A power transmission information model storage medium and method

Abstract

The present invention provides a transmission information model storage medium and method. The method includes collecting and preprocessing transmission network operation data, formatting and filtering node voltage, line power, load distribution and equipment status; constructing a transmission information model including a static topology layer, a dynamic operation status layer and a prediction analysis layer based on the preprocessed data, wherein the topology layer describes the network infrastructure, the dynamic layer updates electrical parameters according to the topological connection, and the prediction layer predicts load changes using time series data and historical records. A distributed storage architecture is used to partition and store model hierarchical data and realize incremental synchronization through a consistency protocol, while generating a global index to support fast query, and providing query, update and analysis result export services through a standardized interface.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and is a power transmission information model storage medium and method. Background Art

[0002] As the scale of power transmission networks continues to expand, the widely used power transmission information management methods in existing technologies are usually based on independent topology analysis, real-time data monitoring and load forecasting tools. These technologies obtain operating data such as node voltage, line power and equipment status through data acquisition equipment, and use offline analysis models to process historical data. Some solutions also try to achieve management and access to massive data through distributed storage technology.

[0003] However, common problems in existing technologies include: data collection and preprocessing have not formed an efficient collaborative mechanism, and the accuracy of data anomaly filtering is low; real-time parameter updates and load forecasts lack deep correlation, resulting in insufficient dynamic response capabilities; data synchronization efficiency in distributed storage architectures is low, and there is a lack of unified interface specifications to support cross-platform access and analysis.

[0004] Therefore, there is an urgent need to provide a new technical solution to improve the shortcomings of the existing methods. Summary of the invention

[0005] The present application provides a power transmission information model storage medium and method to improve the efficiency of power transmission network management.

[0006] The present application provides a power transmission information model storage medium, wherein the storage medium stores computer-readable instructions, and when the computer-readable instructions are executed on a processor, the power transmission information model construction method is used to implement the power transmission information model construction method, and the power transmission information model construction method includes:

[0007] Collecting and preprocessing the operation data of the power transmission network, wherein the real-time operation data includes node voltage, line power, load distribution and equipment operation status; the preprocessing includes formatting the data and filtering out exceptions;

[0008] According to the preprocessed operation data, a transmission information model is constructed, and the transmission information model includes a static topology layer, a dynamic operation status layer and a prediction and analysis layer; wherein the static topology layer is used to describe the basic connection structure of the transmission network; the dynamic operation status layer updates the real-time parameters based on the connection relationship provided by the static topology layer to reflect the changes in electrical parameters; the prediction and analysis layer combines the time series data of the dynamic operation status layer and the historical operation records to predict future transmission load changes;

[0009] Based on a distributed storage architecture, the hierarchical data of the power transmission information model is partitioned and stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, while a global index is generated to support fast query;

[0010] Provide standardized interface services, including model query interface, model update interface and analysis result export interface.

[0011] Furthermore, the collecting and preprocessing of the operation data of the power transmission network includes:

[0012] Create a time segment tag for each batch of newly acquired operation data, and classify and store the node and line records within the same time segment;

[0013] After classification, identification codes are added to data items that span multiple time segments and whose parameters fluctuate continuously to distinguish the special status of these data items in subsequent analysis;

[0014] The data structure with the identification code attached is output to a unified pre-processing index list, so that the subsequent steps can track the timing information of specific lines and nodes based on the identification code.

[0015] Furthermore, based on the distributed storage architecture, the hierarchical data of the power transmission information model is partitioned and stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, including:

[0016] According to the association relationship between nodes and lines in the static topology layer, the hierarchical data of the power transmission information model is divided into multiple logical partitions and a unique identifier is assigned to each partition;

[0017] A version record is created for each partition that has been assigned an identifier. When the data content of any partition changes, the change information is appended in the form of a new version record;

[0018] When a partition update request is received, only the partition data with differences is synchronized according to the change information in the version record, and the version status of the corresponding partition in the global index is updated after the synchronization is completed.

[0019] Furthermore, it includes:

[0020] Create an association table for each partition data, and map the node set and line set corresponding to each partition to a specific data storage location;

[0021] When data is updated, the time stamp and partition identifier of the update event are recorded in the associated table so that the subsequent synchronization step can perform local data refresh for the partition;

[0022] After completing the partial data refresh, the associated table is adjusted accordingly based on the record information of the update event so that the subsequent query step can locate the latest data source through the table.

[0023] Furthermore, the provision of standardized interface services includes:

[0024] Prepare a parameter list for each interface request type, registering the field names, data types and search paths required for query, update and export in the list;

[0025] After receiving the external request, the corresponding partition data or model record is located from the global index according to the matching request parameters of the list;

[0026] After the data is located, the search results are assembled into a response data packet in a unified format and returned to the caller according to the parameter list requirements of the request type.

[0027] Furthermore, the provision of standardized interface services includes:

[0028] Create a configuration file for the analysis result export interface, which defines the time span, node range and line selection rules of the output data;

[0029] When receiving an export request, determine the content to be extracted from the prediction analysis layer or the dynamic operation status layer according to the configuration file, and retrieve the corresponding partition data;

[0030] The retrieved data is packaged into a recognizable export data format according to the order of fields listed in the configuration file, and the data format is output to the caller as a unified result.

[0031] The present application provides a method for constructing a power transmission information model, comprising:

[0032] Collecting and preprocessing the operation data of the power transmission network, wherein the real-time operation data includes node voltage, line power, load distribution and equipment operation status; the preprocessing includes formatting the data and filtering out exceptions;

[0033] According to the preprocessed operation data, a transmission information model is constructed, and the transmission information model includes a static topology layer, a dynamic operation status layer and a prediction and analysis layer; wherein the static topology layer is used to describe the basic connection structure of the transmission network; the dynamic operation status layer updates the real-time parameters based on the connection relationship provided by the static topology layer to reflect the changes in electrical parameters; the prediction and analysis layer combines the time series data of the dynamic operation status layer and the historical operation records to predict future transmission load changes;

[0034] Based on a distributed storage architecture, the hierarchical data of the power transmission information model is partitioned and stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, while a global index is generated to support fast query;

[0035] Provide standardized interface services, including model query interface, model update interface and analysis result export interface.

[0036] The beneficial effects of the present application mainly include: (1) By formatting and filtering the operating data during the collection process, the present application realizes the standardization of real-time data and the elimination of abnormal data, significantly improving the quality of data input and ensuring that the basic data for subsequent model construction has high accuracy and consistency. (2) The present application describes the basic structure, real-time status and future trends of the transmission network through hierarchical modeling of static topology layer, dynamic operation status layer and prediction analysis layer, and combines time series data with historical records for dynamic update and prediction analysis, effectively improving the accuracy and timeliness of transmission load prediction. (3) The present application adopts a distributed storage architecture to partition the hierarchical data of the transmission information model into multiple nodes, and realizes incremental synchronization through a consistency protocol, while supporting fast query through a global index, thereby improving the efficiency and reliability of data management and adapting to the data storage needs of large-scale transmission networks. (4) The present application enables external systems to efficiently access the transmission information model through standardized model query, update and analysis result export interfaces, realizes cross-platform data interaction and function expansion, and provides good support for real-time monitoring, prediction analysis and scheduling optimization of the transmission network. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flowchart of a method for constructing a power transmission information model involved in the first embodiment of the present application. DETAILED DESCRIPTION

[0038] Many specific details are described in the following description to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present application, so the present application is not limited by the specific implementation disclosed below.

[0039] The first embodiment of the present application provides a power transmission information model storage medium. Figure 1 , which is a flow chart of a method for constructing a power transmission information model involved in the first embodiment of the present application. Figure 1 A power transmission information model storage medium is described in detail in the first embodiment of the present application.

[0040] The storage medium stores computer-readable instructions, which, when executed on a processor, are used to implement a method for constructing a power transmission information model. The method for constructing a power transmission information model includes:

[0041] Step S101: collecting and preprocessing the operation data of the power transmission network, wherein the real-time operation data includes node voltage, line power, load distribution and equipment operation status; the preprocessing includes formatting the data and filtering out exceptions.

[0042] Step S101 in the power transmission information model construction method is responsible for collecting and preprocessing the operation data of the power transmission network to ensure the accuracy and standardization of the input data. Step S101 is described in detail below.

