Formula optimization method, system and equipment for accelerant and storage medium

Abstract

The invention discloses a formula optimization method, system and equipment for an accelerant and a storage medium, and belongs to the field of formula optimizing.The formula optimization method comprises the steps that the accelerant needed for generating a target product is obtained, and components of different dosages in the accelerant are combined to obtain a plurality of candidate formulas; determining a plurality of target performances of the target product, obtaining element concentrations of target elements in the first product and the second product after use for the preset time period, and generating a first quantity value of the target performances based on the element concentrations; a performance prediction model is constructed based on the candidate formula and the first quantity value, the performance prediction model predicts the current formula to generate a second quantity value, and an evaluation value of the current formula is generated based on the rating function; if the evaluation value is greater than the first threshold value, determining the formula as a preferred formula, otherwise, determining the formula as a formula to be optimized, adjusting the dosage of the components in the formula to be optimized, calculating the corresponding evaluation value, and repeating the step until the evaluation value is greater than the first threshold value. According to the invention, the optimization efficiency of the accelerator formula is improved.

Description

Technical Field

[0001] The present invention belongs to the field of formula optimization, and in particular relates to a formula optimization method, system, equipment and storage medium for an accelerator. Background Art

[0002] In the production and use of rubber products, accelerators are an indispensable type of chemical additives. They can accelerate the chemical reaction between rubber and vulcanizers, thereby improving production efficiency and product quality. As rubber products are used for a longer time, environmental factors such as sunlight, temperature, and humidity have a significant impact on their performance and appearance. In particular, ultraviolet light has a particularly serious aging effect on light-colored rubber products. The aging resistance, yellowing resistance, and wear resistance of rubber products are all important considerations in formula design.

[0003] Traditional formulation design methods often target all ingredients in rubber products. For example, a Chinese patent application with publication number CN117689275A proposes a rubber material production batching method, device, equipment and storage medium. The method includes: determining an initial ratio of ingredients according to performance quality requirement parameters of a target rubber material and starting production of the target rubber material; acquiring real-time environmental parameters in real time during the production process of the target rubber material, and performing non-destructive testing on the target rubber material generated during the production process to obtain multiple real-time performance quality parameters; calculating multiple performance quality difference values ​​based on multiple performance quality requirement parameters and multiple real-time performance quality parameters and determining optimal batching parameters according to the multiple performance quality difference values; and continuing production of the target rubber material according to the current optimal ratio of each required raw material in the optimal batching parameters and a preset production process.

[0004] The above existing technologies can meet the performance and quality requirements of the target rubber material to a certain extent by adopting a comprehensive analysis of multiple performance and quality requirement parameters and real-time performance and quality parameters, as well as a method of dynamically adjusting the ratio. However, the process efficiency of quality monitoring is low, resulting in low production efficiency. Therefore, a formula optimization method for accelerators is needed to improve the production efficiency of rubber products while ensuring the production quality of rubber products. Summary of the invention

[0005] To solve the above problems, the present invention provides a method, system, device and storage medium for optimizing the formulation of an accelerator to solve the problems in the prior art.

[0006] In order to achieve the above-mentioned object of the invention, the present invention proposes a method for optimizing the formulation of an accelerator, comprising:

[0007] S1: obtaining an accelerator required for producing a target product, the accelerator comprising a plurality of components, determining the type and dosage range of each component, combining the components in different dosages to obtain a plurality of candidate formulas, and producing a first product based on the candidate formulas;

[0008] S2: determining multiple target properties of the target product, wherein the target properties are determined by element concentrations of multiple target elements in the target product, obtaining element concentrations of the target elements in the first product and the second product after a preset time period based on X-rays, and generating a first value of the target property based on the element concentrations, wherein the smaller the first value, the better the target property of the target product;

[0009] S3: constructing a performance prediction model based on the candidate recipe and the corresponding first value, wherein the performance prediction model predicts the current recipe to generate a corresponding second value, setting a rating function, and generating evaluation values ​​of all the target performances of the current recipe based on the rating function;

[0010] S4: If the evaluation value is greater than the first threshold, the current formula is the preferred formula, otherwise it is the formula to be optimized, the dosage of the components in the formula to be optimized is adjusted, the evaluation value corresponding to the adjusted formula is calculated, and this step is repeated until the evaluation value is greater than the first threshold.

