CN117272683B - Model parameter optimization method and system for magnetic hysteresis effect simulation of ferromagnetic material - Google Patents
Model parameter optimization method and system for magnetic hysteresis effect simulation of ferromagnetic material Download PDFInfo
- Publication number
- CN117272683B CN117272683B CN202311498049.XA CN202311498049A CN117272683B CN 117272683 B CN117272683 B CN 117272683B CN 202311498049 A CN202311498049 A CN 202311498049A CN 117272683 B CN117272683 B CN 117272683B
- Authority
- CN
- China
- Prior art keywords
- sample
- model
- hysteresis
- result
- clustering
- 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
- 238000005457 optimization Methods 0.000 title claims abstract description 116
- 238000004088 simulation Methods 0.000 title claims abstract description 82
- 239000003302 ferromagnetic material Substances 0.000 title claims abstract description 62
- 238000000034 method Methods 0.000 title claims abstract description 56
- 230000000694 effects Effects 0.000 title claims abstract description 30
- 230000005291 magnetic effect Effects 0.000 title abstract description 14
- 238000010276 construction Methods 0.000 claims abstract description 36
- 238000004458 analytical method Methods 0.000 claims abstract description 29
- 238000013507 mapping Methods 0.000 claims abstract description 18
- 238000004364 calculation method Methods 0.000 claims description 35
- 230000003044 adaptive effect Effects 0.000 claims description 15
- 230000005855 radiation Effects 0.000 claims description 13
- 230000003313 weakening effect Effects 0.000 claims description 12
- 230000009172 bursting Effects 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 8
- 230000008569 process Effects 0.000 description 7
- 239000013598 vector Substances 0.000 description 6
- 238000004804 winding Methods 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 3
- 230000006698 induction Effects 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 230000002427 irreversible effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000005415 magnetization Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 238000011524 similarity measure Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Probability & Statistics with Applications (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application provides a model parameter optimization method and system for magnetic hysteresis effect simulation of a ferromagnetic material, and relates to the technical field of model optimization, wherein the method comprises the following steps: establishing an initial hysteresis model of a ferromagnetic material, constructing a simulation circuit through simulation software, establishing model connection, executing simulation fitting, updating model parameters, establishing a sample feature space, establishing a sample set, performing trust analysis, establishing a mapping trust value, performing sample clustering, performing self-adaptive optimizing updating of the model parameters according to a clustering result, and finally completing the construction of the hysteresis model. The method solves the problems that the number of parameters is large, the model construction cannot be carried out on ferromagnetic materials with different types and different purposes, and the accuracy of simulation and prediction of hysteresis effects is low. By using the self-adaptive optimizing updating method of the sample clustering center, the parameters of the initial hysteresis model can be effectively optimized, and the more accurate and practical hysteresis model can be obtained.
Description
Technical Field
The application relates to the technical field of model optimization, in particular to a model parameter optimization method and system for magnetic hysteresis effect simulation of a ferromagnetic material.
Background
The application background of the model parameter optimization method for the hysteresis effect simulation of the ferromagnetic material is mainly in the fields of power, electronics, communication and the like. In these fields, ferromagnetic materials such as transformers, motors, generators, etc. are used in a very wide variety of applications. However, the operational performance of these ferromagnetic materials is affected by their hysteresis effects, which makes accurate simulation and prediction of such effects very important. Conventional hysteresis models, such as the Ja model, can be used to describe the hysteresis characteristics of ferromagnetic materials, but the optimization of parameters of these models has been a challenge.
In the process of realizing the technical scheme of the invention in the embodiment of the application, the technology at least has the following technical problems:
the number of parameters is large, the model construction can not be carried out on ferromagnetic materials with different types and different purposes, and the accuracy of the simulation and the prediction of hysteresis effects is low.
Disclosure of Invention
The method solves the problems that the number of parameters is large, the model construction cannot be carried out on ferromagnetic materials with different types and different purposes, and the accuracy of simulation and prediction of hysteresis effects is low.
In view of the above problems, the present application provides a method and a system for optimizing model parameters for magnetic hysteresis effect simulation of a ferromagnetic material, and in a first aspect, the present application provides a method for optimizing model parameters for magnetic hysteresis effect simulation of a ferromagnetic material, where the method includes: establishing an initial hysteresis model of the ferromagnetic material, and constructing a simulation circuit through simulation software; connecting the simulation circuit with the initial hysteresis model, executing simulation fitting of the ferromagnetic material, and updating model parameters of the initial hysteresis model based on simulation fitting results; establishing a sample feature space according to a simulation fitting result, collecting hysteresis samples of sample ferromagnetic materials based on the sample feature space, and establishing a sample set; sample trust analysis is carried out on the sample set, and a mapping trust value of each sample in the sample set is established; sample clustering is carried out on the sample set, a sample clustering result is generated, and clustering characteristics comprise control parameter characteristics, hysteresis result characteristics and trust value characteristics of samples; and carrying out self-adaptive optimizing updating of model parameters by using a clustering center of the sample clustering result, and completing model parameter optimization of an initial hysteresis model according to the self-adaptive optimizing updating result to complete construction of the hysteresis model.
In a second aspect, the present application provides a model parameter optimization system for simulation of hysteresis effects of ferromagnetic materials, the system comprising: the simulation circuit construction module is used for establishing an initial hysteresis model of the ferromagnetic material and constructing a simulation circuit through simulation software; the model parameter updating module is used for connecting the simulation circuit with the initial hysteresis model, executing simulation fitting of the ferromagnetic material and updating model parameters of the initial hysteresis model based on a simulation fitting result; the sample set establishing module is used for establishing a sample feature space according to a simulation fitting result, collecting hysteresis samples of sample ferromagnetic materials based on the sample feature space and establishing a sample set; the mapping trust value establishing module is used for carrying out sample trust analysis on the sample set and establishing a mapping trust value of each sample in the sample set; the sample clustering result generation module is used for carrying out sample clustering on the sample set to generate a sample clustering result, and the clustering characteristics comprise control parameter characteristics, hysteresis result characteristics and trust value characteristics of the sample; and the hysteresis model construction module is used for carrying out self-adaptive optimizing updating of model parameters by using a clustering center of the sample clustering result, completing model parameter optimization of an initial hysteresis model according to the self-adaptive optimizing updating result, and completing the construction of the hysteresis model.