[0043] In step S101, the operation data of the power transmission network is first collected in real time through multiple types of sensor equipment and monitoring systems. The operation data includes but is not limited to the following four main types: node voltage data, line power data, load distribution data and equipment operation status data. Node voltage data refers to the voltage value of each node, which is collected in real time by voltage sensors deployed at transmission nodes; line power data records the actual power flow between transmission lines, which comes from power sensors or measuring devices on transmission lines; load distribution data reflects the dynamic changes of loads at each node of the transmission network, which is obtained through the load monitoring system; equipment operation status data records the operation status of key equipment in the transmission network, such as transformer temperature, circuit breaker position and switch status. These data are transmitted to the data acquisition server in real time through a dedicated communication protocol.

[0044] After the data is collected, it is preprocessed to ensure that it meets the requirements of model building. Preprocessing mainly includes two parts: formatting and anomaly filtering. In the formatting stage, the collected raw data is converted into a unified structured format. For example, for node voltage and line power data, the data is stored in a time series database format with timestamps as indexes; load distribution data is divided into regions and summarized by time period of each region; equipment operation status data is classified and processed by status tags. The purpose of this step is to facilitate subsequent data access and retrieval while ensuring the consistency of the structure of different types of data.

[0045] Abnormal filtering is another important part of preprocessing. By setting the normal range and threshold of the operating data, for example, the fluctuation range of the node voltage can be set to ±5% of the rated voltage, the values ​​outside the range are marked, and combined with the timing information to determine whether it is a continuous abnormality; for line power, if negative values ​​or abnormal values ​​far exceeding the rated power appear continuously, it is judged as data abnormality and recorded. The standards for filtering abnormal load distribution data include whether the load fluctuation amplitude is too large and whether the load exceeds the maximum carrying capacity of the equipment. The abnormality of the equipment operation status data is mainly determined by comparing the actual status with the normal status specified in the equipment manual, such as the transformer temperature exceeding the safety threshold or the circuit breaker switching frequently.

[0046] After detecting abnormal data, interpolation or time series analysis is used to repair some missing or abnormal values. The repaired data is stored in the temporary data area together with the normal data, and the data with abnormal marks is stored separately in the abnormal record table for further analysis and processing in the subsequent steps.

[0047] After completing the above-mentioned collection and preprocessing work, the output of step S101 includes a standardized operation data set and an abnormal record data set, which provides high-quality input data for building a power transmission information model in step S102. Through this collection and preprocessing method, step S101 ensures the integrity and accuracy of the power transmission network operation data, providing a reliable data foundation for subsequent steps.

[0048] Furthermore, the collecting and preprocessing of the operation data of the power transmission network includes:

[0049] Create a time segment tag for each batch of newly acquired operation data, and classify and store the node and line records within the same time segment;

[0050] After classification, identification codes are added to data items that span multiple time segments and whose parameters fluctuate continuously to distinguish the special status of these data items in subsequent analysis;

[0051] The data structure with the identification code attached is output to a unified pre-processing index list, so that the subsequent steps can track the timing information of specific lines and nodes based on the identification code.

[0052] The method for collecting and preprocessing the operation data of the power transmission network involved in this embodiment establishes an accurate and efficient data management system by introducing the mechanisms of time segment marking, identification code attachment and unified index list to support the construction and subsequent application of the power transmission information model.

[0053] In the process of establishing time segment tags for each batch of newly acquired operating data, the real-time collected data first needs to be segmented on the time axis. This can be done by setting fixed time intervals, such as 1 minute or 5 minutes, to group data such as node voltage, line power, load distribution, and equipment operating status. Each time segment tag consists of a timestamp and a segment number. For example, the time segment marked as "2024-12-0715:00:00-T1" represents data collected in the first time period at 3 pm on December 7, 2024. In this way, the operating data of all nodes and lines in the same time segment will be uniformly classified and stored in the corresponding groups. This segmentation tagging mechanism not only reduces the complexity of data retrieval, but also lays the foundation for subsequent time series analysis.

[0054] After completing the time segment tagging and grouping, the system attaches identification codes to data items that span multiple time segments and whose parameters fluctuate continuously. The purpose of this step is to track data anomalies or trend changes. The attachment of identification codes is based on specific detection rules, such as node voltage continuously exceeding ±5% of the rated value, or line power increasing at a constant rate over multiple time segments. For data items that meet the conditions, the system generates a unique identification code, such as "ID-001-VOLTAGE-RISING" to represent the gradual upward trend of node voltage. The identification code not only records the type of data anomaly or characteristic, but also includes the relevant starting time segment, duration, and identification information of the associated node or line. This mechanism provides clear markings for subsequent analysis, allowing specific data to be quickly located and processed in a targeted manner.

[0055] After the identifier is attached, all time segment data together with the marked identifier data structure will be output to a unified pre-processed index list. The index list is a structured data table used to store and organize the data grouping of each time segment and the associated information of the attached identifier. For example, an item in the list may include the time segment tag "2024-12-0715:00:00-T1", the node identifier "Node-A", the identifier "ID-001-VOLTAGE-RISING", the data source and related feature descriptions. The list also supports hierarchical management, such as storing data marked with identifiers in a high-priority index layer for quick retrieval in subsequent model building or anomaly analysis.

[0056] This series of operations forms a complete process from data collection to preprocessing, which has good scalability and flexibility. For example, by adjusting the length of the time segment, the data density requirements of different scenarios can be adapted; by expanding the types of identification codes, more complex abnormal pattern recognition can be supported. In addition, the structured design of the index list makes it easy to integrate with other system modules, such as distributed storage and dynamic operation status analysis modules.

[0057] Through these steps, this embodiment achieves accurate classification, anomaly tracking and structured storage of transmission network operation data, ensuring the efficiency and reliability of subsequent model building and analysis tasks, while also providing a solid foundation for complex time series analysis and multidimensional data mining.

[0058] Step S102: construct a power transmission information model based on the preprocessed operation data, wherein the power transmission information model includes a static topology layer, a dynamic operation status layer and a prediction and analysis layer; wherein the static topology layer is used to describe the basic connection structure of the power transmission network; the dynamic operation status layer updates the real-time parameters based on the connection relationship provided by the static topology layer to reflect the changes in electrical parameters; the prediction and analysis layer combines the time series data of the dynamic operation status layer and the historical operation records to predict future changes in transmission load.

[0059] Step S102 involves constructing a transmission information model based on the preprocessed operating data, which consists of a static topology layer, a dynamic operating status layer, and a predictive analysis layer. First, node voltage, line power, load distribution, and equipment operating status data are extracted from the standardized operating data output by step S101. Combined with the physical connection information of the network, a static topology layer is established, which describes the basic connection relationship of the transmission network in the form of a graph structure. Nodes represent physical devices of the transmission network, such as transformers, circuit breakers, and busbars, and edges represent transmission lines connecting these nodes. During the construction process, an association table is created based on the attribute values ​​of the nodes and edges (for example, rated voltage, line length, resistance, and capacitance) for efficient storage and access to detailed information of each node and line.

[0060] After the static topology layer is completed, the dynamic operation status layer is built based on the static topology layer. By associating the node voltage and line power in the operation data and combining the topological connection relationship, the real-time load distribution of each node and the power flow of the line are calculated. The construction process of the dynamic layer includes an iterative update mechanism, which gradually propagates the dynamic changes of load and power by performing multiple rounds of calculations on the real-time parameters of each node in the topology diagram. These updates are carried out based on the electrical laws in the network (such as power balance conditions) and the time series data output in the preprocessing stage, so as to reflect the changes in electrical parameters of the entire transmission network in real time.

[0061] The prediction and analysis layer uses the time series data and historical operation records of the dynamic operation state layer as the basis, and uses pattern recognition and statistical modeling methods to predict future changes in transmission load. In the specific implementation, the historical parameter data of each node and line is stored in segments, and the input features are extracted in units of time windows. For example, the load change curve over time and the power fluctuation range are used to determine potential trends and abnormal change points. By matching the current state of the dynamic layer with similar operating modes in historical records, the prediction and analysis layer can output the predicted values ​​of load distribution and power flow in the future.

[0062] The construction process of the entire model has a strict hierarchical dependency. The static topology layer provides the physical structure foundation for the dynamic operation status layer, and the output of the dynamic operation status layer provides time series data and status characteristics for the prediction analysis layer. Through this hierarchical structure, the transmission information model can fully cover the physical connection relationship, dynamic electrical parameters and future load trends of the transmission network, forming a logically clear and highly scalable modeling framework. The completed model will serve as the core foundation for distributed storage and query optimization in subsequent steps.

[0063] The following are specific examples of constructing each layer of the model:

[0064] The static topology layer is the basis of the model and is used to describe the physical connection relationship of the transmission network. Its construction starts with the input node and line data, including the attributes of the node (such as equipment type, rated voltage) and the attributes of the line (such as starting node, ending node, line length, resistance, etc.). For example, in a simple network, node A represents a transformer, node B is a bus, and the two are connected by a line (A, B). The attributes of the line include a length of 10 kilometers, a resistance of 0.1 ohms, and a capacitance of 50 microfarads. This information is organized into a topological graph, in which nodes represent transmission equipment, edges represent line connection relationships, and detailed attributes of nodes and lines are stored through association tables. This structure not only clearly describes the basic layout of the transmission network, but also provides support for subsequent dynamic parameter calculations.

[0065] Based on the static topology layer, the dynamic operation status layer reflects the real-time changes of electrical parameters in the transmission network. By combining real-time data, such as node voltage and line power, the dynamic operation status layer gradually updates the status of nodes and lines based on the topology map. First, the initial state of each node is set according to its collected voltage value, while the initial power of the line comes from the corresponding real-time operation data. During the update process, the power balance principle is followed, that is, the total power inflow of a node should be equal to its total power outflow plus its own load. For example, when the power of line (A, B) increases from 20 MW to 25 MW, the load of node B needs to be adjusted accordingly and propagated to the downstream line (B, C), thereby completing the iterative update of the entire network. The updated data, such as node voltage, line power, and load distribution, are written into the dynamic state table to provide real-time input for subsequent predictions.

[0066] The predictive analysis layer combines the time series data and historical operation records of the dynamic operation status layer to predict future load changes in the transmission network. This layer is based on the historical data of nodes and lines, extracts time series features, such as the fluctuation range of node voltage and the trend change of line power, and matches them with the current state. If the current operating state has a similar pattern to a historical record, the predictive analysis layer will use the record to infer future change trends. For example, if historical data shows that when the voltage at node A drops to 105 kV, the power of line (A, B) usually rises to 30 MW, the predictive analysis layer will output this trend forecast when similar conditions are detected. The generated prediction results are recorded in the form of time series for further analysis and verification by the system.

[0067] Furthermore, the construction of the static topology layer includes weighting the initial adjacency matrix representing the connection relationship between the nodes of the power transmission network and their lines, so that each line not only reflects whether there is a connection, but also reflects the stability of the line in long-term operation and the sensitivity to load changes through the weight;

[0068] The weighted value W(i,j) of the line (i,j) is calculated according to the following formula (1):

[0069] W(i,j)=α·ln(1+R(i,j))+β·[L(i,j)] μ (1)

[0070] Wherein, R(i,j) is the line reliability index of line (i,j); L(i,j) is the line sensitivity index of line (i,j); α, β and μ are positive real number parameters; the line reliability index R(i,j) is calculated according to the following formula (2):

[0071]

[0072] Where M is the number of historical observation periods; f m is the proportion of fault-free operation time of line (i, j) in the mth cycle; ε is a small positive number used to avoid zero value; γ is the nonlinear amplification factor;

[0073] The line sensitivity index L(i,j) is calculated according to the following formula 3:

[0074]

[0075] Where T is the number of load data sampling points; P i,j (t) is the measured load value of line (i, j) at time t; P ref (t) is the reference load value of line (i, j) at time t; δ is the adjustment factor; ω t is the time adjustment factor, calculated according to the following formula 4:

[0076]

[0077] Among them, λ is the adjustment factor.

[0078] In order to more intuitively understand how the construction of the static topology layer reflects the characteristics of the transmission network through weighted processing, this process can be illustrated by a specific example.

[0079] Suppose there is a simple power transmission network, including three nodes A, B and C. There is a transmission line between node A and node B, and there is a transmission line between node B and node C, but there is no direct connection between node A and node C. The initial connection relationship of this network can be represented by an adjacency matrix, where 1 indicates that there is a line connection between the two nodes, and 0 indicates no connection:

[0080]

[0081] In traditional topological descriptions, this adjacency matrix can only indicate the existence of a connection, but cannot reflect the characteristics of each line, such as long-term stability or sensitivity to load changes. In order to make the topological structure more meaningful, it is necessary to weight the non-zero values ​​in the matrix so that these values ​​not only indicate the existence of the line, but also quantify the performance of the line.

[0082] For example, suppose the operating data of line (A, B) and line (B, C) are as follows:

[0083] For line (A, B), the trouble-free operation time accounted for 98% in the past year, and the load changes were relatively stable.

[0084] For line (B, C), the fault-free operation time accounts for 75%, and the load changes greatly and fluctuates significantly.

[0085] Through weighted processing, a comprehensive weight can be calculated for each line, which reflects both its operational stability and its sensitivity to load changes. Assumptions:

[0086] Lines with high stability (e.g., longer trouble-free operation time) have larger weights;

[0087] The line weights with larger load fluctuations also increase accordingly.

[0088] After weighted calculation, the value of the matrix may become:

[0089]

[0090] Here, the weight of line (A, B) is 0.85, indicating that its operation is relatively stable and the load fluctuation is small; while the weight of line (B, C) is 0.60, reflecting that its operation stability is relatively low and it is sensitive to load changes.

[0091] In this way, the static topology layer can not only describe the connection relationship of the transmission network, but also quantify the actual operating characteristics of each line. This weighted topology matrix can provide an important basis for subsequent analysis, such as giving priority to monitoring lines with lower weights, or considering the potential impact of highly sensitive lines on the operation of the entire network in predictive analysis.

[0092] The static topology layer in this embodiment constructs the basic connection structure of the power transmission network through a weighted adjacency matrix, and weights the reliability and load sensitivity of the line so that the weight of each line not only indicates whether it is connected, but also reflects its long-term operating stability and sensitivity to load changes. This process is based on the mathematical description of formulas (1) to (4), and the specific implementation steps and the definition and calculation methods of the parameters in the formulas are as follows.

[0093] First, according to formula (1), the weighted value of line (i, j) is obtained by a linear combination of two indicators, namely, line reliability index and line sensitivity index. The coefficient parameters in the weight formula include α, β and μ, where α and β are positive real numbers used to balance the weights of the two indicators, and μ is a nonlinear adjustment factor for the sensitivity index. The specific values ​​of these parameters can be set according to the actual operation of the transmission network. For example, when the system pays more attention to the long-term operation stability of the line, α can be set to 0.7 and β to 0.3; when the load fluctuation sensitivity is more important, the parameter values ​​are adjusted in the opposite direction.

[0094] The line reliability index is calculated according to formula (2), which is obtained by weighted averaging the fault-free operation time of historical observation periods. The number of historical observation periods is usually 30 days, 60 days or other time spans, depending on the monitoring frequency and data availability of the transmission network. The fault-free operation time ratio represents the normal operation ratio of the line in each observation period, and its value ranges from 0 to 1. In order to avoid invalid calculation results when the fault-free operation time is zero, a small positive number ε is introduced, which is usually 0.001 or less. The exponent γ is a nonlinear amplification factor used to perform nonlinear amplification on high-reliability lines. For example, when γ is 2, the weight contribution of reliable lines will be significantly enhanced.

[0095] For the line sensitivity index, its calculation is based on formula (3), which measures the sensitivity of the line to load changes by taking a weighted average of the line load deviation. The number of load data sampling points is usually related to the sampling frequency of the system. For example, a sampling frequency of once every 10 minutes will generate 144 sampling points in a day. The measured load value is the real power flow of the line at a certain moment, while the reference load value is usually determined by the average value of the historical load or the planned design value. The adjustment factor δ is used to adjust the contribution of high load deviation, and its value is generally set between 1 and 3. The time adjustment factor is calculated by formula (4), which gives different degrees of influence to the sampling points based on the time weight. The adjustment factor λ controls the fluctuation range of the time weight. For example, when λ is 0.5, the weight changes from 1 to 1.5, which is suitable for highlighting the load characteristics of a specific time period.

[0096] In the time adjustment factor formula (4), the time point is the index of the current sampling point, ranging from 1 to . By introducing the cosine function, the time weight shows a periodic change within a day, which is often used to emphasize the load characteristics of specific periods such as daytime or nighttime. This weight design is suitable for transmission networks with obvious daytime fluctuations in load.

[0097] Through the above steps, the line weighted value can comprehensively reflect the key characteristics of each line in the transmission network. The reliability index provides an assessment of the long-term stability of the line, while the sensitivity index describes the response characteristics of the line to load changes. The combination of the two provides a comprehensive and dynamically adaptive weighting scheme for the static topology layer, which can further optimize the model construction and calculation efficiency of the transmission network in the subsequent dynamic operation status layer and predictive analysis layer.

[0098] Furthermore, the transmission load forecasting method of the forecasting and analysis layer embeds the line weight information formed by the static topology layer into the training and calculation process of the forecasting model based on the long short-term memory network to improve the accuracy of the load forecasting of specific lines and the pertinence of the scheduling decision, specifically including:

[0099] The multi-time series electrical parameter data output by the dynamic operation state layer and the R(i,j) and L(i,j) parameters corresponding to each line (i,j) calculated in the static topology layer and their weighted values ​​W(i,j) are integrated into an input feature vector, and the feature vector is numerically normalized;

[0100] The data samples corresponding to each line in the input feature vector are sorted in descending order according to the W(i,j) value, and the sorting number r represents the position of the line (i,j) in the sequence, and in the process of updating the hidden layer state and output calculation of the long short-term memory network model, the input data of the line with the sorting number r is assigned the influence factor G(r), so that the data samples corresponding to the line with a larger W(i,j) value obtain a numerically higher influence factor G(r) in the state calculation, thereby realizing the weight enhancement of the characteristics of this type of line in time series processing;

[0101] The impact factor G(r) is calculated using the following formula 5:

[0102]

[0103] where c is a predefined positive real constant.

[0104] The transmission load forecasting method of this embodiment embeds the line weight information formed by the static topology layer into the forecasting model based on the long short-term memory network (LSTM), so as to fully consider the importance of each line in the time series analysis, improve the accuracy of the forecast and the pertinence of the scheduling decision. The following is a detailed description of the specific implementation process of the method and each parameter and weight in the formula.

[0105] First, the dynamic operation status layer outputs multi-time series electrical parameter data, such as the time series of node voltage, line power and load distribution. These data reflect the actual operation status of the transmission network over a period of time. At the same time, the static topology layer calculates the weighted value W(i,j) of each line according to the above method. The weight combines the reliability R(i,j) and sensitivity L(i,j) of the line. Specifically, R(i,j) represents the long-term operation stability of the line, which may be calculated by the proportion of fault-free operation time. For example, R(A,B) = 0.9 of line (A,B) indicates that the past operation reliability is high; while L(i,j) represents the sensitivity of the line to load changes. For example, if the load fluctuation of line (B,C) is large, its L(B,C) = 1.2 reflects its high sensitivity. After these values ​​are processed by formula (1), the weighted value W(i,j) is generated, such as W(A,B) = 0.85 for line (A,B) and W(B,C) = 0.6 for line (B,C).

[0106] Next, the timing data of the dynamic operation state layer and the weight information of the static topology layer are integrated into the input feature vector. The feature vector of each line includes the timing electrical parameters (such as the power value sequence) and the values ​​of R(i,j), L(i,j) and W(i,j). For example, for line (A, B), the feature vector may contain the following information: [P A,B (t1),P A,B (t2), R(A,B), L(A,B), W(A,B)]. The integrated eigenvectors are uniformly normalized to ensure that data of different dimensions can be treated equally in the model. Normalization can be done using common methods, such as linear normalization, to map all eigenvalues ​​to the range of [0,1].

[0107] Subsequently, the data samples corresponding to each line in the input feature vector are sorted in descending order according to the value of W(i,j), and the lines with higher weights are given priority. For example, if W(A,B) of line (A,B) = 0.85, and W(B,C) of line (B,C) = 0.6, then after sorting, line (A,B) is before line (B,C). After sorting, a sorting number r is assigned to each line, for example, r(A,B) = 1, r(B,C) = 2.

[0108] In the process of updating the hidden layer state and calculating the output of the LSTM model, the system assigns an influence factor G(r) to the line with the sorting number r, so that the data samples corresponding to the line with a larger weight gain a higher influence. The influence factor is calculated using formula (5):

[0109]

[0110] Where c is a predefined positive real constant used to control the absolute value range of the impact factor. Specifically, if c = 1.0, the impact factor of the first-ranked line G(1) = 1.0, and the impact factor of the second-ranked line G(2) = 0.5. This design ensures that the importance of the line decreases step by step in the calculation of the model, but still retains a certain influence for the lines with lower weights.

[0111] In the LSTM model, these influencing factors act on the updating process of the hidden state through the gating mechanism. For example, the weight of the feature vector of the line with a higher influencing factor is amplified in the memory unit, which more significantly affects the final prediction output in time series processing. This mechanism can preferentially capture the contribution of high-weight lines to the overall operation state of the network, thereby improving the accuracy of the prediction results.

[0112] Through the above method, this embodiment effectively integrates the weight information of the static topology layer into the LSTM model, which not only optimizes the organization of input data, but also improves the accuracy and pertinence of prediction through the dynamic allocation of influencing factors. This design is particularly suitable for quickly identifying the operating trends of key lines in large-scale power transmission networks and providing a reliable basis for actual scheduling decisions.

[0113] Furthermore, the real-time parameter update of the dynamic operation state layer includes updating the real-time parameters of the power transmission network through a Markov state transition model to reflect the immediate changes and abnormal fluctuations of the electrical parameters, specifically including:

[0114] Using the line weight information determined by the static topology layer, the transfer probability between each node and line in the Markov state transition model is dynamically corrected, so that the key lines with higher weights have more influence on the update of node voltage, line power and load distribution during the state transition process;

[0115] In the process of continuously receiving real-time operation data, when it is observed that key parameters such as node voltage or line power deviate from their expected range, the corresponding state calibration step is triggered to timely suppress abnormal states that may spread along the key lines by redistributing transition probabilities;

[0116] After the operating parameters return to stability, the updated state transfer relationship is used as the basic information for subsequent time steps, so that the dynamic operating status layer can achieve refined management and continuous optimization of the evolution trend of electrical parameters without relying on multiple prediction models.

[0117] The dynamic operation state layer of this embodiment uses a Markov state transition model to update the real-time parameters of the power transmission network to reflect the immediate changes and abnormal fluctuations of electrical parameters. This design combines the weight information of the static topology layer, so that the state transition process can more accurately consider the role of key lines, and at the same time suppress the spread of abnormalities through dynamic correction and state calibration, ensuring the real-time and stability of the network.

[0118] During the implementation process, the calculated line weight information in the static topology layer is first used to initialize and dynamically correct the transition probability between nodes and lines in the Markov state transition model. The line weight W(i,j) reflects its reliability and load sensitivity. For critical lines with higher weights, their influence in the transition probability matrix increases accordingly. For example, assuming that the weight of line (A, B) is 0.85 and the weight of line (B, C) is 0.60, during the state transition process, the transition probability of line (A, B) will be amplified to preferentially reflect its important role in the node voltage or line power update. Specifically, the transition probability can be corrected by the following rule: the higher the line weight, the more significant the impact of the electrical parameter update of its related nodes on other nodes or lines, so that the critical line plays a leading role in state propagation.

[0119] In the process of continuously receiving real-time operation data, the dynamic operation status layer updates the current network status in real time based on the actually observed node voltage, line power and load distribution. For example, when the voltage of node A drops from the rated value of 110kV to 105kV, the system adjusts the parameter changes of the lines and downstream nodes connected to node A according to the transition probability matrix. This adjustment is completed through the Markov state transition formula, in which the next state of each node is a weighted combination of its current state and transition probability. This design ensures that the state update can promptly reflect the real-time changes of the network and reasonably allocate the propagation range and intensity of the changes based on the topological structure.

[0120] When a critical parameter is observed to deviate from its expected range (for example, line power exceeds a safety threshold), the system triggers a state calibration step. The purpose of this process is to suppress the spread of anomalies and prevent the network from triggering a chain reaction due to a single point failure. The core of state calibration is to redistribute transition probabilities to limit the propagation of abnormal states along critical lines. For example, when the power fluctuation of line (A, B) is abnormal and may be transmitted to node B, the calibration step reduces the transition probability of the abnormal state propagating to downstream node C, while enhancing the compensation capabilities of other lines directly connected to node B (such as line (B, D)). This dynamic adjustment can localize the abnormal state at the first time and prevent it from spreading to the entire network.

[0121] After the operating parameters return to stability, the system stores the updated state transition relationship as the basic information for the next time step. This means that the transition probability matrix after state calibration will be used in subsequent state update steps, thereby gradually optimizing the state evolution process. This design has memory and can continuously adjust the transition rules according to the dynamic changes of the actual operation of the network, thereby improving the adaptability and robustness of the model.

[0122] The advantage of this method is that it does not rely on complex prediction models, and the changing trends of electrical parameters can be finely managed only through the Markov state transition model. For example, for a power transmission network with frequent load distribution fluctuations, the model can accurately capture the spatial distribution characteristics of load fluctuations through real-time updates and dynamic calibration, and predict the next change trend to a certain extent. This capability is particularly important for real-time monitoring and exception handling of large-scale power transmission networks. By closely combining static topology weight information, real-time data streams, and state transition models, this embodiment provides an efficient and flexible implementation method for the dynamic operation status layer, which can meet the real-time and stability requirements of complex power transmission networks.

[0123] The following example illustrates how to use the calculated line weight information in the static topology layer to initialize and dynamically correct the transition probability in the Markov state transition model.

[0124] Assume that a transmission network consists of three nodes A, B, and C and two lines (A, B) and (B, C). Node A and node B are connected by line (A, B), and node B and node C are connected by line (B, C). The weight W(i, j) of each line has been calculated in the static topology layer. For example, the weight of line (A, B) is 0.85, and the weight of line (B, C) is 0.60. These weights are calculated by comprehensively considering the operation of the line. The weight is derived from the performance and load sensitivity of the line, where line (A, B) has a higher weight because it is more stable and responds more gently to load changes.

[0125] In the Markov state transition model, the transition probability matrix is ​​used to describe the probability of change of the electrical parameter state between nodes. For example, assume that the current node voltage or line power may be transmitted to other nodes through the connecting line within one time step. In the initial state, the transition probability matrix can be evenly distributed, indicating that all connecting lines contribute equally to the state change. However, in order to more accurately reflect the characteristics of the actual transmission network, the transition probability needs to be corrected according to the line weight W(i,j), so that the line with higher weight plays a more significant role in the state update.

[0126] The correction process is as follows: First, the transfer probability of each node is normalized so that the sum of the probabilities of transferring from a node to all adjacent nodes is 1. Take node B as an example. It is connected to node A through line (A, B) and to node C through line (B, C). In the initial state of uniform distribution, the transfer probability of node B to both node A and node C may be 0.5. However, since the weight of line (A, B) is 0.85 and the weight of line (B, C) is 0.60, these two transfer probabilities need to be redistributed after correction.

[0127] Correction can be achieved by:

[0128] First, take the weights of lines (A, B) and (B, C) as the basic values ​​of weights, and calculate the sum of weights:

[0129] =W(A,B)+W(B,C)=0.85+0.60=1.45

[0130] Next, the transition probability of each line is normalized:

[0131]

[0132] Through this correction, line (A, B) dominates the transfer process, and its weight advantage is directly reflected in the increase of transfer probability, while line (B, C) also participates in state propagation, but its contribution is small. This adjustment of transfer probability can ensure that the state update process is more in line with the actual operation characteristics of the network, and highlight the influence of key lines.

[0133] Step S103: Based on the distributed storage architecture, the hierarchical data partitions of the power transmission information model are stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, while a global index is generated to support fast query.

[0134] Step S103 is the core step of storing the hierarchical data of the power transmission information model in the distributed storage system and realizing the incremental synchronization and fast query functions. This step specifically involves the implementation of the partition storage strategy, data synchronization mechanism and global index generation method for the hierarchical data, ensuring that the model can run efficiently in a large-scale distributed environment.

[0135] When implementing step S103, it is first necessary to partition the data of the static topology layer, the dynamic operating status layer, and the predictive analysis layer. Partitioning is based on the network topology and data association characteristics. Each partition consists of a group of physically or logically closely related nodes and lines, ensuring that the data within the partition can support efficient local storage while minimizing cross-partition access. For example, a partition may include a substation and its directly associated transmission lines and load nodes. In this way, the model data is mapped to multiple partitions and a unique identifier is assigned to each partition.

[0136] After data partitioning is completed, the data of each partition is stored in an independent node in the distributed storage system. During the storage process, metadata records need to be created for each partition's data, including information such as the partition identifier, data range, and storage location. These metadata records are centrally stored in a global index table for subsequent rapid location and access to partition data. The global index table is maintained in the form of key-value pairs, for example, the key is the partition identifier, and the value is the storage node location and data update version of the partition.

[0137] For data synchronization, an incremental synchronization mechanism is used to ensure that all partitions maintain consistency when data is updated. The incremental synchronization mechanism is implemented by maintaining version control records. When the data of a partition changes, the partition generates an update record with a timestamp and propagates the record to other storage nodes in the distributed system. This propagation uses a consistency protocol, such as a two-phase commit protocol or a synchronization strategy based on consistent hashing, to ensure that each node eventually receives the same data version. In addition, incremental synchronization only sends the changed data blocks, which significantly reduces the amount of transmitted data and improves synchronization efficiency.

[0138] The global index table plays a key role in this process. Each time it is updated, the global index table is modified accordingly to reflect the latest partition version information. When a query request arrives, the global index table will be accessed first to quickly determine the partition where the query target is located and its latest data version. In this way, the query can directly locate the data node of the target partition, avoiding unnecessary global scans.

[0139] To ensure the scalability of the storage system, the distributed storage system adopts a dynamic partitioning strategy. When the load increases or the storage nodes are expanded, the partitions can be re-divided and reallocated as needed. The partition migration process will also update the global index table to ensure that data access is always accurate.

[0140] Through these steps, hierarchical data is effectively partitioned and stored, updated synchronously, and supports fast query, so that the transmission information model can run efficiently in a distributed environment, providing a solid foundation for the implementation and practical application of standardized interfaces in subsequent steps.

[0141] Furthermore, based on the distributed storage architecture, the hierarchical data of the power transmission information model is partitioned and stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, including:

[0142] According to the association relationship between nodes and lines in the static topology layer, the hierarchical data of the power transmission information model is divided into multiple logical partitions and a unique identifier is assigned to each partition;

[0143] A version record is created for each partition that has been assigned an identifier. When the data content of any partition changes, the change information is appended in the form of a new version record;

[0144] When a partition update request is received, only the partition data with differences is synchronized according to the change information in the version record, and the version status of the corresponding partition in the global index is updated after the synchronization is completed.

[0145] In this embodiment, the hierarchical data partitions of the power transmission information model are stored in multiple nodes based on a distributed storage architecture, and incremental synchronization is achieved through a consistency protocol, which is a key step to ensure the efficiency and consistency of large-scale power transmission network data management. The specific implementation process includes both the initial steps of logical partitioning and identity allocation, as well as the dynamic maintenance of version records and synchronization processing, forming a complete and flexible data storage and update system.

[0146] First, the data of the power transmission information model is divided into multiple logical partitions based on the association between nodes and lines described in the static topology layer. The division of logical partitions is based on the topological structure and aims to reduce cross-partition data access. Each partition can contain one or more physical nodes and their associated transmission lines. For example, a partition may include substation A and its connected lines (A, B) and (A, C). This partitioning method can aggregate topologically closely related data, thereby reducing the frequency of accessing multiple partitions in query or update operations.