[0011] Further, obtaining the element concentration includes the following steps:

[0012] The first product and the second product are sampled to obtain a first sample and a second sample, respectively. An energy range of X-ray irradiation is set for each of the target elements. The first sample and the second sample are irradiated with X-rays within the energy range, respectively. The absorption of X-rays by the first sample and the second sample is defined as the element concentration of the target element in the first sample and the second sample.

[0013] Further, generating the first value of the target performance based on the element concentration comprises the following steps:

[0014] Performing two-dimensional mapping on the element concentration of the target element of the first sample and the second sample to obtain corresponding first images and second images, identifying image regions in the first image and the second image where the concentration of the target element is higher than a second threshold, defining them as first regions and second regions, calculating the area difference between the area of ​​the first region and the area of ​​the second region, setting a corresponding first weight coefficient for each of the target elements, and calculating a first value α of the target performance of the target product obtained by using the candidate formula based on a first formula, wherein the first formula is: α= Wherein, n is the number of target elements associated with each target performance, A i is the difference between the area of ​​the first region and the area of ​​the second region corresponding to the i-th target element, μ i is the first weight coefficient corresponding to the i-th target element.

[0015] Further, generating the corresponding second value includes the following steps:

[0016] A performance prediction model is established based on a linear regression equation, the performance prediction model generates a corresponding relationship between a first value of each target performance and the candidate recipe, and a second value corresponding to the target performance of each current recipe is predicted and generated based on the performance prediction model.

[0017] Furthermore, setting the rating function includes the following steps:

[0018] Obtain a target range of the first value corresponding to each of the target performances, determine a target value of each of the target performances within the target range, set a second weight coefficient for each of the target performances, and the rating function is: Among them, P i is the second value of the i-th target performance, Y is the evaluation value, m is the total number of all the target performances, is the second weight coefficient of the i-th target performance, q i is the target value of the i-th target performance, f i The target range of the first value corresponding to the i-th target performance.

[0019] Furthermore, adjusting the amount of each component in the formula to be optimized comprises the following steps:

[0020] The number of target properties whose second quantity value in the recipe to be optimized is not within the target range is defined as a first quantity. If the first quantity is greater than a third threshold, a first component is selected from the recipe to be optimized, the first component being a component that affects the target property and the number of target properties affected is greater than a first value. The first component is determined based on experiments. The dosage range is divided into a first range and a second range based on the current dosage of the first component in the recipe to be optimized, the center points of the first range and the second range are obtained respectively, and the dosages at the two center points are defined as a first dosage and a second dosage. On the basis of keeping the dosages of other components unchanged, two adjusted recipes are obtained based on the first dosage and the second dosage respectively, the second quantity value corresponding to each target property in the two adjusted recipes is calculated, and it is determined whether each second quantity value is within the target range. If not, the recipe closer to the target range is defined as the first recipe, the dosage of the first component in the first recipe is defined as a first intermediate dosage, and the first range or the second range is divided into a first interval and a second interval based on the first intermediate dosage. This step is repeated until the second quantity value corresponding to each target property is within the corresponding target range.

[0021] Furthermore, adjusting the amount of each component in the formula to be optimized also includes the following steps:

[0022] If the first number is less than the third threshold, the rating function is determined to be the objective function, and the target range corresponding to each target performance is the constraint condition. On the basis of the constraint condition, the amount of each component in the formula to be optimized is adjusted based on the particle swarm algorithm until the second value of each target performance of the formula to be optimized meets the target range.

[0023] The present invention also provides a formula optimization system for an accelerator, which is used to implement the above method. The system mainly includes:

[0024] A formula generation module, used to obtain an accelerator required for producing a target product, wherein the accelerator includes a plurality of components, determine the type and dosage range of each component, combine the components in different dosages to obtain a plurality of candidate formulas, and produce a first product based on the candidate formulas;

[0025] a target performance quantification module, used to determine multiple target performances of the target product, wherein the target performance is determined by the element concentrations of multiple target elements in the target product, obtain the element concentrations of the target elements in the first product and the second product after a preset time period based on X-rays, and generate a first value of the target performance based on the element concentrations, wherein the smaller the first value, the better the target performance of the target product;

[0026] A performance prediction module, configured to construct a performance prediction model based on the candidate recipes and the corresponding first value, wherein the performance prediction model predicts the current recipe to generate the corresponding second value, sets a rating function, and generates evaluation values ​​of all the target performances of the current recipe based on the rating function;

[0027] A recipe optimization module is used to: if the evaluation value is greater than a first threshold, the current recipe is the preferred recipe; otherwise, it is the recipe to be optimized; the dosage of the components in the recipe to be optimized is adjusted; the evaluation value corresponding to the adjusted recipe is calculated; and this step is repeated until the evaluation value is greater than the first threshold.

[0028] A formulation optimization device for an accelerator, characterized in that the device comprises a processor and a memory, the memory is used to store at least one computer program, and the at least one computer program is loaded by the processor and executes the method described in any one of claims 1 to 7.

[0029] The present invention also provides a computer storage medium, which stores program instructions, wherein when the program instructions are executed, the device where the computer storage medium is located is controlled to execute the above-mentioned method.

[0030] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0031] The present invention obtains the promoter required for generating the target product, determines the composition and dosage range of each promoter, generates multiple candidate formulas, determines multiple target properties of the target product, and uses X-rays to obtain the concentration of the target element. The X-ray technology accurately determines the relationship between the material performance and the element concentration, and quantifies the target performance of the target product through a first value, thereby improving the accuracy of the target performance; a performance prediction model is established using a linear regression equation, and the performance prediction model quickly and accurately predicts the performance of the current formula to obtain a second value by using the relationship between the candidate formula and the corresponding first value, and evaluates whether the current formula is the preferred formula based on the rating function. If it is not the preferred formula, the current formula is optimized based on the objective function, and the proportion of each component in the current formula is adjusted until the optimization target is met. By optimizing the formula and production conditions, the cycle from product design to market can be shortened, the launch of new products can be accelerated, and the production efficiency of the target product can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A flow chart of the steps of the method for optimizing the formulation of the accelerator of the present invention;

[0033] Figure 2 The structure diagram of the system for optimizing the formulation of the accelerator according to the present invention. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0035] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0036] like Figure 1 As shown, a method for optimizing the formulation of an accelerator comprises:

[0037] S1: Obtain the promoter required to produce the target product, the promoter includes multiple components, determine the type and dosage range of each component, combine the components in different dosages to obtain multiple candidate formulas, and produce the first product based on the candidate formulas.

[0038] Specifically, the target product in the present invention is, for example, a rubber product, and the raw material components of the rubber product include, for example, a yellowing-resistant xanthate polysulfide accelerator, referred to as accelerator A, and the formula of accelerator A includes 100 phr of raw rubber; 20 to 100 phr of filler; 0 to 3.0 phr of processing aid; 0 to 20 phr of processing oil; 1.5 to 10 phr of vulcanizer; 0.5 to 2.0 phr of xanthate polysulfide accelerator; 3.0 to 5.0 phr of zinc oxide; 0 to 5.0 phr of activator; and 0 to 3.0 phr of zinc stearate. According to the dosage range of each component, multiple candidate formulas are generated, and other raw material information for generating the target product is obtained, such as antistatic agent, colorant, and extender, etc., and the first product is generated on the basis of ensuring that the dosage of other raw materials remains unchanged.

[0039] S2: Determine multiple target properties of the target product, where the target properties are determined by the element concentrations of multiple target elements in the target product. The element concentrations of the target elements in the first product and the second product after a preset time period are obtained based on X-rays, and a first value of the target property is generated based on the element concentration. The smaller the first value is, the better the target performance of the target product is.

[0040] Specifically, the target properties of rubber products include: yellowing resistance, wear resistance, durability, etc. The target elements related to yellowing resistance include but are not limited to sulfur (sulfur reacts with the rubber molecular chain during the vulcanization process to form cross-linked bonds, affecting the material's yellowing resistance), zinc (zinc oxide is used as an activator for the vulcanization reaction in the formula, which helps to improve the vulcanization efficiency and the yellowing resistance of the rubber), titanium (titanium dioxide) as a pigment, which can improve the material's light resistance and reduce the occurrence of yellowing), etc.; the target elements related to wear resistance include but are not limited to carbon (carbon can increase the hardness of the material, thereby enhancing wear resistance), chromium (chromium can increase the hardness, toughness and tensile strength of steel, and has a positive effect on improving wear resistance), manganese (manganese has a positive effect on wear resistance by improving the wear resistance, hardenability and tensile strength of steel), etc. Other target properties are not given as examples.

[0041] The X-ray absorption spectroscopy (XAS) method is used to obtain the element concentrations of the above elements in the first product and the second product after a preset time period. Compared with traditional chemical analysis methods, X-ray absorption spectroscopy can quickly measure a large number of samples and significantly improve test efficiency. The preset time period is, for example, half a year. The X-ray absorption spectroscopy method obtains the concentration of the target element in the analysis sample by measuring the X-ray absorption of the sample, and uses the element concentration to generate a first value of the target performance. The specific generation method will be explained later. The first value refers to the difference in target performance between the first product and the second product, such as the first value representing yellowing resistance. The smaller the first value, the smaller the difference in yellowing degree of the target product after one year of use, and the stronger the yellowing resistance of the target product. X-ray technology can help researchers better understand the relationship between material properties and element concentrations, and better quantify target performance.

[0042] S3: construct a performance prediction model based on the candidate recipe and the corresponding first value, use the performance prediction model to predict the current recipe and generate the corresponding second value, set a rating function, and generate evaluation values ​​of all target performances of the current recipe based on the rating function.

[0043] Specifically, in order to predict the target performance of the rubber product generated by the current formula, a performance prediction model is established. The specific model construction process will be explained later. Multiple candidate formulas and corresponding first values ​​are input into the performance prediction model. The performance prediction model predicts and generates a second value of the target performance of the target product generated by the current formula based on the relationship between the candidate formula and the first value. The performance prediction model is used to improve work efficiency and reduce energy consumption and time in the production process. A rating function is set to generate a comprehensive evaluation value based on the second values ​​of all target performance of the current formula, which is used to evaluate the overall effect of the candidate formula.

[0044] S4: If the evaluation value is greater than the first threshold, the current formula is the preferred formula, otherwise it is the formula to be optimized. The dosage of the components in the formula to be optimized is adjusted, and the evaluation value corresponding to the adjusted formula is calculated. This step is repeated until the evaluation value is greater than the first threshold.

[0045] Specifically, the amount of each component in the formula to be optimized is adjusted according to the set adjustment method and the objective function in the particle swarm algorithm, the evaluation value corresponding to the adjusted formula is calculated, and this step is repeated until the evaluation value is greater than the first threshold. The objective function can quickly iterate different formula combinations, greatly shortening the time to find the optimal solution.

[0046] The present invention obtains the promoter required for generating the target product, determines the composition and dosage range of each promoter, generates multiple candidate formulas, determines multiple target properties of the target product, and uses X-rays to obtain the concentration of the target element. The X-ray technology accurately determines the relationship between the material performance and the element concentration, and quantifies the target performance of the target product through a first value, thereby improving the accuracy of the target performance; a performance prediction model is established using a linear regression equation, and the performance prediction model quickly and accurately predicts the performance of the current formula to obtain a second value by using the relationship between the candidate formula and the corresponding first value, and evaluates whether the current formula is the preferred formula based on the rating function. If it is not the preferred formula, the current formula is optimized based on the objective function, and the proportion of each component in the current formula is adjusted until the optimization target is met. By optimizing the formula and production conditions, the cycle from product design to market can be shortened, the launch of new products can be accelerated, and the production efficiency of the target product can be improved.

[0047] As a preferred technical solution of the present invention, obtaining the element concentration includes the following steps:

[0048] The first product and the second product are sampled to obtain a first sample and a second sample, respectively. An energy range of X-ray irradiation is set for each target element. The first sample and the second sample are irradiated with X-rays within the energy range, respectively. The absorption of X-rays by the first sample and the second sample is defined as the element concentration of the target element in the first sample and the second sample.

[0049] Specifically, a portion of rubber is extracted from the first product and the second product as the first sample and the second sample. The energy range of X-ray irradiation of each target element is different. For example, the energy range of X-ray irradiation of sulfur (S) element is between 2.3 and 2.5 kiloelectron volts (keV), while the energy range of X-ray irradiation of zinc element is between 9.6 and 9.8 keV. Assuming that the target element is sulfur element, X-rays are irradiated to the first sample and the second sample respectively. Before irradiation, the position of the X-ray detector is fixed to ensure that the distance between the X-ray detector and the sample is fixed so as to perform consistent measurement. The absorption of X-rays by the sample is used to represent the element concentration. The greater the X-ray absorption, the higher the sulfur concentration. The X-ray absorption is measured by the fluorescence method. After obtaining the X-ray absorption spectrum, the data needs to be processed. The X-ray absorption is corrected by the calibration curve to obtain the final element concentration.

[0050] Generating a first measure of a target property based on the element concentration comprises the following steps:

[0051] Perform two-dimensional mapping on the element concentrations of the target elements of the first sample and the second sample to obtain the corresponding first image and second image, identify the image areas in the first image and the second image where the target element concentration is higher than the second threshold, define them as the first area and the second area, calculate the area difference between the first area and the second area, set the corresponding first weight coefficient for each target element, and calculate the first value α of the target performance of the target product obtained by using the candidate formula based on the first formula, the first formula is: Where n is the number of target elements associated with each target performance, A i is the difference between the area of ​​the first region and the area of ​​the second region corresponding to the i-th target element, μ i is the first weight coefficient corresponding to the i-th target element.

[0052] Specifically, in two-dimensional mapping, color mapping (such as heat map) can be used to represent the change in absorption. Different absorption ranges correspond to different colors. If the absorption is large, the pixel will be displayed as a brighter color; if the absorption is small, the pixel will be displayed as a darker color. Since the X-ray absorption may span multiple orders of magnitude, the data needs to be normalized for better representation on the image. Before identifying the first area and the second area, the color image formed by the two-dimensional mapping is converted into a grayscale image to enhance the contrast between the target element area and the background. The threshold segmentation method is used to identify the first area and the second area in the first image and the second image where the sulfur concentration is higher than the set threshold.

[0053] The area difference between the two regions is calculated, and a corresponding first weight coefficient is set for the target elements related to yellowing resistance, such as sulfur and zinc. The first weight coefficient reflects the degree of contribution of the target elements to the yellowing resistance performance. The greater the contribution, the higher the first weight coefficient is set. The weighted sum of the corresponding differences of all target elements related to yellowing resistance is taken as the first quantity value. The first quantity value reflects the degree of change in yellowing resistance of the target product after a period of use. The smaller the first quantity value, the smaller the degree of change in yellowing resistance, and the better the physical properties of the target product regarding yellowing resistance.

[0054] Generating the corresponding second quantity value comprises the following steps:

[0055] A performance prediction model is established based on a linear regression equation. The performance prediction model generates a corresponding relationship between a first value of each target performance and a candidate recipe. A second value corresponding to the target performance of each current recipe is predicted and generated based on the performance prediction model.

[0056] Specifically, a performance prediction model is established. The model is established based on a linear regression equation. The model determines a specific regression equation through training input data. The regression equation is, for example: S=α0+α1X1+α2X2+…+α n X n +ε, where S is the target performance, X1, X2, …, X n is the amount of different components in the candidate formula, α0, α1, …, α n is the model parameter, ε is the error term, and the model parameters and error terms are estimated based on Bayes’ theorem. The current formula is input into the performance prediction model, and the second value corresponding to each target performance is predicted and generated. The performance prediction model can be used to quickly predict the performance of the new formula without the need for actual experimental testing every time, which greatly speeds up the R&D process.

[0057] Setting up the rating function involves the following steps:

[0058] Obtain the target range of the first value corresponding to each target performance, determine the target value of each target performance within the target range, set a second weight coefficient for each target performance, and the rating function is: Among them, P i is the second value of the i-th target performance, Y is the evaluation value, m is the total number of all target performances, is the second weight coefficient of the i-th target performance, q i is the target value of the i-th target performance, f i is the target range of the first value corresponding to the i-th target performance.

[0059] Specifically, the formulation of the accelerator may affect multiple target properties of the target product. Therefore, when optimizing the formulation of the accelerator, a rating function should be set taking into account all target properties. The rating function is the sum of the weighted probabilities that all target properties fall within the corresponding target range. For different rubber products, the second weight coefficient of each target performance is different. For example, for rubber products such as tires, stronger wear resistance is required and weaker yellowing resistance is required, so a larger second weight coefficient will be set for wear resistance, and a smaller second weight coefficient will be set for yellowing resistance. For rubber shoes, stronger yellowing resistance is required, followed by wear resistance, so a larger second weight coefficient will be set for yellowing resistance, and a smaller second weight coefficient will be set for wear resistance. The evaluation value of the target product for all target properties is calculated. The larger the evaluation value, the closer the target performance is to the target value, and the better the corresponding formulation.

[0060] Adjusting the amount of each component in the formulation to be optimized includes the following steps:

[0061] The number of target properties whose second quantity value in the recipe to be optimized is not within the target range is defined as a first quantity. If the first quantity is greater than a third threshold, a first component is selected from the recipe to be optimized, the first component being a component that has an impact on the target property and whose impacted target property quantity is greater than the first value. The first component is determined based on experiments. The dosage range is divided into a first range and a second range based on the current dosage of the first component in the recipe to be optimized, and the center points of the first range and the second range are obtained respectively. The dosages at the two center points are defined as a first dosage and a second dosage. On the basis of keeping the dosages of other components unchanged, two adjusted recipes are obtained based on the first dosage and the second dosage respectively, and the second quantity value corresponding to each target property in the two adjusted recipes is calculated to determine whether each second quantity value is within the target range. If not, the recipe closer to the target range is defined as the first recipe, and the dosage of the first component in the first recipe is defined as a first intermediate dosage. The first range or the second range is divided into a first interval and a second interval based on the first intermediate dosage. This step is repeated until the second quantity value corresponding to each target property is within the corresponding target range.

[0062] Specifically, it is assumed that there are 6 target properties, and the third threshold is set to 3. If the number of target properties whose second value in the formula to be optimized is not within the target range is greater than 3, the first component is selected based on the experiment. In the process of rubber product production, the vulcanization process is an important process. The vulcanization process directly affects the physical and chemical properties of rubber products, including strength, elasticity, wear resistance and aging resistance, so the sulfur element has an impact on most target properties. The component of sulfur in the yellowing-resistant xanthate polysulfide accelerator is a xanthate polysulfide accelerator, and the xanthate polysulfide accelerator is the first component. According to the current dosage of the first component, for example, 1phr, the target range of 0.5-2.0phr is divided into a first range [0.5, 1] ​​and a second range [1, 2]. In the middle of the two ranges, The center points are 0.75phr and 1.5phr. On the basis of keeping the dosages of other components unchanged, two adjusted formulas are obtained based on the first dosage and the second dosage respectively, the second value corresponding to each target performance in the two adjusted formulas is calculated, and it is determined whether each second value is within the target range. If not, the formula closer to the target value is defined as the first formula. Assuming that the formula with the first ingredient of 0.75phr is closer to the target range, the dosage of the first ingredient of 0.75phr is defined as the first intermediate dosage. Based on the first intermediate dosage of 0.75phr, the first range [0.5, 1] ​​is equally divided into a first interval [0.5, 0.75] and a second interval [0.5, 1], and this step is repeated until the second value corresponding to each target performance is within the corresponding target range.

[0063] Adjusting the amount of each component in the formula to be optimized also includes the following steps:

[0064] If the first quantity is less than the third threshold, the rating function is determined as the objective function, and the target range corresponding to each target performance is the constraint condition. Based on the constraint condition, the amount of each component in the formula to be optimized is adjusted based on the particle swarm algorithm until the second value of each target performance of the formula to be optimized meets the target range.

[0065] Specifically, if the first number is less than 3, it is used as the objective function, and the objective function evaluation value corresponding to each recipe to be optimized is calculated based on the particle swarm optimization algorithm. The particle swarm optimization (PSO) algorithm is a heuristic search algorithm that simulates the social behavior of bird flocks to find the optimal solution. In the PSO algorithm, each "particle" represents a potential solution (i.e., a recipe) and adjusts its position (i.e., the proportion of recipe ingredients) based on its own experience and the experience of the group. For each recipe to be optimized, the value of the objective function is calculated. This value reflects the performance of the current recipe relative to the target performance. The preferred recipe is adjusted within the target range to find a recipe that can minimize the target optimization function value.

[0066] like Figure 2 As shown, the present invention also provides a formula optimization system for an accelerator, which is used to implement the above method, and the system mainly includes:

[0067] The formula generation module is used to obtain the promoter required to produce the target product. The promoter includes multiple components, determine the type and dosage range of each component, combine the components in different dosages to obtain multiple candidate formulas, and produce the first product based on the candidate formulas.

[0068] The target performance quantification module is used to determine multiple target performances of the target product. The target performance is determined by the element concentrations of multiple target elements in the target product. The element concentrations of the target elements in the first product and the second product after a preset time period are obtained based on X-rays, and a first value of the target performance is generated based on the element concentration. The smaller the first value, the better the target performance of the target product.

[0069] The performance prediction module is used to build a performance prediction model based on the candidate formula and the corresponding first value. The performance prediction model predicts the current formula to generate the corresponding second value, sets a rating function, and generates evaluation values ​​of all target performances of the current formula based on the rating function.

[0070] The formula optimization module is used to: if the evaluation value is greater than the first threshold, the current formula is the preferred formula, otherwise it is the formula to be optimized, adjust the dosage of the components in the formula to be optimized, calculate the evaluation value corresponding to the adjusted formula, and repeat this step until the evaluation value is greater than the first threshold.

[0071] A device for optimizing the formulation of an accelerator comprises a processor and a memory, wherein the memory is used to store at least one computer program, and the at least one computer program is loaded by the processor and executes the above-mentioned method for optimizing the formulation of an accelerator.

[0072] The present invention also provides a computer storage medium, which stores program instructions, wherein when the program instructions are executed, the device where the computer storage medium is located is controlled to execute the above-mentioned method for optimizing the formulation of the accelerator.

[0073] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0074] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The above-mentioned program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0075] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0076] The above embodiments only express several implementation methods of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

[0077] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing the formulation of an accelerator, characterized in that: The method comprises the following steps: S1: obtaining an accelerator required for producing a target product, the accelerator comprising a plurality of components, determining the type and dosage range of each component, combining the components in different dosages to obtain a plurality of candidate formulas, and producing a first product based on the candidate formulas; S2: determining multiple target properties of the target product, wherein the target properties are determined by element concentrations of multiple target elements in the target product, obtaining element concentrations of the target elements in the first product and the second product after a preset time period based on X-rays, and generating a first value of the target property based on the element concentrations, wherein the smaller the first value, the better the target property of the target product; S3: constructing a performance prediction model based on the candidate recipe and the corresponding first value, wherein the performance prediction model predicts the current recipe to generate a corresponding second value, setting a rating function, and generating evaluation values ​​of all the target performances of the current recipe based on the rating function; S4: If the evaluation value is greater than the first threshold, the current formula is the preferred formula, otherwise it is the formula to be optimized, the dosage of the components in the formula to be optimized is adjusted, the evaluation value corresponding to the adjusted formula is calculated, and this step is repeated until the evaluation value is greater than the first threshold.

2. The method according to claim 1, characterized in that: Obtaining element concentrations involves the following steps: The first product and the second product are sampled to obtain a first sample and a second sample, respectively. An energy range of X-ray irradiation is set for each of the target elements. The first sample and the second sample are irradiated with X-rays within the energy range, respectively. The absorption of X-rays by the first sample and the second sample is defined as the element concentration of the target element in the first sample and the second sample.

3. The method according to claim 2, characterized in that Generating a first value of the target property based on the element concentration comprises the following steps: Perform two-dimensional mapping on the element concentration of the target element of the first sample and the second sample to obtain the corresponding first image and second image, identify the image areas in the first image and the second image where the concentration of the target element is higher than the second threshold, define them as the first area and the second area, calculate the area difference between the first area and the second area, set the corresponding first weight coefficient for each of the target elements, and calculate the first value α of the target performance of the target product obtained by using the candidate formula based on a first formula, the first formula is: Wherein, n is the number of target elements associated with each target performance, A i is the difference between the area of ​​the first region and the area of ​​the second region corresponding to the i-th target element, μ i is the first weight coefficient corresponding to the i-th target element.

