on the target brand data. The generalization ability of the historical brand model is evaluated through the target brand data to determine the similarity between the historical GP model and the target brand. In this step, the target brand data is sorted and evaluated by the historical GP model to determine which historical models are most suitable for migration to the optimization task of the target brand. Next, a sampling method is used to randomly extract part of the data from the target brand data in proportion, and a loss function is constructed to measure the accuracy of the historical model in sorting the target observations. Each calculation result will assign a similarity weight to the model of each historical brand. In this way, through multiple sampling and repeated evaluation, the similarity weight of each historical GP model is finally determined. The higher the weight of the historical model, the greater its similarity with the target brand task. According to the similarity weight of each historical brand model, the integrated acquisition function is used to calculate the recommended sampling point for the next step. This recommended point is the production parameter that is most likely to improve the optimization efficiency at present. At the same time, a parallel strategy is adopted to perform Bayesian optimization in different local search spaces. Parallel processing can accelerate the search for the best sampling point and reduce the reliance on a single historical model. The recommended sampling points obtained in multiple local search spaces are sorted, and the sampling point with the largest recommended value is selected as the best point for actual experiment. After the actual experiment is completed, the actual experimental data is obtained and fed back to the model, the target brand data set is updated, and the previous optimization steps (such as Figure 4 In the cyclic process from "building a Gaussian model" to "determining the recommended sampling location and recommended value", the process parameters are continuously adjusted and optimized until the production parameter combination that best meets the production requirements of the target grade is found, ensuring that the production unit can stably produce polyolefin products that meet the requirements under the new production parameters.

[0143] In a possible implementation, 100 data uniformly distributed on the molecular weight distribution curve are taken as the target task. According to the method for optimizing the production parameters of polyolefins, the present application embodiment can calculate the values ​​of the production parameters that meet the target molecular weight distribution, such as the hydrogen feed amount, the ethylene and propylene feed amounts, and the catalyst and co-catalyst feed amounts. Figure 5 The figure shows a schematic diagram of the simulation optimization results of the production parameters of the propylene multi-stage polymerization process provided by the embodiment of the present application. Figure 5 Displays the average loss and 95% confidence interval at each iteration step. Figure 5 It can be seen that the transfer learning BO algorithm provided by the embodiment of the present application has good optimization performance, especially in the early stage of iteration, a smaller loss value can be quickly obtained. This shows that, especially in the early stage of iteration, the embodiment of the present application can accelerate the optimization solution through the information provided by the source task.

[0144] Please refer to Figure 6 As shown, based on the same inventive concept, the embodiment of the present application also provides a polyolefin production parameter optimization device 60, which includes:

[0145] A construction unit 601 is used to construct corresponding historical product prediction models for historical production parameter sets corresponding to multiple historical polyolefin categories, respectively; each historical production parameter set includes multiple historical production parameters of the corresponding historical polyolefin category and molecular weight distribution of polyolefin products;

[0146] The optimization unit 602 is used to iteratively perform parameter optimization processing on the parameter set to be optimized of the target polyolefin category based on each historical product prediction model until the parameter set to be optimized meets the production conditions of the target category, thereby obtaining the target production parameter set.

[0147] Optionally, the optimization unit 602 is specifically configured to:

[0148] Based on each historical product prediction model, a prediction comparison process is performed on the parameter set to be optimized to obtain the similarity weight of each historical product prediction model;

[0149] Determining recommended production parameters for the target polyolefin class based on the obtained similarity weights;

[0150] Based on the recommended production parameters, the parameter set to be optimized is updated, and based on the updated parameter set to be optimized, the next parameter optimization process is performed.

[0151] Optionally, the optimization unit 602 is specifically configured to:

[0152] Based on the parameter set to be optimized, a corresponding target product prediction model is constructed;

[0153] Based on the model similarity between each historical product prediction model and the target product prediction model, the similarity weight of each historical product prediction model is obtained.

[0154] Optionally, the optimization unit 602 is specifically configured to:

[0155] Based on the acquisition functions corresponding to each historical production prediction model and combined with each similarity weight, a recommendation function for the target polyolefin category is obtained;

[0156] Based on the recommendation function, recommended production parameters are obtained.

[0157] Optionally, the target product prediction model includes local product prediction models in multiple local search spaces, and the optimization unit 602 is further used to:

[0158] From the parameter set to be optimized, multiple parameter subsets to be optimized are obtained, and a local search space corresponding to each parameter subset to be optimized is determined;

[0159] The migration Bayesian processing is performed on each local search space respectively to obtain the local product prediction model corresponding to each local search space.

[0160] Optionally, the optimization unit 602 is further configured to:

[0161] For each local search space, the model similarity between each historical product prediction model and the corresponding local product prediction model is calculated respectively;

[0162] Based on the similarity of each model, the local similarity weight of each historical product prediction model is obtained; the local similarity weight represents the local similarity weight of the historical product prediction model in the corresponding local search space.

[0163] For the convenience of description, the above parts are divided into each unit module (or module) according to the function and described separately. Of course, when implementing the present application, the functions of each unit (or module) can be implemented in the same one or more software or hardware. It can be understood by those skilled in the art that various aspects of the present application can be implemented as a system, method or program product. Therefore, various aspects of the present application can be specifically implemented in the following forms, namely: a complete hardware implementation method, a complete software implementation method (including firmware, microcode, etc.), or an implementation method combining hardware and software, which can be collectively referred to as "circuit", "module" or "system" here.

[0164] The device can be used to execute the methods shown in the various embodiments of the present application. Therefore, for the functions that can be implemented by the various functional modules of the device, reference can be made to the description of the aforementioned embodiments and no further details will be given.

[0165] See also Figure 7As shown, based on the same technical concept, the embodiment of the present application also provides a computer device 70, which can be Figure 1 As shown as a master system or a slave system, the computer device 70 may include a memory 701 and a processor 702 .

[0166] The so-called memory 701 is used to store computer programs executed by the processor 702. The memory 701 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the computer device, etc. The processor 702 may be a central processing unit (CPU), or a digital processing unit, etc. The specific connection medium between the above-mentioned memory 701 and the processor 702 is not limited in the embodiments of the present application. The embodiments of the present application are Figure 4 In the embodiment, the memory 701 and the processor 702 are connected via a bus 703. The bus 703 is connected to the processor 702 via a bus 703. Figure 4 The connection between other components is shown by bold lines, and is not intended to be limiting. The so-called bus 703 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0167] The memory 701 may be a volatile memory, such as a random-access memory (RAM); the memory 701 may also be a non-volatile memory, such as a read-only memory, a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD), or the memory 701 may be any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 701 may be a combination of the above memories.

[0168] The processor 702 is used to call the computer program stored in the so-called memory 701 to execute the polyolefin production parameter optimization method executed by the device in each embodiment of the present application.

[0169] In some possible embodiments, various aspects of the polyolefin production parameter optimization method provided in the present application can also be implemented in the form of a program product, which includes a program code. When the program product is run on a computer device, the program code is used to enable the computer device to execute the steps of the polyolefin production parameter optimization method according to various exemplary embodiments of the present application described above in this specification. For example, the computer device can execute the steps of each embodiment.

[0170] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0171] The program product of the embodiment of the present application may adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a computing device. However, the program product of the present application is not limited thereto. In the present application, the readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with a command execution system, device or device.

[0172] The readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, wherein the readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable signal medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with a command execution system, apparatus, or device.

[0173] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0174] Program code for performing the operations of the present application may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user equipment, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device via any type of network including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0175] It should be noted that, although several units or subunits of the device are mentioned in the above detailed description, this division is merely exemplary and not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more units described above can be embodied in one unit. Conversely, the features and functions of one unit described above can be further divided into multiple units to be embodied.

[0176] In addition, although the operations of the method of the present application are described in a specific order in the drawings, this does not require or imply that the operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

[0177] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0178] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0179] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for optimizing polyolefin production parameters, characterized in that: The method comprises: For each of the historical production parameter sets corresponding to the multiple historical polyolefin categories, corresponding historical product prediction models are constructed respectively; each historical production parameter set includes multiple historical production parameters of the corresponding historical polyolefin category and the molecular weight distribution of the polyolefin product; Based on each historical product prediction model, parameter optimization processing is iteratively performed on the parameter set to be optimized of the target polyolefin category until the parameter set to be optimized meets the production conditions of the target category, thereby obtaining a target production parameter set.

2. The method according to claim 1, characterized in that Each parameter optimization process includes: Based on the historical product prediction models, the parameter set to be optimized is subjected to prediction comparison processing to obtain similarity weights of the historical product prediction models; Determining recommended production parameters for the target polyolefin class based on the obtained similarity weights; Based on the recommended production parameters, the parameter set to be optimized is updated, and based on the updated parameter set to be optimized, the next parameter optimization process is performed.

3. The method according to claim 2, characterized in that The method of performing prediction comparison processing on the parameter set to be optimized based on each historical product prediction model to obtain the similarity weight of each historical product prediction model includes: Based on the parameter set to be optimized, construct a corresponding target product prediction model; Based on the model similarity between each historical product prediction model and the target product prediction model, the similarity weight of each historical product prediction model is obtained.

4. The method according to claim 2, characterized in that The step of determining the recommended production parameters of the target polyolefin category based on the obtained similarity weights includes: Based on the acquisition functions corresponding to the historical production prediction models, and in combination with the similarity weights, a recommendation function for the target polyolefin category is obtained; Based on the recommendation function, the recommended production parameters are obtained.

5. The method according to claim 2, characterized in that The target product prediction model includes a plurality of local product prediction models under local search spaces, and the step of constructing a corresponding target product prediction model based on the parameter set to be optimized further includes: From the parameter set to be optimized, multiple parameter subsets to be optimized are obtained, and a local search space corresponding to each parameter subset to be optimized is determined; The migration Bayesian processing is performed on each local search space respectively to obtain the local product prediction model corresponding to each local search space.

6. The method according to claim 5, characterized in that The method of obtaining the similarity weight of each historical product prediction model based on the model similarity between each historical product prediction model and the target product prediction model also includes: For each local search space, the model similarity between each historical product prediction model and the corresponding local product prediction model is calculated respectively; Based on the similarity of each model, the local similarity weight of each historical product prediction model is obtained; the local similarity weight represents the local similarity weight of the historical product prediction model in the corresponding local search space.

7. A polyolefin production parameter optimization device, characterized in that: The device comprises: A construction unit is used to construct corresponding historical product prediction models for historical production parameter sets corresponding to multiple historical polyolefin categories, respectively; each historical production parameter set includes multiple historical production parameters of the corresponding historical polyolefin category and molecular weight distribution of polyolefin products; The optimization unit is used to iteratively perform parameter optimization processing on the parameter set to be optimized of the target polyolefin category based on each historical product prediction model until the parameter set to be optimized meets the production conditions of the target category, thereby obtaining a target production parameter set.

8. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product comprising computer program instructions, characterized in that When the computer program instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images