CN113283462A - Secondary system fault positioning method based on improved IDNN model - Google Patents
Secondary system fault positioning method based on improved IDNN model Download PDFInfo
- Publication number
- CN113283462A CN113283462A CN202110314718.8A CN202110314718A CN113283462A CN 113283462 A CN113283462 A CN 113283462A CN 202110314718 A CN202110314718 A CN 202110314718A CN 113283462 A CN113283462 A CN 113283462A
- Authority
- CN
- China
- Prior art keywords
- fault
- secondary system
- alarm
- information
- port
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000003745 diagnosis Methods 0.000 claims abstract description 20
- 238000012423 maintenance Methods 0.000 claims abstract description 12
- 238000012549 training Methods 0.000 claims abstract description 12
- 238000007781 pre-processing Methods 0.000 claims abstract description 11
- 238000012545 processing Methods 0.000 claims abstract description 3
- 230000002159 abnormal effect Effects 0.000 claims description 25
- 230000003287 optical effect Effects 0.000 claims description 19
- 230000005856 abnormality Effects 0.000 claims description 18
- 238000013528 artificial neural network Methods 0.000 claims description 14
- 238000005259 measurement Methods 0.000 claims description 11
- 241000272814 Anser sp. Species 0.000 claims description 10
- 230000006870 function Effects 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 10
- 238000005070 sampling Methods 0.000 claims description 10
- 238000012544 monitoring process Methods 0.000 claims description 8
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000005265 energy consumption Methods 0.000 abstract description 2
- 239000013307 optical fiber Substances 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000001960 triggered effect Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/245—Classification techniques relating to the decision surface
- G06F18/2453—Classification techniques relating to the decision surface non-linear, e.g. polynomial classifier
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- General Health & Medical Sciences (AREA)
- Economics (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Tourism & Hospitality (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Business, Economics & Management (AREA)
- Biomedical Technology (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Nonlinear Science (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Primary Health Care (AREA)
- Remote Monitoring And Control Of Power-Distribution Networks (AREA)
- Alarm Systems (AREA)
Abstract
The invention discloses a secondary system fault positioning method based on an improved IDNN model, which comprises the steps of firstly, triggering and preprocessing an alarm signal of a secondary system fault; then, training and constructing a fault diagnosis model based on IDNN by utilizing a historical fault sample set, and performing pair input on feature set X of fault section of a secondary systemiCarrying out fault position diagnosis; and finally, outputting a fault positioning result, reporting the fault positioning result to operation and maintenance personnel for processing, and simultaneously storing the fault positioning result to a historical fault sample set of a background database. The fault positioning method sets trigger pretreatment before the fault diagnosis module, thereby avoiding errorsThe interference of information reduces the energy consumption of the system; the target for fault location of the secondary equipment is detected to the module layer from the equipment layer, so that the economical efficiency of operation and maintenance of the secondary equipment is greatly improved; and a fault section feature set is formed by utilizing the running state of the secondary equipment and the alarm signal, and the fault position obtained after the fault section feature set is processed by the fault diagnosis model is more accurate.
Description
Technical Field
The invention relates to the technical field of secondary systems of intelligent substations, in particular to a fault positioning method of a secondary system of an intelligent substation based on an improved IDNN model.
Background
With the development of science and technology, a substation automation system goes through the development process from a centralized RTU (remote terminal unit), a distributed system, a network-based monitoring system and a digital substation to the current intelligent substation, and the automation level of the substation is higher and higher.
In recent years, the construction of the intelligent power grid in China has risen to the height of the national strategic level, and the intelligent substation is one of the core platforms for realizing energy conversion and control in the strong intelligent power grid construction, and the planning construction of the intelligent substation has been spread comprehensively. The intelligent substation is based on the IEC61850 standard, advanced technologies such as digital sampling, intelligent primary equipment and optical fiber Ethernet are widely adopted, the reliability and the intelligent level of the substation are greatly improved, and meanwhile, new requirements are provided for debugging and testing of the substation.
The primary system of the transformer substation is a system consisting of a generator, a power transmission line, a transformer, a breaker and other equipment for power generation, power transmission, transformation, power distribution and the like; the main body of the power system is used for reducing the voltage of the electric energy generated by the generator step by step to be transmitted to the power distribution system through the power transmission and transformation equipment, and then distributing the electric energy to users through the distribution lines. The secondary System of the substation is a System composed of relay protection, safety automatic Control, System communication, scheduling automation, a Distributed Control System (DCS) and the like; the secondary system is an important component which is indispensable to the power system, and is used for realizing the contact monitoring and control of people and the primary system, so that the primary system can run safely and economically.
The basic functions of the secondary system of the intelligent substation are to automatically complete information acquisition, measurement, control, protection and the like, so that fault positions and types can be divided according to information sources, information sinks, intermediate equipment and optical fiber links.
The information source and the information sink comprise typical secondary equipment, such as a merging unit, an intelligent terminal, a protection device and the like, in the past, fault location for the secondary equipment mostly reaches an equipment layer, but the purpose of fault location is detected to a module layer from the equipment layer in consideration of the fact that the hardware design of the current device basically follows the modularization idea, and the economical efficiency of operation and maintenance of the secondary equipment is greatly improved.
The main failure of the switch as the intermediate equipment is mainly the I/O port abnormality or the complete machine shutdown caused by power loss and the like.
The faults occurring in the optical fiber link include protection device channel faults, GOOSE/SV network link breakage and the like.
For a long time, when a secondary system of a transformer substation breaks down and needs to be overhauled, generally, operation personnel are needed to feed back alarm information to overhaul personnel, the overhaul personnel carry out manual analysis through the alarm information, and then an overhaul strategy is formulated according to an analysis result. However, with the popularization of the intelligent substation, on one hand, the information amount of the substation has been increased by times, a large amount of digital interface information is added, it becomes unrealistic that operation and maintenance personnel want to accurately know the detailed meanings of various kinds of alarm information, on the other hand, with the increase of the information amount, errors of manual fault analysis are increased, because the equipment which gives an alarm is not necessarily fault equipment, nor is the fault position, the alarm information on the surface is probably caused by various reasons, the position and the type of the fault can be determined more accurately only by fully mastering the operation state of each equipment in the secondary system of the substation, however, the misleading of the alarm information on the surface causes the deviation of the manual fault analysis, and the accuracy and the efficiency of fault maintenance of the secondary system of the substation are influenced.
Due to the fact that the intelligent substation equipment is various in types and the communication network is complex, multiple types of faults may occur at different positions in the secondary system.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the intelligent substation secondary system aims at the problems that the secondary system of the intelligent substation is increasingly complex, various types of faults and positions in the secondary system are likely to occur at different positions, and the faults and the positions are difficult to accurately identify.
To solve the above technical problems.
The invention is realized by the following technical scheme:
the invention provides a secondary system fault positioning method based on an improved IDNN model, which comprises the following steps:
s1, triggering and preprocessing the alarm signal of the secondary system fault;
s2, training and constructing fault diagnosis model based on IDNN by utilizing historical fault sample set, and performing fault diagnosis on input secondary system fault section feature set XiCarrying out fault position diagnosis;
and S3, outputting a fault positioning result, reporting the fault positioning result to operation and maintenance personnel for processing, and simultaneously storing the fault positioning result in a historical fault sample set of the background database.
In order to eliminate the interference of a small amount of error information, a trigger preprocessing process of a trigger diagnosis module is set, when the background monitors and detects that the total number of alarm signals is greater than a threshold value, a secondary system fault positioning module is triggered, when the number of the alarm signals is not greater than the threshold value due to the occurrence of the error alarm signals, the system can track to a sending end and then correct, so that the interference of some misoperation or error information can be avoided, the diagnosis module cannot be triggered, and the power consumption is reduced.
Further, preferably, in the method for locating a secondary system fault based on the IDNN model, the triggering preprocessing indicates that S2 is performed when the total number N of alarm signals detected by the background monitoring is greater than or equal to a threshold value; when the background monitoring detects that the total number N of the alarm signals is smaller than a threshold value, tracking the sending end and correcting; the threshold is set to be the minimum value of the total number of the warning signals at the historical fault moment.
Further, the preferable scheme is that the secondary system fault positioning method based on the improved IDNN model is characterized in that the data feature set X of the secondary system fault sectioniUnder the condition of secondary system fault, the alarm signal of the secondary system fault and the operation state data of the secondary system typical equipment are obtained after data preprocessing.
Further, a preferred embodiment is that, the secondary system fault location method based on the improved IDNN model is characterized in that the alarm signal of the secondary system fault includes: switching value abnormality alarm, sampling abnormality alarm and device abnormality alarm;
the main alarm information of the abnormal switching value alarm signal comprises the following steps: alarm information related to GOOSE;
the main alarm information of the sampling abnormal alarm signal comprises the following steps: SV-related warning information;
the main alarm information of the device abnormal alarm signal comprises: and (5) equipment self-checking abnormal alarm information.
Further, preferably, in the method for locating a fault of a secondary system based on an improved IDNN model, the secondary system typical device in S3 includes: the system comprises a merging unit, a protection device, an intelligent terminal, a measurement and control device and a switch;
the equipment state information of the merging unit, the protection device, the intelligent terminal and the measurement and control device comprises: the power module outputs voltage in a direct current mode, and a port receives optical power;
the state information of the switch includes: power supply voltage, port message flow.
Further, in a preferred embodiment, in the method for locating a secondary system fault based on the improved IDNN model, the data preprocessing process is to unify the operation state information data of the typical device of the secondary system and the alarm signal data of the secondary system fault into a general judgment parameter format.
The format of the judgment parameter of the state information data is as follows: when the state data of the typical equipment port in the secondary system is in a specified normal range, the corresponding position element in the feature set is 0, and when the state data of the typical equipment port in the secondary system is not in the specified normal range, the corresponding position element in the feature set is 1;
the format of the judgment parameter of the alarm signal is as follows: when the equipment sends an alarm signal, the element value of the position of the specific alarm information corresponding to the alarm signal is 1, and the element value of the position of the alarm information which does not send the alarm signal is 0.
In a further preferred embodiment, the secondary system fault location method based on the improved IDNN model is characterized in that the data feature set X of the fault sectioniThe expression is:
Xi＝{ZtXi,GjXi},i＝1,2,...N
in the formula: xiRepresenting a set of data features at the occurrence of an ith fault;
ZtXirepresenting the running state information of the secondary equipment;
GjXiindicating secondary system abnormal alarm information;
n represents the total number of faults.
Further preferably, in the method for locating a fault in a secondary system based on the improved IDNN model, the operation state information ZtX of the secondary device isiThe operation state of typical equipment in a secondary system is integrated, and the expression is as follows:
in the formula: MU (Multi-user)iThe operation state of the merging unit at the ith fault is shown, including voltage U and portOptical power Pop (subscript port number);
PRithe operating state of the protection device at the ith fault is represented, and the operating state comprises a voltage U and port optical power Pop (subscript is port number);
ITithe method comprises the steps that the operation state of the intelligent terminal in the ith fault is shown, wherein the operation state comprises voltage U and port optical power Pop (subscript is port number);
MCithe operation state of the measurement and control device in the ith fault is represented, and the operation state comprises a voltage U and port optical power Pop (subscript is port number);
SWithe method comprises the steps that the operation state of the switch at the ith fault is represented, and the operation state comprises message information Mes (subscript is a message number), wherein the message information comprises flow statistics of the message flowing through a switch Port (subscript is a Port number);
further preferred scheme is that the secondary system fault positioning method based on the improved IDNN model is secondary system abnormal alarm information GjXiA series of alarm signals received by background monitoring during fault are integrated, and the expression is as follows:
in the formula: asvai、Aaosvi、DaaiSwitching value abnormality alarm, sampling value abnormality alarm and device abnormality alarm during the ith fault; asvai、Aaosvi、DaaiThe warning information contains concrete warning information Asmes, Aames and Dames, namely warning information related to GOOSE, warning information related to SV and equipment self-checking abnormal warning information.
Further preferably, the method for locating a fault in a secondary system based on an improved IDNN model is characterized in that, for example, the lower limit of a normal range specified by the optical power of a port j of a certain secondary device is (30dBm, -30dBm), and if the optical power received by the port is found to be-40 dBm, Pop is determinedjWhen the optical power received by the port is found to be 20dBm, Pop is 1j＝0。
For example, a secondary device receives GOOSE message interrupt and sends out chain-breaking alarm, and the alarm signal is at AsvaiThe position element in (1) is AsmesjThen AsmesjIf the alarm signal does not appear, Asmes is 1j＝0。
Further preferably, the method for positioning the fault of the secondary system based on the improved IDNN model is characterized in that the construction of the IDNN-based fault diagnosis model comprises the following steps:
a1, inputting the stored historical fault data feature set as an input layer into a deep neural network by a background database, and inputting the fault data feature set X into the deep neural networki＝{x1,x2,...xpP is the total number of the characteristic information;
a2, using a mini-batch learning mode, and simultaneously performing learning training on an input layer by using a cross entropy through a loss function;
a3, the output layer is the secondary system fault type code Y ═ Y1,y2,...,yi,...yqQ is the total number of secondary system fault types, yiWhether the fault is of the ith type or not is indicated, and if yes, y i1, otherwise yi＝0。
And A4, obtaining a historical optimal model after training.
The loss function of the final improved deep neural network is:
in the formula: theta is a network parameter;
n is the total number of data;
tnqthe value of the qth element for the nth data (the data tag of the actual failure);
ynqis the value of the q-th element of the nth data (diagnostic result of the neural network output).
A Drop-out mechanism is added into the model, in addition, an Adam algorithm and an exponential decay method of the learning rate are introduced to update parameters, and finally the optimization process of the network parameter theta is as follows.
mt＝β1mt-1+(1-β1)gt
In the formula: gtIs the gradient of the network parameter;
mtis an exponential moving average of the gradient;
vtis an exponential moving average of the square of the gradient;
β1、β2and beta3The exponential decay rates are respectively 0.9, 0.999 and 0.95;
α0is the initial value of the learning rate; the epoch _ num is the current training frequency;
batch _ size is a batch parameter.
For a long time, when a secondary system of a transformer substation breaks down and needs to be overhauled, generally, operation personnel are needed to feed back alarm information to overhaul personnel, the overhaul personnel carry out manual analysis through the alarm information, and then an overhaul strategy is formulated according to an analysis result. However, with the popularization of the intelligent substation, on one hand, the information amount of the substation has been increased by times, a large amount of digital interface information is added, it becomes unrealistic that operation and maintenance personnel want to accurately know the detailed meanings of various kinds of alarm information, on the other hand, with the increase of the information amount, errors of manual fault analysis are increased, because the equipment which gives an alarm is not necessarily fault equipment, nor is the fault position, the alarm information on the surface is probably caused by various reasons, the position and the type of the fault can be determined more accurately only by fully mastering the operation state of each equipment in the secondary system of the substation, however, the misleading of the alarm information on the surface causes the deviation of the manual fault analysis, and the accuracy and the efficiency of fault maintenance of the secondary system of the substation are influenced. Aiming at the problems in the prior art, the fault positioning method of the secondary system of the intelligent substation provided by the invention firstly summarizes the main fault positions and fault types of the secondary system according to the information flow of a typical secondary system, trains out a historical optimal model by a historical fault sample set and introduces the historical optimal model into a fault diagnosis module; secondly, a fault section data feature set is provided by using the running state of the secondary equipment and an alarm signal; in the case of a secondary system failure, two types of characteristic information are combined: and obtaining a data characteristic set representing a fault section based on the characteristic information of the running state of the secondary equipment and the characteristic information of the abnormal alarm of the secondary system, and outputting a fault position after the data characteristic set of the fault section is processed by a fault diagnosis model, wherein the output result of the fault position is more accurate.
In the past, fault location for secondary equipment mostly reaches an equipment layer, but the hardware design of the current device basically follows the modularization idea, so that the target of fault location is detected to a module layer from the equipment layer, and the economical efficiency of operation and maintenance of the secondary equipment is greatly improved.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. the invention provides a secondary system fault positioning method based on an improved IDNN model, wherein a trigger preprocessing module is arranged in front of a fault diagnosis module, so that the interference of error information is avoided, and the energy consumption of a system is reduced;
2. the invention provides a secondary system fault positioning method based on an improved IDNN model, which summarizes main fault positions and fault types of a secondary system, and utilizes the running state of secondary equipment and an alarm signal to form a fault section characteristic set, wherein the fault section characteristic set is more accurate in fault position obtained after being processed by a fault diagnosis model;
3. the invention provides a secondary system fault positioning method based on an improved IDNN model, which aims at the problem that the target of fault positioning of secondary equipment is detected to a module layer from an equipment layer, and greatly improves the economical efficiency of operation and maintenance of the secondary equipment.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention.
In the drawings:
FIG. 1 is a flow chart of the method for establishing a secondary system fault diagnosis framework.
Fig. 2 is a schematic diagram of various types of faults that may occur at different locations in a secondary system.
Fig. 3 is a schematic diagram of a DNN topology with n hidden layers.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to examples and accompanying drawings, and the exemplary embodiments and descriptions thereof are only used for explaining the present invention and are not used as limitations of the present invention.
Examples
The basic functions of the secondary system of the intelligent substation are to automatically complete information acquisition, measurement, control, protection and the like, so that fault positions and types can be divided according to information sources, information sinks, intermediate equipment and optical fiber links.
The information source and the information sink comprise typical secondary equipment, such as a merging unit, an intelligent terminal, a protection device and the like, in the past, fault location for the secondary equipment mostly reaches an equipment layer, but the purpose of fault location is detected to a module layer from the equipment layer in consideration of the fact that the hardware design of the current device basically follows the modularization idea, and the economical efficiency of operation and maintenance of the secondary equipment is greatly improved.
The main failure of the switch as the intermediate equipment is mainly the I/O port abnormality or the complete machine shutdown caused by power loss and the like.
The faults occurring in the optical fiber link include protection device channel faults, GOOSE/SV network link breakage and the like.
Due to the fact that the intelligent substation devices are various in types and the communication network is complex, multiple types of faults may occur at different positions in the secondary system, and the main situation is as shown in fig. 2.
1) Secondary equipment running state information
Different from the situation that a single state monitoring device needs to be installed on primary equipment, most of the secondary equipment has online self-checking and communication functions, and therefore the running conditions of the secondary equipment can be known only by adding a part of state quantity statistical function on related equipment. According to the method, partial state information of the secondary equipment is selected as data characteristics, as shown in the following table, when a fault occurs in the secondary system, corresponding state data are obviously different from normal working conditions, and therefore the operation state of each IED in the secondary system can be represented.
The secondary system typical apparatus is: the system comprises a merging unit, a protection device, an intelligent terminal, a measurement and control device and a switch; the equipment state information of the merging unit, the protection device, the intelligent terminal and the measurement and control device comprises: the power module outputs voltage in direct current, and a port receives optical power; the state information of the switch includes: power supply voltage, port message flow.
2) Secondary system abnormal alarm information
When the abnormal condition in the secondary system affects the normal work of the device, the related equipment sends out an alarm signal to reflect the influence of the fault on the equipment. The alarm signals of the secondary system can be divided into abnormal on-off value alarm, abnormal sampling alarm and abnormal device alarm, wherein the abnormal on-off value alarm mainly comprises alarm signals related to GOOSE, the abnormal sampling alarm mainly comprises alarm signals related to SV, and the abnormal device alarm mainly takes equipment self-checking abnormality as the main alarm.
In the case of a secondary system failure, two types of characteristic information are combined: based on the characteristic information of the running state of the secondary equipment and the characteristic information of the abnormal alarm of the secondary system, the data characteristic set of the fault section can be represented, as shown in the formula (1)
Xi＝{ZtXi,GjXi},i＝1,2,...N (1)
In the formula: xiA set of data features representing the occurrence of the ith fault including secondary equipment operating status information ZtXiAnd secondary system anomaly alarm information GjXi(ii) a And N is the total number of faults.
Secondary device operating status information ZtXiThe running state of typical equipment in a secondary system is integrated, and the formula (2) is shown as follows:
in the formula: MU (Multi-user)i、PRi、ITi、MCiAnd SWiThe information is the operation state of the merging unit, the protection device, the intelligent terminal, the measurement and control device and the switch in the ith fault, and comprises voltage U, Port optical power Pop (subscript represents Port number) and message information Mes (subscript represents message number), wherein the message information comprises the flow statistics of the message flowing through the switch Port (subscript represents Port number). Such information should be continuously monitored, and if the failure time value exceeds the specified normal range, the corresponding position element will be affected, for example, the optical power lower limit of a certain secondary equipment port j is-30 dBm, if the port is foundThe received optical power crosses this lower limit, Pop j1, otherwise, PopjThe voltage is equal to 0, and the same as the message information.
Secondary system abnormal warning GjXiA series of alarm signals received by background monitoring during fault are integrated, as shown in formula (3):
in the formula: asvai、Aaosvi、DaaiThe switching value abnormality alarm, the sampling value abnormality alarm and the device abnormality alarm in the ith fault respectively comprise specific alarm information Asmes, Aames and Dames. For example, when a fault occurs, a secondary device receives GOOSE message interrupt and sends out chain-breaking alarm, and the alarm signal is AsvaiThe position element in (1) is AsmesjThen AsmesjIf the alarm signal does not appear, Asmes is 1j＝0。
3) The secondary system fault location modeling deep neural network is composed of an input layer, a hidden layer (1 layer or more) and an output layer, and is essentially a nonlinear classifier. Fig. 3 shows a DNN topology comprising n hidden layers. Wherein the input layer is a fault data characteristic set Xi＝{x1,x2,...xpP is the total number of the characteristic information; the output layer is secondary system fault type code Y ═ Y1,y2,...,yi,...yqQ is the total number of secondary system fault types, yiWhether the fault is of the ith type or not is indicated, and if yes, y i1, otherwise yi＝0。
If a conventional Deep Neural Network (DNN) is adopted, the training time of the model is greatly increased along with the increase of future data samples, so that a mini-batch learning mode is used for improving the conventional DNN, and meanwhile, the cross entropy is adopted as the loss function, so that compared with the conventional DNN, the training speed of the model is greatly increased, and the problem of low back propagation convergence speed is avoided.
The loss function of the final improved deep neural network is shown in equation (4).
In the formula: theta is a network parameter; n is the total number of data; t is tnqThe value of the qth element for the nth data (the data tag of the actual failure); y isnqIs the value of the q-th element of the nth data (diagnostic result of the neural network output).
In order to prevent poor prediction effect of the neural network on test data caused by the overfitting phenomenon, a Drop-out mechanism is added into the model, and the structure of the neural network is changed by randomly discarding part of neurons, so that the tolerance among the neurons is reduced. In addition, in order to improve the defect that the learning rate is constant in the iteration process of the traditional gradient descent method, the Adam algorithm and the learning rate exponential decay method are introduced to update parameters so as to accelerate the convergence speed of the model, and therefore the optimization process of the final network parameter theta is shown in formulas (5) to (11).
mt＝β1mt-1+(1-β1)gt (6)
In the formula: gtIs the gradient of the network parameter; m istIs an exponential moving average of the gradient; v. oftIs an exponential moving average of the square of the gradient;
and
is the corrected amount; beta is a1、β2And beta3The exponential decay rates are respectively 0.9, 0.999 and 0.95; alpha is alpha0Is the initial value of the learning rate; the epoch _ num is the current training frequency; batch _ size is a batch parameter.
4) Based on the above, the secondary system fault diagnosis is established, and the specific steps are as follows:
(1) in order to eliminate the interference of a small amount of error information, a judging process of triggering a diagnosis module is set, when the background monitors that the total number of the alarm signals is larger than a threshold value, a secondary system fault positioning module is triggered, and the threshold value of the process is set to be the minimum value of the total number of the alarm signals at the moment of historical fault.
(2) The equipment state information and the alarm signal when the secondary system fails are called to form a feature set X of a failure sectioni。
(3) Set of features XiAnd sending the input to a trained secondary system fault positioning model based on the IDNN to obtain a fault positioning result.
The above-mentioned embodiments, objects, technical solutions and advantages of the present invention are further described in detail, it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A secondary system fault positioning method based on an improved IDNN model is characterized by comprising the following steps:
s1, triggering and preprocessing the alarm signal of the secondary system fault;
s2, training and constructing fault diagnosis model based on IDNN by using historical fault sample set, and performing pair input on feature set X of fault section of secondary systemiCarrying out fault position diagnosis;
and S3, outputting a fault positioning result, reporting the fault positioning result to operation and maintenance personnel for processing, and simultaneously storing the fault positioning result in a historical fault sample set of the background database.
2. The method for secondary system fault location based on the IDNN model improvement according to claim 1, wherein the triggering preprocessing indicates that S2 is performed when the total number N of alarm signals detected by the background monitor is greater than or equal to a threshold value; when the background monitoring detects that the total number N of the alarm signals is smaller than a threshold value, tracking the sending end and correcting; the threshold is set to the minimum value of the total number of alarm signals at the historical fault moment.
3. The method for secondary system fault location based on improved IDNN model as claimed in claim 1, wherein said set of data feature of secondary system fault section XiUnder the condition of secondary system fault, the alarm signal of the secondary system fault and the operation state data of the secondary system typical equipment are obtained after data preprocessing.
4. The method according to claim 3, wherein the alarm signal of the secondary system fault comprises: switching value abnormality alarm, sampling abnormality alarm and device abnormality alarm;
the main alarm information of the abnormal switching value alarm signal comprises the following steps: alarm information related to GOOSE;
the main alarm information of the sampling abnormal alarm signal comprises the following steps: SV-related warning information;
the main alarm information of the device abnormal alarm signal comprises: and (5) equipment self-checking abnormal alarm information.
5. The method as claimed in claim 3, wherein the secondary system fault location method based on the IDNN model comprises, in S3: the system comprises a merging unit, a protection device, an intelligent terminal, a measurement and control device and a switch;
the equipment state information of the merging unit, the protection device, the intelligent terminal and the measurement and control device comprises: the power module outputs voltage in direct current, and a port receives optical power;
the state information of the switch includes: power supply voltage, port message flow.
6. The method according to claim 3, wherein the data preprocessing process is to unify the running state information data of the typical device in the secondary system and the alarm signal data of the secondary system fault into a general decision parameter format.
The format of the judgment parameter of the state information data is as follows: when the state data of the typical equipment port in the secondary system is in the specified normal range, the corresponding position element in the characteristic set is 0, and when the state data of the typical equipment port in the secondary system is not in the specified normal range, the corresponding position element in the characteristic set is 1;
the format of the judgment parameter of the alarm signal is as follows: when the equipment sends an alarm signal, the element value of the position of the specific alarm information corresponding to the alarm signal is 1, and the element value of the position of the alarm information which does not send the alarm signal is 0.
7. The method for secondary system fault location based on improved IDNN model as claimed in claim 1, wherein said data feature set X of fault sectioniThe expression is:
Xi＝{ZtXi,GjXi},i＝1,2,...N
in the formula: xiRepresenting a set of data features at the occurrence of an ith fault;
ZtXirepresenting the running state information of the secondary equipment;
GjXiindicating secondary system abnormal alarm information;
n represents the total number of faults.
Wherein secondary device operational status information ZtXiThe operation state of typical equipment in a secondary system is integrated, and the expression is as follows:
in the formula: MU (Multi-user)iThe operation state of the merging unit at the ith fault is represented, and the operation state comprises a voltage U and a port optical power Pop (subscript is a port number);
PRithe operation state of the protection device at the ith fault is represented, and the operation state comprises a voltage U and a port optical power Pop (subscript is a port number);
ITithe operating state of the intelligent terminal at the ith fault is represented, wherein the operating state comprises a voltage U and a port optical power Pop (subscript is a port number);
MCithe operation state of the measurement and control device in the ith fault is represented, and the operation state comprises a voltage U and a port optical power Pop (subscript is a port number);
SWithe method comprises the steps that the operation state of the switch at the ith fault is represented, and the operation state comprises message information Mes (subscript is a message number), wherein the message information comprises flow statistics of the message flowing through a switch Port (subscript is a Port number);
secondary system abnormal warning GjXiA series of alarm signals received by background monitoring during fault are integrated, and the expression is as follows:
in the formula: asvai、Aaosvi、DaaiSwitching value abnormality alarm, sampling value abnormality alarm and device abnormality alarm during the ith fault; asvai、Aaosvi、DaaiThe warning information contains concrete warning information Asmes, Aames and Dames, namely warning information related to GOOSE, warning information related to SV and equipment self-checking abnormal warning information.
8. The method of claim 7, wherein j optical power of a secondary device port defines a lower limit of a normal range as (30dBm, -30dBm), and if the optical power received by the port is found to be-40 dBm, Pop is determinedjWhen the optical power received by the port is found to be 20dBm, Pop is 1j＝0。
When some secondary equipment receives GOOSE message and is interrupted, a chain breakage alarm is sent out, and an alarm signal of the chain breakage alarm is at AsvaiThe position element in (1) is AsmesjThen AsmesjIf the alarm signal does not appear, Asmes is 1j＝0。
9. The method for positioning the fault of the secondary system based on the improved IDNN model according to claim 1, wherein the constructing of the fault diagnosis model based on the IDNN model comprises the following steps:
a1, inputting the stored historical fault data feature set as an input layer into a deep neural network by a background database, and inputting the fault data feature set X into the deep neural networki＝{x1,x2,...xpP is the total number of the characteristic information;
a2, using a mini-batch learning mode, and simultaneously performing learning training on an input layer by using a cross entropy through a loss function;
a3, the output layer is the secondary system fault type code Y ═ Y1,y2,...,yi,...yqQ is the total number of secondary system fault types, yiWhether the fault is of the ith type or not is indicated, and if yes, yi1, otherwise yi＝0。
And A4, obtaining a historical optimal model after training.
10. The method of claim 9, wherein the final improved deep neural network loss function formula is as follows:
in the formula: theta is a network parameter;
n is the total number of data;
tnqthe value of the qth element for the nth data (the data tag of the actual failure);
ynqis the value of the q-th element of the nth data (diagnostic result of the neural network output).
A Drop-out mechanism is added into the model, in addition, an Adam algorithm and an exponential decay method of the learning rate are introduced to update parameters, and finally the optimization process of the network parameter theta is as follows.
mt＝β1mt-1+(1-β1)gt
In the formula: gtIs the gradient of the network parameter;
mtis an exponential moving average of the gradient;
vtis an exponential moving average of the square of the gradient;
β1、β2and beta3The exponential decay rates are respectively 0.9, 0.999 and 0.95;
α0is the initial value of the learning rate; the epoch _ num is the current training frequency;
batch _ size is a batch parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110314718.8A CN113283462B (en) | 2021-03-24 | 2021-03-24 | Secondary system fault positioning method based on improved IDNN model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110314718.8A CN113283462B (en) | 2021-03-24 | 2021-03-24 | Secondary system fault positioning method based on improved IDNN model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113283462A true CN113283462A (en) | 2021-08-20 |
| CN113283462B CN113283462B (en) | 2022-09-20 |
Family
ID=77276013
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110314718.8A Active CN113283462B (en) | 2021-03-24 | 2021-03-24 | Secondary system fault positioning method based on improved IDNN model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113283462B (en) |
Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7570105B1 (en) * | 2007-10-04 | 2009-08-04 | Altera Corporation | Variable current charge pump with modular switch circuit |
| CN101826860A (en) * | 2009-03-02 | 2010-09-08 | 恩益禧电子股份有限公司 | Duty-cycle correction circuit and duty ratio correction method |
| US20110141913A1 (en) * | 2009-12-10 | 2011-06-16 | Clemens Joseph R | Systems and Methods for Providing Fault Detection and Management |
| US20110267637A1 (en) * | 2009-01-16 | 2011-11-03 | Ecoaxis Systems Pvt. Ltd. | Automatic register control system with intelligent optical sensor and dry presetting facility |
| CN103616579A (en) * | 2013-11-07 | 2014-03-05 | 山东大学 | Fault diagnosis method used for intelligent substation secondary system |
| CN107491375A (en) * | 2017-08-18 | 2017-12-19 | 国网山东省电力公司信息通信公司 | Equipment detection and fault early warning system and method under a kind of cloud computing environment |
| CN107769972A (en) * | 2017-10-25 | 2018-03-06 | 武汉大学 | A kind of power telecom network equipment fault Forecasting Methodology based on improved LSTM |
| CN108808850A (en) * | 2017-11-28 | 2018-11-13 | 电子科技大学 | Intelligent substation based on IEC61850 alerts intelligent analysis method |
| CN108988496A (en) * | 2018-08-01 | 2018-12-11 | 南方电网科学研究院有限责任公司 | Intelligent substation secondary system fault diagnosis method, device and equipment |
| CN109614981A (en) * | 2018-10-17 | 2019-04-12 | 东北大学 | The Power System Intelligent fault detection method and system of convolutional neural networks based on Spearman rank correlation |
| CN109633351A (en) * | 2018-12-13 | 2019-04-16 | 平安普惠企业管理有限公司 | Intelligent IT O&M Fault Locating Method, device, equipment and readable storage medium storing program for executing |
| CN109726200A (en) * | 2018-12-06 | 2019-05-07 | 国网甘肃省电力公司信息通信公司 | Grid information system fault location system and method based on two-way deep neural network |
| US20190189114A1 (en) * | 2017-12-19 | 2019-06-20 | Industry-University Cooperation Foundation Sogang University | Method for beamforming by using maximum likelihood estimation for a speech recognition apparatus |
| CN110326011A (en) * | 2017-02-21 | 2019-10-11 | 万事达卡国际公司 | Determine the lawful condition calculated at equipment |
| CN110674189A (en) * | 2019-09-27 | 2020-01-10 | 国网四川省电力公司电力科学研究院 | Method for monitoring secondary state and positioning fault of intelligent substation |
| CN111259947A (en) * | 2020-01-13 | 2020-06-09 | 国网浙江省电力有限公司信息通信分公司 | Power system fault early warning method and system based on multi-mode learning |
| CN112346093A (en) * | 2020-10-27 | 2021-02-09 | 合肥工业大学 | Method for repairing BDS cycle slip |
-
2021
- 2021-03-24 CN CN202110314718.8A patent/CN113283462B/en active Active
Patent Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7570105B1 (en) * | 2007-10-04 | 2009-08-04 | Altera Corporation | Variable current charge pump with modular switch circuit |
| US20110267637A1 (en) * | 2009-01-16 | 2011-11-03 | Ecoaxis Systems Pvt. Ltd. | Automatic register control system with intelligent optical sensor and dry presetting facility |
| CN101826860A (en) * | 2009-03-02 | 2010-09-08 | 恩益禧电子股份有限公司 | Duty-cycle correction circuit and duty ratio correction method |
| US20110141913A1 (en) * | 2009-12-10 | 2011-06-16 | Clemens Joseph R | Systems and Methods for Providing Fault Detection and Management |
| CN103616579A (en) * | 2013-11-07 | 2014-03-05 | 山东大学 | Fault diagnosis method used for intelligent substation secondary system |
| CN110326011A (en) * | 2017-02-21 | 2019-10-11 | 万事达卡国际公司 | Determine the lawful condition calculated at equipment |
| CN107491375A (en) * | 2017-08-18 | 2017-12-19 | 国网山东省电力公司信息通信公司 | Equipment detection and fault early warning system and method under a kind of cloud computing environment |
| CN107769972A (en) * | 2017-10-25 | 2018-03-06 | 武汉大学 | A kind of power telecom network equipment fault Forecasting Methodology based on improved LSTM |
| CN108808850A (en) * | 2017-11-28 | 2018-11-13 | 电子科技大学 | Intelligent substation based on IEC61850 alerts intelligent analysis method |
| US20190189114A1 (en) * | 2017-12-19 | 2019-06-20 | Industry-University Cooperation Foundation Sogang University | Method for beamforming by using maximum likelihood estimation for a speech recognition apparatus |
| CN108988496A (en) * | 2018-08-01 | 2018-12-11 | 南方电网科学研究院有限责任公司 | Intelligent substation secondary system fault diagnosis method, device and equipment |
| CN109614981A (en) * | 2018-10-17 | 2019-04-12 | 东北大学 | The Power System Intelligent fault detection method and system of convolutional neural networks based on Spearman rank correlation |
| CN109726200A (en) * | 2018-12-06 | 2019-05-07 | 国网甘肃省电力公司信息通信公司 | Grid information system fault location system and method based on two-way deep neural network |
| CN109633351A (en) * | 2018-12-13 | 2019-04-16 | 平安普惠企业管理有限公司 | Intelligent IT O&M Fault Locating Method, device, equipment and readable storage medium storing program for executing |
| CN110674189A (en) * | 2019-09-27 | 2020-01-10 | 国网四川省电力公司电力科学研究院 | Method for monitoring secondary state and positioning fault of intelligent substation |
| CN111259947A (en) * | 2020-01-13 | 2020-06-09 | 国网浙江省电力有限公司信息通信分公司 | Power system fault early warning method and system based on multi-mode learning |
| CN112346093A (en) * | 2020-10-27 | 2021-02-09 | 合肥工业大学 | Method for repairing BDS cycle slip |
Non-Patent Citations (2)
| Title |
|---|
| R. YANG 等: ""An anti-interference MIMU/GPS vehicle integrated navigation algorithm based on IDNN-EKF"", 《2016 IEEE/ION POSITION, LOCATION AND NAVIGATION SYMPOSIUM》 * |
| 潘元进 等: ""IDNN 的GPS 动态测量多周跳探测与修复模型"", 《测绘科学技术学报》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113283462B (en) | 2022-09-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108628291B (en) | Expert intelligent diagnosis system based on simulation platform in waste incineration plant | |
| CN105914881B (en) | A kind of intelligent substation advanced application system | |
| CN114819415B (en) | Power equipment fault prediction system based on data analysis | |
| CN110287543B (en) | Method for predicting service life of relay protection device | |
| CN108732444B (en) | Intelligent substation protection equipment running state evaluation device and maintenance decision method | |
| CN107453485A (en) | A kind of management and running information method for inspecting and system | |
| CN109116181A (en) | A kind of fault detector failure judgment method based on maximum probability | |
| CN117689214A (en) | Dynamic safety assessment method for energy router of flexible direct-current traction power supply system | |
| CN104267294A (en) | IGIS fault online diagnostic method based on rough set theory and Petri network | |
| CN110265998B (en) | Method and device for checking strategy fixed value of stability control system | |
| CN110443481B (en) | Power distribution automation terminal state evaluation system and method based on hybrid K-nearest neighbor algorithm | |
| CN113283462B (en) | Secondary system fault positioning method based on improved IDNN model | |
| CN109962528A (en) | A kind of substation secondary device intelligence managing and control system | |
| CN106373028A (en) | Electric power maintenance system | |
| CN108539740A (en) | A kind of power grid accident anticipation simulation system and its method | |
| CN114254774A (en) | One-key sequential control-based automatic issuing system for maintenance order associated operation order | |
| CN115689532A (en) | Power system fault analysis method and device | |
| CN113807461A (en) | Transformer fault diagnosis method based on Bayesian network | |
| CN115313621A (en) | Automatic operation and maintenance system and method for intelligent substation based on grey theory | |
| Zhang et al. | Research on intelligent operation and maintenance technology of primary equipment in substation | |
| CN106709158B (en) | Performance improvement method of power grid feeder automation system | |
| Wang | Research on Fault Predition and Self-healing Control Technology of Distribution Network | |
| Yang et al. | Trend Analysis and Fault Diagnosis of Equipment State Based on Transformer Operation Mechanism Modeling | |
| Li et al. | Research on Model-based Condition-based Maintenance Technology of Process-Layer Smart Switch in smart substation | |
| CN111162525A (en) | Power supply unit optimizing system of distribution network |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |