Insulation resin performance optimization method based on big data analysis
Optimizing the defoaming process of insulating resin through big data analysis methods, solving the randomness and blindness of the defoaming process optimization analysis in the prior art, improving accuracy and efficiency, and achieving more efficient optimization of the performance of insulating resin.
Patent Information
- Application Number
- CN202410831313.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-26
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2044-06-26
AI Technical Summary
The optimization analysis of the defoaming process of existing insulating resins depends on experimental trial and error, and there is great randomness and blindness, which leads to a long time and difficulty in ensuring accuracy and accuracy.
Using a method based on big data analysis, data mining and screening are carried out by obtaining environmental interference factors, applying electrical components and resin attribute information, building a space for optimizing the parameters of defoaming parameters, and using simplified defoaming quality evaluation function to optimize the parameters and output optimization. The defoaming parameters are collected, and the quality evaluation is carried out through the standard defoaming quality evaluation function to determine the optimal parameters.
The randomness and blindness of defoaming parameters are reduced, the intelligence of analysis is improved, time is saved, accuracy and accuracy are enhanced, and more efficient defoaming parameters are achieved.
Smart Images

Figure CN118609733B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of insulating resin preparation, and in particular to a method for optimizing insulating resin performance based on big data analysis. Background Art
[0002] Epoxy resin is a widely used insulating material and is increasingly used in electrical equipment. However, the formation of bubbles is a common problem in the preparation of epoxy resin, which can affect the electrical insulation, mechanical properties and appearance quality of the resin.
[0003] At present, the existing methods for optimizing the degassing process of insulating resin often rely on experimental trial and error, which has the defects of large randomness and blindness in optimization and takes a long time. At the same time, the precision and accuracy of the optimization results are difficult to guarantee. Summary of the invention
[0004] The purpose of this application is to provide a method for optimizing the performance of insulating resins based on big data analysis, so as to solve the technical problems that the existing methods for optimizing the degassing process of insulating resins often rely on experimental trial and error, have the defects of large randomness and blindness in optimization, take a long time, and the precision and accuracy of the optimization results are difficult to guarantee.
[0005] In view of the above problems, the present application provides an insulating resin performance optimization method based on big data analysis, wherein the method includes: collecting environmental interference factors of the target insulating resin based on predetermined influencing indicators to obtain an environmental interference factor set; reading the applied electrical components and resin property information of the target insulating resin, and using the environmental interference factor set, the applied electrical components and the resin property information as constraints to perform data mining based on industrial big data to obtain a sample data set, wherein the resin property information includes a predetermined raw material ratio and a predetermined mixing process parameter; screening the sample data set according to predetermined constraints, extracting sample degassing parameters from the sample data that meet the predetermined constraints, and constructing a degassing parameter optimization space, wherein the predetermined constraints include a predetermined service life and a predetermined electrical failure frequency; based on the degassing parameter optimization space, optimizing the degassing parameters using a simplified degassing quality evaluation function, and outputting an optimized degassing parameter set, wherein the degassing parameters include vacuum degree, temperature, pressure, stirring speed and degassing time; performing a degassing test on the mixed raw materials based on the optimized degassing parameter set, performing a quality assessment on the degassing test results using a standard degassing quality evaluation function, determining the optimal degassing parameters, and performing subsequent degassing operations according to the optimal degassing parameters.
[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0007] By taking the environmental interference factor set, applied electrical components and resin property information as constraints, obtaining a sample data set based on industrial big data retrieval, and then determining the deaeration parameter optimization space based on the sample data set analysis, the randomness and blindness of the deaeration parameter optimization can be reduced; then, based on the deaeration parameter optimization space, the deaeration parameters are optimized using a simplified deaeration quality evaluation function and a deaeration analysis model to obtain an optimized deaeration parameter set, which can improve the intelligence of the deaeration optimization analysis, reduce the number of experiments, and save optimization analysis time; finally, based on the optimized deaeration parameter set, a deaeration test is carried out on the mixed raw materials, and the quality of the deaeration test results is evaluated using a standard deaeration quality evaluation function, the deaeration parameter with the best quality evaluation result is set as the optimal deaeration parameter, and subsequent deaeration operations are performed according to the optimal deaeration parameter, thereby achieving the technical effect of improving the precision, accuracy and efficiency of the optimal deaeration parameters.
[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0009] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.
[0010] Figure 1 A schematic diagram of a process for optimizing the performance of insulating resin based on big data analysis in this application;
[0011] Figure 2 A schematic diagram of a process for obtaining predetermined influencing indicators in a method for optimizing insulating resin performance based on big data analysis is provided in this application. DETAILED DESCRIPTION
[0012] This application provides a method for optimizing the performance of insulating resins based on big data analysis, which solves the technical problem that the existing methods for optimizing the degassing process of insulating resins often rely on experimental trial and error, have the defects of large randomness and blindness in optimization, take a long time, and the precision and accuracy of the optimization results are difficult to guarantee. By performing associative retrieval based on big data, the randomness and blindness of optimization can be reduced, and at the same time, using intelligent means to optimize the degassing parameters can save optimization analysis time and improve the precision, accuracy and efficiency of the optimization results.
[0013] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.
[0014] For examples, please see the attached Figure 1 The present application provides a method for optimizing the performance of insulating resin based on big data analysis, and the method specifically comprises the following steps:
[0015] Step 1: Collect environmental interference factors of the target insulating resin based on predetermined influencing indicators to obtain an environmental interference factor set.
[0016] Specifically, a predetermined influencing indicator is obtained, where the predetermined influencing indicator refers to an environmental factor that is highly correlated with the service life of the target insulating resin, such as: average ultraviolet intensity, average humidity and other indicators, which are obtained based on sample data analysis; then, the environmental interference factor of the target insulating resin is collected according to the predetermined influencing indicator, wherein the target insulating resin is a resin material with good insulating properties, which are widely used in electrical and electronic equipment for insulating, encapsulating, fixing and protecting electronic components, such as: epoxy resin, polyester resin, polyimide resin, etc. For ease of understanding, in this embodiment, epoxy resin is mainly used as a sample for example; wherein the environmental interference factor refers to the environmental factor in the specific application scenario of the target insulating resin, which can be set according to actual conditions to obtain a set of environmental interference factors.
[0017] Step 2: Read the applied electrical components and resin property information of the target insulating resin, and perform data mining based on industrial big data with the environmental interference factor set, applied electrical components and resin property information as constraints to obtain a sample data set, wherein the resin property information includes a predetermined raw material ratio and predetermined mixing process parameters.
[0018] Specifically, the application electrical components and resin property information of the target insulating resin are read, wherein the application electrical components refer to the specific application equipment or components of the target insulating resin, such as: electromagnet, contactor coil, mutual inductor, and insulation package of dry-type transformer; wherein the resin property information includes a predetermined raw material ratio and predetermined mixing process parameters, the predetermined raw material ratio includes the raw material type and the corresponding proportion, and the predetermined mixing process parameters are the mixing parameters of the completed raw material operation or the expected setting mixing parameters.
[0019] Industrial big data is a series of technologies and methods that enable the value contained in industrial big data to be mined and displayed, including data planning, data collection, analysis and mining and other technical means. Then, with the environmental interference factor set, applied electrical components and resin property information as retrieval constraints, data mining is performed based on industrial big data to obtain a sample data set, where the sample data includes data such as sample resin service life, sample resin failure frequency, and sample resin degassing parameters. By performing data mining based on industrial big data, the accuracy, reliability and comprehensiveness of the sample data can be improved.
[0020] Step three: Screen the sample data set according to predetermined constraints, extract sample degassing parameters from the sample data that meet the predetermined constraints, and construct a degassing parameter optimization space, wherein the predetermined constraints include a predetermined service life and a predetermined electrical failure frequency.
[0021] Specifically, a predetermined constraint condition is obtained, wherein the predetermined constraint condition includes a predetermined service life and a predetermined electrical failure frequency, which are required performance indicators of the target insulating resin and can be set according to actual conditions. Then, the sample data set is screened according to the predetermined service life and the predetermined electrical failure frequency, and the sample degassing parameters in the sample data that meet the predetermined service life and the predetermined electrical failure frequency are extracted to obtain a plurality of sample degassing parameters, and a degassing parameter optimization space is constructed based on the plurality of sample degassing parameters. By constructing a degassing parameter optimization space based on the retrieved sample degassing parameters, the randomness and blindness of the degassing parameter optimization can be reduced, while providing support for the next step of degassing parameter optimization.
[0022] Step 4: Based on the deaeration parameter optimization space, a simplified deaeration quality evaluation function is used to optimize the deaeration parameters, and an optimized deaeration parameter set is output, wherein the deaeration parameters include vacuum degree, temperature, pressure, stirring speed and deaeration time.
[0023] Specifically, a degassing analysis model is constructed based on the BP neural network, which is a neural network model that can be iteratively optimized in machine learning and is obtained through supervised training of a training data set; the degassing analysis model is used to predict the bubble removal rate and resin density after degassing according to the degassing parameters, so as to improve the intelligence of the degassing result analysis, reduce the experimental time, and improve the optimization efficiency; a simplified degassing quality evaluation function is constructed, and the simplified degassing quality evaluation function is used to evaluate the degassing quality according to the bubble removal rate and resin density prediction results.
[0024] Then, based on the deaeration parameter optimization space, deaeration analysis model and simplified deaeration quality evaluation function, the deaeration parameters are optimized using an optimization algorithm, where the deaeration parameters include vacuum degree, temperature, pressure, stirring speed and deaeration time. According to the optimization results, an optimized deaeration parameter set is output, where the optimized deaeration parameter set contains multiple better deaeration parameters.
