Full-process data monitoring method for power distribution network

By collecting and analyzing operating status characteristics in the distribution network and identifying abnormal conditions using machine learning models, the problems of incomplete equipment and insufficient real-time performance are solved, and the data monitoring accuracy and system security of the distribution network are improved.

CN118944286AActive Publication Date: 2024-11-12STATE GRID FUJIAN ELECTRIC POWER CO LTD +1
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202410984914.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-12
Estimated Expiration
2044-07-22

AI Technical Summary

Technical Problem

Distribution network data monitoring has problems such as incomplete equipment, insufficient real-time and low intelligence, which affects the stable operation of the power system and the power supply quality.

Method used

The full-process data monitoring method of the distribution network is adopted, and the node's operating status characteristics are collected through sensors, a data monitoring model is established, and the machine learning model is used to obtain correlation characteristics, divide the monitoring sub-data sets, and abnormal conditions are identified through the abnormal detection model.

Benefits of technology

It improves the accuracy and real-time data monitoring of the distribution network, enhances the reliability and safety of the system, and can promptly detect and resolve abnormal faults.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118944286A_ABST
    Figure CN118944286A_ABST
Patent Text Reader

Abstract

The invention relates to a power distribution network full-process data monitoring method, which comprises the following steps of: acquiring operation state characteristics of each node through a sensor according to a topological relation among the nodes of a power distribution network to obtain a power distribution network data monitoring model; obtaining a monitoring data set according to the power distribution network data monitoring model, and determining the types and the number of the operation state features; inputting the monitoring data set and the different types of monitoring target parameters into a machine learning model, and obtaining an associated feature of each monitoring target parameter; dividing the monitoring data set into a plurality of monitoring sub-data sets based on the associated features of each monitoring target parameter; initiating a monitoring request for any monitoring target parameter, acquiring a sub-data set corresponding to the monitoring target parameter, and acquiring a plurality of monitoring target time sequence data sets based on the monitoring time interval and each associated feature; and inputting the multiple monitoring target time series data sets into the anomaly detection model to identify whether the monitoring parameters are abnormal or not.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a method for monitoring data of a distribution network in its entire process, and belongs to the technical field of data monitoring of a distribution network. Background Art

[0002] With the rapid development of smart grids, the accuracy and real-time requirements for distribution network data monitoring are becoming higher and higher. However, in practical applications, distribution network data monitoring has many defects, which affect the stable operation and power supply quality of the power system.

[0003] First, the imperfection of distribution network data monitoring equipment is one of the main reasons for the defects. The distribution network monitoring equipment in some areas is outdated and has a single function, which cannot meet the needs of the current power system. These devices can often only monitor limited power parameters and cannot fully reflect the operating status of the distribution network. At the same time, due to reasons such as aging equipment and improper maintenance, the accuracy and reliability of monitoring data are also greatly reduced.

[0004] Secondly, the lack of real-time data monitoring is also a major drawback of distribution network data monitoring. In the power system, the operating status of the distribution network changes in real time, and real-time monitoring and processing of data is crucial. However, existing data monitoring systems often have problems such as data transmission delays and slow processing speeds, resulting in the monitoring data being unable to reflect the actual status of the distribution network in a timely manner. This may lead to untimely fault detection and increase the safety risks of the power system.

[0005] In addition, the low level of intelligence in distribution network data monitoring is also a major problem. With the continuous development of artificial intelligence technology, distribution network data monitoring should gradually become intelligent to improve the efficiency and accuracy of data processing and analysis. However, most distribution network data monitoring systems are still at the stage of simple data collection and display, lacking intelligent analysis and prediction functions. This makes it difficult for operation and maintenance personnel to extract valuable information from massive data, affecting the optimized operation of the power system. Summary of the invention

[0006] In order to solve the above problems existing in the prior art, the present invention proposes a method for monitoring the whole process data of a distribution network.

[0007] The technical solution of the present invention is as follows:

[0008] On the one hand, the present invention proposes a method for monitoring data of the entire distribution network, comprising the following steps:

[0009] According to the topological relationship between the nodes of the distribution network and the operation status characteristics of each node collected by sensors, a distribution network data monitoring model is obtained;

[0010] According to the distribution network data monitoring model, obtain the monitoring data set and determine the type and quantity of operating status characteristics;

[0011] Inputting the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain the associated features with each monitoring target parameter, wherein the associated features are operating status features associated with the monitoring target parameters;

[0012] Based on the associated features of each monitoring target parameter, the monitoring data set is divided into a plurality of monitoring sub-data sets, each sub-data set corresponding to a monitoring target parameter and a corresponding associated feature;

[0013] Initiate a monitoring request for any monitoring target parameter, obtain the sub-dataset of the corresponding monitoring target parameter, and obtain multiple monitoring target time series data sets based on the monitoring time interval and various associated features; input multiple monitoring target time series data sets into the anomaly detection model to identify whether there are anomalies in the monitoring parameters.