One or more technical solutions provided in the present application have at least the following technical effects or advantages:
the application provides a model parameter optimization method and system for magnetic hysteresis effect simulation of a ferromagnetic material, and relates to the technical field of model optimization, wherein the method comprises the following steps: establishing an initial hysteresis model of a ferromagnetic material, constructing a simulation circuit through simulation software, establishing model connection, executing simulation fitting, updating model parameters, establishing a sample feature space, establishing a sample set, performing trust analysis, establishing a mapping trust value, performing sample clustering, performing self-adaptive optimizing updating of the model parameters according to a clustering result, and finally completing the construction of the hysteresis model.
The method solves the problems that the number of parameters is large, the model construction cannot be carried out on ferromagnetic materials with different types and different purposes, and the accuracy of simulation and prediction of hysteresis effects is low. By using the self-adaptive optimizing updating method of the sample clustering center, the parameters of the initial hysteresis model can be effectively optimized, and the more accurate and practical hysteresis model can be obtained.
The foregoing description is merely an overview of the technical solutions of the present application, and may be implemented according to the content of the specification in order to make the technical means of the present application more clearly understood, and in order to make the above-mentioned and other objects, features and advantages of the present application more clearly understood, the following detailed description of the present application will be given.
Drawings
For a clearer description of the technical solutions of the present application or of the prior art, the drawings used in the description of the embodiments or of the prior art will be briefly described below, it being obvious that the drawings in the description below are only exemplary and that other drawings can be obtained, without inventive effort, by a person skilled in the art from the drawings provided.
Fig. 1 is a schematic flow chart of a model parameter optimization method for magnetic hysteresis effect simulation of a ferromagnetic material according to an embodiment of the present application.
Fig. 2 is a schematic flow chart of a method for completing construction of a hysteresis model according to an optimization constraint result in a model parameter optimization method for simulation of a hysteresis effect of a ferromagnetic material according to an embodiment of the present application.
Fig. 3 is a schematic flow chart of a method for adaptive optimization updating compensation in a model parameter optimization method for hysteresis effect simulation of a ferromagnetic material according to an embodiment of the present application.
Fig. 4 is a schematic structural diagram of a model parameter optimization system for magnetic hysteresis effect simulation of a ferromagnetic material according to an embodiment of the present application.
Reference numerals illustrate: the system comprises a simulation circuit construction module 10, a model parameter updating module 20, a sample set construction module 30, a mapping trust value construction module 40, a sample clustering result generation module 50 and a hysteresis model construction module 60.
Detailed Description
The method solves the problems that the number of parameters is large, the model construction cannot be carried out on ferromagnetic materials with different types and different purposes, and the accuracy of simulation and prediction of hysteresis effects is low. By using the self-adaptive optimizing updating method of the sample clustering center, the parameters of the initial hysteresis model can be effectively optimized, and the more accurate and practical hysteresis model can be obtained.
For a better understanding of the foregoing technical solutions, the following detailed description will be given with reference to the accompanying drawings and specific embodiments of the present invention:
example 1
A method for optimizing model parameters for a hysteresis effect simulation of a ferromagnetic material, as shown in fig. 1, the method comprising:
establishing an initial hysteresis model of the ferromagnetic material, and constructing a simulation circuit through simulation software;
specifically, an initial hysteresis model of ferromagnetic material is established: the hysteresis is decomposed into an irreversible component mirr with a friction effect and an elastically reversible component mrev. The relationship between the magnetization m and the magnetic field strength h is described by a modified langevin function, and finally the relationship between the magnetic induction b and h is obtained. Constructing a simulation circuit: and constructing a simulation circuit, and using MATLAB or Simulink. In this software, a circuit model may be constructed, which includes a transformer and a core winding. In the simulation process, the hysteresis characteristics of different ferromagnetic materials can be simulated by adjusting parameters in a model, and the influence of the hysteresis characteristics on the circuit performance is observed. A new model window is created in MATLAB/Simulink. The required modules, such as transformers and core windings, are added to the model window. Parameters of the transformer, such as transformation ratio and leakage reactance, are configured. Parameters of the core winding, such as resistance and inductance, are configured. The initial hysteresis model is connected to the core winding and parameters in the initial hysteresis model are set.
Connecting the simulation circuit with the initial hysteresis model, executing simulation fitting of the ferromagnetic material, and updating model parameters of the initial hysteresis model based on simulation fitting results;
specifically, an initial hysteresis model is connected to the core winding, and parameters in the initial hysteresis model are set. These parameters may be coefficients in hysteresis characteristic equations or other relevant parameters. And (3) running simulation, and observing the hysteresis phenomenon of the iron core winding and the change condition of circuit performance. Based on the simulation results, parameters in the initial hysteresis model, such as coefficients in the hysteresis characteristic equation, are adjusted. Repeating the steps until a better simulation result is obtained. And then updating model parameters of the initial hysteresis model based on simulation results, and gradually improving the accuracy and applicability of the model by continuously updating the model parameters of the initial hysteresis model, so that the hysteresis characteristics of the ferromagnetic material can be better described.