[0025] Step 5: Perform a degassing test on the mixed raw material based on the optimized degassing parameter set, use a standard degassing quality evaluation function to perform a quality assessment on the degassing test results, determine the optimal degassing parameters, and perform subsequent degassing operations based on the optimal degassing parameters.
[0026] Specifically, based on the optimized degassing parameter set, the vacuum degassing equipment is controlled to perform degassing tests on the mixed and processed insulating resin, and multiple degassing test results are output, wherein the degassing test results correspond to the optimized degassing parameters one by one; a standard degassing quality evaluation function is constructed, wherein the evaluation indicators of the standard degassing quality evaluation function include bubble removal rate, resin density, transparency, surface smoothness and color uniformity coefficient; the standard degassing quality evaluation function is further used to perform quality evaluation on multiple degassing test results, and the optimized degassing parameter with the largest quality evaluation coefficient is set as the optimal degassing parameter. Finally, subsequent degassing operations are performed according to the optimal degassing parameters, thereby achieving the technical effect of improving the precision, accuracy and efficiency of the optimal degassing parameters.
[0027] By providing an insulating resin performance optimization method based on big data analysis, it is possible to solve the technical problems that the existing methods for optimizing the degassing process of insulating resin often rely on experimental trial and error, have the defects of large randomness and blindness in optimization, take a long time, and the precision and accuracy of the optimization results are difficult to guarantee. First, the environmental interference factors of the target insulating resin are collected based on the predetermined influencing indicators to obtain the environmental interference factor set; then, the applied electrical components and resin property information of the target insulating resin are read, and data mining is performed based on industrial big data with the environmental interference factor set, the applied electrical components and the resin property information as constraints to obtain a sample data set, wherein the resin property information includes a predetermined raw material ratio and a predetermined mixing process parameter; then, the sample data set is screened according to the predetermined constraint conditions, and the sample degassing parameters in the sample data that meet the predetermined constraint conditions are extracted to construct a degassing parameter optimization space, wherein the predetermined constraint conditions include a predetermined service life and a predetermined electrical failure frequency; next, based on the degassing parameter optimization space, a simplified degassing quality evaluation function is used to optimize the degassing parameters, and an optimized degassing parameter set is output, wherein the degassing parameters include vacuum degree, temperature, pressure, stirring speed and degassing time; finally, a degassing test is performed on the mixed raw materials based on the optimized degassing parameter set, and the quality of the degassing test results is evaluated using a standard degassing quality evaluation function to determine the optimal degassing parameters, and subsequent degassing operations are performed according to the optimal degassing parameters. By conducting associative retrieval based on big data, the randomness and blindness of optimization can be reduced. At the same time, using intelligent means to perform degassing parameter optimization analysis can save optimization analysis time and improve the precision, accuracy and efficiency of optimization results.
[0028] Further, obtain the predetermined impact indicators, such as the attached Figure 2 As shown, step one of this application includes:
[0029] Collect and obtain multiple environmental interference factors of the target insulating resin, wherein the multiple environmental interference factors include at least the average ultraviolet intensity, the extreme ultraviolet intensity, the average temperature, the extreme temperature, the temperature difference characteristics, the average humidity, the extreme humidity, the average corrosion intensity and the extreme corrosion intensity; based on industrial big data, perform correlation analysis on the multiple environmental interference factors and the service life of the resin in turn, and select the environmental interference factor with a correlation coefficient greater than a predetermined correlation threshold as the influencing indicator, and add it to the predetermined influencing indicator.
[0030] Specifically, first, multiple environmental interference factors of the target insulating resin are collected and obtained. Environmental interference factors refer to environmental factors that affect the service life of the target insulating resin, wherein the multiple environmental interference factors at least include the mean value of ultraviolet intensity, the extreme value of ultraviolet intensity, the mean value of temperature, the extreme value of temperature, the temperature difference characteristic, the mean value of humidity, the extreme value of humidity, the mean value of corrosion intensity and the extreme value of corrosion intensity. The technicians in this field can also add other environmental interference factors based on the actual situation. Then, based on industrial big data, multiple sample data containing environmental interference factors and resin service life are retrieved; further, based on multiple sample data, based on the principle of single variable analysis, multiple environmental interference factors and resin service life are analyzed in turn for correlation, wherein commonly used correlation analysis methods include statistical analysis, correlation analysis, regression analysis, etc., and the appropriate analysis method can be selected based on actual needs. These methods can help determine the degree of correlation between each factor and the service life of the resin, and obtain multiple correlation coefficients, wherein the higher the degree of correlation, the larger the corresponding correlation coefficient, and the correlation coefficient and the environmental interference factor correspond one to one. Then, environmental interference factors with correlation coefficients greater than a predetermined correlation threshold are selected as influencing indicators, wherein the influencing indicators reflect environmental factors that have a significant impact on resin performance and life in an actual use environment, and multiple influencing indicators are obtained and set as predetermined influencing indicators.
