CN116913435B - High-strength engineering plastic evaluation method and system based on component analysis - Google Patents
High-strength engineering plastic evaluation method and system based on component analysis Download PDFInfo
- Publication number
- CN116913435B CN116913435B CN202310937200.9A CN202310937200A CN116913435B CN 116913435 B CN116913435 B CN 116913435B CN 202310937200 A CN202310937200 A CN 202310937200A CN 116913435 B CN116913435 B CN 116913435B
- Authority
- CN
- China
- Prior art keywords
- performance
- strength engineering
- engineering plastic
- component composition
- application scene
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C60/00—Computational materials science, i.e. ICT specially adapted for investigating the physical or chemical properties of materials or phenomena associated with their design, synthesis, processing, characterisation or utilisation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/20—Identification of molecular entities, parts thereof or of chemical compositions
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Crystallography & Structural Chemistry (AREA)
- General Engineering & Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Injection Moulding Of Plastics Or The Like (AREA)
Abstract
The invention relates to the technical field of high-strength engineering plastics, in particular to a component analysis-based high-strength engineering plastic evaluation method, which comprises the following steps: respectively obtaining component composition information, formula proportion information and process parameter information; analyzing the performance of a structure formed by high-strength engineering plastics to obtain application scene performance analysis information; taking the component composition information, the formula proportion information and the process parameter information as inputs of a neural network model, and performing performance evaluation on the high-strength engineering plastics to obtain a performance evaluation result; and processing the application scene performance analysis result and the performance evaluation result, and optimizing the component composition, the formula proportion and the process parameters of the high-strength engineering plastic according to the processing result or maintaining the original component composition, the formula proportion and the process parameters. The invention can improve the performance of the product, reduce the production cost and reduce the material waste and the trial-and-error time, thereby improving the application efficiency and the cost control capability of engineering plastics.
Description
Technical Field
The invention relates to the technical field of high-strength engineering plastics, in particular to a component analysis-based high-strength engineering plastic evaluation method and system.
Background
Currently, high strength engineering plastics are widely used in engineering applications, however, there is a general problem in the processing process: the existing component composition design cannot be specifically adjusted according to the finally formed structural form, which means that even if high-quality high-strength engineering plastic raw materials are used in the production process, due to the lack of effective adjusting means, it is difficult to ensure that the finally manufactured product can completely meet the requirements of application scenes in performance.
In the conventional production of high-strength engineering plastics, the composition, the formulation proportion and the process parameters are usually set according to experience and basic tests, and the method can meet the requirements of a part of application scenes to some extent, but with the continuous expansion of the application range of the high-strength engineering plastics and the continuous improvement of the performance requirements of products, the conventional experience design method has been limited.
This current situation brings a series of problems to the production and application of high-strength engineering plastics, including but not limited to the fact that the strength and durability of the products may not fully meet the requirements of specific engineering applications, excessive design may lead to waste of materials and increase in production cost, and time waste caused by multiple trial and error in the development period of the products, etc., which all affect the application popularization and market competitiveness of the high-strength engineering plastics to some extent.
Disclosure of Invention
The invention provides a high-strength engineering plastic evaluation method and a system based on component analysis, so that the problems pointed out in the background art are effectively solved.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
a method for evaluating high-strength engineering plastics based on component analysis, comprising:
detecting the component composition, the formula proportion and the process parameters of the high-strength engineering plastic to be detected respectively to obtain component composition information, formula proportion information and process parameter information respectively;
analyzing the performance of the structure formed by the high-strength engineering plastic to obtain an application scene performance analysis result;
taking the component composition information, the formula proportion information and the process parameter information as inputs of a neural network model, and performing performance evaluation of the high-strength engineering plastic to obtain a performance evaluation result;
and processing the application scene performance analysis result and the performance evaluation result, and optimizing the component composition, the formula proportion and the technological parameters of the high-strength engineering plastic according to the processing result or maintaining the original component composition, the original formula proportion and the original technological parameters.
Further, processing the application scene performance analysis result and the performance evaluation result includes:
enumerating all performance characteristics in the performance analysis result and the performance evaluation result of the application scene, and obtaining weight values of importance of the performance characteristics in the application scene;
carrying out normalization processing on each performance characteristic, and mapping the numerical value of each performance characteristic to a uniform range;
multiplying the normalized performance characteristics with the corresponding weight values, and summing after weighting to obtain a comprehensive evaluation index;
setting a comprehensive index threshold according to actual requirements, optimizing the component composition, the formula proportion and the process parameters of the high-strength engineering plastic if the comprehensive evaluation index is lower than the comprehensive index threshold, otherwise, maintaining the original component composition, the original formula proportion and the original process parameters.
Further, enumerating all performance features in the performance analysis result and the performance evaluation result of the application scene, and obtaining a weight value of importance of each performance feature in the application scene, including:
collecting original data of an application scene performance analysis result and a performance evaluation result of the high-strength engineering plastic, and preprocessing the original data;
determining a weight value of the importance of each performance feature in an application scene by using a feature importance evaluation method;
introducing a multi-objective optimization algorithm, finding a group of optimal combinations in all the performance characteristics, and acquiring weight values of the performance characteristics in the optimal combinations.
Further, determining a weight value of importance of each performance feature in the application scene by using a feature importance evaluation method includes:
dividing the original data in the application scene to obtain a data set corresponding to each performance characteristic one by one;
performing variance calculation for data in each data set;
and assigning a weight value to the performance characteristic according to the size of the variance, wherein the weight value is proportional to the variance.