Establishing a sample feature space according to a simulation fitting result, collecting hysteresis samples of sample ferromagnetic materials based on the sample feature space, and establishing a sample set;
specifically, after the simulation fitting result is obtained, it can be used as a basis for the sample feature space. The sample feature space is a vector space composed of a plurality of features, wherein each vector represents the hysteresis characteristics of a ferromagnetic material. Firstly, the simulation fitting result needs to be processed, and hysteresis characteristic parameters of the ferromagnetic material, such as hysteresis curves, coercive force, remanence and the like, are extracted. These parameters can be used as characteristics of the sample feature space to describe the hysteresis characteristics of the ferromagnetic material. Next, hysteresis sample collection of the sample ferromagnetic material is performed based on the sample feature space. Hysteresis samples of different ferromagnetic materials can be acquired by experiment or actual test and expressed as vectors in the sample feature space. These vectors constitute a sample set. By creating a sample set, the hysteresis characteristics of ferromagnetic materials and their impact on circuit performance can be further analyzed and studied.
Sample trust analysis is carried out on the sample set, and a mapping trust value of each sample in the sample set is established;
specifically, sample trust analysis of a sample set may be achieved by a trust evaluation method. And determining characteristic indexes of a sample characteristic space, such as hysteresis curves, coercive force, remanence and the like. Each feature indicator is assigned a weight that indicates how important it is in trust evaluation. For each sample, according to the characteristic index value and the corresponding weight, the distribution of the weight needs to be reasonably adjusted according to the actual situation so as to reflect the importance of different characteristic indexes in trust evaluation. And calculating the trust value thereof. And (3) comprehensively considering a plurality of characteristic indexes of each sample by adopting a fuzzy comprehensive evaluation method, and calculating the comprehensive trust value of the sample. And sorting or classifying each sample according to the comprehensive trust value to establish a mapping trust value of each sample in the sample set. By establishing a mapped trust value for each sample in the sample set, the trust level of each sample in the sample set can be better understood.
Sample clustering is carried out on the sample set, a sample clustering result is generated, and clustering characteristics comprise control parameter characteristics, hysteresis result characteristics and trust value characteristics of samples;
specifically, clustering a sample set, and extracting features: control parameter features, hysteresis result features, and trust value features are extracted from each sample. These characteristics may include control of current, control of magnetic field strength, shape parameters of hysteresis loops, saturation induction, residual induction, etc. The data is then preprocessed: the extracted features are subjected to necessary preprocessing such as missing value filling, outlier processing, standardization and the like so as to ensure the consistency and the effectiveness of the data. Reconstructing a feature vector: all the features of each sample are combined into a feature vector to form a feature matrix. And (3) selecting a clustering algorithm: suitable clustering algorithms, such as K-means, hierarchical clustering, DBSCAN, etc., are selected. K-means are used herein as an example. And (3) running a clustering algorithm: and taking the feature matrix as input, and running a K-means algorithm to perform clustering. The appropriate number of clusters K is selected as required. Evaluating the clustering result: various indexes such as profile coefficients, calinski-Harabasz indexes and the like can be used for evaluating the advantages and disadvantages of the clustering result. Through the steps, the sample set can be clustered, and a sample clustering result comprising the control parameter characteristic, the hysteresis result characteristic and the trust value characteristic is generated. This helps to better understand and master the properties and behavior of ferromagnetic materials, providing guidance for material selection and optimization in practical applications.
And carrying out self-adaptive optimizing updating of model parameters by using a clustering center of the sample clustering result, and completing model parameter optimization of an initial hysteresis model according to the self-adaptive optimizing updating result to complete construction of the hysteresis model.
Specifically, after the sample clustering result is obtained, the clustering center can be further utilized to perform self-adaptive optimization updating of the model parameters so as to optimize the parameters of the initial hysteresis model and complete the construction of the hysteresis model. Extracting a clustering center: the average or median of the samples is extracted from each cluster as the center of the cluster. These centers can be considered as representative of the samples in the cluster. Defining an optimization objective function: an objective function is defined for evaluating the merits of the model parameters. The objective function may be defined according to actual requirements, for example, may be a sum of squares, a sum of absolute values, etc. of the prediction errors. Initializing model parameters: the parameters of the initial hysteresis model are initialized, and can be randomly assigned or set according to priori knowledge. Iterative optimization: and continuously adjusting model parameters in an iterative mode, and calculating the value of the objective function. In each iteration, judging the merits of the model parameters according to the values of the objective functions, and updating the model parameters. Updating model parameters: the model parameters are updated according to the values of the objective function and certain optimization strategies (such as gradient descent, genetic algorithm, etc.). Judging a stopping condition: after each iteration, it is checked whether the stop condition is met. The stop condition may be that the value of the objective function reaches a preset minimum threshold, reaches a preset maximum number of iterations, etc. And (3) optimizing: and when the stopping condition is met, ending the iterative optimization to obtain the optimized model parameters. Building a hysteresis model: the optimized model parameters are applied to an initial hysteresis model, and the hysteresis model after construction can be used for hysteresis characteristic prediction and simulation of ferromagnetic materials. By using the self-adaptive optimizing updating method of the sample clustering center, the parameters of the initial hysteresis model can be effectively optimized, and the more accurate and practical hysteresis model can be obtained. This helps to better understand and predict the hysteresis characteristics of ferromagnetic materials, providing powerful support for material selection, design and optimization in practical applications.
Further, as shown in fig. 2, the method of the present application further includes:
determining a sample extremum based on the sample feature space, and generating demand data, wherein the demand data is a model precision demand for building a hysteresis model;
configuring step length constraint according to the sample extremum and the requirement data, and distributing authentication nodes according to the sample extremum and the step length constraint;
setting a node parameter optimization limit value at the authentication node, and carrying out optimization constraint of the model parameter optimization through the parameter optimization limit value;
and completing the construction of the hysteresis model according to the optimization constraint result.