[0031] Further, the degassing parameter optimization space is constructed, and step three of this application includes:
[0032] The sample data set is screened according to a predetermined service life and a predetermined electrical failure frequency, and the sample degassing parameters in the sample data that meet the predetermined service life and the predetermined electrical failure frequency are set as qualified sample degassing parameters, so as to obtain a plurality of qualified sample degassing parameters that meet a predetermined quantity constraint; the occurrence frequency of the same qualified sample degassing parameters is counted, and the same qualified sample degassing parameters whose occurrence frequency is greater than a predetermined frequency threshold are selected and set as standard degassing parameters, so as to obtain a plurality of standard degassing parameters, and a degassing parameter optimization space is constructed based on the plurality of standard degassing parameters.
[0033] Specifically, first, the sample data set is screened according to the predetermined service life and the predetermined electrical failure frequency, and the sample degassing parameters in the sample data that meet the predetermined service life and the predetermined electrical failure frequency are set as qualified sample degassing parameters, and multiple qualified sample degassing parameters that meet the predetermined number constraints are obtained, wherein if the number of qualified sample degassing parameters obtained after screening does not meet the predetermined number constraints, iterative retrieval and screening can be performed. Then, the occurrence frequency of the same qualified sample degassing parameters is counted, and the same qualified sample degassing parameters with an occurrence frequency greater than the predetermined frequency threshold are selected and set as standard degassing parameters, wherein the standard degassing parameters are qualified sample degassing parameters with a larger occurrence frequency, which can improve the stability and reliability of the standard degassing parameters, and obtain multiple standard degassing parameters; finally, the minimum parameter feature of multiple standard degassing parameters is used as the lower limit of the degassing parameter optimization space, and the maximum parameter feature is used as the upper limit of the degassing parameter optimization space, to generate the degassing parameter optimization space.
[0034] Further, based on the degassing parameter optimization space, the degassing parameter optimization is performed using a simplified degassing quality evaluation function, and an optimized degassing parameter set is output. Step 4 of the present application includes:
[0035] The degassing parameter optimization space includes a vacuum constraint threshold, a temperature constraint threshold, a pressure constraint threshold, a stirring speed constraint threshold and a degassing time constraint threshold; a parameter is randomly selected from the vacuum constraint threshold, the temperature constraint threshold, the pressure constraint threshold, the stirring speed constraint threshold and the degassing time constraint threshold for combination to generate a first initial degassing parameter, and multiple initial degassing parameters are iteratively selected; the multiple initial degassing parameters are predicted using a degassing analysis model to obtain multiple predicted degassing results, wherein the predicted degassing results include a predicted bubble removal rate and a predicted resin density; the multiple predicted degassing results are evaluated using a simplified degassing quality evaluation function to obtain multiple fitnesses, wherein the simplified degassing quality evaluation function includes a first weight for the bubble removal rate and a second weight for the resin density; the initial degassing parameter is regarded as an initial solution, and multiple initial solutions are arranged from large to small according to the fitness to generate an initial solution sequence , select the first N initial solutions of the initial solution sequence as optimal solutions, and the last M initial solutions as inferior solutions, wherein the sum of N and M is the number of initial solutions, and M is an integral multiple of N; with the N optimal solutions as the center, randomly cluster the M inferior solutions to obtain N domains, wherein the number of inferior solutions in each domain is the same; within the N domains, take the optimal solution as the optimization direction, mutate the inferior solutions in the domain once according to the predetermined multivariate optimization step length, and obtain N update domains, wherein if the mutated inferior solution does not satisfy the degassing parameter optimization space, no mutation is performed; identify N update domains, if in the same update domain, there is an inferior solution whose fitness is greater than the fitness of the optimal solution, then the inferior solution is replaced with the optimal solution; perform iterative optimization until the predetermined number of optimization times is met, output N current domains, select the domain with the largest sum of fitness among the N current domains as the optimal domain, and use the optimal domain as the optimized degassing parameter set.
[0036] Specifically, the deaeration parameter optimization space includes a vacuum constraint threshold, a temperature constraint threshold, a pressure constraint threshold, a stirring speed constraint threshold and a deaeration time constraint threshold, and the constraint threshold refers to the adjustment range of the deaeration parameter; then, a parameter is randomly selected from the vacuum constraint threshold, the temperature constraint threshold, the pressure constraint threshold, the stirring speed constraint threshold and the deaeration time constraint threshold for combination to obtain the first initial deaeration parameter, and the iterative selection is continued to obtain multiple initial deaeration parameters.
[0037] A degassing analysis model is constructed based on a BP neural network. The degassing analysis model is a neural network model that can be iteratively optimized in machine learning and is obtained through supervised training of a training data set; the degassing analysis model is used to predict the bubble removal rate and resin density after degassing according to the degassing parameters. Then, the degassing analysis model is used to predict multiple initial degassing parameters to obtain multiple predicted degassing results, wherein the predicted degassing results include predicted bubble removal rate and predicted resin density. A simplified degassing quality evaluation function is constructed, wherein the simplified degassing quality evaluation function includes a first weight for the bubble removal rate and a second weight for the resin density, which is used to perform weighted calculations on the bubble removal rate and the resin density. The sum of the first weight and the second weight is 1, and the corresponding weight can be set according to the influence of the indicator. The greater the influence, the greater the corresponding weight. The simplified degassing quality evaluation function is further used to evaluate multiple predicted degassing results to obtain multiple fitness, wherein the greater the fitness, the better the degassing quality.
[0038] The initial degassing parameters are regarded as the initial solution, and multiple initial solutions are arranged from large to small according to the fitness to generate an initial solution sequence, and the first N initial solutions of the initial solution sequence are selected as the optimal solution, and the last M initial solutions are set as the inferior solution, where the sum of N and M is the number of initial solutions, and M is an integer multiple of N. Then, with the N optimal solutions as the center, the M inferior solutions are randomly clustered to obtain N domains, where the number of inferior solutions in each domain is the same. Then, in the N domains, the optimal solution is used as the optimization direction, and the inferior solutions in the domain are mutated once according to the predetermined multivariate optimization step length, where the predetermined multivariate optimization step length includes the void adjustment step length, the temperature adjustment step length, the pressure adjustment step length, the stirring speed adjustment step length, and the degassing time adjustment step length, which can be set according to the optimization requirements to obtain N updated domains, where if the mutated inferior solution does not meet the degassing parameter optimization space, the tail solution will not be mutated. Then, N update domains are identified. If, in the same update domain, there is an inferior solution whose fitness is greater than that of the optimal solution, the optimal solution is replaced by the inferior solution. The optimization is continued until the predetermined number of optimizations is met, and N current domains are output. The domain with the largest sum of fitness among the N current domains is selected as the optimal domain, and the optimal domain is used as the optimization degassing parameter set.
[0039] By using the above optimization algorithm to optimize the degassing parameters, due to the strong global search ability of the algorithm, it is possible to avoid falling into the local optimum and improve the reliability, accuracy and practicality of the optimized degassing parameter set.
[0040] Further, to construct a degassing analysis model, the present application also includes the following steps:
[0041] Based on big data, with the degassing equipment property information and the resin property information as constraints, a sample training data set is retrieved and obtained, wherein the sample training data includes sample degassing parameters, sample bubble removal rate and sample resin density; the sample training data set is equally divided into K parts, and K parts are selected with replacement for K times to construct a first data set, and K training sets are selected iteratively for K times to obtain K training sets; with the sample degassing parameters as input, and the sample bubble removal rate and sample resin density as supervision, the BP neural network is supervisedly trained and cross-validated using the K training sets to obtain K degassing analysis units, and a degassing analysis model is built in parallel based on the K degassing analysis units, wherein the output of the degassing analysis model is the mode of the output results of the K degassing analysis units.
[0042] Specifically, based on industrial big data, with the degassing equipment attribute information and the resin attribute information as constraints, a sample training data set is retrieved, wherein the degassing equipment attribute information includes the equipment type, equipment specifications, working principle, etc., and the sample training data includes sample degassing parameters, sample bubble removal rate, and sample resin density. Then the sample training data set is divided into K parts to obtain K data sets, and K data sets are selected with replacement K times to construct the first data set, and iterative selection is performed K times to obtain K training sets.
[0043] An initial degassing analysis unit is constructed based on a BP neural network, wherein the input data of the initial degassing analysis unit is the sample degassing parameter, and the output data is the sample bubble removal rate and the sample resin density. Then, K training sets are used to supervise the initial degassing analysis unit, wherein the gradient training can be performed using a gradient descent algorithm and a loss function, and after the predetermined number of training times is met, the data set is exchanged, and the K initial degassing analysis units are verified and trained using the exchanged data set to obtain K degassing analysis units that meet the expected convergence constraints, and a degassing analysis model is built in parallel based on the K degassing analysis units, wherein the output of the degassing analysis model is the mode of the output results of the K degassing analysis units. By building a degassing analysis model in parallel based on K degassing analysis units, the prediction error of a single degassing analysis unit can be reduced, and the accuracy and reliability of the degassing quality prediction can be further improved.
