Data intelligent visualization method based on digital twins

By analyzing the correlation between multi-dimensional data and text information, screening out key data dimensions and visualizing them in combination with display requirements parameters, the visualization problem of large amount of digital twin data is solved, and efficient information display and analysis is achieved.

CN118657001BActive Publication Date: 2024-11-05JIANGSU ZEQINCHUANG CULTURE TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411146705.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-21
Publication Date
2024-11-05
Estimated Expiration
2044-08-21

AI Technical Summary

Technical Problem

Large-scale digital twin data volume makes visualization of data twin models difficult, increasing the difficulty and understanding of users' data analysis.

Method used

By obtaining multi-dimensional data and text information, analyzing the correlation and information density between data dimensions, filtering out the initial data sequence to be displayed, and visualizing data with display requirements parameters, reducing redundant information, improving information readability and ability Explanatory.

Benefits of technology

It realizes intelligent visualization that meets the needs of different audience groups, reduces the burden of users in processing and understanding data, improves analysis efficiency and utilization rate of information output, and obtains a more ideal visual display effect.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118657001B_ABST
    Figure CN118657001B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of electronic digital data pattern analysis and processing, and specifically to a data intelligent visualization method based on digital twins. The present invention analyzes the visualization needs of different audience groups, and analyzes the data correlation and text information correlation between data dimensions, combined with information content characteristics, to preliminarily screen the data dimensions; then obtains visualization parameters, combines the display demand parameters of the audience group, intelligently determines the display data dimensions, reduces the input of prior content, increases the robustness of the visualization system, and the process of screening data of each dimension takes into account the complexity of information acceptance of different audience groups, improves the utilization rate of data output, and obtains a more ideal visualization display effect.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of electrical digital data pattern analysis and processing, and in particular to a data intelligent visualization method based on digital twins. Background Art

[0002] Digital twin technology is a technology that dynamically maps and simulates physical entities in digital space. It combines physical systems and virtual simulation models to create a highly fidelity digital copy through real-time data, sensors, and software algorithms. This digital model can not only replicate the behavior and characteristics of its physical counterpart, but also simulate its possible responses and evolution under various conditions, thereby achieving real-time monitoring, prediction, and optimization of physical entities.

[0003] However, for most digital twin processes, in order to improve the accuracy of model building, a large amount of data is collected and analyzed in the real world, and the amount of data generated during simulation experiments is also large. In the process of visualizing the digital twin model, excessive data volume may make it more difficult for users to capture the key points of analysis when analyzing the data, and increase the cost of understanding. Summary of the invention

[0004] In order to solve the technical problem that the data volume of the data twin model is too large, which is not conducive to the visualization of the data twin model, the purpose of the present invention is to provide a data intelligent visualization method based on digital twins. The technical solution adopted is as follows:

[0005] Obtaining data to be visualized of the digital twin system; the data to be visualized includes at least three data dimensions and text information; obtaining display demand parameters of the current display audience group;

[0006] According to the correlation between different data dimensions, the information summary parameters of each data dimension are obtained; according to the information summary parameters and information content of each data dimension, combined with the data fluctuation characteristics of each data dimension, the information density of each data dimension is obtained;

[0007] According to the information density of each data dimension, an initial data sequence to be displayed is screened out; according to the correlation of the text information corresponding to each data dimension in the initial data sequence to be displayed, combined with the information density, a visualization parameter of each data dimension in the initial data sequence to be displayed is obtained;

[0008] Data visualization is performed according to the visualization parameter of each data dimension in the initial sequence to be displayed in combination with the display requirement parameter.

[0009] Furthermore, the method for obtaining the information summary parameter includes:

[0010] Obtain data correlation coefficients between different data dimensions;

[0011] According to the data correlation coefficient between each data dimension and all other data dimensions, an information summary parameter of each data dimension is obtained; the information summary parameter is positively correlated with the data correlation coefficient.

[0012] Furthermore, the method for obtaining the information density includes:

[0013] Obtain a data fitting curve for each data dimension; obtain the information density of each data dimension according to the fluctuation severity of each data fitting curve, combined with the information summary parameter and the information content; the fluctuation severity of the data fitting curve, the information summary parameter and the information content are all positively correlated with the information density.

[0014] Furthermore, the method for obtaining the intensity of the fluctuation includes:

[0015] Obtaining the ratio of the number of extreme value points of each data fitting curve to the number of all data points on the corresponding data fitting curve as a first fluctuation sub-parameter;

[0016] According to the deviation characteristics of the slope of each data point on the data fitting curve, combined with the change rate of the data point, a second fluctuation sub-parameter is obtained; the deviation characteristics of the slope of the data point and the change rate of the data point are both positively correlated with the second fluctuation sub-parameter;

[0017] The first fluctuation sub-parameter and the second fluctuation sub-parameter are integrated to obtain the fluctuation severity of each of the data fitting curves.

[0018] Furthermore, the method for obtaining the deviation characteristic of the slope of the data point includes:

[0019] The deviation characteristic of the slope of each data point is obtained according to the difference between the slope of each data point and the average value of the slopes of all data points on the corresponding data fitting curve.

[0020] Furthermore, the method for obtaining the initial data sequence to be displayed includes:

[0021] Arrange the data in descending order according to the information density of each data dimension to obtain a first sequence;

[0022] Divide the first sequence into a preset number of parts, and use the divided sequence containing the minimum information density as the second sequence; in the second sequence, obtain the critical information density according to the difference of the information density between adjacent elements;

[0023] The data dimensions corresponding to the information density being less than the critical information density are removed from the first sequence, and the data corresponding to the remaining data dimensions are used as the initial data sequence to be displayed.

[0024] Furthermore, the method for obtaining the critical information density includes:

[0025] The maximum information density of the two elements corresponding to the point where the difference in the information density between adjacent elements in the second sequence is the maximum is taken as the critical information density.

[0026] Furthermore, the method for obtaining the visualization parameters includes:

[0027] Vectorizing the text information to obtain text correlation coefficients between different text information vectors;

[0028] According to the text correlation coefficient between each text information vector and other text information vectors, the other text information vectors are sorted, and the sorting results of the other text information vectors corresponding to each text information vector are used as the correlation sequence of each text information vector;

[0029] According to the difference between the related sequence corresponding to each text information vector and the related sequences corresponding to other text information vectors, combined with the information density corresponding to each text information vector, the visualization parameter of each data dimension is obtained; the difference in related sequences and the information density are both positively correlated with the visualization parameter.

[0030] Furthermore, the method for obtaining the difference of the related sequences includes:

[0031] The mean square error of the correlation sequences corresponding to the two text information vectors is taken as the difference of the correlation sequences corresponding to the two text information vectors.

[0032] Furthermore, the method for data visualization includes:

[0033] The visualization parameters and the display requirement parameters are both normalized; visualization parameters are selected from large to small and gradually added to the intelligent visualization data sequence until the visualization parameters and values ​​of the elements in the intelligent visualization data sequence are greater than the display requirement parameters of the current audience group for the first time, and the data of the data dimension corresponding to the intelligent visualization data sequence is visualized.

[0034] The present invention has the following beneficial effects:

[0035] The present invention first obtains the data to be visualized of the digital twin system to provide a data basis; further obtains the display demand parameters of the current display audience group to meet the needs of different audience groups, so that subsequent intelligent visualization can adapt to different implementation scenarios and be more intelligent; further obtains the information summary parameters of each data dimension, measures the information summary ability of each data dimension, and provides a basis for subsequent information density calculation; further obtains the information density of each data dimension to facilitate subsequent intelligent visualization of data, so that users can capture key information more quickly, thereby reducing the burden of users processing and understanding data, and meeting the visualization needs of different audience groups; further, according to the information density of each data dimension, the initial data sequence to be displayed is screened out to avoid displaying redundant and irrelevant information, improve analysis efficiency, and help users better understand and analyze data, and improve the readability and interpretability of information; further, by analyzing the correlation of text information, data dimensions with similar or complementary contents can be identified, the correlation of data dimensions is supplemented, and the visualization parameters of each data dimension in the initial sequence to be displayed are obtained to prepare for subsequent data intelligent visualization; finally, according to the visualization parameters of each data dimension in the initial sequence to be displayed, combined with the display demand parameters, data visualization is performed. The present invention analyzes the visualization needs of different audience groups, and analyzes the data correlation and text information correlation between data dimensions, combined with information content characteristics, to perform preliminary screening of data dimensions; then obtains visualization parameters, combines the display demand parameters of the audience groups, and intelligently determines the display data dimensions, thereby reducing the input of prior content and increasing the robustness of the visualization system. The process of screening data of each dimension takes into account the complexity of information acceptance of different audience groups, improves data output utilization, and obtains a more ideal visualization display effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 A flowchart of a data intelligent visualization method based on digital twins provided by an embodiment of the present invention;

[0038] Figure 2 A schematic diagram of a training method for a neural network model for obtaining display demand parameters provided by an embodiment of the present invention;

[0039] Figure 3 A flow chart of a method for obtaining fluctuation severity provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0040] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation, structure, features and effects of a data intelligent visualization method based on digital twins proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0041] 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.

[0042] The following is a detailed description of a specific scheme of a data intelligent visualization method based on digital twins provided by the present invention in conjunction with the accompanying drawings.

[0043] The present invention analyzes the visualization needs of the audience group, and analyzes the data correlation and text information correlation between data dimensions, combined with the information content characteristics, to perform a preliminary screening of data dimensions; then obtains visualization parameters, combines the display demand parameters of the audience group, and intelligently determines the display data dimensions, thereby reducing the input of prior content and increasing the robustness of the visualization system. The process of screening data of each dimension takes into account the complexity of information acceptance of different audience groups, improves the utilization rate of data output, and obtains a more ideal visualization display effect.

[0044] See also Figure 1 , which shows a flow chart of a data intelligent visualization method based on digital twins provided by an embodiment of the present invention, specifically including:

[0045] Step S1: Obtain the data to be visualized of the digital twin system; the data to be visualized includes at least three data dimensions and text information; obtain the display demand parameters of the current display audience group.

[0046] A digital twin model is a high-precision digital copy of a physical entity or system that integrates sensor data, operational data, and simulation and analysis tools to monitor, predict, and optimize performance in real time. This method is an existing technology and will not be described in detail here.

[0047] The purpose of visualizing the digital twin model is to enable users to understand and analyze the entities, processes, or systems simulated by the digital twin more intuitively and effectively through intuitive visual expressions and interactive interfaces. Therefore, it is necessary to obtain the data to be visualized of the simulation results of the digital twin model, including the historical data of the digital twin system, real-time data streams, and prediction data for predicting the next state based on the current system state.

[0048] In one embodiment of the present invention, the implementation scenario is a drive motor monitoring system of a sandbox interactive control system based on the Internet of Things, and the data to be visualized include at least temperature, current and voltage data monitored by sensors; data such as speed and power of equipment operating performance; health status and degree of wear of equipment status. Different types of data are sorted in chronological order to obtain multiple data sequences; then each data sequence is cleaned, and the data is standardized by the range normalization method, and each processed data sequence is used as a data dimension. In other embodiments of the present invention, the implementer can set the type and quantity of monitoring data of the digital twin model by himself.

[0049] It should be noted that the data for subsequent analysis and processing are all standardized data, while the data for visual display are original data.

[0050] Taking into account that different audiences of the digital twin model have different demands for the level of visualization sophistication, for example, visitors, managers, and engineering and technical personnel have increasing demands for visualization sophistication, we first obtain the display demand parameters of the current display audience group to facilitate subsequent intelligent visualization to adapt to different implementation scenarios and be more intelligent.

[0051] As an example, based on the pre-trained neural network model, obtain the display demand parameters of the current display audience group; see Figure 2 , which shows a schematic diagram of a training method for a neural network model for obtaining display demand parameters provided by an embodiment of the present invention. The specific training process includes:

[0052] S201. Obtain the occupational label of the audience displayed by the current digital twin model.

[0053] S202. Obtain a set of 600 occupational labels, and score each occupational label data manually. The scoring values ​​are: 0.1, 0.2, 0.3...1, a total of 10 score levels, corresponding to 10 types of data display precision.

[0054] Among them, the scoring result is the display demand parameter of the audience group. The higher the parameter, the more precise the data displayed by the digital twin model.

[0055] S203. The classification network structure is a 5-layer fully connected neural network, and the loss function is a cross entropy function.

[0056] S204. Divide the occupation label data set into a training set and a validation set in a ratio of 7:3, input the training set data into the classification network for training, and use the gradient descent method for training until the loss function converges, completing the classification network training.

[0057] S205: Input the occupational labels of the desired display audience of the current visualization platform into the trained neural network, and output the display demand parameter F of the current display audience group.

[0058] As another example, the implementer may assign empirical values ​​to the display demand parameters of different display audience groups based on experience, such as 0.5, 0.8, and 1 for the visitor group, management personnel, and engineering and technical personnel, respectively.

[0059] Step S2: According to the correlation between different data dimensions, obtain the information summary parameters of each data dimension; according to the information summary parameters and information content of each data dimension, combined with the data fluctuation characteristics of each data dimension, obtain the information density of each data dimension.

[0060] Considering that in the multi-dimensional data generated by the digital twin model, the data of different dimensions have different degrees of precision in the visualization process. In order to ensure that the displayed data can meet the reading needs of different audiences, it is necessary to analyze the information characteristics contained in the data of each dimension.

[0061] Taking into account the existence of certain correlations between different data dimensions, the status of different parts of the model system reflects different degrees of feedback: for example, electric power is correlated with current and voltage. Compared with voltage and current parameters, electric power parameters can better reflect the comprehensive operating status of the electrical equipment. For the visualization process, this type of data has a strong comprehensive expression ability for the operating status of a certain part of the system, and is more conducive to providing key information to the audience. Therefore, according to the correlation between different data dimensions, the information summary parameters of each data dimension are obtained, and the information summary ability of each data dimension is measured, which provides a basis for the subsequent information density calculation.

[0062] Preferably, in one embodiment of the present invention, considering that the stronger the data correlation between the current data dimension and all other data dimensions, the larger the correlation coefficient, the more the current data dimension can reflect the data characteristics of other data dimensions, the stronger the information summary capability, the larger the information summary parameter, based on this, the data correlation coefficient between different data dimensions is obtained; the data correlation coefficient is positively correlated with the correlation between different data dimensions;

[0063] According to the data correlation coefficient between each data dimension and all other data dimensions, the information summary parameter of each data dimension is obtained; the information summary parameter is positively correlated with the data correlation coefficient.

[0064] As an example, the Pearson correlation coefficient between different data dimensions is obtained, the absolute value of the Pearson correlation coefficient is taken as the data correlation coefficient, and the sum of the data correlation coefficients of each data dimension and all other data dimensions is taken as the information summary parameter.

[0065] As another example, implementers can also perform curve fitting on the data of each data dimension by analyzing the absolute value of the difference in the time of occurrence of extreme points of different fitting curves. The smaller the absolute value of the difference, the more the different fitting curves show the characteristics of simultaneous occurrence of extreme points, reflecting that the change correlation between the fitting curves is stronger and the similarity of the data dimensions is stronger, thereby obtaining the correlation coefficient and the information summary parameters.

[0066] It should be noted that in other embodiments of the present invention, the implementer may also obtain information summary parameters based on other methods such as chi-square statistics or mutual information, which and the Pearson correlation coefficient are technical means well known to those skilled in the art and will not be described in detail here.

[0067] Considering that in the data display stage of the digital twin model, the audience usually tends to pay attention to the constantly changing data in a certain period of time, and the fluctuating data can maintain a more three-dimensional digital twin model, the data fluctuation characteristics of the data dimension can reflect the information density characteristics; at the same time, the information summary parameter measures the information summary ability of each data dimension, and the information content reflects the information volume of the data dimension. Therefore, according to the information summary parameters and information content of each data dimension, combined with the data fluctuation characteristics of each data dimension, the information density of each data dimension is obtained, which is convenient for subsequent intelligent visualization of data, enabling users to capture key information faster, thereby reducing the burden on users to process and understand data, and meeting the visualization needs of different audiences.

[0068] Preferably, in one embodiment of the present invention, considering that the fitting curve can more intuitively observe the fluctuation changes of the data and is more conducive to extracting the main trends and change characteristics of the data, a data fitting curve for each data dimension is obtained; considering that the intensity of fluctuation reflects the dynamic changes of the data, the greater the intensity of fluctuation, the more frequent the data changes, the greater the fluctuation amplitude, the more information contained, and the greater the information density; the information summary parameter represents the information summary ability of the data dimension, and the larger the information summary parameter, the stronger the information summary ability, the more information contained, and the greater the information density; the information content represents the complexity and uncertainty of the data, and the greater the information content, the more information contained, and the greater the information density; therefore, according to the intensity of fluctuation of each data fitting curve, combined with the information summary parameter and the information content, the information density of each data dimension is obtained.

[0069] As an example, the data fitting curve is obtained by the least square method; the information entropy is used as the information content; the product of the fluctuation intensity of each data fitting curve, the information summary parameter and the information content is used as the information density of each data dimension. Among them, the least square method and information entropy are technical means well known to those skilled in the art. In other embodiments of the present invention, other methods such as polynomial fitting can also be used to obtain the data fitting curve, which will not be described in detail here.

[0070] As another example, the information density can also be obtained by weighted summation. For example, the fluctuation intensity is 0.8, the information summary parameter is 0.7, the information content is 0.9, the weights are 0.3, 0.4, and 0.3 respectively, and the information density is 0.79. The specific weights can be adjusted by the implementer.

[0071] Preferably, in one embodiment of the present invention, the method for obtaining the severity of fluctuation includes:

[0072] See also Figure 3 , which shows a flow chart of a method for obtaining the severity of fluctuations provided by an embodiment of the present invention, specifically comprising:

[0073] Step S301: obtaining the ratio of the number of extreme value points of each data fitting curve to the number of all data points on the corresponding data fitting curve as the first fluctuation sub-parameter.

[0074] Considering that the more extreme points of the data fitting curve, the greater the proportion of the data they occupy, the richer the dynamic changes of the current dimensional data and the greater the degree of fluctuation, the ratio of the number of extreme points of each data fitting curve to the number of all data points on the corresponding data fitting curve is obtained as the first fluctuation sub-parameter.

[0075] Step S302: Obtain a second fluctuation sub-parameter according to the deviation characteristics of the slope of each data point on the data fitting curve and the change rate of the data point.

[0076] Among them, the deviation characteristics of the slope of the data point and the change rate of the data point are both positively correlated with the second wave sub-parameter.

[0077] Considering that the greater the deviation of the slope of each data point on the data fitting curve, the more obvious the specificity of the current data point in the overall data change process, the more data fluctuation patterns there are and the greater the fluctuation intensity; the greater the change rate of the data point, the more obvious the reflection of the dynamic change of the data and the greater the fluctuation intensity, so according to the deviation characteristics of the slope of each data point on the data fitting curve, combined with the change rate of the data point, the second fluctuation sub-parameter is obtained.

[0078] As an example, the absolute value of the slope of a data point represents the rate of change of the data point, the absolute value of the difference between the slope of each data point on the data fitting curve and the average value of the slope represents the deviation characteristic of the slope, and the average value of the product of the rate of change of all data points and the deviation characteristic is taken as the second volatility parameter.

[0079] Step S303: The first fluctuation sub-parameter and the second fluctuation sub-parameter are integrated to obtain the fluctuation severity of each data fitting curve.

[0080] As an example, the product of the first fluctuation sub-parameter and the second fluctuation sub-parameter is used as the fluctuation severity.

[0081] As another example, after the first fluctuation sub-parameter and the second fluctuation sub-parameter are normalized respectively, a weighted sum method is used to obtain the fluctuation severity, for example, the first fluctuation sub-parameter is 0.4, and the second fluctuation sub-parameter is 0.6.

[0082] Step S3: Filter out the initial data sequence to be displayed according to the information density of each data dimension; obtain the visualization parameter of each data dimension in the initial data sequence to be displayed according to the relevance of the text information corresponding to each data dimension in the initial data sequence to be displayed and the information density.

[0083] After analyzing the information density characteristics of the data in each dimension of the digital twin model, you can preliminarily screen the data dimensions to avoid displaying redundant and irrelevant information, improve analysis efficiency, and help users better understand and analyze the data, and improve the readability and interpretability of information.

[0084] Preferably, in one embodiment of the present invention, considering that the greater the information density, the more meaningful the data dimension is to display, the first sequence is obtained by arranging the data dimensions from large to small according to the information density; considering that the smaller the information density, the less information can be provided for visualization, and the more it can be ignored; based on this, the first sequence is divided into a preset number of parts, and the sequence containing the minimum information density after the division is taken as the second sequence; and in order to avoid the problem of incomplete visualization of the twin model system, considering that the second sequence is the sequence with the smallest overall information density and has the smallest contribution to visualization, the critical information density is obtained in the second sequence according to the difference in information density between adjacent elements;

[0085] The data dimensions corresponding to the information density less than the critical information density are removed from the first sequence, and the data corresponding to the remaining data dimensions are used as the initial data sequence to be displayed, thereby reducing the impact of the data dimensions with lower information density.

[0086] As an example, the preset number is 2, the first sequence is divided into two equal parts, and the sequence containing the minimum information density is used as the second sequence; when the first sequence is an odd number, the middle value is divided into the second sequence, and the second sequence contains the most elements. For example, if the first sequence contains 9 elements, the second sequence contains 5 elements; for example, when the preset number is 3 and the first sequence is 11, the division result is 3, 3, 5, and the second sequence contains 5 elements.

[0087] Preferably, in one embodiment of the present invention, considering that the point where the difference in information density between adjacent elements is the largest, it indicates that the information density of adjacent elements has been greatly reduced, marking an obvious dividing point. The information density of the data dimension before this dividing point is higher, and that after this point is lower. Therefore, the maximum information density of the two elements corresponding to the point where the difference in information density between adjacent elements in the second sequence is the largest is taken as the critical information density.

[0088] As an example, the difference in information density is represented by the absolute value of the difference in information density of adjacent elements, and the maximum information density of the two elements corresponding to the maximum absolute value of the difference is selected as the critical information density.

[0089] Considering that the relevance of text information is used to evaluate the semantic relationship between data dimensions, by analyzing the relevance of text information, we can identify data dimensions with similar or complementary contents, supplement the relevance of data dimensions, and thus better evaluate the visualization parameters of each data dimension. Therefore, according to the relevance of text information corresponding to each data dimension in the initial data sequence to be displayed, combined with information density, the visualization parameters of each data dimension in the initial data sequence to be displayed are obtained to prepare for subsequent intelligent data visualization.

[0090] Preferably, in one embodiment of the present invention, considering that the correlation of text information cannot be directly analyzed, the text information is vectorized to obtain the text correlation coefficient between different text information vectors; the text correlation coefficient is positively correlated with the correlation between different text information vectors;

[0091] Taking into account the text correlation coefficient between the current text information vector and other text information vectors, which represents the correlation between the current text information vector and other text information vectors, the other text information vectors of the current text information vector are sorted by the text correlation coefficient, and the sorting results of other text information vectors corresponding to each text information vector are used as the correlation sequence of each text information vector. The correlation sequence represents the correlation features in the entire data set. The greater the ranking difference of the correlation sequences between the text information vectors, the greater the difference in the correlation features of the text information vectors, the stronger the independence and complementarity, and the smaller the probability that the corresponding two-dimensional data represent the same part of the system. Therefore, according to the difference between the correlation sequence corresponding to each text information vector and the correlation sequence corresponding to other text information vectors, combined with the information density corresponding to each text information vector, the visualization parameters of each data dimension are obtained; the difference in correlation sequences and the information density are both positively correlated with the visualization parameters.

[0092] As an example, the differences between different correlation sequences are analyzed through mean square error. The larger the mean square error, the greater the difference in the ranking of the correlation sequences between the two text information vectors, and the smaller the probability that the corresponding two dimensional data represent the same part of the system. The larger the mean square error between the current text information vector and all other text information vectors, the higher the priority of retaining and displaying the data dimension corresponding to the current text information vector, and the larger the visualization parameter. At the same time, the greater the information density, the more information the data dimension contains, which makes it easier for users to capture key information faster. The higher the display priority, the larger the visualization parameter. The average value of the mean square error of the correlation sequence corresponding to each text information vector and the correlation sequence corresponding to other text information vectors is multiplied by the information density corresponding to each text information vector as the visualization parameter of each data dimension.

[0093] It should be noted that, in one embodiment of the present invention, the method used to vectorize text information is the word2vec method, and the text correlation coefficient is obtained by the absolute value of the Pearson correlation coefficient between two text information vectors. In other embodiments of the present invention, other text vectorization methods such as the bag-of-words model may also be used. These are technical means well known to those skilled in the art and will not be elaborated here.

[0094] Step S4: Perform data visualization according to the visualization parameters of each data dimension in the initial sequence to be displayed and in combination with the display requirement parameters.

[0095] The display requirement parameters represent the current needs and expectations of intelligent visualization, and the visualization parameters indicate the level of visualization detail or quality that each data dimension can provide. Therefore, data visualization is ultimately performed based on the visualization parameters of each data dimension in the initial sequence to be displayed, combined with the display requirement parameters.

[0096] Preferably, in one embodiment of the present invention, both visualization parameters and display requirement parameters are normalized; visualization parameters are selected from large to small and gradually added to the intelligent visualization data sequence until the visualization parameters and values ​​of the elements in the intelligent visualization data sequence are greater than the display requirement parameters of the current audience group for the first time, and the data of the data dimensions corresponding to the intelligent visualization data sequence are visualized.

[0097] Before the maximum value of the visualization parameter is added to the intelligent visualization data sequence, the intelligent visualization data sequence is an empty sequence.

[0098] After obtaining the data dimensions that need to be visualized, statistical tools such as charts and graphs can be used to display the data of each dimension in chronological order, and trend lines, time series analysis and other methods can be used to help the audience understand the dynamic changes of the data, and update the visualization chart based on the real-time data generated by the data twin model. The visualization process can be implemented through Python or MATLAB / Simulink and its visualization library, which is a technical means well known to those skilled in the art and will not be described in detail here.

[0099] In summary, the present invention aims at the technical problem that the amount of data in the data twin model is too large, which is not conducive to the visualization of the data twin model, and proposes a data intelligent visualization method based on digital twins. The present invention first obtains the data to be visualized of the digital twin system; obtains the display demand parameters of different display audience groups; further obtains the information density of each data dimension according to the correlation between different data dimensions and the information content of each data dimension, combined with the data fluctuation characteristics of each data dimension; further obtains the visualization parameters according to the information density of each data dimension and the correlation of the text information corresponding to each data dimension, combined with the information density; finally, according to the visualization parameters of each data dimension, combined with the display demand parameters, data visualization is performed. The present invention analyzes the visualization needs of different audience groups, and analyzes the data correlation and text information correlation between data dimensions, combined with the information content characteristics, to preliminarily screen the data dimensions; then obtains the visualization parameters, combined with the display demand parameters of the audience group, intelligently determines the display data dimensions, reduces the input of prior content, increases the robustness of the visualization system, and the process of screening the data of each dimension takes into account the complexity of information acceptance of different audience groups, improves the utilization rate of data output, and obtains a more ideal visualization display effect.

[0100] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0101] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A data intelligent visualization method based on digital twins, characterized in that: The method comprises: Obtaining data to be visualized of the digital twin system; the data to be visualized includes at least three data dimensions and text information; obtaining display demand parameters of the current display audience group; According to the correlation between different data dimensions, the information summary parameters of each data dimension are obtained; according to the information summary parameters and information content of each data dimension, combined with the data fluctuation characteristics of each data dimension, the information density of each data dimension is obtained; According to the information density of each data dimension, an initial data sequence to be displayed is screened out; according to the correlation of text information corresponding to each data dimension in the initial data sequence to be displayed, combined with the information density, a visualization parameter of each data dimension in the initial data sequence to be displayed is obtained; Performing data visualization according to the visualization parameter of each data dimension in the initial sequence to be displayed in combination with the display requirement parameter; The method for obtaining the information summary parameter includes: Obtain data correlation coefficients between different data dimensions; According to the data correlation coefficient between each data dimension and all other data dimensions, an information summary parameter of each data dimension is obtained; the information summary parameter is positively correlated with the data correlation coefficient; The method for obtaining the information density includes: Obtaining a data fitting curve for each data dimension; obtaining the information density of each data dimension according to the fluctuation intensity of each data fitting curve, in combination with the information summary parameter and the information content; the fluctuation intensity of the data fitting curve, the information summary parameter and the information content are all positively correlated with the information density; The method for obtaining the intensity of the fluctuation includes: Obtaining the ratio of the number of extreme value points of each data fitting curve to the number of all data points on the corresponding data fitting curve as a first fluctuation sub-parameter; According to the deviation characteristics of the slope of each data point on the data fitting curve, combined with the change rate of the data point, a second fluctuation sub-parameter is obtained; the deviation characteristics of the slope of the data point and the change rate of the data point are both positively correlated with the second fluctuation sub-parameter; The first fluctuation sub-parameter and the second fluctuation sub-parameter are integrated to obtain the fluctuation severity of each of the data fitting curves; The method for obtaining the deviation characteristic of the slope of the data point includes: Obtaining the deviation characteristic of the slope of each data point according to the difference between the slope of each data point and the average value of the slopes of all data points on the corresponding data fitting curve; The method for obtaining the initial data sequence to be displayed includes: Arrange the data in descending order according to the information density of each data dimension to obtain a first sequence; Divide the first sequence into a preset number of parts, and use the divided sequence containing the minimum information density as the second sequence; in the second sequence, obtain the critical information density according to the difference of the information density between adjacent elements; The data dimensions corresponding to the information density being less than the critical information density are removed from the first sequence, and the data corresponding to the remaining data dimensions are used as the initial data sequence to be displayed; The method for obtaining the visualization parameters includes: Vectorizing the text information to obtain text correlation coefficients between different text information vectors; According to the text correlation coefficient between each text information vector and other text information vectors, the other text information vectors are sorted, and the sorting results of the other text information vectors corresponding to each text information vector are used as the correlation sequence of each text information vector; Obtaining a visualization parameter for each data dimension according to the difference between the correlation sequence corresponding to each text information vector and the correlation sequence corresponding to other text information vectors, combined with the information density corresponding to each text information vector; the difference in correlation sequences and the information density are both positively correlated with the visualization parameter; The method for data visualization comprises: The visualization parameters and the display requirement parameters are both normalized; visualization parameters are selected from large to small and gradually added to the intelligent visualization data sequence until the visualization parameters and values ​​of the elements in the intelligent visualization data sequence are greater than the display requirement parameters of the current audience group for the first time, and the data of the data dimension corresponding to the intelligent visualization data sequence is visualized.

2. According to claim 1, a data intelligent visualization method based on digital twins is characterized in that: The method for obtaining the critical information density includes: The maximum information density of the two elements corresponding to the point where the difference in the information density between adjacent elements in the second sequence is the maximum is taken as the critical information density.

3. The method for intelligent data visualization based on digital twins according to claim 1, characterized in that: The method for obtaining the difference of the related sequences comprises: The mean square error of the correlation sequences corresponding to the two text information vectors is taken as the difference of the correlation sequences corresponding to the two text information vectors.