Nuclear power electronic information multi-source data fusion and visual analysis method

Abstract

The invention discloses a nuclear power electronic information multi-source data fusion and visual analysis method, and relates to the technical field of nuclear power, and the method comprises the steps: S1, data collection; s2, data fusion; s3, feature extraction and selection; s4, performing data analysis and modeling; s5, carrying out visual design; s6, realizing visualization; and S7, result interpretation and application. And S8, performing feedback and optimization. According to the nuclear power electronic information multi-source data fusion and visual analysis method, multi-source data fusion can be easily realized, so that related personnel are helped to better understand and analyze the operation state of a nuclear power system, the accuracy and efficiency of decision making are improved, and complex data can be further promoted to become intuitive and easy to understand in cooperation with application of data visualization. Meanwhile, the nuclear power data visualization configuration software has flexible data access and processing capabilities, can adapt to continuously changing data and business requirements, and further improves the multi-source data fusion and visualization analysis effects.

Description

Technical Field

[0001] The present invention relates to the technical field of nuclear power, and particularly to a multi-source data fusion and visualization analysis method for nuclear power electronic information. Background Art

[0002] Nuclear power electronic information refers to a technology that uses electronic and information technology means to process, transmit, store, and apply information in the nuclear power field. It combines the advantages of electronic technology and information technology, providing strong support for the safety management, equipment monitoring, data analysis, etc. of the nuclear power industry. In the prior art, nuclear power projects often involve multiple professional fields, such as nuclear reaction engineering, thermal-hydraulics, mechanical design, electrical control, etc. A large amount of data is generated in each field, and the correlation relationships between these data are complex, the data sources are extensive, and the formats are diverse. This makes nuclear power electronic information usually involve multi-source data and requires unified management and integration for subsequent data analysis and application.

[0003] However, due to the fact that multi-source data fusion involves a large amount of complex data, it is very difficult to directly understand and analyze these data. Through visualization analysis, these data can be presented in the form of graphs, images, etc., making the characteristics and trends of the data more intuitive and easy to understand, which is particularly important for the nuclear power industry. However, in the actual application process, because the data in the nuclear power industry not only has multi-source diversity, but more importantly, it has high variability, with constantly changing data and business requirements, which makes the existing visualization analysis tools unable to fully meet the needs of the nuclear power industry for data visualization, lacking visualization components for nuclear power industry data, and having poor applicability and scalability. Summary of the Invention

[0004] The purpose of the present invention is to provide a multi-source data fusion and visualization analysis method for nuclear power electronic information to solve the problems raised in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A multi-source data fusion and visualization analysis method for nuclear power electronic information, including the following steps:

[0006] S1. Data collection: Collect electronic information data during the operation of nuclear power from various sources (such as sensors, monitoring systems, databases, etc.), and preprocess the collected data;

[0007] S2. Data fusion: Fuse the preprocessed multi-source data to form a more comprehensive and accurate data set, including data alignment (ensuring the consistency of data from different sources in terms of time and space), data association (identifying the correlation relationships between different data sources), and data integration (combining data from different sources into a unified data view);

[0008] S3. Feature Extraction and Selection: Extract useful features from the fused data. These features can reflect the operating status or potential problems of the nuclear power system, and select key features according to the analysis objectives to reduce data redundancy and improve analysis efficiency;

[0009] S4. Data Analysis and Modeling: Apply statistical methods, machine learning algorithms or data mining techniques to analyze the extracted features to identify patterns, trends or anomalies in the data, and based on this, establish a prediction model for predicting the future status or performance of the nuclear power system;

[0010] S5. Visualization Design: According to the analysis objectives and audience needs, design a data visualization solution, specifically using nuclear power data visualization configuration software to simultaneously meet the analysis and visualization requirements of relational data (mainly covering the management data of nuclear power plants, such as plans, work orders, personnel, finance, materials, etc., which are the data types processed by traditional IT technologies) and time-series data (mainly the operating process status data of nuclear power plants, such as equipment status, temperature, pressure, flow rate, power, etc., which are usually obtained by sensors, control coordination and other equipment and belong to the field of traditional OT technologies);

[0011] S6. Visualization Implementation: Implement the designed visualization solution, display the analysis results in the form of graphics, images or animations, and verify the visualization results to ensure the effectiveness of data accuracy (by comparing the visualization results with the original data) and visualization effect (by user feedback and expert evaluation to verify the visualization effect);

[0012] S7. Result Interpretation and Application: Interpret the visualization results, reveal the meaning and laws behind the data, and apply the analysis results to the monitoring, diagnosis, optimization or decision support of the nuclear power system;

[0013] S8. Feedback and Optimization: According to the actual application effect and user feedback, continuously improve and optimize the visualization analysis method, continuously update the data source, optimize the data fusion algorithm, and improve the visualization design to improve the analysis efficiency and accuracy.

[0014] Furthermore, in step S1, the data collection channels include: collecting data from various sensors in the nuclear power plant (such as temperature sensors, pressure sensors, liquid level sensors, etc.), obtaining real-time operation data from the monitoring system of the nuclear power plant, and extracting historical data and event records from the database.

[0015] Furthermore, in step S1, data preprocessing includes data cleaning, data transformation and data standardization, specifically as follows:

[0016] Data cleaning: Remove invalid or incorrect data, including handling missing values (using methods such as deletion, replacement, or imputation), correcting data errors (identifying outliers, i.e., incorrect data, using univariate scatter plots or box plots), and deleting duplicate data;

[0017] Data transformation: Convert data into a format suitable for analysis, including operations such as data type conversion, data aggregation, and data grouping to improve data operability and analysis efficiency;

[0018] Data standardization: Unify the format, structure, and units of data to achieve data interoperability and analysis, enabling different data sources to have a unified measurement standard.

[0019] Furthermore, in step S2, data alignment specifically includes:

[0020] Time alignment: Synchronize the time of data from different sensors, specifically using methods such as interpolation and time window matching to make the sampling frequencies or timestamps of each sensor the same;

[0021] Spatial alignment: For data with spatial attributes (such as geographical coordinates), through spatial calibration or geographical coordinate transformation, make the spatial positions between different data sources consistent;

[0022] The data association: By analyzing the correlation, causal relationship, or similarity between data, identify the association relationships between different data sources. In the nuclear power field, specifically identify the mutual influences between different sensors and the data interaction between different systems, etc.;

[0023] The data integration: Use the Kalman filter algorithm to merge data from different sources into a unified data view. The goal of this step is to eliminate data redundancy, improve data quality and reliability, and form a comprehensive and accurate dataset for subsequent data analysis and application.

[0024] Furthermore, in step S3, feature selection is completed by constructing an objective optimization function, establishing feature evaluation indicators, etc. The specific operations are as follows:

[0025] The feature selection defines one or more optimization objectives according to the analysis goals, such as classification accuracy, regression error, etc., and then searches for the optimal feature combination through genetic algorithms or particle swarm algorithms to make the optimization objective reach the optimal;

[0026] The establishment of feature evaluation indicators constructs feature evaluation indicators based on theories such as statistical methods, machine learning algorithms, or information theory, and uses the indicators to measure the contribution degree of features to the analysis target. For example, based on indicators such as information gain and Gini coefficient, the importance of features in classification tasks can be evaluated; based on indicators such as correlation coefficient and mutual information, the linear or non-linear relationship between features and target variables can be evaluated.

[0027] In addition, in the field of nuclear power electronic information, feature selection should also fully consider domain knowledge. For example, according to the operating principles of nuclear power plants, equipment characteristics, etc., features closely related to key tasks such as nuclear safety and equipment status monitoring are screened out, so as to improve the pertinence and practicality of feature selection.

[0028] Further, in step S4, the specific operations of the three methods of statistical methods, machine learning algorithms, or data mining techniques are as follows:

[0029] Statistical modeling: Using statistical methods including linear regression, logistic regression, and time series analysis to analyze historical data, reveal the patterns and trends in the data, model the nuclear power system data, and identify the trends and periodic changes in the data through the model, providing a basis for prediction.

[0030] Machine learning algorithms: Adopting machine learning algorithms including support vector machine (SVM), decision tree, random forest, and neural network to perform tasks such as classification, clustering, and prediction on nuclear power system data to handle non-linear relationships and complex patterns. Specifically, through the training data set, the potential laws in the data are learned to predict the future state or performance of the nuclear power system.

[0031] Data mining techniques: Using data mining algorithms including association rule mining and anomaly detection to discover the potential laws and abnormal behaviors in nuclear power system data. These discoveries help to timely discover potential safety hazards and provide decision-making support for safety management.

[0032] Further, in step S5, the design of the visualization scheme specifically includes:

[0033] Determine the visualization target: Clearly define the data features, laws, and relationships that the visualization aims to display.

[0034] Select the visualization type: According to the data characteristics and target, select the visualization type, such as charts (bar charts, line charts, pie charts, etc.), maps, heat maps, animations, etc.

[0035] Design the layout and style: Design the overall layout, color matching, font size, etc. of the visualization to ensure clear and beautiful visual effects.

[0036] Further, in step S5, the nuclear power data visualization configuration software is designed based on the B / S (Browser / Server) structure, which is convenient for deployment, easy for users to access and for later expansion and maintenance. The server-side uses the cross-platform language JAVA for back-end development, and the front-end uses the Vue + Element UI framework and the open-source mxGraph drawing framework, supporting editing operations such as dragging, alignment, and scaling. At the same time, it provides multiple modules such as data mapping, component development, screen development, and system management, and supports general services such as log management and permission management.

[0037] Further, in step S6, the visualization tool selects FineReport, FineVis or Tableau to provide rich chart types and interactive functions to support the visualization requirements of specific data types in the nuclear power industry, and realizes the designed visualization solution by writing code or using the tool interface.

[0038] The present invention provides a method for multi-source data fusion and visualization analysis of nuclear power electronic information, having the following beneficial effects:

[0039] 1. The present invention can easily achieve multi-source data fusion, thereby helping relevant personnel better understand and analyze the operation status of the nuclear power system, improving the accuracy and efficiency of decision-making. With the application of data visualization, it can further make complex data intuitive and easy to understand, and help to discover potential problems and risks in a timely manner. At the same time, through standardized and modular design, the nuclear power data visualization configuration software improves the development efficiency of nuclear power data display screens, reduces the requirements of business personnel for professional knowledge such as IT and instrument control, enables users to focus more on the nuclear power business itself, and due to its flexible data access and processing capabilities, it can adapt to changing data and business requirements, further enhancing the multi-source data fusion and visualization analysis effects.

[0040] 2. The present invention can not only help understand the information of a single data source, but also reveal the relationships and mutual influences between data through comparison and correlation analysis between different data sources to obtain more comprehensive and accurate data results. At the same time, in the process of multi-source data fusion, there may be data conflicts and inconsistencies between different data sources. Through visualization analysis, it can help discover abnormal situations in the data, thus easily identifying these problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic diagram of the operation step flow of a method for multi-source data fusion and visualization analysis of nuclear power electronic information according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0042] The following further describes the embodiments of the present invention in detail in conjunction with the accompanying drawings and embodiments. The following embodiments are used to illustrate the present invention, but cannot be used to limit the scope of the present invention.

[0043] As Figure 1 shown, a multi-source data fusion and visualization analysis method for nuclear power electronic information includes the following steps:

[0044] S1. Data collection: Collect electronic information data during the operation of nuclear power from various sources, including collecting data from various sensors in the nuclear power plant (such as temperature sensors, pressure sensors, liquid level sensors, etc.), obtaining real-time operation data from the monitoring system of the nuclear power plant, and extracting historical data and event records from the database, and preprocessing the collected data. In this embodiment, data preprocessing includes data cleaning, data conversion, and data standardization, specifically as follows:

[0045] Data cleaning: Remove invalid or incorrect data, including handling missing values (using methods such as deletion, replacement, or interpolation), correcting data errors (identifying outliers, i.e., incorrect data, using univariate scatter plots or box plots), and deleting duplicate data;

[0046] Data conversion: Convert the data into a format suitable for analysis, including operations such as data type conversion, data aggregation, and data grouping, to improve the operability of the data and the efficiency of analysis;

[0047] Data standardization: Unify the format, structure, and unit of the data to achieve the intercommunication and analysis of the data, so that data from different sources have a unified measurement standard.

[0048] S2. Data fusion: Fusion the preprocessed multi-source data to form a more comprehensive and accurate data set, including data alignment (ensuring the consistency of data from different sources in time and space), data association (identifying the association relationships between different data sources), and data integration (merging data from different sources into a unified data view).

[0049] In this step, data alignment specifically includes:

[0050] Time alignment: Synchronize the time of data from different sensors, specifically using methods such as interpolation and time window matching to make the sampling frequencies or timestamps of each sensor the same;

[0051] Space alignment: For data with spatial attributes (such as geographical coordinates), through spatial calibration or geographical coordinate conversion, make the spatial positions between different data sources consistent;

[0052] In this step, data association: By analyzing the correlation, causality, or similarity between data, identify the association relationships between different data sources. In the nuclear power field, specifically, identify the mutual influence between different sensors, data interaction between different systems, etc.;

[0053] In this step, data integration: Use the Kalman filtering algorithm to merge data from different sources into a unified data view. The goal of this step is to eliminate data redundancy, improve data quality and reliability, and form a comprehensive and accurate dataset for subsequent data analysis and applications.

[0054] S3. Feature extraction and selection: Extract useful features from the fused data. These features can reflect the operating state or potential problems of the nuclear power system, and select key features according to the analysis goal to reduce data redundancy and improve analysis efficiency. In this embodiment, feature selection is completed by constructing an objective optimization function, establishing feature evaluation indicators, etc. The specific operations are as follows:

[0055] Feature selection defines one or more optimization goals according to analysis goals such as classification accuracy, regression error, etc., and then searches for the optimal feature combination through genetic algorithms or particle swarm algorithms to make the optimization goal reach the optimal;

[0056] Establish feature evaluation indicators. Based on theories such as statistical methods, machine learning algorithms, or information theory, construct feature evaluation indicators, and use the indicators to measure the contribution degree of features to the analysis goal. For example, based on indicators such as information gain and Gini coefficient, the importance of features in classification tasks can be evaluated; based on indicators such as correlation coefficient and mutual information, the linear or nonlinear relationship between features and target variables can be evaluated;

[0057] In addition, in the field of nuclear power electronic information, feature selection should also fully consider domain knowledge. For example, according to the operating principles of nuclear power plants, equipment characteristics, etc., screen out features closely related to key tasks such as nuclear safety and equipment status monitoring, so as to improve the pertinence and practicality of feature selection.

[0058] S4. Data analysis and modeling: Apply statistical methods, machine learning algorithms, or data mining techniques to analyze the extracted features to identify patterns, trends, or anomalies in the data, and based on this, establish a prediction model for predicting the future state or performance of the nuclear power system.

[0059] Statistical modeling: Use statistical methods including linear regression, logistic regression, time series analysis, etc. to analyze historical data, reveal the patterns and trends in the data, model the nuclear power system data, and identify the trends and periodic changes in the data through the model to provide a basis for prediction;

[0060] Machine learning algorithms: Machine learning algorithms including support vector machines (SVM), decision trees, random forests, neural networks, etc. are used to classify, cluster, predict, etc. the nuclear power system data to handle non-linear relationships and complex patterns. Specifically, through the training data set, the potential laws in the data are learned to predict the future state or performance of the nuclear power system;

[0061] Data mining techniques: Data mining algorithms including association rule mining, anomaly detection, etc. are used to discover the potential laws and abnormal behaviors in the nuclear power system data. These discoveries help to timely detect potential safety hazards and provide decision-making support for safety management.

[0062] S5. Visualization design: According to the analysis objectives and audience needs, design a data visualization solution, specifically including:

[0063] Determine the visualization objectives: Clearly define the data features, laws, and relationships that the visualization aims to display;

[0064] Select the visualization type: According to the data characteristics and objectives, select the visualization type, such as charts (bar charts, line charts, pie charts, etc.), maps, heat maps, animations, etc.;

[0065] Design the layout and style: Design the overall layout, color matching, font size, etc. of the visualization to ensure a clear and beautiful visual effect.

[0066] This step specifically uses nuclear power data visualization configuration software to simultaneously meet the analysis and visualization requirements of relational data (mainly covering the management data of nuclear power plants, such as plans, work orders, personnel, finance, materials, etc., which are the data types processed by traditional IT technologies) and time-series data (mainly the operation process state data of nuclear power plants, such as equipment status, temperature, pressure, flow rate, power, etc., which are usually obtained by sensors, control coordination and other devices and belong to the field of traditional OT technologies).

[0067] In this embodiment, the nuclear power data visualization configuration software is designed based on the B / S (Browser / Server) structure, which is convenient to deploy, easy for users to access and expand and maintain later. The server-side uses the cross-platform language JAVA for back-end development, and the front-end uses the Vue + Element UI framework and the open-source mxGraph drawing framework, supporting editing operations such as dragging, alignment, and scaling. At the same time, it provides multiple modules such as data mapping, component development, screen development, and system management, and supports general services such as log management and permission management.

[0068] S6, Visualization implementation: Implement the designed visualization solution, display the analysis results in the form of graphs, images, animations, etc., and verify the visualization results to ensure the accuracy of the data (by comparing the visualization results with the original data) and the effectiveness of the visualization effect (by user feedback and expert evaluation). In this step, visualization tools such as FineReport, FineVis, or Tableau are selected to provide rich chart types and interactive functions to support the visualization needs of specific data types in the nuclear power industry, and the designed visualization solution is implemented by writing code or using the tool interface.

[0069] S7, Result interpretation and application: Interpret the visualization results to reveal the meaning and laws behind the data, and apply the analysis results to aspects such as monitoring, diagnosis, optimization, or decision support of the nuclear power system.

[0070] S8, Feedback and optimization: Continuously improve and optimize the visualization analysis method according to the actual application effect and user feedback, and continuously update the data source, optimize the data fusion algorithm, and improve the visualization design to improve the analysis efficiency and accuracy.

[0071] The embodiments of the present invention are given for purposes of illustration and description, and are not exhaustive or limit the invention to the disclosed form. Many modifications and variations are obvious to those of ordinary skill in the art. The embodiments are selected and described to better illustrate the principles and practical applications of the present invention, and to enable those of ordinary skill in the art to understand the present invention and design various embodiments with various modifications suitable for specific purposes.

Claims

1. A method for multi-source data fusion and visualization analysis of nuclear power electronic information, characterized in that: The following steps are involved: S1. Data collection: Collect electronic information data during nuclear power operation from various sources and pre-process the collected data; S2, data fusion: fuse the pre-processed multi-source data, including data alignment, data association and data integration; S3, Feature extraction and selection: Extract useful features from the fused data and select key features according to the analysis objectives; S4. Data analysis and modeling: Apply statistical methods, machine learning algorithms or data mining techniques to analyze the extracted features to identify patterns, trends or anomalies in the data, and build predictive models based on them to predict the future state or performance of the nuclear power system; S5. Visualization design: Design data visualization solutions based on analysis objectives and audience needs, using nuclear power data visualization configuration software; S6. Visualization implementation: Implement the designed visualization scheme, display the analysis results, and verify the visualization results.

2. A method for multi-source data fusion and visualization analysis of nuclear power electronic information according to claim 1, characterized in that: In step S1, the data collection methods include: collecting data from various sensors of the nuclear power plant, obtaining real-time operation data from the monitoring system of the nuclear power plant, and extracting historical data and event records from the database.

3. A method for multi-source data fusion and visualization analysis of nuclear power electronic information according to claim 1, characterized in that: In step S1, data preprocessing includes data cleaning, data conversion and data standardization, as follows: Data cleaning: remove invalid or erroneous data, including processing missing values, correcting data errors, and deleting duplicate data; Data conversion: converting data into a format suitable for analysis, including data type conversion, data aggregation, and data grouping operations; Data standardization: unify the format, structure and unit of data to achieve data interoperability and analysis, so that data from different sources have unified measurement standards.

4. A method for multi-source data fusion and visualization analysis of nuclear power electronic information according to claim 1, characterized in that: In step S2, data alignment specifically includes: Time alignment: synchronize the data from different sensors by using interpolation and time window matching methods to make the sampling frequency or timestamp of each sensor the same; Spatial alignment: For data with spatial attributes, spatial calibration or geographic coordinate conversion is used to make the spatial positions of different data sources consistent; The data association is to identify the association between different data sources by analyzing the correlation, causal relationship or similarity between the data. In the field of nuclear power, it is to identify the mutual influence between different sensors and the data interaction between different systems. The data integration: uses a Kalman filter algorithm to merge data from different sources into a unified data view.

5. The method for multi-source data fusion and visualization analysis of nuclear power electronic information according to claim 1 is characterized in that: In step S3, feature selection is completed by constructing a target optimization function and establishing feature evaluation indicators. The specific operations are as follows: The feature selection defines one or more optimization targets according to the analysis target, and then searches for the optimal feature combination through genetic algorithm or particle swarm algorithm to achieve the optimal optimization target; The establishment of the feature evaluation index is based on statistical methods, machine learning algorithms or information theory, and the index is used to measure the contribution of the feature to the analysis target.

6. A method for multi-source data fusion and visualization analysis of nuclear power electronic information according to claim 1, characterized in that: In step S4, the specific operations of the three methods of statistical method, machine learning algorithm or data mining technology are as follows: Statistical modeling: Analyze historical data using statistical methods including linear regression, logistic regression, and time series analysis to reveal patterns and trends in the data. Model the nuclear power system data and use the model to identify trends and cyclical changes in the data. Machine learning algorithms: Use machine learning algorithms including support vector machines, decision trees, random forests, and neural networks to classify, cluster, and predict nuclear power system data to handle nonlinear relationships and complex patterns. Specifically, the training data set is used to learn the underlying laws in the data to predict the future state or performance of the nuclear power system. Data mining technology: Using data mining algorithms including association rule mining and anomaly detection, we can discover potential patterns and abnormal behaviors in nuclear power system data and provide decision support for safety management.

7. A method for multi-source data fusion and visualization analysis of nuclear power electronic information according to claim 1, characterized in that: In step S5, the design of the visualization scheme specifically includes: Determine the visualization goal: clarify the data characteristics, patterns, and relationships that the visualization aims to show; Select visualization type: Select the visualization type based on data characteristics and objectives, including but not limited to charts, maps, heat maps, and animations; Design layout and style: Design the overall layout, color matching, and font size of the visualization.

8. The method for multi-source data fusion and visualization analysis of nuclear power electronic information according to claim 1 is characterized in that: In step S5, the nuclear power data visualization configuration software is designed based on the B / S structure, and the server side uses the cross-platform language JAVA for back-end development. The front-end uses the Vue + Element UI framework and the open source mxGraph drawing framework, which supports editing operations such as dragging, aligning, and scaling.

9. The method for multi-source data fusion and visualization analysis of nuclear power electronic information according to claim 1, characterized in that: In step S6, a visualization tool is selected such as FineReport, FineVis or Tableau, and the designed visualization scheme is implemented by writing code or using the tool interface.

10. The method for multi-source data fusion and visualization analysis of nuclear power electronic information according to claim 1, characterized in that: The method further comprises the steps of: S7. Result interpretation and application: Interpret the visualization results and apply the analysis results to monitoring, diagnosis, optimization or decision support of nuclear power systems; S8. Feedback and optimization: Based on actual application results and user feedback, continuously improve and optimize the visualization analysis method, constantly update data sources, optimize data fusion algorithms, and improve visualization design.

Images