[0044] Furthermore, the degassing test results are evaluated for quality using a standard degassing quality evaluation function to determine the optimal degassing parameters. Step five of this application includes:
[0045] Based on a predetermined detection scheme, multiple tests are performed on the degassing test results to obtain a first degassing test result, wherein the first degassing test result includes bubble removal rate, resin density, surface smoothness and transparency; the degassing test result is imaged by an image sensor to obtain a resin image, and a second degassing test result is obtained based on the resin image analysis, wherein the second degassing test result is a color uniformity coefficient; a standard degassing quality evaluation function is used to perform quality evaluation on the bubble removal rate, resin density, transparency, surface smoothness and color uniformity coefficient to obtain multiple quality evaluation coefficients, and the optimized degassing parameter corresponding to the maximum quality evaluation coefficient is selected and set as the optimal degassing parameter.
[0046] Specifically, based on a predetermined detection scheme, the degassing test result is subjected to multiple tests, such as visually inspecting the smoothness and transparency of the resin surface, measuring the resin density using a densitometer, and calculating the bubble removal rate through experimental data to obtain a first degassing test result, wherein the first degassing test result includes bubble removal rate, resin density, surface smoothness and transparency. Then, the degassing test result is imaged by an image sensor to obtain a resin image, and further analyzed according to the resin image to obtain a second degassing test result, wherein the second degassing test result is a color uniformity coefficient, which represents color uniformity, and the higher the color uniformity coefficient, the better the degassing quality.
[0047] Finally, the standard degassing quality evaluation function is used to evaluate the bubble removal rate, resin density, transparency, surface smoothness and color uniformity coefficient, and multiple quality evaluation coefficients are obtained. The higher the quality evaluation coefficient, the better the overall degassing quality. Finally, the optimized degassing parameter corresponding to the maximum quality evaluation coefficient is selected and set as the optimal degassing parameter.
[0048] Further, based on the resin image analysis to obtain a second degassing detection result, the present application also includes the following steps:
[0049] Grayscale processing is performed on a resin image to obtain a grayscale resin image, and the grayscale resin image is grid-divided according to a preset size to obtain a grayscale grid image set; feature analysis is performed on the grayscale grid image set in turn to obtain a grayscale feature set, wherein the grayscale value of the central area of the grayscale grid image is used as a reference, the grayscale value of the surrounding area is judged, the surrounding area with a grayscale value greater than the central area is marked as 1, and the surrounding area with a grayscale value less than or equal to the central area is marked as 0, and the grayscale feature value of the grayscale grid image is obtained based on a binary algorithm; variance calculation is performed on the grayscale feature set, and a color uniformity coefficient is set based on the variance calculation result, wherein the color uniformity coefficient is negatively correlated with the variance calculation result.
[0050] Specifically, the resin image is gray-processed to obtain a gray-scale resin image, wherein gray-scale processing refers to the process of converting a color image into a monochrome image containing only grayscale levels (0-255). This processing helps to reduce the color information in the image, making it easier to focus on the texture and details of the image. Commonly used image grayscale processing methods include average grayscale, weighted average grayscale, etc., and an appropriate method can be selected according to actual conditions. The grayscale resin image is further gridded according to a preset size, wherein the preset size is a grid size, for example, the grid size can be set to a 3 by 3 pixel grid to obtain a grayscale grid image set.
[0051] The grayscale feature analysis method is as follows: first, the grayscale value of the central area of the grayscale grid image is used as a reference to judge the grayscale value of the surrounding area, and the surrounding area with a grayscale value greater than the central area is marked as 1, and the surrounding area with a grayscale value less than or equal to the central area is marked as 0; then, the binary codes of all surrounding areas are combined based on the binary algorithm to obtain the grayscale feature value of the grayscale grid image; then, the grayscale feature analysis method is used to perform feature analysis on the grayscale grid image set in turn to obtain a grayscale feature set.
[0052] The variance of the grayscale feature set is further calculated, and a color uniformity coefficient is set according to the variance calculation result, wherein the color uniformity coefficient is negatively correlated with the variance calculation result, and the smaller the variance calculation result is, the smaller the overall deviation of the characterization color is, and the larger the color uniformity coefficient is. By using the above method to perform color uniformity analysis, the convenience of uniformity analysis can be improved, and the accuracy and efficiency of obtaining the color uniformity coefficient can be improved.
[0053] Furthermore, the present application also includes the following steps:
[0054] The expression of the standard degassing quality evaluation function is: P = v1 log 0.5 RA+v2 log 0.5 RB+v3 log 0.5 RC+v4 log 0.5 RD+v5 log 0.5 RE; where P is the quality assessment coefficient, v1 is the bubble removal rate weight, RA is the bubble removal rate, v2 is the resin density weight, RB is the resin density, v3 is the transparency weight, RC is the transparency, v4 is the surface smoothness weight, RD is the surface smoothness, v5 is the color uniformity coefficient weight, and RE is the color uniformity coefficient.
[0055] Specifically, in the standard degassing quality evaluation function, P is the quality evaluation coefficient, v1 is the weight of the bubble removal rate, RA is the bubble removal rate, v2 is the weight of the resin density, RB is the resin density, v3 is the weight of the transparency, RC is the transparency, v4 is the weight of the surface smoothness, RD is the surface smoothness, v5 is the weight of the color uniformity coefficient, and RE is the color uniformity coefficient. By constructing a standard degassing quality evaluation function, the reliability and efficiency of the comprehensive evaluation of the degassing quality can be improved.
[0056] In summary, the insulating resin performance optimization method based on big data analysis provided by this application has the following technical effects:
[0057] By taking the environmental interference factor set, applied electrical components and resin property information as constraints, obtaining a sample data set based on industrial big data retrieval, and then determining the deaeration parameter optimization space based on the sample data set analysis, the randomness and blindness of the deaeration parameter optimization can be reduced; then, based on the deaeration parameter optimization space, the deaeration parameters are optimized using a simplified deaeration quality evaluation function and a deaeration analysis model to obtain an optimized deaeration parameter set, which can improve the intelligence of the deaeration optimization analysis, reduce the number of experiments, and save optimization analysis time; finally, based on the optimized deaeration parameter set, a deaeration test is carried out on the mixed raw materials, and the quality of the deaeration test results is evaluated using a standard deaeration quality evaluation function, the deaeration parameter with the best quality evaluation result is set as the optimal deaeration parameter, and subsequent deaeration operations are performed according to the optimal deaeration parameter, thereby achieving the technical effect of improving the precision, accuracy and efficiency of the optimal deaeration parameters.
[0058] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
[0059] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.
Claims
1. A method for optimizing the performance of insulating resin based on big data analysis, characterized in that: The method comprises: Collect environmental interference factors of the target insulating resin based on predetermined influencing indicators to obtain an environmental interference factor set; Reading the applied electrical components and resin property information of the target insulating resin, and performing data mining based on industrial big data with the environmental interference factor set, the applied electrical components and the resin property information as constraints to obtain a sample data set, wherein the resin property information includes a predetermined raw material ratio and a predetermined mixing process parameter; The sample data set is screened according to predetermined constraints, sample degassing parameters in the sample data satisfying the predetermined constraints are extracted, and a degassing parameter optimization space is constructed, wherein the predetermined constraints include a predetermined service life and a predetermined electrical failure frequency; Based on the degassing parameter optimization space, a simplified degassing quality evaluation function is used to optimize the degassing parameters, and an optimized degassing parameter set is output, wherein the degassing parameters include vacuum degree, temperature, pressure, stirring speed and degassing time; Based on the optimized degassing parameter set, a degassing test is performed on the mixed raw material, a quality assessment is performed on the degassing test result using a standard degassing quality evaluation function, an optimal degassing parameter is determined, and subsequent degassing operations are performed according to the optimal degassing parameter; Wherein, based on the degassing parameter optimization space, the degassing parameter optimization is performed using a simplified degassing quality evaluation function, and an optimized degassing parameter set is output, including: The degassing parameter optimization space includes a vacuum degree constraint threshold, a temperature constraint threshold, a pressure constraint threshold, a stirring speed constraint threshold and a degassing time constraint threshold; A parameter is randomly selected from a vacuum degree constraint threshold, a temperature constraint threshold, a pressure constraint threshold, a stirring speed constraint threshold, and a degassing time constraint threshold to combine, to generate a first initial degassing parameter, and multiple initial degassing parameters are obtained by iterative selection; A degassing analysis model is used to predict a plurality of initial degassing parameters to obtain a plurality of predicted degassing results, wherein the predicted degassing results include a predicted bubble removal rate and a predicted resin density; Using a simplified degassing quality evaluation function to evaluate multiple predicted degassing results to obtain multiple fitness levels, wherein the simplified degassing quality evaluation function includes a first weight of bubble removal rate and a second weight of resin density; The initial degassing parameters are regarded as the initial solutions, and multiple initial solutions are arranged from large to small according to fitness to generate an initial solution sequence. The first N initial solutions in the initial solution sequence are selected as the best solutions, and the last M initial solutions are set as the worst solutions, where the sum of N and M is the number of initial solutions, and M is an integer multiple of N; Taking N optimal solutions as the center, randomly cluster M inferior solutions to obtain N domains, where the number of inferior solutions in each domain is the same; In N domains, the optimal solution is used as the optimization direction, and the inferior solution in the domain is mutated once according to the predetermined multivariate optimization step length to obtain N updated domains. If the inferior solution after mutation does not satisfy the degassing parameter optimization space, no mutation is performed; Identify N update areas. If, within the same update area, the fitness of an inferior solution is greater than that of an optimal solution, the optimal solution is replaced by the inferior solution. Iterate optimization until a predetermined number of optimization times is met, output N current fields, select the field with the largest sum of fitness among the N current fields as the optimal field, and use the optimal field as the optimized degassing parameter set; Among them, the degassing analysis model is constructed, including: Based on big data, with the degassing equipment property information and the resin property information as constraints, a sample training data set is retrieved and obtained, wherein the sample training data includes a sample degassing parameter, a sample bubble removal rate, and a sample resin density; Divide the sample training data set into K equal parts, select K parts with replacement, construct the first data set, and iterate and select K parts K times to obtain K training sets; With sample degassing parameters as input, sample bubble removal rate and sample resin density as supervision, K training sets are used to perform supervised training and cross-validation on the BP neural network to obtain K degassing analysis units. A degassing analysis model is built in parallel based on the K degassing analysis units, wherein the output of the degassing analysis model is the mode of the output results of the K degassing analysis units.
2. The insulating resin performance optimization method based on big data analysis according to claim 1 is characterized in that: Get the scheduled impact indicators, including: Collect and obtain multiple environmental interference factors of the target insulating resin, wherein the multiple environmental interference factors at least include ultraviolet intensity average, ultraviolet intensity extreme, temperature average, temperature extreme, temperature difference characteristics, humidity average, humidity extreme, corrosion intensity average and corrosion intensity extreme; Based on industrial big data, correlation analysis is performed on multiple environmental interference factors and resin service life in turn, and environmental interference factors with correlation coefficients greater than a predetermined correlation threshold are selected as influencing indicators and added to the predetermined influencing indicators.
3. The insulating resin performance optimization method based on big data analysis according to claim 1 is characterized in that: Construct the degassing parameter optimization space, including: The sample data set is screened according to a predetermined service life and a predetermined electrical failure frequency, and the sample degassing parameters in the sample data that meet the predetermined service life and the predetermined electrical failure frequency are set as qualified sample degassing parameters, so as to obtain a plurality of qualified sample degassing parameters that meet a predetermined quantity constraint; The occurrence frequencies of degassing parameters of the same qualified samples are counted, and the degassing parameters of the same qualified samples with an occurrence frequency greater than a predetermined frequency threshold are selected as standard degassing parameters to obtain multiple standard degassing parameters, and a degassing parameter optimization space is constructed based on the multiple standard degassing parameters.
4. The insulating resin performance optimization method based on big data analysis according to claim 1 is characterized in that: The standard degassing quality evaluation function is used to evaluate the quality of the degassing test results and determine the optimal degassing parameters, including: Based on a predetermined detection scheme, multiple detections are performed on the degassing test results to obtain a first degassing test result, wherein the first degassing test result includes bubble removal rate, resin density, surface smoothness and transparency; The degassing test result is imaged by an image sensor to obtain a resin image, and a second degassing test result is obtained based on the resin image analysis, wherein the second degassing test result is a color uniformity coefficient; The standard degassing quality evaluation function is used to perform quality evaluation on the bubble removal rate, resin density, transparency, surface smoothness and color uniformity coefficient to obtain multiple quality evaluation coefficients, and the optimized degassing parameter corresponding to the maximum quality evaluation coefficient is selected and set as the optimal degassing parameter.
5. The insulating resin performance optimization method based on big data analysis according to claim 4 is characterized in that: The second degassing test result is obtained based on the resin image analysis, including: grayscale processing is performed on the resin image to obtain a grayscale resin image, and the grayscale resin image is gridded according to a preset size to obtain a grayscale grid image set; Perform feature analysis on the grayscale grid image set in turn to obtain a grayscale feature set, wherein the grayscale value of the central area of the grayscale grid image is used as a reference to judge the grayscale value of the surrounding area, mark the surrounding area with a grayscale value greater than the central area as 1, and mark the surrounding area with a grayscale value less than or equal to the central area as 0, and obtain the grayscale feature value of the grayscale grid image based on a binary algorithm; A variance calculation is performed on the grayscale feature set, and a color uniformity coefficient is set based on the variance calculation result, wherein the color uniformity coefficient is negatively correlated with the variance calculation result.
6. The insulating resin performance optimization method based on big data analysis according to claim 4 is characterized in that: The expression of the standard degassing quality evaluation function is: P=v1 log 0.5 RA+v2 log 0.5 RB+v3 log 0.5 RC+v4 log 0.5 RD+ v5 log 0.5 RE; Among them, P is the quality assessment coefficient, v1 is the bubble removal rate weight, RA is the bubble removal rate, v2 is the resin density weight, RB is the resin density, v3 is the transparency weight, RC is the transparency, v4 is the surface smoothness weight, RD is the surface smoothness, v5 is the color uniformity coefficient weight, and RE is the color uniformity coefficient.
Citation Information
Patent Citations
Light weapon part and environmental adaptability evaluation method of composite material and high polymer material thereof
CN110299192A
High polymer material service life prediction method based on environmental big data and machine learning
CN112331281A
