CN106296440A - Based on transformer station's warning information analysis and decision system integrated for ANN and ES and method - Google Patents
Based on transformer station's warning information analysis and decision system integrated for ANN and ES and method Download PDFInfo
- Publication number
- CN106296440A CN106296440A CN201510253001.1A CN201510253001A CN106296440A CN 106296440 A CN106296440 A CN 106296440A CN 201510253001 A CN201510253001 A CN 201510253001A CN 106296440 A CN106296440 A CN 106296440A
- Authority
- CN
- China
- Prior art keywords
- reasoning
- knowledge
- fault
- module
- information
- 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
Links
Classifications
-
- 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
Abstract
Based on transformer station's warning information analysis and decision system integrated for ANN and ES and method, including: data base, the Real-time Alarm information received for storage system and equipment static configuration data；Reasoning module, coordinates the reasoning process of warning information and controls；Explanation module, explains the reasoning results of reasoning module；Implicit expression knowledge extraction module, for by the knowledge explicit expression of numerical expression；Knowledge base, for depositing explicit knowledge's rule and being included in the implicit expression knowledge connected in weights；Human-machine interface module, for being converted into the form of user's understanding and exporting by the information that explanation module and implicit expression knowledge extraction module export；Reasoning based on numerical operation in ANN is introduced ES system, instead of the expression based on symbol that traditional E S is single, improve the execution efficiency of ES；And take different analysis strategies for different faults；Not only increase the maintainability of whole system and fault-tolerance and expand the diagnostic area of warning information analysis and decision system.
Description
Technical field
The present invention relates to a kind of system and method, be specifically related to based on transformer station's warning information analysis and decision system integrated for ANN and ES and method.
Background technology
Along with the scale expanding day of power system, various potential safety hazards are the most serious, and especially simple accident causes power failure on a large scale, the probability of even area power grid paralysis being continuously increased.Under the complexity and requirement of real-time of modern power network, need fast detecting failure, accurate failure judgement during system jam, rationally fix a breakdown, reduce power failure range, to safeguard the stable operation of whole electrical network.Therefore, transformer station break down after how fast searching abort situation, identify that fault category is the key processing power outage.
In order to judge fast and accurately and fix a breakdown, guarantee the safe and stable operation of system, strengthen seriality and the reliability of power supply, realize the comprehensive analysis decision to transformer station's internal fault information, accurately failure judgement type and abort situation, to improve work efficiency, shorten power off time, to reduce loss of outage significant.
And conventional transformer station is after unattended, full detail is aggregated into Surveillance center, shows in chronological order, does not do any reasoning and judging and processes.In recent years, " running greatly " system that proposes requires to realize regulation and control integration, scheduling business blends with monitoring of tools business, thus need to send a large amount of substation data information, various types of signal acts frequently, operations staff's monitor task is heavier, it is easy to omits significant alarm signal, thus is delayed process and causes the accident.Once have an accident, not only exacerbate the sharp increase of semaphore, there is also roll screen, brush screen phenomenon so that operations staff is the most dazzled, at a loss as to what to do, be difficult to concentrate on crucial points, affect the correct process to accident.
nullAt present，Substantial amounts of research work is had been carried out both at home and abroad in terms of warning information analyzing and processing and fault diagnosis，Propose a lot of method，Such as artificial neural network、Specialist system、Petri net、Data mining、Tabu search etc.，When but single intelligent algorithm carries out accident analysis to transformer station，Always there will be reasoning matching conflict、The problem of fault-tolerant ability difference，Easily cause erroneous judgement or fail to judge，And in terms of power system failure diagnostic，Major part is the diagnosis of research power system global fault or the diagnosis of a certain concrete element fault，And seldom study Fault Diagnosis for Substation，The most abundant to the fault-tolerance research of fault diagnosis system，The fault diagnosis system of diagnosis core is made almost without fault-tolerance research especially with specialist system，Do not account for the impact on network structure of the substation operation mode simultaneously yet，Somebody's artificial neural networks is not on the actual application in power system failure diagnostic，Great majority are only as the aid of a kind of off-line.
Therefore need a kind of warning information analysis and decision system of foundation badly, be used for optimizing transformer station's substantial amounts of original alarm information, it is provided that the means of a kind of inductive decision judge the fault occurred, and give an explaination and express.
Summary of the invention
For solving the problems referred to above, the present invention proposes based on the comprehensive analysis and decision system of transformer station's warning information integrated for ANN and ES and method, reasoning based on numerical operation in ANN is introduced ES system, instead of the expression based on symbol that traditional E S is single, thus improve the execution efficiency of ES.
It is an object of the invention to use following technical proposals to realize:
Based on transformer station's warning information analysis and decision system integrated for ANN and ES, described system includes:
Data base, for depositing Real-time Alarm information and the equipment static configuration data that described system receives；
Reasoning module, for coordinating the reasoning process of warning information and control；
Explanation module, for explaining the reasoning results of reasoning module；
Implicit expression knowledge extraction module, for by the knowledge explicit expression of numerical expression；
Knowledge base, for depositing explicit knowledge's rule and being included in the implicit expression knowledge connected in weights；
Human-machine interface module, for being converted into the form of user's understanding and exporting by the information that explanation module and implicit expression knowledge extraction module export.
Preferably, described reasoning module includes numerical operation reasoning element and symbolic logic reasoning element；
Described symbolic logic reasoning element is used for the heuristic computing of symbol；Described numerical operation reasoning element is used for mathematical reasoning computing.
Preferably, described knowledge base includes ES explicit knowledge module and ANN implicit expression knowledge module；Wherein,
Described ES explicit knowledge module is used for transformer station's conventional fault reasoning；
Described ANN implicit expression knowledge module is for depositing and the reasoning results of the knowledge base unmatched fault of rule.
Further, when there is conventional fault in transformer station, described symbolic logic reasoning element, explanation module and the intercommunication of human-machine interface module；System start-up symbolic logic reasoning element, is triggered explanation module, and is exported by human-machine interface module；When there is unconventional fault in transformer station, described numerical operation reasoning element, ANN implicit expression knowledge module, knowledge base and the intercommunication of human-machine interface module；Described system start-up numerical operation reasoning element, is triggered described ANN implicit expression knowledge module and described knowledge base, and is exported by human-machine interface module.
The warning information that data base deposits is after alarm pretreatment rejecting is sent out by mistake and missed warning information, if the knowledge deposited with ES explicit knowledge module matches, directly give ES process, by setting up fault reasoning model, fault reasoning mechanism and selecting inference direction to make inferences, export the reasoning results.
Based on warning information analysis decision method integrated for ANN and ES, described method includes:
(1) warning information input database are received；
(2) warning information described in pretreatment, rejects announcement information by mistake；
(3) effective warning information is extracted；
(4) when pretreated warning information is with when any bar knowledge rule mates in knowledge base, directly made inferences by ES and export result；
(5) when pretreated warning information knowledge rule all with knowledge base neither mates, then ANN process is gone to；
(6) output result is explained, and be knowledge rule by the implicit expression knowledge transformation corresponding to the equipment fault of ANN computing reasoning, be stored in knowledge base.
Preferably, warning information described in described step (2) pretreatment includes, judges according to the remote measurement change of remote signalling displacement, and concrete steps include:
A) the false remote signalling of definition differentiates storehouse；
According to the actual requirements, define described false remote signalling and differentiate storehouse, including plant stand number, remote signalling sequence number, remote measurement sequence number, remote measurement undulating value, remote manipulation time and effective time length；
B) when a certain remote signalling occurs displacement, start this vacation remote signalling and differentiate storehouse, find the remote signalling sequence number of change and corresponding remote measurement sequence number；
C) by this remote measurement value and false remote signalling, the remote measurement value that in acquisition effective time length, described remote measurement sequence number is corresponding from data base, differentiates that in storehouse, remote measurement undulating value compares；If more than remote measurement undulating value, then this remote signalling is normal displacement, otherwise for announcement information by mistake.
Preferably; effective warning information in described step (3), including protection act information, prepared auto restart information, switch changed position information, reclosing information, protection exit information, protection pressing plate information, observing and controlling distant place information on the spot, Threshold Crossing Alert information, communications status information and abnormality alarming information.
Preferably, in described step (4), ES reasoning includes, sets up fault reasoning model and fault reasoning mechanism, and determines inference direction.
Further, described fault reasoning model, including fault type, time window and reasoning enabling signal；
Wherein, described fault type includes simple fault and combined fault；
It is 3～10s that described time window is adjusted；
Described reasoning enabling signal includes protection act signal, breaker actuation signal and accident resultant signal.
Further, described fault reasoning mechanism uses the multiple inference mechanism that individual event reasoning combines with multiple correlating event reasonings.
Further, described inference direction includes, selects forward and reverse mixed inference mode, first forward reasoning to propose it is assumed that then backward reasoning confirms it is assumed that and export the reasoning results；Detailed process includes:
11-1 searches for knowledge base, extract the knowledge rule mated with the fault message of data base: first extract the premise part of the arbitrary knowledge rule of knowledge base, contrast with the fault message of data base, if described fault message comprises this premise part, the match is successful for this knowledge rule, and adds it in knowledge set of matches；Otherwise, next knowledge rule coupling, circulation operation are carried out；
BP neural network algorithm is used to make inferences for the fault that it fails to match；
11-2 carries out ES reasoning to each knowledge rule in knowledge set of matches successively；
The conclusion that reasoning is drawn by 11-3 puts into dynamic data base, and the fault message for reaching a conclusion uses BP neural network algorithm to make inferences；Until no longer producing new conclusion.
The premise part of all knowledge rules is added fault as new hypothesis and assumes set by 11-4, until fault assumes that in set, a certain bar is assumed to set up, if all of assuming all to be false, system enters BP neural computing program.
Preferably, in described step (5), ANN carries out process and includes, arranges the main protection of substation equipment before treatment and the probability coefficent of back-up protection malfunction, tripping is ai, the fault safety factor of protection device is θi, the integrated protection coefficient of equipment is obtained according to following formula
Switch changed position that this FACTOR P is associated together with equipment and fault message are together as the input information of ANN.
Further, described input information is carried out ANN computing, including；
Selected neural network structure, sets all threshold values and is connected weights as the relatively fractional value being evenly distributed；
Use forward-propagating and error back propagation, this neural network structure is learnt.
Further, described forward-propagating includes, input signal is through sigmoid functionSuccessively forward-propagating, is exported information processing result by output layer, if actual output is not inconsistent with expectation, then carries out error back propagation.
Further, described error back propagation includes, the error of output signal is along backtracking, the mode declined by error gradient revises weights and the threshold value of each layer neuron, through repeatedly propagating, each layer weights constantly adjust so that within signal errors is down to claimed range；
When the output result trying to achieve certain equipment after ANN computing is 1, then device fails being described, otherwise this equipment is normal.
Compared with prior art, the present invention reaches to provide the benefit that:
1) transformer station's warning information analysis and decision system of present invention design provides the effective means of a kind of quick fault location, the intelligent inference algorithm integrated for ANN with ES used is compared other alarm information processing methods and is had the advantage that speed is fast, accuracy is high and fault-tolerance is strong, thus be conducive to seizing the first chance of substation intelligent alarm, have the right of speech in this field.
2) present invention uses remote signalling displacement and the alarm preprocessing means of remote measurement change associating judgement, differentiate storehouse by increasing false remote signalling, from the mechanism mechanism of research warning signal by mistake, can effectively suppress the mistake taken place frequently to send out warning signal by mistake from root.
3) intelligent substation warning information optimization and diagnostics architecture are advantageously formed, operator on duty quickly can be grasped the key link information from substantial amounts of original alarm signal, operations staff is greatly reduced and monitors pressure, it is greatly improved the level of intelligence of whole system, meets intelligent substation from now on and use few people, unwatched pattern.
4) this method for designing and scheduling combine, and having promoted stands based on boss works in coordination with research and the application that interactive distributed intelligence alerts, and is conducive to the development that preferably support regulation and control are integrated.
5) the present invention is directed to different faults and take different analysis strategies；Not only increase the maintainability of whole system and fault-tolerance and expand the diagnostic area of warning information analysis and decision system.
Accompanying drawing explanation
Fig. 1 is structural representation based on transformer station's warning information analysis and decision system integrated for ANN and ES；
Fig. 2 is based on transformer station's warning information analysis decision method flow diagram integrated for ANN and ES；
Fig. 3 is ES inference method flow chart.
Detailed description of the invention
Below in conjunction with the accompanying drawings the detailed description of the invention of the present invention is described in further detail.
As it is shown in figure 1, based on transformer station's warning information analysis and decision system integrated for ANN and ES, described system includes:
Data base, for depositing Real-time Alarm information and the equipment static configuration data that described system receives；
Reasoning module, for coordinating the reasoning process of warning information and control；
Explanation module, for explaining the reasoning results of reasoning module；
Implicit expression knowledge extraction module, for by the knowledge explicit expression of numerical expression；
Knowledge base, for depositing explicit knowledge's rule and being included in the implicit expression knowledge connected in weights；
Human-machine interface module, for being converted into the form of user's understanding and exporting by the information that explanation module and implicit expression knowledge extraction module export.
Described reasoning module includes numerical operation reasoning element and symbolic logic reasoning element；
Described symbolic logic reasoning element is used for the heuristic computing of symbol；Described numerical operation reasoning element is used for mathematical reasoning computing.
Described knowledge base includes ES explicit knowledge module and ANN implicit expression knowledge module；Wherein,
Described ES explicit knowledge module is used for transformer station's conventional fault reasoning；
Described ANN implicit expression knowledge module is for depositing and the reasoning results of the knowledge base unmatched fault of rule.
When there is conventional fault in transformer station, described symbolic logic reasoning element, explanation module and the intercommunication of human-machine interface module；System start-up symbolic logic reasoning element, is triggered explanation module, and is exported by human-machine interface module；When there is unconventional fault in transformer station, described numerical operation reasoning element, ANN implicit expression knowledge module, knowledge base and the intercommunication of human-machine interface module；Described system start-up numerical operation reasoning element, is triggered described ANN implicit expression knowledge module and described knowledge base, and is exported by human-machine interface module.
The warning information that data base deposits is after alarm pretreatment rejecting is sent out by mistake and missed warning information, if the knowledge deposited with ES explicit knowledge module matches, directly give ES process, by setting up fault reasoning model, fault reasoning mechanism and selecting inference direction to make inferences, export the reasoning results.
As in figure 2 it is shown, based on warning information analysis decision method integrated for ANN and ES, described method includes:
(1) warning information input database are received；
(2) warning information described in pretreatment, rejects announcement information by mistake；Warning information described in described step (2) pretreatment includes, judges according to the remote measurement change of remote signalling displacement, and concrete steps include:
A) the false remote signalling of definition differentiates storehouse；
According to the actual requirements, define described false remote signalling and differentiate storehouse, including plant stand number, remote signalling sequence number, remote measurement sequence number, remote measurement undulating value, remote manipulation time and effective time length；
B) when a certain remote signalling occurs displacement, start this vacation remote signalling and differentiate storehouse, find the remote signalling sequence number of change and corresponding remote measurement sequence number；
C) by this remote measurement value and false remote signalling, the remote measurement value that in acquisition effective time length, described remote measurement sequence number is corresponding from data base, differentiates that in storehouse, remote measurement undulating value compares；If more than remote measurement undulating value, then this remote signalling is normal displacement, otherwise for announcement information by mistake.
(3) effective warning information is extracted；Effective warning information in described step (3), including protection act information, prepared auto restart information, switch changed position information, reclosing information, protection exit information, protection pressing plate information, observing and controlling distant place information on the spot, Threshold Crossing Alert information, communications status information and abnormality alarming information.
(4) when pretreated warning information is with when any bar knowledge rule mates in knowledge base, directly made inferences by ES and export result；In described step (4), ES reasoning includes, sets up fault reasoning model and fault reasoning mechanism, and determines inference direction.
Described fault reasoning model, including fault type, time window and reasoning enabling signal；
Wherein, described fault type includes simple fault and combined fault；
It is 3～10s that described time window is adjusted；
Described reasoning enabling signal includes protection act signal, breaker actuation signal and accident resultant signal.
Described fault reasoning mechanism uses the multiple inference mechanism that individual event reasoning combines with multiple correlating event reasonings.
Described inference direction includes, selects forward and reverse mixed inference mode, first forward reasoning to propose it is assumed that then backward reasoning confirms it is assumed that and export the reasoning results；Detailed process includes:
11-1 searches for knowledge base, extract the knowledge rule mated with the fault message of data base: first extract the premise part of the arbitrary knowledge rule of knowledge base, contrast with the fault message of data base, if described fault message comprises this premise part, the match is successful for this knowledge rule, and adds it in knowledge set of matches；Otherwise, next knowledge rule coupling, circulation operation are carried out；
BP neural network algorithm is used to make inferences for the fault that it fails to match；
As it is shown on figure 3,11-2 carries out ES reasoning to each knowledge rule in knowledge set of matches successively；
The conclusion that reasoning is drawn by 11-3 puts into dynamic data base, and the fault message for reaching a conclusion uses BP neural network algorithm to make inferences；Until no longer producing new conclusion.
The premise part of all knowledge rules is added fault as new hypothesis and assumes set by 11-4, until fault assumes that in set, a certain bar is assumed to set up, if all of assuming all to be false, system enters BP neural computing program.
(5) when pretreated warning information knowledge rule all with knowledge base neither mates, then ANN process is gone to；
In described step (5), ANN carries out process and includes, arranges the main protection of substation equipment before treatment and the probability coefficent of back-up protection malfunction, tripping is ai, the fault safety factor of protection device is θi, the integrated protection coefficient of equipment is obtained according to following formulaSwitch changed position that this FACTOR P is associated together with equipment and fault message are together as the input information of ANN.
Described input information is carried out ANN computing, including；
Selected neural network structure, sets all threshold values and is connected weights as the relatively fractional value being evenly distributed；
Use forward-propagating and error back propagation, this neural network structure is learnt.
Described forward-propagating includes, input signal is through sigmoid functionSuccessively forward-propagating, is exported information processing result by output layer, if actual output is not inconsistent with expectation, then carries out error back propagation.
Described error back propagation includes, the error of output signal revises weights and the threshold value of each layer neuron along backtracking, the mode declined by error gradient, and through repeatedly propagating, each layer weights constantly adjust so that within signal errors is down to claimed range；
When the output result trying to achieve certain equipment after ANN computing is 1, then device fails being described, otherwise this equipment is normal.
(6) output result is explained, and be knowledge rule by the implicit expression knowledge transformation corresponding to the equipment fault of ANN computing reasoning, be stored in knowledge base.
Finally should be noted that: above example is only in order to illustrate the technical scheme of the application rather than restriction to its protection domain; although the application being described in detail with reference to above-described embodiment; those of ordinary skill in the field are it is understood that those skilled in the art still can carry out all changes, amendment or equivalent to the detailed description of the invention of application after reading the application; these changes, amendment or equivalent, it is all within the right that its application is awaited the reply.
Claims (15)
1. based on transformer station's warning information analysis and decision system integrated for ANN and ES, it is characterised in that described system includes:
Data base, for depositing Real-time Alarm information and the equipment static configuration data that described system receives；
Reasoning module, for coordinating the reasoning process of warning information and control；
Explanation module, for explaining the reasoning results of reasoning module；
Implicit expression knowledge extraction module, for by the knowledge explicit expression of numerical expression；
Knowledge base, for depositing explicit knowledge's rule and being included in the implicit expression knowledge connected in weights；
Human-machine interface module, for being converted into, by the information that explanation module and implicit expression knowledge extraction module export, the form that user understands
And export.
2. the system as claimed in claim 1, it is characterised in that described reasoning module includes numerical operation reasoning element and symbol
Logical reasoning unit；
Described symbolic logic reasoning element is used for the heuristic computing of symbol；Described numerical operation reasoning element is used for mathematical reasoning computing.
3. system as claimed in claim 2, it is characterised in that described knowledge base includes that ES explicit knowledge module and ANN are hidden
Formula knowledge module；Wherein,
Described ES explicit knowledge module is used for transformer station's conventional fault reasoning；
Described ANN implicit expression knowledge module is for depositing and the reasoning results of the knowledge base unmatched fault of rule.
4. system as claimed in claim 3, it is characterised in that when transformer station occurs conventional fault, described symbolic logic pushes away
Reason unit, explanation module and the intercommunication of human-machine interface module；System start-up symbolic logic reasoning element, triggers and explains mould
Block, and exported by human-machine interface module；When there is unconventional fault in transformer station, described numerical operation reasoning element, ANN
The intercommunication of implicit expression knowledge module, knowledge base and human-machine interface module；Described system start-up numerical operation reasoning element, touches
Send out ANN implicit expression knowledge module and described knowledge base described, and exported by human-machine interface module；
The warning information that data base deposits is rejected through alarm pretreatment and is sent out by mistake after warning information by mistake, if deposited with ES explicit knowledge module
The knowledge put matches, and directly gives ES process, by setting up fault reasoning model, fault reasoning mechanism and selecting inference direction
Make inferences, export the reasoning results.
5. based on transformer station's warning information analysis decision method integrated for ANN and ES, it is characterised in that described method includes:
(1) warning information input database are received；
(2) warning information described in pretreatment, rejects announcement information by mistake；
(3) effective warning information is extracted；
(4) when pretreated warning information is with when in knowledge base, any bar knowledge rule mates, directly made inferences by ES
And export result；
(5) when pretreated warning information knowledge rule all with knowledge base neither mates, then ANN process is gone to；
(6) output result is explained, and by the implicit expression knowledge transformation corresponding to equipment fault of ANN computing reasoning for knowing
Know rule, be stored in knowledge base.
6. method as claimed in claim 5, it is characterised in that warning information described in described step (2) pretreatment includes,
Remote measurement change according to remote signalling displacement judges, concrete steps include:
A) the false remote signalling of definition differentiates storehouse；
According to the actual requirements, define described false remote signalling and differentiate storehouse, including plant stand number, remote signalling sequence number, remote measurement sequence number, remote measurement fluctuation
Value, remote manipulation time and effective time length；
B) when a certain remote signalling occurs displacement, start this vacation remote signalling and differentiate storehouse, find the remote signalling sequence number of change and corresponding remote measurement sequence
Number；
C) the remote measurement value that in acquisition effective time length, described remote measurement sequence number is corresponding from data base, sentences this remote measurement value with false remote signalling
In other storehouse, remote measurement undulating value compares；If more than remote measurement undulating value, then this remote signalling is normal displacement, otherwise for accusing by mistake
Information.
7. method as claimed in claim 5, it is characterised in that effective warning information in described step (3), including protection
Action message, prepared auto restart information, switch changed position information, reclosing information, protection exit information, protection pressing plate information, observing and controlling
Distant place information on the spot, Threshold Crossing Alert information, communications status information and abnormality alarming information.
8. method as claimed in claim 5, it is characterised in that in described step (4), ES reasoning includes, sets up fault and pushes away
Reason model and fault reasoning mechanism, and determine inference direction.
9. method as claimed in claim 8, it is characterised in that described fault reasoning model, including fault type, time window
Mouth and reasoning enabling signal；
Wherein, described fault type includes simple fault and combined fault；
It is 3～10s that described time window is adjusted；
Described reasoning enabling signal includes protection act signal, breaker actuation signal and accident resultant signal.
10. as claimed in claim 8 based on warning information analysis decision method integrated for ANN and ES, it is characterised in that institute
State the multiple inference mechanism that fault reasoning mechanism uses individual event reasoning to combine with multiple correlating event reasonings.
11. methods as claimed in claim 8, it is characterised in that described inference direction includes, selects forward and reverse mixed inference
Mode, first forward reasoning propose it is assumed that then backward reasoning confirms it is assumed that and export the reasoning results；Detailed process includes:
11-1 searches for knowledge base, extracts the knowledge rule mated with the fault message of data base: first extract the arbitrary knowledge of knowledge base
The premise part of rule, contrasts with the fault message of data base, if comprising this premise part in described fault message, this is known
Know rule match success, and add it in knowledge set of matches；Otherwise, next knowledge rule coupling, circulation operation are carried out；
BP neural network algorithm is used to make inferences for the fault that it fails to match；
11-2 carries out ES reasoning to each knowledge rule in knowledge set of matches successively；
The conclusion that reasoning is drawn by 11-3 puts into dynamic data base, uses BP nerve net for the fault message that cannot reach a conclusion
Network algorithm makes inferences；Until no longer producing new conclusion.
The premise part of all knowledge rules is added fault as new hypothesis and assumes set, until fault is assumed in set by 11-4
Till a certain bar is assumed to set up, if all of assuming all to be false, system enters BP neural computing program.
12. want the method as described in 5 such as right, it is characterised in that in described step (5), ANN carries out process and includes, are processing
The probability coefficent of the front main protection arranging substation equipment and back-up protection malfunction, tripping is ai, the fault of protection device is reliably
Number is θi, the integrated protection coefficient of equipment is obtained according to following formula
Switch changed position that this FACTOR P is associated together with equipment and fault message are together as the input information of ANN.
13. methods as claimed in claim 12, it is characterised in that described input information is carried out ANN computing, including；
Selected neural network structure, sets all threshold values and is connected weights as the relatively fractional value being evenly distributed；
Use forward-propagating and error back propagation, this neural network structure is learnt.
14. methods as claimed in claim 13, it is characterised in that described forward-propagating includes, input signal is through sigmoid
FunctionSuccessively forward-propagating, is exported information processing result by output layer, if actual output is not inconsistent with expectation, then enters
Row error back propagation.
15. methods as claimed in claim 14, it is characterised in that described error back propagation includes, the error of output signal
Along backtracking, the mode declined by error gradient revises weights and the threshold value of each layer neuron, and through repeatedly propagating, each layer is weighed
Value constantly adjusts so that within signal errors is down to claimed range；
When the output result trying to achieve certain equipment after ANN computing is 1, then device fails being described, otherwise this equipment is normal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510253001.1A CN106296440A (en) | 2015-05-18 | 2015-05-18 | Based on transformer station's warning information analysis and decision system integrated for ANN and ES and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510253001.1A CN106296440A (en) | 2015-05-18 | 2015-05-18 | Based on transformer station's warning information analysis and decision system integrated for ANN and ES and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106296440A true CN106296440A (en) | 2017-01-04 |
Family
ID=57631236
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510253001.1A Pending CN106296440A (en) | 2015-05-18 | 2015-05-18 | Based on transformer station's warning information analysis and decision system integrated for ANN and ES and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106296440A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109557414A (en) * | 2018-11-30 | 2019-04-02 | 国家电网有限公司技术学院分公司 | Integrated power system fault diagnosis alarming processing system and method |
| CN109902373A (en) * | 2019-02-21 | 2019-06-18 | 国网山东省电力公司临沂供电公司 | A kind of area under one's jurisdiction Fault Diagnosis for Substation, localization method and system |
| CN110531742A (en) * | 2019-09-16 | 2019-12-03 | 重庆华能水电设备制造有限公司 | A kind of generator current collecting equipment real time monitoring and method for diagnosing faults |
| 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 |
| CN110687473A (en) * | 2019-09-27 | 2020-01-14 | 国网四川省电力公司电力科学研究院 | Fault positioning method and system for relay protection test of intelligent substation |
| CN113743764A (en) * | 2021-08-26 | 2021-12-03 | 贵州电网有限责任公司 | Spare power automatic switching action judgment method for dispatching master station |
-
2015
- 2015-05-18 CN CN201510253001.1A patent/CN106296440A/en active Pending
Non-Patent Citations (1)
| Title |
|---|
| 王保民: "智能变电站故障信息综合分析决策系统的研究与应用", 《中国优秀硕士学位论文全文数据库信息科技II辑》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109557414A (en) * | 2018-11-30 | 2019-04-02 | 国家电网有限公司技术学院分公司 | Integrated power system fault diagnosis alarming processing system and method |
| CN109557414B (en) * | 2018-11-30 | 2021-04-13 | 国家电网有限公司技术学院分公司 | Fault diagnosis alarm processing system and method for integrated power system |
| CN109902373A (en) * | 2019-02-21 | 2019-06-18 | 国网山东省电力公司临沂供电公司 | A kind of area under one's jurisdiction Fault Diagnosis for Substation, localization method and system |
| CN109902373B (en) * | 2019-02-21 | 2023-06-23 | 国网山东省电力公司临沂供电公司 | Fault diagnosis and positioning method and system for district transformer substation |
| CN110531742A (en) * | 2019-09-16 | 2019-12-03 | 重庆华能水电设备制造有限公司 | A kind of generator current collecting equipment real time monitoring and method for diagnosing faults |
| 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 |
| CN110687473A (en) * | 2019-09-27 | 2020-01-14 | 国网四川省电力公司电力科学研究院 | Fault positioning method and system for relay protection test of intelligent substation |
| CN110687473B (en) * | 2019-09-27 | 2021-08-03 | 国网四川省电力公司电力科学研究院 | Fault positioning method and system for relay protection test of intelligent substation |
| CN110674189B (en) * | 2019-09-27 | 2022-05-17 | 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 |
| CN113743764A (en) * | 2021-08-26 | 2021-12-03 | 贵州电网有限责任公司 | Spare power automatic switching action judgment method for dispatching master station |
| CN113743764B (en) * | 2021-08-26 | 2024-04-05 | 贵州电网有限责任公司 | Method for judging spare power automatic switching action of scheduling master station |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106296440A (en) | Based on transformer station's warning information analysis and decision system integrated for ANN and ES and method | |
| CN107358366B (en) | Distribution transformer fault risk monitoring method and system | |
| CN103336222B (en) | Power system fault diagnosis method based on fuzzy reasoning pulse neurolemma system | |
| Peng et al. | Fault diagnosis of power systems using fuzzy tissue-like P systems | |
| CN107346466A (en) | A kind of control method and device of electric power dispatching system | |
| CN103793853A (en) | Overhead power transmission line running state assessment method based on bidirectional Bayesian network | |
| CN110296124A (en) | Remote failure diagnosis system and method based on expert system | |
| CN109490702A (en) | A kind of method for diagnosing faults based on adaptive optimization pulse nerve membranous system | |
| CN109445411A (en) | A kind of Substation fault diagnosis | |
| CN110175324A (en) | A kind of operation of power networks operational order method of calibration and system based on data mining | |
| CN112415331A (en) | Power grid secondary system fault diagnosis method based on multi-source fault information | |
| Liu et al. | Improvement of fault diagnosis efficiency in nuclear power plants using hybrid intelligence approach | |
| CN112926752B (en) | Intelligent power transmission network fault line detection and fault recovery method considering energy storage system | |
| Singh et al. | Power system fault diagnosis using fuzzy decision tree | |
| CN106802599A (en) | A kind of diagnosing fault of power transformer system based on expert database | |
| CN117034149A (en) | Fault processing strategy determining method and device, electronic equipment and storage medium | |
| Yang et al. | Research on the evaluation of public emergency management intelligence capability in probabilistic language environment | |
| CN110333709A (en) | A kind of rolling process fault diagnosis method and system | |
| Ma et al. | A probabilistic cascading failure model for dynamic operating conditions | |
| CN112147974B (en) | Alarm root cause diagnosis method based on chemical process knowledge automation | |
| Wei et al. | Short text data model of secondary equipment faults in power grids based on LDA topic model and convolutional neural network | |
| Xiao et al. | Research on the construction and implementation of power grid fault handling knowledge graphs | |
| Lin et al. | A method of satellite network fault synthetic diagnosis based on C4. 5 algorithm and expert knowledge database | |
| Jingyu et al. | Statistical analysis of distribution network fault information based on multi-source heterogeneous data mining | |
| Xinjie et al. | A Construction Method for the Knowledge Graph of Power Grid Supervision Business |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170104 |