A method and apparatus for optimizing a mica composite formulation

Abstract

The application provides a mica composite material formula optimization method and device, relates to the mica material optimization technical field, and includes the following steps: obtaining formula data and performance data of an initial mica composite material, performing correlation analysis on the formula data and the performance data, and obtaining core data; setting an initial performance threshold of the performance data, adjusting the initial performance threshold through an entropy weight algorithm, obtaining a final performance threshold, and setting a final performance constraint according to the final performance threshold; constructing a multi-objective optimization model according to the core data and the final performance constraint, iteratively solving the multi-objective optimization model through a chaotic sparrow search algorithm, and obtaining an initial optimized material formula; obtaining target scene parameters, constructing a machine learning random forest model, correcting the initial optimized material formula through the machine learning random forest model based on the target scene parameters, and obtaining a final optimized material formula.

Description

Technical Field

[0001] This application relates to the field of mica material optimization technology, and in particular to a method and apparatus for optimizing the formulation of mica composite materials. Background Technology

[0002] Mica composites, due to their excellent mechanical properties, dielectric properties, and temperature stability, are widely used in electrical insulation, mechanical structures, and high-temperature applications. The rationality of their formulation directly determines the material's application effect and market competitiveness. However, current methods for optimizing mica composite formulations still face many unresolved issues, severely hindering the improvement of material performance and the efficiency of industrial applications.

[0003] Existing optimization methods fail to systematically analyze the correlation between formulation data (such as raw material composition and preparation process) and performance data (such as mechanical strength and dielectric strength), resulting in an inability to identify the core data affecting mica composites. Furthermore, the lack of constraints on performance data and consideration of multiple optimization objectives during optimization model construction leads to poor formulation optimization results. The limited optimization capabilities of existing algorithms result in poor stability of the optimized formulations and limited performance improvements. Traditional optimization methods also fail to consider the impact of the target environment on material properties, leading to significant deviations between actual and theoretical performance in specific scenarios, insufficient stability, and an inability to meet the application requirements of multiple scenarios. Summary of the Invention

[0004] To address the aforementioned problems, in a first aspect, the present invention provides a method for optimizing the formulation of mica composite materials, comprising: Obtain initial formulation and performance data of mica composite materials, perform correlation analysis on the formulation and performance data, and obtain core data; Set an initial performance threshold for the performance data, adjust the initial performance threshold using an entropy weight algorithm to obtain a final performance threshold, and set final performance constraints based on the final performance threshold. A multi-objective optimization model is constructed based on core data and final performance constraints. The multi-objective optimization model is iteratively solved using the chaotic sparrow search algorithm to obtain the initial optimized material formulation. Obtain the target scene parameters, construct a machine learning random forest model, and revise the initial optimized material formula based on the target scene parameters using the machine learning random forest model to obtain the final optimized material formula.

[0005] Optionally, the formulation data includes raw material component data and preparation process data; The raw material composition data includes the average particle size of mica, mica purity, type of reinforcing phase, content of reinforcing phase, type of binder phase and binder phase ratio. The preparation process data includes mixing speed, mixing time, molding pressure, sintering temperature and holding time. The performance data includes mechanical strength, dielectric strength, resistivity and temperature resistance.

[0006] Optionally, the correlation analysis of the formulation data and performance data to obtain core data includes: The Pearson correlation coefficients and significance test values ​​between the average particle size of mica, mica purity, type of reinforcing phase, content of reinforcing phase, type of binder phase, binder phase ratio, mixing speed, mixing time, molding pressure, sintering temperature and holding time in the formulation data and the mechanical strength, dielectric strength, resistivity and temperature resistance in the performance data were calculated respectively. Formula data whose absolute values ​​of Pearson correlation coefficients are all greater than or equal to the first threshold and whose significance test values ​​are all less than the second threshold are used as core data.

[0007] Optionally, the step of setting an initial performance threshold for performance data, adjusting the initial performance threshold using an entropy weight algorithm to obtain a final performance threshold, and setting final performance constraints based on the final performance threshold includes: Set initial performance thresholds for performance data. Initial performance thresholds include initial mechanical strength threshold, initial dielectric strength threshold, initial resistivity threshold, and initial temperature resistance threshold. The entropy values ​​of mechanical strength, dielectric strength, resistivity, and temperature resistance are obtained by calculating the entropy values ​​using the entropy weight algorithm. The initial mechanical strength threshold, initial dielectric strength threshold, initial resistivity threshold, and initial temperature resistance threshold are adjusted according to the mechanical strength entropy value, dielectric strength entropy value, resistivity entropy value, and temperature resistance entropy value, respectively, to obtain the final performance threshold. The final performance threshold includes the final mechanical strength threshold, the final dielectric strength threshold, the final resistivity threshold, and the final temperature resistance threshold. The final performance constraints are defined as mechanical strength, dielectric strength, resistivity, and temperature resistance being greater than the final mechanical strength threshold, final dielectric strength threshold, final resistivity threshold, and final temperature resistance threshold, respectively.

