CN106886651A - Insulator metal accessory corrodes quantity of electric charge Forecasting Methodology - Google Patents
Insulator metal accessory corrodes quantity of electric charge Forecasting Methodology Download PDFInfo
- Publication number
- CN106886651A CN106886651A CN201710127276.XA CN201710127276A CN106886651A CN 106886651 A CN106886651 A CN 106886651A CN 201710127276 A CN201710127276 A CN 201710127276A CN 106886651 A CN106886651 A CN 106886651A
- Authority
- CN
- China
- Prior art keywords
- metal
- electric charge
- ware
- corrosion
- result
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/04—Ageing analysis or optimisation against ageing
Abstract
The invention discloses a kind of insulator metal accessory corrosion quantity of electric charge Forecasting Methodology, data are obtained from on-line monitoring system, the data to obtaining carry out missing values treatment, obtain continuous data；Metal-ware is corroded the quantity of electric charge and characteristic quantity is extracted from continuous data；The metal-ware corrosion quantity of electric charge and characteristic quantity are processed with neural network algorithm；The metal-ware corrosion quantity of electric charge is processed with time series analysis method；By the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term and the Comparative result of the time series quantity of electric charge prediction of metal-ware corrosion in short-term, the result of the quantity of electric charge prediction of metal-ware corrosion in short-term is obtained；Result finally according to the corrosion quantity of electric charge prediction of metal-ware in short-term is estimated to the running status of insulator; the result that the quantity of electric charge is predicted is corroded according to average annual metal-ware; metal-ware protection device to insulator is designed, and solves the problems, such as that the insulator protection device life-span is reduced and waste of materials.
Description
Technical field
The present invention relates to technical field of ultrahigh voltage direct current, more particularly to insulator metal accessory corrosion quantity of electric charge prediction
Method.
Background technology
In recent years, China's high voltage dc transmission technology quickly grows, but consequently also brings some problems, is being related to high pressure
The technical field of DC Insulator, glass and porcelain insulator on a plurality of HVDC transmission line occur in that the metal of large area
Annex corrosion phenomenon.
Insulator metal accessory corrosion is broadly divided into the corrosion of steel pin and chapeau de fer corrodes two kinds.The corrosion of steel pin can be directly contributed absolutely
The decline of edge mechanical strength, meanwhile, corrosion product can cause to be produced between steel pin and cement one than larger stress, and then
Cause the damage of insulator；And chapeau de fer corrosion can then cause occur the rusty stain passage caused by erosion on insulator, the rusty stain
Passage can speed up the accumulation of filth.During metal-ware seriously corroded, the problems such as can also cause insulator to come off, and then influence defeated
The safe operation of electric line.In order to protect insulator metal accessory not corroded, most effective method is in insulation at present
Install zinc protection device on son additional.
However, zinc protection device typically all empirically value is designed, i.e., the average annual corrosion quantity of electric charge is less than
1500 DEG C of area protection zinc is arranged when being calculated as 4mm thickness, and zinc set service life is 30 years.But the actual average annual corrosion quantity of electric charge
Typically change with time and occur changing, this will result in, and zinc set actual life is shorter than projected life or zinc cover material is wasted
Situation occur.
The content of the invention
The invention provides a kind of insulator metal accessory corrosion quantity of electric charge Forecasting Methodology, to solve insulator protection device
Life-span reduces the problem with waste of materials.
The embodiment provides a kind of insulator metal accessory corrosion quantity of electric charge Forecasting Methodology, methods described bag
Include：
Data are obtained from insulator metal accessory corrosion quantity of electric charge on-line monitoring system；
Missing values treatment is carried out to the data, continuous data is obtained；
The extraction of the metal-ware corrosion quantity of electric charge and the extraction of characteristic quantity are carried out to the continuous data, the metal is obtained
Annex corrodes the quantity of electric charge and the characteristic quantity, and the characteristic quantity includes：Relative humidity, the temperature difference and rain fall；
The metal-ware is corroded into the treatment that the quantity of electric charge and the characteristic quantity pass through neural network algorithm, nerve net is obtained
The result of the network quantity of electric charge prediction of metal-ware corrosion in short-term；
The metal-ware is corroded into treatment of the quantity of electric charge by time series analysis method, time series metal in short-term is obtained
The result of annex corrosion quantity of electric charge prediction and the result of average annual metal-ware corrosion quantity of electric charge prediction；
By result and the time series of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term metal in short-term
The result of annex corrosion quantity of electric charge prediction is contrasted, and obtains the result of the quantity of electric charge prediction of metal-ware corrosion in short-term.
Preferably, it is described that the metal-ware is corroded into the place that the quantity of electric charge and the characteristic quantity pass through neural network algorithm
Reason, during obtaining the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term, the neural network algorithm includes：
Using the characteristic quantity as variable, the metal-ware is corroded the quantity of electric charge as dependent variable, set up neutral net
Model；
The characteristic quantity and the metal-ware corrosion quantity of electric charge are normalized, the data after being processed；
Data after the treatment are classified, training data and test data is obtained；
The training data is input into the neural network model, the test metal-ware corrosion quantity of electric charge is obtained；
The metal-ware corrosion quantity of electric charge in the test data is subtracted each other with the test metal-ware corrosion quantity of electric charge,
Obtain difference；
The difference is compared with preset range, master pattern is obtained；
According to the master pattern, the neutral net quantity of electric charge of metal-ware corrosion in short-term is predicted, obtains god
Through the result that the network quantity of electric charge of metal-ware corrosion in short-term is predicted.
Preferably, it is described to compare the difference with preset range, including：
Judge the difference whether in the preset range；
If the difference is in the preset range, the neural network model is the master pattern；
If the difference is not in the preset range, obtain described after being revised to the neural network model
Master pattern.
Preferably, it is described that the metal-ware is corroded into treatment of the quantity of electric charge by time series analysis method, obtain the time
The process of the result of the sequence quantity of electric charge prediction of metal-ware corrosion in short-term and the result of average annual metal-ware corrosion quantity of electric charge prediction
In, the time series analysis method includes：
The metal-ware corrosion quantity of electric charge is arranged sequentially in time, setup time sequence；
Influence of the seasonal variations to the time series is determined, seasonal variations factor of influence is obtained；
According to the seasonal variations factor of influence, the time series is revised, the time after the influence that is eliminated
The change curve that sequence and the metal-ware corrosion quantity of electric charge are changed over time；
Time series after described eliminating the effects of the act is fitted with the change curve；
The cyclic swing amplitude and Cycle Length of the time series after digital simulation；
According to the cyclic swing amplitude and the Cycle Length to the time series electric charge of metal-ware corrosion in short-term
Amount and the average annual metal-ware corrosion quantity of electric charge are predicted, and obtain the time series quantity of electric charge prediction of metal-ware corrosion in short-term
Result and average annual metal-ware corrosion the quantity of electric charge prediction result.
Preferably, the result and the time sequence by the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term
The result of the row quantity of electric charge prediction of metal-ware corrosion in short-term is contrasted, and obtains the knot of the quantity of electric charge prediction of metal-ware corrosion in short-term
During fruit, the method for the contrast, including：
If the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term is golden in short-term with the time series
The result of category annex corrosion quantity of electric charge prediction is identical, then corrode electric charge using the identical result as the metal-ware in short-term
Measure the result of prediction；
If the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term is more than the time series in short-term
The result of metal-ware corrosion quantity of electric charge prediction, then by the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term
As the result that the metal-ware in short-term corrosion quantity of electric charge is predicted；
If the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term is less than the time series in short-term
The result of metal-ware corrosion quantity of electric charge prediction, then by the result of the time series quantity of electric charge prediction of metal-ware corrosion in short-term
As the result that the metal-ware in short-term corrosion quantity of electric charge is predicted.
Preferably, methods described also includes：Described by the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term
The result predicted of result and the time series quantity of electric charge of metal-ware corrosion in short-term contrasted, obtain metal-ware in short-term
After the result of corrosion quantity of electric charge prediction, methods described also includes：Corrode what the quantity of electric charge was predicted according to the metal-ware in short-term
As a result, the running status to the insulator is estimated.
Preferably, the metal-ware is corroded into treatment of the quantity of electric charge by time series analysis method described, when obtaining
Between sequence in short-term metal-ware corrosion the quantity of electric charge prediction result and average annual metal-ware corrosion the quantity of electric charge prediction result after,
Methods described also includes：According to the result that the average annual metal-ware corrosion quantity of electric charge is predicted, the metal to the insulator is attached
Part protection device is designed.
Preferably, the method that data are carried out with missing values treatment is interpolation method nearby.
From above technical scheme, a kind of insulator metal accessory corrosion quantity of electric charge prediction is the embodiment of the invention provides
Method, first obtains data from insulator metal accessory corrosion quantity of electric charge on-line monitoring system, and the data to obtaining are lacked
Value treatment, makes data have continuity；Metal-ware is corroded the quantity of electric charge and characteristic quantity is extracted from continuous data；With
Neural network algorithm corrodes the quantity of electric charge to metal-ware and characteristic quantity is processed, and obtains neutral net metal-ware corrosion in short-term
The result of quantity of electric charge prediction；The metal-ware corrosion quantity of electric charge is processed with time series analysis method, obtains time series
The result of the quantity of electric charge prediction of metal-ware corrosion in short-term and the result of average annual metal-ware corrosion quantity of electric charge prediction；Again by nerve net
The result of the network quantity of electric charge prediction of metal-ware corrosion in short-term and the result of the time series quantity of electric charge prediction of metal-ware corrosion in short-term
Contrast, obtains the result of the quantity of electric charge prediction of metal-ware corrosion in short-term；Finally according to the corrosion quantity of electric charge prediction of metal-ware in short-term
Result the running status of insulator is estimated, according to average annual metal-ware corrode the quantity of electric charge predict result, to insulation
The metal-ware protection device of son is designed, and solves the problems, such as that the insulator protection device life-span is reduced and waste of materials.
Brief description of the drawings
In order to illustrate more clearly of technical scheme, letter will be made to the accompanying drawing to be used needed for embodiment below
Singly introduce, it should be apparent that, for those of ordinary skills, without having to pay creative labor,
Other accompanying drawings can also be obtained according to these accompanying drawings.
Fig. 1 corrodes the flow chart of quantity of electric charge Forecasting Methodology for the insulator metal accessory provided according to a preferred embodiment；
Fig. 2 is the flow chart of neural network algorithm provided according to a preferred embodiment；
Fig. 3 is the flow chart of time series analysis method provided according to a preferred embodiment.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Whole description, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiments.It is based on
Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
As described in Figure 1, it is a kind of insulator metal accessory corrosion quantity of electric charge Forecasting Methodology provided in an embodiment of the present invention
Flow chart, methods described includes：
S101, data are obtained from insulator metal accessory corrosion quantity of electric charge on-line monitoring system.
Be stored with interior discharge capacity per minute, accumulation in the insulator metal accessory corrosion quantity of electric charge on-line monitoring system
The data such as discharge capacity, one minute maximum discharge current, temperature and relative humidity, the insulator metal accessory corrodes the quantity of electric charge
In Forecasting Methodology, first have to obtain out from the on-line monitoring system by data.
The data are carried out missing values treatment by S102, obtain continuous data.
During data acquisition, it is likely that because the failure problems of the on-line monitoring system, cause some data
Missing, and during data processing, the data value of missing can affect to the treatment of data, if such as
The on-line monitoring system lost substantial amounts of useful information, can be difficult to determine the certainty of the on-line monitoring system, also can
Make the packet of output containing null value, make output data insincere, it is therefore necessary to carry out missing values treatment to the data, obtain
Continuous data, strengthens data reliability, makes the result of prediction more accurate.
S103, the extraction of the metal-ware corrosion quantity of electric charge and the extraction of characteristic quantity are carried out to the continuous data, obtain institute
State metal-ware the corrosion quantity of electric charge and the characteristic quantity；The characteristic quantity includes：Relative humidity, the temperature difference and rain fall.
Correlation analysis are carried out by the continuous data, insulator can be determined using correlation figure and coefficient correlation
The metal-ware corrosion quantity of electric charge influences larger factor, with reference to actual theoretical case, can obtain relative humidity, the temperature difference and rainfall
Situation is the principal element for influenceing insulator metal accessory to corrode the quantity of electric charge, so relative humidity, the temperature difference and rain fall are made
The amount of being characterized is extracted；Studied in units of one hour in embodiments of the invention, with average relative in one hour
Humidity as the relative humidity in this time period, using one day in maximum temperature difference as the temperature difference in this time period, in addition also
Count daily whether rainfall.
S104, the treatment that the quantity of electric charge and the characteristic quantity pass through neural network algorithm is corroded by the metal-ware, is obtained
The result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term.
S105, treatment of the quantity of electric charge by time series analysis method is corroded by the metal-ware, obtains time series short
When metal-ware corrosion the quantity of electric charge prediction result and average annual metal-ware corrosion the quantity of electric charge prediction result.
S106, by the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term with the time series in short-term
The result of metal-ware corrosion quantity of electric charge prediction is contrasted, and obtains the result of the quantity of electric charge prediction of metal-ware corrosion in short-term.
From above technical scheme, a kind of insulator metal accessory corrosion quantity of electric charge prediction is the embodiment of the invention provides
Method, first obtains data from insulator metal accessory corrosion quantity of electric charge on-line monitoring system, and the data to obtaining are lacked
Value treatment, makes data have continuity；Metal-ware is corroded the quantity of electric charge and characteristic quantity is extracted from continuous data；With
Neural network algorithm corrodes the quantity of electric charge to metal-ware and characteristic quantity is processed, and obtains neutral net metal-ware corrosion in short-term
The result of quantity of electric charge prediction；The metal-ware corrosion quantity of electric charge is processed with time series analysis method, obtains time series
The result of the quantity of electric charge prediction of metal-ware corrosion in short-term and the result of average annual metal-ware corrosion quantity of electric charge prediction；Again by nerve net
The result of the network quantity of electric charge prediction of metal-ware corrosion in short-term and the result of the time series quantity of electric charge prediction of metal-ware corrosion in short-term
Contrast, obtains the result of the quantity of electric charge prediction of metal-ware corrosion in short-term；Finally according to the corrosion quantity of electric charge prediction of metal-ware in short-term
Result the running status of insulator is estimated, according to average annual metal-ware corrode the quantity of electric charge predict result, to insulation
The metal-ware protection device of son is designed, and solves the problems, such as that the insulator protection device life-span is reduced and waste of materials.
As shown in Fig. 2 being a kind of flow chart of neural network algorithm provided in an embodiment of the present invention, methods described is included such as
Lower step：
S201, using the characteristic quantity as variable, the quantity of electric charge as dependent variable is corroded using the metal-ware, sets up nerve
Network model.
Human brain rely primarily on neuron in information process between interaction, artificial neutral net pair
The simulation of this process, and then Treatment Analysis some complicated problems；Neutral net is by input layer, hidden layer and output layer structure
Into again each layer has multiple neurons to constitute, and the output of neuron can be determined by multiple input, be input into the relation with output
It is shown below：
yi=f (Xi)
In formula, xj, j=1,2 ..., n are the input signal of neuron；θiIt is the threshold value of neuron；kjiBe from cell j to
The connection weight of cell i；N is the number of input signal；yiIt is the output of neuron.
Neural network algorithm is set up on neutral net, according to above formula, using variable as input value, using dependent variable as
Output valve, can obtain the relational expression of dependent variable and variable, that is, obtain the characteristic quantity and corrode the quantity of electric charge with the metal-ware
Relation, sets up neural network model.
S202, the characteristic quantity and the metal-ware corrosion quantity of electric charge is normalized, after being processed
Data.
The characteristic quantity and the metal-ware corrosion quantity of electric charge are normalized, according to the following formula data is mapped
To in the interval of (0,1)：
In formula, xiIt is initial data, x 'iNumerical value after being normalized for initial data, xmax、xminIt is respectively former
The maximum and minimum value of beginning data.The characteristic quantity and the metal-ware corrosion quantity of electric charge are normalized
Purpose is to accelerate data convergence rate in the training process.
S203, the data after the treatment are classified, and obtain training data and test data.
S204, the training data is input into the neural network model, obtains the test metal-ware corrosion quantity of electric charge.
S205, by the metal-ware corrosion quantity of electric charge in the test data and the test metal-ware corrosion quantity of electric charge
Subtract each other, obtain difference.
S206, the difference is compared with preset range, obtains master pattern.
S207, according to the master pattern, is predicted to the neutral net quantity of electric charge of metal-ware corrosion in short-term, obtains
To the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term.
According to the master pattern, it is possible to use the weather forecast situation in following a period of time, relative humidity, temperature are drawn
Difference and rain fall, the relative humidity that will be obtained, the temperature difference and rain fall draw future as the input value of the master pattern
The metal-ware corrosion quantity of electric charge in a period of time.
Further, it is described to compare the difference with preset range, including：
Judge the difference whether in the preset range；
If the difference is in the preset range, the neural network model is the master pattern；
If the difference is not in the preset range, obtain described after being revised to the neural network model
Master pattern.
As shown in figure 3, being a kind of flow chart of time series analysis method provided in an embodiment of the present invention, methods described bag
Include：
S301, the metal-ware corrosion quantity of electric charge in the data is arranged sequentially in time, setup time sequence.
Establish after the time series, it is necessary to find the variable for influenceing the time series, and then determine to change
Type.The variation type is divided into：Long-term trend, seasonal move, cyclical variations and erratic variation.It is determined that it is favourable to change type
It is analyzed in the time series.
S302, determines influence of the seasonal variations to the time series, obtains seasonal variations factor of influence.
S303, according to the seasonal variations factor of influence, revises to the time series, is eliminated after influence
The change curve that time series and the metal-ware corrosion quantity of electric charge are changed over time.
Because the time series changes with the change in season, seasonal variations are eliminated here to the time series
Influence, obtains a more correct time sequence, is conducive to corroding metal-ware the prediction of the quantity of electric charge.
S304, the time series after described eliminating the effects of the act is fitted with the change curve.
There is more accurate periodicity through the time series after over-fitting, to the pre- of the metal-ware corrosion quantity of electric charge
It is more accurate to survey.
S305, the cyclic swing amplitude and Cycle Length of the time series after digital simulation.
S306, according to the cyclic swing amplitude and the Cycle Length to time series metal-ware corrosion in short-term
The quantity of electric charge and the average annual metal-ware corrosion quantity of electric charge are predicted, and obtain the time series electricity of metal-ware corrosion in short-term
Lotus amount and the result of the average annual metal-ware corrosion quantity of electric charge prediction.
The data in the on-line monitoring system are analyzed using time series analysis method, the average annual metal for obtaining is attached
Part corrosion quantity of electric charge predicted value more can accurately reflect the situation of change of the metal-ware corrosion quantity of electric charge in years ahead, more
Stick on nearly actual conditions.
Further, it is described by the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term and the time
The result of the sequence quantity of electric charge prediction of metal-ware corrosion in short-term is contrasted, and obtains the quantity of electric charge prediction of metal-ware corrosion in short-term
During result, the method for the contrast, including：
If the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term is golden in short-term with the time series
The result of category annex corrosion quantity of electric charge prediction is identical, then corrode electric charge using the identical result as the metal-ware in short-term
Measure the result of prediction；
If the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term is more than the time series in short-term
The result of metal-ware corrosion quantity of electric charge prediction, then by the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term
As the result that the metal-ware in short-term corrosion quantity of electric charge is predicted；
If the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term is less than the time series in short-term
The result of metal-ware corrosion quantity of electric charge prediction, then by the result of the time series quantity of electric charge prediction of metal-ware corrosion in short-term
As the result that the metal-ware in short-term corrosion quantity of electric charge is predicted.
The method of the contrast, be take the neutral net in short-term metal-ware corrosion the quantity of electric charge prediction result with it is described
Higher value between the result of the time series quantity of electric charge prediction of metal-ware corrosion in short-term, corrodes as the metal-ware in short-term
The result of quantity of electric charge prediction, to increase the reliability of predicted value.
Further, it is described by the neutral net in short-term metal-ware corrosion the quantity of electric charge prediction result with it is described when
Between sequence in short-term metal-ware corrosion the quantity of electric charge prediction result contrasted, obtain in short-term metal-ware corrosion the quantity of electric charge prediction
Result after, methods described also includes：The result that the quantity of electric charge is predicted is corroded according to the metal-ware in short-term, to the insulation
The running status of son is estimated, and then ensures the safe and stable operation of power network.
Further, the metal-ware is corroded into treatment of the quantity of electric charge by time series analysis method described, is obtained
Time series in short-term metal-ware corrosion the quantity of electric charge prediction result and average annual metal-ware corrosion the quantity of electric charge prediction result it
Afterwards, methods described also includes：According to the result that the average annual metal-ware corrosion quantity of electric charge is predicted, to the metal of the insulator
Annex protection device is designed, it is possible to prevente effectively from wasting or projected life situation not up to standard occurs, that is, ensure that electricity
The safe operation of net, saves lot of materials again.
Further, the method that data are carried out with missing values treatment is interpolation method nearby.
Conventional method to missing values treatment is interpolation processing and delete processing, because in statistics metal-ware corrosion electric charge
Need to be added up the corrosion quantity of electric charge of record during amount, therefore only consider the situation of interpolation processing.Interpolation processing is again
It is divided into：Random interpolation, according to probability interpolation, nearby interpolation and classification interpolation.Typically have in view of the metal-ware corrosion quantity of electric charge
Continuity, will not nearby insert in minute to occur very big mutation in the time of unit, therefore being used in the treatment of missing values
Data are carried out interpolation polishing by value method according to other record informations near missing values.
From above technical scheme, a kind of insulator metal accessory corrosion quantity of electric charge prediction is the embodiment of the invention provides
Method, first obtains data from insulator metal accessory corrosion quantity of electric charge on-line monitoring system, and the data to obtaining are lacked
Value treatment, makes data have continuity；Metal-ware is corroded the quantity of electric charge and characteristic quantity is extracted from continuous data；With
Neural network algorithm corrodes the quantity of electric charge to metal-ware and characteristic quantity is processed, and obtains neutral net metal-ware corrosion in short-term
The result of quantity of electric charge prediction；The metal-ware corrosion quantity of electric charge is processed with time series analysis method, obtains time series
The result of the quantity of electric charge prediction of metal-ware corrosion in short-term and the result of average annual metal-ware corrosion quantity of electric charge prediction；Again by nerve net
The result of the network quantity of electric charge prediction of metal-ware corrosion in short-term and the result of the time series quantity of electric charge prediction of metal-ware corrosion in short-term
Contrast, obtains the result of the quantity of electric charge prediction of metal-ware corrosion in short-term；Finally according to the corrosion quantity of electric charge prediction of metal-ware in short-term
Result the running status of insulator is estimated, according to average annual metal-ware corrode the quantity of electric charge predict result, to insulation
The metal-ware protection device of son is designed, and solves the problems, such as that the insulator protection device life-span is reduced and waste of materials.
Those skilled in the art considering specification and after putting into practice invention disclosed herein, will readily occur to it is of the invention its
Its embodiment.It is contemplated that cover any modification of the invention, purposes or adaptations, these modifications, purposes or
Person's adaptations follow general principle of the invention and including undocumented common knowledge in the art of the invention
Or conventional techniques.Description and embodiments are considered only as exemplary, and true scope and spirit of the invention are by following
Claim is pointed out.
It should be appreciated that the invention is not limited in the precision architecture being described above and be shown in the drawings, and
And can without departing from the scope carry out various modifications and changes.The scope of the present invention is only limited by appended claim.
Claims (8)
1. a kind of insulator metal accessory corrodes quantity of electric charge Forecasting Methodology, it is characterised in that methods described includes：
Data are obtained from insulator metal accessory corrosion quantity of electric charge on-line monitoring system；
Missing values treatment is carried out to the data, continuous data is obtained；
The extraction of the metal-ware corrosion quantity of electric charge and the extraction of characteristic quantity are carried out to the continuous data, the metal-ware is obtained
The corrosion quantity of electric charge and the characteristic quantity；
The metal-ware is corroded into the treatment that the quantity of electric charge and the characteristic quantity pass through neural network algorithm, neutral net is obtained short
When metal-ware corrosion the quantity of electric charge prediction result；
The metal-ware is corroded into treatment of the quantity of electric charge by time series analysis method, time series metal-ware in short-term is obtained
The result of corrosion quantity of electric charge prediction and the result of average annual metal-ware corrosion quantity of electric charge prediction；
By result and the time series of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term metal-ware in short-term
The result of corrosion quantity of electric charge prediction is contrasted, and obtains the result of the quantity of electric charge prediction of metal-ware corrosion in short-term.
2. method according to claim 1, it is characterised in that described that the metal-ware is corroded into the quantity of electric charge and the spy
The treatment that the amount of levying passes through neural network algorithm, obtains the process of the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term
In, the neural network algorithm includes：
Using the characteristic quantity as variable, the metal-ware is corroded the quantity of electric charge as dependent variable, set up neural network model；
The characteristic quantity and the metal-ware corrosion quantity of electric charge are normalized, the data after being processed；
Data after the treatment are classified, training data and test data is obtained；
The training data is input into the neural network model, the test metal-ware corrosion quantity of electric charge is obtained；
The metal-ware corrosion quantity of electric charge in the test data is subtracted each other with the test metal-ware corrosion quantity of electric charge, is obtained
Difference；
The difference is compared with preset range, master pattern is obtained；
According to the master pattern, the neutral net quantity of electric charge of metal-ware corrosion in short-term is predicted, obtains nerve net
The result of the network quantity of electric charge prediction of metal-ware corrosion in short-term.
3. method according to claim 2, it is characterised in that described to compare the difference with preset range, including：
Judge the difference whether in the preset range；
If the difference is in the preset range, the neural network model is the master pattern；
If the difference obtains the standard not in the preset range after being revised to the neural network model
Model.
4. method according to claim 1, it is characterised in that described that the metal-ware corrosion quantity of electric charge is passed through into the time
The treatment of sequence analysis, the result and average annual metal-ware for obtaining the time series quantity of electric charge prediction of metal-ware corrosion in short-term is rotten
During the result of erosion quantity of electric charge prediction, the time series analysis method includes：
The metal-ware corrosion quantity of electric charge is arranged sequentially in time, setup time sequence；
Influence of the seasonal variations to the time series is determined, seasonal variations factor of influence is obtained；
According to the seasonal variations factor of influence, the time series is revised, the time series after the influence that is eliminated
The change curve changed over time with the metal-ware corrosion quantity of electric charge；
Time series after described eliminating the effects of the act is fitted with the change curve；
The cyclic swing amplitude and Cycle Length of the time series after digital simulation；
According to the cyclic swing amplitude and the Cycle Length to the time series in short-term the metal-ware corrosion quantity of electric charge and
The average annual metal-ware corrosion quantity of electric charge is predicted, and obtains the knot of the time series quantity of electric charge prediction of metal-ware corrosion in short-term
Fruit and the result of average annual metal-ware corrosion quantity of electric charge prediction.
5. method according to claim 1, it is characterised in that described by the neutral net electricity of metal-ware corrosion in short-term
The result that the result of lotus amount prediction is predicted with the time series quantity of electric charge of metal-ware corrosion in short-term is contrasted, and is obtained in short-term
During the result of metal-ware corrosion quantity of electric charge prediction, the method for the contrast, including：
If metal is attached in short-term with the time series for the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term
The result of part corrosion quantity of electric charge prediction is identical, then the identical result is pre- as the metal-ware in short-term corrosion quantity of electric charge
The result of survey；
If the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term is more than time series metal in short-term
Annex corrosion the quantity of electric charge prediction result, then using the neutral net in short-term metal-ware corrosion the quantity of electric charge prediction result as
The result of the corrosion of the metal-ware in short-term quantity of electric charge prediction；
If the result of the neutral net quantity of electric charge prediction of metal-ware corrosion in short-term is less than time series metal in short-term
Annex corrosion the quantity of electric charge prediction result, then using the time series in short-term metal-ware corrosion the quantity of electric charge prediction result as
The result of the corrosion of the metal-ware in short-term quantity of electric charge prediction.
6. method according to claim 1, it is characterised in that described by neutral net metal-ware corrosion in short-term
The result that the result of quantity of electric charge prediction is predicted with the time series quantity of electric charge of metal-ware corrosion in short-term is contrasted, and obtains short
When metal-ware corrosion the quantity of electric charge prediction result after, methods described also includes：Electricity is corroded according to the metal-ware in short-term
The result of lotus amount prediction, the running status to the insulator is estimated.
7. method according to claim 1, it is characterised in that it is described the metal-ware corrosion quantity of electric charge is passed through when
Between sequence analysis treatment, obtain time series in short-term metal-ware corrosion the quantity of electric charge prediction result and average annual metal-ware
After the result of corrosion quantity of electric charge prediction, methods described also includes：Predicted according to the average annual metal-ware corrosion quantity of electric charge
As a result, the metal-ware protection device to the insulator is designed.
8. method according to claim 1, it is characterised in that the method that data are carried out with missing values treatment is for nearby
Interpolation method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710127276.XA CN106886651B (en) | 2017-03-06 | 2017-03-06 | Method for predicting corrosion charge quantity of insulator metal accessory |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710127276.XA CN106886651B (en) | 2017-03-06 | 2017-03-06 | Method for predicting corrosion charge quantity of insulator metal accessory |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106886651A true CN106886651A (en) | 2017-06-23 |
| CN106886651B CN106886651B (en) | 2020-07-31 |
Family
ID=59179144
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710127276.XA Active CN106886651B (en) | 2017-03-06 | 2017-03-06 | Method for predicting corrosion charge quantity of insulator metal accessory |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106886651B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117368799A (en) * | 2023-12-07 | 2024-01-09 | 山西思极科技有限公司 | Diagnosis method for short-circuit fault of power transmission line of power system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102736025A (en) * | 2012-06-29 | 2012-10-17 | 沈阳工业大学 | Device and method for predicting electric remaining service life of circuit breaker |
| CN102809568A (en) * | 2012-08-28 | 2012-12-05 | 广东电网公司佛山供电局 | Method and system for monitoring contamination distribution of insulator |
| CN103046109A (en) * | 2012-12-29 | 2013-04-17 | 南方电网科学研究院有限责任公司 | Testing device and testing method for direct current porcelain insulator cap corrosion mechanism research |
| CN105510525A (en) * | 2015-07-23 | 2016-04-20 | 深圳供电局有限公司 | Electric-transmission-line dirt monitoring apparatus and method |
| CN105740582A (en) * | 2016-03-16 | 2016-07-06 | 华北电力大学 | Prediction method for ageing state of composite insulator |
-
2017
- 2017-03-06 CN CN201710127276.XA patent/CN106886651B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102736025A (en) * | 2012-06-29 | 2012-10-17 | 沈阳工业大学 | Device and method for predicting electric remaining service life of circuit breaker |
| CN102809568A (en) * | 2012-08-28 | 2012-12-05 | 广东电网公司佛山供电局 | Method and system for monitoring contamination distribution of insulator |
| CN103046109A (en) * | 2012-12-29 | 2013-04-17 | 南方电网科学研究院有限责任公司 | Testing device and testing method for direct current porcelain insulator cap corrosion mechanism research |
| CN105510525A (en) * | 2015-07-23 | 2016-04-20 | 深圳供电局有限公司 | Electric-transmission-line dirt monitoring apparatus and method |
| CN105740582A (en) * | 2016-03-16 | 2016-07-06 | 华北电力大学 | Prediction method for ageing state of composite insulator |
Non-Patent Citations (4)
| Title |
|---|
| CHENLONG ZHAO: "Development of Contamination Flashover Pre-warning", 《DEVELOPMENT OF CONTAMINATION FLASHOVER PRE-WARNING SYSTEM AND ANALYSIS OF OPERATING EXPERIENCE》 * |
| FANGCHENG LÜ等: "Particle Swarm Optimization-based BP Neural Network for UHV DC Insulator Pollution Forecasting", 《JOURNAL OF ENGINEERING SCIENCE AND TECHNOLOGY REVIEW》 * |
| 毛颖科: "基于 BP 人工神经网络的绝缘子泄漏电流预测", 《中国电机工程学报》 * |
| 陈升等: "基于 BP 神经网络的绝缘子泄漏电流量预测方法", 《浙江电力》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117368799A (en) * | 2023-12-07 | 2024-01-09 | 山西思极科技有限公司 | Diagnosis method for short-circuit fault of power transmission line of power system |
| CN117368799B (en) * | 2023-12-07 | 2024-02-23 | 山西思极科技有限公司 | Diagnosis method for short-circuit fault of power transmission line of power system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106886651B (en) | 2020-07-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106600463B (en) | Local shape similarity ultra-short term load prediction method and device | |
| CN109613338B (en) | Low-voltage user loop impedance estimation method based on unitary model | |
| CN112149873B (en) | Low-voltage station line loss reasonable interval prediction method based on deep learning | |
| CN110807550A (en) | Distribution transformer overload identification early warning method based on neural network and terminal equipment | |
| CN106295207B (en) | Insulator dirty degree appraisal procedure based on meteorological data statistics | |
| CN106452931A (en) | Monitoring index, domain value discovery method, domain value adjusting method and automatic monitoring system | |
| CN105117538A (en) | Method for warning waving of power transmission channels | |
| CN116523140A (en) | Method and device for detecting electricity theft, electronic equipment and storage medium | |
| CN115293584A (en) | Overhead transmission line risk prediction method, system and terminal | |
| CN114742283A (en) | Medium-voltage line loop closing current estimation and risk assessment method based on deep learning network | |
| CN114580829A (en) | Power utilization safety sensing method, equipment and medium based on random forest algorithm | |
| CN106529741A (en) | Space relevant characteristic-based ultra-short-period wind power prediction method | |
| CN106886651A (en) | Insulator metal accessory corrodes quantity of electric charge Forecasting Methodology | |
| CN114896872A (en) | Comprehensive evaluation method for icing state of high-voltage transmission line | |
| CN104732107B (en) | Using medium parameter as the residual life of sleeve of transformer Forecasting Methodology for assessing parameter | |
| CN109858668B (en) | Coordination prediction method for power load region in thunder and lightning climate | |
| CN114357670A (en) | Power distribution network power consumption data abnormity early warning method based on BLS and self-encoder | |
| CN110728395A (en) | Main transformer short-term power load calculation method and device, computer and storage medium | |
| CN104915727B (en) | Various dimensions allomer BP neural network optical power ultra-short term prediction method | |
| CN116894151B (en) | Intelligent power grid abnormal data cleaning method based on artificial intelligence | |
| CN117150808A (en) | Method, system and equipment for evaluating toughness of power transmission line in strong convection weather | |
| CN107563641B (en) | Disaster-resistant multi-scene differentiation planning method for power distribution network considering disaster preference | |
| CN112597629B (en) | Method for establishing decision tree model for judging whether icing exists on wire and method for predicting icing duration | |
| CN106570561B (en) | A kind of insoluble sediment density forecasting system of insulator surface and method | |
| CN115224684A (en) | Intelligent power distribution network risk state identification method and system based on immune hazard theory |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210207 Address after: 650217 No. 105 Yunda West Road, Kunming Economic and Technological Development Zone, Yunnan Province Patentee after: YUNNAN POWER GRID CO., LTD. ELECTRIC POWER Research Institute Address before: 650217 No. 105 Yunda West Road, Kunming Economic and Technological Development Zone, Yunnan Province Patentee before: YUNNAN POWER GRID CO., LTD. ELECTRIC POWER Research Institute Patentee before: Tsinghua University Shenzhen Graduate School |
|
| TR01 | Transfer of patent right |