[0014] As a preferred implementation, the steps of obtaining the distribution network data monitoring model according to the topological relationship between the nodes of the distribution network and collecting the operating status characteristics of each node through sensors are specifically as follows:

[0015] Build the state model:

[0016] F(QN)=P(Q1)·P(Q2 / Q1)·P(Q3 / Q2)·…·P(Qn / Qn-1);

[0017] Among them, F(QN) represents the state model of the set of nodes in the distribution network, Qn represents the state of the nth node in the distribution network; P(Qn / Qn-1) represents the conditional probability distribution between the n-1th node in the distribution network and the nth node in the distribution network;

[0018] And there exists the following relationship between the state Qn of the nth node and the state Qn-1 of the n-1th node in the distribution network:

[0019] Qn=Qn-1+s;

[0020] Among them, s is the state noise that conforms to the Gaussian distribution;

[0021] Then P(Qn / Qn-1)=G(Qn-1,σ 2 );G(·,σ 2 ) represents Gaussian distribution;

[0022] Build a sampling model:

[0023] F(xn / Qn)=G(J(xn),σ 2 );

[0024] Among them, F(xn / Qn) is the probability distribution of the sampling value of the distribution network node in state Qn, xn is the sensor sampling value of the nth node, and J(xn) is the sampling function;

[0025] And J(xn)=xn+τ;

[0026] Among them, τ is the sampling noise that conforms to the Gaussian distribution;

[0027] The distribution network data monitoring model is obtained by combining the state model and the sampling model:

[0028] Y(QN,xN)=P(Q1)F(x1 / Q1)·P(Q2 / Q1)F(x2 / Q2)·P(Q3 / Q2)F(x3 / Q3)·…·P(Qn / Qn-1)F(xn / Qn).

[0029] As a preferred implementation, the step of inputting the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain the associated features with each monitoring target parameter specifically includes:

[0030] Preprocess the monitoring data set, including data cleaning and normalization;

[0031] Based on the preprocessed monitoring data set, a training sample set is established, and each training sample in the training sample set includes an operating state feature and a corresponding monitoring target parameter label;

[0032] Construct a machine learning model based on a neural network; iteratively train the machine learning model using a training sample set to obtain a trained correlation feature classification model;

[0033] The monitoring data set and different categories of monitoring target parameters are input into the trained associated feature classification model, and the associated features corresponding to each input monitoring target parameter are output.

[0034] As a preferred implementation manner, the operating status characteristics include voltage, current, power factor, load rate, temperature and humidity.

[0035] On the other hand, the present invention also proposes a distribution network full-process data monitoring system, comprising:

[0036] The monitoring model building module is used to obtain the distribution network data monitoring model based on the topological relationship between the nodes of the distribution network and the operating status characteristics of each node collected by sensors;

[0037] A feature classification module is used to obtain a monitoring data set and determine the type and quantity of operating status features according to the distribution network data monitoring model;

[0038] An associated feature acquisition module is used to input the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain associated features with each monitoring target parameter, wherein the associated features are operating status features associated with the monitoring target parameters;

[0039] A data set division module, which divides the monitoring data set into multiple monitoring sub-data sets based on the associated features of each monitoring target parameter, each sub-data set corresponding to a monitoring target parameter and the corresponding associated features;

[0040] The anomaly diagnosis module is used to initiate a monitoring request for any monitoring target parameter, obtain the sub-dataset of the corresponding monitoring target parameter, and obtain multiple monitoring target time series data sets based on the monitoring time interval and various associated features; the multiple monitoring target time series data sets are input into the anomaly detection model to identify whether there are anomalies in the monitoring parameters.

[0041] As a preferred implementation, the steps of obtaining the distribution network data monitoring model according to the topological relationship between the nodes of the distribution network and collecting the operating status characteristics of each node through sensors are specifically as follows:

[0042] Build the state model:

[0043] F(QN)=P(Q1)·P(Q2 / Q1)·P(Q3 / Q2)·…·P(Qn / Qn-1);

[0044] Among them, F(QN) represents the state model of the set of nodes in the distribution network, Qn represents the state of the nth node in the distribution network; P(Qn / Qn-1) represents the conditional probability distribution between the n-1th node in the distribution network and the nth node in the distribution network;

[0045] And there exists the following relationship between the state Qn of the nth node and the state Qn-1 of the n-1th node in the distribution network:

[0046] Qn=Qn-1+s;

[0047] Among them, s is the state noise that conforms to the Gaussian distribution;

[0048] Then P(Qn / Qn-1)=G(Qn-1,σ 2 );G(·,σ 2 ) represents Gaussian distribution;

[0049] Build a sampling model:

[0050] F(xn / Qn)=G(J(xn),σ 2 );

[0051] Among them, F(xn / Qn) is the probability distribution of the sampling value of the distribution network node in state Qn, xn is the sensor sampling value of the nth node, and J(xn) is the sampling function;

[0052] And J(xn)=xn+τ;

[0053] Among them, τ is the sampling noise that conforms to the Gaussian distribution;

[0054] The distribution network data monitoring model is obtained by combining the state model and the sampling model:

[0055] Y(QN,xN)=P(Q1)F(x1 / Q1)·P(Q2 / Q1)F(x2 / Q2)·P(Q3 / Q2)F(x3 / Q3)·…·P(Qn / Qn-1)F(xn / Qn).

[0056] As a preferred implementation, the step of inputting the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain the associated features with each monitoring target parameter specifically includes:

[0057] Preprocess the monitoring data set, including data cleaning and normalization;

[0058] Based on the preprocessed monitoring data set, a training sample set is established, and each training sample in the training sample set includes an operating state feature and a corresponding monitoring target parameter label;

[0059] Construct a machine learning model based on a neural network; iteratively train the machine learning model using a training sample set to obtain a trained correlation feature classification model;

[0060] The monitoring data set and different categories of monitoring target parameters are input into the trained associated feature classification model, and the associated features corresponding to each input monitoring target parameter are output.

[0061] As a preferred implementation manner, the operating status characteristics include voltage, current, power factor, load rate, temperature and humidity.

[0062] On the other hand, the present invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for monitoring the full-process data of a distribution network as described in any embodiment of the present invention is implemented.

[0063] On the other hand, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for monitoring the full-process data of a distribution network as described in any embodiment of the present invention.

[0064] The present invention has the following beneficial effects:

[0065] The present invention provides a distribution network full-process data monitoring method, which establishes a distribution network data monitoring model through topological relationships and data collected by sensors; constructs a monitoring data set through the distribution network data monitoring model, and classifies the collected features; uses a machine learning model to obtain the associated features of the monitoring target parameters to divide each monitoring target parameter into a sub-data set; finally, uses an anomaly detection model and each monitoring target parameter sub-data set to identify abnormal conditions and improve the reliability and safety of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 The figure is a schematic diagram of the method flow of the first embodiment of the present invention. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] It should be understood that the step numbers used in this document are only for convenience of description and are not intended to limit the order in which the steps are executed.

[0069] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0070] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0071] The term "and / or" means and includes any and all possible combinations of one or more of the associated listed items.

[0072] Embodiment 1:

[0073] See also Figure 1In the power system, the distribution network is a key link in power transmission and distribution, and its stable operation is crucial to ensuring the reliability of the entire power supply. In order to achieve comprehensive monitoring and precise management of the operating status of the distribution network, this embodiment provides a distribution network full-process data monitoring method based on topological relationships and machine learning technology, which specifically includes the following steps:

[0074] S100. Based on the topological relationship between the nodes of the distribution network, the structure and connection mode of the entire distribution network are clearly understood, and the operating status characteristics of each node are collected in real time through sensors, including key indicators such as voltage, current, and power factor. The distribution network data monitoring model is established by integrating the node topological relationship and operating status characteristics.

[0075] S200. According to the distribution network data monitoring model, a monitoring data set is obtained, and the types and quantities of operating status characteristics are determined; the operating status of the distribution network can be understood, and basic data resources can be provided for subsequent machine learning model training.

[0076] S300, inputting the monitoring data set and different types of monitoring target parameters into the machine learning model, and obtaining the associated features with each monitoring target parameter, wherein the associated features are operating state features associated with the monitoring target parameters;

[0077] S400, dividing the monitoring data set into a plurality of monitoring sub-data sets based on the associated features of each monitoring target parameter, each sub-data set corresponding to a monitoring target parameter and a corresponding associated feature;

[0078] S500. Initiate a monitoring request for any monitoring target parameter to quickly obtain the sub-dataset of the corresponding monitoring target parameter. Then, according to the set monitoring time interval and associated features, we can construct multiple monitoring target time series data sets. These time series data sets can intuitively show the changes of monitoring target parameters over time; input multiple monitoring target time series data sets into the anomaly detection model to identify whether there are anomalies in the monitoring parameters.

[0079] Based on the above implementation methods, the method provided in this embodiment can realize the full-process monitoring of the operation status of the distribution network through topological relationships and machine learning technology. By deeply analyzing the operation status characteristics, building monitoring sub-datasets, and using anomaly detection models, abnormal faults can be discovered and resolved in a timely manner, thereby improving the reliability and safety of the distribution network.

[0080] As a preferred implementation of this embodiment, the steps of obtaining the distribution network data monitoring model according to the topological relationship between the nodes of the distribution network and collecting the operating status characteristics of each node through sensors are specifically as follows:

[0081] Build the state model:

[0082] F(QN)=P(Q1)·P(Q2 / Q1)·P(Q3 / Q2)·…·P(Qn / Qn-1);

[0083] Among them, F(QN) represents the state model of the set of nodes in the distribution network, Qn represents the state of the nth node in the distribution network; P(Qn / Qn-1) represents the conditional probability distribution between the n-1th node in the distribution network and the nth node in the distribution network;

[0084] And there exists the following relationship between the state Qn of the nth node and the state Qn-1 of the n-1th node in the distribution network:

[0085] Qn=Qn-1+s;

[0086] Among them, s is the state noise that conforms to the Gaussian distribution;

[0087] Then P(Qn / Qn-1)=G(Qn-1,σ 2 );G(·,σ 2 ) represents Gaussian distribution;

[0088] Build a sampling model:

[0089] F(xn / Qn)=G(J(xn),σ 2 );

[0090] Among them, F(xn / Qn) is the probability distribution of the sampling value of the distribution network node in state Qn, xn is the sensor sampling value of the nth node, and J(xn) is the sampling function;

[0091] And J(xn)=xn+τ;

[0092] Among them, τ is the sampling noise that conforms to the Gaussian distribution;

[0093] The distribution network data monitoring model is obtained by combining the state model and the sampling model:

[0094] Y(QN,xN)=P(Q1)F(x1 / Q1)·P(Q2 / Q1)F(x2 / Q2)·P(Q3 / Q2)F(x3 / Q3)·…·P(Qn / Qn-1)F(xn / Qn).

[0095] As a preferred implementation of this embodiment, the step of inputting the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain the associated features with each monitoring target parameter specifically includes:

[0096] The monitoring data set is preprocessed, including data cleaning and normalization; the purpose is to eliminate noise, errors and redundant information in the original data, and enable data of different dimensions to be compared and calculated on the same scale.

[0097] Data cleaning involves checking for missing values, outliers, duplicate values, and other issues in the data, and taking appropriate action. For missing values, interpolation, mean filling, median filling, and other methods can be used to fill them; for outliers, it is necessary to determine whether to remove or modify them based on the data distribution and business background; for duplicate values, deduplication is performed directly.

[0098] Since data of different features may have different dimensions and units, if directly used for training a machine learning model, the model may not be able to correctly learn the inherent laws and relationships of the data. Therefore, this embodiment normalizes the data and scales the data of each feature to the same range, such as [0, 1] or [-1, 1], thereby eliminating the impact of dimensions and units on model training.

[0099] Based on the preprocessed monitoring data set, a training sample set is established. Each training sample in the training sample set contains an operating state feature and a corresponding monitoring target parameter label; the monitoring target parameter label is obtained by feature extraction and labeling of the monitoring data set.

[0100] Construct a machine learning model based on a neural network; use the training sample set to iteratively train the machine learning model, and continuously adjust the parameters and structure of the model to minimize the prediction error and improve the generalization ability of the model, so as to obtain a trained correlation feature classification model;

[0101] The monitoring data set and different categories of monitoring target parameters are input into the trained associated feature classification model, and the associated features corresponding to each input monitoring target parameter are output.

[0102] As a preferred implementation of this embodiment, the operating status characteristics include voltage, current, power factor, load rate, temperature and humidity. Voltage and current are important indicators of the working status of power equipment in the node. The stability of voltage is directly related to whether the equipment can work normally. If the voltage is too high or too low, it may damage the equipment and even cause failure. The size and change of current can reflect the load condition of the equipment and reveal the power consumption of the equipment.

[0103] The higher the power factor, the higher the power utilization rate of the equipment. Conversely, it means that there is a lot of reactive power loss. Therefore, by monitoring the power factor, the energy efficiency operation status of the equipment can be discovered in time.

[0104] The load rate indicates the ratio between the actual load borne by the equipment and the rated load. The load rate directly reflects the utilization of the equipment. Too high or too low a load rate may have an adverse effect on the equipment.

[0105] Temperature and humidity are important environmental factors that affect the operating status of equipment. Equipment generates heat during operation. If the heat dissipation is poor or the ambient temperature is too high, the temperature of the equipment may rise, thus affecting its normal operation. At the same time, high humidity may also cause condensation or corrosion inside the equipment.

[0106] Embodiment 2:

[0107] This embodiment provides a distribution network full-process data monitoring system, including:

[0108] A monitoring model building module is used to obtain a distribution network data monitoring model based on the topological relationship between the nodes of the distribution network and the operating status characteristics of each node collected by sensors; this module is used to implement the function of step S100 in the first embodiment, which will not be repeated here;

[0109] A feature classification module is used to obtain a monitoring data set according to the distribution network data monitoring model, and determine the type and quantity of the operating status features; this module is used to implement the function of step S200 in the first embodiment, and will not be repeated here;

[0110] An associated feature acquisition module is used to input the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain associated features with each monitoring target parameter, wherein the associated features are operating state features associated with the monitoring target parameters; this module is used to implement the function of step S300 in the first embodiment, and will not be described in detail here;

[0111] A data set division module, which divides the monitoring data set into a plurality of monitoring sub-data sets based on the associated features of each monitoring target parameter, each sub-data set corresponding to a monitoring target parameter and a corresponding associated feature; this module is used to implement the function of step S400 in the first embodiment, and will not be described in detail here;

[0112] The abnormal diagnosis module is used to initiate a monitoring request for any monitoring target parameter, obtain a sub-dataset of the corresponding monitoring target parameter, and obtain multiple monitoring target time series data sets based on the monitoring time interval and various associated features; the multiple monitoring target time series data sets are input into the anomaly detection model to identify whether there are abnormalities in the monitoring parameters; this module is used to implement the function of step S500 in Example 1, which will not be repeated here.

[0113] As a preferred implementation of this embodiment, the steps of obtaining the distribution network data monitoring model according to the topological relationship between the nodes of the distribution network and collecting the operating status characteristics of each node through sensors are specifically as follows:

[0114] Build the state model:

[0115] F(QN)=P(Q1)·P(Q2 / Q1)·P(Q3 / Q2)·…·P(Qn / Qn-1);

[0116] Among them, F(QN) represents the state model of the set of nodes in the distribution network, Qn represents the state of the nth node in the distribution network; P(Qn / Qn-1) represents the conditional probability distribution between the n-1th node in the distribution network and the nth node in the distribution network;

[0117] And there exists the following relationship between the state Qn of the nth node and the state Qn-1 of the n-1th node in the distribution network:

[0118] Qn=Qn-1+s;

[0119] Among them, s is the state noise that conforms to the Gaussian distribution;

[0120] Then P(Qn / Qn-1)=G(Qn-1,σ 2 );G(·,σ 2 ) represents Gaussian distribution;

[0121] Build a sampling model:

[0122] F(xn / Qn)=G(J(xn),σ 2 );

[0123] Among them, F(xn / Qn) is the probability distribution of the sampling value of the distribution network node in state Qn, xn is the sensor sampling value of the nth node, and J(xn) is the sampling function;

[0124] And J(xn)=xn+τ;

[0125] Among them, τ is the sampling noise that conforms to the Gaussian distribution;

[0126] The distribution network data monitoring model is obtained by combining the state model and the sampling model:

[0127] Y(QN,xN)=P(Q1)F(x1 / Q1)·P(Q2 / Q1)F(x2 / Q2)·P(Q3 / Q2)F(x3 / Q3)·…·P(Qn / Qn-1)F(xn / Qn).

[0128] As a preferred implementation of this embodiment, the step of inputting the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain the associated features with each monitoring target parameter specifically includes:

[0129] Preprocess the monitoring data set, including data cleaning and normalization;

[0130] Based on the preprocessed monitoring data set, a training sample set is established, and each training sample in the training sample set includes an operating state feature and a corresponding monitoring target parameter label;

[0131] Construct a machine learning model based on a neural network; iteratively train the machine learning model using a training sample set to obtain a trained correlation feature classification model;

[0132] The monitoring data set and different categories of monitoring target parameters are input into the trained associated feature classification model, and the associated features corresponding to each input monitoring target parameter are output.

[0133] As a preferred implementation manner of this embodiment, the operating status characteristics include voltage, current, power factor, load rate, temperature and humidity.

[0134] Embodiment three:

[0135] This embodiment proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method for monitoring the full-process data of a distribution network as described in any embodiment of the present invention is implemented.

[0136] Embodiment 4:

[0137] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method for monitoring the full-process data of a distribution network as described in any embodiment of the present invention is implemented.

[0138] In the embodiments of the present application, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. Among them, A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can be represented by: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, c can be single or multiple.

[0139] Those of ordinary skill in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented in a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0140] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0141] In several embodiments provided in the present application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), disk or optical disk, and other media that can store program codes.

[0142] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for monitoring data of the entire distribution network, characterized in that: The following steps are involved: According to the topological relationship between the nodes of the distribution network and the operation status characteristics of each node collected by sensors, a distribution network data monitoring model is obtained; According to the distribution network data monitoring model, obtain the monitoring data set and determine the type and quantity of operating status characteristics; Inputting the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain the associated features with each monitoring target parameter, wherein the associated features are operating status features associated with the monitoring target parameters; Based on the associated features of each monitoring target parameter, the monitoring data set is divided into a plurality of monitoring sub-data sets, each sub-data set corresponding to a monitoring target parameter and a corresponding associated feature; Initiate a monitoring request for any monitoring target parameter, obtain the sub-dataset of the corresponding monitoring target parameter, and obtain multiple monitoring target time series data sets based on the monitoring time interval and various associated features; input multiple monitoring target time series data sets into the anomaly detection model to identify whether there are anomalies in the monitoring parameters.

2. A method for monitoring the whole process data of a distribution network according to claim 1, characterized in that: The steps of obtaining the distribution network data monitoring model according to the topological relationship between the nodes of the distribution network and collecting the operating status characteristics of each node through sensors are specifically as follows: Build the state model: F(QN)=P(Q1)·P(Q2 / Q1)·P(Q3 / Q2)·…·P(Qn / Qn-1); Among them, F(QN) represents the state model of the set of nodes in the distribution network, Qn represents the state of the nth node in the distribution network; P(Qn / Qn-1) represents the conditional probability distribution between the n-1th node in the distribution network and the nth node in the distribution network; And there exists the following relationship between the state Qn of the nth node and the state Qn-1 of the n-1th node in the distribution network: Qn=Qn-1+s; Among them, s is the state noise that conforms to the Gaussian distribution; Then P(Qn / Qn-1)=G(Qn-1,σ 2 );G(·,σ 2 ) represents Gaussian distribution; Build a sampling model: F(xn / Qn)=G(J(xn),σ 2 ); Among them, F(xn / Qn) is the probability distribution of the sampling value of the distribution network node in state Qn, xn is the sensor sampling value of the nth node, and J(xn) is the sampling function; And J(xn)=xn+τ; Among them, τ is the sampling noise that conforms to the Gaussian distribution; The distribution network data monitoring model is obtained by combining the state model and the sampling model: Y(QN,xN)=P(Q1)F(x1 / Q1)·P(Q2 / Q1)F(x2 / Q2)·P(Q3 / Q2)F(x3 / Q3)·…·P(Qn / Qn-1)F(xn / Qn).

3. A method for monitoring the whole process data of a distribution network according to claim 1, characterized in that: The step of inputting the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain the associated features with each monitoring target parameter specifically includes: Preprocess the monitoring data set, including data cleaning and normalization; Based on the preprocessed monitoring data set, a training sample set is established, and each training sample in the training sample set includes an operating state feature and a corresponding monitoring target parameter label; Construct a machine learning model based on a neural network; iteratively train the machine learning model using a training sample set to obtain a trained correlation feature classification model; The monitoring data set and different categories of monitoring target parameters are input into the trained associated feature classification model, and the associated features corresponding to each input monitoring target parameter are output.

4. A method for monitoring data of the entire distribution network according to claim 1, characterized in that: The operating status characteristics include voltage, current, power factor, load rate, temperature and humidity.

5. A distribution network full-process data monitoring system, characterized in that: include: The monitoring model building module is used to obtain the distribution network data monitoring model based on the topological relationship between the nodes of the distribution network and the operating status characteristics of each node collected by sensors; A feature classification module is used to obtain a monitoring data set and determine the type and quantity of operating status features according to the distribution network data monitoring model; An associated feature acquisition module is used to input the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain associated features with each monitoring target parameter, wherein the associated features are operating status features associated with the monitoring target parameters; A data set division module, which divides the monitoring data set into multiple monitoring sub-data sets based on the associated features of each monitoring target parameter, each sub-data set corresponding to a monitoring target parameter and the corresponding associated features; The anomaly diagnosis module is used to initiate a monitoring request for any monitoring target parameter, obtain the sub-dataset of the corresponding monitoring target parameter, and obtain multiple monitoring target time series data sets based on the monitoring time interval and various associated features; the multiple monitoring target time series data sets are input into the anomaly detection model to identify whether there are anomalies in the monitoring parameters.

6. A distribution network full-process data monitoring system according to claim 5, characterized in that: The steps of obtaining the distribution network data monitoring model according to the topological relationship between the nodes of the distribution network and collecting the operating status characteristics of each node through sensors are specifically as follows: Build the state model: F(QN)=P(Q1)·P(Q2 / Q1)·P(Q3 / Q2)·…·P(Qn / Qn-1); Among them, F(QN) represents the state model of the set of nodes in the distribution network, Qn represents the state of the nth node in the distribution network; P(Qn / Qn-1) represents the conditional probability distribution between the n-1th node in the distribution network and the nth node in the distribution network; And there exists the following relationship between the state Qn of the nth node and the state Qn-1 of the n-1th node in the distribution network: Qn=Qn-1+s; Among them, s is the state noise that conforms to the Gaussian distribution; Then P(Qn / Qn-1)=G(Qn-1,σ 2 );G(·,σ 2 ) represents Gaussian distribution; Build a sampling model: F(xn / Qn)=G(J(xn),σ 2 ); Among them, F(xn / Qn) is the probability distribution of the sampling value of the distribution network node in state Qn, xn is the sensor sampling value of the nth node, and J(xn) is the sampling function; And J(xn)=xn+τ; Among them, τ is the sampling noise that conforms to the Gaussian distribution; The distribution network data monitoring model is obtained by combining the state model and the sampling model: Y(QN,xN)=P(Q1)F(x1 / Q1)·P(Q2 / Q1)F(x2 / Q2)·P(Q3 / Q2)F(x3 / Q3)·…·P(Qn / Qn-1)F(xn / Qn).

7. A distribution network full-process data monitoring system according to claim 5, characterized in that: The step of inputting the monitoring data set and different types of monitoring target parameters into the machine learning model to obtain the associated features with each monitoring target parameter specifically includes: Preprocess the monitoring data set, including data cleaning and normalization; Based on the preprocessed monitoring data set, a training sample set is established, and each training sample in the training sample set includes an operating state feature and a corresponding monitoring target parameter label; Construct a machine learning model based on a neural network; iteratively train the machine learning model using a training sample set to obtain a trained correlation feature classification model; The monitoring data set and different categories of monitoring target parameters are input into the trained associated feature classification model, and the associated features corresponding to each input monitoring target parameter are output.

8. A distribution network full-process data monitoring system according to claim 5, characterized in that: The operating status characteristics include voltage, current, power factor, load rate, temperature and humidity.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the distribution network full-process data monitoring method as described in any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for monitoring the full-process data of a distribution network as described in any one of claims 1 to 4 is implemented.

Citation Information

Patent Citations

  • Method and system for monitoring distributed flexible resource access power distribution network

    CN117937731A

  • Computer-implemented method and system for anomaly detection and anomaly location in an energy distribution network

    EP4297222A1

  • Data Drive Active Power Grid Abnormal Condition Detection Method and System

    JP7240691B1

  • Machine learnable system with conditional normalizing flow

    US20210019619A1

  • Database exception monitoring method and device, computer device, and storage medium

    WO2021184554A1