Further, introducing a multi-objective optimization algorithm to find a set of optimal combinations among all of the performance characteristics, comprising:
defining an objective function, wherein the objective function takes a plurality of performance characteristics and corresponding weight values as input, and outputs a comprehensive performance index of the high-strength engineering plastic;
according to the requirements and optimization targets of the application scene, comprehensively considering the maximized comprehensive performance index and the minimized comprehensive performance index;
setting constraint conditions, wherein the constraint conditions are the constraint conditions of the optimal combination;
selecting and running a multi-objective optimization algorithm, inputting the objective function, the optimization objective and the constraint condition into the selected multi-objective optimization algorithm, and running the multi-objective optimization algorithm to search for an optimal combination.
Further, selecting and running a multi-objective optimization algorithm, inputting the objective function, the optimization objective and the constraint condition into the selected multi-objective optimization algorithm, and running the multi-objective optimization algorithm to search for an optimal combination, including:
selecting a multi-objective optimization algorithm;
randomly generating an initial population, each individual within the initial population representing a set of performance feature combinations;
substituting each individual into an objective function, and calculating the fitness value of each individual on a plurality of objective functions, wherein the calculation mode of determining the fitness value according to an optimization target is maximized or minimized;
generating new individuals through genetic algorithm operation, selecting excellent individuals according to fitness values and Pareto optimal solution sets, and enabling the excellent individuals to reproduce new individuals so as to gradually optimize population;
checking whether the genetic algorithm converges to a Pareto optimal solution set, and judging whether to terminate the optimization process according to a preset stopping condition;
and when the genetic algorithm meets a preset stopping condition, outputting a Pareto optimal solution set as the optimal combination.
Further, the neural network model is a multi-layer perceptron model and comprises an input layer, a plurality of hidden layers and an output layer;
the neuron number of the hidden layer is regulated according to the activation value of the activation function, and the regulating step comprises the following steps:
tracking activation values of neurons in each of the hidden layers during training of the model;
after the training period of every set number is finished, according to the activation value condition of each neuron in each hidden layer, the following steps are carried out:
setting an upper threshold and a lower threshold for an activation value, and comparing the activation value with the upper threshold and the lower threshold; increasing the number of neurons when the activation value of the neurons exceeds the upper threshold, decreasing the number of neurons when the activation value of the neurons is below the lower threshold, otherwise, maintaining the number of neurons unchanged.
Further, the method further comprises the following steps: in the model construction stage, setting the maximum neuron number upper limit of each hidden layer, and when the neuron number of the hidden layer reaches the upper limit:
stopping to continue increasing the number of neurons;
or setting a rejection threshold value lower than the lower limit threshold value, and rejecting the neuron corresponding to the activation value when the activation value is smaller than the rejection threshold value.
A component analysis-based high strength engineering plastic evaluation system, comprising:
the detection module is used for respectively detecting the component composition, the formula proportion and the process parameters of the high-strength engineering plastic to be detected to obtain component composition information, formula proportion information and process parameter information;
the performance analysis module is used for analyzing the performance of the structure formed by the high-strength engineering plastic to obtain an application scene performance analysis result;
the neural network module takes the component composition information, the formula proportion information and the process parameter information as input to realize the performance evaluation of the high-strength engineering plastic and obtain a performance evaluation result;
and the optimizing processing module is used for processing the application scene performance analysis result and the performance evaluation result, and optimizing the component composition, the formula proportion and the process parameters of the high-strength engineering plastic according to the processing result or maintaining the original component composition, the original formula proportion and the original process parameters.
Further, the optimization processing module includes:
the feature importance evaluation module enumerates all the performance features in the application scene performance analysis result and the performance evaluation result, and obtains the weight value of the importance of each performance feature in the application scene;
the normalization processing module performs normalization processing on each performance characteristic and maps the numerical value of each performance characteristic to a uniform range;
the comprehensive evaluation module multiplies the normalized performance characteristics with corresponding weight values, and sums the weighted performance characteristics to obtain a comprehensive evaluation index;
the optimization decision module is used for setting a comprehensive index threshold according to actual requirements, and triggering to optimize the component composition, the formula proportion and the process parameters of the high-strength engineering plastic if the comprehensive evaluation index is lower than the comprehensive index threshold; otherwise, the original component composition, the formula proportion and the technological parameters are maintained.
By the technical scheme of the invention, the following technical effects can be realized:
the invention can improve the performance of the product, reduce the production cost and reduce the material waste and the trial-and-error time, thereby improving the application efficiency and the cost control capability of engineering plastics.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present invention, and other drawings may be obtained according to the drawings without inventive effort to those skilled in the art.
FIG. 1 is a flow chart of a method of evaluating high strength engineering plastics based on component analysis;
FIG. 2 is a flowchart showing step S400 in FIG. 1;
FIG. 3 is a flowchart showing the step S410 in FIG. 2;
FIG. 4 is a flowchart showing step S412 in FIG. 3;
fig. 5 is a specific flowchart of step S413 in fig. 3;
FIG. 6 is a flowchart showing the step B40 in FIG. 5;
FIG. 7 is a flow chart of the adjustment of the number of neurons of the hidden layer according to the activation value of the activation function;
FIG. 8 is a frame diagram of a high strength engineering plastic evaluation system based on component analysis;
FIG. 9 is a block diagram of an optimization processing module;
reference numerals: 01. a detection module; 02. a performance analysis module; 03. a neural network module; 04. an optimization processing module; 041. a feature importance assessment module; 042. a normalization processing module; 043. a comprehensive evaluation module; 044. and (5) optimizing the decision module.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.
