CN104166718B - A kind of bad data detection and identification method suitable for bulk power grid - Google Patents

A kind of bad data detection and identification method suitable for bulk power grid Download PDF

Info

Publication number
CN104166718B
CN104166718B CN201410407189.6A CN201410407189A CN104166718B CN 104166718 B CN104166718 B CN 104166718B CN 201410407189 A CN201410407189 A CN 201410407189A CN 104166718 B CN104166718 B CN 104166718B
Authority
CN
China
Prior art keywords
data
suspicious
sus
association
bad
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.)
Active
Application number
CN201410407189.6A
Other languages
Chinese (zh)
Other versions
CN104166718A (en
Inventor
赵昆
窦成龙
颜磊
邹昱
徐杰
张印
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
State Grid Corp of China SGCC
China Electric Power Research Institute Co Ltd CEPRI
State Grid Liaoning Electric Power Co Ltd
Original Assignee
State Grid Corp of China SGCC
China Electric Power Research Institute Co Ltd CEPRI
State Grid Liaoning Electric Power Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by State Grid Corp of China SGCC, China Electric Power Research Institute Co Ltd CEPRI, State Grid Liaoning Electric Power Co Ltd filed Critical State Grid Corp of China SGCC
Priority to CN201410407189.6A priority Critical patent/CN104166718B/en
Publication of CN104166718A publication Critical patent/CN104166718A/en
Application granted granted Critical
Publication of CN104166718B publication Critical patent/CN104166718B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/903—Querying
    • G06F16/90335—Query processing
    • 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

Abstract

The present invention relates to a kind of bad data detection and identification method suitable for bulk power grid, methods described is used in during Power system state estimation；It the described method comprises the following steps：Detection bad data simultaneously determines suspicious data collection therein；Determine the association suspicious data collection that the suspicious data is concentrated；Determine the bad data that suspicious data is concentrated；Delete the bad data that the suspicious data is concentrated.The present invention can be widely used in current net provincial power network and nationwide integrated power grid, and can adapt to following scale grid line analysis demand, can fast and accurately reject the bad data in measurement information, improve precision of state estimation.