[0147] Assign a unique identifier to each logical partition. The identifier is usually generated based on the physical or logical attributes of the partition, such as the node number or the start and end nodes of the line. The purpose of the partition identifier is to create a separate namespace for each partition, which is used to identify and access the partition data between storage nodes. For example, the partition containing the data of substation A may be assigned the identifier "PART-A", while the data containing the line (A, B) may be identified as "PART-AB".

[0148] After completing the partition and identification assignment, the system creates a version record for each partition. The version record is used to track the update status of the partition data, including the updated timestamp, summary information of the updated content, and the source of the update operation. For example, when the rated voltage parameter of substation A in the partition "PART-A" is modified, the system generates a version record, such as "VERSION-002", which records the time of the update, the specific fields modified, and the new values. These version records are stored in a chain structure, and each update appends a new version node to the existing record chain.

[0149] When the storage system receives a partition update request, the system first retrieves the current version information of the target partition from the global index table. Then, by comparing the data in the new request with the version records of the existing data, the system identifies the parts that are different. For example, if the impedance value of line (A, B) is adjusted in partition "PART-AB", the system will detect that the value of this field is different from that in the previous version. Incremental synchronization processing only operates on these identified differences, avoiding the waste of resources caused by full data synchronization.

[0150] After synchronization is completed, the system will record the new data status to the target partition and update the version information of the partition in the global index table. For example, the version of the partition "PART-AB" is updated from "VERSION-003" to "VERSION-004", and the new timestamp is written to the index table. This process ensures the consistency of data in the entire distributed system and improves synchronization efficiency.

[0151] The flexibility of this architectural design is also reflected in its scalability. When a new storage node is added or the partition needs to be repartitioned, the logical partition can be reallocated to different physical nodes, and the maintenance mechanism of the partition identification and version record ensures the integrity and consistency of the data. For example, when the storage node load increases, the partition "PART-A" can be split into two sub-partitions and assigned to different storage nodes, and the existing version records can be seamlessly inherited to the sub-partitions.

[0152] Through the partitioned storage and incremental synchronization mechanism of this hierarchical data, this embodiment can effectively support the storage and management of the power transmission information model in a distributed environment, and provide a solid technical foundation for subsequent efficient query and model update.

[0153] Furthermore, the hierarchical data of the power transmission information model is partitioned and stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, including:

[0154] Create an association table for each partition data, and map the node set and line set corresponding to each partition to a specific data storage location;

[0155] When data is updated, the time stamp and partition identifier of the update event are recorded in the associated table so that the subsequent synchronization step can perform local data refresh for the partition;

[0156] After completing the partial data refresh, the associated table is adjusted accordingly based on the record information of the update event so that the subsequent query step can locate the latest data source through the table.

[0157] This embodiment involves the management and synchronization of hierarchical data of a power transmission information model based on a distributed storage architecture, specifically including creating an association table for each partition data, recording event information and performing local data refresh when updating, and adjusting the association table according to the update record to achieve an efficient and consistent data storage and update mechanism.

[0158] First, in order to achieve efficient management and access of partitioned data in a distributed environment, an association table needs to be created for each partitioned data. The association table is a structured data structure used to store the mapping relationship between the partition and its corresponding node set and line set. The division of partitions is based on the topological characteristics of the transmission network. For example, a partition can contain a substation and its directly connected transmission line. The association table maps each partition identifier to the nodes and lines it contains. For example, the partition identifier "PART-A" may correspond to substation A and lines (A, B) and (A, C). The association table also stores the physical storage location of this data, such as the IP address of the storage node or the logical volume identifier, so as to quickly locate the target data in query or update operations.

[0159] When the data in a partition is updated, the system records the details of the update event in the association table, including the time stamp of the event, the partition ID, and a summary of the update content. For example, if the impedance value of line (A, B) is modified, the system will add a new record in the association table, including the partition ID "PART-AB", the update time stamp "2024-12-07 14:30:00" and the update field "Impedance". This recording mechanism enables the system to clearly track the data change history of each partition, providing the necessary basis for subsequent synchronization operations.

[0160] After the data is updated, the system will perform a partial data refresh based on the records in the associated table. Partial data refresh means that only the partition data that has changed is synchronized without fully refreshing all partitions. For example, if the line (A, B) of the partition "PART-AB" is updated, the system will only synchronize the data of this partition after checking the associated table to avoid unnecessary operations on irrelevant data in other partitions. The synchronization process is completed through a consistency protocol (such as a two-phase commit protocol or a consistent hashing protocol) to ensure that the data in all storage nodes remains consistent after the update. After the synchronization is complete, the system will write the latest version of the information back to the associated table, such as updating the version number of the partition "PART-AB" from "VERSION-003" to "VERSION-004".

[0161] After synchronization is complete, the system will adjust the association table accordingly based on the record information of the update event. For example, if the data of a partition is reallocated to a new storage node, or some data blocks are split into multiple sub-partitions, the association table needs to update the mapping information of the storage location and partition structure. Through this dynamic adjustment, the association table always keeps the latest partition structure and storage location records, thereby supporting the subsequent query steps to accurately and quickly locate the latest data.

[0162] This implementation not only provides precise management and efficient synchronization of partitioned data, but also has good scalability. For example, when the storage system is expanded to more nodes, the new nodes can be quickly integrated into the existing distributed architecture by dynamically adjusting the association table. In addition, the system can reduce the storage usage of the association table and improve access efficiency by regularly cleaning up outdated association records.

[0163] Through the above steps, this embodiment provides an accurate and dynamic management mechanism for the partitioned storage and synchronization of hierarchical data, ensuring the efficiency and consistency of the power transmission information model in a large-scale distributed environment, while providing stable support for subsequent query and update operations.

[0164] Step S104: providing standardized interface services, including a model query interface, a model update interface, and an analysis result export interface.

[0165] Step S104 involves providing standardized interface services, including model query interface, model update interface and analysis result export interface. These interface services are intended to establish a unified data interaction method between users and the system to achieve efficient query, dynamic update and result export.

[0166] First, when implementing the model query interface, it is necessary to design a flexible query function based on the hierarchical structure of the transmission information model. The query interface supports on-demand access to data in the static topology layer, dynamic operating status layer, and predictive analysis layer. The user sends a query request to the system by specifying query parameters, such as node identification, line number, time range, or topological area. The query interface quickly locates the data node of the target partition by accessing the global index table, retrieves the corresponding data, and returns it to the user in a predefined format. For example, when the user requests the voltage and power data of node A, the query interface finds the partition location that stores the relevant data of node A from the global index table, directly accesses the partition, and returns the time series data of node A. The query results can be organized according to time series or topological structure so that users can intuitively understand the operating status of the transmission network.

[0167] Secondly, the model update interface allows users to dynamically modify and expand the data at each layer of the transmission information model. Users submit modification requests through the update interface, such as changing the impedance value of the line (A, B) or adding a new transformer node. After receiving the request, the system first verifies the legitimacy of the update operation, such as ensuring that the electrical parameters of the new node meet the system specifications, or that the connection relationship of the new line does not destroy the integrity of the topology. After the verification is passed, the update interface calls the distributed storage system to write the modified data to the corresponding partition and synchronize it to other storage nodes. At the same time, the update interface will also modify the global index table to record the timestamp and version number of the data change to ensure that subsequent query operations can access the latest data version.

[0168] Finally, the analysis result export interface is used to provide access to the output data of the prediction analysis layer. Users can specify the time range, node or line range, and output format through the export interface, and the system extracts the corresponding result data from the prediction analysis layer based on these parameters. For example, if the user wants to export the load forecast data for the past 24 hours and the power flow forecast for the next 6 hours, the export interface will summarize the data from the relevant partitions according to the user's needs and organize it in the form of tables, charts, or visual reports. The export process also supports result compression and segmented downloading to improve data transmission efficiency.

[0169] To ensure the universality and scalability of the interface, all interfaces follow a unified standardized design. The data format of requests and responses uses a structured markup language (such as JSON or XML) to support data interoperability between different platforms. In addition, the interface service includes access control functions. Users need to pass identity authentication before they can query, update, or export operations to ensure data security and system stability.

[0170] Through the collaborative work of the model query interface, the model update interface and the analysis result export interface, step S104 realizes efficient interaction between the transmission information model and the user system, providing flexible and convenient tool support for real-time monitoring, dynamic adjustment and predictive analysis of the transmission network.

[0171] Furthermore, the provision of standardized interface services includes:

[0172] Prepare a parameter list for each interface request type, registering the field names, data types and search paths required for query, update and export in the list;

[0173] After receiving the external request, the corresponding partition data or model record is located from the global index according to the matching request parameters of the list;

[0174] After the data is located, the search results are assembled into a response data packet in a unified format and returned to the caller according to the parameter list requirements of the request type.

[0175] This embodiment involves providing standardized interface services. By compiling a parameter list for the interface request type, matching the request parameters and locating the target data, and returning the response data packet in a unified format, an efficient and flexible data interaction mechanism is constructed, which is suitable for the query, update and result export requirements of the transmission information model.

[0176] When implementing standardized interface services, it is first necessary to compile a parameter list for each type of interface request. The parameter list is a structured configuration file that defines the field names, data types, and data retrieval paths required for each request. For example, for a model query request list, you can define the node identifier, time range, and return fields required for the query, such as node voltage or line power; the update request list may include the partition identifier of the update target, the name of the modified field, and the corresponding new value; the export request list will include the export time span, target data type (such as predictive analysis results or dynamic operation status), and the required output format. Each parameter in the list can also be accompanied by constraints, such as the node identifier must be a unique identifier and the time range must be expressed in the form of a timestamp. This predefined list not only standardizes the input structure of the interface, but also reduces ambiguity in request processing and improves the automation of request processing.

[0177] When the system receives an external request, it automatically extracts the defined field information from the parameter list and matches it with the parameters submitted in the request. For example, if the query request contains the node identifier "Node-A" and the time range "2024-12-07 00:00:00 to 2024-12-07 12:00:00", the system verifies whether these fields are complete and in the expected format according to the parameter list. Once the parameter matching is completed, the system calls the global index table to locate the partition data or model record related to the request. The global index table records the storage location, data version, and update status of each partition. For example, partition "PART-A" may contain the voltage and power records of node "Node-A" and is stored in the specified storage node. By matching the request parameters with the contents of the index table, the system can quickly find the location of the target data.

[0178] After the data is located, the system retrieves the data of the relevant partitions according to the requirements of the request type and assembles the results into a response data packet in a unified format. The format of the response data packet is defined by the parameter list. For example, the result of a query request may be organized in a table with fields including timestamp, node voltage, and line power; the response to an update request may contain the updated confirmation status and the modified new version number; the result of an export request may be a compressed file containing the complete time series of the target data and related metadata. The response data packet also supports multiple transmission protocols, such as HTTP or File Transfer Protocol (FTP), to meet the integration requirements of different user systems.

[0179] This mechanism is designed to be extensible. For example, more types of interface requests can be supported by adding new parameter list entries, such as a visual export interface for predictive analysis results or a real-time monitoring interface for dynamic operating status. In addition, the parameter list can be dynamically updated to adapt to changes in the transmission information model, such as adding new fields or adjusting the search path.

[0180] By compiling parameter lists for interface request types, matching request parameters and locating data, and returning response data packets in a unified format, the standardized interface service not only achieves efficient data interaction, but also provides a flexible and reliable solution for querying, updating, and exporting transmission information models. This mechanism ensures the smooth interaction between users and the system, while ensuring the standardization and scalability of data access, providing important support for real-time monitoring and analysis of transmission networks.

[0181] Furthermore, the provision of standardized interface services includes:

[0182] Create a configuration file for the analysis result export interface, which defines the time span, node range and line selection rules of the output data;

[0183] When receiving an export request, determine the content to be extracted from the prediction analysis layer or the dynamic operation status layer according to the configuration file, and retrieve the corresponding partition data;

[0184] The retrieved data is packaged into a recognizable export data format according to the order of fields listed in the configuration file, and the data format is output to the caller as a unified result.

[0185] The analysis result export interface of this embodiment realizes a flexible and standardized analysis result export process by establishing a configuration file, extracting data from the prediction analysis layer or the dynamic operation status layer, and organizing the data into a standardized format, which can adapt to the diverse application requirements of the transmission network.

[0186] In order to achieve efficient operation of the export interface, it is first necessary to establish a configuration file for the analysis result export interface. The configuration file is a structured document that defines the key parameters and rules of the export request. Specifically, the configuration file consists of three main parts: time span, node range, and line selection rules. The time span clarifies the time interval of the required data, for example, it can be set to the forecast period of the past 24 hours or the next 48 hours; the node range identifies the set of nodes involved in the exported data, such as specifying a substation and its connected main load nodes; the line selection rules further refine the scope of the exported data, such as selecting lines with voltage levels higher than 220 kV or only exporting lines with power fluctuations exceeding a threshold. This rule-based configuration file allows users to clearly specify the required data range and dimensions in the export request, improving the pertinence and efficiency of data extraction.

[0187] When the system receives an export request, the export interface parses the request content and determines the specific data that needs to be extracted from the model based on the corresponding configuration file. The extraction process involves accessing partitioned data of the predictive analysis layer or the dynamic operating status layer. For the predictive analysis layer, the system retrieves prediction results that match the time span and node range, such as trend curves of node load changes or timing prediction data of line power; for the dynamic operating status layer, the system extracts real-time operating parameters, such as actual measurements of node voltage and line power. To ensure the speed of the retrieval process, the export interface relies on the global index table to locate the data nodes of the target partition, and directly accesses the storage location to extract the required content after locating it.

[0188] After the retrieved raw data is extracted, the system will organize and package the data according to the field order defined in the configuration file. The organization process includes data filtering, formatting and merging. For example, if the configuration file requires the changes in node voltage and load to be exported in chronological order, the system will first merge the data of different partitions into a unified time series structure and arrange them according to the field order specified in the configuration file (such as "timestamp-node voltage-load"). During the packaging process, the system will generate output files according to the preset export data format, such as table files (such as CSV or Excel), visual charts (such as load trend charts in PDF format) or data exchange formats (such as JSON or XML). The packaged output data not only meets the requirements of the configuration file, but also has good readability and cross-platform compatibility.

[0189] After completing the data organization, the export interface transmits the generated output data to the caller. This process can be completed in a variety of ways, such as directly returning the result file through the HTTP protocol, or storing the result on a specified server for users to download through the File Transfer Protocol (FTP). For requests with larger data volumes, the system supports segmented export and compressed transmission to improve export efficiency and reduce network bandwidth usage.

[0190] This design is also highly flexible and extensible. The configuration file can be dynamically adjusted according to different application scenarios, such as adding export rules for specific lines or expanding the selection conditions for the node range. The system also supports the management and switching of multiple configuration files to meet the needs of different users or applications. In addition, the export interface can be combined with a data permission control mechanism to ensure that only authorized users can access a specific range of analysis results.

[0191] By establishing a configuration file for the export interface, standardizing data extraction rules, and outputting data in a unified format, the present invention achieves efficient export of analysis results, which can not only adapt to complex power transmission network scenarios, but also provide users with flexible, accurate and reliable data services. This export mechanism provides important support for the operation monitoring, trend prediction and fault analysis of the power transmission network, and significantly improves the practicality and ease of use of the system.

[0192] In the above-mentioned embodiment, a power transmission information model storage medium is provided, and correspondingly, the present application also provides a power transmission information model construction method. Since this embodiment, i.e., the second embodiment, is basically similar to the storage medium embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the storage medium embodiment. The system embodiment described below is merely illustrative.

[0193] A second embodiment of the present application provides a method for constructing a power transmission information model, including:

[0194] Collecting and preprocessing the operation data of the power transmission network, wherein the real-time operation data includes node voltage, line power, load distribution and equipment operation status; the preprocessing includes formatting the data and filtering out exceptions;

[0195] According to the preprocessed operation data, a transmission information model is constructed, and the transmission information model includes a static topology layer, a dynamic operation status layer and a prediction and analysis layer; wherein the static topology layer is used to describe the basic connection structure of the transmission network; the dynamic operation status layer updates the real-time parameters based on the connection relationship provided by the static topology layer to reflect the changes in electrical parameters; the prediction and analysis layer combines the time series data of the dynamic operation status layer and the historical operation records to predict future transmission load changes;

[0196] Based on a distributed storage architecture, the hierarchical data of the power transmission information model is partitioned and stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, while a global index is generated to support fast query;

[0197] Provide standardized interface services, including model query interface, model update interface and analysis result export interface.

[0198] Although the present application is disclosed as above in the form of a preferred embodiment, it is not intended to limit the present application. Any technical personnel in this field may make possible changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be based on the scope defined by the claims of the present application.

Claims

1. A power transmission information model storage medium, characterized in that: The storage medium stores computer-readable instructions, which, when executed on a processor, are used to implement a method for constructing a power transmission information model. The method for constructing a power transmission information model includes: Collecting and preprocessing the operation data of the power transmission network, wherein the real-time operation data includes node voltage, line power, load distribution and equipment operation status; the preprocessing includes formatting the data and filtering out exceptions; According to the preprocessed operation data, a transmission information model is constructed, and the transmission information model includes a static topology layer, a dynamic operation status layer and a prediction and analysis layer; wherein the static topology layer is used to describe the basic connection structure of the transmission network; the dynamic operation status layer updates the real-time parameters based on the connection relationship provided by the static topology layer to reflect the changes in electrical parameters; the prediction and analysis layer combines the time series data of the dynamic operation status layer and the historical operation records to predict future transmission load changes; Based on a distributed storage architecture, the hierarchical data of the power transmission information model is partitioned and stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, while a global index is generated to support fast query; Provide standardized interface services, including model query interface, model update interface and analysis result export interface.

2. The power transmission information model storage medium according to claim 1, characterized in that: The collecting and preprocessing of the operation data of the power transmission network includes: Create a time segment tag for each batch of newly acquired operation data, and classify and store the node and line records within the same time segment; After classification, identification codes are added to data items that span multiple time segments and whose parameters fluctuate continuously to distinguish the special status of these data items in subsequent analysis; The data structure with the identification code attached is output to a unified pre-processing index list, so that the subsequent steps can track the timing information of specific lines and nodes based on the identification code.

3. The power transmission information model storage medium according to claim 1, characterized in that: Based on the distributed storage architecture, the hierarchical data of the power transmission information model is partitioned and stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, including: According to the association relationship between nodes and lines in the static topology layer, the hierarchical data of the power transmission information model is divided into multiple logical partitions and a unique identifier is assigned to each partition; A version record is created for each partition that has been assigned an identifier. When the data content of any partition changes, the change information is appended in the form of a new version record; When a partition update request is received, only the partition data with differences is synchronized according to the change information in the version record, and the version status of the corresponding partition in the global index is updated after the synchronization is completed.

4. The power transmission information model storage medium according to claim 3, characterized in that: Based on the distributed storage architecture, the hierarchical data of the power transmission information model is partitioned and stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, including: Create an association table for each partition data, and map the node set and line set corresponding to each partition to a specific data storage location; When data is updated, the time stamp and partition identifier of the update event are recorded in the associated table so that the subsequent synchronization step can perform local data refresh for the partition; After completing the partial data refresh, the associated table is adjusted accordingly based on the record information of the update event so that the subsequent query step can locate the latest data source through the table.

5. The power transmission information model storage medium according to claim 1, characterized in that: The standardized interface service provided includes: Prepare a parameter list for each interface request type, registering the field names, data types and search paths required for query, update and export in the list; After receiving the external request, the corresponding partition data or model record is located from the global index according to the matching request parameters of the list; After the data is located, the search results are assembled into a response data packet in a unified format and returned to the caller according to the parameter list requirements of the request type.

6. The power transmission information model storage medium according to claim 1, characterized in that: The standardized interface service provided includes: Create a configuration file for the analysis result export interface, which defines the time span, node range and line selection rules of the output data; When receiving an export request, determine the content to be extracted from the prediction analysis layer or the dynamic operation status layer according to the configuration file, and retrieve the corresponding partition data; The retrieved data is packaged into a recognizable export data format according to the order of fields listed in the configuration file, and the data format is output to the caller as a unified result.

7. The power transmission information model storage medium according to claim 1, characterized in that: The construction of the static topology layer includes weighting the initial adjacency matrix representing the connection relationship between the nodes and the lines of the power transmission network, so that each line not only reflects whether there is a connection, but also reflects the stability of the line in long-term operation and the sensitivity to load changes through the weight; The weighted value W(i,j) of the line (i,j) is calculated according to the following formula (1): W(i,j)=α·ln(1+R(i,j))+β·[L(i,j)] μ (1) Wherein, R(i,j) is the line reliability index of line (i,j); L(i,j) is the line sensitivity index of line (i,j); α, β and μ are positive real number parameters; the line reliability index R(i,j) is calculated according to the following formula (2): Where M is the number of historical observation periods; f m is the proportion of fault-free operation time of line (i, j) in the mth cycle; δ is a small positive number used to avoid zero value; γ is the nonlinear amplification factor; The line sensitivity index L(i,j) is calculated according to the following formula 3: Where T is the number of load data sampling points; P i,j (t) is the measured load value of line (i, j) at time t; P ref (t) is the reference load value of line (i, j) at time t; δ is the adjustment factor; ω t is the time adjustment factor, calculated according to the following formula 4: Among them, λ is the adjustment factor.

8. The power transmission information model storage medium according to claim 7, characterized in that: The transmission load forecasting method of the forecasting and analysis layer embeds the line weight information formed by the static topology layer into the training and calculation process of the forecasting model based on the long short-term memory network to improve the accuracy of the load forecast of a specific line and the pertinence of the dispatching decision. Specifically, it includes: The multi-time series electrical parameter data output by the dynamic operation state layer and the R(i,j) and L(i,j) parameters corresponding to each line (i,j) calculated in the static topology layer and their weighted values ​​W(i,j) are integrated into an input feature vector, and the feature vector is numerically normalized; The data samples corresponding to each line in the input feature vector are sorted in descending order according to the W(i,j) value, and the sorting number r represents the position of the line (i,j) in the sequence, and in the process of updating the hidden layer state and output calculation of the long short-term memory network model, the input data of the line with the sorting number r is assigned the influence factor G(r), so that the data samples corresponding to the line with a larger W(i,j) value obtain a numerically higher influence factor G(r) in the state calculation, thereby realizing the weight enhancement of the characteristics of this type of line in time series processing; The impact factor G(r) is calculated using the following formula 5: where c is a predefined positive real constant.

9. The power transmission information model storage medium according to claim 7, characterized in that: The real-time parameter update of the dynamic operation state layer includes updating the real-time parameters of the power transmission network through a Markov state transition model to reflect the immediate changes and abnormal fluctuations of the electrical parameters, specifically including: Using the line weight information determined by the static topology layer, the transfer probability between each node and line in the Markov state transition model is dynamically corrected, so that the key lines with higher weights have more influence on the update of node voltage, line power and load distribution during the state transition process; In the process of continuously receiving real-time operation data, when it is observed that key parameters such as node voltage or line power deviate from their expected range, the corresponding state calibration step is triggered to timely suppress abnormal states that may spread along the key lines by redistributing transition probabilities; After the operating parameters return to stability, the updated state transfer relationship is used as the basic information for subsequent time steps, so that the dynamic operating status layer can achieve refined management and continuous optimization of the evolution trend of electrical parameters without relying on multiple prediction models.

10. A method for constructing a power transmission information model, characterized in that: include: Collecting and preprocessing the operation data of the power transmission network, wherein the real-time operation data includes node voltage, line power, load distribution and equipment operation status; the preprocessing includes formatting the data and filtering out exceptions; According to the preprocessed operation data, a transmission information model is constructed, and the transmission information model includes a static topology layer, a dynamic operation status layer and a prediction and analysis layer; wherein the static topology layer is used to describe the basic connection structure of the transmission network; the dynamic operation status layer updates the real-time parameters based on the connection relationship provided by the static topology layer to reflect the changes in electrical parameters; the prediction and analysis layer combines the time series data of the dynamic operation status layer and the historical operation records to predict future transmission load changes; Based on a distributed storage architecture, the hierarchical data of the power transmission information model is partitioned and stored in multiple nodes, and incremental synchronization is achieved through a consistency protocol, while a global index is generated to support fast query; Provide standardized interface services, including model query interface, model update interface and analysis result export interface.

Images