Example 1
As shown in fig. 1, a method for evaluating high-strength engineering plastics based on component analysis includes:
s100: detecting the component composition, the formula proportion and the process parameters of the high-strength engineering plastic to be detected respectively to obtain component composition information, formula proportion information and process parameter information respectively;
in the step, detailed component detection and process parameter detection are carried out on the high-strength engineering plastic material to be detected, component composition information relates to the component quantity and names of the plastic material, formula proportion information refers to the proportion relation of different components in the high-strength engineering plastic, and process parameter information comprises operating conditions such as temperature, pressure and the like in the processing process. Through the detection of the information, the specific components of the high-strength engineering plastic and important parameters in the preparation process can be obtained, and basic data are provided for subsequent performance evaluation and optimization;
s200: analyzing the performance of a structure formed by high-strength engineering plastics to obtain an application scene performance analysis result;
in this step, a molded structure made of a high-strength engineering plastic, such as a part, a product, etc., is subjected to performance analysis, including performance characteristics such as strength, wear resistance, corrosion resistance, rigidity, etc., of the material. Through the performance analysis of the molding structure, the performance of the high-strength engineering plastic in the actual application scene can be known, and the performance requirement and limitation of the high-strength engineering plastic in specific engineering application can be determined;
s300: taking the component composition information, the formula proportion information and the process parameter information as inputs of a neural network model, and performing performance evaluation on the high-strength engineering plastics to obtain a performance evaluation result;
in the implementation process, the obtained component composition information, formula proportion information and process parameter information are used as input, a neural network model is established to evaluate the performance of the high-strength engineering plastics, and the neural network model predicts the performance of the high-strength engineering plastics, such as strength, toughness and the like, according to the input information.
S400: and processing the application scene performance analysis result and the performance evaluation result, and optimizing the component composition, the formula proportion and the process parameters of the high-strength engineering plastic according to the processing result or maintaining the original component composition, the formula proportion and the process parameters.
In the invention, the components, the formula proportion and the process parameters of the high-strength engineering plastic are detected in detail, so that the method can comprehensively know the components of the plastic material and the important parameters in the preparation process, thereby providing sufficient basic data for subsequent performance evaluation; meanwhile, the performance of the high-strength engineering plastic in practical application can be comprehensively evaluated by combining the performance analysis of the application scene, the performance of the high-strength engineering plastic can be evaluated by adopting a neural network model, the performance of the plastic can be predicted by learning the rule of input data, the neural network model has flexibility and intelligence, the performance of the plastic can be predicted more accurately, the traditional experience design method is not relied on, the performance of the high-strength engineering plastic can be quantitatively measured by comparing the performance analysis result of the application scene with the performance evaluation result output by the neural network model, and whether the performance of the high-strength engineering plastic meets the application requirement can be quantitatively measured, so that the optimization and adjustment of the composition, the formula proportion and the technological parameters can be purposefully performed, and the requirement of a specific application scene can be met to the greatest extent.
The invention can improve the performance of the product, reduce the production cost and reduce the material waste and the trial-and-error time, thereby improving the application efficiency and the cost control capability of engineering plastics.
In step S1, component composition, formulation proportion and process parameters are detected, and the obtained performance characteristics are mainly information about material formulation and process parameters, such as content proportion of different components in the material, temperature, pressure and other operating conditions, and these information are basic data for subsequent evaluation, and are used for building a neural network model; in step S2, the performance of the molded structure is analyzed, and the obtained performance characteristics are about the actual performance of the molded structure made of the high-strength engineering plastic, such as strength, wear resistance, corrosion resistance, etc., and are used for understanding the performance of the high-strength engineering plastic in the actual application scene; in step S3, component composition information, recipe proportion information and process parameter information are used as inputs of a neural network model, and performance evaluation is performed, wherein the neural network model predicts performance of the high-strength engineering plastic, and predicted performance characteristics include strength, toughness and the like, and the predicted performance characteristics are based on calculation results of the model.
In the actual working process, the performance characteristics obtained in steps S2 and S3 are often not identical, or are not identical; based on the above-described problems, as a preferable example of the above-described embodiment, as shown in fig. 2, processing the application scenario performance analysis result and the performance evaluation result in step S400 includes:
s410: enumerating all performance characteristics in the performance analysis result and the performance evaluation result of the application scene, and obtaining weight values of importance of each performance characteristic in the application scene; the strength, wear resistance, corrosion resistance, rigidity, toughness performance characteristics of the materials as in the above embodiments, the weight values are used to represent their relative importance to subsequent comprehensive performance evaluation, and in particular, the weight values may be determined according to expert evaluation, experimental data, or user requirements;
s420: normalizing the performance characteristics, and mapping the numerical values of the performance characteristics to a uniform range;
since the range of values for different performance characteristics may be different, the normalization process described above is required to map the values of each performance characteristic to a uniform range, e.g., between 0 and 1, in order for them to function in balance in the subsequent overall performance evaluation.
S430: multiplying the normalized performance characteristics with corresponding weight values, and summing after weighting to obtain a comprehensive evaluation index;
the steps S420 and S403 may be performed simultaneously or in sequence;
s440: setting a comprehensive index threshold according to actual requirements, optimizing the component composition, the formula proportion and the process parameters of the high-strength engineering plastic if the comprehensive evaluation index is lower than the comprehensive index threshold, otherwise, maintaining the original component composition, the formula proportion and the process parameters.
The optimization scheme comprehensively considers the importance of different performance characteristics and unifies the importance of the performance characteristics into one comprehensive index, so that the performance and the application applicability of the high-strength engineering plastic are more comprehensively evaluated.
The following are specific examples:
weight values are set for all performance characteristics: the strength weight w1=0.4, the wear resistance weight w2=0.3, the corrosion resistance weight w3=0.2, the rigidity weight w4=0.1, and the toughness weight w5=0.2 are set.
Normalization processing: assuming that the strength of the material ranges from [0,100], the wear resistance ranges from [0,10], the corrosion resistance ranges from [0,5], the rigidity ranges from [0,50], and the toughness ranges from [0,20]; the normalized performance characteristic values are calculated as follows:
normalized intensity value n1=intensity/100;
normalized wear resistance value n2=wear resistance/10;
normalized corrosion resistance value n3=corrosion resistance/5
Normalized stiffness value n4=stiffness/50;
normalized toughness value (N5): n5=toughness/20.
Calculating a comprehensive evaluation index ei=w1×n1+w2×n2+w3×n3+w4×n4+w5×n5, wherein w1, w2, w3, w4 and w5 are weight values set before, and N1, N2, N3, N4 and N5 are normalized performance characteristic values.
Setting a comprehensive index threshold value: assuming that the comprehensive index threshold is 0.6, if EI is smaller than 0.6, the performance of the high-strength engineering plastic is not satisfied with the application requirement and needs to be optimized; otherwise, the performance meets the requirements, and the original component composition, formula proportion and process parameters can be maintained.
It should be noted that the weight value, the performance range, and the comprehensive index threshold are only exemplary values, and in practical application, adjustment and determination are required according to specific situations, the determination of the weight value may be performed by methods such as expert evaluation, experimental data, and user requirements, and the performance range and the comprehensive index threshold should be set based on the actual application scenario and requirements.
As a specific optimization implementation, as shown in fig. 3, in step S410, all performance features in the performance analysis result and the performance evaluation result of the application scenario are enumerated, and a weight value of importance of each performance feature in the application scenario is obtained, including:
s411: collecting original data of an application scene performance analysis result and a performance evaluation result of the high-strength engineering plastic, and preprocessing the original data;
s412: determining a weight value of the importance of each performance feature in the application scene by using a feature importance assessment method;
s413: introducing a multi-objective optimization algorithm, finding a group of optimal combinations in all the performance characteristics, and acquiring weight values of all the performance characteristics in the optimal combinations.
In practical applications, high strength engineering plastics may contain a large number of performance characteristics, not all of which play an important role in the requirements of the application scenario; in the optimization scheme, after the performance characteristics with high importance are screened out, the important points are concentrated on the performance characteristics to evaluate and optimize, so that the complexity and the calculated amount of a model can be reduced, and the evaluation process is more efficient; through feature importance assessment and multi-objective optimization, the performance features of the optimal combination are selected, the performance features have higher importance to the requirements of application scenes, and the performance features with high importance are involved in the calculation of the comprehensive assessment index, so that the accuracy of subsequent assessment can be improved, and the actual application requirements can be better met; the optimization scheme can also reduce the feature quantity, reduce the risk of overfitting and improve the generalization capability of the model.
As a preference of the above-described embodiment, as shown in fig. 4, in step S412, a weight value of importance of each performance feature in an application scene is determined using a feature importance evaluation method, including:
a10: dividing original data in an application scene to obtain a data set corresponding to each performance characteristic one by one;
a20: performing variance calculation for data in each data set; variance calculation can measure the discrete degree of data in a data set, and fluctuation conditions of each performance characteristic in an application scene can be known by calculating the variance;
a30: and assigning a weight value to the performance characteristic according to the size of the variance, wherein the weight value is proportional to the variance.
Specifically, performance features with larger variances are given higher weight values, while performance features with smaller variances are given lower weight values, and such weight assignment can be regarded as a way of reflecting the importance of the performance features in the application scene, and the performance features with larger variances have larger influence on the application scene, so that the weight values are correspondingly larger.
As shown in fig. 5, in step S413, a multi-objective optimization algorithm is introduced to find a set of optimal combinations among all performance characteristics, including:
b10: defining an objective function, wherein the objective function takes a plurality of performance characteristics and corresponding weight values as input, and outputs a comprehensive performance index of the high-strength engineering plastic; in the step, the comprehensive performance index can be a weighted sum of all performance characteristics in an application scene, and is used for measuring the overall performance of the high-strength engineering plastic;
b20: according to the requirements and optimization targets of the application scene, comprehensively considering the maximized comprehensive performance index and the minimized comprehensive performance index; through the steps, performance of the high-strength engineering plastics in multiple aspects can be comprehensively considered in an optimization scheme, so that a balance point can be found in different application scenes, a solution with more excellent comprehensive performance is obtained, as different application scenes possibly have different requirements on the high-strength engineering plastics, optimization pursuing a certain performance index can not necessarily meet the requirements of all applications, and the formula and the component composition of the high-strength engineering plastics can be customized according to different application scenes by simultaneously considering maximization and minimization of a plurality of performance indexes, so that optimal performance can be realized, excessive optimization in specific performance can be avoided, and materials and production cost can be saved;
b30: setting constraint conditions, wherein the constraint conditions are the constraint conditions of the optimal combination; these constraints may be upper and lower limits on performance characteristics, limits on the range of weight values, or other practical application-related constraints, by setting constraints, it may be ensured that the optimal combination meets practical application requirements and limits;
b40: selecting and running a multi-objective optimization algorithm, inputting an objective function, an optimization objective and constraint conditions into the selected multi-objective optimization algorithm, and running the multi-objective optimization algorithm to search for an optimal combination.
By the scheme, a group of optimal performance characteristic combinations can be found, so that the high-strength engineering plastic can achieve better performance on a plurality of performance indexes; the use of a multi-objective optimization algorithm can find a more comprehensive solution in a complex multi-objective optimization problem, and the above-mentioned optimization scheme provides a systematic method to search for the optimal performance combination of high-strength engineering plastics to meet the requirements of specific application scenarios, while considering the requirements of multiple objectives.
In step B40, a multi-objective optimization algorithm is selected and run, the objective function, the optimization objective and the constraint condition are input into the selected multi-objective optimization algorithm, and the running multi-objective optimization algorithm searches for an optimal combination, as shown in fig. 6, preferably including:
b41: selecting a multi-objective optimization algorithm;
and B42: randomly generating an initial population, wherein each individual in the initial population represents a group of performance characteristic combinations;
b43: substituting each individual into the objective function, and calculating the fitness value of each individual on a plurality of objective functions, wherein the calculation mode of determining the fitness value according to the optimization objective is maximized or minimized;
and B44: generating new individuals through crossover, mutation and other genetic algorithm operations, selecting excellent individuals according to fitness values and Pareto optimal solution sets, and enabling the excellent individuals to reproduce the new individuals so as to gradually optimize the population;
b45: checking whether the genetic algorithm converges to the Pareto optimal solution set, and judging whether to terminate the optimization process according to a preset stopping condition, such as reaching a maximum algebra or meeting convergence accuracy;
b46: and when the genetic algorithm meets the preset stopping condition, outputting the Pareto optimal solution set as an optimal combination.
Through the above preferred scheme, the optimal performance characteristic combination of the high-strength engineering plastics can be found by using the multi-objective optimization algorithm under the premise of comprehensively considering the maximization and minimization of the comprehensive performance indexes, and the optimization result comprises a group of solutions of the high-strength engineering plastics, wherein each solution is relatively excellent in different performance characteristics, and flexible selection and decision basis is provided for you.
It is assumed that it is desirable to maximize strength and durability while minimizing cost in implementation; by gradually optimizing the population, a solution will be created that is relatively excellent in terms of performance characteristics such as strength, durability, and cost; when the algorithm finds a set of non-dominant solutions that are relatively excellent in strength, durability, and cost, the algorithm can be terminated, which contains an optimal combination of performance characteristics for a set of high strength engineering plastics, where each solution is relatively excellent in strength, durability, and cost, and can meet application requirements and constraints.
As a preference of the above embodiment, as shown in fig. 7, the neural network model is a multi-layer sensor model, including an input layer, several hidden layers, and an output layer;
the number of neurons of the hidden layer is adjusted according to the activation value of the activation function, and the adjusting step comprises the following steps:
c10: tracking activation values of neurons in each hidden layer during training of the model;
c20: after the training period of every set number is finished, according to the activation value condition of each neuron in each hidden layer, the following steps are carried out: in this step, the set number may be 1, that is, after one training period is finished, the subsequent operation is performed, or in order to avoid too frequent adjustment, the set number may be set to a number of 2 or more;
c30: setting an upper limit threshold and a lower limit threshold for the activation value, and comparing the activation value with the upper limit threshold and the lower limit threshold;
when the activation value of the neuron exceeds the upper threshold, step C40 is performed: increasing the number of neurons;
when the activation value of the neuron is lower than the lower threshold, step C50 is performed: reducing the number of neurons;
otherwise, the number of neurons is maintained unchanged.
In the optimization scheme, the number of neurons of the hidden layer is dynamically regulated, so that the model has self-adaptability according to the activation value condition of each neuron, and in the training process, the number of neurons can be regulated according to the distribution and the complexity degree of data, so that the model is better suitable for different input data. In the implementation process, the number of neurons is increased or reduced according to the activation values, so that the structure of the model can be optimized, when the activation values of some neurons are higher, the influence of the neurons in the model can be enhanced by increasing the number of the neurons, and when the activation values of some neurons are lower, the complexity of the model can be reduced by reducing the number of the neurons, and overfitting is avoided.
Through the optimization scheme, unnecessary calculation expenditure can be avoided by adaptively adjusting the number of the hidden layer neurons, and for neurons with lower activation values, calculation resources can be saved by reducing the number of the neurons, and the efficiency of model training and reasoning is improved; meanwhile, the model can be better adapted to different input distribution and complexity, so that the generalization capability of the model is improved, and the model is better applicable to unseen data.
In general, the optimization scheme can improve the self-adaptability of the model, optimize the model structure, save the computing resources and improve the generalization capability, thereby enhancing the performance and the efficiency of the multi-layer perceptron model in the high-strength engineering plastic evaluation method.
Preferably, in the process of adjusting the number of neurons of the hidden layer according to the activation value of the activation function, the method further includes: in the model construction stage, setting the maximum neuron number upper limit of each hidden layer, and when the neuron number of the hidden layer reaches the upper limit:
stopping to continue increasing the number of neurons;
or setting a rejection threshold value lower than the lower limit threshold value, and rejecting the neuron corresponding to the activation value when the activation value is smaller than the rejection threshold value.
By setting the upper limit of the maximum neuron number of each hidden layer or eliminating unimportant neurons, the complexity of the model can be effectively controlled, the computing resources can be saved, and the performance and generalization capability of the model can be improved, so that the multi-layer perceptron model in the high-strength engineering plastic evaluation method is more effective and reliable.
In the evaluation of high-strength engineering plastics, the input data involved are mainly component composition information, recipe proportion information and process parameter information, which are typically structured data, and in the above-described embodiments, the multilayer perceptron model is adapted to process the structured data; meanwhile, in the performance evaluation of the high-strength engineering plastic, nonlinear relations between input features such as component composition and process parameters and performance output are mainly focused, the multi-layer perceptron model P has good fitting capability, can learn complex nonlinear mapping, can better capture complex relations between input and output, and can comprise the following steps in a specific establishment process:
data preparation: and collecting component composition information, formula proportion information and process parameter information of the high-strength engineering plastic sample, and finishing the information into a structured data set, and meanwhile, collecting an application scene performance analysis result corresponding to the structured data set as tag data.
Data preprocessing: preprocessing the input data, including data normalization, missing value processing, feature selection, etc., ensures that the input data is on the same scale and processes any abnormal or missing data.
Determining a model structure: each layer of the multi-layer perceptron contains a plurality of neurons, and in determining the model structure, the dimensions of the input features and the number of target properties need to be considered, and the number of neurons per layer and the choice of activation functions need to be determined.
And (3) constructing a model: and establishing a multi-layer perceptron model according to the determined model structure. The first layer of the model is an input layer, and the number of neurons of the first layer corresponds to the characteristic number of input data. The middle layer of the model is a hidden layer, the number of the neurons of each hidden layer can be adjusted according to the embodiment, the activation function can be adjusted according to the complexity of the problem, the last layer of the model is an output layer, and the number of the neurons corresponds to the number of the output performance evaluation results.
Compiling a model: the optimizers, loss functions and evaluation indexes of the multi-layer perceptron model are configured, and an appropriate optimization algorithm (such as random gradient descent), loss functions (such as mean square error or cross entropy) and evaluation indexes (such as accuracy or root mean square error) are selected to train and evaluate the model.
Training a model: and inputting the prepared data set into a multi-layer perceptron model, training the model, and continuously adjusting the network weight and bias through a back propagation algorithm to gradually fit the relation between the input data and the output label by the model.
And (3) verifying a model: and monitoring and adjusting the training process by using the verification set to prevent overfitting, and determining whether the model needs to be adjusted by monitoring a loss function and an evaluation index of the model.
Test model: and evaluating the performance of the trained multi-layer perceptron model on unseen data by using a test set, comparing the model prediction result with real performance data, and evaluating the accuracy and generalization capability of the model.
Performance evaluation results: according to the performance evaluation result on the test set, the predicted values of the high-strength engineering plastic sample on the application scene performance are obtained, and the predicted values can be used as the performance evaluation result to carry out further optimization and decision.
Example two
As shown in fig. 8, a high-strength engineering plastic evaluation system based on component analysis includes:
the detection module 01 is used for respectively detecting the component composition, the formula proportion and the process parameters of the high-strength engineering plastic to be detected to obtain component composition information, formula proportion information and process parameter information;
the performance analysis module 02 is used for analyzing the performance of the structure formed by the high-strength engineering plastic to obtain an application scene performance analysis result;
the neural network module 03 takes component composition information, formula proportion information and process parameter information as input to realize performance evaluation of the high-strength engineering plastics and obtain a performance evaluation result;
and the optimization processing module 04 processes the application scene performance analysis result and the performance evaluation result, and optimizes the component composition, the formula proportion and the process parameters of the high-strength engineering plastic or maintains the original component composition, the formula proportion and the process parameters according to the processing result.
As a preference of the present embodiment, as shown in fig. 9, the optimization processing module 04 includes:
the feature importance evaluation module 041 enumerates all the performance features in the application scene performance analysis result and the performance evaluation result, and obtains the weight value of the importance of each performance feature in the application scene;
the normalization processing module 042 performs normalization processing on each performance characteristic, and maps the numerical value of each performance characteristic to a uniform range so as to facilitate subsequent calculation;
the comprehensive evaluation module 043 multiplies the normalized performance characteristics by corresponding weight values, and sums the weighted performance characteristics to obtain a comprehensive evaluation index;
the optimization decision module 044 sets a comprehensive index threshold according to actual demands, and if the comprehensive evaluation index is lower than the threshold, the optimization of the component composition, the formula proportion and the process parameters of the high-strength engineering plastic is triggered; otherwise, the original component composition, the formula proportion and the technological parameters are maintained.
The technical effects that can be achieved in this embodiment are the same as those in the above embodiment, and will not be described here again.
The foregoing has outlined and described the basic principles, features, and advantages of the present invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, and that the above embodiments and descriptions are merely illustrative of the principles of the present invention, and various changes and modifications may be made without departing from the spirit and scope of the invention, which is defined in the appended claims. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (5)
1. The high-strength engineering plastic evaluation method based on component analysis is characterized by comprising the following steps of:
detecting the component composition, the formula proportion and the process parameters of the high-strength engineering plastic to be detected respectively to obtain component composition information, formula proportion information and process parameter information respectively;
analyzing the performance of the structure formed by the high-strength engineering plastic to obtain an application scene performance analysis result;
taking the component composition information, the formula proportion information and the process parameter information as inputs of a neural network model, and performing performance evaluation of the high-strength engineering plastic to obtain a performance evaluation result; the neural network model is a multi-layer perceptron model and comprises an input layer, a plurality of hidden layers and an output layer; the neuron number of the hidden layer is regulated according to the activation value of the activation function, and the regulating step comprises the following steps: tracking activation values of neurons in each of the hidden layers during training of the model; after the training period of every set number is finished, according to the activation value condition of each neuron in each hidden layer, the following steps are carried out: setting an upper threshold and a lower threshold for an activation value, and comparing the activation value with the upper threshold and the lower threshold; increasing the number of neurons when the activation value of the neurons exceeds the upper threshold, decreasing the number of neurons when the activation value of the neurons is below the lower threshold, otherwise, maintaining the number of neurons unchanged; processing the application scene performance analysis result and the performance evaluation result, including enumerating all performance characteristics in the application scene performance analysis result and the performance evaluation result, and obtaining weight values of importance of the performance characteristics in the application scene; carrying out normalization processing on each performance characteristic, and mapping the numerical value of each performance characteristic to a uniform range; multiplying the normalized performance characteristics with the corresponding weight values, and summing after weighting to obtain a comprehensive evaluation index; setting a comprehensive index threshold according to actual requirements, optimizing the component composition, the formula proportion and the process parameters of the high-strength engineering plastic if the comprehensive evaluation index is lower than the comprehensive index threshold, otherwise, maintaining the original component composition, the original formula proportion and the original process parameters; optimizing the component composition, the formula proportion and the process parameters of the high-strength engineering plastic according to the processing result, or maintaining the original component composition, the original formula proportion and the original process parameters;
enumerating all performance characteristics in the performance analysis result and the performance evaluation result of the application scene, and obtaining a weight value of importance of each performance characteristic in the application scene, wherein the weight value comprises the following steps:
collecting original data of an application scene performance analysis result and a performance evaluation result of the high-strength engineering plastic, and preprocessing the original data; determining a weight value of the importance of each performance feature in an application scene by using a feature importance evaluation method;
introducing a multi-objective optimization algorithm, finding a group of optimal combinations in all the performance characteristics, wherein the method comprises the steps of defining an objective function, taking a plurality of the performance characteristics and corresponding weight values as input by the objective function, and outputting a comprehensive performance index of the high-strength engineering plastic; according to the requirements and optimization targets of the application scene, comprehensively considering the maximized comprehensive performance index and the minimized comprehensive performance index; setting constraint conditions, wherein the constraint conditions are the constraint conditions of the optimal combination; selecting and running a multi-objective optimization algorithm, inputting the objective function, the optimization objective and the constraint condition into the selected multi-objective optimization algorithm, and running the multi-objective optimization algorithm to search for an optimal combination; and obtaining the weight value of each performance characteristic in the optimal combination; selecting and running a multi-objective optimization algorithm, inputting the objective function, the optimization objective and the constraint condition into the selected multi-objective optimization algorithm, and running the multi-objective optimization algorithm to search for an optimal combination, wherein the method comprises the following steps: selecting a multi-objective optimization algorithm; randomly generating an initial population, each individual within the initial population representing a set of performance feature combinations; substituting each individual into an objective function, and calculating the fitness value of each individual on a plurality of objective functions, wherein the calculation mode of determining the fitness value according to an optimization target is maximized or minimized; generating new individuals through genetic algorithm operation, selecting excellent individuals according to fitness values and Pareto optimal solution sets, and enabling the excellent individuals to reproduce new individuals so as to gradually optimize population; checking whether the genetic algorithm converges to a Pareto optimal solution set, and judging whether to terminate the optimization process according to a preset stopping condition; and when the genetic algorithm meets a preset stopping condition, outputting a Pareto optimal solution set as the optimal combination.
2. The component analysis-based high-strength engineering plastic evaluation method according to claim 1, wherein determining a weight value of importance of each of the performance characteristics in an application scene using a characteristic importance evaluation method comprises:
dividing the original data in the application scene to obtain a data set corresponding to each performance characteristic one by one;
performing variance calculation for data in each data set;
and assigning a weight value to the performance characteristic according to the size of the variance, wherein the weight value is proportional to the variance.
3. The component analysis-based high strength engineering plastic evaluation method according to claim 1, further comprising: in the model construction stage, setting the maximum neuron number upper limit of each hidden layer, and when the neuron number of the hidden layer reaches the upper limit:
stopping to continue increasing the number of neurons;
or setting a rejection threshold value lower than the lower limit threshold value, and rejecting the neuron corresponding to the activation value when the activation value is smaller than the rejection threshold value.
4. A component analysis-based high-strength engineering plastic evaluation system employing a component analysis-based high-strength engineering plastic evaluation method according to claim 1, comprising:
the detection module is used for respectively detecting the component composition, the formula proportion and the process parameters of the high-strength engineering plastic to be detected to obtain component composition information, formula proportion information and process parameter information;
the performance analysis module is used for analyzing the performance of the structure formed by the high-strength engineering plastic to obtain an application scene performance analysis result;
the neural network module takes the component composition information, the formula proportion information and the process parameter information as input to realize the performance evaluation of the high-strength engineering plastic and obtain a performance evaluation result;
and the optimizing processing module is used for processing the application scene performance analysis result and the performance evaluation result, and optimizing the component composition, the formula proportion and the process parameters of the high-strength engineering plastic according to the processing result or maintaining the original component composition, the original formula proportion and the original process parameters.
5. The component analysis-based high strength engineering plastic evaluation system of claim 4, wherein the optimization processing module comprises:
the feature importance evaluation module enumerates all the performance features in the application scene performance analysis result and the performance evaluation result, and obtains the weight value of the importance of each performance feature in the application scene;
the normalization processing module performs normalization processing on each performance characteristic and maps the numerical value of each performance characteristic to a uniform range;
the comprehensive evaluation module multiplies the normalized performance characteristics with corresponding weight values, and sums the weighted performance characteristics to obtain a comprehensive evaluation index;
the optimization decision module is used for setting a comprehensive index threshold according to actual requirements, and triggering to optimize the component composition, the formula proportion and the process parameters of the high-strength engineering plastic if the comprehensive evaluation index is lower than the comprehensive index threshold; otherwise, the original component composition, the formula proportion and the technological parameters are maintained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310937200.9A CN116913435B (en) | 2023-07-27 | 2023-07-27 | High-strength engineering plastic evaluation method and system based on component analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310937200.9A CN116913435B (en) | 2023-07-27 | 2023-07-27 | High-strength engineering plastic evaluation method and system based on component analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116913435A CN116913435A (en) | 2023-10-20 |
| CN116913435B true CN116913435B (en) | 2024-01-26 |
Family
ID=88364591
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310937200.9A Active CN116913435B (en) | 2023-07-27 | 2023-07-27 | High-strength engineering plastic evaluation method and system based on component analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116913435B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116837422B (en) * | 2023-07-24 | 2024-01-26 | 扬中凯悦铜材有限公司 | A production process and system for high-purity oxygen-free copper materials |
| CN117371341B (en) * | 2023-12-07 | 2024-04-09 | 小米汽车科技有限公司 | Process parameter determining method, target piece manufacturing device and medium |
| CN117610435B (en) * | 2024-01-22 | 2024-04-09 | 天津路联智通交通科技有限公司 | Automatic proportioning method and system for civil and architectural construction mixed materials |
| CN118155760B (en) * | 2024-03-13 | 2024-10-22 | 镇江先锋植保科技有限公司 | Preparation process and system of pyribenzoxim-pyrithiobac-sodium composition |
| CN118917582B (en) * | 2024-07-12 | 2025-02-25 | 浙江祺阳新材料有限公司 | Unsaturated polyester structure adjustment method combined with application demand analysis |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110423444A (en) * | 2019-06-21 | 2019-11-08 | 朱忠义 | A kind of preparation method of high intensity thermosetting plastics |
| CN110837523A (en) * | 2019-10-29 | 2020-02-25 | 山东大学 | High-confidence reconstruction quality and false-transient-reduction quantitative evaluation method based on cascade neural network |
| CN112037011A (en) * | 2020-08-14 | 2020-12-04 | 百维金科(上海)信息科技有限公司 | Credit scoring method based on FOA-RBF neural network |
| CN112115579A (en) * | 2020-08-12 | 2020-12-22 | 江苏师范大学 | Multi-target optimization method for injection molding process parameters of glass fiber reinforced plastics |
| WO2021007812A1 (en) * | 2019-07-17 | 2021-01-21 | 深圳大学 | Deep neural network hyperparameter optimization method, electronic device and storage medium |
| CN113733521A (en) * | 2020-05-30 | 2021-12-03 | 湖北盛智塑料包装有限公司 | Production process of high-density high-strength plastic milk bottle |
| CN114193884A (en) * | 2021-12-30 | 2022-03-18 | 福建恒安卫生材料有限公司 | High-strength lightweight antibacterial plastic film and preparation method thereof |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100210745A1 (en) * | 2002-09-09 | 2010-08-19 | Reactive Surfaces, Ltd. | Molecular Healing of Polymeric Materials, Coatings, Plastics, Elastomers, Composites, Laminates, Adhesives, and Sealants by Active Enzymes |
-
2023
- 2023-07-27 CN CN202310937200.9A patent/CN116913435B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110423444A (en) * | 2019-06-21 | 2019-11-08 | 朱忠义 | A kind of preparation method of high intensity thermosetting plastics |
| WO2021007812A1 (en) * | 2019-07-17 | 2021-01-21 | 深圳大学 | Deep neural network hyperparameter optimization method, electronic device and storage medium |
| CN110837523A (en) * | 2019-10-29 | 2020-02-25 | 山东大学 | High-confidence reconstruction quality and false-transient-reduction quantitative evaluation method based on cascade neural network |
| CN113733521A (en) * | 2020-05-30 | 2021-12-03 | 湖北盛智塑料包装有限公司 | Production process of high-density high-strength plastic milk bottle |
| CN112115579A (en) * | 2020-08-12 | 2020-12-22 | 江苏师范大学 | Multi-target optimization method for injection molding process parameters of glass fiber reinforced plastics |
| CN112037011A (en) * | 2020-08-14 | 2020-12-04 | 百维金科(上海)信息科技有限公司 | Credit scoring method based on FOA-RBF neural network |
| CN114193884A (en) * | 2021-12-30 | 2022-03-18 | 福建恒安卫生材料有限公司 | High-strength lightweight antibacterial plastic film and preparation method thereof |
Non-Patent Citations (1)
| Title |
|---|
| 1992年我国工程塑料进展;孙安垣, 刘传新, 贾崇明, 张晓云;工程塑料应用(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116913435A (en) | 2023-10-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116913435B (en) | High-strength engineering plastic evaluation method and system based on component analysis | |
| Rankovic et al. | A new approach to software effort estimation using different artificial neural network architectures and Taguchi orthogonal arrays | |
| Attarzadeh et al. | Proposing a new software cost estimation model based on artificial neural networks | |
| CN113408869A (en) | Power distribution network construction target risk assessment method | |
| CN108241901B (en) | Transformer early warning evaluation method and device based on prediction data | |
| CN111028100A (en) | Refined short-term load prediction method, device and medium considering meteorological factors | |
| CN119597493A (en) | Distributed computing resource intelligent evolution method and system based on digital twin | |
| CN113012766A (en) | Self-adaptive soft measurement modeling method based on online selective integration | |
| CN110659767A (en) | Stock trend prediction method based on LSTM-CNN deep learning model | |
| CN119493393B (en) | Intelligent model data processing method applied to metal smelting | |
| CN119476878B (en) | An AGV-based intelligent material distribution management optimization method and system | |
| CN119941312A (en) | Concrete operation robot industry scale prediction method, system, equipment and medium | |
| CN119002605B (en) | A mass spectrometry detection kit storage environment control method and system | |
| CN112149870B (en) | Pig house ammonia concentration combination prediction method based on ISODATA clustering and Elman neural network | |
| CN114330818A (en) | Dynamic water demand prediction method based on main driving factor screening and deep learning | |
| CN117724425B (en) | Industrial process real-time prediction method and system based on recursion random configuration network | |
| CN113722098A (en) | Server capacity prediction method, device, equipment and medium based on artificial intelligence | |
| CN118965223A (en) | Abnormal diagnosis method and system for oil-free dry vacuum unit | |
| CN118428557A (en) | Multi-target-oriented self-adaptive threat degree ordering method and system | |
| CN116956160A (en) | Data classification prediction method based on self-adaptive tree species algorithm | |
| Ordieres-Meré et al. | Comparison of models created for the prediction of the mechanical properties of galvanized steel coils | |
| CN116822292A (en) | Data-driven-based automobile seat slide rail peel strength prediction method | |
| Thomas et al. | The evolution of forecasting techniques traditional versus machine learning methods | |
| Musial et al. | Improving the Efficiency of Production Processes by Reducing Human Errors Using Intelligent Methods | |
| Liu et al. | Use clustering to improve neural network in financial time series prediction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |