CN111340326A - Power terminal monitoring method and device based on neural network fault prediction - Google Patents

Power terminal monitoring method and device based on neural network fault prediction Download PDF

Info

Publication number
CN111340326A
CN111340326A CN201911350335.5A CN201911350335A CN111340326A CN 111340326 A CN111340326 A CN 111340326A CN 201911350335 A CN201911350335 A CN 201911350335A CN 111340326 A CN111340326 A CN 111340326A
Authority
CN
China
Prior art keywords
power terminal
neural network
state information
convolutional neural
power
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
CN201911350335.5A
Other languages
Chinese (zh)
Inventor
郑伟军
邵炜平
陈鼎
方景辉
吴国庆
唐锦江
刘哲
刘军雨
邓伟
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.)
State Grid Corp of China SGCC
State Grid Information and Telecommunication Co Ltd
Jiaxing Power Supply Co of State Grid Zhejiang Electric Power Co Ltd
Beijing Zhongdian Feihua Communication Co Ltd
Original Assignee
State Grid Corp of China SGCC
State Grid Information and Telecommunication Co Ltd
Jiaxing Power Supply Co of State Grid Zhejiang Electric Power Co Ltd
Beijing Zhongdian Feihua Communication 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 State Grid Corp of China SGCC, State Grid Information and Telecommunication Co Ltd, Jiaxing Power Supply Co of State Grid Zhejiang Electric Power Co Ltd, Beijing Zhongdian Feihua Communication Co Ltd filed Critical State Grid Corp of China SGCC
Priority to CN201911350335.5A priority Critical patent/CN111340326A/en
Publication of CN111340326A publication Critical patent/CN111340326A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • 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
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06311—Scheduling, planning or task assignment for a person or group
    • 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
    • G06Q10/20—Administration of product repair or maintenance
    • 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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Health & Medical Sciences (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Biomedical Technology (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • Primary Health Care (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Game Theory and Decision Science (AREA)
  • Remote Monitoring And Control Of Power-Distribution Networks (AREA)

Abstract

The invention discloses a method and a device for monitoring a power terminal based on neural network fault prediction, wherein the method comprises the following steps: inputting the state information of each power terminal in the power private network into a convolutional neural network; if the prediction result of the power terminal output by the convolutional neural network is judged to be 'fault of automatic recovery type', controlling the power terminal to automatically reset so as to prevent the fault; if the prediction result of the power terminal output by the convolutional neural network is judged to be the 'failure of the automatic unrecoverable type', the power terminal is judged to be a failure terminal to be dispatched and maintained; the convolutional neural network is obtained by utilizing the historical state information of each power terminal through training in advance. The invention can make the power terminal respond to the fault quickly in time when abnormal.

Description

Power terminal monitoring method and device based on neural network fault prediction
Technical Field
The invention relates to the technical field of power terminal monitoring, in particular to a method and a device for monitoring a power terminal based on neural network fault prediction.
Background
In the long-term operation of the power grid system, the performance is gradually reduced, the reliability is reduced, the failure rate is increased, and the safe operation of the system is endangered. In order to ensure safe and stable operation of a power grid, a power terminal in the administrative field is often monitored to ensure normal operation of the power terminal.
The application of the new technology of the smart power grid and the standardization and integrated promotion of power grid equipment management greatly improve the operation level of the power grid equipment, and a relatively perfect power secondary system is established for assisting the operation of primary equipment of a power system in domestic and foreign power enterprises. However, in the power production operation, the reliability of the operation of the secondary system terminal is reduced due to factors such as equipment quality, artificial misoperation, natural disasters and the like, and the safe and stable operation of the primary power terminal equipment is threatened directly.
Therefore, the operation state of the power terminal is monitored in real time, so that the fault can be responded quickly in time when the power terminal is abnormal, and the method becomes an indispensable step for overhauling the power terminal.
Disclosure of Invention
In view of the above, the present invention provides a method and an apparatus for monitoring an electric power terminal based on neural network fault prediction, so that the electric power terminal can respond to a fault quickly in time when the electric power terminal is abnormal.
Based on the above purpose, the present invention provides a method for monitoring an electric power terminal based on neural network fault prediction, which includes:
inputting the state information of each power terminal in the power private network into a convolutional neural network;
if the prediction result of the power terminal output by the convolutional neural network is judged to be 'fault of automatic recovery type', controlling the power terminal to automatically reset so as to prevent the fault;
if the prediction result of the power terminal output by the convolutional neural network is judged to be the 'failure of the automatic unrecoverable type', the power terminal is judged to be a failure terminal to be dispatched and maintained;
the convolutional neural network is obtained by utilizing the historical state information of each power terminal through pre-training; wherein the historical state information of each power terminal includes: historical state information of normal power terminals, historical state information of power terminals that confirm an anomaly and that are successfully remotely reset, and historical state information of power terminals that are dispatched for maintenance and that solve a fault are confirmed on site.
The state information of the power terminal comprises one or any combination of the following state parameters: bidirectional wireless monitoring heartbeat, online time, offline time, retransmission times, reference signal power/reference signal quality, bidirectional application layer monitoring heartbeat and various application layer alarm information; and
the automatically recoverable type failure specifically includes: communication module failure, application module failure.
The specific training method of the convolutional neural network comprises the following steps:
generating a training sample set: marking the prediction outputs of training samples formed by the state information of the power terminal in the normal working state as 'normal'; the prediction output of a training sample formed by the state information of the power terminal which is abnormal and is recovered after restarting is marked as 'automatically recoverable type failure'; marking the prediction output of a training sample which is abnormal and cannot be formed by restarting the state information of the automatically recovered power terminal as 'failure of the automatic recovery impossible type';
taking a training sample from a training sample set and inputting the training sample into the convolutional neural network;
and calculating the error between the actual output of the convolutional neural network and the output of the mark, and adjusting the parameters of the convolutional neural network according to the calculated error in a method of minimizing the error.
The invention also provides a power terminal monitoring device based on neural network fault prediction, which comprises:
the neural network control module is used for inputting the state information of each power terminal communicated with the base station into a convolutional neural network;
the terminal monitoring module is used for controlling the power terminal to automatically reset to prevent faults when the prediction result of the power terminal output by the convolutional neural network is 'fault capable of automatically recovering type'; when the prediction result of the power terminal output by the convolutional neural network is 'failure of the type which cannot be automatically recovered', judging that the power terminal is a failure terminal to be dispatched and maintained;
the convolutional neural network is obtained by utilizing the historical state information of each power terminal through pre-training; wherein the historical state information of each power terminal includes: historical state information of normal power terminals, historical state information of power terminals that confirm an anomaly and that are successfully remotely reset, and historical state information of power terminals that are dispatched for maintenance and that solve a fault are confirmed on site.
Further, the neural network control module is further configured to train the convolutional neural network using the historical state information of each power terminal.
The invention also provides a terminal monitoring system of the private power network, which comprises a plurality of power terminal monitoring devices based on neural network fault prediction and arranged in the mobile edge computing server configured for each base station of the private power network.
In the technical scheme of the invention, the state information of each power terminal in the power private network is input into a convolutional neural network; if the prediction result of the power terminal output by the convolutional neural network is judged to be 'fault of automatic recovery type', controlling the power terminal to automatically reset so as to prevent the fault; if the prediction result of the power terminal output by the convolutional neural network is judged to be the 'failure of the automatic unrecoverable type', the power terminal is judged to be a failure terminal to be dispatched and maintained; the convolutional neural network is obtained by utilizing the historical state information of each power terminal through pre-training; wherein the historical state information of each power terminal includes: historical state information of normal power terminals, historical state information of power terminals that confirm an anomaly and that are successfully remotely reset, and historical state information of power terminals that are dispatched for maintenance and that solve a fault are confirmed on site. Therefore, the working conditions of all power terminals in the power private network are predicted by utilizing the convolutional neural network; for the electric power terminal which is judged to be the automatic recovery type fault, an automatic fault evasion flow can be sent to the electric power terminal, and the electric power terminal is restarted to achieve the optimization goals of preventing the fault, responding to the fault quickly in time and improving the working reliability of the whole network; when the fault which cannot be automatically recovered is confirmed, for example, the terminal parameters continuously deteriorate, and the number of continuous failures of automatic reset and dial test check reaches a threshold value, the power terminal is judged to be a fault terminal to be dispatched for maintenance, a dispatching order of the power terminal is automatically generated according to dispatching strategy planning, and the aims of reducing the total dispatching amount and improving the dispatching efficiency and the like are achieved through a work optimization module on the premise of ensuring the overall reliability of the system.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of a method for monitoring an electric power terminal based on neural network fault prediction according to an embodiment of the present invention;
fig. 2 is a flowchart of a training method of a convolutional neural network according to an embodiment of the present invention;
fig. 3 is a schematic network topology diagram of a private power network according to an embodiment of the present invention;
fig. 4 is a block diagram of an internal structure of a power terminal monitoring device based on neural network fault prediction according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to specific embodiments and the accompanying drawings.
It is to be noted that technical terms or scientific terms used in the embodiments of the present invention should have the ordinary meanings as understood by those having ordinary skill in the art to which the present disclosure belongs, unless otherwise defined. The use of "first," "second," and similar terms in this disclosure is not intended to indicate any order, quantity, or importance, but rather is used to distinguish one element from another. The word "comprising" or "comprises", and the like, means that the element or item listed before the word covers the element or item listed after the word and its equivalents, but does not exclude other elements or items. The terms "connected" or "coupled" and the like are not restricted to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "upper", "lower", "left", "right", and the like are used merely to indicate relative positional relationships, and when the absolute position of the object being described is changed, the relative positional relationships may also be changed accordingly.
The inventor of the invention considers that the convolutional neural network is utilized to monitor each power terminal in the special power grid, and track and predict the working state, the fault expression, the maintenance and the recovery process of each power terminal. At the initial stage of deployment, the convolutional neural network is trained by utilizing the online data of each power terminal. After certain historical data are accumulated, particularly after a certain amount of state information data of 'confirming abnormal power terminals which are successfully reset remotely' and 'dispatching and confirming power terminals which solve faults on site' are recorded, the prediction capability of the convolutional neural network is gradually improved.
When the convolutional neural network achieves certain credibility for 'identification and management of the power terminal with higher fault risk probability', the convolutional neural network can be used for predicting the working condition of each power terminal in the power private network; for the electric power terminal which is judged to be the automatic recovery type fault, an automatic fault evasion flow can be sent to the electric power terminal, and the electric power terminal is restarted to achieve the optimization goals of preventing the fault, responding to the fault quickly in time and improving the working reliability of the whole network; when the fault which cannot be automatically recovered is confirmed, for example, the terminal parameters continuously deteriorate, and the number of continuous failures of automatic reset and dial test check reaches a threshold value, the power terminal is judged to be a fault terminal to be dispatched for maintenance, a dispatching order of the power terminal is automatically generated according to dispatching strategy planning, and the aims of reducing the total dispatching amount and improving the dispatching efficiency and the like are achieved through a work optimization module on the premise of ensuring the overall reliability of the system.
The technical solution of the embodiments of the present invention is described in detail below with reference to the accompanying drawings.
The invention provides a method for monitoring a power terminal based on neural network fault prediction, which has a specific flow shown in figure 1 and comprises the following steps:
step S101: and inputting the state information of each power terminal in the power private network into the convolutional neural network.
In this step, the state information of each power terminal in the power private network is input into the convolutional neural network obtained by training in advance. The convolutional neural network is obtained by utilizing the historical state information of each power terminal through pre-training; wherein the historical state information of each power terminal includes: historical state information of normal power terminals, historical state information of power terminals that confirm an anomaly and that are successfully remotely reset, and historical state information of power terminals that are dispatched for maintenance and that solve a fault are confirmed on site.
Specifically, the state information of the power terminal may include one or any combination of the following state parameters: bidirectional wireless monitoring heartbeat, online time, offline time, retransmission times, reference signal power/reference signal quality, bidirectional application layer monitoring heartbeat and various application layer alarm information; the vector format description of the state information input to the power terminals of the convolutional neural network may be as shown in table 1:
TABLE 1
Figure BDA0002334501750000061
The vector of state information collected by the kth terminal ue (k) at time t _ i is represented as:
Input(t_i,k)＝
{HtBeatFWd(t_i,k)；HtBeatBWd(t_i,k)；InLine(t_i,k)；OffLinet(t_i,k)；HRQ_T(t_i,k)；RSRP(t_i,k)；RSRQ(t_i,k)……}
because the original physical meaning and dimension difference of the input vector are large, each dimension of the vector is normalized to be a [0,1] interval by adopting an experience-based scale when input data is normalized.
Generally, to improve input dimensionality and improve parallelism, the single-input vector for each terminal can be constructed as:
Input_UE(t_i,k)＝
{Input(t_i,k)；Input(t_i-1,k)；Input(t_i-2,k)；Input(t_i-3,k)；……Input(t_i-n,k)}
that is, the state information input to the power terminals of the convolutional neural network has temporal continuity; the state information of the power terminal input to the convolutional neural network at one time comprises the state information of the power terminal collected at the time t _ i, t _ i-1, t _ i-2, … … and t _ i-n; the state information of the power terminal input to the convolutional neural network at one time is a matrix formed by a group of continuously acquired state information of the power terminal; wherein n can be set by a person skilled in the art according to actual conditions.
The convolution kernel adopted by the convolution neural network comprises convolution kernels with the size of 3 × 3 or 5 × 5, and is mainly used for carrying out feature extraction on input gradient change, identifying parameter jitter conditions through convolution kernels in different gradient directions, and extracting and filtering parameter change modes conforming to a fault training sample set through a pooling layer, so that the convolution neural network can identify a fault power terminal.
Step S102: if the prediction result of the power terminal output by the convolutional neural network is judged to be the automatic recovery type fault, executing the following step S103; if the prediction result of the power terminal output by the convolutional neural network is judged to be the "failure of the non-automatic recovery type", the following step S104 is executed.
Specifically, the common automatically recoverable type failures of the power terminals include: the communication module failure and the application module failure are two major types, and the state logic is shown in the following table 2:
TABLE 2
Power terminal status Communication module Application module
Is normal Good taste Good taste
Exception 1 Bad Good taste
Anomaly 2 Good taste Bad
Anomaly 3 Bad Bad
Wherein, the 'abnormity 1' indicates the fault of the communication module, when the communication module of the power terminal has the fault, the 'communication module working parameter/state management' module at the base station side can not obtain the updating information of the terminal, and immediately confirms the abnormal mode that the terminal state is 'abnormity 1'; and the power terminal side communication module working parameter/state management module can also detect the fault according to the terminal state. Generally, power terminals are provided with built-in batteries or large-capacity capacitors, and can provide power-off process protection and other capabilities for the power terminals. If the power terminal is normally powered and the OS and the application management module of the terminal are normal, the communication module working parameter/state management module of the power terminal enters a fault processing flow: firstly, checking reference signal strength RSRP/RSRQ of a base station before failure and a downlink signal-to-noise ratio SNR, if a network signal is lower than a set threshold, setting a modem to be in a cell search state by a terminal, and according to a configuration period, for example, if cell search is unsuccessful within 30 seconds, resetting the terminal modem immediately.
The recovery of "exception 1" generally relies on self-recovery capability, and after the power terminal communication module is reset, the communication module will re-initiate cell search and attachment. When the attachment is successful, the "Exception 1" fault is removed.
"anomaly 2" indicates that the application module is in failure, the communication module of the power terminal works normally when the anomaly 2 occurs, and the "bidirectional wireless monitoring heartbeat" can be detected at both the "application management module" at the base station side and the "application management module" of the power terminal. At this time, the "application management module" of the power terminal cannot communicate with the power application software of the terminal, or cannot obtain the response information of the power application software through a handshake monitoring mechanism. Once the state exceeds the preset delay time, the terminal application management module informs the terminal Operating System (OS) to initiate a reset procedure to the terminal application layer. And when the terminal is restarted, an application management module of the power terminal continuously monitors the state of the terminal according to the related information in the local record file.
In general, the recovery of "exception 2" may also depend on the self-recovery capability, and after the power terminal is automatically reset and restarted, the "exception 2" fault may be eliminated.
The 'abnormity 3' indicates that both the communication module and the application module are in failure, and the 'abnormity 3' failure can be eliminated by controlling the power terminal to automatically reset and restart.
Step S103: and controlling the power terminal to perform automatic reset so as to prevent faults.
Specifically, an 'automatic fault evasion flow' is sent to the power terminal, and a suspected fault module of the power terminal is restarted to achieve the purposes of preventing faults and improving the working reliability of the whole network.
Step S104: and judging that the power terminal is a fault terminal to be dispatched and maintained, putting the terminal into a maintenance queue to be dispatched, and generating a dispatching order of the power terminal.
Specifically, after the prediction result of the power terminal output by the convolutional neural network is judged to be a failure of the type which cannot be automatically recovered, or after the judged failure of the type which can be automatically recovered reaches a threshold value in automatic resetting and dial testing, the terminal is put into a maintenance queue to be dispatched according to a dispatching strategy plan, a dispatching order of the power terminal is automatically generated, and on the premise of ensuring the overall reliability of the system, the aims of reducing the total dispatching amount and improving the dispatching efficiency and the like are achieved through a work optimization module.
The convolutional neural network described above is pre-trained: through training of a large number of normal sample sets and training of a small number of fault sample sets, a data pattern which is likely to have faults is separated from a high-dimensional sample space in an attempt, and when the prediction confidence of the convolutional neural network is higher than a threshold value, such as 60%, the convolutional neural network can be used for providing reference information for predicting the faults. The normal sample set is a set of historical state information of a normal power terminal; the fault sample set is a set including historical state information of the power terminals for which an anomaly is confirmed and a success in remote reset, and historical state information of the power terminals for which a fault is dispatched for maintenance and confirmed on site.
The flow of the above-mentioned training method of the convolutional neural network is shown in fig. 2, and includes the following steps:
step S201: and generating a training sample set, and initializing parameters of the convolutional neural network.
Specifically, the state information reported by each power terminal in the private power grid can be used as a training sample, and because most power terminals are in normal working states, the prediction outputs of the training samples formed by the state information of the power terminals in the normal working states are all marked as "normal";
when an electric terminal in the private electric power network is abnormal and is recovered after restarting is executed, a training sample formed by state information of the electric terminal which is abnormal and is recovered after restarting is taken as an abnormal/automatic restarting sample of a fault sample set, the prediction output of the training sample is marked as an automatic recoverable type fault, and the prediction output of the training sample can be subdivided and marked as 'abnormal 1', 'abnormal 2' and 'abnormal 3' according to the states;
when an abnormal power terminal in the private power network is abnormal and cannot be automatically recovered through restarting, a training sample formed by state information of the abnormal power terminal which cannot be automatically recovered through restarting is used as a 'failure to be recovered' sample in a failure sample set, and the prediction output of the training sample is marked as 'failure of the type which cannot be automatically recovered'; for the 'unrecoverable fault' training sample, the training sample needs to be used for training after dispatching, processing and fault repair summary completion.
The training samples input to the convolutional neural network are also time-continuous in the manner described above, that is, each training sample is a matrix formed by a set of continuously acquired state information of the power terminal, that is, as one training sample input to the convolutional neural network, a set of continuously acquired state information of the power terminal.
When initializing the parameters of the convolutional neural network, setting the weights, thresholds and the like of the convolutional neural network to random values close to 0, and initializing the precision control parameters and the learning rate.
Step S202: and (3) taking a training sample from the training sample set, inputting the training sample into the convolutional neural network, and obtaining the output of the convolutional neural network through a forward propagation stage.
Specifically, a training sample is taken from a training sample set and input into the convolutional neural network, and in a forward propagation stage, input information is transmitted to an output layer from an input layer of the convolutional neural network through gradual transformation.
Since most of the power terminals within the network do not fail, the set of fault samples has relatively few fault samples. In the actual training process, a sample relaxation method is usually considered for a fault sample, namely, a fault flow pattern of the sample trained in a convolutional neural network is subjected to certain relaxation treatment on each dimension of a flow. The area occupied by the fault manifold in the high-dimensional space is appropriately increased according to the calculation accuracy. The range of failure prediction is appropriately increased. The relaxation method can be used for clearing the incorrect region by the sample set in the subsequent iterative training.
Step S203: through the back propagation stage, the parameters of the convolutional neural network are adjusted.
In this step, an error between the actual output of the convolutional neural network and the output of the tag is calculated, and parameters of the convolutional neural network are adjusted according to the calculated error by a method of minimizing the error.
Specifically, each parameter of the convolutional neural network is adjusted by adopting a gradient adjustment algorithm, namely the actual output of the convolutional neural network is compared with the output of the mark, and an output error is calculated; errors also need to be calculated for hidden units in the middle layer; and sequentially calculating the adjustment quantity of each weight and the adjustment quantity of the threshold value, and adjusting the weight and the threshold value of the convolutional neural network.
Step S204: judging whether the training of a group of training samples is finished or not, and whether the indexes of the convolutional neural network for fault prediction meet the precision requirement or not; if yes, go to step S205; otherwise, the iterative training is continued by returning to step S202.
Step S205: when the accuracy of the network for identifying the fault prediction sample reaches a certain index, the network can be used for predicting the fault of the current network terminal.
It should be noted that the method of the embodiment of the present invention may be executed by a single device, such as a computer or a server. The method of the embodiment can also be applied to a distributed scene and completed by the mutual cooperation of a plurality of devices. In the case of such a distributed scenario, one of the multiple devices may only perform one or more steps of the method according to the embodiment of the present invention, and the multiple devices interact with each other to complete the method.
In fact, a plurality of base stations and power terminals communicating with the base stations can be included in the private power network, and the network topology is shown in fig. 3; based on the above method for monitoring a power terminal based on neural network fault prediction, the terminal monitoring system for a power private network provided by the embodiment of the invention comprises: the plurality of power terminal monitoring devices based on neural network failure prediction are respectively arranged in an MEC (Mobile Edge Computing) configured in each base station of the private power network. Fig. 4 shows a block diagram of an internal structure of the power terminal monitoring device based on neural network fault prediction, which includes: a neural network control module 401, a terminal monitoring module 402 and a convolutional neural network 403.
The neural network control module 401 is configured to input status information of each power terminal communicating with the base station to the convolutional neural network 403;
the terminal monitoring module 402 is configured to control the power terminal to perform automatic reset to prevent a fault when a prediction result of the power terminal output by the convolutional neural network 403 is an "automatically recoverable type fault"; when the prediction result of the power terminal output by the convolutional neural network is 'failure of the type which cannot be automatically recovered', judging that the power terminal is a failure terminal to be dispatched and maintained;
the convolutional neural network 403 is obtained by pre-training by using historical state information of each power terminal; wherein the historical state information of each power terminal includes: historical state information of normal power terminals, historical state information of power terminals that confirm an anomaly and that are successfully remotely reset, and historical state information of power terminals that are dispatched for maintenance and that solve a fault are confirmed on site.
Further, the neural network control module 401 may be further configured to train the convolutional neural network 403 with the historical state information of each power terminal: the neural network control module 401 generates a training sample set: marking the prediction outputs of training samples formed by the state information of the power terminal in the normal working state as 'normal'; the prediction output of a training sample formed by the state information of the power terminal which is abnormal and is recovered after restarting is marked as 'automatically recoverable type failure'; marking the prediction output of a training sample which is abnormal and cannot be formed by restarting the state information of the automatically recovered power terminal as 'failure of the automatic recovery impossible type'; then, training samples are sequentially taken out from the training sample set and input into the convolutional neural network 403; an error between an actual output of the convolutional neural network 403 and an output of the tag is calculated, and a parameter of the convolutional neural network 403 is adjusted according to the calculated error in a method of minimizing the error.
Specifically, as shown in fig. 4, an "application management module" and a "communication management module" are provided in the power terminal; an application management module in the power terminal acquires position information in a mode of 'configuration files' or 'reading terminal position information of a GIS module carried by the terminal';
the working parameters/states of the power terminal acquired by the communication management module in the power terminal include: bidirectional wireless monitoring heartbeat, "online", "offline", "number of retransmissions", "RSRP/RSRQ", etc.; the self-checking data acquired by the application management module in the power terminal comprises the following steps: the bidirectional application layer monitors heartbeat, various application layer alarm information and the like.
An application management module and a communication management module are correspondingly arranged in a mobile edge computing server configured at a base station side;
a communication management module in a mobile edge computing server at a base station side acquires registration information of the electric power terminal and report information of related states of the electric power terminal; an application management module in a mobile edge computing server at a base station side records the geographic position of an electric power terminal according to GIS data reported by the electric power terminal accessed to the base station.
The application management module in the mobile edge computing server at the base station side can also inquire the topological logic relationship among the nodes of each power terminal in the network through a superior management node; and calculating the accuracy of the data information reported by each node in the topological logical relationship through the logical relationship, and predicting and tracking the accuracy of the related data of each node. The application management module preferentially updates monitoring data and automatically dials and tests the acquired terminal in the area of the possible lightning stroke position through an interface with an external lightning stroke observation system, and actively updates the state of the terminal in the area of the possible fault. And updating monitoring data and automatically testing the reported suspected surge area, actively updating the state of the terminal in the possible fault area, and actively monitoring and maintaining the possible network fault.
In the scheduling process, the base station side scheduler updates the state of the private network terminal, such as on-line, off-line, retransmission times, RSRP/RSRQ and the like, in quasi-real time through a scheduling algorithm such as polling and the like, regularly copies the terminal state information buffer area in the base station scheduler in a slicing mode by using an interface provided by system Operation Maintenance (OM), periodically provides the terminal state information buffer area to an application management module in a mobile edge computing server at the base station side, and monitors the state of a modem of an electric power terminal in the network. If the communication module of the power terminal fails, the 'communication management module' in the mobile edge computing server at the base station side cannot acquire the update information of the power terminal, immediately changes the state of the power terminal into an abnormal mode, and notifies the 'application management module'; and the 'communication management module' on the side of the fault terminal can also detect the fault according to the state of the terminal. Generally, power terminals are provided with built-in batteries or large-capacity capacitors, and can provide power-off process protection and other capabilities for the terminals.
If the power supply of the power terminal is normal, and the OS and the application management module of the terminal are normal, the communication management module of the terminal enters a fault processing flow, firstly, the reference signal strength RSRP/RSRQ and the downlink signal-to-noise ratio SNR of a base station before fault are checked, if the network signal is lower than a set threshold, the terminal sets the modem to be in a cell search state, and according to a configuration period, for example, the cell search is unsuccessful within 30 seconds, the terminal modem is reset immediately.
The 'application management module' and the 'application management module' of the power terminal in the mobile edge computing server at the base station side wait for the terminal communication module to reset, and after the terminal communication module resets, the communication module initiates cell search and attachment again. When the attachment is successful, the "Exception 1" fault is removed. If the terminal communication module can not be connected again, the application management module in the mobile edge computing server at the base station side puts the power terminal into a maintenance queue to be dispatched after the time delay exceeds a certain threshold.
The bidirectional wireless heartbeat monitoring can be detected at an application management module in a mobile edge computing server at the base station side and an application management module in a power terminal. At this time, if the power terminal "application management module" cannot communicate with the terminal power application software or cannot obtain the response information of the power application software through a handshake monitoring mechanism, once the state exceeds the preset delay time. The 'application management module' of the power terminal then informs the terminal Operating System (OS) to initiate a reset procedure to the terminal application layer. And when the terminal is restarted, the terminal application management module continuously monitors the state of the terminal according to the related information in the local record file.
If the power terminal application management module finds that the application and the communication module are in a fault state, a terminal reset command is initiated immediately. If the terminal can not be connected again, the application management module in the mobile edge computing server at the base station side puts the power terminal into a maintenance queue to be dispatched after the time delay exceeds a certain threshold.
The neural network control module 401 obtains status information of each power terminal communicating with the base station from an "application management module" in the mobile edge computing server on the base station side.
In the technical scheme of the invention, the electric power terminal monitoring device based on neural network fault prediction is deployed in the mobile edge computing server at the base station side, which is beneficial to automatic terminal management and fault prediction at the cell level, and the weights and network structures of the convolutional neural networks trained on the mobile edge computing server at each base station side are different. Because the fault data of different regional terminals in the network has the characteristics of regional difference and the like due to factors such as environmental difference and the like, the regional division mode based on the mobile edge computing server at the base station side is adopted, and the balance can be achieved on the calculated amount and the prediction effect.
A scheduler in the electric power wireless private network base station carries out repeated dial testing on the electric power terminal which is judged to be the automatic recovery type fault by the electric power terminal monitoring device based on the neural network fault prediction in a polling mode, and further analyzes the fault type through dial testing data, so that a basis is provided for subsequent fault elimination processing.
In the technical scheme of the invention, the state information of each power terminal in the power private network is input into a convolutional neural network; if the prediction result of the power terminal output by the convolutional neural network is judged to be 'fault of automatic recovery type', controlling the power terminal to automatically reset so as to prevent the fault; if the prediction result of the power terminal output by the convolutional neural network is judged to be the 'failure of the automatic unrecoverable type', the power terminal is judged to be a failure terminal to be dispatched and maintained; the convolutional neural network is obtained by utilizing the historical state information of each power terminal through pre-training; wherein the historical state information of each power terminal includes: historical state information of normal power terminals, historical state information of power terminals that confirm an anomaly and that are successfully remotely reset, and historical state information of power terminals that are dispatched for maintenance and that solve a fault are confirmed on site. Therefore, the working conditions of all power terminals in the power private network are predicted by utilizing the convolutional neural network; for the electric power terminal which is judged to be the automatic recovery type fault, an automatic fault evasion flow can be sent to the electric power terminal, and the electric power terminal is restarted to achieve the optimization goals of preventing the fault, responding to the fault quickly in time and improving the working reliability of the whole network; when the fault which cannot be automatically recovered is confirmed, for example, the terminal parameters continuously deteriorate, and the number of continuous failures of automatic reset and dial test check reaches a threshold value, the power terminal is judged to be a fault terminal to be dispatched for maintenance, a dispatching order of the power terminal is automatically generated according to dispatching strategy planning, and the aims of reducing the total dispatching amount and improving the dispatching efficiency and the like are achieved through a work optimization module on the premise of ensuring the overall reliability of the system.
Computer-readable media of the present embodiments, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device.
Those of ordinary skill in the art will understand that: the discussion of any embodiment above is meant to be exemplary only, and is not intended to intimate that the scope of the disclosure, including the claims, is limited to these examples; within the idea of the invention, also features in the above embodiments or in different embodiments may be combined, steps may be implemented in any order, and there are many other variations of the different aspects of the invention as described above, which are not provided in detail for the sake of brevity.
In addition, well known power/ground connections to Integrated Circuit (IC) chips and other components may or may not be shown within the provided figures for simplicity of illustration and discussion, and so as not to obscure the invention. Furthermore, devices may be shown in block diagram form in order to avoid obscuring the invention, and also in view of the fact that specifics with respect to implementation of such block diagram devices are highly dependent upon the platform within which the present invention is to be implemented (i.e., specifics should be well within purview of one skilled in the art). Where specific details (e.g., circuits) are set forth in order to describe example embodiments of the invention, it should be apparent to one skilled in the art that the invention can be practiced without, or with variation of, these specific details. Accordingly, the description is to be regarded as illustrative instead of restrictive.
While the present invention has been described in conjunction with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art in light of the foregoing description. For example, other memory architectures (e.g., dynamic ram (dram)) may use the discussed embodiments.
The embodiments of the invention are intended to embrace all such alternatives, modifications and variances that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, substitutions, improvements and the like that may be made without departing from the spirit and principles of the invention are intended to be included within the scope of the invention.

Claims (10)

1. A power terminal monitoring method based on neural network fault prediction is characterized by comprising the following steps:
inputting the state information of each power terminal in the power private network into a convolutional neural network;
if the prediction result of the power terminal output by the convolutional neural network is judged to be 'fault of automatic recovery type', controlling the power terminal to automatically reset so as to prevent the fault;
if the prediction result of the power terminal output by the convolutional neural network is judged to be the 'failure of the automatic unrecoverable type', the power terminal is judged to be a failure terminal to be dispatched and maintained;
the convolutional neural network is obtained by utilizing the historical state information of each power terminal through pre-training; wherein the historical state information of each power terminal includes: historical state information of normal power terminals, historical state information of power terminals that confirm an anomaly and that are successfully remotely reset, and historical state information of power terminals that are dispatched for maintenance and that solve a fault are confirmed on site.
2. The method according to claim 1, wherein the status information of the power terminal comprises one or any combination of the following status parameters: bidirectional wireless monitoring heartbeat, online time, offline time, retransmission times, reference signal power/reference signal quality, bidirectional application layer monitoring heartbeat and various application layer alarm information; and
the automatically recoverable type failure specifically includes: communication module failure, application module failure.
3. The method of claim 2, wherein the specific training method of the convolutional neural network comprises:
generating a training sample set: marking the prediction outputs of training samples formed by the state information of the power terminal in the normal working state as 'normal'; the prediction output of a training sample formed by the state information of the power terminal which is abnormal and is recovered after restarting is marked as 'automatically recoverable type failure'; marking the prediction output of a training sample which is abnormal and cannot be formed by restarting the state information of the automatically recovered power terminal as 'failure of the automatic recovery impossible type';
taking a training sample from a training sample set and inputting the training sample into the convolutional neural network;
and calculating the error between the actual output of the convolutional neural network and the output of the mark, and adjusting the parameters of the convolutional neural network according to the calculated error in a method of minimizing the error.
4. The method according to any one of claims 1 to 3, wherein the state information of the power terminal input to the convolutional neural network at a time is a matrix formed by a set of continuously collected state information of the power terminal.
5. An electric power terminal monitoring device based on neural network fault prediction, characterized in that the device comprises:
the neural network control module is used for inputting the state information of each power terminal communicated with the base station into a convolutional neural network;
the terminal monitoring module is used for controlling the power terminal to automatically reset to prevent faults when the prediction result of the power terminal output by the convolutional neural network is 'fault capable of automatically recovering type'; when the prediction result of the power terminal output by the convolutional neural network is 'failure of the type which cannot be automatically recovered', judging that the power terminal is a failure terminal to be dispatched and maintained;
the convolutional neural network is obtained by utilizing the historical state information of each power terminal through pre-training; wherein the historical state information of each power terminal includes: historical state information of normal power terminals, historical state information of power terminals that confirm an anomaly and that are successfully remotely reset, and historical state information of power terminals that are dispatched for maintenance and that solve a fault are confirmed on site.
6. The apparatus of claim 5, wherein the status information of the power terminal comprises one or any combination of the following status parameters: bidirectional wireless monitoring heartbeat, online time, offline time, retransmission times, reference signal power/reference signal quality, bidirectional application layer monitoring heartbeat and various application layer alarm information; and
the automatically recoverable type failure specifically includes: communication module failure, application module failure.
7. The apparatus of claim 5,
the neural network control module is also used for training the convolutional neural network by utilizing the historical state information of each power terminal.
8. The apparatus of claim 5,
the neural network control module is specifically configured to generate a training sample set when training the convolutional neural network: marking the prediction outputs of training samples formed by the state information of the power terminal in the normal working state as 'normal'; the prediction output of a training sample formed by the state information of the power terminal which is abnormal and is recovered after restarting is marked as 'automatically recoverable type failure'; marking the prediction output of a training sample which is abnormal and cannot be formed by restarting the state information of the automatically recovered power terminal as 'failure of the automatic recovery impossible type'; then, training samples are sequentially taken out from the training sample set and input into the convolutional neural network; and calculating the error between the actual output of the convolutional neural network and the output of the mark, and adjusting the parameters of the convolutional neural network according to the calculated error in a method of minimizing the error.
9. The apparatus of claim 5, wherein the state information of the power terminal inputted to the convolutional neural network at a time by the neural network control module is a matrix formed by a set of continuously collected state information of the power terminal.
10. A terminal monitoring system of a private power network is characterized by comprising: a plurality of electric power terminal monitoring devices based on neural network failure prediction according to any one of claims 5 to 9, respectively installed in the mobile edge computing server configured for each base station of the private electric power network.
CN201911350335.5A 2019-12-24 2019-12-24 Power terminal monitoring method and device based on neural network fault prediction Pending CN111340326A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911350335.5A CN111340326A (en) 2019-12-24 2019-12-24 Power terminal monitoring method and device based on neural network fault prediction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911350335.5A CN111340326A (en) 2019-12-24 2019-12-24 Power terminal monitoring method and device based on neural network fault prediction

Publications (1)

Publication Number Publication Date
CN111340326A true CN111340326A (en) 2020-06-26

Family

ID=71181385

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911350335.5A Pending CN111340326A (en) 2019-12-24 2019-12-24 Power terminal monitoring method and device based on neural network fault prediction

Country Status (1)

Country Link
CN (1) CN111340326A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112506690A (en) * 2020-12-09 2021-03-16 北京地平线信息技术有限公司 Method and device for controlling processor
CN112948205A (en) * 2021-02-22 2021-06-11 国网新疆电力有限公司营销服务中心(资金集约中心、计量中心) Novel electric power measurement monitored control system
CN113525714A (en) * 2021-07-30 2021-10-22 浙江工商大学 Civil aviation aircraft fault prediction system based on neural network
CN114205011A (en) * 2021-12-15 2022-03-18 浙江华云信息科技有限公司 Wireless signal quality early warning method, device and medium

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101702537A (en) * 2009-11-10 2010-05-05 深圳市科陆电子科技股份有限公司 Method for processing failures on adaptive basis in terminal of distribution network
CN102709889A (en) * 2012-05-29 2012-10-03 长园深瑞继保自动化有限公司 Self-healing method and self-healing system for faults of power distribution network
CN105510865A (en) * 2015-11-27 2016-04-20 国家电网公司 Power user onsite electric energy meter fault remote determination method
CN107886168A (en) * 2017-11-07 2018-04-06 歌拉瑞电梯股份有限公司 One kind carries out elevator faults using multilayer perceptron neutral net and knows method for distinguishing
CN109615086A (en) * 2018-10-11 2019-04-12 国网浙江省电力有限公司电力科学研究院 A kind of generation method and system of O&M assisted tag
CN109713795A (en) * 2019-01-08 2019-05-03 烟台科大正信电气有限公司 Distributed feeder automation method based on network peer-to-peer communications
CN109768625A (en) * 2019-03-29 2019-05-17 国网山东省电力公司费县供电公司 A kind of electric system overhaul management terminal and method

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101702537A (en) * 2009-11-10 2010-05-05 深圳市科陆电子科技股份有限公司 Method for processing failures on adaptive basis in terminal of distribution network
CN102709889A (en) * 2012-05-29 2012-10-03 长园深瑞继保自动化有限公司 Self-healing method and self-healing system for faults of power distribution network
CN105510865A (en) * 2015-11-27 2016-04-20 国家电网公司 Power user onsite electric energy meter fault remote determination method
CN107886168A (en) * 2017-11-07 2018-04-06 歌拉瑞电梯股份有限公司 One kind carries out elevator faults using multilayer perceptron neutral net and knows method for distinguishing
CN109615086A (en) * 2018-10-11 2019-04-12 国网浙江省电力有限公司电力科学研究院 A kind of generation method and system of O&M assisted tag
CN109713795A (en) * 2019-01-08 2019-05-03 烟台科大正信电气有限公司 Distributed feeder automation method based on network peer-to-peer communications
CN109768625A (en) * 2019-03-29 2019-05-17 国网山东省电力公司费县供电公司 A kind of electric system overhaul management terminal and method

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
何勇枢;陈赣;: "基于BP神经网络模型的故障预测分析", no. 16, pages 227 - 229 *
周午纵: "《旅英学人谈科技热点》", 复旦大学出版社, pages: 29 - 31 *
张旭;程雪婷;赵冬梅;张东英;刘燕华;: "电网故障恢复的发展、现状及展望", no. 02, pages 13 - 18 *
陈灵;: "人工智能化在电气自动化中应用分析", no. 17, pages 28 - 29 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112506690A (en) * 2020-12-09 2021-03-16 北京地平线信息技术有限公司 Method and device for controlling processor
CN112948205A (en) * 2021-02-22 2021-06-11 国网新疆电力有限公司营销服务中心(资金集约中心、计量中心) Novel electric power measurement monitored control system
CN113525714A (en) * 2021-07-30 2021-10-22 浙江工商大学 Civil aviation aircraft fault prediction system based on neural network
CN114205011A (en) * 2021-12-15 2022-03-18 浙江华云信息科技有限公司 Wireless signal quality early warning method, device and medium

Similar Documents

Publication Publication Date Title
CN111340326A (en) Power terminal monitoring method and device based on neural network fault prediction
US10886736B2 (en) Post-disaster topology detection and energy flow recovery in power distribution network
US10389117B2 (en) Dynamic modeling and resilience for power distribution
US11748185B2 (en) Multi-factor cloud service storage device error prediction
CN109313475A (en) Processor device voltage characterization
CN110164501A (en) A kind of hard disk detection method, device, storage medium and equipment
US20210019653A1 (en) Failure analysis device, failure analysis method, and failure analysis program
CN112487210A (en) Abnormal device identification method, electronic device, and medium
CN116256661B (en) Battery fault detection method, device, electronic equipment and storage medium
US11429467B2 (en) Method and system for prediction of correct discrete sensor data based on temporal uncertainty
CN116794510A (en) Fault prediction method, device, computer equipment and storage medium
CN113261328A (en) Apparatus and method for monitoring performance of network device in wireless communication system
CN105224416B (en) Restorative procedure and related electronic device
CN113406439A (en) Power distribution network fault positioning method
CN104158843A (en) Storage unit invalidation detecting method and device for distributed file storage system
CN115766402B (en) Method and device for filtering server fault root cause, storage medium and electronic device
CN112836843A (en) Method and device for predicting base station out-of-service alarm
CN110855522A (en) Electric energy meter communication quality detection method based on analytic hierarchy process
CN115713038A (en) Distribution and utilization fault detection method and system based on deep circulation neural network
CN114444886A (en) Analysis method based power distribution network damage probability evaluation method under typhoon disaster
US20210168638A1 (en) Methods, apparatus and computer-readable mediums relating to detection of sleeping cells in a cellular network
EP3836599A1 (en) Method for detecting permanent failures in mobile telecommunication networks
Segovia et al. Connected system for monitoring electrical power transformers using thermal imaging
CN117749819B (en) Cloud service-oriented disaster recovery data synchronization method and system
WO2019223118A1 (en) Method and apparatus for evaluating reliability of wide area protection system

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: 20200626