4. The method according to claim 1, characterized in that: Generating the corresponding second quantity value comprises the following steps: A performance prediction model is established based on a linear regression equation, the performance prediction model generates a corresponding relationship between a first value of each target performance and the candidate recipe, and a second value corresponding to the target performance of each current recipe is predicted and generated based on the performance prediction model.

5. The method according to claim 4, characterized in that Setting up the rating function involves the following steps: Obtain a target range of the first value corresponding to each of the target performances, determine a target value of each of the target performances within the target range, set a second weight coefficient for each of the target performances, and the rating function is: Among them, P i is the second value of the i-th target performance, Y is the evaluation value, m is the total number of all the target performances, is the second weight coefficient of the i-th target performance, q i is the target value of the i-th target performance, f i The target range of the first value corresponding to the i-th target performance.

6. The method according to claim 5, characterized in that Adjusting the dosage of each component in the formula to be optimized comprises the following steps: The number of target properties whose second quantity value in the recipe to be optimized is not within the target range is defined as a first quantity. If the first quantity is greater than a third threshold, a first component is selected from the recipe to be optimized, the first component being a component that affects the target property and the number of target properties affected is greater than a first value. The first component is determined based on experiments. The dosage range is divided into a first range and a second range based on the current dosage of the first component in the recipe to be optimized, the center points of the first range and the second range are obtained respectively, and the dosages at the two center points are defined as a first dosage and a second dosage. On the basis of keeping the dosages of other components unchanged, two adjusted recipes are obtained based on the first dosage and the second dosage respectively, the second quantity value corresponding to each target property in the two adjusted recipes is calculated, and it is determined whether each second quantity value is within the target range. If not, the recipe closer to the target range is defined as the first recipe, the dosage of the first component in the first recipe is defined as a first intermediate dosage, and the first range or the second range is divided into a first interval and a second interval based on the first intermediate dosage. This step is repeated until the second quantity value corresponding to each target property is within the corresponding target range.

7. The method according to claim 6, characterized in that Adjusting the dosage of each component in the formula to be optimized also includes the following steps: If the first number is less than the third threshold, the rating function is determined to be the objective function, and the target range corresponding to each target performance is the constraint condition. On the basis of the constraint condition, the amount of each component in the formula to be optimized is adjusted based on the particle swarm algorithm until the second value of each target performance of the formula to be optimized meets the target range.

8. A formulation optimization system for accelerators, used to implement the method according to any one of claims 1 to 7, characterized in that: The system includes the following modules: A formula generation module, used to obtain an accelerator required for producing a target product, wherein the accelerator includes a plurality of components, determine the type and dosage range of each component, combine the components in different dosages to obtain a plurality of candidate formulas, and produce a first product based on the candidate formulas; a target performance quantification module, used to determine multiple target performances of the target product, wherein the target performance is determined by the element concentrations of multiple target elements in the target product, obtain the element concentrations of the target elements in the first product and the second product after a preset time period based on X-rays, and generate a first value of the target performance based on the element concentrations, wherein the smaller the first value, the better the target performance of the target product; A performance prediction module, configured to construct a performance prediction model based on the candidate recipes and the corresponding first value, wherein the performance prediction model predicts the current recipe to generate the corresponding second value, sets a rating function, and generates evaluation values ​​of all the target performances of the current recipe based on the rating function; A recipe optimization module is used to: if the evaluation value is greater than a first threshold, the current recipe is the preferred recipe; otherwise, it is the recipe to be optimized; the dosage of the components in the recipe to be optimized is adjusted; the evaluation value corresponding to the adjusted recipe is calculated; and this step is repeated until the evaluation value is greater than the first threshold.

9. A formulation optimization device for accelerators, characterized in that: The device comprises a processor and a memory, wherein the memory is used to store at least one computer program, and the at least one computer program is loaded by the processor to execute the method according to any one of claims 1 to 7.

10. A computer storage medium, characterized in that: The computer storage medium stores program instructions, wherein when the program instructions are executed, the device where the computer storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

Images