CN111884236A - Intelligent transient stability evaluation system for power grid - Google Patents

Intelligent transient stability evaluation system for power grid Download PDF

Info

Publication number
CN111884236A
CN111884236A CN202010789805.4A CN202010789805A CN111884236A CN 111884236 A CN111884236 A CN 111884236A CN 202010789805 A CN202010789805 A CN 202010789805A CN 111884236 A CN111884236 A CN 111884236A
Authority
CN
China
Prior art keywords
stability
generator set
generator
key generator
output
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202010789805.4A
Other languages
Chinese (zh)
Inventor
姚海成
黄济宇
管霖
苏寅生
徐光虎
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.)
China Southern Power Grid Co Ltd
Original Assignee
China Southern Power Grid 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 China Southern Power Grid Co Ltd filed Critical China Southern Power Grid Co Ltd
Priority to CN202010789805.4A priority Critical patent/CN111884236A/en
Publication of CN111884236A publication Critical patent/CN111884236A/en
Pending legal-status Critical Current

Links

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
    • H02J3/24—Arrangements for preventing or reducing oscillations of power in networks
    • H02J3/242—Arrangements for preventing or reducing oscillations of power in networks using phasor measuring units [PMU]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • 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
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Power Engineering (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses an intelligent transient stability evaluation system for a power grid, which mainly comprises a phasor measurement unit, a preprocessing unit, a stability classification module, a key generator module, a stability classification simple logic processing unit, a key generator simple logic processing unit and an auxiliary decision module. The evaluation system provided by the invention takes the same PMU sampling as input, respectively takes transient stability classification and key generator identification as output, establishes a dual-task stability evaluation model, performs simple logic processing on the stability classification and key generator identification given by the dual-task stability evaluation model, and then is used as the input of an auxiliary decision system, adopts dual-task interactive verification and discrimination logic to realize mutual check of prediction results, obtains a final stability discrimination result and a key generator set, and simultaneously screens uncertain samples which are difficult to discriminate, thereby obviously improving the prediction accuracy and the result reliability of intelligent transient stability evaluation and being beneficial to safe operation and risk regulation of a power grid.

Description

Intelligent transient stability evaluation system for power grid
Technical Field
The invention relates to the technical field of electric power, in particular to an intelligent transient stability evaluation system for a power grid.
Background
The intelligent transient stability evaluation algorithm applying the machine learning technology does not need to depend on a power grid model and simulation calculation, and a data driving method is adopted to learn the mapping relation between the power grid characteristics and the stability result in an off-line mode, so that rapid on-line judgment can be realized. However, due to the fact that complex dynamic characteristics and strong nonlinearity of transient stability are difficult to accurately capture, the output information of intelligent transient stability assessment is generally single, and the intelligent transient stability assessment is mostly stable/unstable binary classification or stable indexes such as critical clearing time. When the operation mode or fault with a larger difference with the training sample set is faced, the reliability of the prediction result of the data-driven stability evaluation model can be obviously reduced, but the reliability of the given result can not be detected by self, so that the application of the stability evaluation model in the actual engineering is limited, and the safety operation and risk regulation of a power grid can not be facilitated.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides an intelligent transient stability evaluation system for a power grid, so that the prediction accuracy of the evaluation system is improved, and the safe operation and risk regulation of the power grid are facilitated.
In order to achieve the purpose, the technical scheme of the invention is as follows:
a power grid intelligent transient stability assessment system comprises:
the phasor measurement unit is used for measuring and collecting the disturbed power grid state quantity at a set frequency;
the preprocessing unit is used for preprocessing the power grid state quantity acquired by the phasor measurement unit;
the stable classification module is used for taking the output of the preprocessing unit as the input and outputting a stable classification vector;
the key generator module is used for taking the output of the preprocessing unit as the input and outputting a generator stability degree vector;
the stable classification simple logic processing unit is used for carrying out logic judgment processing on the stable classification vector and outputting a stable classification conversion signal;
the key generator simple logic processing unit is used for carrying out logic judgment processing on the generator stability degree vector and outputting a generator stability degree conversion signal;
and the auxiliary decision module is used for taking the stable classification conversion signal and the generator stability conversion signal as input and outputting a final stable state judgment result and a key generator set.
Further, the grid state quantity comprises the voltage amplitude and the phase angle of each bus, and for the grid with N nodes, the PMU data matrix with C characteristic quantities is arranged as follows for the t-th sampling time step:
Figure BDA0002623356310000021
wherein, Fi＝[x1i,x2i,…,xNi]TRepresenting the ith feature column vector.
Further, the preprocessing unit normalizes each characteristic of the power grid state quantity acquired by the phasor measurement unit by adopting zscore:
Figure BDA0002623356310000022
wherein, mu and sigma are respectively the mean value and variance corresponding to the feature vector, and F' is the normalized feature column vector.
Further, the stability classification module adopts a deep learning network model, and if the input of the corresponding disturbance causes system instability, the output of the module is [1,0 ]]T(ii) a Otherwise, if the system is stable, the output is [0,1 ]]T；(ii) a The deep learning network model is formed by off-line training of samples.
Furthermore, the key power generation module adopts a deep learning network model, and the total number of the generators in the power grid is recorded as NGThe output of the key generator identification module is a generator serial number arranged according to the length NGColumn vector of-1
Figure BDA0002623356310000023
Each element in c takes the value of [0, 1%]In between, namely have
Figure BDA0002623356310000024
The deep learning network model is formed by off-line training of samples.
Further, the stable classification simple logic processing unit is configured to perform logic judgment processing on the stable classification vector, and outputting a stable classification conversion signal includes:
in online operation, each element in the output vector of the stable classification module is no longer 0/1 two values, but [0,1 ]]Between the successive values, noting the stable classification vector of its output
Figure BDA0002623356310000025
Is provided with
Figure BDA0002623356310000026
If it is
Figure BDA0002623356310000027
Output signal S1(ii) a Otherwise, the signal S is output2；uIs a threshold value.
Further, the simple logic processing unit of the key generator is configured to perform logic judgment processing on the generator stability vector, and outputting a generator stability conversion signal includes:
in online operation, the vector output by the key generator prediction module is N in lengthG-1 stability vector
Figure BDA0002623356310000028
Wherein each element also takes on the value of [0,1]Between, there are
Figure BDA0002623356310000029
Two threshold values are defined for each of the two,agiven 0 < >aLess than 1; the method is mainly used for screening key generators; whileaThen the method is used for screening important generators;
if it is
Figure BDA00026233563100000210
All the elements in the composition satisfy
Figure BDA00026233563100000211
Then the signal S is output5(ii) a Otherwise, if all elements are satisfied
Figure BDA00026233563100000212
And at least one unit with insufficient stability
Figure BDA00026233563100000213
Then the signal S is output4(ii) a Otherwise, the signal S is output3.
Further, the input signal of the assistant decision module is one of the following six combinations:
(S1,S3)，(S1,S4)，(S1,S5)，(S2,S3)，(S2,S4)，(S2,S5)。
further, the assistant decision module outputs a final stable state judgment result which is one of three judgments of instability, uncertainty and stability; the key generator set comprises two categories of generator subsets which are respectively key generator sets
Figure BDA0002623356310000031
And/or a collection of vital generators
Figure BDA0002623356310000032
Their elements are the generator numbers belonging to the set, which can be an empty set.
Further, the logic of the final steady state decision result and the key generator set output by the auxiliary decision module is as follows:
1) when the input is a signal (S)2,S3) The assistant decision system outputs a destabilizing signal and simultaneously outputs
Figure BDA0002623356310000033
The generator serial number i is recorded into a key generator set
Figure BDA0002623356310000034
In will satisfy
Figure BDA0002623356310000035
The serial number of the generator is recorded in the important generator set
Figure BDA0002623356310000036
Performing the following steps; and outputting the key generator set and the important generator set after the scanning is finished. In the input category, the important generator set is an empty set;
2) when the input is a signal (S)1,S3) When the system is in use, the assistant decision-making system outputs 'uncertain' signals, and simultaneously scans the stability degree of the generator one by one, and when the stability degree is in use
Figure BDA0002623356310000037
The sequence numbers are recorded in the key generator set,
Figure BDA0002623356310000038
the time is recorded in the important generator set, the key generator set and the important generator set are output after the scanning is finished, and if the key generator set and the important generator set do not exist, the key generator set and the important generator set are output
Figure BDA0002623356310000039
The set of important generators is an empty set;
3) when the input is a signal (S)2,S4) When the system is in use, the assistant decision-making system outputs 'uncertain' signals, and simultaneously scans the stability degree of the generator one by one, and when the stability degree is in use
Figure BDA00026233563100000310
The sequence numbers are recorded in the key generator set,
Figure BDA00026233563100000311
the time is recorded in the important generator set, and the key generator set and the important generator set are output after the scanning is finished. At the moment, the key generator set is an empty set;
4) when the input is a signal (S)2,S5) When the system is in use, the assistant decision system outputs a 'stable' signal; while scanning the generator stability one by one, while
Figure BDA00026233563100000312
The sequence numbers are recorded in the key generator set,
Figure BDA00026233563100000313
the time is recorded in the important generator set, and the key generator set and the important generator set are output after the scanning is finished; at the moment, the two sets are both empty sets;
5) when the input is a signal (S)1,S4) Or (S)1,S5) The decision-making aid system outputs a "stable" signal. While scanning the generator stability one by one, while
Figure BDA00026233563100000314
The sequence numbers are recorded in the key generator set,
Figure BDA00026233563100000315
the time is recorded in the important generator set, and the key generator set and the important generator set are output after the scanning is finished; at the moment, the key generator set is an empty set, and if the key generator set does not exist
Figure BDA00026233563100000316
The vital generator set is also empty.
Compared with the prior art, the invention has the beneficial effects that:
the intelligent transient stability evaluation system of the power grid provided by the invention takes the same PMU sampling as input, respectively takes transient stability classification and key generator identification as output, establishes a dual-task stability evaluation model, performs simple logic processing on the stability classification and key generator identification given by the dual-task stability evaluation model, and then is used as the input of an auxiliary decision-making system, adopts dual-task interactive verification and discrimination logic to realize mutual check of prediction results, obtains a final stability discrimination result and a key generator set, and simultaneously screens uncertain samples which are difficult to discriminate, thereby obviously improving the prediction accuracy and the result reliability of the intelligent transient stability evaluation and being beneficial to safe operation and risk regulation of the power grid.
Drawings
Fig. 1 is a schematic composition diagram of an intelligent transient stability evaluation system of a power grid according to an embodiment of the present invention;
FIG. 2 is a decision logic diagram of an auxiliary decision module;
in the figure: 1. a phasor measurement unit; 2. a pre-processing unit; 3. a stable classification module; 4. a key generator module; 5. a stable classification simple logic processing unit; 6. a key generator simple logic processing unit; 7. and an assistant decision module.
Detailed Description
Example (b):
the technical solution of the present invention is further described below with reference to the accompanying drawings and examples.
Referring to fig. 1, the system for evaluating the intelligent transient stability of the power grid provided in the present embodiment mainly includes a phasor measurement unit 1, a preprocessing unit 2, a stable classification module 3, a key generator module 4, a stable classification simple logic processing unit 5, a key generator simple logic processing unit 6, and an auxiliary decision module 7.
After an actual power grid fails, the phasor measurement unit 1 (PMU) measures and collects the disturbed power grid state quantity at a certain frequency; the preprocessing unit 2 is used for preprocessing the power grid state quantity acquired by the phasor measurement unit; the stable classification module 3 is used for taking the data preprocessed by the preprocessing unit as input and outputting a stable classification vector; the key generator module 4 is used for outputting a generator stability degree vector by taking data preprocessed by the preprocessing unit as input; the stable classification simple logic processing unit 5 is used for performing logic judgment processing on the stable classification vector and outputting a stable classification conversion signal; the key generator simple logic processing unit 6 is used for performing logic judgment processing on the generator stability degree vector and outputting a generator stability degree conversion signal; the auxiliary decision module 7 is used for taking the stable classification conversion signal and the generator stability conversion signal as input and outputting a final stable state judgment result and a key generator set.
Therefore, the power grid intelligent transient stability evaluation system provided by the embodiment takes the same PMU sampling as input, respectively takes the transient stability classification and the key generator identification as output, establishes the dual-task stability evaluation model, performs simple logic processing on the stability classification and the key generator identification given by the dual-task stability evaluation model, and then takes the processed stability classification and the key generator identification as input of the auxiliary decision making system, adopts the dual-task interactive verification and discrimination logic to realize mutual check of prediction results, obtains a final stability discrimination result and a key generator set, and simultaneously screens uncertain samples which are difficult to discriminate, so that the prediction accuracy and the result reliability of the intelligent transient stability evaluation can be obviously improved, and the power grid safe operation and risk regulation are facilitated.
Specifically, the grid state quantity includes the amplitude and phase angle of each bus voltage, and the like. Generally, for a grid with N nodes, for the t-th sampling time step, PMU data matrices with C kinds of characteristic quantities are arranged as follows:
Figure BDA0002623356310000051
wherein, Fi＝[x1i,x2i,…,xNi]TRepresenting the ith feature column vector. The collected data are sent to a preprocessing unit, normalized by zscore for each feature:
Figure BDA0002623356310000052
wherein, mu and sigma are respectively the mean value and variance corresponding to the feature vector, and F' is the normalized feature column vector.
The above-mentioned stable classification and key generator identification modules can be implemented by any data-driven artificial intelligence model, and in this embodiment, a deep learning network model is adopted.
The output of the stable classification module is a bivalued vector. If the input is unstable under the corresponding disturbance, the output of the module is [1,0 ]]T(ii) a Otherwise, if the system is stable, the output is [0,1 ]]T. The model is trained offline using a large number of samples. And setting an output label of the training sample according to the time domain stable simulation curve. Specifically, if the absolute value of the maximum value of the power angle difference between any two generators in the simulation time exceeds 180 degrees, the stable condition of the sample is recorded as instability, and the given label is [1,0 ]]T(ii) a Otherwise, the stable condition is recorded as stable, and the given label is [0,1 ]]T。
The total number of generators in the power grid is recorded as NGThe output of the key generator identification module is a generator serial number arranged according to the length NGColumn vector of-1
Figure BDA0002623356310000053
Each element in c takes the value of [0, 1%]In between, namely have
Figure BDA0002623356310000054
Wherein if the absolute value of the power angle difference between the generator i and the reference generator exceeds 180 degrees, the element c in the corresponding ciAnd if not, taking 1, and performing off-line training on the model by using the same sample. The output labels of the training samples are given according to the simulation curve.
Specifically, the work flow of the stable classification simple logic processing unit and the key generator simple logic processing unit is as follows:
in online operation, due to the existence of prediction error and the output characteristics of the neural network, each element in the output vector of the stable classification module is not 0/1 two values any more, but is [0,1 ]]BetweenA continuous value of (c). Noting the stable classification vector of its output
Figure BDA0002623356310000055
Is provided with
Figure BDA0002623356310000056
Similarly, the vector output by the key generator prediction module is of length NG-1 stability vector
Figure BDA0002623356310000057
Wherein each element also takes on the value of [0,1]Between, there are
Figure BDA0002623356310000058
For this purpose, their outputs will first enter a simple logic processing element, generating the inputs required by the aid of a decision-making system. The conversion logic of the link is as follows:
defining threshold value for stable classification link outputuGenerally, 0.5 is used. The logical processing link implements the following transformations: if it is
Figure BDA0002623356310000059
Output signal S1(ii) a Otherwise, the signal S is output2。
Two thresholds are respectively defined for the key generator identification link,agiven 0 < >aIs less than 1. The method is mainly used for screening key generators, and generally 0.5 is selected; whileaThe method is used for screening important generators, and the selection is generally 0.5-0.95. The logical processing link implements the following transformations: if it is
Figure BDA0002623356310000061
All the elements in the composition satisfy
Figure BDA0002623356310000062
Then the signal S is output5. Otherwise, if all elements are satisfied
Figure BDA0002623356310000063
And at least one unit with insufficient stability
Figure BDA0002623356310000064
Then the signal S is output4. Otherwise, the signal S is output3。
The predicted output signal after the simple logic processing link can only be one of the following six combinations: (S)1,S3)，(S1,S4)，(S1,S5)，(S2,S3)，(S2,S4)，(S2,S5)。
The assistant decision module takes the output after the simple logic processing link as input, and the output information comprises a final stable state judgment result and a key generator set. Wherein, the judgment result of the stable state is one of three judgments of instability, uncertainty and stability. The key generator set comprises two categories of generator subsets which are respectively key generator sets
Figure BDA0002623356310000065
Or (and) the important generator set
Figure BDA0002623356310000066
Their elements are the generator serial numbers belonging to the set, which may be an empty set.
The decision logic of the assistant decision module is shown as 2, and the specific processing logic is as follows:
1) when the input is a signal (S)2,S3) The assistant decision system outputs a destabilizing signal and simultaneously outputs
Figure BDA0002623356310000067
The generator serial number i is recorded into a key generator set
Figure BDA0002623356310000068
In will satisfy
Figure BDA0002623356310000069
The serial number of the generator is recorded in the important generator set
Figure BDA00026233563100000610
In (1). And outputting the key generator set and the important generator set after the scanning is finished. In this input category, the set of important generators is an empty set.
2) When the input is a signal (S)1,S3) When the system is in use, the assistant decision-making system outputs 'uncertain' signals, and simultaneously scans the stability degree of the generator one by one, and when the stability degree is in use
Figure BDA00026233563100000611
The sequence numbers are recorded in the key generator set,
Figure BDA00026233563100000612
the time is recorded in the important generator set, the key generator set and the important generator set are output after the scanning is finished, and if the key generator set and the important generator set do not exist, the key generator set and the important generator set are output
Figure BDA00026233563100000613
The set of vital generators is an empty set.
3) When the input is a signal (S)2,S4) When the system is in use, the assistant decision-making system outputs 'uncertain' signals, and simultaneously scans the stability degree of the generator one by one, and when the stability degree is in use
Figure BDA00026233563100000614
The sequence numbers are recorded in the key generator set,
Figure BDA00026233563100000615
the time is recorded in the important generator set, and the key generator set and the important generator set are output after the scanning is finished. The critical generator set is now an empty set.
4) When the input is a signal (S)2,S5) The decision-making aid system outputs a "stable" signal. While scanning the generator stability one by one, while
Figure BDA00026233563100000616
The sequence numbers are recorded in the key generator set,
Figure BDA00026233563100000617
the time is recorded in the important generator set, and the key generator set and the important generator set are output after the scanning is finished. Both sets are now empty sets.
5) When the input is a signal (S)1,S4) Or (S)1,S5) The decision-making aid system outputs a "stable" signal. While scanning the generator stability one by one, while
Figure BDA00026233563100000618
The sequence numbers are recorded in the key generator set,
Figure BDA00026233563100000619
the time is recorded in the important generator set, and the key generator set and the important generator set are output after the scanning is finished. At the moment, the key generator set is an empty set, and if the key generator set does not exist
Figure BDA00026233563100000620
The vital generator set is also empty.
The effect of the system is further explained below with reference to an application example:
and (3) taking an IEEE 10 machine 39 node system as a test system, finishing simulation data generation based on PSD-BPA, and verifying the validity of the model on a Pythrch developed by Facebook. The experimental equipment is configured as Intel Core i7-97003.0GHz CPU, 16GB RAM, GTX 1660Ti 6G GPU.
The IEEE 10 machine 39-node system comprises 39 nodes, 10 generators and 46 transmission lines, wherein the No. 10 generator is a reference motor. All generators are six-order models and are provided with an IEEE I type excitation system and an IEEE G1 type speed regulating system. Setting the simulation time length to be 4s, generating 29500 samples in total after transient simulation is carried out on the PSD-BPA, wherein the numbers of stable samples and unstable samples are 25168 and 4332 respectively. The sampling frequency is set as 100Hz, and the sampling interval covers the disturbance occurrence time t0+To the disturbance removal time tc+(including t)0+And tc+) The total sampling time is 0.1s, and the total number of time steps is 11. The sampling state quantity is set as the voltage amplitude U, the phase angle theta and the rotating speed omega of the generator of each bus, the node serial number is represented by the following marks, and the characteristic arrangement of a single time step is as follows:
Figure BDA0002623356310000071
the proportion of the training set, the verification set and the test set is about 6: 2: 2, all sample sets need to be marked with simulation curves to obtain off-line labels, and table 1 shows the sample labels, actual stability conditions and key generator sets in five scenes of the test set, wherein "-" represents an empty set.
Table 1 test set sample label presentation
Figure BDA0002623356310000072
Figure BDA0002623356310000081
The deep learning model comprises a 3-layer graph volume network (GCN) and a 1-layer long-short term memory (LSTM), and all parameters are optimized by an Adam algorithm. And setting a plurality of groups of parameters, training in the training set to obtain the optimal performance of each group of parameters in the verification set, and selecting the parameters with the best performance from the optimal performance as final model parameters.
Setting a threshold using the optimized depth modelu＝0.5，＝0.5，aAnd (5) verifying five scenes based on the condition that the test set simulates the online application, namely 0.9. The outputs of the stable classification module and the key generator identification module, the signal set and the decision output of the assistant decision system under five scenes are recorded, as shown in tables 2 and 3.
TABLE 2 Stable Classification and Key Generator outputs and Signal sets
Figure BDA0002623356310000082
TABLE 3 decision output of aided decision system
Figure BDA0002623356310000083
Figure BDA0002623356310000091
1) For scene 1, the stable classification module outputs a stable classification vector [1,0 ]]TIs provided with
Figure BDA0002623356310000092
The key generator identification module outputs a stability vector, and the stability of the generator is not satisfied
Figure BDA0002623356310000093
Thus, the logical element is detected and converted into a signal set (S)2,S3) And the auxiliary decision system outputs a destabilization signal and a generator set, wherein the important generator set is an empty set. The stable case, the set of critical generators in this decision, is consistent with the actual cases of table 1.
2) For scene 2, the stable classification module outputs a stable classification vector [0.15, 0.85 ]]TIs provided with
Figure BDA0002623356310000094
The key generator identification module outputs a stability vector, and the stability of the generator is not satisfied
Figure BDA0002623356310000095
Thus, the logical element is detected and converted into a signal set (S)1,S3) And outputting an uncertain signal, a key generator and an important generator set by the aid of an auxiliary decision system. It is noted that when a stability assessment is performed using only the stability classification module, this module will give a "stable" signal, whereas the sample in Table 1 is actually in a destabilized state. Auxiliary blockThe strategy system avoids missing warnings, gives stable conditions, and the key generator set is consistent with the reality.
3) For scene 3, the stable classification module outputs a stable classification vector [0.93, 0.07 ]]TIs provided with
Figure BDA0002623356310000096
The key generator identification module outputs stability degree vectors, and all stability degrees meet the requirements
Figure BDA0002623356310000097
But the stability degree is not satisfied
Figure BDA0002623356310000098
Thus, the logical element is detected and converted into a signal set (S)2,S4) And after the decision is made by the aid of an auxiliary decision system, outputting an uncertain signal and a generator set, wherein the key generator set is an empty set. It should be noted that when only the stable classification module is used for stability evaluation, this module will give a "destabilization" signal, whereas the sample in table 1 is actually in a stable state, and the decision-making aid system decision avoids the stable classification module giving a false alarm, but still gives an important set of generators that need attention.
4) For scene 4, the stable classification module outputs a stable classification vector [0.90, 0.10 ]]TThat is to say have
Figure BDA0002623356310000099
The key generator identification module outputs stability degree vectors, and all stability degrees meet
Figure BDA00026233563100000910
Thus, the logical element is detected and converted into a signal set (S)2,S5) And after the decision is made by the aid of the auxiliary decision system, a stable signal and a generator set are output. At this time, the generator sets are all empty sets. It should be noted that when the stability evaluation is performed only by the stability classification module, this module will give a "destabilization" signal, whereas the sample in Table 1 is actually in a stable state, which the aid decision system successfully avoidsThe warning is missed and the given stable situation is consistent with reality.
5) For scene 5, the stable classification module in the two sub-scenes respectively outputs a stable classification vector [0.00, 1.00 ]]TAnd [0.02, 0.98]TAll are provided with
Figure BDA00026233563100000911
The key generator identification module outputs stability degree vectors, and all stability degrees meet the requirements
Figure BDA00026233563100000912
Unsatisfied stability degree in current scene
Figure BDA0002623356310000101
Then, the logic element is detected and converted into a signal set (S)1,S4) Otherwise, it is converted into a signal set (S)1,S5). And finally, outputting a stable signal and a generator set after the decision of the assistant decision system is made. At this time, the signal set (S)1,S5) The generator sets in the corresponding sub-scenes are all empty sets, signal sets (S)1,S4) Only the set of critical generators in the corresponding sub-scenario is an empty set. The sample is actually in a stable state, and the stable condition given by the aid of the decision-making system is consistent with the actual condition.
Through the checking of the plurality of scenes, the robustness and the flexibility of the method in the transient stability decision of the large power grid are verified, and the method has important application significance for safe operation and risk regulation of the power grid.
The above embodiments are only for illustrating the technical concept and features of the present invention, and the purpose thereof is to enable those skilled in the art to understand the contents of the present invention and implement the present invention accordingly, and not to limit the protection scope of the present invention accordingly. All equivalent changes or modifications made in accordance with the spirit of the present disclosure are intended to be covered by the scope of the present disclosure.

Claims (10)

1. A power grid intelligent transient stability assessment system, comprising:
the phasor measurement unit is used for measuring and collecting the disturbed power grid state quantity at a set frequency;
the preprocessing unit is used for preprocessing the power grid state quantity acquired by the phasor measurement unit;
the stable classification module is used for taking the output of the preprocessing unit as the input and outputting a stable classification vector;
the key generator module is used for taking the output of the preprocessing unit as the input and outputting a generator stability degree vector;
the stable classification simple logic processing unit is used for carrying out logic judgment processing on the stable classification vector and outputting a stable classification conversion signal;
the key generator simple logic processing unit is used for carrying out logic judgment processing on the generator stability degree vector and outputting a generator stability degree conversion signal;
and the auxiliary decision module is used for taking the stable classification conversion signal and the generator stability conversion signal as input and outputting a final stable state judgment result and a key generator set.
2. The system of claim 1, wherein the grid state variables include bus voltage magnitudes and phase angles, and for a grid with N nodes, for the t-th sampling time step, the PMU data matrices with C characteristic variables are arranged as follows:
Figure FDA0002623356300000011
wherein, Fi＝[x1i,x2i,…,xNi]TRepresenting the ith feature column vector.
3. The system according to claim 2, wherein the preprocessing unit normalizes each characteristic of the grid state quantities acquired by the phasor measurement unit by zscore:
Figure FDA0002623356300000012
wherein, mu and sigma are respectively the mean value and variance corresponding to the feature vector, and F' is the normalized feature column vector.
4. The system according to claim 1, wherein the stability classification module employs a deep learning network model, and if the input of the system is unstable due to corresponding disturbance, the output of the module is [1,0 ]]T(ii) a Otherwise, if the system is stable, the output is [0,1 ]]T；(ii) a The deep learning network model is formed by off-line training of samples.
5. The system of claim 4, wherein the key power generation module employs a deep learning network model, and records the total number of generators in the grid as NGThe output of the key generator identification module is a generator serial number arranged according to the length NGColumn vector of-1
Figure FDA0002623356300000013
Each element in c takes the value of [0, 1%]In between, namely have
Figure FDA0002623356300000014
The deep learning network model is formed by off-line training of samples.
6. The system according to claim 4, wherein the stability classification simple logic processing unit is configured to perform logic judgment processing on the stability classification vector, and the outputting the stability classification conversion signal includes:
in online operation, each element in the output vector of the stable classification module is no longer 0/1 two values, but [0,1 ]]Between the successive values, noting the stable classification vector of its output
Figure FDA0002623356300000021
Is provided with
Figure FDA0002623356300000022
If it is
Figure FDA0002623356300000023
Output signal S1(ii) a Otherwise, the signal S is output2；uIs a threshold value.
7. The system according to claim 5, wherein the key generator simple logic processing unit is configured to perform logic judgment processing on a generator stability vector, and the outputting the generator stability conversion signal includes:
in online operation, the vector output by the key generator prediction module is N in lengthG-1 stability vector
Figure FDA0002623356300000024
Wherein each element also takes on the value of [0,1]Between, there are
Figure FDA0002623356300000025
Two threshold values are defined for each of the two,agiven 0 < >aLess than 1; the method is mainly used for screening key generators; whileaThen the method is used for screening important generators;
if it is
Figure FDA0002623356300000026
All the elements in the composition satisfy
Figure FDA0002623356300000027
Then the signal S is output5(ii) a Otherwise, if all elements are satisfied
Figure FDA0002623356300000028
And at least one unit with insufficient stability
Figure FDA0002623356300000029
Then the signal S is output4(ii) a Otherwise, the signal S is output3。
8. The grid intelligent transient stability assessment system of claim 7, wherein the input signal of the auxiliary decision module is one of the following six combinations:
(S1,S3)，(S1,S4)，(S1,S5)，(S2,S3)，(S2,S4)，(S2,S5)。
9. the system for evaluating intelligent transient stability of a power grid according to claim 8, wherein the auxiliary decision module outputs one of three determinations, namely "instability", "uncertainty" and "stability", as the final stable state determination result; the key generator set comprises two categories of generator subsets which are respectively key generator sets
Figure FDA00026233563000000210
And/or a collection of vital generators
Figure FDA00026233563000000211
Their elements are the generator numbers belonging to the set, which can be an empty set.
10. The system according to claim 9, wherein the logic of the auxiliary decision module outputting the final steady state decision result and the set of key generators is:
1) when the input is a signal (S)2,S3) The assistant decision system outputs a destabilizing signal and simultaneously outputs
Figure FDA00026233563000000212
The generator serial number i is recorded into a key generator set
Figure FDA00026233563000000213
In will satisfy
Figure FDA00026233563000000214
The serial number of the generator is recorded in the important generator set
Figure FDA00026233563000000215
Performing the following steps; outputting a key generator set and an important generator set after the scanning is finished; in the input category, the important generator set is an empty set;
2) when the input is a signal (S)1,S3) When the system is in use, the assistant decision-making system outputs 'uncertain' signals, and simultaneously scans the stability degree of the generator one by one, and when the stability degree is in use
Figure FDA00026233563000000216
The sequence numbers are recorded in the key generator set,
Figure FDA00026233563000000217
the time is recorded in the important generator set, the key generator set and the important generator set are output after the scanning is finished, and if the key generator set and the important generator set do not exist, the key generator set and the important generator set are output
Figure FDA00026233563000000218
The set of important generators is an empty set;
3) when the input is a signal (S)2,S4) When the system is in use, the assistant decision-making system outputs 'uncertain' signals, and simultaneously scans the stability degree of the generator one by one, and when the stability degree is in use
Figure FDA0002623356300000031
The sequence numbers are recorded in the key generator set,
Figure FDA0002623356300000032
the time is recorded in the important generator set, and the key generator set and the important generator set are output after the scanning is finished; at the moment, the key generator set is an empty set;
4) when the input is a signal (S)2,S5) When the system is in use, the assistant decision system outputs a 'stable' signal; while scanning the generator stability one by one, while
Figure FDA0002623356300000033
The sequence numbers are recorded in the key generator set,
Figure FDA0002623356300000034
the time is recorded in the important generator set, and the key generator set and the important generator set are output after the scanning is finished; at the moment, the two sets are both empty sets;
5) when the input is a signal (S)1,S4) Or (S)1,S5) When the system is in use, the assistant decision system outputs a 'stable' signal; while scanning the generator stability one by one, while
Figure FDA0002623356300000035
The sequence numbers are recorded in the key generator set,
Figure FDA0002623356300000036
the time is recorded in the important generator set, and the key generator set and the important generator set are output after the scanning is finished; at the moment, the key generator set is an empty set, and if the key generator set does not exist
Figure FDA0002623356300000037
The vital generator set is also empty.
CN202010789805.4A 2020-08-07 2020-08-07 Intelligent transient stability evaluation system for power grid Pending CN111884236A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010789805.4A CN111884236A (en) 2020-08-07 2020-08-07 Intelligent transient stability evaluation system for power grid

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010789805.4A CN111884236A (en) 2020-08-07 2020-08-07 Intelligent transient stability evaluation system for power grid

Publications (1)

Publication Number Publication Date
CN111884236A true CN111884236A (en) 2020-11-03

Family

ID=73211282

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010789805.4A Pending CN111884236A (en) 2020-08-07 2020-08-07 Intelligent transient stability evaluation system for power grid

Country Status (1)

Country Link
CN (1) CN111884236A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114091705A (en) * 2021-11-26 2022-02-25 国网四川省电力公司电力科学研究院 Power system instability analysis method and device, electronic equipment and storage medium
CN116361668A (en) * 2023-06-02 2023-06-30 北京安天网络安全技术有限公司 Monitoring method, device, equipment and medium for multiple SDR equipment

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106504116A (en) * 2016-10-31 2017-03-15 山东大学 Based on the stability assessment method that operation of power networks is associated with transient stability margin index
CN110417005A (en) * 2019-07-23 2019-11-05 清华大学 In conjunction with the transient stability catastrophe failure screening technique of deep learning and simulation calculation

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106504116A (en) * 2016-10-31 2017-03-15 山东大学 Based on the stability assessment method that operation of power networks is associated with transient stability margin index
CN110417005A (en) * 2019-07-23 2019-11-05 清华大学 In conjunction with the transient stability catastrophe failure screening technique of deep learning and simulation calculation

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
JIYU HUANG; LIN GUAN; YINSHENG SU: "Recurrent Graph Convolutional Network-Based Multi-Task Transient Stability Assessment Framework in Power System", IEEE ACCESS *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114091705A (en) * 2021-11-26 2022-02-25 国网四川省电力公司电力科学研究院 Power system instability analysis method and device, electronic equipment and storage medium
CN116361668A (en) * 2023-06-02 2023-06-30 北京安天网络安全技术有限公司 Monitoring method, device, equipment and medium for multiple SDR equipment
CN116361668B (en) * 2023-06-02 2023-08-11 北京安天网络安全技术有限公司 Monitoring method, device, equipment and medium for multiple SDR equipment

Similar Documents

Publication Publication Date Title
CN110829417B (en) Electric power system transient stability prediction method based on LSTM double-structure model
CN110994604B (en) Power system transient stability assessment method based on LSTM-DNN model
CN109324604A (en) A kind of intelligent train resultant fault analysis method based on source signal
CN114021433B (en) Construction method and application of dominant instability mode identification model of power system
CN111914486A (en) Power system transient stability evaluation method based on graph attention network
CN110375987A (en) One kind being based on depth forest machines Bearing Fault Detection Method
CN113376516A (en) Medium-voltage vacuum circuit breaker operation fault self-diagnosis and early-warning method based on deep learning
CN111884236A (en) Intelligent transient stability evaluation system for power grid
CN115859077A (en) Multi-feature fusion motor small sample fault diagnosis method under variable working conditions
CN110991472A (en) Micro fault diagnosis method for high-speed train traction system
CN114264915A (en) Power distribution network cable joint operation condition assessment early warning device and method
CN112308148A (en) Defect category identification and twin neural network training method, device and storage medium
CN115409335A (en) Electric power system disturbance identification method based on deep learning and considering unknown disturbance types
CN115712871A (en) Power electronic system fault diagnosis method combining resampling and integrated learning
Guo et al. Fault detection and diagnosis using statistic feature and improved broad learning for traction systems in high-speed trains
CN115600136A (en) High-voltage bushing fault diagnosis method, system and medium based on multiple sensors
CN111985528A (en) PDGAN-based cable partial discharge data enhancement method
CN111062569A (en) Low-current fault discrimination method based on BP neural network
CN113610119A (en) Method for identifying power transmission line developmental fault based on convolutional neural network
CN111898446A (en) Single-phase earth fault studying and judging method based on multi-algorithm normalization analysis
Taylor et al. Adaptive local fusion systems for novelty detection and diagnostics in condition monitoring
CN116911161A (en) Data-enhanced deep learning transient voltage stability evaluation method
CN115684786A (en) Inverter switching tube health diagnosis method, device and system based on gram angular field and parallel CNN
CN115236272A (en) Gas sensor fault diagnosis method and device under multi-working condition and storage medium
CN112966345B (en) Rotary machine residual life prediction hybrid shrinkage method based on countertraining and transfer learning

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20201103