CN108876163B - Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning - Google Patents
Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning Download PDFInfo
- Publication number
- CN108876163B CN108876163B CN201810674177.8A CN201810674177A CN108876163B CN 108876163 B CN108876163 B CN 108876163B CN 201810674177 A CN201810674177 A CN 201810674177A CN 108876163 B CN108876163 B CN 108876163B
- Authority
- CN
- China
- Prior art keywords
- operation mode
- angle stability
- power angle
- transient
- historical
- 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
- 230000001052 transient effect Effects 0.000 title claims abstract description 115
- 238000004458 analytical method Methods 0.000 title claims abstract description 46
- 238000010801 machine learning Methods 0.000 title claims abstract description 19
- 238000011156 evaluation Methods 0.000 title claims abstract description 16
- 230000001364 causal effect Effects 0.000 title claims abstract description 15
- 238000000034 method Methods 0.000 claims abstract description 35
- 238000004364 calculation method Methods 0.000 claims abstract description 24
- 238000013136 deep learning model Methods 0.000 claims abstract description 22
- 238000004445 quantitative analysis Methods 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims abstract description 11
- 238000004088 simulation Methods 0.000 claims description 22
- 230000008859 change Effects 0.000 claims description 9
- 238000013097 stability assessment Methods 0.000 claims description 7
- 238000012216 screening Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 abstract description 5
- 238000013473 artificial intelligence Methods 0.000 description 6
- 238000011160 research Methods 0.000 description 3
- 230000003068 static effect Effects 0.000 description 3
- 201000004569 Blindness Diseases 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000013145 classification model Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Educational Administration (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Game Theory and Decision Science (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a transient power angle stability rapid evaluation method integrating causal analysis and machine learning, which is characterized in that a historical operation mode sample is divided into a plurality of operation mode clusters based on historical data stored by an online safety analysis application transient stability evaluation function, transient power angle stability margin is described by a linearization formula according to the difference between the modes, and key characteristic quantity of a power grid is extracted; establishing a deep learning model by taking the key characteristic quantity as an input quantity and taking the formed cluster classification condition of the historical operation mode as an output quantity, utilizing historical data samples for training, establishing a relation between the current real-time operation mode and the historical operation mode, and estimating the transient power angle stability of the current real-time operation mode; the method can effectively reduce the time consumed by calculation while ensuring the accuracy of the transient power angle stability analysis, quickly obtain the quantitative analysis result of the transient power angle stability of the power grid, help to find the transient operation risk in the power grid in time and improve the safe operation level of the power grid.
Description
Technical Field
The invention relates to the technical field of safety and stability analysis of a power system, in particular to a transient state power angle stability rapid evaluation method for comprehensive causal analysis and machine learning.
Background
In order to ensure safe and stable operation of a power grid, particularly an extra-high voltage alternating current and direct current hybrid power grid, each level of scheduling mechanism needs to be capable of rapidly mastering safe and stable operation conditions of the power grid and rapidly positioning risk points and risk degrees of the power grid, so that a basis is provided for risk prevention and control decision of the power grid. At present, most provincial and above power grid dispatching control mechanisms in China build a power grid online safety analysis application function, the application function is based on a power grid model and real-time operation mode data, the power grid operation state is periodically analyzed and calculated, the safety and stability of the power grid are comprehensively evaluated from multiple aspects such as static state, transient state and dynamic state, an auxiliary decision is made aiming at the found safety and stability problems or potential safety and stability hazards, and a reasonable regulation scheme suggestion is provided for dispatching operation personnel.
Transient stability analysis is a key function of online safety analysis application, the core of the transient stability analysis is to analyze whether a transient power angle is unstable, at present, a time domain simulation analysis method or an EEAC (Extended Equal-Area Criterion) quantitative analysis method based on the time domain simulation analysis is mainly adopted, and a transient stability result of a power grid is obtained based on strict numerical calculation of a power grid model and real-time operation mode data. Usually, the number of transient stability faults needing to be calculated by a provincial power grid is hundreds to thousands, and hundreds of CPU (Central processing Unit) core number calculation resources are required to be deployed to meet the calculation speed requirement of completing one-time whole-grid transient stability analysis within 5-10 minutes. With the rapid expansion of the grid structure scale, the addition of a large amount of various novel devices such as wind power, photovoltaic devices and UPFC, the calculation complexity tends to rise exponentially, and the required calculation resources or calculation time consumption further increases.
Meanwhile, the 'man-machine war' of Google AlphaGo and the top high-hand of the domestic and foreign go initiates a new round of heat tide of artificial intelligence, and various reports, comments and prospects about the artificial intelligence are seen in all newspaper ends and networks, so that people can recognize the huge energy of new technologies such as artificial intelligence and the like. Machine learning is a branch of the field of artificial intelligence, and by enabling a computer to mine required information from a large amount of historical data and learn rules from the information, new samples are intelligently identified or the future is predicted, so that the computer can make correct response or judgment without specific programming in advance. Machine learning has brought a great deal of help in autodriving cars, practical speech recognition, genome recognition, and the like. In the field of power systems, how to combine the requirements of power system scheduling operation and with the help of new technologies such as big data and artificial intelligence, a new solution is provided for power system scheduling operation and analysis decision, and the method becomes a hotspot of research in the field of power system analysis decision.
Relevant research and industrial institutions of the power system combine the actual requirements of analysis and decision of the large power grid, relevant research of big data and artificial intelligence of the power system is developed, and certain achievements are obtained. In the patent, "power system stability fast judging method based on historical data" (patent No. 2015110301902), a characteristic quantity under each fault is obtained by calculating the correlation between each state quantity of a power grid and the statistic of the electric quantity and the critical fault removal time, then a measurement distance between a current real-time mode and a historical mode characteristic quantity is calculated, and a stability degree index of the current mode is obtained by adopting a K-nearest neighbor algorithm according to the measurement distance so as to judge the stability of the power grid at the current time. In the patent ' an electric power system transient stability assessment method based on-line historical data ' (patent number: 2015108372153) ', firstly, a plurality of static state quantities are preliminarily screened out manually, then, the selected static state quantities are compressed to obtain power grid characteristic quantities, on the basis, historical data samples are subjected to expansion of instability samples and compression of stable samples to form calculation samples, classification model training and parameter optimization are carried out by utilizing an SVM algorithm to form a classification model, and therefore the transient stability of the power grid is assessed. The method achieves certain effect, but the application of the method to historical analysis results is not sufficient, a pure machine learning method has certain blindness, and meanwhile, the method of manually screening the state quantity has great dependence on the experience of personnel.
Disclosure of Invention
The invention aims to provide a transient power angle stability rapid evaluation method integrating causal analysis and machine learning, which realizes rapid and accurate online judgment of the transient power angle stability of the current real-time operation mode of a power grid, effectively reduces the calculation time consumption of transient power angle stability analysis, rapidly obtains a quantitative analysis result of the transient power angle stability of the power grid, is beneficial to timely finding transient stable operation risks in the power grid and improves the safe operation level of the power grid.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
the transient state power angle stability rapid evaluation method integrating causal analysis and machine learning comprises the following steps:
1) acquiring a historical data sample; on-line safety analysis application based on the EEAC quantitative analysis method obtains historical data stored in application operation as an analyzed data sample; the historical data comprises historical operation mode data and transient state power angle stability simulation results corresponding to all examination faults in the operation mode;
2) clustering historical operation modes; considering the regularity and repeatability of actual power grid operation, classifying historical operation mode data of the power grid according to assessment faults and safety and stability modes, judging whether transient power angle stability margin change caused by operation mode difference is suitable for a linear calculation formula, and clustering the historical operation modes to corresponding operation mode clusters;
3) selecting key characteristic quantity; analyzing the influence of the power grid operation state quantity on the transient power angle stability under each examination fault based on historical operation mode data and a transient power angle stability simulation result, and selecting a key characteristic quantity under the fault;
4) deep learning model construction and training: establishing a deep learning model by taking the key characteristic quantity as an input quantity and taking the operation mode cluster as an output quantity, and training the established deep learning model by utilizing the selected key characteristic quantity and the operation mode cluster corresponding to each historical sample to obtain a trained deep learning model;
5) calculating the operation mode cluster to which the real-time operation mode belongs: inputting the current real-time operation mode of the power grid into a deep learning model obtained through training to obtain an operation mode cluster corresponding to the current real-time operation mode under each examination fault;
6) estimating the transient power angle stability margin in the current real-time mode: selecting any historical operation mode data in an operation mode cluster corresponding to the current real-time operation mode and a transient power angle stability simulation result of the historical operation mode, analyzing the difference between the current real-time operation mode and the selected historical operation mode in the cluster, obtaining a transient power angle stability margin of the current power grid operation mode based on a transient power angle stability margin quick estimation method, judging the transient power angle stability according to the power angle stability margin, and outputting a transient power angle stability result;
7) current real-time mode simulation analysis: simulating the current real-time operation mode based on an EEAC quantitative analysis method, and solving the transient power angle stability result under the current real-time mode, including transient power angle stability margin, grouping mode and participation factor;
8) and (3) simulation result sample processing: adding current real-time mode data and a transient power angle stability result in the real-time mode into a historical sample; calculating an operation mode cluster to which the current real-time mode belongs under each assessment fault according to a transient power angle stability result, judging whether the obtained operation mode cluster is consistent with a mode cluster obtained by adopting a deep learning model, and if not, retraining the deep learning model by using a historical sample;
9) and (5) acquiring the real-time operation mode data of the power grid in the new round again, and entering the step 5) to realize the evaluation of the transient power angle stability of the operation mode of the power grid in the new round.
The historical operation mode clustering in the step 2) specifically comprises the following steps:
2-1) classifying the historical operation modes according to the assessment fault and safety and stability modes, and regarding each historical operation mode in the same class as a same cluster;
2-2) judging whether the transient power angle stability margin change caused by the difference of the operation modes between the clusters is suitable for a linear calculation formula or not, and if the transient power angle stability margin change is suitable for the linear calculation formula, combining the two clusters to be classified into the same cluster;
2-3) repeating steps 2-1) and 2-2) until the treatment of all clusters under all classifications is completed.
Step 2-2) is detailedThe method comprises the following steps: in the same examination of the fault FcNext, historical operating mode SAThe transient power angle stability margin obtained by quantitative analysis is etaA.cHistorical operating mode SBThe transient power angle stability margin obtained by quantitative analysis is etaB.cIf the difference of the transient power angle stability margin is etaA.c-ηB.cEstimation difference eta of transient power angle stability margin caused by operation mode differenceestThe difference between them is less than a set threshold etaline-setNamely: l (. eta.)A.c-ηB.c)-ηest|＜ηline-setThen, the fault is considered to be in examination FcNext, historical operating mode SAAnd SBThe transient power angle stability margin of (2) is described by a linear formula.
The step 3) specifically comprises the following steps:
3-1) selecting a transient power angle stable participation factor larger than a set threshold value lambda based on the EEAC quantitative analysis resultset-AAs a key characteristic quantity;
3-2) calculating the correlation I (X, eta) of each historical operation mode state quantity of the power grid and the transient power angle stabilityA) Selecting correlation greater than a predetermined threshold value Iset-AAs a key feature quantity;
3-3) selecting power grid statistics closely related to the assessment faults as key characteristic quantities, wherein the power grid statistics closely related to the assessment faults comprise power grid total output, total load and key section power;
3-4) obtaining the union of the key characteristic quantities of 3-1), 3-2) and 3-3), and screening the top N with the maximum correlationkeyThe individual characteristic quantity is used as a key characteristic quantity of the corresponding fault; n is a radical ofkeyAnd setting according to the scale of the power grid and the calculation performance requirement.
Preferably, in the step 1), a certain number of typical operation modes of the power grid are set in the initial stage of lack of the data sample by combining the operation characteristics of the power grid, and transient simulation analysis is performed on the typical operation modes by adopting an EEAC quantitative analysis method to form a historical data sample.
The power grid running state quantity comprises a unit starting state, a unit output, a power plant output, a bus voltage, a phase angle, a connection line power flow, a connection transformer power flow, a load level, a direct current outgoing/incoming power, a section power and a line switching state.
The beneficial effects of the invention include:
the invention discloses a transient state power angle stability rapid evaluation method integrating causal analysis and machine learning, which is based on EEAC quantitative analysis, fully utilizes historical data (including power grid historical operation mode data and corresponding transient state stability result data) stored by an online safety analysis application transient state stability evaluation function, integrates a causal analysis method based on strict mathematical model derivation and a machine learning method based on big data, and realizes rapid evaluation of transient state power angle stability in a power grid real-time operation mode. Because time domain simulation of the power grid fault process is avoided, the calculation speed of transient stability evaluation can be greatly improved by adopting the method; meanwhile, transient stability quantitative information in a historical analysis result is adopted, so that the blindness of pure machine learning is reduced, and the discovery of accidental association is reduced. The method can be used as an effective supplement of the existing transient state power angle stability analysis method based on digital simulation analysis.
Drawings
The invention is further explained below with reference to the figures and examples;
fig. 1 is a flowchart of a transient state power angle stability rapid evaluation method integrating causal analysis and machine learning according to the present invention.
Detailed Description
The present invention will be described in further detail with reference to the following embodiments, which are illustrative only and not limiting, and the scope of the present invention is not limited thereby.
In order to achieve the objectives and effects of the technical means, creation features, working procedures and using methods of the present invention, and to make the evaluation methods easy to understand, the present invention will be further described with reference to the following embodiments.
The invention will now be further described, by way of example, with reference to the accompanying drawings.
As shown in fig. 1, the method for rapidly evaluating the transient state power angle stability of the combined causal analysis and machine learning provided by the present invention includes the following steps:
step 1) obtaining a historical data sample; on-line safety analysis application based on the EEAC quantitative analysis method obtains historical data stored in application operation as an analyzed data sample; the historical data comprises historical operation mode data and transient state power angle stability simulation results corresponding to all examination faults in the operation mode. In the initial stage of lacking enough data samples, a certain number of representative power grid typical operation modes are set by combining the power grid operation characteristics, and transient simulation analysis is carried out on the typical operation modes by adopting an EEAC quantitative analysis method to form historical data samples;
the typical operation mode of the representative power grid is to cover different scenes such as big summer, small summer, big winter, small winter, overhaul, open loop, closed loop and the like, and ensure the diversity of data samples by setting different power generation-load levels.
Step 2)) clustering historical operation modes; considering that the actual power grid operation has regularity and repeatability, classifying historical operation mode data of the power grid according to assessment faults and safety and stability modes, judging whether transient power angle stability margin change caused by operation mode difference is suitable for a linear calculation formula, and clustering the historical operation modes to corresponding operation mode clusters. Consistent assessment faults and the same safe and stable mode are arranged in the same operation mode cluster, and the transient power angle stability margin can be described by a linear formula according to mode difference. The same operation mode cluster may include a plurality of operation modes and their analysis calculation results, or may include only 1 operation mode and its analysis calculation results.
Dividing the historical operating modes into corresponding operating mode clusters includes the following:
2-1) classifying the historical operation modes according to the assessment fault and safety and stability modes, and taking each historical operation mode in the same class as 1 cluster;
2-2) analyzing whether stability margin change caused by difference of operation modes between the cluster A and the cluster B is suitable for linear calculation formula description, and if so, merging the two clusters (the cluster A and the cluster B) to be classified into the same cluster; wherein, the cluster A and the cluster B refer to any two operation mode clusters;
2-3) repeating the steps until all the classifications and all the clusters are combined.
Wherein, the judging method of whether the change of the stability margin caused by the difference of the operation modes is described by a linear calculation formula is as follows: set at the same examination fault FcNext, historical operating mode SAThe transient power angle stability margin obtained by quantitative analysis is etaA.cHistorical operating mode SBThe transient power angle stability margin obtained by quantitative analysis is etaB.cIf the difference of the transient power angle stability margin is etaA.c-ηB.cAnd transient power angle stability margin estimation difference eta estimated by operation mode difference through formula (5) in a patent ' on-line transient safety stability assessment method based on automatic screening of expected fault set ' (patent number: 201710247481X) 'estThe difference between them is less than a set threshold etaline-setNamely: l (. eta.)A.c-ηB.c)-ηest|＜ηline-setThen, the fault is considered to be in examination FcNext, historical operating mode SAAnd SBThe transient power angle stability margin of (2) is described by a linear formula.
Step 3), selecting key characteristic quantity; analyzing the influence of the power grid operation state quantity on the transient power angle stability under each examination fault based on historical operation mode data and a transient power angle stability simulation result, and selecting a key characteristic quantity under the fault;
selecting the key feature quantities includes the following:
3-1) selecting a transient power angle stable participation factor larger than a set threshold value lambda based on the EEAC quantitative analysis resultset-AAs a key characteristic quantity;
3-2) calculating the correlation I (X, eta) of each historical operation mode state quantity of the power grid and the transient power angle stabilityA) Selecting correlation greater than a predetermined threshold value Iset-AAs a key feature quantity;
3-3) manually selecting part of power grid statistics closely related to the assessment fault, including total output, total load and key section power of the power grid, as key characteristic quantities;
3-4) obtaining the union of the three key characteristic quantities, and screening the top N with the maximum correlationkeyThe individual characteristic quantities serve as key characteristic quantities for the corresponding fault. N is a radical ofkeyAnd setting according to the scale of the power grid and the calculation performance requirement.
The power grid running state quantity comprises a unit starting state, a unit output, a power plant output, a bus voltage, a phase angle, a connection line power flow, a connection transformer power flow, a load level, a direct current outgoing/incoming power, a section power and a line switching state.
Step 4), deep learning model construction and training: establishing a deep learning model by taking the key characteristic quantity as an input quantity and taking the operation mode cluster as an output quantity, and training the established deep learning model by utilizing the selected key characteristic quantity and the operation mode cluster corresponding to each historical sample to obtain a trained deep learning model;
step 5), calculating the operation mode cluster to which the real-time operation mode belongs: inputting the current real-time operation mode of the power grid into a deep learning model obtained through training to obtain an operation mode cluster corresponding to the current real-time operation mode under each examination fault;
step 6), estimating the transient power angle stability margin of the current real-time mode: selecting any historical operation mode data in an operation mode cluster corresponding to the current real-time operation mode and a transient power angle stability simulation result of the historical operation mode, analyzing the difference between the current real-time operation mode and the selected historical operation mode in the cluster, obtaining the transient power angle stability margin of the current power grid operation mode by using a transient power angle stability margin quick estimation method in an online transient power angle safety stability assessment method (patent number: 201710247481X) based on automatic screening of an expected fault set, judging the transient power angle stability according to the power angle stability margin, and outputting the transient power angle stability result;
step 7), current real-time mode simulation analysis: simulating the current real-time operation mode based on an EEAC quantitative analysis method, and solving the transient power angle stability result under the current real-time mode, including transient power angle stability margin, grouping mode and participation factor;
step 8) simulation result sample processing: adding current real-time mode data and a transient power angle stability result in the real-time mode into a historical sample; calculating an operation mode cluster to which the current real-time mode belongs under each assessment fault according to a transient power angle stability result, judging whether the obtained operation mode cluster is consistent with a mode cluster obtained by adopting a deep learning model, and if not, retraining the deep learning model by using a historical sample;
and 9) acquiring the real-time operation mode data of the new power grid, and entering the step 5) to realize the evaluation of the transient power angle stability of the operation mode of the new power grid.
Those skilled in the art can design the invention to be modified or varied without departing from the spirit and scope of the invention. Therefore, if such modifications and variations of the present invention fall within the technical scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (5)
1. The method for rapidly evaluating the stability of the transient power angle by combining causal analysis and machine learning is characterized by comprising the following steps of:
1) acquiring a historical data sample; on-line safety analysis application based on the EEAC quantitative analysis method obtains historical data stored in application operation as an analyzed data sample; the historical data comprises historical operation mode data and transient state power angle stability simulation results corresponding to all examination faults in the operation mode;
2) clustering historical operation modes; considering the regularity and repeatability of actual power grid operation, classifying historical operation mode data of the power grid according to assessment faults and safety and stability modes, judging whether transient power angle stability margin change caused by operation mode difference is suitable for a linear calculation formula, and clustering the historical operation modes to corresponding operation mode clusters;
3) selecting key characteristic quantity; analyzing the influence of the power grid operation state quantity on the transient power angle stability under each examination fault based on historical operation mode data and a transient power angle stability simulation result, and selecting a key characteristic quantity under the fault;
4) deep learning model construction and training: establishing a deep learning model by taking the key characteristic quantity as an input quantity and taking the operation mode cluster as an output quantity, and training the established deep learning model by utilizing the selected key characteristic quantity and the operation mode cluster corresponding to each historical sample to obtain a trained deep learning model;
5) calculating the operation mode cluster to which the real-time operation mode belongs: inputting the current real-time operation mode of the power grid into a deep learning model obtained through training to obtain an operation mode cluster corresponding to the current real-time operation mode under each examination fault;
6) estimating the transient power angle stability margin in the current real-time mode: selecting any historical operation mode data in an operation mode cluster corresponding to the current real-time operation mode and a transient power angle stability simulation result of the historical operation mode, analyzing the difference between the current real-time operation mode and the selected historical operation mode in the cluster, obtaining a transient power angle stability margin of the current power grid operation mode based on a transient power angle stability margin quick estimation method, judging the transient power angle stability according to the power angle stability margin, and outputting a transient power angle stability result;
7) current real-time mode simulation analysis: simulating the current real-time operation mode based on an EEAC quantitative analysis method, and solving the transient power angle stability result under the current real-time mode, including transient power angle stability margin, grouping mode and participation factor;
8) and (3) simulation result sample processing: adding current real-time mode data and a transient power angle stability result in the real-time mode into a historical sample; calculating an operation mode cluster to which the current real-time mode belongs under each assessment fault according to a transient power angle stability result, judging whether the obtained operation mode cluster is consistent with a mode cluster obtained by adopting a deep learning model, and if not, retraining the deep learning model by using a historical sample;
9) re-acquiring the real-time operation mode data of the power grid in the new round, and entering the step 5) to realize the evaluation of the transient power angle stability of the operation mode of the power grid in the new round;
the historical operation mode clustering in the step 2) specifically comprises the following steps:
2-1) classifying the historical operation modes according to the assessment fault and safety and stability modes, and regarding each historical operation mode in the same class as a same cluster;
2-2) judging whether the transient power angle stability margin change caused by the difference of the operation modes between the clusters is suitable for a linear calculation formula or not, and if the transient power angle stability margin change is suitable for the linear calculation formula, combining the two clusters to be classified into the same cluster;
2-3) repeating steps 2-1) and 2-2) until the treatment of all clusters under all classifications is completed.
2. The method for rapid transient power angle stability assessment by combined causal analysis and machine learning of claim 1,
the step 2-2) specifically comprises the following steps: in the same examination of the fault FcNext, historical operating mode SAThe transient power angle stability margin obtained by quantitative analysis is etaA.cHistorical operating mode SBThe transient power angle stability margin obtained by quantitative analysis is etaB.cIf the difference of the transient power angle stability margin is etaA.c-ηB.cEstimation difference eta of transient power angle stability margin caused by operation mode differenceestThe difference between them is less than a set threshold etaline-set，|(ηA.c-ηB.c)-ηest|＜ηline-setThen, the fault is considered to be in examination FcNext, historical operating mode SAAnd SBThe transient power angle stability margin of (2) is described by a linear formula.
3. The method for rapid transient power angle stability assessment by combined causal analysis and machine learning of claim 1,
the step 3) specifically comprises the following steps:
3-1) based on the amount of EEACAnalyzing the analysis result, selecting the transient power angle stable participation factor larger than the set threshold value lambdaset-AAs a key characteristic quantity;
3-2) calculating the correlation I (X, eta) of each historical operation mode state quantity of the power grid and the transient power angle stabilityA) Selecting correlation greater than a predetermined threshold value Iset-AAs a key feature quantity;
3-3) selecting power grid statistics closely related to the assessment faults as key characteristic quantities, wherein the power grid statistics closely related to the assessment faults comprise power grid total output, total load and key section power;
3-4) obtaining the union of the key characteristic quantities of 3-1), 3-2) and 3-3), and screening the top N with the maximum correlationkeyThe individual characteristic quantity is used as a key characteristic quantity of the corresponding fault; n is a radical ofkeyAnd setting according to the scale of the power grid and the calculation performance requirement.
4. The method for rapid transient power angle stability assessment by combined causal analysis and machine learning of claim 1,
step 1) setting a certain number of typical operation modes of the power grid in combination with the operation characteristics of the power grid at the initial stage of lacking of the data sample, and performing transient simulation analysis on the typical operation modes by adopting an EEAC quantitative analysis method to form a historical data sample.
5. The method for rapid transient power angle stability assessment by combined causal analysis and machine learning of claim 1,
the power grid running state quantity comprises a unit starting state, a unit output, a power plant output, a bus voltage, a phase angle, a connection line power flow, a connection transformer power flow, a load level, a direct current outgoing/incoming power, a section power and a line switching state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810674177.8A CN108876163B (en) | 2018-06-27 | 2018-06-27 | Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810674177.8A CN108876163B (en) | 2018-06-27 | 2018-06-27 | Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108876163A CN108876163A (en) | 2018-11-23 |
| CN108876163B true CN108876163B (en) | 2021-07-23 |
Family
ID=64294952
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810674177.8A Active CN108876163B (en) | 2018-06-27 | 2018-06-27 | Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108876163B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109861206B (en) * | 2018-12-29 | 2022-08-09 | 国电南瑞科技股份有限公司 | Transient power angle stability fault screening system and method based on support vector machine |
| CN109902352B (en) * | 2019-01-24 | 2022-07-01 | 南瑞集团有限公司 | Extraction method and system for key characteristic quantity of transient power angle stability analysis |
| CN110163540B (en) * | 2019-06-28 | 2021-06-15 | 清华大学 | Power system transient stability prevention control method and system |
| CN110443724B (en) * | 2019-07-19 | 2022-08-16 | 河海大学 | Electric power system rapid state estimation method based on deep learning |
| CN110323743B (en) * | 2019-07-24 | 2020-08-25 | 国电南瑞科技股份有限公司 | Clustering method and device for transient power angle stability evaluation historical data |
| CN110348540B (en) * | 2019-07-24 | 2021-06-01 | 国电南瑞科技股份有限公司 | Clustering-based method and device for screening transient power angle stability faults of power system |
| CN110707695B (en) * | 2019-11-12 | 2021-01-26 | 国电南瑞科技股份有限公司 | Transient power angle stability margin calculation method and system based on artificial intelligence |
| CN110969214B (en) * | 2019-12-18 | 2023-06-23 | 天津大学 | Transient security domain online construction method based on support vector machine comprehensive model |
| CN113114527B (en) * | 2021-03-15 | 2022-11-18 | 河池学院 | Micro-grid transient stability rapid discrimination system and method based on machine learning |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102074955B (en) * | 2011-01-20 | 2015-06-10 | 中国电力科学研究院 | Method based on knowledge discovery technology for stability assessment and control of electric system |
| WO2012122310A1 (en) * | 2011-03-08 | 2012-09-13 | Trilliant Networks, Inc. | System and method for managing load distribution across a power grid |
| CN104578053B (en) * | 2015-01-09 | 2017-01-04 | 北京交通大学 | Power system transient stability Forecasting Methodology based on disturbed voltage trace bunch feature |
| CN106504116B (en) * | 2016-10-31 | 2020-05-22 | 山东大学 | Stability evaluation method based on power grid operation and transient stability margin index correlation |
| CN107800140B (en) * | 2017-10-18 | 2021-08-13 | 天津大学 | Large user power supply access decision method considering load characteristics |
-
2018
- 2018-06-27 CN CN201810674177.8A patent/CN108876163B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108876163A (en) | 2018-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108876163B (en) | Transient state power angle stability rapid evaluation method integrating causal analysis and machine learning | |
| CN107516170B (en) | Difference self-healing control method based on equipment failure probability and power grid operation risk | |
| CN102663412B (en) | Power equipment current-carrying fault trend prediction method based on least squares support vector machine | |
| CN104809658B (en) | A kind of rapid analysis method of low-voltage distribution network taiwan area line loss | |
| CN106909933A (en) | A kind of stealing classification Forecasting Methodology of three stages various visual angles Fusion Features | |
| CN112149873B (en) | Low-voltage station line loss reasonable interval prediction method based on deep learning | |
| CN103488869A (en) | Wind power generation short-term load forecast method of least squares support vector machine | |
| CN104408562A (en) | Photovoltaic system generating efficiency comprehensive evaluation method based on BP (back propagation) neural network | |
| CN110837915B (en) | Low-voltage load point prediction and probability prediction method for power system based on hybrid integrated deep learning | |
| CN106611243A (en) | Residual correction method for wind speed prediction based on GARCH (Generalized ARCH) model | |
| CN109242136A (en) | A kind of micro-capacitance sensor wind power Chaos-Genetic-BP neural network prediction technique | |
| CN111652478B (en) | Umbrella algorithm-based power system voltage stability evaluation misclassification constraint method | |
| Xue et al. | Adaptive ultra-short-term wind power prediction based on risk assessment | |
| CN115859099A (en) | Sample generation method and device, electronic equipment and storage medium | |
| CN112508254B (en) | Method for determining investment prediction data of transformer substation engineering project | |
| Wang et al. | Analysis of network loss energy measurement based on machine learning | |
| Guo et al. | Wind speed forecasting of genetic neural model based on rough set theory | |
| CN112052985A (en) | Lightgbm-based medium-short term low voltage prediction method | |
| Xue et al. | An efficient and robust case sorting algorithm for transient stability assessment | |
| CN113837486B (en) | RNN-RBM-based distribution network feeder long-term load prediction method | |
| CN116663393A (en) | Random forest-based power distribution network continuous high-temperature fault risk level prediction method | |
| CN116317937A (en) | Distributed photovoltaic power station operation fault diagnosis method | |
| Zhao et al. | Short-term Wind Power Prediction Method Based on GCN-LSTM | |
| CN108959719B (en) | AC-DC hybrid distributed renewable energy system test scene simulation method | |
| CN112633565A (en) | Photovoltaic power aggregation interval prediction method |
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 |