Specifically, a sample extremum is determined based on a sample feature space, demand data is generated, step length constraint is configured according to the sample extremum and the demand data, authentication nodes are distributed according to the sample extremum and the step length constraint, node parameter optimization limit values are set at the authentication nodes, and optimization constraint of model parameter optimization is carried out through the parameter optimization limit values, so that hysteresis model construction can be completed. Determining a sample extremum based on the sample feature space: by analyzing the sample feature space, the maximum and minimum values of each feature, i.e., sample extremum, can be determined. These sample extrema may reflect the characteristic distribution and range of each sample in the sample set. Generating demand data: and generating a model precision requirement for building the hysteresis model according to the actual requirement and the target. These requirements may include the sum of squares, the sum of absolute values, etc. of the prediction errors. Configuring step length constraint: from the sample extremum and model accuracy requirements, the step size constraints used in the optimization process can be determined. The step size constraint can ensure that the set range is not exceeded each time the model parameters are updated in the optimization process. Distribution authentication node: from the sample extremum and the step size constraint, it can be determined at which nodes to perform parameter optimization. These nodes can be regarded as key points in the optimization process, and parameter authentication and optimization are required on these nodes. Setting node parameter optimization limit values: at the authentication node, a limit value for node parameter optimization may be set. These limits can ensure that during the optimization process, the model is not over-fitted or unstable due to over-optimization. And (3) performing optimization constraint of model parameter optimization: according to the set parameter optimization limit value, optimization constraint can be carried out on model parameters. In the optimization process, the variation range of the model parameters is limited, and the set limit value is prevented from being exceeded. And (3) completing the construction of a hysteresis model: and according to the result of the optimization constraint, the construction of the hysteresis model can be completed. The optimized model parameters are applied to a hysteresis model, and can be used for the hysteresis characteristic prediction and simulation of ferromagnetic materials. Through the steps, the sample extremum can be determined based on the sample feature space, the demand data is generated, step length constraint is configured according to the sample extremum and the demand data, authentication nodes are distributed according to the sample extremum and the step length constraint, node parameter optimization limit values are set at the authentication nodes, optimization constraint of model parameter optimization is carried out through the parameter optimization limit values, and finally the construction of a hysteresis model is completed. This helps to better understand and predict the hysteresis characteristics of ferromagnetic materials, providing powerful support for material selection, design and optimization in practical applications.
Further, as shown in fig. 3, the method of the present application further includes:
invoking the clustering center through the authentication node, and executing association matching;
performing authentication node bursting analysis of the clustering center corresponding to the association matching result based on the node parameter optimization limit value to generate a bursting value;
performing radiation analysis of the adjacent area through the burst value, and generating parameter optimization compensation based on a radiation analysis result;
and executing the self-adaptive optimizing updating compensation of the model parameters through the parameter optimizing compensation.
Specifically, the clustering center is called by the authentication node: the cluster center can now be invoked by the authentication node for association matching. Performing association matching: and carrying out matching analysis through the association relation between the authentication node and the clustering center. The similarity degree between the sample and the clustering center can be evaluated by means of similarity measurement, a classification algorithm and the like, so that matching is performed. And carrying out authentication node burst analysis of the clustering center corresponding to the association matching result based on the node parameter optimization limit value: after the associative matching result is obtained, the validity and feasibility of the matching result can be evaluated based on the node parameter optimization limit value. The node parameter optimization limit value here refers to a parameter variation range set when parameter optimization is performed. If the parameter of a certain authentication node exceeds a set limit value, this authentication node is considered to break through the parameter optimization limit. Generating a burst value: for each broken authentication node, its break-through value may be calculated. The burst value can reflect the contribution degree of the node to the whole hysteresis model and can also be used for subsequent compensation analysis. Radiation analysis of the vicinity was performed by burst value: for each broken authentication node, radiation analysis can be performed towards the adjacent area with the authentication node as a center. The adjacent region may refer to an adjacent region in the physical space or an adjacent region in the parameter space. Through radiation analysis, other nodes related to the burst node can be found out, and the influence degree of the other nodes on the whole hysteresis model can be estimated. Generating a parameter optimization compensation based on the radiation analysis result: based on the results of the radiation analysis, a corresponding parameter optimization compensation may be generated for each broken authentication node. The parameter optimization compensation may be an increment or a correction amount for adjusting the model parameters, or may be a compensation value calculated by integrating the burst value and other factors. Executing self-adaptive optimizing updating compensation of model parameters through parameter optimizing compensation: and applying the generated parameter optimization compensation to the corresponding parameters of the hysteresis model, and then carrying out self-adaptive optimization updating of the model again. The adaptive optimizing update refers to re-optimizing the parameters of the model according to new parameter optimization compensation so that the model fits the actual data better. Through the steps, the clustering center can be called by the authentication node and the association matching is executed, and then the authentication node bursting analysis of the association matching result corresponding to the clustering center is executed based on the node parameter optimization limit value and the bursting value is generated. And then carrying out radiation analysis of the adjacent area through the burst value, generating parameter optimization compensation based on a radiation analysis result, and finally executing self-adaptive optimizing updating compensation of the model parameters through the parameter optimization compensation. Therefore, parameters of the hysteresis model can be further optimized, and the fitting effect and the prediction accuracy of the hysteresis model can be improved.
Further, the method of the present application further comprises:
dividing the control parameter characteristics, and determining real control parameters and other control parameters, wherein the real control parameters are simulation fit control parameters;
performing data splitting of the sample set by the real control parameters to generate an initial data splitting set;
and carrying out cluster center search in each initial data splitting set, carrying out sample clustering according to the cluster center search result, and generating the sample clustering result.
Specifically, the control parameter feature is divided: first, the extracted control parameter features are divided. This may be achieved by feature selection, feature transformation or machine learning algorithms. The control parameter features are divided into two parts, namely a real control parameter and other control parameters. The real control parameters refer to parameters that have an important effect on the simulation fit, while other control parameters may have less effect on the simulation fit or have other parameters that are less relevant. Determining real control parameters and other control parameters: after feature segmentation is completed, the actual control parameters and other control parameters may be determined. The real control parameters usually have more remarkable influence, and have important influence on simulation results. Other control parameters may have less impact on simulation results or other less relevant parameters may be present. Data splitting of the sample set is performed with real control parameters: and carrying out data fracturing on the sample set by using the real control parameters to generate an initial data fracturing set. Each initial data burst contains a set of samples with similar real control parameters. Cluster center search is performed in each initial data fracture set: and (5) in each initial data splitting set, searching a clustering center. The cluster center search generally adopts a clustering algorithm such as K-means and the like to search for the cluster center in each initial data splitting set. Sample clustering is carried out according to the search result of the clustering center: and clustering samples in each initial data splitting set according to the searching result of the clustering center. The same clustering algorithm can be adopted to cluster each initial data fracture set, and different clustering algorithms can be adopted to cluster different initial data fracture sets. Generating a sample clustering result: after the sample clustering is completed, a sample clustering result can be obtained. Each cluster represents a set of samples with similar characteristics that can be used for subsequent model construction and optimization. Through the steps, the control parameter characteristics can be segmented, the real control parameters and other control parameters are determined, and data fracturing of the sample set is carried out according to the real control parameters, so that an initial data fracturing set is generated. And then carrying out cluster center search in each initial data splitting set, and carrying out sample clustering according to the cluster center search result to generate a sample clustering result. These results can be used for subsequent model construction and optimization tasks.
Further, the method of the present application further comprises:
calculating parameter fluctuation values of the other control parameters, and carrying out additional feature calculation according to parameter fluctuation value calculation results and the trust value features through preset distribution weights to generate additional feature calculation results, wherein the additional feature calculation results have a mapping relation with the hysteresis result features;
and carrying out feature transparency weakening of hysteresis result features according to the additional feature calculation results, and determining a cluster center search result according to the feature transparency weakening results.
Specifically, parameter fluctuation values of other control parameters are calculated: first, parameter fluctuation values of other control parameters are calculated. This may be achieved by calculating the variance, standard deviation or other relevant index of the control parameter. The parameter fluctuation value can reflect the fluctuation condition of other control parameters, and has certain importance on the influence of hysteresis characteristics. Carrying out additional feature calculation according to the parameter fluctuation value calculation result and the trust value feature through a preset distribution weight: and taking the parameter fluctuation value calculation result and the trust value characteristic as inputs, and carrying out additional characteristic calculation through a preset distribution weight. The predetermined distribution weight may be a predetermined weight distribution pattern for balancing the contributions of the two features in the additional feature calculation. Generating an additional feature calculation result: and generating an additional feature calculation result according to the parameter fluctuation value calculation result, the input of the trust value feature and the distribution of the preset distribution weight. The additional feature calculation result may be a new feature combining both features for characterizing the hysteresis characteristics of the sample. The additional feature calculation result and the hysteresis result feature have a mapping relation: and ensuring that a certain mapping relation exists between the additional feature calculation result and the hysteresis result feature. That is, the change of the calculation result of the additional feature can reflect the change of the hysteresis result feature, which is helpful to improve the accuracy and stability of model prediction. Feature transparent weakening of hysteresis result features is performed by additional feature calculation results: and carrying out characteristic transparency weakening treatment on the hysteresis result characteristic by utilizing the additional characteristic calculation result. This can reduce the complexity of hysteresis characteristics and improve the generalization ability of the model. Determining a cluster center search result according to the feature transparency weakening result: and according to the characteristic transparent weakening result, the result of the cluster center search can be redetermined. The clustering center search generally adopts a clustering algorithm such as K-means and the like, and samples are clustered again according to new characteristic transparent weakening results, so that a more accurate clustering center is obtained. Through the steps, the parameter fluctuation value calculation of other control parameters can be realized, and the additional feature calculation is carried out through the preset distribution weight according to the parameter fluctuation value calculation result and the trust value feature, so that the additional feature calculation result is generated. The additional feature calculation result can be used for reducing the complexity of hysteresis result features and improving the generalization capability of the model. Meanwhile, through feature transparent weakening processing, the result of cluster center searching can be determined again, and the model performance is further optimized.
Further, the method of the present application further comprises:
performing sample searching and clustering by taking the searching result of the clustering center as a datum point;
and taking the characteristic distance of the hysteresis result characteristic as a length constraint, and obtaining the sample clustering result according to the length constraint result, wherein the sample clustering result is provided with a cluster stable value mark.
Specifically, with the cluster center search result as a reference point, sample search clustering is performed: and searching and clustering the sample set by using the search result of the clustering center as a reference point. This may be achieved by calculating the distance or similarity between the sample and the cluster center, assigning the sample to the cluster in which the cluster center closest thereto is located. Taking the characteristic distance of hysteresis result characteristics as a length constraint: and when the sample search clustering is performed, taking the characteristic distance of hysteresis result characteristics as a length constraint condition. The feature distance may be a similarity measure or distance measure that measures the degree of similarity or distance between samples. By setting appropriate length constraints, the range and scale of sample clusters can be limited. Obtaining a sample clustering result according to the length constraint result: and clustering the samples according to the set length constraint conditions. In each cluster, the feature distance between samples should meet the requirements of the length constraint. Sample clustering results have cluster stability value identification: after the sample clustering result is obtained, a cluster stability value identifier is allocated to each cluster. The cluster stability value can be an index for evaluating the stability and quality of the clustering result, and a proper evaluation method can be selected according to actual conditions. For example, the number of samples within each cluster, the average feature distance between samples, or other relevant metrics may be calculated to assess the stability of the clusters. Through the steps, the sample searching and clustering can be executed by taking the searching result of the clustering center as a reference point, the characteristic distance of the hysteresis result characteristic is taken as a length constraint, and the sample clustering result is obtained according to the length constraint result. The sample clustering result is provided with a cluster stability value identifier, and can be used for evaluating the stability and quality of the clustering result.
Further, the method of the present application further comprises:
establishing an adaptive optimization unit, and fitting the adaptive optimization unit to the hysteresis model;
when the self-adaptive optimization unit receives newly added data, data comparison between the newly added data and the sample set is executed;
when the data comparison result meets a preset threshold value, temporarily storing the newly added data into an update space;
and if the data quantity in the update space meets a preset data threshold, executing sample replacement of the update space and the sample set, and completing the parameter reconstruction of the hysteresis model.
Specifically, an adaptive optimization unit is established: the adaptive optimization unit is a mechanism that can dynamically adjust model parameters based on data. It may be an algorithm, model or module that adjusts and optimizes itself according to the characteristics and distribution of the data. Fitting an adaptive optimization unit to the hysteresis model: an adaptive optimization unit is integrated into the hysteresis model to dynamically adjust parameters of the hysteresis model in accordance with changes in the data. This can be achieved by training and optimizing the adaptive optimization unit with a hysteresis model. Receiving newly added data and performing data comparison: when the adaptive optimization unit receives the newly added data, the adaptive optimization unit compares the newly added data with the sample set. This may be accomplished by calculating a similarity, distance, or other related indicator between the newly added data and the sample set. Temporarily storing newly added data meeting a preset threshold value into an update space: and if the similarity, the distance or other related indexes between the newly added data and the sample set meet a preset threshold value, temporarily storing the newly added data into an updating space. The preset threshold can be set according to actual requirements, and is used for judging whether the newly added data are similar or related enough to update the model. Checking whether the data quantity in the update space meets a preset data threshold value: after the newly added data is temporarily stored in the update space, whether the data volume in the update space meets a preset data threshold value is checked. The preset data threshold may be set according to actual requirements to determine when a model update is required. Sample replacement of the update space and the sample set is performed and hysteresis model parameter reconstruction is completed: if the amount of data within the update space meets a preset data threshold, a sample replacement of the update space with the sample set is performed. This may be achieved by selecting a part of the old samples to be replaced from the sample set with new samples in the update space. After the sample replacement is completed, the hysteresis model is retrained and optimized to realize the reconstruction of the hysteresis model parameters.
Example two
Based on the same inventive concept as the model parameter optimization method for the hysteresis effect simulation of the ferromagnetic material in the foregoing embodiment, as shown in fig. 4, the present application provides a model parameter optimization system for the hysteresis effect simulation of the ferromagnetic material, the system comprising:
the simulation circuit construction module 10 is used for establishing an initial hysteresis model of the ferromagnetic material and constructing a simulation circuit through simulation software;
a model parameter updating module 20, wherein the model parameter updating module 20 is used for connecting the simulation circuit with the initial hysteresis model, executing simulation fitting of ferromagnetic materials, and updating model parameters of the initial hysteresis model based on simulation fitting results;
the sample set establishing module 30 is used for establishing a sample characteristic space according to a simulation fitting result, and carrying out hysteresis sample collection of a sample ferromagnetic material based on the sample characteristic space to establish a sample set;
the mapping trust value establishing module 40 is configured to perform sample trust analysis on the sample set, and establish a mapping trust value of each sample in the sample set;
the sample clustering result generating module 50 is used for performing sample clustering on the sample set to generate a sample clustering result, wherein the clustering characteristic comprises a control parameter characteristic, a hysteresis result characteristic and a trust value characteristic of a sample;
and the hysteresis model construction module 60 is used for carrying out self-adaptive optimizing updating of model parameters by using a clustering center of the sample clustering result, completing model parameter optimization of an initial hysteresis model according to the self-adaptive optimizing updating result, and completing the construction of the hysteresis model.
Further, the system further comprises:
the demand data generation module is used for determining a sample extremum based on the sample feature space and generating demand data, wherein the demand data is the model precision demand for building a hysteresis model;
the step length constraint configuration module is used for configuring step length constraint according to the sample extreme value and the requirement data, and distributing authentication nodes according to the sample extreme value and the step length constraint;
the constraint optimization module is used for setting a node parameter optimization limit value at the authentication node and carrying out optimization constraint of the model parameter optimization through the parameter optimization limit value;
and the model construction module is used for completing the construction of the hysteresis model according to the optimization constraint result.
Further, the system further comprises:
the association matching execution module is used for calling the clustering center through the authentication node and executing association matching;
the breaking value generation module is used for carrying out authentication node breaking analysis corresponding to the clustering center by the association matching result based on the node parameter optimization limit value to generate a breaking value;
the parameter optimization compensation generation module is used for carrying out radiation analysis on the adjacent area through the burst value and generating parameter optimization compensation based on a radiation analysis result;
and the compensation updating module is used for executing the self-adaptive optimizing updating compensation of the model parameters through the parameter optimizing compensation.
Further, the system further comprises:
the feature segmentation module is used for carrying out feature segmentation on the control parameters and determining real control parameters and other control parameters, wherein the real control parameters are simulation fit control parameters;
the splitting set generation module is used for carrying out data splitting on the sample set according to the real control parameters to generate an initial data splitting set;
and the sample clustering result generation module is used for carrying out cluster center search in each initial data splitting set, carrying out sample clustering according to the cluster center search result and generating the sample clustering result.
Further, the system further comprises:
the additional feature calculation result generation module is used for calculating parameter fluctuation values of the other control parameters, carrying out additional feature calculation according to parameter fluctuation value calculation results and the trust value features through preset distribution weights, and generating additional feature calculation results, wherein the additional feature calculation results have a mapping relation with the hysteresis result features;
and the clustering and central search result determining module is used for carrying out feature transparent weakening of hysteresis result features according to the additional feature calculation results and determining a clustering central search result according to the feature transparent weakening results.
Further, the system further comprises:
the searching and clustering execution module is used for executing sample searching and clustering by taking the searching result of the clustering center as a datum point;
the clustering result acquisition module is used for taking the characteristic distance of the hysteresis result characteristic as a length constraint, and acquiring the sample clustering result according to the length constraint result, wherein the sample clustering result is provided with a cluster stable value mark.
Further, the system further comprises:
the optimization unit establishment module is used for establishing an adaptive optimization unit and fitting the adaptive optimization unit to the hysteresis model;
the data comparison module is used for executing data comparison of the newly added data and the sample set when the self-adaptive optimization unit receives the newly added data;
the space updating module is used for temporarily storing the newly added data to an updating space when the data comparison result meets a preset threshold value;
and the parameter reconstruction module is used for executing sample replacement of the update space and the sample set if the data quantity in the update space meets a preset data threshold value, and completing the parameter reconstruction of the hysteresis model.
The foregoing detailed description of a model parameter optimization method for magnetic hysteresis simulation of ferromagnetic material will be clear to those skilled in the art, and the description of a model parameter optimization system for magnetic hysteresis simulation of ferromagnetic material in this embodiment is relatively simple for the system disclosed in the embodiments, and the relevant points refer to the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. A method for model parameter optimization of hysteresis effect simulation of ferromagnetic material, the method comprising:
establishing an initial hysteresis model of the ferromagnetic material, and constructing a simulation circuit through simulation software;
connecting the simulation circuit with the initial hysteresis model, executing simulation fitting of the ferromagnetic material, and updating model parameters of the initial hysteresis model based on simulation fitting results;
establishing a sample feature space according to a simulation fitting result, collecting hysteresis samples of sample ferromagnetic materials based on the sample feature space, and establishing a sample set;
sample trust analysis is carried out on the sample set, and a mapping trust value of each sample in the sample set is established;
sample clustering is carried out on the sample set, a sample clustering result is generated, and clustering characteristics comprise control parameter characteristics, hysteresis result characteristics and trust value characteristics of samples;
and carrying out self-adaptive optimizing updating of model parameters by using a clustering center of the sample clustering result, and completing model parameter optimization of an initial hysteresis model according to the self-adaptive optimizing updating result to complete construction of the hysteresis model.
2. The method of claim 1, wherein the method further comprises:
determining a sample extremum based on the sample feature space, and generating demand data, wherein the demand data is a model precision demand for building a hysteresis model;
configuring step length constraint according to the sample extremum and the requirement data, and distributing authentication nodes according to the sample extremum and the step length constraint;
setting a node parameter optimization limit value at the authentication node, and carrying out optimization constraint of the model parameter optimization through the parameter optimization limit value;
and completing the construction of the hysteresis model according to the optimization constraint result.
3. The method of claim 2, wherein the method further comprises:
invoking the clustering center through the authentication node, and executing association matching;
performing authentication node bursting analysis of the clustering center corresponding to the association matching result based on the node parameter optimization limit value to generate a bursting value;
performing radiation analysis of the adjacent area through the burst value, and generating parameter optimization compensation based on a radiation analysis result;
and executing the self-adaptive optimizing updating compensation of the model parameters through the parameter optimizing compensation.
4. The method of claim 1, wherein the method further comprises:
dividing the control parameter characteristics, and determining real control parameters and other control parameters, wherein the real control parameters are simulation fit control parameters;
performing data splitting of the sample set by the real control parameters to generate an initial data splitting set;
and carrying out cluster center search in each initial data splitting set, carrying out sample clustering according to the cluster center search result, and generating the sample clustering result.
5. The method of claim 4, wherein the method further comprises:
calculating parameter fluctuation values of the other control parameters, and carrying out additional feature calculation according to parameter fluctuation value calculation results and the trust value features through preset distribution weights to generate additional feature calculation results, wherein the additional feature calculation results have a mapping relation with the hysteresis result features;
and carrying out feature transparency weakening of hysteresis result features according to the additional feature calculation results, and determining a cluster center search result according to the feature transparency weakening results.
6. The method of claim 5, wherein the method further comprises:
performing sample searching and clustering by taking the searching result of the clustering center as a datum point;
and taking the characteristic distance of the hysteresis result characteristic as a length constraint, and obtaining the sample clustering result according to the length constraint result, wherein the sample clustering result is provided with a cluster stable value mark.
7. The method of claim 1, wherein the method further comprises:
establishing an adaptive optimization unit, and fitting the adaptive optimization unit to the hysteresis model;
when the self-adaptive optimization unit receives newly added data, data comparison between the newly added data and the sample set is executed;
when the data comparison result meets a preset threshold value, temporarily storing the newly added data into an update space;
and if the data quantity in the update space meets a preset data threshold, executing sample replacement of the update space and the sample set, and completing the parameter reconstruction of the hysteresis model.
8. A model parameter optimization system for a hysteresis effect simulation of a ferromagnetic material, the system comprising:
the simulation circuit construction module is used for establishing an initial hysteresis model of the ferromagnetic material and constructing a simulation circuit through simulation software;
the model parameter updating module is used for connecting the simulation circuit with the initial hysteresis model, executing simulation fitting of the ferromagnetic material and updating model parameters of the initial hysteresis model based on a simulation fitting result;
the sample set establishing module is used for establishing a sample feature space according to a simulation fitting result, collecting hysteresis samples of sample ferromagnetic materials based on the sample feature space and establishing a sample set;
the mapping trust value establishing module is used for carrying out sample trust analysis on the sample set and establishing a mapping trust value of each sample in the sample set;
the sample clustering result generation module is used for carrying out sample clustering on the sample set to generate a sample clustering result, and the clustering characteristics comprise control parameter characteristics, hysteresis result characteristics and trust value characteristics of the sample;
and the hysteresis model construction module is used for carrying out self-adaptive optimizing updating of model parameters by using a clustering center of the sample clustering result, completing model parameter optimization of an initial hysteresis model according to the self-adaptive optimizing updating result, and completing the construction of the hysteresis model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311498049.XA CN117272683B (en) | 2023-11-13 | 2023-11-13 | Model parameter optimization method and system for magnetic hysteresis effect simulation of ferromagnetic material |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311498049.XA CN117272683B (en) | 2023-11-13 | 2023-11-13 | Model parameter optimization method and system for magnetic hysteresis effect simulation of ferromagnetic material |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117272683A CN117272683A (en) | 2023-12-22 |
| CN117272683B true CN117272683B (en) | 2024-01-26 |
Family
ID=89206622
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311498049.XA Active CN117272683B (en) | 2023-11-13 | 2023-11-13 | Model parameter optimization method and system for magnetic hysteresis effect simulation of ferromagnetic material |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117272683B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107103163A (en) * | 2017-05-26 | 2017-08-29 | 华中科技大学 | A kind of method that current transformer J A simulation models are determined based on Physical Experiment |
| CN107515992A (en) * | 2017-09-06 | 2017-12-26 | 武汉大学 | A kind of JA Model Parameter Optimization self-adapted genetic algorithms in ferromagnetic material hysteresis characteristic emulation |
| CN109446681A (en) * | 2018-11-02 | 2019-03-08 | 积成电子股份有限公司 | The optimization implementation method of analogue transformer magnetic hysteresis based on J-A model |
| CN110188480A (en) * | 2019-05-31 | 2019-08-30 | 华北电力大学（保定） | The hysteresis characteristic sunykatuib analysis system and method for ferromagnetic material under a kind of Direct Current Bias |
| CN111523236A (en) * | 2020-04-24 | 2020-08-11 | 哈尔滨工业大学 | Piezoelectric ceramic hysteresis model linearization identification method based on Koopman operator |
| CN112364490A (en) * | 2020-10-28 | 2021-02-12 | 许继集团有限公司 | Model parameter identification method for hysteresis characteristics of ferromagnetic material |
-
2023
- 2023-11-13 CN CN202311498049.XA patent/CN117272683B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107103163A (en) * | 2017-05-26 | 2017-08-29 | 华中科技大学 | A kind of method that current transformer J A simulation models are determined based on Physical Experiment |
| CN107515992A (en) * | 2017-09-06 | 2017-12-26 | 武汉大学 | A kind of JA Model Parameter Optimization self-adapted genetic algorithms in ferromagnetic material hysteresis characteristic emulation |
| CN109446681A (en) * | 2018-11-02 | 2019-03-08 | 积成电子股份有限公司 | The optimization implementation method of analogue transformer magnetic hysteresis based on J-A model |
| CN110188480A (en) * | 2019-05-31 | 2019-08-30 | 华北电力大学（保定） | The hysteresis characteristic sunykatuib analysis system and method for ferromagnetic material under a kind of Direct Current Bias |
| CN111523236A (en) * | 2020-04-24 | 2020-08-11 | 哈尔滨工业大学 | Piezoelectric ceramic hysteresis model linearization identification method based on Koopman operator |
| CN112364490A (en) * | 2020-10-28 | 2021-02-12 | 许继集团有限公司 | Model parameter identification method for hysteresis characteristics of ferromagnetic material |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117272683A (en) | 2023-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104899457B (en) | Based on the satellite data Forecasting Methodology for improving unbiased GM (1,1) model | |
| CN111046930A (en) | Power supply service satisfaction influence factor identification method based on decision tree algorithm | |
| CN113377964B (en) | Knowledge graph link prediction method, device, equipment and storage medium | |
| CN117668622B (en) | Training method of equipment fault diagnosis model, fault diagnosis method and device | |
| CN115496144A (en) | Power distribution network operation scene determining method and device, computer equipment and storage medium | |
| CN114005015A (en) | Model training method, electronic device, and computer-readable storage medium | |
| CN117272683B (en) | Model parameter optimization method and system for magnetic hysteresis effect simulation of ferromagnetic material | |
| CN111222669A (en) | Overhead line cost prediction method and device | |
| CN112016598B (en) | SDAE-based partial discharge pattern recognition method, computer device and storage medium | |
| CN114118370A (en) | Model training method, electronic device, and computer-readable storage medium | |
| CN113656707A (en) | Financing product recommendation method, system, storage medium and equipment | |
| CN113408622A (en) | Non-invasive load identification method and system considering characteristic quantity information expression difference | |
| CN113095501A (en) | Deep reinforcement learning-based unbalanced classification decision tree generation method | |
| CN117170980A (en) | Early warning method, device, equipment and storage medium for server hardware abnormality | |
| CN110825583A (en) | Energy efficiency qualitative assessment technology for multi-index fusion of cloud data center | |
| CN116880867A (en) | Policy big model-based decision engine updating method and device | |
| CN113132482B (en) | Distributed message system parameter adaptive optimization method based on reinforcement learning | |
| WO2022162839A1 (en) | Learning device, learning method, and recording medium | |
| CN108596781A (en) | Data mining and prediction integration method for power system | |
| Gavrilov et al. | Time series prediction using the adaptive resonance theory algorithm ART-2 | |
| CN114429172A (en) | Load clustering method, device, equipment and medium based on transformer substation user constitution | |
| CN111026661A (en) | Method and system for comprehensively testing usability of software | |
| CN115473232B (en) | Load parameter adjusting method, system, equipment and storage medium | |
| Pal et al. | A Robust Flight Control Design of an Unmanned Aerial Vehicle Using Harmony Search | |
| CN113283174B (en) | Reservoir productivity prediction method, system and terminal based on algorithm integration and self-control |
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 |