Description

A kind of bad data detection and identification method suitable for bulk power grid
Technical field：
The present invention relates to a kind of bad data detection and identification method, it is more particularly to a kind of suitable for the bad of bulk power grid Data Detection and discrimination method.
Background technology：
Power system state estimation is the important component of modern energy management systems, and its metric data is largely originated In SCADA system, information is in addition to containing normal measurement noise, it is also possible to contain bad data.The presence of bad data, will Cause estimated result to be contaminated, or even be allowed to serious distortion.The detection of bad data and identification are Power system state estimations One of critical function, its object is to exclude to measure a small amount of bad data accidentally occurred in sampled data, to improve state The reliability of estimation.
Power system state estimation is that data precision is improved using the redundancy of real-time measurement system, excludes random disturbances Caused error message, and then the running status of estimation or forecasting system.Raw data detection based on state estimation is with distinguishing The method of knowledge mainly has residual error search method, Non quadratic criteria method, zero residual error method and estimation identification method.These methods will mainly add Power residual error or residual value are used as characteristic value, it is assumed that it obeys a certain probability distribution, and true according to certain level of confidence A fixed threshold value, carries out hypothesis testing.Find after suspicious measurement data, it is excluded from measurement data or reduces its power Value, obtains new state estimation.Shortcoming that may be present in above detection and identification method：It is possible that residual contamination and residual error Phenomenon is flooded, so as to cause missing inspection or flase drop, the effect of identification is influenceed, further influences estimation effect.Because algorithm is using non- Need to carry out multiple state estimation in linear residual equation, identification process, therefore amount of calculation is greatly, and state estimation has to real-time Certain requirement, should not be used as online real-time estimation method.In addition using linearisation residual equation, residual sensitivity matrix is utilized Submatrix calculate the estimate of measurement, because sensitivity matrix is the full battle array of higher-dimension, therefore this method amount of calculation is still very big. In addition, in the case of there are multiple bad datas, the phenomenon for the identification that makes often to make a mistake in this way.For various Defect in method, state estimate also occurs in that many improved research branches.
Traditional detection, using weighted residual rw or residual rn as bad data recognition method, obtains suspicious with discrimination method After metric data collection, its weights is reduced one by one or is directly rejected from metric data, then re-starts state estimation, so circulation Untill meeting the condition of convergence.The shortcoming of such method is that amount of calculation is very big, and calculating speed is slow；If defining detection number of times, Easily there is residual contamination again and residual error floods phenomenon, cause missing inspection or miss detection, influence identification effect.Especially measuring number According to few, and data value is inaccurate, it is insecure in the case of (such as in power distribution network), it is more tight that residual contamination and residual error flood phenomenon Weight.
Power system state estimation has at home and abroad developed decades, and bad data detection and identification research also never Stagnate, but still without the effective ways for proposing a good detection and identification bad data, it is proposed that one kind is applied to bulk power grid Bad data detection and identification method to overcome drawbacks described above.
The content of the invention：
It is an object of the invention to provide a kind of bad data detection and identification method suitable for bulk power grid, this method is extensive Applied to current net provincial power network and nationwide integrated power grid, and following scale grid line analysis demand is can adapt to, can be quickly accurate The true bad data rejected in measurement information, improves precision of state estimation.
To achieve the above object, the present invention uses following technical scheme：A kind of raw data detection suitable for bulk power grid With discrimination method, methods described is used in during Power system state estimation；It the described method comprises the following steps：
(1) the suspicious data collection SUS of the whole network is determined；
(2) the association suspicious data collection sus [n] that the suspicious data is concentrated is determined；
(3) bad data in identification association suspicious data collection sus [n].
The present invention provides a kind of bad data detection and identification method suitable for bulk power grid, the determination of the step (1) Process is：
(1-1) carries out the state estimation of the whole network；
(1-2) passes through rnDetection method determines suspicious data collection SUS；
The rnThe detection process of detection method is：
In formula, H0It is not suspicious metric data, H for i-th of metric data1It is suspicious measurement number for i-th of metric data According to；rN,iFor i-th of standardized residual,For the threshold value of i-th of standardized residual.
The present invention provides a kind of bad data detection and identification method suitable for bulk power grid, the door of the standardized residual Threshold value is determined by following steps：
Define W and WNRespectively m × m ranks residual sensitivity matrix and the sensitivity matrix of m × m rank standardized residuals, then Have
Standardized residual rNFor：
rN=WNv (3)
Wherein, diagonal matrix D=diag [WR], R are the weight matrix corresponding with measurement, and v is the m dimensions containing bad data Error vector；
Under the conditions of normal measure, there is standardized residual rNCovariance matrix be：
Wherein, E represents expectation function, rN,zFor the standardized residual of all measurements, therefore have
Probability of false detection P is taken so working ase=0.005, take the threshold value of the standardized residualFor
The present invention provides a kind of bad data detection and identification method suitable for bulk power grid, the determination of the step (2) Process is：
The search of (2-1) incidence matrix；
(2-2) ultimately forms association suspicious data collection.
The present invention provides a kind of bad data detection and identification method suitable for bulk power grid, and the step (2-1) is in electricity In Force system state estimation, the association suspicious data search procedure based on Jacobian matrix is：
Step (2-1-1)：The number n for setting association suspicious data collection is 1；
Step (2-1-2)：Newly-built n-th of association suspicious data collection sus [n]；
Step (2-1-3)：Whether be empty, if not being idle running step (2-1- if judging all suspicious data collection SUS of the whole network 4), if sky, then search procedure terminates；
Step (2-1-4)：The maximum data of standardized residual are taken out from SUS, it is assumed that for i-th of data, then count this Reject, and be added in sus [n] according to from SUS；
Step (2-1-5)：The nonzero element searched in the row element of Jacobian matrix i-th, and record these nonzero elements Row number, forms set LOR；
Step (2-1-6)：Take out row number successively from LOR, and search for the corresponding row of Jacobian matrix, record non-in each row The line number of 0 element, forms set ROW；
Step (2-1-7)：Judge whether the corresponding data of line number belong to set SUS in ROW successively, if there is belonging to collection SUS data are closed, then all data for belonging to set SUS are added in association suspicious data collection sus [n], while these The data for belonging to set SUS are rejected from SUS, are gone to step (2-1-8)；If there is no the data for belonging to set SUS, then n-th Individual suspicious measurement collection sus [n] formation of association is finished, and performs n=n+1, and go to step (2-1-2)；
Step (2-1-8)：Take out in step (2-1-7) and belong to the set SUS corresponding line number of data, and search for Jacobi Non-zero element in matrix correspondence row element, records row number where non-zero element, forms LOR, go to step (2-1-6)；
8 steps more than, form n association suspicious data collection of the step (2-2).
The present invention provides a kind of bad data detection and identification method suitable for bulk power grid, according to the suspicious number of association Each sus [n] weighted residual quadratic sum J (x) is calculated according to collection sus [n]：
J (x)=[z-h (x)]TR-1[z-h(x)] (7)
Wherein, z is association suspicious data, and R is corresponding diagonal weight matrix, and h (x) is the calculating letter for associating suspicious data Number.
The present invention provides a kind of bad data detection and identification method suitable for bulk power grid, and the step (3) is by double Layer bad data recognition method determines bad data.
The present invention provides a kind of bad data detection and identification method suitable for bulk power grid, and the double-deck bad data is distinguished Knowledge method process is：
Step (8-1)：Set current association to be processed is suspicious to measure collection sequence number k=1；
Step (8-2)：K-th of association suspicious data collection is taken out, the weighted residual square of all data in this set is calculated Be designated as J (x)；
Step (8-3)：Take out k-th of association suspicious data and concentrate the data sus that standardized residual is maximum and is not identified [k, i], it is quick to correct state estimation factor table and quickly delete the suspicious data sus based on Givens orthogonal row converter techniques [k,i]；
Step (8-4)：The weighted residual that all data in suspicious data collection sus [k] are associated described in rapid solving is put down Side and J ' (x)；
Step (8-5)：Whether recognize the suspicious data is bad data；
If the suspicious data sus [k, i] is bad data, and goes to step (8-6)；If the suspicious data sus [k, i] is no It is bad data, recovers the suspicious data sus [k, i] into association suspicious data collection sus [k], put identification mark, and pass through The quick modifying factor sublist of Givnes orthogonal row converter techniques, then goes to step (8-3)；
Step (8-6)：The identification association suspicious data concentrates whether also there is bad data；
If the association suspicious data is concentrated without bad data, continue next association suspicious data set analysis, perform k =k+1, and go to step (8-7), k is association suspicious data collection number；If the association suspicious data, which is concentrated, also has bad data, Identification need to be continued, gone to step (8-3)；
Step (8-7)：Judge whether k is more than n, if it is, bad data recognition process terminates；If it is not, then turning Step (8-2).
The present invention provides a kind of bad data detection and identification method suitable for bulk power grid, and the step (8-5) passes through Following formula judges whether bad data：
| J (x)-J ' (x) | ＞ ε1, ε1For threshold value,；If inequality is set up, the data are bad data, if inequality Invalid, then bad data is not present in the data；
The step (8-6) judges the association suspicious data concentrates whether also there is bad data by following formula：
J ' (x) ＜ ε2, ε2For threshold value；If inequality is set up, the association suspicious data is concentrated without bad data； If inequality is invalid, the association suspicious data, which is concentrated, also has bad data.
The present invention provides a kind of bad data detection and identification method suitable for bulk power grid, completes detection and was recognizing The data respective weights of the Power system state estimation are modified, gone forward side by side by Cheng Hou according to the bad data picked out Row iteratively faster is calculated, and obtains the accurate Power system state estimation data section.
With immediate prior art ratio, the present invention, which provides technical scheme, has following excellent effect
1st, the present invention is the important component of power state estimation, is requisite measure and the pass for improving precision of state estimation Key technology；
2nd, the present invention can fast and accurately reject the larger bad data of error, can while bad data is accurately positioned Avoid unnecessary search；
3rd, the present invention can improve raw data detection identification sensitivity and the degree of accuracy, and the final state estimation that improves calculates essence Degree；
4th, the present invention has than wide application prospect, is widely used in current net provincial power network and nationwide integrated power grid, And can adapt to following scale grid line analysis demand；
5th, the present invention can further lift intelligent grid dispatching technique and support system in the Demonstration Application of scheduling institutions at different levels Accuracy and calculating speed that state estimation of uniting is calculated, the becoming more meticulous of comprehensive support intelligent grids at different levels scheduling, lean and one The ability of bodyization running；
6th, the present invention can effectively improve the real-time that ultra-large powernet analysis is calculated, and be extra-high voltage bulk power grid Safety, high-quality and economical operation provide strong technical support；
7th, the present invention enables dispatcher to track operation of power networks state in time by accurate status estimated result, finds different Normal operation conditions simultaneously makes control decision, it is to avoid fault spread, further lifting intelligent grid supporting system technology technology Level and operation stability, further lifting dispatching of power netwoks are controlled the ability of bulk power grid；
8th, the present invention ensures safe and stable bulk power grid, high-quality, economical operation, to lifting electrical power services quality and guarantee society The stable development of meeting has important realistic meaning.
Brief description of the drawings
Fig. 1 is straightforward procedure flow chart of the invention；
Fig. 2 is specific method flow chart of the invention.
Embodiment
With reference to embodiment, the invention will be described in further detail.
Embodiment 1：
As shown in Figure 1-2, a kind of bad data detection and identification method suitable for bulk power grid of invention of this example, the side Method is used in during Power system state estimation；It the described method comprises the following steps：
(1) the suspicious data collection SUS of the whole network is determined；
(2) the association suspicious data collection sus [n] that the suspicious data is concentrated is determined；
(3) bad data in identification association suspicious data collection sus [n].
The determination process of the step (1) is：
(1-1) carries out the state estimation of the whole network.
(1-2)rnDetection method determines suspicious data collection SUS
The step (1-1) is done the state estimation an of the whole network by original measurement.
The rnThe detection process of detection method is：
In formula, H0It is not suspicious metric data, H for i-th of metric data1It is suspicious measurement number for i-th of metric data According to；rN,iFor i-th of standardized residual,For the threshold value of i-th of standardized residual.
The threshold value of the standardized residual is determined by following steps：
Define W and WNRespectively m × m ranks residual sensitivity matrix and the sensitivity matrix of m × m rank standardized residuals, then Have
Standardized residual rNFor：
rN=WNv (3)
Wherein, diagonal matrix D=diag [WR], R are the weight matrix corresponding with measurement, and v is the m dimensions containing bad data Error vector；
Therefore, under the conditions of normal measure, there is standardized residual rNCovariance matrix be：
Wherein, E represents expectation function, rN,zFor the standardized residual of all measurements, therefore have
Probability of false detection P is taken so working ase=0.005, take the threshold value of the standardized residualFor
The determination process of the step (2) is：
The search of (2-1) incidence matrix；
(2-2) ultimately forms association suspicious data collection.
The step (2-1) is in Power system state estimation, and the association suspicious data based on Jacobian matrix was searched for Cheng Wei：
Step (2-1-1)：The number n for setting association suspicious data collection is 1；
Step (2-1-2)：Newly-built n-th of association suspicious data collection sus [n]；
Step (2-1-3)：Whether be empty, if not being idle running step (2-1- if judging all suspicious data collection SUS of the whole network 4), if sky, then search procedure terminates；
Step (2-1-4)：The maximum data of standardized residual are taken out from SUS, it is assumed that for i-th of data, then count this Reject, and be added in sus [n] according to from SUS；
Step (2-1-5)：The nonzero element searched in the row element of Jacobian matrix i-th, and record these nonzero elements Row number, forms set LOR；
Step (2-1-6)：Take out row number successively from LOR, and search for the corresponding row of Jacobian matrix, record non-in each row The line number of 0 element, forms set ROW；
Step (2-1-7)：Judge whether the corresponding data of line number belong to set SUS in ROW successively, if there is belonging to collection SUS data are closed, then all data for belonging to set SUS are added in association suspicious data collection sus [n], while these The data for belonging to set SUS are rejected from SUS, are gone to step (2-1-8)；If there is no the data for belonging to set SUS, then n-th Individual suspicious measurement collection sus [n] formation of association is finished, and performs n=n+1, and go to step (2-1-2)；
Step (2-1-8)：Take out in step (2-1-7) and belong to the set SUS corresponding line number of data, and search for Jacobi Non-zero element in matrix correspondence row element, records row number where non-zero element, forms LOR, go to step (2-1-6)；
8 steps more than, form n association suspicious data collection of the step (2-2).
Each sus [n] weighted residual quadratic sum J (x) is calculated according to the association suspicious data collection sus [n]：
J (x)=[z-h (x)]TR-1[z-h(x)] (7)
Wherein, z is association suspicious data, and R is corresponding diagonal weight matrix, and h (x) is the calculating letter for associating suspicious data Number.
The double-deck bad data recognition method process is：
Step (8-1)：Set current association to be processed is suspicious to measure collection sequence number k=1；
Step (8-2)：K-th of association suspicious data collection is taken out, the weighted residual square of all data in this set is calculated Be designated as J (x)；
Step (8-3)：Take out k-th of association suspicious data and concentrate the data sus that standardized residual is maximum and is not identified [k, i], it is quick to correct state estimation factor table and quickly delete the suspicious data sus based on Givens orthogonal row converter techniques [k,i]；
Step (8-4)：The weighted residual that all data in suspicious data collection sus [k] are associated described in rapid solving is put down Side and J ' (x)；
Step (8-5)：Whether recognize the suspicious data is bad data；
If the suspicious data sus [k, i] is bad data, and goes to step (8-6)；If the suspicious data sus [k, i] is no It is bad data, recovers the suspicious data sus [k, i] into association suspicious data collection sus [k], put identification mark, and pass through The quick modifying factor sublist of Givnes orthogonal row converter techniques, then goes to step (8-3)；
Step (8-6)：The identification association suspicious data concentrates whether also there is bad data；
If the association suspicious data is concentrated without bad data, continue next association suspicious data set analysis, perform k =k+1, and go to step (8-7), k is the suspicious collection number of association；, need to be after if the association suspicious data is concentrated and also there is bad data Continuous identification, goes to step (8-3)；
Step (8-7)：Judge whether k is more than n, if it is, bad data recognition process terminates；If it is not, then turning Step (8-2).
The step (8-5) judges whether bad data by following formula：
| J (x)-J ' (x) | ＞ ε1, ε1For threshold value,；If inequality is set up, the data are bad data, if inequality Invalid, then bad data is not present in the data；
The step (8-6) judges the association suspicious data concentrates whether also there is bad data by following formula：
J ' (x) ＜ ε2, ε2For threshold value；If inequality is set up, the association suspicious data is concentrated without bad data； If inequality is invalid, the association suspicious data, which is concentrated, also has bad data.
After detection and identification process is completed, according to the bad data picked out to the Power system state estimation Data respective weights are modified, and carry out iteratively faster calculating, are obtained the accurate Power system state estimation data and are broken Face.
Using suspicious measurement Fast search technique is associated, bad data recognition scope is reduced, bad data identification speed is improved Degree and sensitivity；By many related double-deck identification techniques of bad data, many related bad data recognition accuracys are substantially increased And precision, finally give the accurate Power system state estimation data section.
Finally it should be noted that：The above embodiments are merely illustrative of the technical scheme of the present invention and are not intended to be limiting thereof, institute The those of ordinary skill in category field with reference to above-described embodiment although should be understood：Still can be to embodiment of the invention Modify or equivalent substitution, these any modifications or equivalent substitution without departing from spirit and scope of the invention, in Shen Within claims of the invention that please be pending.

Claims (8)

1. a kind of bad data detection and identification method suitable for bulk power grid, methods described is used in Power system state estimation During；It is characterized in that：It the described method comprises the following steps：
(1) the suspicious data collection SUS of the whole network is determined；
(2) the association suspicious data collection sus [n] that the suspicious data is concentrated is determined；
(3) bad data in identification association suspicious data collection sus [n]；
The determination process of the step (2) is：
The search of (2-1) incidence matrix；
(2-2) ultimately forms association suspicious data collection；
The step (2-1) is in Power system state estimation, and the association suspicious data search procedure based on Jacobian matrix is：
Step (2-1-1)：The number n for setting association suspicious data collection is 1；
Step (2-1-2)：Newly-built n-th of association suspicious data collection sus [n]；
Step (2-1-3)：Whether be empty, if not being idle running step (2-1-4), such as if judging all suspicious data collection SUS of the whole network Fruit is sky, then search procedure terminates；
Step (2-1-4)：The maximum data of standardized residual are taken out from SUS, it is assumed that for i-th of data, then by the data from Reject, and be added in sus [n] in SUS；
Step (2-1-5)：The nonzero element searched in the row element of Jacobian matrix i-th, and record the row number of these nonzero elements, Form set LOR；
Step (2-1-6)：Take out row number successively from LOR, and search for the corresponding row of Jacobian matrix, record non-zero member in each row The line number of element, forms set ROW；
Step (2-1-7)：Judge whether the corresponding data of line number belong to set SUS in ROW successively, if there is belonging to set SUS data, then be added to all data for belonging to set SUS in association suspicious data collection sus [n], while these are belonged to Reject, go to step (2-1-8) from SUS in set SUS data；If there is no the data for belonging to set SUS, then n-th Suspicious measurement collection sus [n] formation of association is finished, and performs n=n+1, and go to step (2-1-2)；
Step (2-1-8)：Take out in step (2-1-7) and belong to the set SUS corresponding line number of data, and search for Jacobian matrix Non-zero element in correspondence row element, records row number where non-zero element, forms LOR, goes to step (2-1-6)；
8 steps more than, form n association suspicious data collection of the step (2-2).
2. a kind of bad data detection and identification method suitable for bulk power grid as claimed in claim 1, it is characterised in that：Institute The determination process for stating step (1) is：
(1-1) carries out the state estimation of the whole network；
(1-2) passes through rnDetection method determines suspicious data collection SUS；
The rnThe detection process of detection method is：
In formula, H0It is not suspicious metric data, H for i-th of metric data1It is suspicious metric data for i-th of metric data；rN,i For i-th of standardized residual, r 'N,iFor the threshold value of i-th of standardized residual.
3. a kind of bad data detection and identification method suitable for bulk power grid as claimed in claim 2, it is characterised in that： The threshold value of the standardized residual is determined by following steps：
Define W and WNRespectively m × m ranks residual sensitivity matrix and the sensitivity matrix of m × m rank standardized residuals, then have
Standardized residual rNFor：
rN=WNv (3)
Wherein, diagonal matrix D=diag [WR], R are the weight matrix corresponding with measurement, and v is the m dimension errors containing bad data Vector；
Under the conditions of normal measure, there is standardized residual rNCovariance matrix be：
Wherein, E represents expectation function, rN,zFor the standardized residual of all measurements, therefore have
Probability of false detection P is taken so working ase=0.005, take the threshold value r of the standardized residualN′,iFor
r′N,i=2.81 (i=1,2 ..., m) (6).
4. a kind of bad data detection and identification method suitable for bulk power grid as claimed in claim 1, it is characterised in that：Root Each sus [n] weighted residual quadratic sum J (x) is calculated according to the association suspicious data collection sus [n]：
J (x)=[z-h (x)]TR-1[z-h(x)] (7)
Wherein, z is association suspicious data, and R is corresponding diagonal weight matrix, and h (x) is the calculating function for associating suspicious data.
5. a kind of bad data detection and identification method suitable for bulk power grid as claimed in claim 4, it is characterised in that：Institute State step (3) and bad data is determined by double-deck bad data recognition method.
6. a kind of bad data detection and identification method suitable for bulk power grid as claimed in claim 5, it is characterised in that：Institute Stating double-deck bad data recognition method process is：
Step (8-1)：Set current association to be processed is suspicious to measure collection sequence number k=1；
Step (8-2)：K-th of association suspicious data collection is taken out, the weighted residual quadratic sum of all data in this set, note is calculated For J (x)；
Step (8-3)：Take out data sus that k-th of association suspicious data concentrate standardized residual maximum and be not identified [k, I], based on Givens orthogonal row converter techniques, quick amendment state estimation factor table and quickly delete the suspicious data sus [k, i]；
Step (8-4)：The weighted residual quadratic sum of all data in suspicious data collection sus [k] is associated described in rapid solving J′(x)；
Step (8-5)：Whether recognize the suspicious data is bad data；
If the suspicious data sus [k, i] is bad data, and goes to step (8-6)；If the suspicious data sus [k, i] is not bad Data, recover the suspicious data sus [k, i] into association suspicious data collection sus [k], put identification mark, and pass through The quick modifying factor sublist of Givnes orthogonal row converter techniques, then goes to step (8-3)；
Step (8-6)：The identification association suspicious data concentrates whether also there is bad data；
If the association suspicious data is concentrated without bad data, continue next association suspicious data set analysis, perform k=k+ 1, and go to step (8-7), k is association suspicious data collection number；, need to be after if the association suspicious data is concentrated and also there is bad data Continuous identification, goes to step (8-3)；
Step (8-7)：Judge whether k is more than n, if it is, bad data recognition process terminates；If it is not, then going to step (8-2)。
7. a kind of bad data detection and identification method suitable for bulk power grid as claimed in claim 6, it is characterised in that：Institute State step (8-5) and bad data is judged whether by following formula：
| J (x)-J ' (x) | ＞ ε1, ε1For threshold value；If inequality set up, the data be bad data, if inequality not into Vertical, then bad data is not present in the data；
The step (8-6) judges the association suspicious data concentrates whether also there is bad data by following formula：
J ' (x) ＜ ε2, ε2For threshold value；If inequality is set up, the association suspicious data is concentrated without bad data；If no Equation is invalid, then the association suspicious data, which is concentrated, also has bad data.
8. a kind of bad data detection and identification method suitable for bulk power grid as claimed in claim 1, it is characterised in that： Detection is completed with after identification process, being weighed according to the bad data picked out to the data correspondence of the Power system state estimation It is modified again, and carries out iteratively faster calculating, obtains the accurate Power system state estimation data section.
CN201410407189.6A 2014-08-18 2014-08-18 A kind of bad data detection and identification method suitable for bulk power grid Active CN104166718B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410407189.6A CN104166718B (en) 2014-08-18 2014-08-18 A kind of bad data detection and identification method suitable for bulk power grid

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410407189.6A CN104166718B (en) 2014-08-18 2014-08-18 A kind of bad data detection and identification method suitable for bulk power grid

Publications (2)

Publication Number Publication Date
CN104166718A CN104166718A (en) 2014-11-26
CN104166718B true CN104166718B (en) 2017-11-03

Family

ID=51910531

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410407189.6A Active CN104166718B (en) 2014-08-18 2014-08-18 A kind of bad data detection and identification method suitable for bulk power grid

Country Status (1)

Country Link
CN (1) CN104166718B (en)

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104766175A (en) * 2015-04-16 2015-07-08 东南大学 Power system abnormal data identifying and correcting method based on time series analysis
CN106295850A (en) * 2016-07-15 2017-01-04 广东电网有限责任公司电力科学研究院 A kind of bad data recognition method and system based on historical metrology
CN106887842A (en) * 2017-04-11 2017-06-23 广东电网有限责任公司电力调度控制中心 A kind of bad data discrimination method based on XML/E format multiple-source electric network datas
CN107133721A (en) * 2017-04-17 2017-09-05 国网江苏省电力公司 Power distribution network based on gray theory measures bad data relation factor analysis method
CN107506824B (en) * 2017-08-31 2021-01-26 广东工业大学 Method and device for detecting bad observation data of power distribution network
CN108182257A (en) * 2017-12-29 2018-06-19 东北电力大学 A kind of GSA bad data detection and identification methods based on the optimization of areal concentration statistical method
CN108333468B (en) * 2018-01-05 2019-08-06 南京邮电大学 The recognition methods of bad data and device under a kind of active power distribution network
CN111931318B (en) * 2020-07-07 2022-09-27 中国电力科学研究院有限公司 Power supply path analysis method and system based on graph calculation
CN112230087B (en) * 2020-10-13 2022-08-05 全球能源互联网研究院有限公司 Linear state estimation method and device, electronic equipment and storage medium
CN114372235A (en) * 2021-12-30 2022-04-19 东方电子股份有限公司 Power system state estimation method for preventing singular transformation matrix

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103324847A (en) * 2013-06-17 2013-09-25 西南交通大学 Method for detecting and identifying dynamic bad data of electric power system
CN103745109A (en) * 2014-01-10 2014-04-23 国家电网公司 Bad data detection and identification method based on measurement of PMU (Phasor Measurement Unit) and measurement of SCADA (Supervisory Control and Data Acquisition)

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5048625B2 (en) * 2008-10-09 2012-10-17 株式会社日立製作所 Anomaly detection method and system

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103324847A (en) * 2013-06-17 2013-09-25 西南交通大学 Method for detecting and identifying dynamic bad data of electric power system
CN103745109A (en) * 2014-01-10 2014-04-23 国家电网公司 Bad data detection and identification method based on measurement of PMU (Phasor Measurement Unit) and measurement of SCADA (Supervisory Control and Data Acquisition)

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
电力系统不良数据检测和辨识方法研究;张永超;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20100215(第02期);第2-7、11-12、42-48页，图1-1 *
电力系统状态估计不良数据检测与辨识方法研究与应用;周小宝;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20140515(第05期);第9-10页 *

Also Published As

Publication number Publication date
CN104166718A (en) 2014-11-26

Similar Documents

Publication Publication Date Title
CN104166718B (en) A kind of bad data detection and identification method suitable for bulk power grid
CN104836223B (en) Power grid parameter error and bad data coordinated identification and estimation method
CN103324847B (en) Electrical Power System Dynamic bad data detection and identification method
CN104077445B (en) Accelerated life test statistical analysis technique based on fuzzy theory
CN108536971A (en) A kind of Structural Damage Identification based on Bayesian model
CN104221475B (en) Fault detection, localization and performance monitoring of photosensors for lighting controls
CN103973697B (en) A kind of thing network sensing layer intrusion detection method
WO2022021726A1 (en) Pmu-based power system state estimation performance evaluation method
CN110472671B (en) Multi-stage-based fault data preprocessing method for oil immersed transformer
CN106407555B (en) Acceleration degraded data analysis method based on accelerated factor principle of invariance
CN105930900B (en) The Forecasting Methodology and system of a kind of hybrid wind power generation
CN107370150A (en) The Power system state estimation Bad data processing method measured based on synchronized phasor
CN107704992A (en) The method and device of transmission line lightning stroke risk assessment
CN105406471A (en) Bad data identification and estimation method for power grid
CN108960334A (en) A kind of multi-sensor data Weighted Fusion method
CN105183938A (en) Bad data identification and estimation method for power grid
CN110334478A (en) Machinery equipment abnormality detection model building method, detection method and model
CN113721182B (en) Method and system for evaluating reliability of online performance monitoring result of power transformer
CN114330486A (en) Power system bad data identification method based on improved Wasserstein GAN
CN107689015A (en) A kind of improved power system bad data recognition method
CN114035086A (en) Battery pack multi-fault diagnosis method based on signal processing
CN109784777B (en) Power grid equipment state evaluation method based on time sequence information fragment cloud similarity measurement
CN106655170A (en) Power system adverse data identification method
CN112803402B (en) Radiation network forward-push back substitution robust state estimation method containing bad data preprocessing
CN115795222A (en) Grid bad parameter identification and correction method based on synchronous phasor measurement

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant