A polymer material formula design optimization method and system

Abstract

The present application relates to the technical field of high polymer material design, and particularly relates to a high polymer material formula design optimization method and system.The method comprises: performing formula iteration according to multiple preliminary experimental data to obtain a target material formula; wherein the i-th iteration in the formula iteration comprises: according to multiple real data of the i-th iteration, performing prediction on multiple preset material formulas of the i-th iteration respectively to obtain multiple virtual data of the i-th iteration; and then performing model training to obtain a prediction model of the i-th iteration; and then determining a candidate material formula of the i-th iteration, and analyzing the difference between the real performance and the model prediction performance of the candidate material formula of the i-th iteration, and terminating the iteration when the difference is less than a preset deviation threshold.The present application can reduce the number of real experiments on material formulas under the premise of ensuring the search accuracy of the material formulas, and reduce the search cost of the optimal formula of the high polymer material.

Description

Technical Field

[0001] This invention relates to the technical field of polymer material design, and specifically to a method and system for optimizing polymer material formulation design. Background Technology

[0002] Polymer materials such as polyethylene (PE) and polyvinyl chloride (PVC) are widely used in cable sheathing, pipes, and automotive parts. Their final performance depends not only on the base resin but also on the synergistic effects of various additives such as flame retardants, antioxidants, light stabilizers, fillers, and plasticizers. To obtain the optimal formulation that satisfies a range of properties including strength, thermal stability, flame retardancy, and abrasion resistance, it is typically necessary to test multiple additives and their different levels.

[0003] Related technologies often employ an exhaustive approach to test the specific performance of possible material formulations. However, with the increase in material systems and types of additives, the search space for material formulations expands rapidly. If the original method is still used to explore the optimal formulation, a large number of experiments are required, which dramatically increases the search cost for the optimal formulation. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for optimizing polymer material formulations, which solves the technical problem of high search costs for optimal polymer material formulations in existing technologies.

[0005] In a first aspect, one embodiment of the present invention provides a method for optimizing polymer material formulation design, the method comprising:

[0006] Formula iterations were conducted based on multiple preliminary experimental data to obtain the target material formula;

[0007] The i-th iteration in the formula iteration includes:

[0008] Based on multiple real data from the i-th iteration, multiple preset material formulations for the i-th iteration are predicted to obtain multiple virtual data for the i-th iteration.

[0009] Based on multiple real data and multiple virtual data from the i-th iteration, the prediction model for the (i-1)-th iteration is trained to obtain the prediction model for the i-th iteration.

[0010] The candidate material formulation for the i-th iteration is determined based on the prediction model of the i-th iteration, and the difference between the actual performance and the model prediction performance of the candidate material formulation for the i-th iteration is analyzed to obtain the prediction bias index of the prediction model for the i-th iteration.

[0011] Where i is a positive integer greater than 1, and when i is 2, the multiple real data of the i-th iteration are obtained from the multiple preliminary experimental data, and the prediction model of the (i-1)-th iteration is a preset model; the target material formulation is: the candidate material formulation corresponding to the prediction deviation index being less than the preset deviation threshold.

[0012] In one embodiment, the step of predicting multiple preset material formulations for the i-th iteration based on multiple real data from the i-th iteration to obtain multiple virtual data for the i-th iteration includes:

[0013] In the multiple real data in the i-th iteration, analyze the formula difference between each real data and each preset material formula to obtain the formula difference information of each preset material formula in the i-th iteration;

[0014] Based on the formulation difference information of each preset material formulation in the i-th iteration, and the material performance data included in each of the multiple real data in the i-th iteration, the material performance of each preset material formulation in the i-th iteration is predicted to obtain multiple virtual data in the i-th iteration.

[0015] In one embodiment, the step of analyzing the formulation differences between each real data point and each preset material formulation in the multiple real data points of the i-th iteration to obtain the formulation difference information of each preset material formulation in the i-th iteration includes:

[0016] In the multiple real data of the i-th iteration, analyze the difference between each real data and each preset material formula under each formula parameter, so as to obtain the multiple parameter difference data of each preset material formula in the i-th iteration.

[0017] Based on the effect index of each formulation parameter on each performance parameter, the parameter difference data corresponding to each preset material formulation in the i-th iteration is corrected to obtain the formulation difference information of each preset material formulation in the i-th iteration, wherein the effect index is used to indicate the degree to which the corresponding formulation parameter affects the corresponding performance parameter.

[0018] In one embodiment, determining the candidate material formulation for the i-th iteration based on the prediction model of the i-th iteration includes:

[0019] Based on multiple real and virtual data from the i-th iteration, analyze the correlation between each parameter value of each formula parameter and the parameter target of each performance parameter to obtain multiple correlation indices.

[0020] Multiple new material formulations are generated based on the aforementioned multiple correlation indices;

[0021] The prediction model of the i-th iteration is used to predict the multiple new material formulations respectively, so as to obtain the material performance data of each new material formulation;

[0022] Among the multiple new material formulations, the new material formulation whose matching degree between the material performance data and the material performance target meets the preset matching conditions is determined as the candidate material formulation for the i-th iteration.

[0023] In one embodiment, the correlation index is determined based on the difference between the corresponding formula parameter value and the corresponding performance parameter target.

[0024] In one embodiment, generating multiple new material formulations based on the multiple correlation indices includes:

[0025] Based on the multiple correlation indices, interpolation is performed on the multiple parameter values ​​corresponding to each formula parameter in the i-th iteration to obtain the multiple parameter interpolation values ​​corresponding to each formula parameter in the i-th iteration;

[0026] Based on the multiple parameter values ​​and multiple parameter insertion values ​​corresponding to each formula parameter in the i-th iteration, the multiple new material formulas are generated;

[0027] Among them, the parameter value corresponding to the formula parameter in the i-th iteration and the parameter insertion value are the parameter values ​​corresponding to the formula parameter in the (i+1)-th iteration.

[0028] In one embodiment, the parameter insertion value is obtained by dividing the numerical interval formed by two adjacent parameter values ​​of the corresponding formula parameter equally, and the equalization index corresponding to the parameter insertion value is positively correlated with the correlation index of the two corresponding parameter values ​​of the corresponding formula parameter.

[0029] In one embodiment, the step of obtaining the mean index corresponding to the parameter insertion value includes:

[0030] Calculate the mean of the correlation index of two corresponding parameter values ​​of the corresponding formula parameter to obtain the average strength index of the numerical interval formed by the two corresponding parameter values ​​of the corresponding formula parameter;

[0031] The average intensity index of the numerical interval formed by the two corresponding parameter values ​​of the corresponding formula parameter is integerized to obtain the average index of the numerical interval formed by the two corresponding parameter values ​​of the corresponding formula parameter.

[0032] In one embodiment, the degree of matching between the material performance data of the new material formulation and the material performance target is: the Euclidean distance between the material performance data of the new material formulation and the material performance target;

[0033] The material performance data of the candidate material formulation matches the material performance target better than other newly added material formulations in the corresponding iteration.

[0034] Secondly, another embodiment of the present invention also provides a polymer material formulation design optimization system, the system comprising:

[0035] The formulation iteration module is used to iterate the formulation based on multiple preliminary experimental data to obtain the formulation of the target material.

[0036] The i-th iteration in the formula iteration includes:

[0037] Based on multiple real data from the i-th iteration, multiple preset material formulations for the i-th iteration are predicted to obtain multiple virtual data for the i-th iteration.

[0038] Based on multiple real data and multiple virtual data from the i-th iteration, the prediction model for the (i-1)-th iteration is trained to obtain the prediction model for the i-th iteration.

[0039] The candidate material formulation for the i-th iteration is determined based on the prediction model of the i-th iteration, and the difference between the actual performance and the model prediction performance of the candidate material formulation for the i-th iteration is analyzed to obtain the prediction bias index of the prediction model for the i-th iteration.

[0040] Where i is a positive integer greater than 1, and when i is 2, the multiple real data of the i-th iteration are obtained from the multiple preliminary experimental data, and the prediction model of the (i-1)-th iteration is a preset model; the target material formulation is: the candidate material formulation corresponding to the prediction deviation index being less than the preset deviation threshold.

[0041] Thirdly, in another embodiment of the present invention, an electronic device is provided, including a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method described in the first aspect.

[0042] Fourthly, in another embodiment of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0043] The present invention has the following beneficial effects:

[0044] This invention first obtains multiple sets of preliminary experimental data through a limited number of preliminary experiments. Then, it iterates the formulation based on this data. In each iteration, based on the real data of the tested material formulations, it predicts multiple preset material formulations to determine the corresponding virtual data. It then trains a prediction model by combining the real and virtual data. The prediction model is then used to determine the candidate material formulation for the current iteration. The difference between the real performance and the model prediction performance of the candidate material formulation in the i-th iteration is analyzed to determine whether to terminate the iteration. The target material formulation is determined based on the candidate material formulation at the time of iteration termination. This can minimize the number of real experiments on the material formulation while ensuring the accuracy of the determined target material formulation, thereby reducing the search cost for the optimal formulation of polymer materials. Attached Figure Description

[0045] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a flowchart illustrating a method for optimizing polymer material formulations according to an embodiment of the present invention.

[0047] Figure 2 This is a schematic diagram of the structure of a polymer material formulation design and optimization system provided in an embodiment of the present invention;

[0048] Figure 3 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0049] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a polymer material formulation design optimization method and system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0050] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0051] The following description, in conjunction with the accompanying drawings, details the specific scheme of the polymer material formulation design optimization method and system provided by the present invention.

[0052] This invention proposes a method for optimizing polymer material formulation design. Please refer to [link / reference]. Figure 1 The diagram illustrates a flowchart of a polymer material formulation design optimization method according to an embodiment of the present invention, the method comprising:

[0053] Step S1: Iterate the formulation based on multiple preliminary experimental data to obtain the target material formulation.

[0054] The i-th iteration in the formula iteration includes:

[0055] Based on multiple real data from the i-th iteration, multiple preset material formulations for the i-th iteration are predicted to obtain multiple virtual data for the i-th iteration.

[0056] Based on multiple real data and multiple virtual data from the i-th iteration, the prediction model for the (i-1)-th iteration is trained to obtain the prediction model for the i-th iteration.

[0057] The candidate material formulation for the i-th iteration is determined based on the prediction model of the i-th iteration, and the difference between the actual performance and the model prediction performance of the candidate material formulation for the i-th iteration is analyzed to obtain the prediction bias index of the prediction model for the i-th iteration.

[0058] Where i is a positive integer greater than 1, and when i is 2, the multiple real data of the i-th iteration are obtained from the multiple preliminary experimental data, and the prediction model of the (i-1)-th iteration is a preset model; the target material formulation is: the candidate material formulation corresponding to the prediction deviation index being less than the preset deviation threshold.

[0059] For example, the polymeric materials for which the optimal formulation is to be explored in this invention can be automotive tire tread rubber (the formulation parameters involved in the material formulation include: natural rubber, styrene-butadiene rubber, cis-butadiene rubber, carbon black, silica, sulfur, accelerator, antioxidant, zinc oxide, aromatic oil), flame-retardant plastics for mobile phone casings (the formulation parameters involved in the material formulation include: polycarbonate, acrylonitrile-butadiene-styrene copolymer, brominated flame retardant, antimony synergist, glass fiber, toughening agent, antioxidant, lubricant, color masterbatch), and high-end sports shoe midsole foam materials (the formulation parameters involved in the material formulation include: polyamide block elastomer, supercritical nitrogen / carbon dioxide, nucleating agent, crosslinking agent, activator, structural reinforcing agent), etc.

[0060] The above preliminary experimental data can be understood as relevant data obtained from experiments based on a material formulation of the polymer material to be explored (specific performance parameter values ​​under multiple performance parameters). The above multiple performance parameters may include: tensile strength, thermal decomposition temperature, coefficient of thermal expansion, wear resistance coefficient, flame retardancy, etc.

[0061] The aforementioned preliminary experimental data correspond one-to-one with multiple material formulations, and the material formulations corresponding to different preliminary experimental data are different (at least one formulation parameter has a different value).

[0062] In application, a corresponding parameter value range (based on experience) can be preset for each formulation parameter involved in the formulation of the polymer material to be explored. Then, a search space is constructed by combining the parameter value ranges of multiple formulation parameters. In the search space, a corresponding search step size is configured for each formulation parameter. The formulation search is performed in the search space according to the search step size to obtain multiple coarse-grained material formulations (for example, if the parameter value range of a certain formulation parameter is 1-100, and the search step size corresponding to the formulation parameter is 10, then the number of values ​​that can be searched for by the formulation parameter is 10).

[0063] Then, a set number (e.g., 50 or 100) or a set proportion (e.g., 5% or 10%) of the material formulations to be tested are randomly searched among multiple coarse-grained material formulations. Based on the searched material formulations, material tests are carried out (referring to the production of polymer materials based on the corresponding material formulations to be tested, and the specific performance of the produced polymer materials under various performance tests) to obtain the above-mentioned preliminary experimental data.

[0064] The aforementioned multiple preset material formulations can be understood as: among the aforementioned multiple coarse-grained material formulations, in addition to the multiple coarse-grained material formulations corresponding to the multiple preliminary experimental data, the other coarse-grained material formulations.

[0065] In this invention, after determining the candidate material formulation for each iteration, a real test of the candidate material formulation is carried out in order to obtain the real performance of the corresponding candidate material formulation.

[0066] It should be noted that after determining the candidate material formulation for the i-th iteration based on the prediction model of the i-th iteration, material experiments will be conducted based on the candidate material formulation for the i-th iteration to obtain more valuable real experimental data (compared to simply exhaustively searching the search space) as a supplement to the aforementioned multiple preliminary experimental data. Furthermore, the multiple real data from the i-th iteration and the experimental data obtained from the material experiments based on the candidate material formulation for the i-th iteration will together constitute the multiple real data for the (i+1)-th iteration (if it is necessary to continue with the (i+1)-th iteration).

[0067] Based on the above setup, this invention first obtains multiple real preliminary experimental data through a limited number of preliminary experiments, and then iterates the formulation accordingly. In each iteration, based on the real data of the tested material formulations, multiple preset material formulations are predicted to determine the corresponding virtual data. The prediction model is then trained by combining the real data and the virtual data. The prediction model is then used to determine the candidate material formulation for the current iteration, and the difference between the real performance and the model prediction performance of the candidate material formulation in the i-th iteration is analyzed to determine whether to terminate the iteration. The target material formulation is determined based on the candidate material formulation at the time of termination of the iteration. This can minimize the number of real experiments on the material formulation while ensuring the accuracy of the determined target material formulation, thereby reducing the search cost of the optimal formulation of polymer materials.

[0068] For example, the aforementioned preset model can be a random forest model.

[0069] In one embodiment, the step of predicting multiple preset material formulations for the i-th iteration based on multiple real data from the i-th iteration to obtain multiple virtual data for the i-th iteration includes:

[0070] In the multiple real data in the i-th iteration, analyze the formula difference between each real data and each preset material formula to obtain the formula difference information of each preset material formula in the i-th iteration;

[0071] Based on the formulation difference information of each preset material formulation in the i-th iteration, and the material performance data included in each of the multiple real data in the i-th iteration, the material performance of each preset material formulation in the i-th iteration is predicted to obtain multiple virtual data in the i-th iteration.

[0072] Specifically, the step of analyzing the formulation differences between each real data point and each preset material formulation in the multiple real data points of the i-th iteration to obtain the formulation difference information of each preset material formulation in the i-th iteration includes:

[0073] In the multiple real data of the i-th iteration, analyze the difference between each real data and each preset material formula under each formula parameter, so as to obtain the multiple parameter difference data of each preset material formula in the i-th iteration.

[0074] Based on the effect index of each formulation parameter on each performance parameter, the parameter difference data corresponding to each preset material formulation in the i-th iteration is corrected to obtain the formulation difference information of each preset material formulation in the i-th iteration, wherein the effect index is used to indicate the degree to which the corresponding formulation parameter affects the corresponding performance parameter.

[0075] The parameter difference data described above is used to characterize the degree of difference between the parameter values ​​of the corresponding preset material formula and the corresponding actual data under the corresponding formula parameters. The value of the parameter difference data is positively correlated with the degree of difference. For example, the value of the parameter difference data can be the sum of the squares of the parameter values ​​of the corresponding preset material formula and the corresponding actual data under the corresponding formula parameters.

[0076] The above formula difference information includes the corrected parameter difference data for each preset material formula in the i-th iteration.

[0077] It should be understood that if the total number of formula parameters is set to a1 and the total number of actual data is a2, then the total number of parameter difference data for the preset material formula in the i-th iteration is... .

[0078] Based on the above settings, the differences in parameter values ​​between each real data material formulation and the preset material formulation under each formulation parameter are analyzed one by one. This allows for a comprehensive evaluation of the differences between each preset material formulation and the material formulations of multiple real data at each formulation parameter. Based on the distribution of parameter values ​​and corresponding material performance values ​​of multiple real data at each formulation parameter, the material performance achievable by each preset material formulation under its corresponding formulation parameter can be accurately estimated. The introduction of an effect index to correct for parameter difference data is to accommodate the varying degrees of influence of different formulation parameters on different material performance parameters (e.g., the mechanical strength of polyethylene (PE) is more easily affected by filler content and processing aids, while flame retardants have a negative effect on its toughness; therefore, when estimating the mechanical strength of preset material formulations, greater attention should be paid to the similarity of filler content, processing aid content, and flame retardant content), further improving the accuracy of the predicted virtual formulation.

[0079] For example, for the first The preset material formula and the i-th iteration Regarding the material formulations corresponding to the actual data, both are in the first place of polymer materials. The coefficient values ​​corresponding to each performance parameter It can be represented as:

[0080]

[0081] in, Let V represent the normalization function, and let V represent the total number of performance parameters of the polymer material. Indicates the first The formula parameter affects the first The effect index of each performance parameter Indicates the first The preset material formula is in the first Parameter values ​​under each formula parameter. Denotes the i-th iteration. The material formulation corresponding to each real data point is in the [number]th [year]. Parameter values ​​under each formula parameter. This can be understood as the first The preset material formula and the i-th iteration The material formulation corresponding to the first real data is for the first The formula parameter in the first Corrected parameter difference data for each performance parameter.

[0082] Correspondingly, for the first The first preset material formulation in polymer materials The parameter values ​​corresponding to each performance parameter can be expressed as:

[0083]

[0084] Where Q represents the total number of real data in the i-th iteration. Represents the normalization function. The i-th iteration The first real data in the first The parameter values ​​under each performance parameter.

[0085] In applications, based on relevant historical experimental information of polymer materials, the relevant numerical data between each formulation parameter and each performance parameter can be obtained, and variance analysis (or correlation analysis) can be performed on them to determine the effect index of each formulation parameter on each performance parameter.

[0086] In one embodiment, determining the candidate material formulation for the i-th iteration based on the prediction model of the i-th iteration includes:

[0087] Based on multiple real and virtual data from the i-th iteration, analyze the correlation between each parameter value of each formula parameter and the parameter target of each performance parameter to obtain multiple correlation indices.

[0088] Multiple new material formulations are generated based on the aforementioned multiple correlation indices;

[0089] The prediction model of the i-th iteration is used to predict the multiple new material formulations respectively, so as to obtain the material performance data of each new material formulation;

[0090] Among the multiple new material formulations, the new material formulation whose matching degree between the material performance data and the material performance target meets the preset matching conditions is determined as the candidate material formulation for the i-th iteration.

[0091] Furthermore, the correlation index is determined based on the difference between the corresponding formula parameter value and the corresponding performance parameter target.

[0092] The above parameters are set by the user according to actual usage needs, and this invention does not limit them.

[0093] Specifically, when the parameter target is a specific value, the closer the corresponding formula parameter value is to the corresponding performance parameter value, the better the corresponding formula parameter value matches the parameter target, and the larger the corresponding correlation index (the correlation index ranges from 0 to 1). Conversely, when the parameter target is a range of values, the closer the corresponding formula parameter value is to the center value of that range, the better the corresponding formula parameter value matches the parameter target, and the larger the corresponding correlation index.

[0094] It should be noted that when there are multiple parameter values ​​for the corresponding formula under the corresponding performance parameter, the average of these multiple parameter values ​​under the corresponding performance parameter is used to analyze the degree of matching with the target parameter.

[0095] Furthermore, the step of generating multiple new material formulations based on the multiple correlation indices includes:

[0096] Based on the multiple correlation indices, interpolation is performed on the multiple parameter values ​​corresponding to each formula parameter in the i-th iteration to obtain the multiple parameter interpolation values ​​corresponding to each formula parameter in the i-th iteration;

[0097] Based on the multiple parameter values ​​and multiple parameter insertion values ​​corresponding to each formula parameter in the i-th iteration, the multiple new material formulas are generated;

[0098] Among them, the parameter value corresponding to the formula parameter in the i-th iteration and the parameter insertion value are the parameter values ​​corresponding to the formula parameter in the (i+1)-th iteration. The material formula set in the i-th iteration is formed by arranging and combining the parameter values ​​corresponding to the multiple formula parameters in the (i+1)-th iteration. The multiple preset material formulas in the i-th iteration are the other material formulas in the material formula set in the i-th iteration, excluding the material formulas corresponding to the multiple real data in the i-th iteration.

[0099] Among the multiple formula parameter values ​​corresponding to the newly added material formula, at least one of the parameter insertion values ​​shall be included.

[0100] The above method utilizes the existence of correlation indices to identify high-value parameter values ​​(the corresponding performance parameter values ​​that are close to the parameter target) among multiple parameter values ​​of the formulation parameters. Interpolation is then performed on the high-value parameter values ​​to dynamically achieve a progressive subdivision of the parameter values ​​of each formulation parameter through iterative interpolation. While ensuring the accuracy of the search for the optimal material formulation, the growth rate of the parameter values ​​of the formulation parameters to be explored is suppressed in multiple iterations, thereby reducing the number of material formulations that need to be experimentally verified, and thus reducing the exploration cost of the optimal material formulation.

[0101] Furthermore, the parameter insertion value is obtained by equally dividing the numerical interval formed by two adjacent parameter values ​​of the corresponding formula parameter, and the average index corresponding to the parameter insertion value is positively correlated with the correlation index of the two corresponding parameter values ​​of the corresponding formula parameter.

[0102] The step of obtaining the mean index corresponding to the parameter insertion value includes:

[0103] Calculate the mean of the correlation index of two corresponding parameter values ​​of the corresponding formula parameter to obtain the average strength index of the numerical interval formed by the two corresponding parameter values ​​of the corresponding formula parameter;

[0104] The average intensity index of the numerical interval formed by the two corresponding parameter values ​​of the corresponding formula parameter is integerized to obtain the average index of the numerical interval formed by the two corresponding parameter values ​​of the corresponding formula parameter.

[0105] For example, the first The average index corresponding to the parameter interpolation value between the p-th parameter value and the (p+1)-th parameter value in a certain iteration of a formula parameter. It can be represented as:

[0106]

[0107] in, This represents the floor function; This represents the amplification factor (used to amplify the correlation index corresponding to the formula parameter value; the k value can be 3). Indicates the first The correlation index of the p-th parameter value of a formula parameter Indicates the first The correlation index of the (p+1)th parameter value of a formula parameter.

[0108] The equal division index is used to characterize the number of new parameter values ​​to be inserted between two vector parameter values ​​of a corresponding formulation parameter. The parameter insertion value can be understood as: the value of one or more division points that equally divide the numerical interval of two adjacent parameter values ​​of a given formulation parameter according to the corresponding equal division index.

[0109] Furthermore, the degree of matching between the material performance data and the material performance target of the new material formulation is: the Euclidean distance between the material performance data and the material performance target of the new material formulation;

[0110] The material performance data of the candidate material formulation matches the material performance target better than other newly added material formulations in the corresponding iteration.

[0111] The aforementioned material performance targets include the parameter targets for each of the multiple performance parameters. The Euclidean distance between the material performance data of the new material formulation and the material performance targets is: the sum of the Euclidean distances between the material performance data of the new material formulation and the material performance targets under each performance parameter (after normalization, the normalized value range is (0, 1)).

[0112] The candidate material formulations for the corresponding iteration should be understood as: based on the degree of matching between the material performance data of the new material formulation and the material performance target, the top-ranked new materials after sorting all the new material formulations for the corresponding iteration in descending order (a specific number (e.g., 10) or a specific proportion (e.g., 5%) can be specified).

[0113] The difference between the actual performance and the model prediction performance of the candidate material formulation in each iteration can be understood as the sum of the Euclidean distances between the actual performance and the model prediction performance of the candidate material formulation in each iteration for each performance parameter (after normalization, the normalized value range is (0, 1)). The above deviation threshold can be set to 0.15 based on experience.

[0114] Among them, the target material formulation is the candidate material formulation with the highest degree of matching between the actual performance and the material performance target (calculated by referring to the aforementioned method based on Euclidean distance; the larger the distance, the lower the degree of matching).

[0115] This invention proposes a polymer material formulation design optimization system. Please refer to [link / reference]. Figure 2 The diagram illustrates a structural schematic of a polymer material formulation design optimization system 200 provided in an embodiment of the present invention. The system includes:

[0116] The formulation iteration module 201 is used to iterate the formulation based on multiple preliminary experimental data to obtain the formulation of the target material.

[0117] The i-th iteration in the formula iteration includes:

[0118] Based on multiple real data from the i-th iteration, multiple preset material formulations for the i-th iteration are predicted to obtain multiple virtual data for the i-th iteration.

[0119] Based on multiple real data and multiple virtual data from the i-th iteration, the prediction model for the (i-1)-th iteration is trained to obtain the prediction model for the i-th iteration.

[0120] The candidate material formulation for the i-th iteration is determined based on the prediction model of the i-th iteration, and the difference between the actual performance and the model prediction performance of the candidate material formulation for the i-th iteration is analyzed to obtain the prediction bias index of the prediction model for the i-th iteration.

[0121] Where i is a positive integer greater than 1, and when i is 2, the multiple real data of the i-th iteration are obtained from the multiple preliminary experimental data, and the prediction model of the (i-1)-th iteration is a preset model; the target material formulation is: the candidate material formulation corresponding to the prediction deviation index being less than the preset deviation threshold.

[0122] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the polymer material formulation design and optimization system and the polymer material formulation design and optimization method embodiment provided in the above embodiments belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0123] This invention also provides an electronic device. Please refer to [link to relevant documentation]. Figure 3 The electronic device may include a processor 301, a memory 302, and a program 3021 stored in the memory 302 and capable of running on the processor 301.

[0124] When program 3021 is executed by processor 301, it can achieve the following: Figure 1 Any steps in the method embodiments and the achievement of the same beneficial effects will not be described in detail here.

[0125] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by hardware related to program instructions, and the program can be stored in a readable medium.

[0126] This invention also provides a readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described functions. Figure 1 Any step in the method embodiment can achieve the same technical effect, and will not be repeated here to avoid repetition.

[0127] The computer-readable storage medium of this invention can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. For example, a computer-readable storage medium can be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0128] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0129] The program code contained on the storage medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0130] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or terminal. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0131] This invention also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned steps to achieve the polymer material formulation design optimization method provided in the above embodiments.

[0132] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0133] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for optimizing polymer material formulations, characterized in that, The method includes: Formula iterations were conducted based on multiple preliminary experimental data to obtain the target material formula; The i-th iteration in the formula iteration includes: Based on multiple real data from the i-th iteration, multiple preset material formulations for the i-th iteration are predicted to obtain multiple virtual data for the i-th iteration. This includes: analyzing the formulation differences between each real data point and each preset material formulation in the multiple real data points of the i-th iteration to obtain the formulation difference information of each preset material formulation in the i-th iteration; and predicting the material properties of each preset material formulation in the i-th iteration based on the formulation difference information of each preset material formulation in the i-th iteration and the material property data included in each real data point in the multiple real data points of the i-th iteration to obtain multiple virtual data for the i-th iteration. Based on multiple real data and multiple virtual data from the i-th iteration, the prediction model for the (i-1)-th iteration is trained to obtain the prediction model for the i-th iteration. The candidate material formulation for the i-th iteration is determined based on the prediction model of the i-th iteration, and the difference between the actual performance and the model prediction performance of the candidate material formulation for the i-th iteration is analyzed to obtain the prediction bias index of the prediction model for the i-th iteration. Where i is a positive integer greater than 1, and when i is 2, the multiple real data of the i-th iteration are obtained from multiple preliminary experimental data, and the prediction model of the (i-1)-th iteration is a preset model; the target material formulation is: the candidate material formulation corresponding to the prediction deviation index being less than the preset deviation threshold. Specifically, in the multiple real data sets of the i-th iteration, the formulation differences between each real data set and each preset material formulation are analyzed to obtain the formulation difference information of each preset material formulation in the i-th iteration. This includes: analyzing the parameter value differences between each real data set and each preset material formulation under each formulation parameter in the multiple real data sets of the i-th iteration to obtain multiple parameter difference data of each preset material formulation in the i-th iteration; and correcting the parameter difference data corresponding to each preset material formulation in the i-th iteration based on the effect index of each formulation parameter on each performance parameter to obtain the formulation difference information of each preset material formulation in the i-th iteration. The effect index is used to indicate the degree to which the corresponding formulation parameter affects the corresponding performance parameter.

2. The method for optimizing polymer material formulations according to claim 1, characterized in that, The candidate material formulation for the i-th iteration is determined based on the prediction model of the i-th iteration, including: Based on multiple real and virtual data from the i-th iteration, analyze the correlation between each parameter value of each formula parameter and the parameter target of each performance parameter to obtain multiple correlation indices. Multiple new material formulations are generated based on multiple correlation indices; Based on the prediction model of the i-th iteration, predictions are made for multiple new material formulations to obtain the material performance data of each new material formulation. Among multiple new material formulations, the new material formulation whose material performance data and material performance target meet the preset matching conditions is determined as the candidate material formulation for the i-th iteration.

3. The method for optimizing polymer material formulations according to claim 2, characterized in that, The correlation index is determined based on the difference between the corresponding formula parameter value and the corresponding performance parameter target.

4. The method for optimizing polymer material formulations according to claim 2, characterized in that, Multiple new material formulations were generated based on several correlation indices, including: Based on multiple correlation indices, interpolation is performed on multiple parameter values ​​corresponding to each formulation parameter in the i-th iteration to obtain multiple parameter interpolation values ​​corresponding to each formulation parameter in the i-th iteration; Based on the multiple parameter values ​​and multiple parameter insertion values ​​corresponding to each formula parameter in the i-th iteration, multiple new material formulas are generated. Among them, the parameter value corresponding to the formula parameter in the i-th iteration and the parameter insertion value are the parameter values ​​corresponding to the formula parameter in the (i+1)-th iteration.

5. The method for optimizing polymer material formulations according to claim 4, characterized in that, The parameter interpolation value is obtained by dividing the numerical interval formed by two adjacent parameter values ​​of the corresponding formula parameter equally. The average index of the parameter interpolation value is positively correlated with the correlation index of the two corresponding parameter values ​​of the corresponding formula parameter.

6. The method for optimizing polymer material formulations according to claim 5, characterized in that, The steps for obtaining the mean index corresponding to the parameter interpolation value include: Calculate the mean of the correlation index of two corresponding parameter values ​​of the corresponding formula parameter to obtain the average strength index of the numerical interval formed by the two corresponding parameter values ​​of the corresponding formula parameter; The average intensity index of the numerical interval formed by the two corresponding parameter values ​​of the corresponding formula parameter is integerized to obtain the average index of the numerical interval formed by the two corresponding parameter values ​​of the corresponding formula parameter.

7. The method for optimizing polymer material formulations according to claim 2, characterized in that, The degree of matching between the material performance data and the material performance target of the new material formulation is: the Euclidean distance between the material performance data and the material performance target of the new material formulation; The material performance data of the candidate material formulations matched the material performance targets better than other newly added material formulations in the corresponding iteration.

8. A polymer material formulation design and optimization system, characterized in that, The system includes: The formulation iteration module is used to iterate the formulation based on multiple preliminary experimental data to obtain the formulation of the target material. The i-th iteration in the formula iteration includes: Based on multiple real data from the i-th iteration, multiple preset material formulations for the i-th iteration are predicted to obtain multiple virtual data for the i-th iteration. This includes: analyzing the formulation differences between each real data point and each preset material formulation in the multiple real data points of the i-th iteration to obtain the formulation difference information of each preset material formulation in the i-th iteration; and predicting the material properties of each preset material formulation in the i-th iteration based on the formulation difference information of each preset material formulation in the i-th iteration and the material property data included in each real data point in the multiple real data points of the i-th iteration to obtain multiple virtual data for the i-th iteration. Based on multiple real data and multiple virtual data from the i-th iteration, the prediction model for the (i-1)-th iteration is trained to obtain the prediction model for the i-th iteration. The candidate material formulation for the i-th iteration is determined based on the prediction model of the i-th iteration, and the difference between the actual performance and the model prediction performance of the candidate material formulation for the i-th iteration is analyzed to obtain the prediction bias index of the prediction model for the i-th iteration. Where i is a positive integer greater than 1, and when i is 2, the multiple real data of the i-th iteration are obtained from multiple preliminary experimental data, and the prediction model of the (i-1)-th iteration is a preset model; the target material formulation is: the candidate material formulation corresponding to the prediction deviation index being less than the preset deviation threshold. Specifically, in the multiple real data sets of the i-th iteration, the formulation differences between each real data set and each preset material formulation are analyzed to obtain the formulation difference information of each preset material formulation in the i-th iteration. This includes: analyzing the parameter value differences between each real data set and each preset material formulation under each formulation parameter in the multiple real data sets of the i-th iteration to obtain multiple parameter difference data of each preset material formulation in the i-th iteration; and correcting the parameter difference data corresponding to each preset material formulation in the i-th iteration based on the effect index of each formulation parameter on each performance parameter to obtain the formulation difference information of each preset material formulation in the i-th iteration. The effect index is used to indicate the degree to which the corresponding formulation parameter affects the corresponding performance parameter.

Images