CN109146093A - A kind of electric power equipment on-site exploration method based on study - Google Patents
A kind of electric power equipment on-site exploration method based on study Download PDFInfo
- Publication number
- CN109146093A CN109146093A CN201810895997.XA CN201810895997A CN109146093A CN 109146093 A CN109146093 A CN 109146093A CN 201810895997 A CN201810895997 A CN 201810895997A CN 109146093 A CN109146093 A CN 109146093A
- Authority
- CN
- China
- Prior art keywords
- data
- power equipment
- prediction model
- information
- fault
- 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 34
- 238000003745 diagnosis Methods 0.000 claims abstract description 37
- 238000007405 data analysis Methods 0.000 claims abstract description 28
- 230000008929 regeneration Effects 0.000 claims abstract description 6
- 238000011069 regeneration method Methods 0.000 claims abstract description 6
- 230000007613 environmental effect Effects 0.000 claims description 42
- 238000003860 storage Methods 0.000 claims description 25
- 238000012549 training Methods 0.000 claims description 6
- 230000000295 complement effect Effects 0.000 claims description 4
- 239000000284 extract Substances 0.000 claims description 2
- 238000012423 maintenance Methods 0.000 abstract description 9
- 230000002159 abnormal effect Effects 0.000 abstract description 8
- 230000008859 change Effects 0.000 abstract description 8
- 230000000694 effects Effects 0.000 abstract description 6
- 238000009826 distribution Methods 0.000 description 28
- 238000011835 investigation Methods 0.000 description 19
- 239000000306 component Substances 0.000 description 16
- 238000012544 monitoring process Methods 0.000 description 12
- 208000024891 symptom Diseases 0.000 description 10
- 238000004458 analytical method Methods 0.000 description 9
- 238000004891 communication Methods 0.000 description 9
- 230000005540 biological transmission Effects 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 238000007689 inspection Methods 0.000 description 7
- 230000009466 transformation Effects 0.000 description 7
- 230000004888 barrier function Effects 0.000 description 5
- 238000001514 detection method Methods 0.000 description 5
- 230000005611 electricity Effects 0.000 description 5
- 238000010801 machine learning Methods 0.000 description 4
- 230000006855 networking Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000013480 data collection Methods 0.000 description 3
- 230000017525 heat dissipation Effects 0.000 description 3
- 239000012212 insulator Substances 0.000 description 3
- 206010011906 Death Diseases 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 2
- 239000008358 core component Substances 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000003066 decision tree Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000009413 insulation Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000010224 classification analysis Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 239000000356 contaminant Substances 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000001012 protector Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000011144 upstream manufacturing Methods 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
Classifications
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—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
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
A kind of electric power equipment on-site exploration method based on study, Data Analysis Services module are configured as the operating mode handled according to the following steps data set: extracting the fault signature of data set and establish prediction model based on fault signature to determine the first information that power equipment breaks down；Fault type is determined based on fault signature and generates the second information including at least fault solution, in the case where the second information failed regeneration, Data Analysis Services module and inline diagnosis module communicate to connect and generate second information by inline diagnosis module in the way of mutually accessing, wherein, in the case where the second information failed regeneration, the associated data based on fault type corresponding with the second information corrects the prediction model.The present invention can identify the abnormal sign of equipment and carry out early warning, can provide corresponding measure in advance to avoid failure generation, reach the passive maintenance of change and overhaul into active, change non-programmed halt is the technical effect of planned shut-down.
Description
Technical field
The invention belongs to technical field of power systems more particularly to a kind of electric power equipment on-site sides of reconnoitring based on study
Method.
Background technique
With constantly expanded using extra-high voltage as the power grid scale of skeleton, trans-regional transmission line of electricity over long distances increases rapidly, intelligence
Energy transmission of electricity, intelligent scheduling scale are growing day by day, and electric network composition, which optimizes first stage upgrading, to be completed, and just need matched daily fortune
Row maintenance, maintenance and fault pre-alarming are adapted therewith.Electric inspection process is that one of guarantee power equipment safety stable operation is daily
Production work, it can improve the reliability of power equipment, it is ensured that equipment is under minimum failure rate state and runs.Power equipment covers
Lid wide scope, many kinds of, quantity is big, and current existing power equipment inspection generallys use the work of manual patrol, manual typing
Make method.There are great human interference factors for which, that is, are directed to identical equipment fault, are based on patrol officer's working experience
Difference may obtain entirely different inspection as a result, serious consequence will be generated to the erroneous detection of equipment fault.
In order to reduce the false detection rate of inspection and the workload of inspection, the electric network fault inline diagnosis system based on various data sources
System comes into being.When grid collapses, fast and accurately fault diagnosis can effectively shorten power off time, reduce faulty section
Domain is reduced to the adverse effect of user.It mainly include expert system, people currently, having much for the method for electric network failure diagnosis
Artificial neural networks, fuzzy theory, optimisation technique, petri network, rough set theory etc..But all there is certain lack in these methods
It falls into, as expert system knowledge base is formed, complicated, diagnosis speed is slow, fault-tolerant ability is poor, and artificial neural network convergence rate is slow, lacks
Interpretability etc. all can not fast and accurately obtain diagnostic result.
In addition, in traditional electric power equipment on-site exploration method, although have been disclosed using data acquisition module come
The state of power equipment is observed, and carries out early warning if necessary.However, this early warning is all subsequent, and usually need
Want artificial intervention that can just make correct solution.The patent document of Publication No. CN106646030A discloses one kind and is based on
The electric network failure diagnosis method in multi-data source and Expert Rules library, comprising the following steps: receive each announcement in data collection system
The signal that alert source is sent；Suspect device is determined according to the signal, if judgement knows that the signal of the suspect device meets in advance
If Fault diagnosis expert rule base in the diagnostic rule classified belonging to suspect device, it is determined that have the suspect device for failure
Equipment, and the comprehensive intelligent analysis into data collection system is alerted with alarm module push.Above-mentioned power grid diagnostic system is to line
The alarm on road carries out after the failure occurred, can not solve the problems, such as the early warning that current device faces, can not identify and set
Standby abnormal sign simultaneously carries out early warning.It can not be provided takes corresponding measure to avoid failure generation in advance, and it is passive to reach change
Maintenance is actively overhauls, and change non-programmed halt is the technical effect of planned shut-down.Meanwhile to solve the above-mentioned problems, energy
It is enough that the data from power equipment are analyzed, but this alanysis is generally also all only limited to subsequent analysis, lacks online
Diagnosis capability, especially its analysis result often excessively lag, and the diagnosis of timeliness cannot be made for field condition, very
The experience of technical staff is depended in big degree.
On the other hand, in the state of the art, online diagnosing technique of support shaft has been disclosed, but inline diagnosis is often straight
It connects read failure code on the data bus or fault message existing for current system is read by data-interface, then according to event
Barrier code or fault message extract diagnostic result relevant to this code or information out of online library, online library cannot self more
Newly, it is to rely on collecting and analyze afterwards for operator, is unfavorable for controlling in face of system complicated and changeable, such as power distribution
System.
Summary of the invention
Word " module " as used herein describes any hardware, software or combination thereof, is able to carry out and " mould
The associated function of block ".
For the deficiencies of the prior art, the present invention provides a kind of electric power equipment on-site exploration method based on study, is based on
The respective operation data of several power equipments of data collecting module collected to form data set, matched by Data Analysis Services module
Being set to according to the following steps to be handled the data set: extracting the fault signature of the data set and based on the event
Barrier feature establishes prediction model to determine the first information that power equipment breaks down.Extracting fault signature not from data set is
From system read failure code, more it is different from downloading fault message offline by data-interface, but Data Analysis Services module is logical
Cross the mode of self study at least based on the fault data made a definite diagnosis that it has been obtained, to current data concentrate it is potential or it is current
It is analyzed through existing abnormal data and predictive or supervisory export fault signature, this fault signature can represent very
Real failure, it is also possible to belong to erroneous judgement.Belong to two modelings process based on the prediction model that predictive fault signature is established again,
The predictive feature in fault signature can largely be reinforced, also just because of prediction fault signature uncertainty,
The prediction result of prediction model after two modelings needs to ensure its accuracy by other steps.Such as it adopts in the present invention
Subsequent judgement is carried out with the second information.It determines fault type based on the fault signature and generates and include at least failure solution party
Second information of case, in the case where the second information failed regeneration, the Data Analysis Services module and inline diagnosis mould
Block communication connection simultaneously generates second information by the inline diagnosis module in the way of mutually accessing, wherein described
In the case where second information failed regeneration, based on the associated associated data correction of fault type corresponding with second information
The prediction model.Associated data associated with fault type is that owned by inline diagnosis module according to what fault signature determined
Fault type and its corresponding solution.The fault signature determined by prediction model exists for the specific of the fault type
Just the second information can be generated in the case where solution.But for doubtful fault signature or the failure classes not occurred
Type, due to not with the presence of specific solution, so corresponding second information can not be generated.It is needed at this time by examining mould online
The accuracy for the predictive fault signature that the actual conditions of data of the block based on acquisition are obtained with inline diagnosis.Pass through inline diagnosis
The second information that module generates can be with predictability.For example, in the case where the second information failed regeneration, failure at this time
Do not occur really, is based only on prediction model and predicts the risk occurred with failure, but specific failure can not be obtained
The data of type and corresponding solution.At this point, the data acquired by inline diagnosis module based on prediction model and in real time
It carries out inline diagnosis and specific fault type and solution is determined with progress.
According to a kind of preferred embodiment, the first information includes at least the time and probability that failure occurs, wherein
In the case where the setting for completing desired estimation range, the Data Analysis Services module at least obtains described according to the following steps
The first information: power equipment specified by estimation range based on expectations determines main association object associated with it and secondary association pair
As.It obtains the first historical data of the main association object and forms the first data set in the way of extracting fault signature, obtain
It takes the second historical data of the secondary association object and first data set is corrected to obtain based on second historical data
Second data set, establishes the first prediction model based on second data set, wherein determines institute based on first prediction model
State the first information.
According to a kind of preferred embodiment, first historical data and second historical data are set including at least electric power
Standby operational parameter data, running state data and environmental factor data, wherein the second historical data of the secondary association object
It is at least determined in accordance with the following steps: determining the environmental factor index value of power equipment based on the environmental factor data, and
Environmental factor index value is divided into n subinterval, calculates fault rate of the power equipment in each index value section；
Index value sequence is determined based on the mean value in each subinterval, and environmental factor index value and failure are calculated based on distance correlation algorithm
Related coefficient between probability of happening, and determine the maximum first environment factor of correlation coefficient value.The secondary association object be by
The first environment factor influences maximum power equipment.
According to a kind of preferred embodiment, the third historical data of the specified power equipment is obtained to extract failure spy
The mode of sign forms third data set, and establishes the second prediction model based on the third data set, wherein based on described the
In the case that two prediction models obtain prediction result, the first information is corrected based on the prediction result.
According to a kind of preferred embodiment, first prediction model and second prediction model are in complementary manner
Combination forecasting is formed to complete the amendment based on the prediction result to the first information, wherein in the first information
Probability value be greater than the probability value in the prediction result in the case where, first prediction model and second prediction model
The combination forecasting is formed in such a way that mixed proportion is greater than one.
According to a kind of preferred embodiment, second information at least obtains according to the following steps: information management stores mould
The operation data of power equipment is classified as the first operation data and the second operation data by block, wherein at least by described
One operation data is stored in the way of classification storage to fault data module based on the difference of fault type.The inline diagnosis
Module can generate second information according to the fault type in a manner of Artificial Diagnosis, wherein second information is pressed
It stores according to the corresponding associated mode of fault type to the fault data module.
According to a kind of preferred embodiment, the main association object at least follows the steps below determination: based on described
Specified power equipment determines several first associate devices being directly connected, and is based on the desired estimation range,
The first environment factor is obtained by the distance correlation algorithm.The main association object is by the first environment factor
Maximum first associate device.
According to a kind of preferred embodiment, the calculation formula of the distance correlation algorithm are as follows:
Wherein, the index value sequence that X is made of the mean value in each subinterval of environmental factor index value, the sequence that Y is made of failure rate
Column, dCov (X, Y) indicate the distance between environmental factor index and failure rate covariance, dVar (X) dVar (Y) expression environment because
The covariance of plain index and failure rate.
According to a kind of preferred embodiment, the probability value in the first information is less than the probability value in the prediction result
In the case of, first prediction model and second prediction model form the combination in such a way that mixed proportion is less than one
Prediction model, wherein the secondary association object is to be influenced maximum second associate device by the first environment factor, through described
First associate device and the specified power equipment are indirectly connected several power equipments and limit second associate device.
According to a kind of preferred embodiment, the Data Analysis Services module obtains first letter according further to following steps
Breath: power equipment specified by estimation range based on expectations determines main association object and secondary association object associated with it.It obtains
It takes the first historical data of the main association object and forms the first data set in the way of extracting fault signature, described in acquisition
Second historical data of secondary association object simultaneously corrects first data set based on second historical data to obtain the second number
According to collection, the first prediction model is established based on second data set, obtains the third historical data to extract fault signature
Mode forms the third data set, and establishes second prediction model based on the third data set.Training described first
Prediction model and second prediction model simultaneously determine optimal the first prediction model and optimal the based on prediction result respectively
Two prediction models.The first information is determined based on the first optimal prediction model, and based on the second optimal prediction model
Prediction result corrects the first information.
Advantageous effects of the invention:
(1) present invention can identify the abnormal sign of equipment and carry out early warning, can provide in advance corresponding measure to avoid
Failure occur and reduce after failure occurs caused by loss, reach the passive maintenance of change actively to overhaul, change non-programmed halt into
The technical effect of planned shut-down.
(2) present invention can be carried out when the failure to power equipment is predicted based on the estimation range that user specifies
Accurate prediction, during prediction, fully considered power equipment itself factor and relation factor associated with it, be based on
The factor of power equipment itself establishes the second prediction model, and establishes the first prediction model, the second prediction mould based on relation factor
The prediction result of type and the first prediction model is exported according to complementary modified mode, so that prediction result is more accurate.
(3) prediction model of the present invention can be based on can be based on new power equipment and new in the operational process of power equipment
Failure carry out autonomous update to correct its hyper parameter and weighted value so that its prediction result is in optimum state always.
Detailed description of the invention
Fig. 1 is the modularization connection relationship diagram of currently preferred electric power equipment on-site investigation system.
Reference signs list
1: data acquisition module 2: information storage tube manages module 3: Data Analysis Services module
4: power equipment 5: communication module 6: fault data module
7: inline diagnosis module
Specific embodiment
It is described in detail with reference to the accompanying drawing.
Embodiment 1
Fig. 1 shows the modularization connection relationship diagram of electric power equipment on-site investigation system.As shown in Figure 1, of the invention
A kind of electric power equipment on-site investigation system based on study is provided, includes at least data acquisition module 1, information storage tube manages module
2 and data analysis and processing module 3.Several power equipments 4 can be communicatively coupled to electric power equipment on-site investigation system with reality
Now to the real time monitoring of power equipment operating status.Electric power equipment on-site investigation system sets electric power by the realization of communication module 5
Standby data access enables the operating condition data of power equipment automatically or based on requirements for access by communication module
It is transferred in investigation system and is analyzed and processed.For example, power equipment can be by being arranged in power equipment erecting bed
Cloud gateway realizes the communication connection with investigation system, and power equipment is in the way of physics, WLAN or radio link
It is connected to cloud gateway.The operating parameter of power equipment include at least with locating for the associated operation data of power equipment itself and its
Environmental data, wherein can communicatedly connect in power equipment with communication module with the associated operation data of power equipment itself
Investigation system is transmitted in the case where connecing, for example, the continuous working period of power equipment itself, current operating parameter or system are matched
Investigation system can not be directly transferred under the effect of the assisted acquisition of external sensor by setting parameter.E.g. environment temperature, when
The surrounding enviroment data of the power equipment in preceding geographical location, ambient humidity or Changes in weather trend can be by being arranged corresponding biography
The mode of sensor or image acquisition device obtains.It just can be real-time for example, temperature sensor can be set in the distribution box of power grid
The temperature changing trend inside distribution box is obtained, setting GPS positioning chip just can be realized to the geographical position coordinates locating for it
Determination.For example, monitor the quantitative filthy state of insulator salt density by installation optical sensor, because insulator is in cleaning, greatly
Part luminous energy is propagated in the core of optical waveguide, and covering of the small part luminous energy along core packet interface transmits, the loss during light propagation
Very little.When filthy, contaminant particles play the role of absorption and scattering to luminous energy, and luminous energy is caused to be lost, and sensor is in insulator
Under same environment, quantitative detection dunghill can be achieved the purpose that by loss of the detection light in transmission process.For example, logical
Double spectrum mountain fire survey meters are crossed, round-the-clock 360 ° of continual level for 24 hours, vertical ± 45 ° of comprehensive transmission line of electricity regions are carried out
Mountain fire detection.
Preferably, the multiple electric power of all grid equipments e.g. in certain area coverage or setting in the factory are set
It is standby to need by configuring several communication modules and several data collectors complete reality to all devices operating status
When monitor.Information storage tube reason module is configured according to class otherwise to the operation shape of the power equipment of data collector acquisition
State data carry out classification storage, can be convenient for the management of data, calling and inquiry by classification storage.For example, for setting
Several distribution boxs within the scope of setting regions assign every station power distribution when installing distribution box and by its typing investigation system
Case is uniquely numbered to distinguish different distribution boxs.The relevant data of every station power distribution case operating status with its respective number
Corresponding mode carries out classification storage.For example, individual memory space, relative example is arranged in the distribution box that number is " 1 "
Geographical location in this way or the data in continuous working period are associated storage with number in the way of associated storage, are needing to look into
When seeing the operating condition of distribution box, corresponding number need to be only inputted in a manner of keyword just can access relevant data.It is preferred that
, information storage tube reason module can for example, by be api interface receive and store other data acquisition equipments acquisition monitoring
Data can be led to the data of acquisition based on intelligent terminal for more needing manually to participate in data caused by monitoring process
It crosses api interface and is transmitted in information storage tube reason module and carry out unified storage.Information storage tube manages module can depositing magnanimity
Storage data are classified, standardized and are stored, and the effective information data after storage are presented to the people of demand according to the actual situation
Member, bodies and agencies.Preferably, e.g. the existing system of administrative department, government department or power department can also be by connecing
The mode access information memory management module of mouth connection is to realize the shared of data information.
Preferably, Data Analysis Services module can be based on the operating status phase of the power equipment of data collecting module collected
Data are closed to model with the abnormal scene to power equipment.Data Analysis Services module can based on data acquisition module when
Between a large amount of monitoring information data for accumulating in gradient, to the O&M of power equipment in the way of e.g. automatic Iterative machine learning
Situation is predicted, for example, Data Analysis Services module can be realized the service life of e.g. transmission line wire by taking power grid as an example
Prediction, fitting life prediction, tower are inclined prediction or Channel Prediction.
Preferably, electric power equipment on-site investigation system further includes fault data module 6, and fault data module can be in determination
In the case that power equipment breaks down, data information relevant to electrical equipment fault is carried out unified storage and is converted to certainly
Plan tree carries out the rule of fault diagnosis to be formed.For example, fault data module can by the number of the power equipment to break down,
Device type or title, failure symptom, failure cause and solution are stored in the way of forming table, then by table
Be converted to the Failure Diagnostic Code that fault data module is just capable of forming after decision tree.Fault data module can be configured as by
Data information relevant to electrical equipment fault needed for being obtained according to following manner:
S1: relevant data information is intercepted from internet big data based on the mode of networking.For example, electric power equipment on-site
Investigation system can be, for example, administration of power networks platform or powernet monitoring platform by private network or internet access, access
Platform disclose the most common failure to e.g. distribution box, failure symptom and usual solution in the case where, number of faults
According to the most common failure of the announced distribution box of module acquisition and usual solution and stored.Work as Data Analysis Services
When operating status defined by operating status related data of the module monitors to data collecting module collected meets failure symptom, just
It can judge fault type caused by distribution box, and push solution for service personnel.Preferably, based on the mode of networking
Relevant data information is intercepted from internet big data, fault data module can determine the solution of partial fault, but
When there is new failure, since no new failure will generate the failure of the second information in fault data module.
S2: electric power equipment on-site investigation system can be acquired directly during monitoring in real time to power equipment
The data relevant to failure that power equipment generates in the process of running.The relevant data of failure include at least the class of power equipment
Environmental factor before type, the time of failure generation, the place of failure generation, failure generation is directly related when occurring to failure
Other power equipments abnormal conditions.For example, having e.g. fuse, breaker, wave by taking distribution box as an example, in distribution box
The several electronic component such as surge protector and radiator leads to distribution box not when the electronic component in distribution box breaks down
When can work normally, electric power equipment on-site investigation system just can obtain the when and where of the distribution box to break down first.?
In the case where monitoring that e.g. fuse breaks down, can not represent fuse failure is that distribution box is caused to break down
Basic reason.The abnormal work of electronic component usually by periphery environmental factor and other components associated with it is comprehensive
Group photo is rung.For example, break down may be related with the temperature of its working environment for fuse, when the temperature in distribution box is super in summer
The case where crossing limiting temperature just most likely results in the failure of fuse, adopts in the temperature sensor by being arranged in distribution box
Collecting in the case that its real-time environment temperature goes beyond the limit of temperature just can be judged as distribution for the basic reason that distribution box generates failure
Case heat dissipation is bad.The operating status with the radiator for the heat dissipation for influencing fuse is monitored again in turn, when discovery radiator is in mistake
Just the first cause that distribution box generates failure directly can be attributed to radiator when imitating working condition and operation irregularity occurs, in turn
Provide the distribution box maintenance program of emphasis maintenance radiator.By acquiring environmental factor relevant to failure and associate power equipment
Operating condition so that electric power equipment on-site investigation system can not only determine that specifically which platform power equipment generates failure, more
It can be accurate to that specifically which electronic component generates failure, so that the diagnostic result of failure is more accurate.For example, also
Can use automatic classification technology and classification analysis carried out to the thousands of a data that sensor collection uploads, can find with
The relevant variable of failure.Such as when the temperature of power equipment slowly increases, air pressure increases therewith, and the amplitude shaken also can be slow
It is slow to increase.When the data of above-mentioned variable are more than the threshold value of diagnostic rule, system is by automatic trigger pre-warning signal and makes related event
Barrier prompt.
S3: it usually will appear various failures during power equipment operation, not all failure is equal
It can correctly be detected by investigation system, for example, the Unrecorded fault type in the fault data module of investigation system.At this point,
The determination of electrical equipment fault needs to complete by field service personnel, and field service personnel is completed to electrical equipment fault really
Its fault type, failure symptom and solution can be uploaded to fault data module in the way of networking after fixed and solution
In saved.Preferably, power equipment investigation system further includes inline diagnosis module 7, in e.g. patrol officer in inspection
When discovery power equipment generates failure in the process, patrol officer can judge the failure of discovery now and carry out helpdesk
Reason, when patrol officer can not judge the root of failure, can access in the way of networking for example, by being mobile terminal
Inline diagnosis module 7.Inline diagnosis module is the inline diagnosis platform being made of veteran equipment expert, and patrol officer can
Directly to be exchanged by inline diagnosis module with equipment expert, by describing field condition and failure symptom to equipment expert
Etc. online direction carried out by equipment expert after information to determine basic reason and corresponding solution that failure generates.Number of faults
Related data in inline diagnosis module progress failure diagnostic process can be obtained according to module to be stored.Preferably, fault data
Module can be data and update library, and the data relevant to failure that fault data module receives can be carried out according to fault type
Classification storage, and then the list of a failure and symptom can be formed, when power equipment breaks down, user only needs to retrieve
Symptom in list just can determine that fault type and solution.List is configured to the operating mode being updated, for example,
When occurring the description about the most common failure of new power equipment, symptom and solution in internet, list can be based on
Data in internet are automatically updated.Fault data module can be continuously updated new failure over time,
The large database concept with abundant fault type is just capable of forming by the accumulation of a period of time.Data, which update in library, is stored with magnanimity
Fault data information, comparison judgement by way of just can quickly obtain power equipment fault type and solve hand
Section.
Preferably, the operation data of power equipment is classified as the first operation data and the second fortune by information management memory module
Row data, wherein at least store the first operation data to failure in the way of classification storage based on the difference of fault type
Data module.First operation data is the normal operation data of power equipment, and the second operation data is that power equipment breaks down
When generated misoperation data.Inline diagnosis module can generate the second letter according to fault type in a manner of Artificial Diagnosis
Breath, wherein the second information is stored according to the corresponding associated mode of fault type to fault data module.
In order to make it easy to understand, the working principle of investigation system of the invention is discussed in detail.
Data acquisition module is monitored the operating status of power equipment in the way of real-time data collection, acquisition
Monitoring data are transmitted in information storage tube reason module and carry out unified storage.The monitoring number of magnanimity in information storage tube reason module
According to storage is classified, Data Analysis Services module just can be in a manner of being, for example, tendency chart the processing to same class data
The operating status of power equipment is intuitively shown.It is for instance possible to obtain the core component of power equipment is becoming at any time
The wave pattern of the current or voltage of change.The tendency chart that environment temperature locating for core component changes over time can be obtained.Therefore
Barrier data module is built-in with the common failure symptom of power equipment and solution just in the initial state to form initial determine
Plan tree-model.Data Analysis Services module to the data of power equipment in the process of processing, Data Analysis Services module
The data that processing obtains are compared with the data in fault data module, in the matched situation of the two, electric power is generated and sets
The standby predictive information that can be broken down in the coming period of time, and prevention inspection is sent to staff by way of early warning
Legislature is built, power equipment is timely overhauled before breaking down.Alternatively, after the failure occurred, scene
Service personnel or staff are by way of the symptom of typing failure, type and/or scene to fault data module
Access retrieval, and fault data module can feed back the solution of failure automatically just for its reference.When new electric power is set
In the case where coming into operation, new fault type also can be following, for new fault type, can pass through inline diagnosis
Module provides online diagnosis, analysis, guidance and/or summary.
Embodiment 2
The present embodiment is the further improvement to embodiment 1, and duplicate content repeats no more.
Data Analysis Services module can carry out storage management to data, and can be acquired based on machine learning algorithm to data
The real-time stream of module acquisition carries out modeling analysis.Data Analysis Services module can establish prediction model to power equipment
Failure predicted.For example, utilizing automatic Iterative machine learning on the basis of accumulating a large amount of monitoring information data for a long time
Can incline to e.g. transmission line wire life prediction, fitting life prediction, tower the power transmission line road transport of prediction and Channel Prediction
Dimension situation is predicted.Specifically, can be by the various sensors of installation on electric power line pole tower, sensor can will for example
It is shaft tower tilt data, microclimate data, icing data, shaft tower shuffle data store, the conducting wire sag data, insulation on transmission line of electricity
Sub- filth monitoring data, image information or video information data are transmitted in Data Analysis Services module.It is defeated by monitoring in real time
The running state data of electric line, when discovery potential faults being capable of automatic trigger alarm or repairing order.
Preferably, Data Analysis Services module can also analyze data in the way of time series to obtain its variation rule
Rule, and then can predict the possibility occurrence of electrical equipment fault.For example, the time occurred by failure to same equipment,
The analysis of the composite factors such as state, fault type when failure occurs just can be realized the prediction that failure occurs.Failure occurs
When state include at least ambient condition locating for the equipment that breaks down, electric current, the voltage, temperature of the equipment to break down itself
Degree or other running parameters, running parameter state and ambient condition with the associated other equipment of equipment to break down.Specifically
, it is frequently not the working condition in continuous operation within e.g. one month time cycle by taking power transformation device as an example.
Power transformation device may be out of service due to its own failure, power-off maintenance or associated line fault.Inside power transformation device
The life cycle of electronic component often just generates failure after continuous operation a certain period of time, so that power transformation device
Generation time based on specific fault caused by electronic component end-of-life has certain regularity, data analysis module
Above-mentioned temporal regularity can be determined by the time that fault type and failure occur, and then the specific fault can be shifted to an earlier date
Prediction.Meanwhile and not all change apparatus failure be all based on caused by the end-of-life of its own electronic component, power transformation
The status consideration that device generates when failure possible breakdown occurs is related.For example, summer climate tropical temperatures are higher, power load is sharply
Increase, transformer is run under full load condition for a long time or even overload state.When data analysis module is according to data acquisition module
Block, which collects environmental status data locating for equipment and predicts transformer, to be in the state of long-term overload or high oil temperature,
Just it can predict in a certain interval range, insulation ag(e)ing causes the system failure or electrical a possibility that puncturing.
Preferably, Data Analysis Services module is configured as carrying out prediction point to fault occurrences according to following step
Analysis.
S1: prediction model is established.
Data Analysis Services module can be configured as establishing the operating mode of several prediction models simultaneously.Prediction model
Can be Kalman prediction model, BP neural network prediction model, based on regression prediction method building model, using more
Kind prediction according to etc. power combination or differential weights combination constitute combination forecasting.Prediction model can be set according to actual electric power
The particularity of standby failure predication carries out flexible choice, for example, prediction model can be according to the actual situation from grey forecasting model, difference
Value with fitting prediction model, time series predicting model, model-naive Bayesian, decision-tree model, Markov prediction model,
Flexible choice is carried out in difference equation prediction model and differential equation prediction model.
S2: extracting the fault signature in the operation data of data collecting module collected, debugs in input prediction model pre-
Survey the parameter in model.
Prediction model needs to be trained it according to available data the debugging e.g. BP mind optimal with determination after establishing
Through the initial connection weight and the parameters such as threshold value in Network Prediction Model, so that prediction model can be received at faster speed
It holds back and reaches higher precision of prediction.For example, the power equipment that data acquisition module acquires in a certain period of time is all
Fault data can be according to being, for example, that the method for salary distribution divided equally is divided into two groups, and first group for constructing prediction model, second group of use
In being trained the optimized parameter initial with determination to prediction model.For example, the running state data of data collecting module collected
It may include the e.g. slow varying signals such as temperature, pressure, electricity, flow and Mechanical Fault Vibration Signals.Operation can be extracted
As predicted time sequence, prediction model passes through to each in predicted time sequence Fault-Sensitive characterization factor in status data
Kind signal data is predicted just to can be realized the debugging to Parameters in Forecasting Model in the forecast interval of setting.
S3: optimum prediction model is determined.
In the case where the training of all prediction models is completed, by the real-time of the power equipment of data collecting module collected
Operation data is directed respectively into different prediction models, just can obtain different prediction results.For example, according to each prediction model
Prediction step, the output result of true value and prediction model just can obtain the prediction error and forecasting efficiency of prediction model.In advance
Surveying efficiency is that real-time running data is imported to prediction model and obtains the whole time-consuming of prediction result, time-consuming shorter, then predicts mould
The forecasting efficiency of type is higher.Prediction error can be prediction model and mistakenly predict that the ratio of failure will be generated.
S4: the prediction result accuracy updated to guarantee prediction model is corrected to prediction model.
Prediction model has a life cycle, and prediction result accuracy is higher when initial launch, when value data is gradually sent out
When changing, prediction model can gradually appear deviation, lead to the reduction of prediction accuracy.Determining optimal prediction model it is pre-
Error and/or forecasting efficiency are surveyed more than in the case where given threshold, can will be handled in the optimum prediction model running time cycle
The history data of the power equipment of completion, which imports, is corrected update with the parameter to prediction model in all prediction models
It optimizes.Optimal prediction model can be redefined according to the prediction error and forecasting efficiency of prediction model, subsequent
The prediction model redefined is all made of in forecast analysis.Prediction model can effectively guarantee pre- in the way of regularly updating
The parameter for surveying model is in Optimal State, and then ensure that the Optimality of prediction error and forecasting efficiency.
Embodiment 3
The present embodiment is the further improvement to embodiment 1 and embodiment 2, and duplicate content repeats no more.
The present invention also provides a kind of electric power equipment on-site exploration methods, establish prediction model based on machine learning method, adopt
Collect the running state data of power equipment, and training set and verifying collection are established based on running state data, wherein is based on training set
Prediction model is trained, the prediction mould optimal with determination is analyzed the prediction result of prediction model based on verifying collection
Type instructs prediction model according to the running state data of the new power equipment acquired in the time cycle of setting again
Practice, it is preferred that prediction model is the time series predicting model established based on time series data, can predict power equipment
The probability of certain failure occurs in some following period.In the case where specifying desired estimation range, at least according to
Lower step generates the prediction result being consistent with desired estimation range:
S1: determining affiliated partner involved in desired estimation range, and based on affiliated partner determine main association object and
Secondary association object.
Preferably, it is specified to refer to that specified power equipment occurs within the scope of specified future time for desired estimation range
Failure probability.For example, service personnel is needed based on maintenance plan at following the tenth day in the operational process of substation
The arrester of substation is overhauled, at this point, service personnel can propose prediction application to prediction model, application prediction is not
A possibility that e.g. winding failure, occurs for power transformation device in the tenth day come, and service personnel is enabled to be based on prediction result right
Maintenance forecasting is carried out to power transformation device while arrester is overhauled.What the related object of desired estimation range can specify
Power equipment and/or specified future time range.
Preferably, power equipment is generally not to be in autonomous working state, is set in a certain range of several electric power
It is standby to there is communication connection to each other, it works jointly in the way of collaboration.The collaborative work mode of power equipment to each other
It influences each other so that existing each other, for example, when being located at a power equipment of upstream there are when the occurrence tendency of failure, position
In downstream and being fitted close the associate device of its work and may be affected by it and gradually generate the occurrence tendency of failure.It closes
Join equipment fault generation time compared to be, for example, its component service life caused by failure generate have it is in-advance, that is, exist
It will break down in advance in its yuan of unclosed situation of component service life.The topology connection structure of power equipment is used for table
Show that power equipment communicates to connect relationship each other, tends to effectively distinguish key by the topology connection structure of power equipment
Equipment, conventional equipment or the associate device being closely related with key equipment.For example, e.g. substation have it is many for example
It is the power equipment of transformer, arrester, breaker, switchgear and communication scheduling equipment.Pass through function division or connection
The mode that relationship divides can determine key equipment, for example, then preliminary judgement most with power equipment that transformer is connected directly
Transformer is the key equipment of substation.
Preferably, based on the topology connection structure of power equipment, for the fault type of power equipment, several masters are determined
Affiliated partner and several secondary association objects, wherein main association object is greater than secondary association object to the contribution that failure generates.Example
Such as, in distribution box, fan generates failure and distribution box is radiated in undesirable situation, and the first electronic component therein is based on
The rising of temperature and be continuously in abnormal working position, with associated second electronic component of the first electronic component to heat dissipation
It is required that harsh degree is lower, the generation rate of failure is smaller.At this point, fan is the main association object of the first electronic component, the
Two electronic components are the secondary association objects of the first electronic component.Preferably, the quantity of main association object and secondary association object
Two can be more than or equal to.Preferably, be connected to each other the topology connection structure to be formed based on power equipment, direct with designated equipment
Connected equipment is the first associate device, is second to be associated with and set by the first associate device and the equipment that designated equipment is indirectly connected
It is standby, wherein in the case where generating maximally related first environment factor with failure based on association algorithm is determining, to choose the first association
At least one power equipment most influenced by first environment factor in equipment is chosen in the second associate device as main association object
At least one power equipment most influenced by first environment factor is as secondary association object.In the connection for considering specified power equipment
On the basis of topological structure, determine that main association object and secondary association object to enhance environment in the training data of prediction model
The importance of factor, so that prediction model is suitable for the failure predication by the more serious grid equipment of such environmental effects.
Preferably, maximum is influenced by first environment factor and refer to that power equipment most easily breaks down under the conditions of the environmental factor, by
The maximum power equipment of one such environmental effects can calculate first environment factor based on e.g. relevance algorithm and set with electric power
The standby relative coefficient to break down chooses the maximum power equipment of relative coefficient as main association object or secondary association pair
As.
S2: forming third data set in the third historical data for obtaining designated equipment in a manner of extracting fault signature, and
In the case where establishing the second prediction model based on third data set, the first historical data of the main association object of designated equipment is obtained
To form the first data set, third historical data and the first historical data include at least operational parameter data, operating status number
According to environmental factor data, wherein the relevance of environmental factor data and fault signature is determined based on association algorithm.
Association algorithm at least includes the following steps:
A1: the environmental factor index value of power equipment is determined based on environmental factor data, and by ring in the way of dividing equally
Border factor index value is divided into n subinterval, wherein the historical data based on power equipment calculates in each index value section
Fault rate.For example, index value section can be determined as [0,40], work as n for the temperature factor for being, for example, 40 DEG C
It is 4 is that can be divided into [0,10], [10,20], [20,30] and [30,40] four sections.When the value of n is too small, meeting
The regularity of data is reduced, therefore according to the actual situation, n value can be selected as biggish numerical value as far as possible.
A2: determining index value sequence based on the mean value in each subinterval, calculates environmental factor using distance correlation algorithm
Related coefficient between index value and rate of breakdown.For example, [0,10], [10,20], [20,30] and [30,40] four sections
The index value sequence of formation is 5,15,25,35.Wherein, the calculation formula of distance correlation algorithm are as follows:
Wherein, the index value sequence that X is made of the mean value in each subinterval of environmental factor index value, Y are failure rate institute
The sequence of composition, dCov (X, Y) indicate the distance between environmental factor index and failure rate covariance, dVar (X) dVar (Y) table
Show the covariance of environmental factor index and failure rate.
Preferably, in the case that the related coefficient dCor (X, Y) of environmental factor is greater than the set value, event is just determined it as
The causing factors of barrier have very big relevance with the environmental factor with the generation for showing the failure.
S3: the correlation generated with failure based on environmental factor is based on secondary association to determine at least one secondary association object
Second historical data of object corrects the first data set to obtain the second data set, wherein establishes first based on the second data set
Prediction model.Secondary association object is most vulnerable to the equipment of the such environmental effects, and the determination of secondary association object can be according to for example
It is to carry out running state data and the mode that operational parameter data compares.For example, the operational parameter data of power equipment
The technological parameter data, specified operating voltage or rated operational current, running state data that can be setting can be electric power
Actual technological parameter data, operating voltage or the operating current of equipment in the process of running.When e.g. temperature environment because
Element in different temperature range variation, the operating status of power equipment can there is a phenomenon where e.g. current fluctuation, pass through by
Rated current is compared with operating current, and the maximum power equipment of current fluctuation range value is determined as secondary association object.
Preferably, fault type based on a specified, the relevance values occurred according to each environmental factor and the failure it is big
It is small to determine that the highest environmental factor of correlation degree, the second historical data include at least secondary association object by correlation degree highest
Environmental factor influence degree be more than given threshold running state data.First data set will be by the highest ring of correlation degree
The influence degree of border factor is deleted lower than the relevant historical data of given threshold to optimize the first data set.After it will optimize
First data set and the second historical data form the second data set after merging.The amendment of first data set can exclude and event
Hinder the not high data of correlation degree, improves the prediction accuracy of the first prediction model.
S4: the prediction result based on the first prediction model is corrected the prediction result of the second prediction model.Second is pre-
The prediction result for surveying model is independently of what the historical data being directly linked with its fault type obtained, is by it in power equipment
In the case where the caused failure generation of itself the reason of, the second prediction model has good prediction accuracy.Power equipment
Failure is in situation caused by other power equipments associated with it or environmental factor associated with it, and prediction result may
It can generate biggish error, this error refreshing correction by prediction model can be slowed down, but be held in sample
Under conditions of amount is very big, prediction model refresh time-consuming universal longer, is when being unable to satisfy in the case where it is pressed for time
Between demand.It is caused by other power equipments associated with it or environmental factor associated with it in the failure of power equipment
In the case of, the prediction result of the first prediction model can accuracy with higher, two prediction models can be by counting parallel
The mode of calculation is calculated simultaneously to obtain different prediction results, by by the pre- of the second prediction model and the first prediction model
Survey result is mixed according to certain mode just can obtain accurate prediction result.Such as.Second prediction model it is pre-
Surveying result is that can break down between following the 10th day to the 12nd day, probability of happening 80%, the first prediction model it is pre-
Surveying result is that can break down between following the 9th day to the 10th day, and probability of happening 70% can after being mixed the two
It is that can break down at the 10th day to obtain prediction result, the probability of happening of failure is 80%*0.7+70%*0.4=84%.
Preferably, the second prediction model and the first prediction model can also form combination in the way of certain mixed proportion
Prediction model forms combination forecasting by way of complementary and enables to prediction result more accurate.Preferably, the first letter
Probability value in breath is greater than probability value in prediction result and shows that designated equipment is influenced to be greater than it by environmental factor and/or associate device
The influence of itself, the generation of failure are largely as caused by its associate device and/or environmental factor, by the first prediction model
Combination forecasting is formed in such a way that mixed proportion is greater than one with the second prediction model, combination forecasting is enabled to have
There is higher prediction accuracy.By the way that designated equipment and equipment associated with it and environmental factor data are inputted hybrid predicting
In model again carry out analysis prediction, just can obtain the more accurate prediction result of precision of prediction.
It should be noted that above-mentioned specific embodiment is exemplary, those skilled in the art can disclose in the present invention
Various solutions are found out under the inspiration of content, and these solutions also belong to disclosure of the invention range and fall into this hair
Within bright protection scope.It will be understood by those skilled in the art that description of the invention and its attached drawing are illustrative and are not
Constitute limitations on claims.Protection scope of the present invention is defined by the claims and their equivalents.
Claims (10)
1. a kind of electric power equipment on-site exploration method based on study acquires several electric power based on data acquisition module (1) and sets
Standby (4) respective operation data is to form data set, which is characterized in that Data Analysis Services module (3) can be configured as by
The data set is handled according to following steps:
It extracts the fault signature of the data set and prediction model is established based on the fault signature to determine that power equipment (4) are sent out
The first information of raw failure；
Determine that fault type and generating includes at least the second information of fault solution based on the fault signature, described the
In the case where two information failed regenerations, the Data Analysis Services module (3) and inline diagnosis module (7) communicate to connect and according to
The mode mutually accessed generates second information by the inline diagnosis module (7), wherein generates and loses in second information
In the case where losing, the prediction model is corrected based on the associated associated data of fault type corresponding with second information.
2. electric power equipment on-site exploration method as described in claim 1, which is characterized in that the first information includes at least event
Hinder the time occurred and probability, wherein in the case where completing the setting of desired estimation range, the Data Analysis Services mould
Block (3) at least obtains the first information according to the following steps:
Power equipment specified by estimation range based on expectations (4) determines main association object associated with it and secondary association pair
As；
It obtains the first historical data of the main association object and forms the first data set in the way of extracting fault signature, obtain
It takes the second historical data of the secondary association object and first data set is corrected to obtain based on second historical data
Second data set establishes the first prediction model based on second data set, wherein
The first information is determined based on first prediction model.
3. electric power equipment on-site exploration method as claimed in claim 2, which is characterized in that first historical data and described
Second historical data includes at least operational parameter data, running state data and the environmental factor data of power equipment, wherein institute
The second historical data for stating secondary association object is at least determined in accordance with the following steps:
The environmental factor index value of power equipment (4) is determined based on the environmental factor data, and environmental factor index value is drawn
It is divided into n subinterval, calculates fault rate of the power equipment in each index value section；
Determine index value sequence based on the mean value in each subinterval, based on distance correlation algorithm calculate environmental factor index value with
Related coefficient between fault rate, and determine the maximum first environment factor of correlation coefficient value；
The secondary association object is to be influenced maximum power equipment (4) by the first environment factor.
4. electric power equipment on-site exploration method as claimed in claim 3, which is characterized in that obtain the specified power equipment
Third historical data form third data set in a manner of extracting fault signature, and establish second based on the third data set
Prediction model, wherein
In the case where obtaining prediction result based on second prediction model, based on prediction result amendment first letter
Breath.
5. electric power equipment on-site exploration method as claimed in claim 4, which is characterized in that first prediction model and described
Second prediction model forms combination forecasting in complementary manner to complete based on the prediction result to first letter
The amendment of breath, wherein
In the case that probability value in the first information is greater than the probability value in the prediction result, first prediction model with
Second prediction model forms the combination forecasting in such a way that mixed proportion is greater than one.
6. the electric power equipment on-site exploration method as described in one of claim 1 to 5, which is characterized in that second information is extremely
It is few to obtain according to the following steps:
Information storage tube manages module (2) and the operation data of power equipment (4) is classified as the first operation data and the second fortune
Row data, wherein at least by first operation data based on the difference of fault type store in the way of classification storage to
Fault data module (6)；
The inline diagnosis module (7) can generate second information according to the fault type in a manner of Artificial Diagnosis,
Wherein, second information is stored according to the corresponding associated mode of fault type to the fault data module (6).
7. electric power equipment on-site exploration method as claimed in claim 6, which is characterized in that the main association object at least according to
Following steps are determined:
Several first associate devices being directly connected are determined based on the specified power equipment (4), based on described
Desired estimation range obtains the first environment factor by the distance correlation algorithm；
The main association object is by maximum first associate device of the first environment factor.
8. electric power equipment on-site exploration method as claimed in claim 7, which is characterized in that the meter of the distance correlation algorithm
Calculate formula are as follows:
Wherein, the index value sequence that X is made of the mean value in each subinterval of environmental factor index value, Y are made of failure rate
Sequence, dCov (X, Y) indicate the distance between environmental factor index and failure rate covariance, dVar (X) dVar (Y) expression ring
The covariance of border factor index and failure rate.
9. electric power equipment on-site exploration method as claimed in claim 8, which is characterized in that the probability value in the first information is small
In the case where probability value in the prediction result, first prediction model and second prediction model are according to mixing ratio
Example forms the combination forecasting less than one mode, wherein
The secondary association object is to be influenced maximum second associate device by the first environment factor, is set through first association
It is standby to be indirectly connected several power equipments (4) restriction second associate device with the specified power equipment.
10. the electric power equipment on-site exploration method stated such as claim 8, which is characterized in that the Data Analysis Services module (3)
The first information is obtained according further to following steps:
Power equipment specified by estimation range based on expectations (4) determines main association object associated with it and secondary association pair
As；
It obtains the first historical data of the main association object and forms the first data set in the way of extracting fault signature, obtain
It takes the second historical data of the secondary association object and first data set is corrected to obtain based on second historical data
Second data set establishes the first prediction model based on second data set, obtains the third historical data to extract failure
The mode of feature forms the third data set, and establishes second prediction model based on the third data set；
Training first prediction model and second prediction model simultaneously determine that optimal first is pre- based on prediction result respectively
Model and the second optimal prediction model are surveyed,
The first information, and the prediction result based on the second optimal prediction model are determined based on the first optimal prediction model
Correct the first information.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810895997.XA CN109146093B (en) | 2018-08-08 | 2018-08-08 | Power equipment field investigation method based on learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810895997.XA CN109146093B (en) | 2018-08-08 | 2018-08-08 | Power equipment field investigation method based on learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109146093A true CN109146093A (en) | 2019-01-04 |
| CN109146093B CN109146093B (en) | 2020-08-11 |
Family
ID=64792104
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810895997.XA Expired - Fee Related CN109146093B (en) | 2018-08-08 | 2018-08-08 | Power equipment field investigation method based on learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109146093B (en) |
Cited By (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111314137A (en) * | 2020-02-18 | 2020-06-19 | 国家电网有限公司 | Information communication network automatic operation and maintenance method, device, storage medium and processor |
| CN111487950A (en) * | 2020-04-24 | 2020-08-04 | 西安交通大学 | 'prediction-verification-feedback-optimization' closed-loop system for online early warning and offline diagnosis |
| CN111525694A (en) * | 2020-05-08 | 2020-08-11 | 广东电网有限责任公司惠州供电局 | Intelligent switch cabinet marking device and system based on Internet of things |
| CN111901158A (en) * | 2020-07-14 | 2020-11-06 | 广东科徕尼智能科技有限公司 | Intelligent home distribution network fault data analysis method, equipment and storage medium |
| CN111921377A (en) * | 2020-09-25 | 2020-11-13 | 大唐环境产业集团股份有限公司 | SCR denitration ammonia injection control method and system based on mechanism and data driving |
| CN111967535A (en) * | 2020-09-04 | 2020-11-20 | 安徽大学 | Fault diagnosis method and device for temperature sensor in grain storage management scene |
| CN112241511A (en) * | 2019-07-19 | 2021-01-19 | 上海杰之能软件科技有限公司 | Performance detection method and device of auxiliary electrical system of vehicle and terminal |
| CN112330494A (en) * | 2020-11-26 | 2021-02-05 | 沈阳惠东管道控制有限公司 | Real-time data acquisition and processing system for natural gas transmission station |
| CN112324414A (en) * | 2020-11-03 | 2021-02-05 | 盐城市崇达石化机械有限公司 | Fracturing manifold remotely controlled by hydraulic system |
| CN112348699A (en) * | 2020-11-02 | 2021-02-09 | 国网山东省电力公司博兴县供电公司 | Power supply system power equipment life cycle management method and system |
| CN112446137A (en) * | 2020-11-03 | 2021-03-05 | 北京清能互联科技有限公司 | Multi-time-gradient undercut load scheme automatic generation method, system and storage medium |
| WO2021057576A1 (en) * | 2019-09-29 | 2021-04-01 | 中兴通讯股份有限公司 | Method for constructing cloud network alarm root cause relational tree model, device, and storage medium |
| CN112633527A (en) * | 2020-12-18 | 2021-04-09 | 威海天鑫现代服务技术研究院有限公司 | Enterprise maintenance network management method, system, computer equipment and storage medium |
| CN112766396A (en) * | 2021-01-27 | 2021-05-07 | 昆仑数智科技有限责任公司 | System, method, computer device and medium for detecting device abnormality |
| CN112862347A (en) * | 2021-03-02 | 2021-05-28 | 同济大学 | Equipment abnormity monitoring method and system based on federal learning, storage medium and terminal |
| CN113065580A (en) * | 2021-03-17 | 2021-07-02 | 国能大渡河大数据服务有限公司 | Power plant equipment management method and system based on multi-information fusion |
| CN113983983A (en) * | 2021-10-27 | 2022-01-28 | 深圳飞赛精密钣金技术有限公司 | Wire temperature and sag actual measurement method and system |
| CN114637351A (en) * | 2022-03-14 | 2022-06-17 | 江西省农业科学院园艺研究所 | A kind of facility crop greenhouse environment regulation method and system |
| CN114879622A (en) * | 2022-07-12 | 2022-08-09 | 珠海市鸿瑞信息技术股份有限公司 | Industrial control log auditing system and method based on multi-source data |
| CN115356592A (en) * | 2022-08-29 | 2022-11-18 | 南京乾鑫电器设备有限公司 | Novel fault monitoring method and system for low-voltage distribution equipment |
| CN117192269A (en) * | 2023-09-20 | 2023-12-08 | 珠海高新区铭越科技有限公司 | Big data monitoring and early warning system and method for electric room environment control box |
| CN117474342A (en) * | 2023-11-27 | 2024-01-30 | 北京国网信通埃森哲信息技术有限公司 | Power grid flood prevention and disaster prevention monitoring and early warning system based on geographical meteorological information |
| CN117590149A (en) * | 2023-11-14 | 2024-02-23 | 南方电网调峰调频发电有限公司检修试验分公司 | Fault solution generation method, device and equipment based on big data technology |
| CN117826712A (en) * | 2023-11-09 | 2024-04-05 | 中通服网盈科技有限公司 | Equipment operation intelligent scheduling system of industrial Internet |
| CN117972602A (en) * | 2024-01-24 | 2024-05-03 | 中国电建集团山东电力建设第一工程有限公司 | Power plant equipment operation life cycle early warning system based on big data analysis |
| CN118017697A (en) * | 2024-04-09 | 2024-05-10 | 国网浙江省电力有限公司宁波供电公司 | A method and device for monitoring power transmission lines based on optical fiber sensing technology |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110010119A1 (en) * | 2008-01-23 | 2011-01-13 | Hsu Pi-Lin | Apparatus for monitoring the maintenance of systems in a nuclear power plant |
| CN103714401A (en) * | 2014-01-21 | 2014-04-09 | 国家电网公司 | Method for predicting distribution and transformation fault frequency of electric network |
| CN104200288A (en) * | 2014-09-18 | 2014-12-10 | 山东大学 | Equipment fault prediction method based on factor-event correlation recognition |
| CN108051709A (en) * | 2017-11-30 | 2018-05-18 | 南京力通达电气技术有限公司 | Transformer state online evaluation analysis method based on artificial intelligence technology |
-
2018
- 2018-08-08 CN CN201810895997.XA patent/CN109146093B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110010119A1 (en) * | 2008-01-23 | 2011-01-13 | Hsu Pi-Lin | Apparatus for monitoring the maintenance of systems in a nuclear power plant |
| CN103714401A (en) * | 2014-01-21 | 2014-04-09 | 国家电网公司 | Method for predicting distribution and transformation fault frequency of electric network |
| CN104200288A (en) * | 2014-09-18 | 2014-12-10 | 山东大学 | Equipment fault prediction method based on factor-event correlation recognition |
| CN108051709A (en) * | 2017-11-30 | 2018-05-18 | 南京力通达电气技术有限公司 | Transformer state online evaluation analysis method based on artificial intelligence technology |
Cited By (36)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112241511A (en) * | 2019-07-19 | 2021-01-19 | 上海杰之能软件科技有限公司 | Performance detection method and device of auxiliary electrical system of vehicle and terminal |
| WO2021057576A1 (en) * | 2019-09-29 | 2021-04-01 | 中兴通讯股份有限公司 | Method for constructing cloud network alarm root cause relational tree model, device, and storage medium |
| CN111314137A (en) * | 2020-02-18 | 2020-06-19 | 国家电网有限公司 | Information communication network automatic operation and maintenance method, device, storage medium and processor |
| CN111487950A (en) * | 2020-04-24 | 2020-08-04 | 西安交通大学 | 'prediction-verification-feedback-optimization' closed-loop system for online early warning and offline diagnosis |
| CN111525694A (en) * | 2020-05-08 | 2020-08-11 | 广东电网有限责任公司惠州供电局 | Intelligent switch cabinet marking device and system based on Internet of things |
| CN111525694B (en) * | 2020-05-08 | 2021-12-31 | 广东电网有限责任公司惠州供电局 | Intelligent switch cabinet marking device and method based on Internet of things |
| CN111901158A (en) * | 2020-07-14 | 2020-11-06 | 广东科徕尼智能科技有限公司 | Intelligent home distribution network fault data analysis method, equipment and storage medium |
| CN111967535A (en) * | 2020-09-04 | 2020-11-20 | 安徽大学 | Fault diagnosis method and device for temperature sensor in grain storage management scene |
| CN111967535B (en) * | 2020-09-04 | 2023-11-14 | 安徽大学 | A temperature sensor fault diagnosis method and its diagnostic device in grain storage management scenarios |
| CN111921377A (en) * | 2020-09-25 | 2020-11-13 | 大唐环境产业集团股份有限公司 | SCR denitration ammonia injection control method and system based on mechanism and data driving |
| CN112348699A (en) * | 2020-11-02 | 2021-02-09 | 国网山东省电力公司博兴县供电公司 | Power supply system power equipment life cycle management method and system |
| CN112446137A (en) * | 2020-11-03 | 2021-03-05 | 北京清能互联科技有限公司 | Multi-time-gradient undercut load scheme automatic generation method, system and storage medium |
| CN112446137B (en) * | 2020-11-03 | 2023-10-03 | 北京清能互联科技有限公司 | Multi-time gradient down-cut load scheme automatic generation method, system and storage medium |
| CN112324414A (en) * | 2020-11-03 | 2021-02-05 | 盐城市崇达石化机械有限公司 | Fracturing manifold remotely controlled by hydraulic system |
| CN112324414B (en) * | 2020-11-03 | 2023-06-13 | 盐城市崇达石化机械有限公司 | Fracturing manifold remotely controlled by hydraulic system |
| CN112330494A (en) * | 2020-11-26 | 2021-02-05 | 沈阳惠东管道控制有限公司 | Real-time data acquisition and processing system for natural gas transmission station |
| CN112633527A (en) * | 2020-12-18 | 2021-04-09 | 威海天鑫现代服务技术研究院有限公司 | Enterprise maintenance network management method, system, computer equipment and storage medium |
| CN112633527B (en) * | 2020-12-18 | 2022-09-09 | 威海天鑫现代服务技术研究院有限公司 | Enterprise maintenance network management method, system, computer equipment and storage medium |
| CN112766396A (en) * | 2021-01-27 | 2021-05-07 | 昆仑数智科技有限责任公司 | System, method, computer device and medium for detecting device abnormality |
| CN112862347A (en) * | 2021-03-02 | 2021-05-28 | 同济大学 | Equipment abnormity monitoring method and system based on federal learning, storage medium and terminal |
| CN113065580A (en) * | 2021-03-17 | 2021-07-02 | 国能大渡河大数据服务有限公司 | Power plant equipment management method and system based on multi-information fusion |
| CN113065580B (en) * | 2021-03-17 | 2024-04-16 | 国能大渡河大数据服务有限公司 | Power plant equipment management method and system based on multi-information fusion |
| CN113983983A (en) * | 2021-10-27 | 2022-01-28 | 深圳飞赛精密钣金技术有限公司 | Wire temperature and sag actual measurement method and system |
| CN114637351A (en) * | 2022-03-14 | 2022-06-17 | 江西省农业科学院园艺研究所 | A kind of facility crop greenhouse environment regulation method and system |
| CN114879622A (en) * | 2022-07-12 | 2022-08-09 | 珠海市鸿瑞信息技术股份有限公司 | Industrial control log auditing system and method based on multi-source data |
| CN115356592A (en) * | 2022-08-29 | 2022-11-18 | 南京乾鑫电器设备有限公司 | Novel fault monitoring method and system for low-voltage distribution equipment |
| CN115356592B (en) * | 2022-08-29 | 2024-03-08 | 南京乾鑫电器设备有限公司 | Novel fault monitoring method and system for low-voltage distribution equipment |
| CN117192269A (en) * | 2023-09-20 | 2023-12-08 | 珠海高新区铭越科技有限公司 | Big data monitoring and early warning system and method for electric room environment control box |
| CN117192269B (en) * | 2023-09-20 | 2024-04-05 | 珠海高新区铭越科技有限公司 | Big data monitoring and early warning system and method for electric room environment control box |
| CN117826712A (en) * | 2023-11-09 | 2024-04-05 | 中通服网盈科技有限公司 | Equipment operation intelligent scheduling system of industrial Internet |
| CN117826712B (en) * | 2023-11-09 | 2024-06-07 | 中通服网盈科技有限公司 | Equipment operation intelligent scheduling system of industrial Internet |
| CN117590149A (en) * | 2023-11-14 | 2024-02-23 | 南方电网调峰调频发电有限公司检修试验分公司 | Fault solution generation method, device and equipment based on big data technology |
| CN117474342A (en) * | 2023-11-27 | 2024-01-30 | 北京国网信通埃森哲信息技术有限公司 | Power grid flood prevention and disaster prevention monitoring and early warning system based on geographical meteorological information |
| CN117474342B (en) * | 2023-11-27 | 2025-03-14 | 北京国网信通埃森哲信息技术有限公司 | Power grid flood prevention and disaster relief monitoring and early warning system based on geographic meteorological information |
| CN117972602A (en) * | 2024-01-24 | 2024-05-03 | 中国电建集团山东电力建设第一工程有限公司 | Power plant equipment operation life cycle early warning system based on big data analysis |
| CN118017697A (en) * | 2024-04-09 | 2024-05-10 | 国网浙江省电力有限公司宁波供电公司 | A method and device for monitoring power transmission lines based on optical fiber sensing technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109146093B (en) | 2020-08-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109146093A (en) | A kind of electric power equipment on-site exploration method based on study | |
| CN118157330B (en) | Panoramic dynamic topology monitoring and control method and system for distribution network based on graph computing | |
| CN112803592B (en) | Intelligent fault early warning method and system suitable for distributed power station | |
| CN105353702B (en) | High-voltage equipment intelligent monitoring system | |
| CN118739184A (en) | A method for automatic fault detection and repair of self-healing intelligent power lines | |
| CN118487276B (en) | A method and system for dynamic control of power grid security for power security objects | |
| CN113537415A (en) | Convertor station inspection method and device based on multi-information fusion and computer equipment | |
| CN109828182A (en) | A kind of network system accident analysis method for early warning based on failure modes processing | |
| CN116614177B (en) | Optical fiber state multidimensional parameter monitoring system | |
| CN110716103A (en) | Early warning method for distribution network fault | |
| CN118758377A (en) | A bridge structure health monitoring and early warning system | |
| CN118839617A (en) | Periodic digital twin auxiliary management platform for photovoltaic construction | |
| CN119066398B (en) | Distributed intelligent micro-grid fault diagnosis method and system based on edge calculation | |
| CN118709611A (en) | A transmission line distribution management simulation method and system based on virtual operation | |
| CN117390403A (en) | Power grid fault detection method and system for new energy lighthouse power station | |
| CN114142614A (en) | Highway power distribution room intelligent operation and maintenance management system based on SD-WAN network | |
| CN112653242A (en) | Power transmission line monitoring system based on cloud platform service | |
| CN119024094A (en) | Fault location and restoration decision method and system for distribution network | |
| CN117639230A (en) | Active service monitoring method based on electric power | |
| KR20240001975A (en) | Systems for predicting and monitoring solar power generation using artificial intelligence | |
| CN118508871A (en) | Photovoltaic power station autonomous regulation system based on GIS data analysis | |
| CN112016739A (en) | Fault detection method, device, electronic device and storage medium | |
| CN111857015B (en) | Power transmission and transformation cloud intelligent controller | |
| CN114389361B (en) | A power grid panoramic control method and system for zero-carbon operation of county power grid | |
| Wu et al. | Research and applications of the key technology of intelligent operation and maintenance system architecture for transmission power cables |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200811 |