CN112733450A - Method and device for analyzing node faults in power network - Google Patents

Method and device for analyzing node faults in power network Download PDF

Info

Publication number
CN112733450A
CN112733450A CN202110034232.9A CN202110034232A CN112733450A CN 112733450 A CN112733450 A CN 112733450A CN 202110034232 A CN202110034232 A CN 202110034232A CN 112733450 A CN112733450 A CN 112733450A
Authority
CN
China
Prior art keywords
node
state
analysis model
power network
fault analysis
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110034232.9A
Other languages
Chinese (zh)
Inventor
李仕林
李梅玉
韦迪潇
王先培
田猛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Electric Power Research Institute of Yunnan Power Grid Co Ltd
Original Assignee
Electric Power Research Institute of Yunnan Power Grid Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Electric Power Research Institute of Yunnan Power Grid Co Ltd filed Critical Electric Power Research Institute of Yunnan Power Grid Co Ltd
Priority to CN202110034232.9A priority Critical patent/CN112733450A/en
Publication of CN112733450A publication Critical patent/CN112733450A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]

Abstract

The application provides a method and a device for analyzing node faults in a power network, which belong to the technical field of power, and the method comprises the following steps: acquiring node states corresponding to a plurality of historical moments of nodes in a power network to be predicted; inputting the node state into a node fault analysis model to obtain a node prediction state at a prediction moment; the node fault analysis model comprises a corresponding relation between a node state and a node prediction state; the node fault analysis model is obtained based on reinforcement learning algorithm training. The method provided by the application simulates the node state of the node in the power network after the fault according to the actual situation, is efficient and accurate, and can effectively estimate the state of each node in the power network at the next moment.

Description

Method and device for analyzing node faults in power network
Technical Field
The present disclosure relates to the field of power technologies, and in particular, to a method and an apparatus for analyzing a node fault in a power network.
Background
The scale of the current power network is becoming huge, and power stations, substations and loads are all important equipment in the power network. These devices often affect each other, and once a device fails, the state of the relevant device changes together. Due to the complex association relationship among the devices, it is difficult for the staff to presume whether the other devices may actually fail according to the real-time device failure condition.
In contrast, the working personnel establish various models, abstract the power network into a complex network, use the important equipment as nodes in the complex network, establish various models to simulate the power network and predict the fault condition of each important equipment in the power network. Common models comprise an OPA model, a blackout accident model based on transient stability constraint optimal power flow and the like. However, if the current model is used in a large and complex power network, once the amount of equipment in the power network exceeds a certain amount, the accuracy of the prediction result is low, and it is difficult to determine the condition of each node in time.
Therefore, there is a need for a method for analyzing node faults in a power network, which is used to solve the problem of low accuracy when predicting the conditions of each node in the power network in the prior art.
Disclosure of Invention
The application provides a method and a device for analyzing node faults in a power network, which can be used for solving the problem of low accuracy when the condition of each node in the power network is predicted in the prior art.
In a first aspect, the present application provides a method for analyzing a node fault in an electrical power network, where the method includes:
acquiring node states corresponding to a plurality of historical moments of nodes in a power network to be predicted;
inputting the node state into a node fault analysis model to obtain a node prediction state at a prediction moment; the node fault analysis model comprises a corresponding relation between a node state and a node prediction state; the node fault analysis model is obtained based on reinforcement learning algorithm training.
With reference to the first aspect, in an implementation manner of the first aspect, the node fault analysis model is trained by using the following method:
acquiring P sample states of the node at P moments;
taking N sample states as input and M states as output, and training to obtain the node fault analysis model; and P is equal to M + N, and the time corresponding to any one of the N sample states is earlier than the time corresponding to any one of the M sample states.
With reference to the first aspect, in an implementation manner of the first aspect, after obtaining node states corresponding to a plurality of historical times of a node in the power network to be predicted, the method further includes:
and carrying out normalization processing on the node state.
With reference to the first aspect, in one implementation manner of the first aspect, the node includes a generator, a substation, and a load.
In a second aspect, the present application provides an apparatus for analyzing a node fault in an electrical power network, the apparatus comprising:
the system comprises an acquisition module, a prediction module and a prediction module, wherein the acquisition module is used for acquiring node states corresponding to a plurality of historical moments of nodes in the power network to be predicted;
the processing module is used for inputting the node state into a node fault analysis model to obtain a node prediction state at a prediction moment; the node fault analysis model comprises a corresponding relation between a node state and a node prediction state; the node fault analysis model is obtained based on reinforcement learning algorithm training.
With reference to the second aspect, in an implementation manner of the second aspect, the node fault analysis model is trained by the following method:
acquiring P sample states of the node at P moments;
taking N sample states as input and M states as output, and training to obtain the node fault analysis model; and P is equal to M + N, and the time corresponding to any one of the N sample states is earlier than the time corresponding to any one of the M sample states.
With reference to the second aspect, in an implementable manner of the second aspect, the apparatus further includes a normalization module;
and the normalization module is used for performing normalization processing on the node state.
With reference to the second aspect, in one implementation manner of the second aspect, the node includes a generator, a substation, and a load.
The method provided by the application simulates the node state of the node in the power network after the fault according to the actual situation, is efficient and accurate, and can effectively estimate the state of each node in the power network at the next moment.
Drawings
Fig. 1 is a schematic flowchart of a method for analyzing a node fault in an electrical power network according to an embodiment of the present disclosure;
fig. 2 is a schematic flowchart of a node fault analysis model trained by the following method according to an embodiment of the present disclosure;
fig. 3 is a schematic structural diagram of an apparatus for analyzing a node fault in an electrical power network according to an embodiment of the present disclosure.
Detailed Description
To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
Fig. 1 is a schematic flow chart of a method for analyzing a node fault in an electrical power network according to an embodiment of the present application. The method provided by the embodiment of the application comprises the following steps:
step 101, obtaining node states corresponding to a plurality of historical moments of nodes in the power network to be predicted.
Specifically, the nodes include generators, substations, and loads.
The corresponding node states include a normal operating state and a fault state.
The power network is abstracted into a complex network, wherein a power station, a transformer substation or a load is regarded as a node of the complex network, a power line is regarded as an edge of the complex network, and finally the complex network with a fixed node number N is obtained.
For a network with a fixed number of nodes N, the load of a node is defined as the betweenness of the node, representing the number of shortest paths through the node in the network. The capacity (maximum load that can be carried) of a node is determined by:
Cj＝(1+α)Ljj 1,2, N equation (1)
In the formula (1), CjIs the capacity of the node; l isjIs the initial load of the node; α is an allowable coefficient.
In the embodiment of the present application, the severity of the fault may be reflected by the maximum connectivity index, and specifically, the maximum connectivity index is determined by the following method:
Figure BDA0002893539780000021
in the formula (2), G is the maximum communication index; n is a radical of,The number of the nodes in the normal working state after the power network is attacked is represented; and N represents the number of nodes in a normal working state before the power network is attacked.
Each node is in a normal working state under the initial condition, at the moment, each node bears a certain initial load, and when the power network fails, the node in the power network is overloaded, namely Lj≥CjThis results in the load being redistributed among the nodes, if some of the nodes cannot continue to be burdenedExcess load may be redistributed again or cause cascading failures due to failures.
And 102, inputting the node state into a node fault analysis model to obtain a node prediction state at the prediction moment.
Before step 102 is executed, the node state needs to be normalized, so that subsequent processing is facilitated.
Specifically, the node fault analysis model includes a correspondence between the node state and the node prediction state. And the node fault analysis model is obtained based on reinforcement learning algorithm training.
For an actual power network, the number of nodes is large, and the high latitude space is provided. The node state is determined by directly utilizing the Bellman optimal equation, and the problem of dimension disaster exists. According to the embodiment of the application, the node fault analysis model is fitted by adopting the multilayer feedforward deep neural network.
In the embodiment of the application, the related multilayer feedforward deep neural network comprises two hidden layers, and the number of the hidden layer neural units is 32 and 16 respectively. In the embodiment of the application, the activation function adopts a linear rectification function.
In the embodiment of the application, xiiRepresenting a complex network, represented by the adjacency matrix of the network, the index i indicates the propagation phase of the cascading failure,
Figure BDA0002893539780000032
represents a node hiA fault, i.e. a node in a fault state, occurs at stage k, indicating that the topology of the power network may change in the fault state. The cascading failure process may be described by a markov decision process, and specifically, the cascading failure may be expressed as:
equation (3) of < S, A, R, p, gamma >
In equation (3), S represents the power network state, and for a power network with N nodes, the state of the system at time t is represented as St＝{x1,x2,…,xNH, if node i fails, then x isi0, otherwise xi＝1。
A denotes the action space, i.e. the node that can fail at time t.
R represents a revenue function, related to the power network state and actions taken, and may be further represented as Rt＝R(st,at,st+1) Wherein s ist,st+1∈S,atE.g. A. The profit is defined as r considering the system loss node and the attack costt＝Gt-c, wherein c represents the cost of the attack,
Figure BDA0002893539780000031
ratio of nodes, N, representing loss at time t due to cascading failuretIndicating the number of nodes lost to cascading failure at time t of the system.
p denotes the transition probability between system states, i.e. take action atFrom state stIs transferred to st+1Is expressed as p(s)t+1|st,at) And the failure probability corresponding to the node is determined.
Gamma represents the discount factor of the system.
As shown in fig. 2, a schematic flow chart of a node fault analysis model provided in the embodiment of the present application is trained by the following method.
The training process of the node fault analysis model provided by the embodiment of the application comprises the following steps:
step 201, acquiring P sample states of a node at P moments.
The sample state is the historical node state of the node.
And step 202, taking the N sample states as input and the M states as output, and training to obtain a node fault analysis model.
And P is equal to M + N, and the time corresponding to any one of the N sample states is earlier than the time corresponding to any one of the M sample states.
Specifically, the training process comprises:
first, the neural network Q, the target network θ, and the cache space D are initialized.
Then, at time t, according to the greedyAvaricious policy selection action atI.e. selecting action a according to probability epsilontOtherwise, choose to make neural network Q (o)t,at(ii) a θ) maximum action at。
Then, in a complex network state otLower execution action atObserving the Complex network State ot+1To obtain a profit rt(ii) a And will information (o)t,at,rt,ot+1) And storing the data into a buffer space D.
Then randomly bulk sampling information (o) from the buffer space Dt,at,rt,ot+1) Sending the data to a neural network, and training the neural network by taking the target network output as a target; and copying parameters of the neural network to the target network at intervals of C.
And finally, stopping training when no node which normally works exists in the complex network.
Repeating the steps for many times, and taking the final neural network as a trained neural network, namely a node fault analysis model.
The method provided by the application simulates the node state of the node in the power network after the fault according to the actual situation, is efficient and accurate, and can effectively estimate the state of each node in the power network at the next moment.
The following are embodiments of the apparatus of the present application that may be used to perform embodiments of the method of the present application. For details which are not disclosed in the embodiments of the apparatus of the present application, reference is made to the embodiments of the method of the present application.
Fig. 3 schematically shows a structural diagram of an analysis apparatus for node faults in an electric power network according to an embodiment of the present application. As shown in fig. 3, the apparatus has a function of implementing the method for analyzing a node fault in the power network, and the function may be implemented by hardware, or may be implemented by hardware executing corresponding software. The apparatus may include: an acquisition module 301, a processing module 302 and a normalization module 303.
An obtaining module 301, configured to obtain node states corresponding to multiple historical times of a node in an electric power network to be predicted;
the processing module 302 is configured to input the node state into the node fault analysis model to obtain a node prediction state at a prediction time; the node fault analysis model comprises a corresponding relation between a node state and a node prediction state; the node fault analysis model is obtained based on reinforcement learning algorithm training.
Optionally, the node fault analysis model is trained by using the following method:
and P sample states of the node at P moments are obtained.
Taking N sample states as input and M states as output, and training to obtain a node fault analysis model; and P is equal to M + N, and the time corresponding to any one of the N sample states is earlier than the time corresponding to any one of the M sample states.
Optionally, the apparatus of the present application further includes a normalization module 303, configured to perform normalization processing on the node state.
Optionally, the nodes include generators, substations and loads.
The method provided by the application simulates the node state of the node in the power network after the fault according to the actual situation, is efficient and accurate, and can effectively estimate the state of each node in the power network at the next moment.
The invention is operational with numerous general purpose or special purpose computing system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like.
The invention may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the invention and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.
It will be understood that the invention is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the invention is limited only by the appended claims.

Claims (8)

1. A method of analyzing node faults in an electrical power network, the method comprising:
acquiring node states corresponding to a plurality of historical moments of nodes in a power network to be predicted;
inputting the node state into a node fault analysis model to obtain a node prediction state at a prediction moment; the node fault analysis model comprises a corresponding relation between a node state and a node prediction state; the node fault analysis model is obtained based on reinforcement learning algorithm training.
2. The method of claim 1, wherein the node fault analysis model is trained using the following method:
acquiring P sample states of the node at P moments;
taking N sample states as input and M states as output, and training to obtain the node fault analysis model; and P is equal to M + N, and the time corresponding to any one of the N sample states is earlier than the time corresponding to any one of the M sample states.
3. The method of claim 1, wherein after obtaining node states corresponding to a plurality of historical times of nodes in the power network to be predicted, the method further comprises:
and carrying out normalization processing on the node state.
4. The method of claim 1, wherein the nodes comprise generators, substations, and loads.
5. An apparatus for analyzing node faults in an electrical power network, the apparatus comprising:
the system comprises an acquisition module, a prediction module and a prediction module, wherein the acquisition module is used for acquiring node states corresponding to a plurality of historical moments of nodes in the power network to be predicted;
the processing module is used for inputting the node state into a node fault analysis model to obtain a node prediction state at a prediction moment; the node fault analysis model comprises a corresponding relation between a node state and a node prediction state; the node fault analysis model is obtained based on reinforcement learning algorithm training.
6. The apparatus of claim 5, wherein the node fault analysis model is trained using the following method:
acquiring P sample states of the node at P moments;
taking N sample states as input and M states as output, and training to obtain the node fault analysis model; and P is equal to M + N, and the time corresponding to any one of the N sample states is earlier than the time corresponding to any one of the M sample states.
7. The apparatus of claim 5, further comprising a normalization module;
and the normalization module is used for performing normalization processing on the node state.
8. The apparatus of claim 5, wherein the nodes comprise generators, substations, and loads.
CN202110034232.9A 2021-01-12 2021-01-12 Method and device for analyzing node faults in power network Pending CN112733450A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110034232.9A CN112733450A (en) 2021-01-12 2021-01-12 Method and device for analyzing node faults in power network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110034232.9A CN112733450A (en) 2021-01-12 2021-01-12 Method and device for analyzing node faults in power network

Publications (1)

Publication Number Publication Date
CN112733450A true CN112733450A (en) 2021-04-30

Family

ID=75590575

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110034232.9A Pending CN112733450A (en) 2021-01-12 2021-01-12 Method and device for analyzing node faults in power network

Country Status (1)

Country Link
CN (1) CN112733450A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113743681A (en) * 2021-11-03 2021-12-03 广东电网有限责任公司惠州供电局 Fault line searching method, device, system and medium based on reinforcement learning

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106909989A (en) * 2017-02-24 2017-06-30 国网河南省电力公司电力科学研究院 A kind of grid disturbance Forecasting Methodology and device
CN107145959A (en) * 2017-03-23 2017-09-08 北京国电通网络技术有限公司 A kind of electric power data processing method based on big data platform
CN111461392A (en) * 2020-01-23 2020-07-28 华中科技大学 Power failure prediction method and system based on graph neural network
CN111912611A (en) * 2020-07-10 2020-11-10 王亮 Method and device for predicting fault state based on improved neural network
CN112054510A (en) * 2020-08-19 2020-12-08 厦门盈盛捷电力科技有限公司 Method for estimating abnormal operation state of power system

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106909989A (en) * 2017-02-24 2017-06-30 国网河南省电力公司电力科学研究院 A kind of grid disturbance Forecasting Methodology and device
CN107145959A (en) * 2017-03-23 2017-09-08 北京国电通网络技术有限公司 A kind of electric power data processing method based on big data platform
CN111461392A (en) * 2020-01-23 2020-07-28 华中科技大学 Power failure prediction method and system based on graph neural network
CN111912611A (en) * 2020-07-10 2020-11-10 王亮 Method and device for predicting fault state based on improved neural network
CN112054510A (en) * 2020-08-19 2020-12-08 厦门盈盛捷电力科技有限公司 Method for estimating abnormal operation state of power system

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
付强 等: "采用自组织RBF网络算法的变压器故障诊断", 《高电压技术》 *
刘晓琴 等: "基于溯因推理网络的电网故障预测方法研究", 《控制工程》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113743681A (en) * 2021-11-03 2021-12-03 广东电网有限责任公司惠州供电局 Fault line searching method, device, system and medium based on reinforcement learning
CN113743681B (en) * 2021-11-03 2022-03-18 广东电网有限责任公司惠州供电局 Fault line searching method, device, system and medium based on reinforcement learning

Similar Documents

Publication Publication Date Title
Kandaperumal et al. Resilience of the electric distribution systems: concepts, classification, assessment, challenges, and research needs
Kang et al. Multimedia knowledge‐based bridge health monitoring using digital twin
Balu et al. On-line power system security analysis
Zhang et al. A comprehensive overview of modeling approaches and optimal control strategies for cyber-physical resilience in power systems
Wang et al. On machine learning-based techniques for future sustainable and resilient energy systems
EP2863509A1 (en) System and method for analyzing oscillatory stability in electrical power transmission systems
Shittu et al. Meta-analysis of the strategies for self-healing and resilience in power systems
Liu et al. Reliability mathematical models of repairable systems with uncertain lifetimes and repair times
Gao et al. Fault diagnosis system for electric vehicle charging devices based on fault tree analysis
Aikhuele Intuitionistic fuzzy model for reliability management in wind turbine system
CN115640748A (en) Dynamic frequency response prediction method for generators after disturbance of power system
Khuntia et al. Risk‐based security assessment of transmission line overloading considering spatio‐temporal dependence of load and wind power using vine copula
Zhang et al. Robust classification model for PMU‐based on‐line power system DSA with missing data
Noorazar et al. Data‐driven operation of the resilient electric grid: A case of COVID‐19
Magoua et al. Incorporating the human factor in modeling the operational resilience of interdependent infrastructure systems
CN112733450A (en) Method and device for analyzing node faults in power network
CN113722303A (en) Electric energy equipment and information operation and maintenance system based on visualization
Shahzad Application of machine learning for optimal wind farm location
Aminev et al. Comparative analysis of reliability prediction models for a distributed radio direction finding telecommunication system
Lin et al. Hpt-rl: Calibrating power system models based on hierarchical parameter tuning and reinforcement learning
Taheri et al. Toward operational resilience of smart energy networks in complex infrastructures
CN115115131A (en) Multi-center power system fault prediction method and system based on transfer learning
Zhang et al. Model and Data Driven Machine Learning Approach for Analyzing the Vulnerability to Cascading Outages With Random Initial States in Power Systems
Xiang et al. A preliminary study of power system reliability considering cloud service reliability
Khalyasmaa et al. Machine learning application for the high-voltage equipment lifecycle forecasting

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20210430

RJ01 Rejection of invention patent application after publication