[0008] Optionally, the step of constructing a multi-objective optimization model based on core data and final performance constraints includes: Based on the core data, a performance objective function and a manufacturing cost function are constructed. The performance objective function includes a mechanical strength prediction function, a dielectric strength prediction function, a resistivity prediction function, and a temperature resistance prediction function. Obtain the value range of the core data, use the core data that is within the value range as the core data constraint, and use the core data constraint and the final performance constraint as the constraint conditions; Based on the constraints, a multi-objective optimization model is constructed with the optimization objectives of maximizing the performance objective function and minimizing the preparation cost function as the optimization goals.

[0009] Optionally, the step of iteratively solving the multi-objective optimization model using the chaotic sparrow search algorithm to obtain the initial optimized material formula includes: S1: Generate a population of multi-objective optimization models using the Logistic chaotic mapping algorithm, where individuals in the population are candidate material formulations; S2: Divide individuals in the population into discoverers, followers, and early warning systems according to a preset ratio; obtain exploration factors and construct a position function based on the exploration factors; S3: Update the discoverer's current position using a position function to obtain the discoverer's updated position; S4: Update the current position of the follower based on the updated position of the discoverer, and obtain the updated position of the follower; S5: Calculate individual fitness, population fitness fluctuation range, and population fitness change rate based on the updated position of the discoverer, the updated position of the follower, and the current position of the alerter; S6: If the rate of change of population fitness is less than the first preset value, then update the current position of the alerter using a normal distribution random function; otherwise, retain the current position of the alerter. S7: Repeat steps S3-S6 until the population fitness fluctuation amplitude converges to the second preset value, and use the candidate material formula corresponding to the individual with the highest individual fitness as the initial optimized material formula.

[0010] Optionally, the step of refining the initial optimized material formulation based on the target scene parameters using a machine learning random forest model to obtain the final optimized material formulation includes: Feature extraction is performed on the target scene parameters to obtain scene feature vectors. The scene feature vectors are then input into a machine learning random forest model for bias prediction to obtain the mechanical strength bias, dielectric strength bias, resistivity bias, and temperature resistance bias of the initial optimized material formulation under the target scene. Obtain the parameter correction mapping database, and retrieve the correction parameter vector from the parameter correction mapping database based on the mechanical strength deviation, dielectric strength deviation, resistivity deviation, and temperature resistance deviation; The initial optimized material formulation is modified by adjusting the parameter vector to obtain the final optimized material formulation.

[0011] Secondly, the present invention provides a mica composite material formulation optimization device for implementing the aforementioned mica composite material formulation optimization method, the device comprising: The core data acquisition module is used to acquire the initial formulation and performance data of mica composite materials, perform correlation analysis on the formulation and performance data, and obtain core data. The final performance constraint acquisition module is used to set the initial performance threshold of the performance data, adjust the initial performance threshold through the entropy weight algorithm to obtain the final performance threshold, and set the final performance constraint based on the final performance threshold. The initial optimized material formulation acquisition module is used to construct a multi-objective optimization model based on core data and final performance constraints, and to iteratively solve the multi-objective optimization model using a chaotic sparrow search algorithm to obtain the initial optimized material formulation. The final optimized material formulation acquisition module is used to obtain target scene parameters, construct a machine learning random forest model, and revise the initial optimized material formulation based on the target scene parameters to obtain the final optimized material formulation.

[0012] Thirdly, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the mica composite material formulation optimization method.

[0013] Fourthly, the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the mica composite material formulation optimization method.

[0014] The present invention has the following beneficial effects: 1. By performing correlation analysis on formulation and performance data, redundant parameters with negligible impact on performance are eliminated, significantly reducing the computational load of optimization while ensuring that no core influencing factors are overlooked, thus improving the efficiency and accuracy of subsequent optimization. The importance of performance indicators is quantified using an entropy weight algorithm, and the initial performance threshold is dynamically adjusted to ensure that the final performance constraints both meet actual needs and balance optimization difficulty, avoiding the problems of over-constraint or under-constraint caused by traditional fixed thresholds. A multi-objective optimization model for maximizing performance and minimizing cost is constructed based on core data and final performance constraints. This ensures that core performance indicators such as mechanical strength and dielectric strength meet standards while controlling raw material and process costs through a cost function, solving the problem of excessively high costs caused by single performance optimization. The multi-objective optimization model is iteratively solved using a chaotic sparrow search algorithm. The chaotic sparrow search algorithm has strong global search capabilities and fast convergence speed, achieving optimal material formulation. A machine learning random forest model is used to predict scenario deviations, and targeted adjustments are made using a parameter correction mapping database to adapt the formulation to the environmental and operating conditions of the target scenario, ensuring that the final optimized material formulation is suitable for the target application scenario.

[0015] 2. The multi-objective optimization model consists of a performance objective function and a fabrication cost function. The performance objective function covers four core properties: mechanical strength, dielectric strength, resistivity, and temperature resistance. Through multi-dimensional prediction functions, it accurately quantifies the correlation between performance and core data, achieving simultaneous maximization of multiple properties and solving the problem of insufficient material application adaptability caused by traditional single-performance optimization. The fabrication cost function incorporates costs such as raw material procurement and process energy consumption into the optimization objective, avoiding the difficulty of industrializing formulations due to prioritizing performance over cost, and achieving an optimal balance between high performance and low cost. Core data constraints limit the reasonable range of parameter values, avoiding invalid searches. Finally, the performance constraints are aligned with the priority of actual needs. The combination of these two aspects makes the constraints precise and controllable, reducing the blindness of optimization.

[0016] 3. The Chaotic Sparrow Search algorithm generates a population through Logistic chaotic mapping. The ergodicity and randomness of the chaotic sequence ensure that candidate recipes evenly cover the core data value range, avoiding blind spots in optimization caused by initial individual concentration. At the same time, redundant individuals are removed through validity verification to ensure population quality and provide a sufficient and high-quality search starting point for subsequent global optimization. Discoverers, followers, and early warning providers are divided according to a preset ratio to form a collaborative mechanism of exploration-follow-up-obstacle avoidance. The linearly decreasing exploration factor dynamically adjusts the search intensity with iteration, focusing on global exploration in the early stage and local development in the later stage. By calculating the individual fitness, the fluctuation range and change rate of population fitness, the progress and status of population optimization are accurately quantified, significantly improving the algorithm's global optimization ability and ensuring the optimality of the initial optimized recipe. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a method according to an embodiment of the present invention; Figure 2 This is a structural diagram of the device according to an embodiment of the present invention. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0020] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0021] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0022] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0023] Reference Figure 1 This invention provides a method for optimizing the formulation of mica composite materials, comprising: S100 acquires the initial formulation and performance data of mica composite materials, performs correlation analysis on the formulation and performance data, and obtains core data.

[0024] In some embodiments, the formulation data includes raw material component data and preparation process data; The raw material composition data includes the average particle size of mica, mica purity, type of reinforcing phase, content of reinforcing phase, type of binder phase and binder phase ratio. The preparation process data includes mixing speed, mixing time, molding pressure, sintering temperature and holding time. The performance data includes mechanical strength, dielectric strength, resistivity and temperature resistance.

[0025] In some embodiments, formulation data is a collection of key parameters characterizing the composition of raw materials and the preparation process of mica composite materials. It is clearly divided into two categories: raw material component data and preparation process data. Both types of data jointly determine the final performance of the material and must be collected and recorded in accordance with standardized specifications.

[0026] The average particle size of mica refers to the average equivalent particle size of mica particles, measured using a laser particle size analyzer. Before measurement, the mica sample is dispersed in anhydrous ethanol and ultrasonically treated for 10 minutes to ensure uniform particle dispersion. The purity of mica refers to the mass fraction of the effective components in the mica raw material, detected by X-ray fluorescence spectroscopy, and quantitatively analyzed by comparing the spectral intensity with that of standard samples. The type of reinforcing phase is selected from commonly used industrial-grade reinforcing phases, categorized according to performance requirements. The content of the reinforcing phase refers to the proportion of the reinforcing phase to the total mass of the composite material, ranging from 5% to 30%, and adjusted according to the reinforcement target. The type of binder is a thermosetting resin binder, selected according to the compatibility with the application environment and process. The binder ratio refers to the mass ratio of the matrix resin to curing agents, diluents, and other additives in the binder phase, standardized according to the resin type.

[0027] Mixing speed refers to the rotational speed of the mixing paddle in the mixing equipment, measured in r / min; mixing time refers to the total time from material input to uniform mixing, measured in min; molding pressure refers to the pressure exerted on the material during molding in the mold, measured in MPa; sintering temperature refers to the temperature at which the material is heat-treated in a high-temperature furnace after molding, measured in ℃, with a range of 800℃-1200℃; holding time refers to the duration of maintaining the set temperature during sintering, measured in h, with a range of 1h-5h.

[0028] Performance data are the core indicators for evaluating the application suitability of mica composite materials and must be tested according to national standards or industry specifications to ensure data accuracy and comparability. Mechanical strength is tested using a universal testing machine, including tensile strength, flexural strength, and impact strength tests; dielectric strength is tested according to GB / T1408.1-2016 standard, with units of kV / mm; resistivity is tested according to GB / T1410-2006 standard, using a high-resistivity meter to measure volume resistivity, with units of Ω×m. Temperature resistance is tested according to GB / T11026.1-2014 standard, including long-term service temperature testing and short-term withstand temperature testing.

[0029] In some embodiments, the correlation analysis of formulation data and performance data to obtain core data includes: The Pearson correlation coefficients and significance test values ​​between the average particle size of mica, mica purity, type of reinforcing phase, content of reinforcing phase, type of binder phase, binder phase ratio, mixing speed, mixing time, molding pressure, sintering temperature and holding time in the formulation data and the mechanical strength, dielectric strength, resistivity and temperature resistance in the performance data were calculated respectively. Formula data whose absolute values ​​of Pearson correlation coefficients are all greater than or equal to the first threshold and whose significance test values ​​are all less than the second threshold are used as core data.

[0030] In some embodiments, at least 30 sets of valid sample data are collected, and a two-dimensional data table is established according to the formula data-performance data. Each row corresponds to one set of complete experimental data, and the columns contain 11 formula data items and 4 performance data items. Continuous data directly retains the original values ​​and uses a unified unit. Categorical data uses dummy variable encoding.

[0031] The calculation tool is Python. The scipy.stats library of Python is imported. Each formula data x and performance data y is traversed. The pearsonr(x,y) function is called to output the Pearson correlation coefficient r and the significance test value p. The value range of r is [-1,1] and the value range of p is [0,1].

[0032] Output results organization: Construct a four-column result table consisting of formula data, performance data, r-value, and p-value. For example, mica average particle size - mechanical strength - r=0.72 - p=0.003.

[0033] The first threshold is set to 0.5 by default and can be adjusted according to the scenario: it can be increased to 0.6 for high-precision optimization scenarios and reduced to 0.4 for general scenarios; the larger the absolute value of r, the stronger the linear correlation between the formula data and the performance data. |r|≥ the first threshold indicates that the formula data has a significant impact on the performance.

[0034] The second threshold is uniformly set to 0.05, and can be set to 0.01 in strict scenarios; the smaller the p-value, the higher the statistical reliability of the correlation coefficient. p < the second threshold indicates that the association is not random and has statistical significance.

[0035] S200 sets the initial performance threshold for performance data, adjusts the initial performance threshold using an entropy weight algorithm to obtain the final performance threshold, and sets the final performance constraint based on the final performance threshold.

[0036] In some embodiments, setting an initial performance threshold for performance data, adjusting the initial performance threshold using an entropy weight algorithm to obtain a final performance threshold, and setting final performance constraints based on the final performance threshold includes: Set initial performance thresholds for performance data. Initial performance thresholds include initial mechanical strength threshold, initial dielectric strength threshold, initial resistivity threshold, and initial temperature resistance threshold. The entropy values ​​of mechanical strength, dielectric strength, resistivity, and temperature resistance are obtained by calculating the entropy values ​​using the entropy weight algorithm. The initial mechanical strength threshold, initial dielectric strength threshold, initial resistivity threshold, and initial temperature resistance threshold are adjusted according to the mechanical strength entropy value, dielectric strength entropy value, resistivity entropy value, and temperature resistance entropy value, respectively, to obtain the final performance threshold. The final performance threshold includes the final mechanical strength threshold, the final dielectric strength threshold, the final resistivity threshold, and the final temperature resistance threshold. The final performance constraints are defined as mechanical strength, dielectric strength, resistivity, and temperature resistance being greater than the final mechanical strength threshold, final dielectric strength threshold, final resistivity threshold, and final temperature resistance threshold, respectively.

[0037] In some embodiments, the core of the entropy weight algorithm is to quantify the importance weight of each indicator by the dispersion of performance data. The greater the data dispersion, the smaller the entropy value, and the higher the weight, indicating that the indicator has a more critical impact on formula optimization. A data matrix is ​​constructed from the collected n sets of performance data. i=1-n are sample numbers, and j=1-4 correspond to mechanical strength, dielectric strength, resistivity and temperature resistance.

[0038] The entropy value of each performance data point is calculated using the following formula: in, Let the entropy value be the value of the j-th type of performance data. Entropy The smaller the value, the greater the dispersion of the indicator, and the higher its importance. Let be the standardized probability of the performance data of the j-th class for the i-th sample.

[0039] The entropy weights for each performance data point are calculated using the following formula: in, For the entropy weight of the j-th type of performance data, , , The larger the value, the higher the weight of the indicator.

[0040] Based on entropy weight adjustment of the initial threshold, the final performance threshold is obtained, and performance metrics with higher logical weights are adjusted accordingly. The larger the value, the more critical its impact on formulation optimization, requiring a higher threshold to strengthen the constraint; for indicators with lower weights, the threshold can be relaxed or maintained to balance the optimization difficulty. The adjustment formula is: in, For the final performance threshold of the j-th type of performance data, is the initial performance threshold; k is the adjustment coefficient, which defaults to 0.2 and can be adjusted according to the scenario. For high-precision scenarios, k=0.3, and for general scenarios, k=0.1.

[0041] The final performance constraint is that all four performance metrics must exceed their respective final thresholds, i.e.: in, For actual mechanical strength, For actual dielectric strength, This is the actual resistivity. This refers to the actual temperature resistance.

[0042] The final performance constraint serves as a hard constraint for the subsequent multi-objective optimization model. That is, during the iterative solution process, only candidate recipes that satisfy all the above inequalities are retained to ensure that the optimization results meet the performance requirements.

[0043] S300 constructs a multi-objective optimization model based on core data and final performance constraints, and iteratively solves the multi-objective optimization model using a chaotic sparrow search algorithm to obtain the initial optimized material formulation.

[0044] In some embodiments, constructing a multi-objective optimization model based on core data and final performance constraints includes: Based on the core data, a performance objective function and a manufacturing cost function are constructed. The performance objective function includes a mechanical strength prediction function, a dielectric strength prediction function, a resistivity prediction function, and a temperature resistance prediction function. Obtain the value range of the core data, use the core data that is within the value range as the core data constraint, and use the core data constraint and the final performance constraint as the constraint conditions; Based on the constraints, a multi-objective optimization model is constructed with the optimization objectives of maximizing the performance objective function and minimizing the preparation cost function as the optimization goals.

[0045] In some embodiments, at least 30 sets of samples containing core data and corresponding real performance data are selected, with the core data used as feature variables. , For the m-th core data point, four performance metrics are used as label variables. Four independent random forest models are constructed using Python's scikit-learn library, each corresponding to one of the four performance metrics. After training, the models are encapsulated as performance prediction functions, as shown in the following expression: Mechanical strength prediction function: ,in The predicted mechanical strength output by the random forest model; Dielectric strength prediction function: ,in This represents the predicted dielectric strength output by the random forest model. Resistivity prediction function: ,in The resistivity prediction value output by the random forest model; Temperature resistance prediction function: ,in These are the predicted temperature tolerance values ​​output by the random forest model.

[0046] The preparation cost function covers raw material procurement costs and process energy consumption costs, and is constructed based on the quantitative relationships of core data, including: mica raw material cost , For mica density, This represents the percentage of mica used. This is the unit price of mica.

[0047] Enhanced phase cost , To increase phase content, To increase the unit price.

[0048] Cost of binder phase , For the binder phase ratio, This is the unit price of the binder phase.

[0049] Hybrid energy costs , For mixed time, The power corresponding to the mixed speed. The price is the unit energy consumption electricity price.

[0050] Molding cost F is the molding pressure, and S is the molding area. This is the unit pressure equipment loss coefficient.

[0051] Sintering energy consumption cost , For heat preservation time, The sintering temperature is... This is the energy consumption coefficient per unit temperature.

[0052] Total cost function integration: K represents fixed costs, such as labor and packaging, which are amortized over batch production.

[0053] The upper and lower limits of each core data point are determined based on historical experimental data extremes, industrial production process limits, and raw material supply specifications; for each core data point... ,set up The binder phase ratio satisfies Mixing time The final performance threshold determined based on the entropy weight algorithm. ,set up: , , , .

[0054] Under the premise of satisfying all constraints, the multi-objective optimization model aims to maximize performance and minimize cost. Its mathematical expression is as follows: In some embodiments, the step of iteratively solving the multi-objective optimization model using a chaotic sparrow search algorithm to obtain an initial optimized material formulation includes: S1: Generate a population of multi-objective optimization models using the Logistic chaotic mapping algorithm, where individuals in the population are candidate material formulations; S2: Divide individuals in the population into discoverers, followers, and early warning systems according to a preset ratio; obtain exploration factors and construct a position function based on the exploration factors; S3: Update the discoverer's current position using a position function to obtain the discoverer's updated position; S4: Update the current position of the follower based on the updated position of the discoverer, and obtain the updated position of the follower; S5: Calculate individual fitness, population fitness fluctuation range, and population fitness change rate based on the updated position of the discoverer, the updated position of the follower, and the current position of the alerter; S6: If the rate of change of population fitness is less than the first preset value, then update the current position of the alerter using a normal distribution random function; otherwise, retain the current position of the alerter. S7: Repeat steps S3-S6 until the population fitness fluctuation amplitude converges to the second preset value, and use the candidate material formula corresponding to the individual with the highest individual fitness as the initial optimized material formula.

[0055] In some embodiments, a core data list and value range are imported, a chaotic matrix is ​​generated using the Logistic chaotic mapping formula, and the chaotic matrix is ​​converted into a population of the recipe according to the mapping rules.

[0056] The expression for the discoverer's location function is: in, Let t+1 be the update position of the discoverer. Let t be the current position of the discoverer. Let be the exploration factor at time t. Let be a standard normally distributed random number at time t. Let t be the position of the globally optimal individual at time t.

[0057] The expression for the position function of the follower is: in, Let be the updated position of the follower at time t+1. Let be the current position of the follower at time t. The update position of the discoverer at time t+1 is randomly selected; The expression for the location function of the early warning system is: in, The updated position of the alerter at time t+1. Let t be the current position of the person issuing the warning. For the intensity of variation, is a random number that follows a normal distribution with respect to the intensity of variation.

[0058] Performance target weighted sum The expression is: in, For performance weights.

[0059] Cost target normalization , The maximum preparation cost for the population.

[0060] The formula for calculating individual fitness F is: in, This indicates the adjustment parameter.

[0061] Population fitness fluctuation range ,in Let be the average fitness of the population at time t. The average fitness of the population at time t-1, and the rate of change of population fitness. .

[0062] The S400 acquires the target scene parameters, constructs a machine learning random forest model, and corrects the initial optimized material formula based on the target scene parameters using the machine learning random forest model to obtain the final optimized material formula.

[0063] In some embodiments, the step of refining the initial optimized material formulation based on the target scene parameters using a machine learning random forest model to obtain the final optimized material formulation includes: Feature extraction is performed on the target scene parameters to obtain scene feature vectors. The scene feature vectors are then input into a machine learning random forest model for bias prediction to obtain the mechanical strength bias, dielectric strength bias, resistivity bias, and temperature resistance bias of the initial optimized material formulation under the target scene. Obtain the parameter correction mapping database, and retrieve the correction parameter vector from the parameter correction mapping database based on the mechanical strength deviation, dielectric strength deviation, resistivity deviation, and temperature resistance deviation; The initial optimized material formulation is modified by adjusting the parameter vector to obtain the final optimized material formulation.

[0064] In some embodiments, the feature vector of the target scene Input a random forest machine learning model and output four performance deviations of the initial optimized recipe in this scenario. .

[0065] Based on the prediction First, determine the interval to which each deviation belongs; then, based on the interval combination, search the database to obtain the corresponding correction parameter vector. If multiple matching intervals are found, the vector with the lowest preparation cost after correction is taken as the optimal solution, and the final core data is obtained. ,in, Core data for initial optimization of material formulation. For correction factors; Substitute the values ​​into the performance objective function to calculate the theoretical performance of the modified formulation, ensuring that all four performance parameters meet the final performance constraints and that the performance deviation in the scenario is ≤0.05. If the cost exceeds the limit, a new correction vector is retrieved. After passing the triple verification of constraints, performance, and cost, The corresponding complete formula is the final optimized material formula.

[0066] Reference Figure 2 This invention provides a mica composite material formulation optimization device 20, used to implement a mica composite material formulation optimization method. The device includes: The core data acquisition module 21 is used to acquire the initial mica composite material formulation data and performance data, perform correlation analysis on the formulation data and performance data, and obtain core data. The final performance constraint acquisition module 22 is used to set the initial performance threshold of the performance data, adjust the initial performance threshold through the entropy weight algorithm to obtain the final performance threshold, and set the final performance constraint according to the final performance threshold. The initial optimized material formulation acquisition module 23 is used to construct a multi-objective optimization model based on core data and final performance constraints, and to iteratively solve the multi-objective optimization model using a chaotic sparrow search algorithm to obtain the initial optimized material formulation. The final optimized material formula acquisition module 24 is used to acquire target scene parameters, construct a machine learning random forest model, and correct the initial optimized material formula based on the target scene parameters through the machine learning random forest model to obtain the final optimized material formula.

[0067] This application provides an electronic device, including a processor and a memory; the memory stores a computer program, wherein the computer program, when executed by the processor, implements the mica composite material formulation optimization method of any of the above schemes.

[0068] Specifically, the processor may include, for example, a general-purpose microprocessor, an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor may also include onboard memory for caching purposes. The processor may be a single processing unit or multiple processing units for performing different actions of the method flow according to embodiments of this application.

[0069] Memory can be any medium capable of containing, storing, transmitting, propagating, or transmitting instructions. For example, memory can include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, instruments, or propagation media. Specific examples of memory include: magnetic storage devices such as magnetic tape or hard disk drives (HDDs); optical storage devices such as optical discs (CD-ROMs); and also random access memory (RAM) or flash memory; and / or wired / wireless communication links.

[0070] This application also provides a computer-readable medium storing a computer program thereon, which, when executed by a processor, implements the mica composite material formulation optimization method of any of the above-described schemes. This computer-readable medium may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The aforementioned computer-readable medium carries one or more programs, which, when executed, implement the method as described in the embodiments of this application.

[0071] According to embodiments of this application, a computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: 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 application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal 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 can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wired, optical fiber, radio frequency signals, etc., or any suitable combination thereof.

[0072] Those skilled in the art will understand that the features described in the various embodiments and / or claims of this application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in this application. In particular, the features described in the various embodiments and / or claims of this application can be combined and / or combined in various ways without departing from the spirit and teachings of this application. All such combinations and / or combinations fall within the scope of this application. Therefore, the scope of this application should not be limited to the above embodiments, but should be defined not only by the appended claims, but also by their equivalents. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for optimizing the formulation of mica composite materials, characterized in that, include: Obtain initial formulation and performance data of mica composite materials, perform correlation analysis on the formulation and performance data, and obtain core data; Set an initial performance threshold for the performance data, adjust the initial performance threshold using an entropy weight algorithm to obtain a final performance threshold, and set final performance constraints based on the final performance threshold. A multi-objective optimization model is constructed based on core data and final performance constraints. The multi-objective optimization model is iteratively solved using the chaotic sparrow search algorithm to obtain the initial optimized material formulation. Obtain the target scene parameters, construct a machine learning random forest model, and revise the initial optimized material formula based on the target scene parameters using the machine learning random forest model to obtain the final optimized material formula.

2. The method for optimizing the formulation of mica composite materials according to claim 1, characterized in that, The formulation data includes raw material component data and preparation process data; The raw material composition data includes the average particle size of mica, mica purity, type of reinforcing phase, content of reinforcing phase, type of binder phase and binder phase ratio. The preparation process data includes mixing speed, mixing time, molding pressure, sintering temperature and holding time. The performance data includes mechanical strength, dielectric strength, resistivity and temperature resistance.

3. The method for optimizing the formulation of mica composite materials according to claim 1, characterized in that, The correlation analysis of the formulation data and performance data to obtain core data includes: The Pearson correlation coefficients and significance test values ​​between the average particle size of mica, mica purity, type of reinforcing phase, content of reinforcing phase, type of binder phase, binder phase ratio, mixing speed, mixing time, molding pressure, sintering temperature and holding time in the formulation data and the mechanical strength, dielectric strength, resistivity and temperature resistance in the performance data were calculated respectively. Formula data whose absolute values ​​of Pearson correlation coefficients are all greater than or equal to the first threshold and whose significance test values ​​are all less than the second threshold are used as core data.

4. The method for optimizing the formulation of mica composite materials according to claim 1, characterized in that, The initial performance threshold for setting performance data is adjusted using an entropy weight algorithm to obtain a final performance threshold. Final performance constraints are then set based on this final performance threshold, including: Set initial performance thresholds for performance data. Initial performance thresholds include initial mechanical strength threshold, initial dielectric strength threshold, initial resistivity threshold, and initial temperature resistance threshold. The entropy values ​​of mechanical strength, dielectric strength, resistivity, and temperature resistance are obtained by calculating the entropy values ​​using the entropy weight algorithm. The initial mechanical strength threshold, initial dielectric strength threshold, initial resistivity threshold, and initial temperature resistance threshold are adjusted according to the mechanical strength entropy value, dielectric strength entropy value, resistivity entropy value, and temperature resistance entropy value, respectively, to obtain the final performance threshold. The final performance threshold includes the final mechanical strength threshold, the final dielectric strength threshold, the final resistivity threshold, and the final temperature resistance threshold. The final performance constraints are defined as mechanical strength, dielectric strength, resistivity, and temperature resistance being greater than the final mechanical strength threshold, final dielectric strength threshold, final resistivity threshold, and final temperature resistance threshold, respectively.

5. The method for optimizing the formulation of mica composite materials according to claim 1, characterized in that, The construction of a multi-objective optimization model based on core data and final performance constraints includes: Based on the core data, a performance objective function and a manufacturing cost function are constructed. The performance objective function includes a mechanical strength prediction function, a dielectric strength prediction function, a resistivity prediction function, and a temperature resistance prediction function. Obtain the value range of the core data, use the core data that is within the value range as the core data constraint, and use the core data constraint and the final performance constraint as the constraint conditions; Based on the constraints, a multi-objective optimization model is constructed with the optimization objectives of maximizing the performance objective function and minimizing the preparation cost function as the optimization goals.

6. The method for optimizing the formulation of mica composite materials according to claim 1, characterized in that, The step of iteratively solving the multi-objective optimization model using the chaotic sparrow search algorithm to obtain the initial optimized material formula includes: S1: Generate a population of multi-objective optimization models using the Logistic chaotic mapping algorithm, where individuals in the population are candidate material formulations; S2: Divide individuals in the population into discoverers, followers, and early warning systems according to a preset ratio; obtain exploration factors and construct a position function based on the exploration factors; S3: Update the discoverer's current position using a position function to obtain the discoverer's updated position; S4: Update the current position of the follower based on the updated position of the discoverer, and obtain the updated position of the follower; S5: Calculate individual fitness, population fitness fluctuation range, and population fitness change rate based on the updated position of the discoverer, the updated position of the follower, and the current position of the alerter; S6: If the rate of change of population fitness is less than the first preset value, then update the current position of the alerter using a normal distribution random function; otherwise, retain the current position of the alerter. S7: Repeat steps S3-S6 until the population fitness fluctuation amplitude converges to the second preset value, and use the candidate material formula corresponding to the individual with the highest individual fitness as the initial optimized material formula.

7. The method for optimizing the formulation of mica composite materials according to claim 1, characterized in that, The process of refining the initial optimized material formulation based on target scene parameters using a machine learning random forest model to obtain the final optimized material formulation includes: Feature extraction is performed on the target scene parameters to obtain scene feature vectors. The scene feature vectors are then input into a machine learning random forest model for bias prediction to obtain the mechanical strength bias, dielectric strength bias, resistivity bias, and temperature resistance bias of the initial optimized material formulation under the target scene. Obtain the parameter correction mapping database, and retrieve the correction parameter vector from the parameter correction mapping database based on the mechanical strength deviation, dielectric strength deviation, resistivity deviation, and temperature resistance deviation; The initial optimized material formulation is modified by adjusting the parameter vector to obtain the final optimized material formulation.

8. A mica composite material formulation optimization device, used to implement the mica composite material formulation optimization method as described in any one of claims 1 to 7, characterized in that, The device includes: The core data acquisition module is used to acquire the initial formulation and performance data of mica composite materials, perform correlation analysis on the formulation and performance data, and obtain core data. The final performance constraint acquisition module is used to set the initial performance threshold of the performance data, adjust the initial performance threshold through the entropy weight algorithm to obtain the final performance threshold, and set the final performance constraint based on the final performance threshold. The initial optimized material formulation acquisition module is used to construct a multi-objective optimization model based on core data and final performance constraints, and to iteratively solve the multi-objective optimization model using a chaotic sparrow search algorithm to obtain the initial optimized material formulation. The final optimized material formulation acquisition module is used to obtain target scene parameters, construct a machine learning random forest model, and revise the initial optimized material formulation based on the target scene parameters to obtain the final optimized material formulation.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the mica composite material formulation optimization method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the mica composite material formulation optimization method as described in any one of claims 1 to 7.

Images