CN108988347B - Method and system for adjusting class imbalance of transient voltage stabilization sample set of power grid - Google Patents
Method and system for adjusting class imbalance of transient voltage stabilization sample set of power grid Download PDFInfo
- Publication number
- CN108988347B CN108988347B CN201810866566.0A CN201810866566A CN108988347B CN 108988347 B CN108988347 B CN 108988347B CN 201810866566 A CN201810866566 A CN 201810866566A CN 108988347 B CN108988347 B CN 108988347B
- Authority
- CN
- China
- Prior art keywords
- sample set
- power grid
- historical
- prediction
- samples
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 230000001052 transient effect Effects 0.000 title claims abstract description 39
- 230000006641 stabilisation Effects 0.000 title claims description 13
- 238000011105 stabilization Methods 0.000 title claims description 13
- 238000012360 testing method Methods 0.000 claims abstract description 47
- 238000003066 decision tree Methods 0.000 claims abstract description 34
- 238000012549 training Methods 0.000 claims abstract description 32
- 238000004088 simulation Methods 0.000 claims abstract description 31
- 238000013210 evaluation model Methods 0.000 claims abstract description 12
- 238000012544 monitoring process Methods 0.000 claims abstract description 6
- 230000000368 destabilizing effect Effects 0.000 claims description 12
- 230000002194 synthesizing effect Effects 0.000 claims description 8
- 230000001687 destabilization Effects 0.000 claims description 7
- 238000007689 inspection Methods 0.000 claims description 4
- 230000033228 biological regulation Effects 0.000 claims description 3
- 230000010354 integration Effects 0.000 claims description 3
- 230000001105 regulatory effect Effects 0.000 claims description 2
- 238000011156 evaluation Methods 0.000 abstract description 6
- 238000005065 mining Methods 0.000 abstract description 2
- 238000012797 qualification Methods 0.000 abstract description 2
- 238000007418 data mining Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 102100029469 WD repeat and HMG-box DNA-binding protein 1 Human genes 0.000 description 1
- 101710097421 WD repeat and HMG-box DNA-binding protein 1 Proteins 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/12—Circuit arrangements for ac mains or ac distribution networks for adjusting voltage in ac networks by changing a characteristic of the network load
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Water Supply & Treatment (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method and a system for adjusting class unbalance of a transient voltage stability sample set of a power grid, and belongs to the field of stability analysis and evaluation of power systems. The method comprises the steps of taking a historical sample set collected from a power grid historical record as a main component of a training sample set, generating a prediction sample by utilizing time domain simulation of an expected fault set under a dispatching operation plan, and testing the qualification of the prediction sample by comparing Euclidean distances among samples of different classes; combining the prediction sample with the historical sample to form a training sample set for mining and learning, and realizing the alleviation and adjustment of the class unbalance problem of the historical sample set; and carrying out classification learning on the training sample set by using a decision tree algorithm to obtain a decision tree model, and taking the decision tree model as a power grid transient voltage stability evaluation model in the real-time monitoring process.
Description
Technical Field
The invention belongs to the field of stability analysis and evaluation of power systems, and particularly relates to a method and a system for adjusting class imbalance of a transient voltage stability sample set of a power grid.
Background
With the rapid development of technologies such as data mining and machine learning, many data mining methods are adopted in the power grid to perform online monitoring and evaluation of transient voltage stability of the power grid at present. In the traditional power grid transient voltage stability monitoring and evaluation, a training sample of data mining is mainly obtained through offline time domain simulation of a power grid, but due to inherent errors of power grid modeling and time domain simulation, the training sample generated through the time domain simulation may bring about the problem of insufficient reliability. In order to improve the reliability of the training sample data source, historical samples of the power grid under various historical faults can be collected from the historical operation records of the power grid. However, in most cases, the power grid can maintain the transient voltage stability in actual operation, and the instability cases are relatively few, which causes serious imbalance of the category of the historical sample set collected from the historical records. If the category imbalance problem is not processed, the attention degree of the grid transient voltage stability evaluation model to the instability sample in the excavation training process is too low, and therefore the recall degree of the finally obtained grid transient voltage stability evaluation model to the instability sample is too low.
The power grid dispatching operation platform has a mature dispatching operation plan arrangement method, and if the dispatching operation plan can be utilized to perform time domain simulation on various expected faults of the power grid in a short period in the future, the reliability of a prediction sample obtained by the time domain simulation can be greatly improved. If a certain number of prediction samples are properly synthesized in this way, the class imbalance problem of the historical sample set can be effectively alleviated and adjusted.
Patent document CN105139289A discloses "a power grid transient voltage stability evaluation method based on misclassification cost classification learning", which is based on dynamic measurement data of a synchronous phasor measurement unit, and extracts a key subsequence closely related to a power grid state from a time sequence formed by a large amount of dynamic measurement data; introducing a weight coefficient into a learning sample by setting different misclassification costs of stable and unstable states of a power grid; the method comprises the steps of utilizing a decision tree algorithm blended with sample weight coefficients to conduct classification learning to obtain a decision tree model, utilizing the decision tree model for online monitoring, and evaluating the transient voltage stability condition of a power grid.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a method for adjusting class unbalance of a power grid transient voltage stability sample set so as to enhance the recall capability of a power grid transient voltage stability evaluation model on a destabilization sample.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention has the beneficial effects that:
a method for adjusting class imbalance of a power grid transient voltage stabilization sample set comprises
S1, collecting historical samples from the historical operation records of the power grid to form a historical sample set Sp0；
S2, generating instability prediction samples through time domain simulation of expected faults under the scheduling operation plan, wherein all instability prediction samples form an initial prediction sample set SpFirst stage；
S3, using the historical sample set S in step S1p0For the initial prediction sample set Sp in step S2First stageAll the instability prediction samples in the test sample set are subjected to eligibility test, unqualified instability prediction samples are removed, and all the instability prediction samples qualified through test form a qualified prediction sample set Sp1；
S4, using the qualified prediction sample set S obtained in step S3p1To adjust the history sample set S in the above step S1p0Will Sp1Is combined to Sp0In the method, a training sample set is formed to overcome the class imbalance problem; (ii) a
And S5, performing classification learning on the training sample set in the step S4 by adopting a decision tree algorithm to obtain a decision tree model, and monitoring and evaluating the transient voltage stability state of the power grid in real time by taking the decision tree model as a transient voltage stability evaluation model of the power grid.
The step S1 specifically includes:
acquiring a historical fault set, a node set and a characteristic variable set of a power grid from a historical operation record of the power grid, collecting characteristic variable values of nodes in the power grid under each historical fault and an operation state Z of the power grid, recording a stable operation state of the power grid as Z-1, recording a unstable operation state of the power grid as Z-1, collecting data collected under one historical fault into a historical sample, and N is total0An integration of historical samples into a historical sample set Sp0Respectively counting the total number N of stable historical samples in the historical sample set1And total number of unstability history samples N-1In which N is1+N-1＝N0。
The step S2 specifically includes:
acquiring a node set and a characteristic variable set of a power grid in the current time period, a scheduling operation plan in N hours in the future and an expected fault set in N hours in the future from a scheduling operation platform of the power grid, and performing N on various expected faults of the power grid under the scheduling operation plan in N hours in the future by using a computer time domain simulation methodpAnd secondary time domain simulation, respectively recording each characteristic variable value of each node in the power grid after encountering a fault and the running state Z of the power grid in each time domain simulation process, wherein Z is 1 to represent that the power grid is in a stable running state, Z is-1 to represent that the power grid is in a destabilizing running state, judging Z, synthesizing data recorded in the time domain simulation process into a destabilizing prediction sample if Z is-1, and not synthesizing the destabilizing prediction sample if Z is 1, and counting all synthesized destabilizing prediction samplesNumber of stationary prediction samples Np0，Np0The unstabilized prediction samples form an initial set of prediction samples SpFirst stage。
The step S3 includes:
s31, using the historical sample set S in step S1p0And the initial prediction sample set Sp in step S2First stageSynthesizing a test sample set S1Wherein the total number of test samples is Nt＝N0+Np0；
S32, arbitrarily selecting the initial prediction sample set Sp of the step S2First stageAnd from the historical sample set S of step S1p0All of N1Randomly selecting a stable historical sample j from the stable historical samples, and calculating the Euclidean distance d (i, j) between the two samples, wherein i is more than or equal to 1 and is less than or equal to Np0，1≤j≤N1；
S33, calculating the test sample set S of the step S31 in sequence1All of N intThe Euclidean distance between each test sample and the instability prediction sample i in the step S32 is calculated to obtain all NtThe maximum of the Euclidean distance values is recorded as d (i, N)t)；
S34, calculating the test sample set S in the step S31 in sequence1All of NtAll the calculated Euclidean distances between the test samples and the stable history sample j in the step S32 are obtainedtThe maximum of the Euclidean distance values is recorded as d (j, N)t)；
S35, comparing and judging the Euclidean distance obtained in the steps S32-S34, if d (i, N)t) D (i, j) and d (j, N)t) D (i, j) or more, the instability prediction sample i is unqualified, and the step S37 is performed, if d (i, N)t) < d (i, j) and d (j, N)t) D (i, j) or d (i, N)t) D (i, j) and d (j, N)t) < d (i, j), or d (i, N)t) < d (i, j) and d (j, N)t) If d (i, j) is less than d, the instability prediction sample i is qualified in the current inspection, and the step S36 is carried out;
s36 traversing history sample set Sp0All of N in (1)1Repeating the above steps S32-S35 for each stable history sample,obtaining a test result of the instability prediction sample i;
s37, traversing the initial prediction sample set SpFirst stageAll of N in (1)p0Repeating the steps S32-S36 for each instability prediction sample to obtain all Np0The test results of the individual instability prediction samples;
s38, reading all N obtained in the step S37p0If the test result of the instability prediction sample is unqualified, the instability prediction sample is selected from the initial prediction sample set SpFirst stageRemoving, if the test result of the instability prediction sample is qualified, placing the instability prediction sample in an initial prediction sample set SpFirst stageReserving the sum of the number of the qualified instability prediction samples Np1All of Np1Forming a qualified prediction sample set S by the reserved qualified instability prediction samplesp1。
Correspondingly, the invention also provides a system for regulating the class imbalance of the transient voltage stabilization sample set of the power grid, which comprises the following steps:
a historical sample set generation module for collecting historical samples from the historical operation records of the power grid to generate a set historical sample set Sp0；
An initial prediction sample set generation module, configured to generate instability prediction samples through time-domain simulation of expected faults under the schedule operation plan, where all the instability prediction samples form an initial prediction sample set SpFirst stage；
A qualified prediction sample set generation module for generating a history sample set S by the history sample set generation modulep0For the initial prediction sample set Sp generated by the initial prediction sample set generation moduleFirst stageAll the instability prediction samples in the test sample set are subjected to eligibility test, unqualified instability prediction samples are removed, and all the instability prediction samples qualified through test form a qualified prediction sample set Sp1；
A training sample set generation module for generating a qualified prediction sample set S by using the qualified prediction sample set generation modulep1To adjust the historical sample set S generated by the initial prediction sample set generation module in the above stepsp0Will beSp1Is combined to Sp0Forming a training sample set;
and the decision tree model generation module is used for classifying and learning the training sample set generated by the training sample set generation module by adopting a decision tree algorithm to obtain a decision tree model, and the decision tree model is used as a transient voltage stability evaluation model of the power grid to monitor and evaluate the transient voltage stability state of the power grid in real time.
Compared with the prior art, the invention has the following beneficial effects:
the invention provides a method for adjusting class unbalance of a transient voltage stabilization sample set of a power grid, which has the advantages that compared with the traditional method for obtaining a training sample completely based on time domain simulation, the training sample in the method is obtained on the basis of the actual operation mode of the power grid, the reliability of the training sample set is ensured from a data source, a destabilization prediction sample is generated by time domain simulation of expected faults under a power grid dispatching operation plan aiming at the class unbalance problem of a history sample obtained initially, the qualification of the destabilization prediction sample is checked according to the Euclidean distance relation between the destabilization prediction sample and a stable history sample, the quality of the destabilization prediction sample is further ensured, the destabilization prediction sample set and the history sample set are combined into the training sample set for mining and learning, and while the class unbalance of the history sample set is effectively adjusted, the bias of the classification learning process on the instability samples can be improved, and the recall capability of the power grid transient voltage stability evaluation model on the instability samples is enhanced.
Drawings
FIG. 1 is a schematic diagram of a single-line structure of a power grid according to an embodiment of the present invention;
FIG. 2 is a block flow diagram of an embodiment of the method of the present invention;
FIG. 3 is a decision tree model obtained by performing classification learning on a training sample set in the method of the present invention.
Detailed Description
The invention will be further described with reference to the accompanying drawings and the detailed description below:
a schematic diagram of a single-line structure of a power grid related in the evaluation method of the present invention is shown in fig. 1, a middle power grid shown in fig. 1 is an application embodiment of the present invention, the power grid is a heavy-load receiving end area in fig. 1, the power grid totally includes 13 nodes, the total load amount is 6190 megawatts, and a flow of the method of the present embodiment is shown in fig. 2, and the method includes the following steps:
(1) acquiring a historical fault set, a node set and a characteristic variable set of a power grid from a historical operation record of the power grid, collecting characteristic variable values of nodes in the power grid under each historical fault and an operation state Z of the power grid, recording a stable operation state of the power grid as Z-1, recording a unstable operation state of the power grid as Z-1, collecting data collected under one historical fault into a historical sample, and N is total0(720 samples in this embodiment) are integrated into a history sample set Sp0Respectively counting the total number N of stable historical samples in the historical sample set1(665 samples in this embodiment) and the total number of unstability history samples N-1(55 in this embodiment), where N1+N-1＝N0；
(2) Acquiring a node set and a characteristic variable set of a power grid in the current time period, a scheduling operation plan in N hours in the future (72 hours in an embodiment) and an expected fault set in N hours in the future from a scheduling operation platform of the power grid, and performing N-time simulation on various expected faults of the power grid under the scheduling operation plan in N hours in the future by using a computer time domain simulation methodpAnd secondary time domain simulation, respectively recording each characteristic variable value of each node in the power grid after encountering a fault and the running state Z of the power grid in each time domain simulation process, wherein Z is 1 to represent that the power grid is in a stable running state, Z is-1 to represent that the power grid is in a destabilizing running state, judging Z, synthesizing data recorded in the time domain simulation process into a destabilizing prediction sample if Z is-1, not synthesizing the destabilizing prediction sample if Z is 1, and counting the number N of all synthesized destabilizing prediction samplesp0(140 in this embodiment), Np0The unstabilized prediction samples form an initial set of prediction samples SpFirst stage；
(3) Utilizing the historical sample set S in step (1)p0For the initial prediction sample set Sp in step (2)First stageAll of N inp0The eligibility of each instability prediction sample is checked, unqualified samples are removed, and all the instability prediction samples qualified through checking form a qualified prediction sample set Sp1The specific process is as follows:
(3-1) Using the historical sample set S in step (1)p0And the initial prediction sample set Sp in step (2)First stageSynthesizing a test sample set S1Wherein the total number of test samples is Nt＝N0+Np0(860);
(3-2) arbitrarily selecting the initial prediction sample set Sp in the step (2)First stageAnd (2) from the historical sample set S of step (1)p0All of N1Randomly selecting a stable historical sample j from the stable historical samples, and calculating the Euclidean distance d (i, j) between the two samples, wherein i is more than or equal to 1 and is less than or equal to Np0，1≤j≤N1；
(3-3) sequentially calculating the test sample set S of the step (3-1)1All of N intThe Euclidean distance between each test sample and the instability prediction sample i in the step (3-2) is calculated to obtain all NtThe maximum of the Euclidean distance values is recorded as d (i, N)t)；
(3-4) sequentially calculating the test sample set S in the step (3-1)1All of NtThe Euclidean distance between each test sample and the stable history sample j in the step (3-2) is calculated to obtain all NtThe maximum of the Euclidean distance values is recorded as d (j, N)t)；
(3-5) comparing and judging the Euclidean distances obtained in the steps (3-2) to (3-4) if d (i, N)t) D (i, j) and d (j, N)t) D (i, j) is more than or equal to d (i, j), the instability prediction sample i is unqualified, and the step (3-7) is carried out, if d (i, N)t) < d (i, j) and d (j, N)t) D (i, j) or d (i, N)t) D (i, j) and d (j, N)t) < d (i, j), or d (i, N)t) < d (i, j) and d (j, N)t) If d (i, j) is less than d, the instability prediction sample i is qualified in the current inspection, and the step (3-6) is carried out;
(3-6) traversing the historical sample set Sp0All of N in (1)1A stable historical sample, repeatObtaining the test result of the instability prediction sample i in the steps (3-2) - (3-5);
(3-7) traversing the initial prediction sample set SpFirst stageAll of N in (1)p0Repeating the steps (3-2) to (3-6) for each instability prediction sample to obtain all Np0The test results of the individual instability prediction samples;
(3-8) reading all N obtained in the step (3-7)p0If the test result of the instability prediction sample is unqualified, the instability prediction sample is selected from the initial prediction sample set SpFirst stageRemoving, if the test result of the instability prediction sample is qualified, placing the instability prediction sample in an initial prediction sample set SpFirst stageReserving the sum of the number of the qualified instability prediction samples Np1(127 in this example), all Np1Forming a qualified prediction sample set S by the reserved qualified instability prediction samplesp1；
(4) Utilizing the qualified prediction sample set S obtained in the step (3)p1To adjust the historical sample set S in the step (1)p0Will Sp1Is combined to Sp0In the method, a training sample set is formed to overcome the problem of class unbalance, statistics show that the proportions of instability samples in a history sample set and the training sample set are 7.63% and 21.49% respectively, so that the problem of class unbalance of the history sample set is greatly relieved after instability prediction samples are blended;
(5) performing classification learning on the training sample set in the step (4) by using a decision tree algorithm to obtain a decision tree model, where the decision tree model is shown in fig. 2, reference numerals 1 and-1 of terminal nodes in the decision tree model shown in fig. 2 represent output power grid states Z, where Z ═ 1 represents a power grid stable state, Z ═ 1 represents a power grid instability state, an internal node U _ k in the decision tree model represents a voltage characteristic variable of a node k, P _ k represents an active power characteristic variable of the node k, Q _ k represents a reactive power characteristic variable of the node k, and a classification accuracy P of the decision tree model is determined by using a cross validation methodreAnd degree of recall RecTest was carried out, Pre＝96.7％，Rec96.2%, is further oneStep comparison and verification, namely directly taking the historical sample set in the step (1) as a training sample set for classification learning, and adopting a cross verification mode to obtain the classification accuracy P of the decision tree model obtained by classification learningreAnd degree of recall RecTest was carried out, Pre＝92.2％，Rec85.5%, so that the instability prediction sample has a good regulation effect on the class unbalance problem of the historical sample set, the integral classification accuracy is improved, the recall capability of the decision tree model on the instability sample is obviously enhanced, and the decision tree model shown in fig. 2 is used as the transient voltage stability evaluation model of the power grid to monitor and evaluate the transient voltage stability state of the power grid in real time.
Therefore, the method collects historical samples from the historical operation records of the power grid, generates the instability prediction samples through time domain simulation of expected faults under the scheduling operation plan, fuses the instability prediction samples qualified through inspection with the historical samples, adjusts the class unbalance of the historical sample set, and derives a more reliable power grid transient voltage stability evaluation model through learning the sample set with the class unbalance problem relieved.
Correspondingly, this embodiment also provides a system for adjusting grid transient voltage stabilization sample set class imbalance, including:
a historical sample set generation module for collecting historical samples from the historical operation records of the power grid to generate a set historical sample set Sp0；
An initial prediction sample set generation module, configured to generate instability prediction samples through time-domain simulation of expected faults under the schedule operation plan, where all the instability prediction samples form an initial prediction sample set SpFirst stage；
A qualified prediction sample set generation module for generating a history sample set S by the history sample set generation modulep0For the initial prediction sample set Sp generated by the initial prediction sample set generation moduleFirst stageAll the instability prediction samples in the test are subjected to eligibility test, unqualified instability prediction samples are removed, and all the instability prediction samples qualified through test form qualified prediction samplesCollection Sp1；
A training sample set generation module for generating a qualified prediction sample set S by using the qualified prediction sample set generation modulep1To adjust the historical sample set S generated by the initial prediction sample set generation module in the above stepsp0Will Sp1Is combined to Sp0Forming a training sample set;
and the decision tree model generation module is used for classifying and learning the training sample set generated by the training sample set generation module by adopting a decision tree algorithm to obtain a decision tree model, and the decision tree model is used as a transient voltage stability evaluation model of the power grid to monitor and evaluate the transient voltage stability state of the power grid in real time.
Since the working principle of each module is the same as the flow principle of the method, it is not described in detail in this embodiment.
Various other modifications and changes may be made by those skilled in the art based on the above-described technical solutions and concepts, and all such modifications and changes should fall within the scope of the claims of the present invention.
Claims (7)
1. A method for adjusting class imbalance of a power grid transient voltage stabilization sample set is characterized by comprising the following steps
S1, collecting historical samples from the historical operation records of the power grid to form a historical sample set Sp0；
S2, generating instability prediction samples through time domain simulation of expected faults under the scheduling operation plan, wherein all instability prediction samples form an initial prediction sample set SpFirst stage；
S3, using the historical sample set S in step S1p0For the initial prediction sample set Sp in step S2First stageAll the instability prediction samples in the test sample set are subjected to eligibility test, unqualified instability prediction samples are removed, and all the instability prediction samples qualified through test form a qualified prediction sample set Sp1；
S4, using the qualified prediction sample set S obtained in step S3p1To adjust the aboveHistorical sample set S in step S1p0Will Sp1Is combined to Sp0Forming a training sample set;
and S5, performing classification learning on the training sample set in the step S4 by adopting a decision tree algorithm to obtain a decision tree model, and monitoring and evaluating the transient voltage stability state of the power grid in real time by taking the decision tree model as a transient voltage stability evaluation model of the power grid.
2. The method for adjusting class imbalance of the grid transient voltage stabilization sample set according to claim 1, wherein the step S1 specifically includes:
acquiring a historical fault set, a node set and a characteristic variable set of a power grid from a historical operation record of the power grid, collecting characteristic variable values of nodes in the power grid under each historical fault and an operation state Z of the power grid, recording a stable operation state of the power grid as Z-1, recording a unstable operation state of the power grid as Z-1, collecting data collected under one historical fault into a historical sample, and N is total0An integration of historical samples into a historical sample set Sp0Respectively counting the total number N of stable historical samples in the historical sample set1And total number of unstability history samples N-1In which N is1+N-1＝N0。
3. The method for adjusting class imbalance of the grid transient voltage stabilization sample set according to claim 2, wherein the step S2 specifically includes:
acquiring a node set and a characteristic variable set of a power grid in the current time period, a scheduling operation plan in N hours in the future and an expected fault set in N hours in the future from a scheduling operation platform of the power grid, and performing N on various expected faults of the power grid under the scheduling operation plan in N hours in the future by using a computer time domain simulation methodpAnd secondary time domain simulation, respectively recording each characteristic variable value of each node in the power grid after encountering a fault and the operation state Z of the power grid in each time domain simulation process, wherein Z-1 represents that the power grid is in a stable operation state, Z-1 represents that the power grid is in a destabilization operation state, judging Z, and if Z-plus1, synthesizing data recorded in the time domain simulation process into one instability prediction sample, if Z is 1, not synthesizing the instability prediction samples, and counting the number N of all synthesized instability prediction samplesp0，Np0The unstabilized prediction samples form an initial set of prediction samples SpFirst stage。
4. The method for adjusting class imbalance of grid transient voltage stabilization sample sets according to claim 3, wherein the step S3 includes:
s31, using the historical sample set S in step S1p0And the initial prediction sample set Sp in step S2First stageSynthesizing a test sample set S1Wherein the total number of test samples is Nt＝N0+Np0；
S32, arbitrarily selecting the initial prediction sample set Sp of the step S2First stageAnd from the historical sample set S of step S1p0All of N1Randomly selecting a stable historical sample j from the stable historical samples, and calculating the Euclidean distance d (i, j) between the two samples, wherein i is more than or equal to 1 and is less than or equal to Np0，1≤j≤N1；
S33, calculating the test sample set S of the step S31 in sequence1All of N intAll the calculated N are the euclidean distances between the respective test samples and the instability prediction sample i in step S32tThe maximum of the Euclidean distance values is recorded as d (i, N)t)；
S34, calculating the test sample set S in the step S31 in sequence1All of NtAll the calculated Euclidean distances between the test samples and the stable history sample j in the step S32 are obtainedtThe maximum of the Euclidean distance values is recorded as d (j, N)t)；
S35, comparing and judging the Euclidean distance obtained in the steps S32-S34, if d (i, N)t) D (i, j) and d (j, N)t) D (i, j) or more, the instability prediction sample i is unqualified, and the step S37 is performed, if d (i, N)t) < d (i, j) and d (j, N)t) D (i, j) or d (i, N)t) D (i, j) and d (j, N)t) < d (i, j), ord(i,Nt) < d (i, j) and d (j, N)t) If d (i, j) is less than d, the instability prediction sample i is qualified in the current inspection, and the step S36 is carried out;
s36 traversing history sample set Sp0All of N in (1)1Repeating the steps S32-S35 for each stable historical sample to obtain the test result of the instability prediction sample i;
s37, traversing the initial prediction sample set SpFirst stageAll of N in (1)p0Repeating the steps S32-S36 for each instability prediction sample to obtain all Np0The test results of the individual instability prediction samples;
s38, reading all N obtained in the step S37p0If the test result of the instability prediction sample is unqualified, the instability prediction sample is selected from the initial prediction sample set SpFirst stageRemoving, if the test result of the instability prediction sample is qualified, placing the instability prediction sample in an initial prediction sample set SpFirst stageReserving the sum of the number of the qualified instability prediction samples Np1All of Np1Forming a qualified prediction sample set S by the reserved qualified instability prediction samplesp1。
5. A system for regulating class imbalance of a power grid transient voltage stabilization sample set is characterized by comprising:
a historical sample set generation module for collecting historical samples from the historical operation records of the power grid to generate a set historical sample set Sp0；
An initial prediction sample set generation module, configured to generate instability prediction samples through time-domain simulation of expected faults under the schedule operation plan, where all the instability prediction samples form an initial prediction sample set SpFirst stage；
A qualified prediction sample set generation module for generating a history sample set S by the history sample set generation modulep0For the initial prediction sample set Sp generated by the initial prediction sample set generation moduleFirst stageAll the instability prediction samples in the test are subjected to eligibility test, unqualified instability prediction samples are removed, and all the qualified instability prediction samples are testedStable prediction sample forming qualified prediction sample set Sp1；
A training sample set generation module for generating a qualified prediction sample set S by using the qualified prediction sample set generation modulep1To adjust the history sample set S generated by the history sample set generation modulep0Will Sp1Is combined to Sp0Forming a training sample set;
and the decision tree model generation module is used for classifying and learning the training sample set generated by the training sample set generation module by adopting a decision tree algorithm to obtain a decision tree model, and the decision tree model is used as a transient voltage stability evaluation model of the power grid to monitor and evaluate the transient voltage stability state of the power grid in real time.
6. The system for regulation of grid transient voltage stabilization sample set class imbalance of claim 5, wherein the historical sample set generation module generates a set historical sample set Sp0The specific process comprises the following steps:
acquiring a historical fault set, a node set and a characteristic variable set of a power grid from a historical operation record of the power grid, collecting characteristic variable values of nodes in the power grid under each historical fault and an operation state Z of the power grid, recording a stable operation state of the power grid as Z-1, recording a unstable operation state of the power grid as Z-1, collecting data collected under one historical fault into a historical sample, and N is total0An integration of historical samples into a historical sample set Sp0Respectively counting the total number N of stable historical samples in the historical sample set1And total number of unstability history samples N-1In which N is1+N-1＝N0。
7. A system for regulation of class imbalance in power grid Transient Voltage Stabilization (TVS) sample sets as claimed in claim 5 or 6, wherein said initial prediction sample set generation module forms an initial prediction sample set SpFirst stageThe specific process comprises the following steps:
acquiring a node set and a characteristic variable set of the power grid in the current time period from a dispatching operation platform of the power grid, wherein the node set and the characteristic variable set are within n hours in the futureThe scheduling operation plan and the expected fault set in the future N hours are carried out on various expected faults of the power grid in the scheduling operation plan in the future N hours by utilizing a computer time domain simulation methodpAnd secondary time domain simulation, respectively recording each characteristic variable value of each node in the power grid after encountering a fault and the running state Z of the power grid in each time domain simulation process, wherein Z is 1 to represent that the power grid is in a stable running state, Z is-1 to represent that the power grid is in a destabilizing running state, judging Z, synthesizing data recorded in the time domain simulation process into a destabilizing prediction sample if Z is-1, not synthesizing the destabilizing prediction sample if Z is 1, and counting the number N of all synthesized destabilizing prediction samplesp0，Np0The unstabilized prediction samples form an initial set of prediction samples SpFirst stage。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810866566.0A CN108988347B (en) | 2018-08-01 | 2018-08-01 | Method and system for adjusting class imbalance of transient voltage stabilization sample set of power grid |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810866566.0A CN108988347B (en) | 2018-08-01 | 2018-08-01 | Method and system for adjusting class imbalance of transient voltage stabilization sample set of power grid |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108988347A CN108988347A (en) | 2018-12-11 |
| CN108988347B true CN108988347B (en) | 2020-09-22 |
Family
ID=64554847
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810866566.0A Active CN108988347B (en) | 2018-08-01 | 2018-08-01 | Method and system for adjusting class imbalance of transient voltage stabilization sample set of power grid |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108988347B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110336270B (en) * | 2019-04-22 | 2021-02-02 | 清华大学 | Updating method of transient stability prediction model of power system |
| CN111244937B (en) * | 2020-01-09 | 2021-01-05 | 清华大学 | Method for screening serious faults of transient voltage stability of power system |
| CN111628501B (en) * | 2020-06-18 | 2023-04-18 | 国网山东省电力公司济南供电公司 | AC/DC large power grid transient voltage stability assessment method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102074955A (en) * | 2011-01-20 | 2011-05-25 | 中国电力科学研究院 | Method based on knowledge discovery technology for stability assessment and control of electric system |
| CN104617574A (en) * | 2015-01-19 | 2015-05-13 | 清华大学 | Assessment method for transient voltage stabilization of load area of electrical power system |
| CN105139289A (en) * | 2015-09-06 | 2015-12-09 | 清华大学 | Power system transient state voltage stability evaluating method based on misclassification cost classified-learning |
| CN106849069A (en) * | 2017-03-13 | 2017-06-13 | 东北电力大学 | A kind of transient stability evaluation in power system method based on Pin SVM |
| JP6174271B2 (en) * | 2014-10-31 | 2017-08-02 | 株式会社日立製作所 | System stabilization control apparatus and method |
| CN107482621A (en) * | 2017-08-02 | 2017-12-15 | 清华大学 | A kind of Transient Voltage Stability in Electric Power System appraisal procedure based on voltage sequential track |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6463772B2 (en) * | 2013-10-22 | 2019-02-06 | ビット・トール・エルエルシー | Method and apparatus for detecting and correcting instability within a distribution system |
-
2018
- 2018-08-01 CN CN201810866566.0A patent/CN108988347B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102074955A (en) * | 2011-01-20 | 2011-05-25 | 中国电力科学研究院 | Method based on knowledge discovery technology for stability assessment and control of electric system |
| JP6174271B2 (en) * | 2014-10-31 | 2017-08-02 | 株式会社日立製作所 | System stabilization control apparatus and method |
| CN104617574A (en) * | 2015-01-19 | 2015-05-13 | 清华大学 | Assessment method for transient voltage stabilization of load area of electrical power system |
| CN105139289A (en) * | 2015-09-06 | 2015-12-09 | 清华大学 | Power system transient state voltage stability evaluating method based on misclassification cost classified-learning |
| CN106849069A (en) * | 2017-03-13 | 2017-06-13 | 东北电力大学 | A kind of transient stability evaluation in power system method based on Pin SVM |
| CN107482621A (en) * | 2017-08-02 | 2017-12-15 | 清华大学 | A kind of Transient Voltage Stability in Electric Power System appraisal procedure based on voltage sequential track |
Non-Patent Citations (4)
| Title |
|---|
| Imbalance Learning Machine-Based Power System Short-Term Voltage Stability Assessment;LiPeng Zhu，Chao Lu et al.;《 IEEE Transactions on Industrial Informatics》;20170420;全文 * |
| Stability Analysis and Controller Design of a Wide-Area Time-Delay System Based on the Expectation Model Method;Xinran Zhang，Chao Lu et al.;《IEEE Transactions on Smart Grid》;20151109;全文 * |
| 基于广域时序数据挖掘策略的暂态电压稳定评估;朱利鹏，陆超等;《电网技术》;20160131;第40卷(第1期);全文 * |
| 基于数据挖掘的区域暂态电压稳定评估;朱利鹏，陆超等;《电网技术》;20150430;第39卷(第4期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108988347A (en) | 2018-12-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108988347B (en) | Method and system for adjusting class imbalance of transient voltage stabilization sample set of power grid | |
| Wang et al. | Decision tree based online stability assessment scheme for power systems with renewable generations | |
| CN104915747A (en) | Electricity generation performance evaluation method of generator set and equipment thereof | |
| CN111401749A (en) | Dynamic safety assessment method based on random forest and extreme learning regression | |
| CN103825269B (en) | Rapid probabilistic load flow calculation method considering static power frequency characteristics of electric power system | |
| CN106570790B (en) | Wind power plant output data restoration method considering wind speed data segmentation characteristics | |
| CN113675878B (en) | Offshore wind farm grid-connected performance testing method and system, computer equipment and medium | |
| CN110570122A (en) | Offshore wind power plant reliability assessment method considering wind speed seasonal characteristics and current collection system element faults | |
| CN114897331A (en) | Power transformer risk assessment method based on three-parameter interval gray number decision | |
| CN110738253A (en) | short-term wind power prediction method based on FCM and AFSA-Elman | |
| CN111950131B (en) | Wind power plant output equivalent aggregation model construction method considering electricity limiting factors | |
| CN110489852A (en) | Improve the method and device of the wind power system quality of data | |
| CN113300373A (en) | Stability margin value prediction method and device based on PRMSE evaluation index | |
| CN116992362A (en) | Transformer fault characterization feature quantity screening method and device based on Xia Puli value | |
| CN105701265A (en) | Double-fed wind generator modeling method and apparatus | |
| CN108536877A (en) | The appraisal procedure and system of vibration of wind generating set degree | |
| CN116307838A (en) | Accurate equivalence method, system and medium for power distribution network in power grid reliability evaluation | |
| Alzubaidi et al. | Probabilistic voltage stability analysis considering variable wind generation and different control modes | |
| CN110943485B (en) | Index evaluation method for simulation reliability of equivalent model of doubly-fed wind power plant | |
| CN113067352B (en) | Power grid critical removal time prediction method based on EEAC and ELM | |
| CN112910006B (en) | Universal electromagnetic transient modeling method for direct-drive wind turbine generator | |
| CN113836125A (en) | Distribution transformer data cleaning and repairing method and system for distribution internet of things edge side | |
| CN112347655A (en) | Wind power plant theoretical power calculation method based on unit operation performance evaluation | |
| Cheng et al. | Sequential short-term reliability evaluation considering repair time distribution | |
| Ripp et al. | A first shot at time-dependent CO 2 intensities in multi-modal energy systems |
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 |