A data asset knowledge graph construction method and system for an integrated energy system

Abstract

The application discloses a kind of data asset knowledge graph construction methods and systems for integrated energy system, belong to integrated energy system information physical fusion technical field.The construction method disclosed in the present application can obtain the physical correlation between the numerical point of equipment and equipment, the numerical point inside equipment by analyzing the structure of integrated energy system and classifying equipment, read the historical data of all equipment numerical points, and obtain the correlation coefficient between points by calculation, combined with the threshold value set, the point group with data correlation is screened out, finally based on the interactive knowledge graph, the numerical point of equipment in energy system is stored as node, the physical correlation and data correlation are stored as edge, the overall situation of system is presented in an intuitive and structured manner, which greatly improves the comprehensiveness and accuracy of data correlation discovery, avoids the technical problems of subsequent research failure or low efficiency caused by insufficient data association mining.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of information and physical fusion of integrated energy systems, and particularly relates to a data asset knowledge graph construction method and system for integrated energy systems. BACKGROUND

[0002] In the process of rapid development of information technology, industrial automation and intelligence, integrated energy systems gradually occupy a core position in various industrial scenarios, and their complexity continues to grow with industrial upgrading. From the coordinated supply and deployment of various energies such as electricity, heat, and gas required by different production lines in large manufacturing plants, to the complex energy conversion and transmission equipment cluster in the integrated energy hub that integrates power generation, power transmission, power distribution, heat supply, gas supply, and other functions, the integrated energy system is like a highly integrated and precise intelligent network, with close coordination and state correlation between various energy subsystems and internal devices. The stability, efficiency and flexibility of its operation play a crucial role in the sustainable development of the entire industrial system.

[0003] Traditional energy system operation and maintenance, fault diagnosis, and operation optimization methods often focus on single device monitoring and maintenance, with device self-monitoring instruments, simple sensors recording basic parameters, and isolated storage in local controllers or upper computer databases. Data is fragmented and islanded. When the operator detects device anomalies, he relies on personal experience and device manuals to manually check associated devices, which is time-consuming and labor-intensive and may miss key clues, causing production losses in actual industrial production.

[0004] With the rise of industrial Internet of Things, the deployment of a large number of sensors has expanded the data collection dimension of integrated energy systems and increased the data collection frequency, resulting in an explosive growth in data volume. Current data analysis methods are limited to single-device time series analysis (statistical indicators, spectral analysis, etc.), or simple comparison of mean values and threshold values to determine the running state of the same type of device, ignoring the deep logic of physical connections and data interactions between devices.

[0005] Knowledge graph technology has emerged in other fields. It is a data structure and a way of expressing and storing knowledge. The core is to organize the objects and concepts in the real world and their various relationships in the form of a graph, providing rich, structured background knowledge for machine learning and improving the understanding and reasoning ability of algorithms. It is widely used in semantic search (search engine analyzes entity relationships to accurately present results), social network analysis (mining complex social ties between users), intelligent question and answer systems, recommendation systems, and other practical applications.

[0006] However, in the industrial field of integrated energy systems, the existing knowledge graph construction scheme has poor universality and insufficient pertinence. On the one hand, the general knowledge graph focuses on the integration of encyclopedic knowledge, and the entity relationship focuses on common sense semantics, which is not consistent with the professional terms and physical association rules of the energy system. The construction of the energy system professional knowledge graph is mostly based on theoretical design documents and standard specifications, and does not fully integrate a large amount of dynamic and multi-source data generated under actual operating conditions. The constructed graph is difficult to accurately map the complex whole of the real integrated energy system, and cannot mine the complex correlations between physical and data points of equipment, which cannot provide effective decision support for system operation, energy efficiency control, etc. SUMMARY

[0007] The purpose of the present application is to provide a data asset knowledge graph construction method and system for integrated energy systems, which solves the technical problems of subsequent research errors or low efficiency caused by insufficient data correlation mining in the existing knowledge graph construction scheme.

[0008] In order to achieve the above purpose, the following technical scheme is adopted:

[0009] The present application discloses a data asset knowledge graph construction method for integrated energy systems, comprising the following steps:

[0010] Performing structural analysis and equipment classification on the integrated energy system to obtain the physical correlation between the numerical points in the integrated energy system;

[0011] Reading the historical data of all numerical points in the integrated energy system, and solving the correlation coefficient between all numerical points according to the historical data; then setting a correlation coefficient threshold, comparing the correlation coefficient and the correlation coefficient threshold, and screening the correlation between the numerical points;

[0012] Storing all numerical points as nodes in a graph database, storing the physical correlation and correlation between numerical points as edges in the graph database, and establishing an interactive data asset knowledge graph.

[0013] Further, the physical correlation between the numerical points includes the physical correlation between the numerical points of the equipment and the physical correlation of the numerical points inside the equipment;

[0014] The physical correlation between the numerical points of the equipment is obtained after structural analysis of the integrated energy system;

[0015] The physical correlation of the numerical points inside the equipment is obtained after equipment classification of the integrated energy system.

[0016] Further, the step of analyzing the structure of the integrated energy system to obtain the physical correlation of the numerical point positions between the devices is as follows:

[0017] According to the physical topological relationship between the devices in the integrated energy system and the actual functions of the devices, the connection relationship between the devices is determined.

[0018] On the basis of the device connection relationship, the flow direction of the material flow in the connection relationship and the device numerical point positions corresponding to the flow direction of the material flow are analyzed, so as to obtain the physical correlation of the numerical point positions between the devices.

[0019] Further, the step of classifying the devices of the integrated energy system to obtain the physical correlation of the numerical point positions inside the devices is as follows:

[0020] All devices in the integrated energy system are classified according to the device type, the physical quantity of each type of device is extracted, and the node relationship between the physical quantities is constructed according to the physical relationship, so as to obtain the physical correlation of the numerical point positions inside the devices.

[0021] Further, the integrated energy system is an electrical system; the specific steps of classifying the devices of the integrated energy system to obtain the physical correlation of the numerical point positions inside the devices are as follows:

[0022] All devices in the integrated energy system are classified into electric meters, inverters and DC choppers according to the device type; all internal numerical point positions of the electric meters, inverters and DC choppers are extracted, and the node relationship is constructed according to the basic physical relationship between the current, voltage and power, so as to obtain the physical correlation of the numerical point positions inside the devices in the electrical system.

[0023] Further, the specific steps of constructing the node relationship according to the basic physical relationship between the current, voltage and power to obtain the physical correlation of the numerical point positions inside the devices in the electrical system are as follows:

[0024] For the electric meter:

[0025] Line voltage = 3 x phase voltage;

[0026] Apparent total power 2 = Active total power 2 + Reactive total power 2 ;

[0027] Total power factor = Active total power / Apparent total power;

[0028] For the inverter:

[0029] DC side power = DC current x DC lower half bus voltage;

[0030] For the DC chopper:

[0031] Total power = battery side voltage x battery side total current.

[0032] Further, the step of reading historical data of all numerical point positions in the integrated energy system and solving correlation coefficients between all numerical point positions according to the historical data comprises:

[0033] Reading data collection time, equipment name, equipment point position name and collected numerical value in the integrated energy system; for each equipment, taking the data collection time as the horizontal coordinate and the corresponding collected numerical value as the vertical coordinate, drawing a time sequence line chart of the equipment, visualizing the running state of the equipment, so as to analyze the running state of the equipment;

[0034] Aggregating all time sequence point position data of all equipment into a matrix, denoted as Z t×n ; wherein the row of the matrix represents the data collection time, denoted as t, and the column of the matrix represents the sum of all point positions of all equipment, denoted as n, and then calculating the correlation coefficient matrix between the columns in the matrix;

[0035] The step of calculating the correlation coefficient matrix between the columns in the matrix comprises the following steps:

[0036] For two column sample points X and Y in the matrix Z t×n , based on the sample expectation, variance and covariance, the following is estimated:

[0037]

[0038] Wherein: X i represents the i-th observation value in the sample, n represents the size of the sample, i.e. the total number of observation values, E(X) is the expectation of the random variable X, indicating the average value of all observation values; The value is the sample mean, cov(X) is the variance of the random variable X, indicating the dispersion degree of the observation value relative to the mean value; cov(X, Y) is the covariance of the random variables X and Y, indicating how two variables change together;

[0039] Substituting the above formula into the definition of the correlation coefficient, the definition formula is as follows:

[0040]

[0041] Wherein: σ X σ Y are the standard deviations of X and Y, respectively, E(XY) is the expectation of the product of X and Y, and p X,Y is the correlation coefficient;

[0042] Further processing obtains the calculation formula of the correlation coefficient as follows:

[0043]

[0044] Further, when the absolute value of the correlation coefficient between certain two point positions is greater than a threshold value, it is considered that the point position group has data level correlation, and accordingly the correlation between the numerical point positions is obtained.

[0045] Further, the setting standard of the correlation coefficient threshold value is as follows:

[0046]

[0047] Wherein r represents the correlation coefficient between the two groups of data;

[0048] The correlation coefficient threshold value is 0.5;

[0049] The interactive data asset knowledge graph is a data asset knowledge graph established based on the interaction of python and neo4j.

[0050] The application also discloses a data asset knowledge graph construction system for a comprehensive energy system, which is used for realizing the construction method.

[0051] Compared with the prior art, the application has the following beneficial effects:

[0052] The application discloses a data asset knowledge graph construction method for a comprehensive energy system, which can obtain the physical correlation relationship between numerical point positions of devices and devices and internal numerical point positions of devices through analysis of the structure of the comprehensive energy system and classification of the device types, read the historical data of all device numerical point positions, and obtain the correlation coefficient between the point positions through calculation, and in combination with a set threshold value, the point position groups with data correlation are screened out, the above process can be based on big data analysis, and potential and not easily perceived data contact can be fully mined, compared with the traditional technology which only relies on limited rules or experience judgment, the comprehensiveness and accuracy of data correlation discovery are greatly improved, and the technical problems of subsequent research mistakes or low efficiency caused by insufficient data correlation mining are avoided.

[0053] Further, the graph construction method disclosed by the application is a knowledge graph based on the interaction of python and neo4j, the numerical point positions of the devices in the energy system are stored as nodes, the physical correlation and data correlation relationship are stored as edges, and the whole system is presented in an intuitive and structured manner. BRIEF DESCRIPTION OF DRAWINGS

[0054] Figure 1 The method block diagram of the data asset knowledge graph construction method for the comprehensive energy system of the application is shown in the figure;

[0055] Figure 2 The device topology graph of the electrical system example of the application is shown in the figure;

[0056] Figure 3 The node label information and edge category information generated after the storage of nodes and edges in embodiment 1 of the present invention is completed;

[0057] Figure 4 This is the final knowledge graph visualization result of Example 1 of the present invention. DETAILED DESCRIPTION

[0058] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0059] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0060] The present invention is described in further detail below with reference to the accompanying drawings:

[0061] like Figure 1 As shown, the present invention discloses a method for constructing a data asset knowledge graph for an integrated energy system. The method can establish a knowledge graph of the system based on the system's physical topology, equipment categories, and historical data information. The method mainly includes the following steps:

[0062] Conduct structural analysis on the integrated energy system to obtain the physical correlation between the numerical points of each device; classify the equipment in the integrated energy system into categories and, based on the principles of physics, determine the physical correlation between the numerical points within each type of equipment;

[0063] Read the historical data of all numerical points in the integrated energy system and solve the correlation coefficients between all numerical points based on the historical data; then set the correlation coefficient threshold, and filter out the correlation between the numerical points based on the comparison between the correlation coefficient and the correlation coefficient threshold;

[0064] All numerical points are stored as nodes (entities) in the graph database, and the physical correlation and correlation between numerical points are stored as edges (relationships) in the graph database, to establish a comprehensive energy system knowledge graph based on the interaction of python and neo4j.

[0065] Preferably, the method for analyzing the structure of the comprehensive energy system to obtain the physical correlation of the numerical points between the system devices is as follows:

[0066] According to the physical topological relationship between the devices in the system and the actual function of the devices, the connection relationship between the devices is determined.

[0067] On the basis of the device connection relationship, the flow direction of the material flow in the connection relationship and the corresponding device numerical points are analyzed, so as to obtain the physical correlation of the numerical points between the devices.

[0068] Preferably, the method for classifying the devices of the comprehensive energy system and obtaining the physical correlation of the internal numerical points of each type of device according to the principles of physics is as follows:

[0069] Taking the electrical system as an example, the devices in the electrical system are classified into three types: ammeter, inverter (DC / AC), and DC chopper (DC / DC).

[0070] Extract all internal numerical points of each device, and construct a node relationship formula according to the basic physical relationship between current, voltage, power, etc., so as to obtain the physical relationship between the internal numerical points of the device.

[0071] More preferably, the method for constructing a point relationship formula according to the basic physical relationship between current, voltage, power, etc., so as to obtain the physical relationship between the internal numerical points of the device is as follows:

[0072] For the ammeter device, it is easy to know that:

[0073] Line voltage = √3 × phase voltage;

[0074] Apparent total power 2 = active total power 2 + reactive total power 2 ;

[0075] Total power factor = active total power / apparent total power;

[0076] For the inverter (DC / AC), it is easy to know that:

[0077] DC side power = DC current × DC lower half bus voltage;

[0078] For the DC chopper (DC / DC), it is easy to know that:

[0079] Total power = battery side voltage x battery side total current.

[0080] Preferably, the historical data of all numerical point positions in the system is read from the database, and the correlation coefficient between all point positions is solved according to the historical data as follows:

[0081] Taking the electrical system as an example, all numerical point position data of all devices in a month is exported from the database of the electrical system according to the device category by using Navicat, including data collection time, device name, device point position name, collected numerical value and other basic elements;

[0082] For each device, the data collection time is taken as the abscissa, and the corresponding collected numerical value is taken as the ordinate, and the time sequence line chart of the device is drawn, the running state of the device is visualized, so as to analyze the running state of the device, if the collected data of the device is missing or abnormal for a long time, the data of other months is exported from the database again, and finally the data set of all devices in normal running state is obtained;

[0083] All time sequence point position data of all devices are aggregated into a new matrix Z t×n , wherein the row t of the matrix represents the data collection time (i.e. the number of point position data), and the column n of the matrix represents the number of point positions (i.e. the sum of all point positions of all devices of the electrical system), and the correlation coefficient matrix between each column of the matrix is calculated.

[0084] For two column sample points X and Y in the matrix Z t×n , the expectation, variance and covariance of the sample are estimated based on the sample:

[0085]

[0086] Wherein: X i represents the i-th observation value in the sample, n represents the size of the sample, i.e. the total number of observation values, E(X) is the expectation of the random variable X, which represents the average value of all observation values; The value is the sample mean, is the variance of the random variable X, which represents the dispersion degree of the observation value relative to the mean value; cov(X, Y) is the covariance of the random variables X and Y, which represents how two variables change together;

[0087] The above formula is substituted into the definition of the correlation coefficient, and the definition formula is as follows:

[0088]

[0089] Wherein: σ X σ Y are the standard deviations of X and Y, respectively, and E(XY) is the expectation of the product of X and Y ρX,Y i.e. the correlation coefficient;

[0090] The calculation formula of the correlation coefficient is:

[0091]

[0092] The correlation coefficient between the two sets of point data can be calculated according to the above formula.

[0093] Preferably, the suitable correlation coefficient threshold h is set, and the method of screening all numerical point groups with data correlation in the system according to the threshold is as follows:

[0094] From historical experience, we know that:

[0095]

[0096] Where r represents the correlation coefficient between the two sets of data;

[0097] According to historical experience and actual system requirements, the threshold is selected as h = 0.5;

[0098] If the absolute value of the correlation coefficient between two points is greater than the threshold, it is considered that the point group has data correlation, and thus the data correlation between all numerical points in the system is obtained.

[0099] Preferably, the method of storing all numerical points of all devices in the system as nodes (entities) in the graph database is as follows:

[0100] Before performing specific operations, install the third-party library py2neo in the python environment to operate neo4j;

[0101] Connect the database according to the browser access address, username and password of neo4j;

[0102] Create point nodes of all devices in each type according to the device type, and the following process takes the type of electric meter as an example;

[0103] Create a node list of all devices in the electric meter in the python project, and the object in the list is:

[0104] Node(*labels, **properties)

[0105] Where *labels is the label of the node, representing the type of the node, which is set to the device name here;

[0106] **properties is the attribute of the node, and the data structure is a dictionary, which can store the attributes of the node, and is set to the device point name here;

[0107] For example, the specific objects obtained are:

[0108] Node('Energy Station No. 1 Electricity Meter', name = 'Phase A Voltage')

[0109] Use Subgraph to create a node list to create nodes in batches;

[0110] By using graph.create() to store all the created batch nodes into the graph database, all the numerical points of all the equipment in the electrical system can be stored as nodes in the graph database.

[0111] Preferably, the method for storing the physical correlation and data correlation relationship between numerical points as edges (relationships) in a graph database is as follows:

[0112] The structure for creating an edge is:

[0113] Relationship(start_node,type,end_node)

[0114] Where start_node represents the starting node of the edge, end_node represents the end point of the edge, and type represents the attribute of the edge, which is used here to distinguish the physical relevance and data relevance of the edge;

[0115] Create physical-related edges and data-related edges between nodes respectively;

[0116] Use graph.create() to store the created edges in the graph database.

[0117] Example 1

[0118] A method for constructing a data asset knowledge graph for an integrated energy system, using an electrical system in this embodiment, specifically includes the following steps:

[0119] Step 1: Analyze the overall system structure and obtain the physical correlation between the numerical points of the system devices;

[0120] The equipment diagram structure of the electrical system in this embodiment is as follows Figure 2 As shown, the connection relationship between each device is determined based on the physical topological relationship between the devices in the electrical system and the actual function of the devices. Based on the device connection relationship, the flow direction of electricity in the connection relationship and the corresponding device numerical points are analyzed to obtain the physical association between the numerical points of the devices;

[0121] Table 1 shows some examples of numerical points with physical associations between specific devices obtained using the above method:

[0122] Table 1 Physical correlation between numerical points of devices

[0123]

[0124]

[0125] Step 2, all devices in the system are classified according to device type, and the physical relationship between internal numerical point positions of each type of device is obtained according to physical principles (physical correlation between internal numerical point positions of the device) ;

[0126] For the devices in the electrical system, they are divided into three categories according to type: electric meter, inverter (DC / AC), and DC chopper (DC / DC). Extract all internal numerical point positions of each device, and construct node relationship according to the basic physical relationship between current, voltage, power and other physical quantities, so as to obtain the physical relationship between the internal numerical point positions of the device;

[0127] For the electric meter device, it is easy to know that:

[0128] Line voltage = √3 × phase voltage;

[0129] Apparent total power 2 = Active total power 2 + Reactive total power 2 ;

[0130] Total power factor = Active total power / Apparent total power;

[0131] For the inverter (DC / AC), it is easy to know that:

[0132] DC side power = DC current × DC lower half bus voltage;

[0133] For the DC chopper (DC / DC), it is easy to know that:

[0134] Total power = Battery side voltage × Battery side total current;

[0135] According to the above physical relationship, part of the numerical point position of the device with physical relationship is shown in Table 2:

[0136] Table 2 Physical correlation of internal numerical point positions of the device

[0137]

[0138]

[0139] Step 3, read the historical data of all numerical point positions in the system from the database, and solve the correlation coefficient between all point positions according to the historical data; use Navicat to export all numerical point position data of all devices in a month from the database of the electrical system according to the device category, including data collection time, device name, device point position name, collected numerical value and other basic elements, the sampling interval is 5 minutes, and Table 3 is part of the data information of the device;

[0140] Table 3 Data information of the device

[0141] load_time system_name device_name position_name load_value 2024 / 9 / 1 0:00:00 photovoltaic system DC / DC - 1st DC / DC 1# DCDC DC bus voltage 2.4 2024 / 9 / 1 0:00:00 photovoltaic system DC / DC - 1st DC / DC 1# DCDC battery side voltage 8.8 2024 / 9 / 1 0:00:00 photovoltaic system DC / DC - 1st DC / DC 1# DCDC total discharge capacity 48.5 2024 / 9 / 1 0:00:00 photovoltaic system DC / DC - 2nd DC / DC 2# DCDC DC bus voltage 759.1 2024 / 9 / 1 0:00:00 photovoltaic system DC / DC - 2nd DC / DC 2# DCDC battery side voltage 10 2024 / 9 / 1 0:00:00 distribution network system DC / AC - 1st DC / AC 1# DC bus voltage 375.3 2024 / 9 / 1 0:00:00 distribution network system DC / AC - 1st DC / AC 1# reactive power 2.9 2024 / 9 / 1 0:00:00 distribution network system DC / AC - 1st DC / AC 1# DC bus voltage 749.4 2024 / 9 / 1 0:00:00 distribution network system DC / AC - 2nd DC / AC 2# Uab voltage 388.6 2024 / 9 / 1 0:00:00 distribution network system DC / AC - 2nd DC / AC 2# Ubc voltage 387.3 2024 / 9 / 1 0:00:00 electricity meter DK1 electricity meter - 1st DK1 electricity meter DK1_1 isolated transformer switchgear A-phase voltage 241.5 2024 / 9 / 1 0:00:00 electricity meter DK1 electricity meter - 1st DK1 electricity meter DK1_1 isolated transformer switchgear AB-line voltage 419 2024 / 9 / 1 0:00:00 electricity meter DK1 electricity meter - 1st DK1 electricity meter DK1_1 isolated transformer switchgear A-phase current 2.4 2024 / 9 / 1 0:00:00 electricity meter DK1 electricity meter - 2nd DK1 electricity meter DK1_2 isolated transformer switchgear total active power 28.8 2024 / 9 / 1 0:00:00 electricity meter DK1 electricity meter - 2nd DK1 electricity meter DK1_2 isolated transformer switchgear total apparent power 29.6

[0142] For each device, the data collection time is taken as the abscissa, and the corresponding collected numerical value is taken as the ordinate, and the time series line graph of the device is drawn, the running state of the device is visualized, and the running state of the device is analyzed. If the collected data of the device is missing or abnormal for a long time, the data of other months is exported from the database again, and finally the data set of all devices in normal running state is obtained; all time series point position data of all devices are aggregated into a new matrix Z t×n , where the row t of the matrix represents the data collection time (i.e. the number of point data), and the column n of the matrix represents the number of points (i.e. the sum of all points of all devices in the electrical system). The correlation coefficient matrix between each column in the matrix is calculated.

[0143] For the matrix Z t×n , the sample points X and Y in two columns are estimated based on the sample expectation, variance and covariance:

[0144]

[0145] Substitute the above formula into the definition of the correlation coefficient, and the definition formula is as follows:

[0146]

[0147] The calculation formula of the correlation coefficient is:

[0148]

[0149] The correlation coefficient between two groups of point position data can be calculated according to the above formula.

[0150] Step 4, set a suitable correlation coefficient threshold h, and filter all numerical point position groups with data correlation in the system according to the threshold. According to historical experience:

[0151]

[0152] According to historical experience and actual demand of the electrical system, the threshold is selected as h=0.5;

[0153] If the absolute value of the correlation coefficient between two points is greater than the threshold, it is considered that the point group has data-level correlation, and the data correlation between all numerical points in the electrical system is obtained based on this.

[0154] Step 5: Store all numerical points of all devices in the system as nodes (entities) in the graph database, store the physical and data-related relationships between the numerical points as edges (relationships) in the graph database, and establish an electrical system knowledge graph based on the interaction between Python and Neo4J.

[0155] Before performing specific operations, install the third-party library py2neo in the Python environment to operate neo4j;

[0156] Connect to the database using the neo4j browser access address, username, and password;

[0157] Create point nodes for all devices of each type according to the device type. The following process takes the electric meter type as an example.

[0158] In the Python project, create a node list of all devices in the meter. The objects in the list are:

[0159] Node(*labels, **properties)

[0160] *labels is the label of the node, representing the type of the node, which is set to the device name here;

[0161] **properties is the property of the node. The data structure is a dictionary, which can store the properties of the node. Here it is set to the device point name;

[0162] For example, the specific objects obtained are:

[0163] Node('Energy Station No. 1 Electricity Meter', name = 'Phase A Voltage')

[0164] Use Subgraph to create a node list to create nodes in batches;

[0165] On the basis of creating all point nodes, a new node is created for each device, named device node, which is used to connect to all point nodes of the device in the future for easy management. The label of the device node is set to device;

[0166] By using graph.create() to store all the created batch nodes into the graph database, all the numerical points of all the equipment in the electrical system can be stored as nodes in the graph database.

[0167] The structure for creating an edge is:

[0168] Relationship(start_node, type, end_node)

[0169] wherein start_node represents the starting node of the edge, end_node represents the end node of the edge, and type represents the attribute of the edge, which is used to distinguish the physical correlation and the data correlation of the edge here;

[0170] The edges of the physical correlation and the edges of the data correlation between the nodes are respectively created;

[0171] The created edges are stored into the graph database by using graph.create();

[0172] The node label information and the edge type information after storage are shown as in Figure 3 The node label information is shown in the Node labels part, each node is labeled with the device name to which the node belongs, the device node is labeled as a device, and different types of devices are distinguished by different colors. The edge type information is shown in the Relationship types part, which is the physical correlation and the data correlation, and is distinguished by different colors.

[0173] The final knowledge graph visualization result is shown as in Figure 4 .

[0174] The above content is only for illustrating the technical idea of the present application, and cannot limit the protection scope of the present application. Any modification made according to the technical idea of the present application on the basis of the technical scheme falls within the protection scope of the claims of the present application.

Claims

1. A method for constructing a data asset knowledge graph for an integrated energy system, characterized in that: The following steps are involved: Conduct structural analysis and equipment classification of the integrated energy system to obtain the physical correlation between numerical points in the integrated energy system; Read the historical data of all numerical points in the integrated energy system and solve the correlation coefficients between all numerical points based on the historical data; then set the correlation coefficient threshold, and filter out the correlation between the numerical points based on the comparison between the correlation coefficient and the correlation coefficient threshold; All numerical points are stored as nodes in a graph database, and the physical correlations and correlations between the numerical points are stored as edges in the graph database to build an interactive data asset knowledge graph; The physical correlation between the numerical points includes the physical correlation between the numerical points between devices and the physical correlation between the numerical points within the devices; The physical correlation between the numerical points of the devices is obtained after structural analysis of the integrated energy system; The physical correlation relationship between the numerical points inside the equipment is obtained by classifying the equipment in the integrated energy system; The steps of reading the historical data of all numerical points in the integrated energy system and solving the correlation coefficients between all numerical points based on the historical data include: Read the data collection time, device name, device location name, and collected value in the integrated energy system; for each device, draw a time series line chart with the data collection time as the horizontal axis and the corresponding collected value as the vertical axis to visualize the device's operating status and analyze the device's operating status; Aggregate all time series point data of all devices into a matrix, expressed as ; The rows of the matrix represent the data acquisition time, denoted as t, and the columns of the matrix represent the sum of all points of all devices, denoted as n. Then the correlation coefficient matrix between each column in the matrix is ​​calculated; Calculating the correlation coefficient matrix between the columns in the matrix includes the following steps: For the matrix Two columns of sample points 、 , estimate the expectation, variance, and covariance based on the sample: ； ； ； in: represents the i-th observation in the sample, n represents the size of the sample, that is, the total number of observations, is a random variable The expectation of , which represents the average of all observations; The value is the sample mean, is a random variable The variance indicates the dispersion of the observations relative to the mean; is a random variable and The covariance of , which indicates how the two variables vary together; Substituting the above formula into the definition of the correlation coefficient, its definition is as follows: ； in: They are and The standard deviation of yes and The expectation of the product, is the correlation coefficient; The calculation formula of the correlation coefficient is further processed as follows: 。 2. A method for constructing a data asset knowledge graph for an integrated energy system according to claim 1, characterized in that: The steps for structural analysis of the integrated energy system and obtaining the physical correlation between the numerical points of the equipment are as follows: Determine the connection relationship between devices based on the physical topology relationship and actual functions of the devices in the integrated energy system; Based on the equipment connection relationship, the flow direction of the material flow in the connection relationship and the equipment numerical points corresponding to the flow direction of the material flow are analyzed, so as to obtain the physical correlation relationship between the numerical points between the equipment.

3. The method for constructing a data asset knowledge graph for an integrated energy system according to claim 1, characterized in that: The steps for classifying the equipment in the integrated energy system and obtaining the physical correlation between the internal numerical points of the equipment are as follows: All equipment in the integrated energy system is classified according to equipment categories, the physical quantities of each category of equipment are extracted, and the node relationship between the physical quantities is constructed based on the physical relationship, so as to obtain the physical correlation relationship between the numerical points inside the equipment.

4. A method for constructing a data asset knowledge graph for an integrated energy system according to claim 3, characterized in that: The integrated energy system is an electrical system. The specific steps for classifying the equipment in the integrated energy system and obtaining the physical correlation between the internal numerical points of the equipment are as follows: All equipment in the integrated energy system is divided into electricity meters, inverters, and DC choppers according to equipment categories; all internal numerical points of the electricity meters, inverters, and DC choppers are extracted, and node relationship equations are constructed based on the basic physical relationship between current, voltage, and power, thereby obtaining the physical correlation between the internal numerical points of the equipment in the electrical system.

5. The method for constructing a data asset knowledge graph for an integrated energy system according to claim 4, characterized in that: The specific steps for constructing node relationships based on the basic physical relationships between current, voltage, and power to obtain the physical correlation between the internal numerical points of the equipment in the electrical system are as follows: For electric meters: Line voltage = √3 × phase voltage; Apparent total power 2 = total active power 2 +Total reactive power 2 ; Total power factor = total active power / total apparent power; For inverters: DC side power = DC current × DC lower half bus voltage; For DC chopper: Total power = battery side voltage × battery side total current.

6. The method for constructing a data asset knowledge graph for an integrated energy system according to claim 1, characterized in that: When the absolute value of the correlation coefficient between two points is greater than the threshold, it is considered that the point group has data-level correlation, and the correlation relationship between the numerical points is obtained accordingly.

7. The method for constructing a data asset knowledge graph for an integrated energy system according to claim 1, characterized in that: The setting criteria for the correlation coefficient threshold are as follows: ； Where r represents the correlation coefficient between two sets of data; The correlation coefficient threshold is 0.5; The interactive data asset knowledge graph is a data asset knowledge graph established based on the interaction between python and neo4j.

8. A data asset knowledge graph construction system for integrated energy systems, characterized by: Used to implement the construction method described in any one of claims 1 to 7.

Images