CN111582571A - Power grid operation situation sensing method and system with model driving and data driving integrated - Google Patents
Power grid operation situation sensing method and system with model driving and data driving integrated Download PDFInfo
- Publication number
- CN111582571A CN111582571A CN202010364427.5A CN202010364427A CN111582571A CN 111582571 A CN111582571 A CN 111582571A CN 202010364427 A CN202010364427 A CN 202010364427A CN 111582571 A CN111582571 A CN 111582571A
- Authority
- CN
- China
- Prior art keywords
- power grid
- index
- situation
- formula
- driven
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 68
- 230000008447 perception Effects 0.000 claims abstract description 26
- 230000007246 mechanism Effects 0.000 claims abstract description 21
- 230000008569 process Effects 0.000 claims abstract description 21
- 238000011156 evaluation Methods 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 14
- 238000005065 mining Methods 0.000 claims abstract description 13
- 238000012847 principal component analysis method Methods 0.000 claims abstract description 12
- 239000013598 vector Substances 0.000 claims description 42
- 239000011159 matrix material Substances 0.000 claims description 29
- 238000004364 calculation method Methods 0.000 claims description 24
- 238000004590 computer program Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 11
- 238000009826 distribution Methods 0.000 claims description 10
- 230000015654 memory Effects 0.000 claims description 10
- 230000004913 activation Effects 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 4
- 210000002569 neuron Anatomy 0.000 claims description 4
- 230000009471 action Effects 0.000 claims description 3
- 230000000295 complement effect Effects 0.000 claims description 3
- 230000001186 cumulative effect Effects 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000000513 principal component analysis Methods 0.000 claims description 3
- 230000004304 visual acuity Effects 0.000 claims description 3
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 2
- 238000012163 sequencing technique Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 11
- 238000013528 artificial neural network Methods 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 5
- 210000004027 cell Anatomy 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000003203 everyday effect Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 238000009412 basement excavation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000010219 correlation analysis Methods 0.000 description 1
- 230000003631 expected effect Effects 0.000 description 1
- 210000002837 heart atrium Anatomy 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000013526 transfer learning Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 238000013024 troubleshooting Methods 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention discloses a power grid operation situation perception method and system integrating model driving and data driving, which comprises the following steps of mining key elements: removing the factor of excessive overlapping of operation index information by using a principal component analysis method based on model driving, and constructing a power grid operation situation evaluation system for representing the operation track of a power grid; situation understanding: evaluating the current running state by using a model-driven based fuzzy analytic hierarchy process; and (3) situation prediction: the perception of the power grid safety situation is completed by training and learning through sample data by utilizing an LSTM-attention mechanism based on data driving; the perception of the security situation of the power grid is finished through sample data training and learning; the result proves that the model-driven and data-driven integrated power grid operation situation perception method can be used as a scientific, reasonable and comprehensive power grid regulation situation perception system and can be accurately suitable for perceiving the current operation state of a power system.
Description
Technical Field
The invention belongs to the technical field of power grid regulation and control, and particularly relates to a power grid operation situation sensing method and system integrating model driving and data driving.
Background
In recent years, with the continuous development of extra-high voltage power grids and energy internet, the time-varying property and complexity of power grid operation modes are increasingly enhanced, and the difficulty in recognizing the power grid is increasingly greater. The traditional regulation and control method based on mechanism analysis and a power grid model depends on decision, execution and other links of experience analysis of regulation and control personnel, and when the problems of nonlinearity, discontinuity and prediction uncertainty of a large power grid are solved, the expected effect is difficult to achieve, and the power grid regulation and control field faces unprecedented opportunities and challenges; in addition, the existing power grid operation focuses on controlling the real-time operation state of the power grid, and the understanding of the whole-process continuous state and possible operation risks of the power grid in a future period is lacked. Therefore, a scientific, reasonable and comprehensive power grid regulation situation perception system is urgently needed to be constructed, and the current operation state and the future development trend of the power system are perceived.
The basis of grid situation awareness is to correctly evaluate the grid operation state. To date, many experts and scholars at home and abroad have made a lot of remarkable results in the aspect of power system operation evaluation. In some prior arts, a set of five major indexes including power system safety power supply capacity, static voltage safety, topological structure vulnerability, transient safety and risk index are constructed; the state of the power system is evaluated. In some prior art, a smart grid strategic index set and a process index set are respectively constructed from a macro strategy and a micro process, and the association of the two is realized in an index chain form, so that the state evaluation of the power system is finally completed. Some prior arts construct a set of key index systems for evaluating the operation state of the power grid based on a large amount of data provided by a support system of the smart power grid dispatching technology from 4 aspects of safety, economy, goodness and cleanness. Some prior arts propose a power grid safety index calculation and display method based on an analytic hierarchy process and a variable weight mechanism, and represent power grid safety by a comprehensive index. Some prior arts provide a power grid dispatching operation Key Performance Indicator (KPI) construction method based on an extension theory and correlation analysis, and evaluate the operation performance of a power system.
With the access of high-proportion renewable energy, the development and change conditions of the running state of the power grid are difficult to control. Under the background, scholars at home and abroad put forward the concept of situation awareness of the power system and try to improve the understanding of the power system by means of the situation awareness theory. Some prior arts establish a grid index system in five aspects of a current state, a development state, an ability state, a controllable state and an evaluation state based on a dispatcher thinking mode, and the grid index system is used as a characteristic quantity of a running track. Some prior arts describe the focus of examining whether the operation state and trend of the power grid are abnormal or not from the aspects of the abnormal change of the power grid structure, the severe fluctuation of intermittent power supply power, the rapid rising of load in the peak period, the rapid falling of load in the valley period, and the like. Some prior arts provide a trend identification model of the operation state of the power system according to the real-time monitoring information of the system operation. Comprehensive evaluation is carried out on the running state of the power grid through a fuzzy analytic hierarchy process, and comprehensive trend identification of the running state of the power grid is achieved. In order to improve the sensing and controlling capability of an extra-high voltage alternating current-direct current hybrid large power grid, a full-dimensional multi-level power grid operation track index system is provided in some prior art.
With the rapid development of artificial intelligence technologies such as deep learning, the artificial intelligence technology based on the data driving mode has potential technical advantages in the aspect of solving the perception of the operation situation of the power grid. In recent years, expert scholars have gained certain results from researches on load prediction, emergency control of power grid, automatic voltage control, automatic power generation control and the like by adopting deep learning, reinforcement learning and transfer learning, and have also made preliminary exploration on the aspect of power grid operation situation perception. In some prior art, a large power grid intelligent regulation and control system framework based on big data and artificial intelligence is designed, and the purpose is to realize intelligent scheduling of 'measuring-identifying-controlling' of a large power grid, and comprehensively improve power grid situation perception and cooperative control capability. Some prior art combines deep learning with security situation awareness of a power grid, and provides power grid security situation awareness based on deep learning. Some prior arts propose a method for constructing a knowledge graph of a dispatching automation system by combining bottom-up and top-down, and introduce the application of the knowledge graph by taking the case that the telemetering data in a D5000 system does not refresh the fault, so as to assist related personnel in fault positioning and problem troubleshooting.
Disclosure of Invention
The invention provides a power grid operation situation sensing method and system integrating model driving and data driving, which utilize a principal component analysis method to realize key element mining, eliminate factors of excessive overlapping of operation index information, mine key elements capable of reflecting power grid situation operation tracks and establish an index system capable of correctly reflecting the power grid tracks. Alternatively, a fuzzy analytic hierarchy process may be used to obtain a weight coefficient of each index, so as to realize a method for understanding the current situation. As an alternative, the method of adding the Attention mechanism into the LSTM model predicts the future power grid situation.
The invention discloses a power grid operation situation perception method integrating model driving and data driving, which comprises the following steps,
the method comprises the following steps of firstly, establishing a basic framework for evaluating the operation situation of the power grid, and dividing the perception of the situation of the power grid into: a key element mining stage, a situation understanding stage and a situation prediction stage;
secondly, in a key element mining stage, removing the factor of excessive overlapping of operation index information by using a principal component analysis method based on model driving, and constructing a power grid operation situation evaluation system for representing the operation track of a power grid;
thirdly, in a situation understanding stage, evaluating the current operation state by using a model-driven fuzzy analytic hierarchy process;
and fourthly, in the situation prediction stage, training and learning through sample data by using a data-driven LSTM-attention mechanism, and finishing the perception of the safety situation of the power grid.
In the above scheme, preferably, in the second step, in the stage of mining the key elements, the factor of excessive overlapping of operation index information is removed by using a principal component analysis method, and the key elements capable of reflecting the power grid situation operation track are mined, so that excessive calculation amount is avoided.
Preferably, in the second step, in the stage of mining the key elements, the factor of excessive overlapping of the operation index information is removed by using a principal component analysis method, and the key elements capable of reflecting the power grid situation operation track are mined, and the specific steps are as follows:
step one, setting n attribute indexes of an index system, and evaluating by adopting m samples to form an index matrix B as formula (1):
step two, preprocessing the index matrix B: the selected index types are different and comprise positive indexes and negative indexes, and the index types need to be unified; if the index types are unified into the negative-direction index, taking the reciprocal of the positive-direction index for calculation; the index data is normalized by the following equation (2):
in the formula: z is a radical ofijFor the purpose of the index data after the normalization,
the j-th index value is the mean value and the mean square error respectively. All index data after normalization form a matrix of Z ═ Z (Z)ij)m×n；
Step three, establishing Z ═ Z by formula (3)ij)m×nIs given by the correlation coefficient matrix R ═ (rij)m×nElement r in the matrixijReflecting index ZiAnd ZjThe degree of correlation of (a) with (b),
in formula (3): cov (Z)i,Zj) Is ZiAnd ZjThe covariance of (a); d (Z)i)、D(Zj) Are each ZiAnd ZjThe variance of (a);
step four, solving q characteristic roots of the correlation matrix R according to the formula (4):
λ1≥λ2≥λ3≥…≥λq≥0 (4)
the corresponding feature vector is as shown in formula (5):
ej＝(l1j,l2j,…,lnj),j＝1,…q (5)
step five, calculating the variance contribution rate
And to wjIn descending order, the cumulative variance contribution rate of the first l (l ≦ q) principal components
Meanwhile, the main component is considered to comprehensively embody n indexes;
step six, calculating the principal components of the m samples according to the formula (6):
Zi,j＝Zm,n×[e1e2…eq]′ (6)
step seven, selecting the first main components, and calculating by the formula (7) to obtain indexes:
preferably, in the seventh step, a power grid operation situation evaluation index system is constructed from three aspects of a power grid structure, power grid equipment and a system state through principal component analysis; the power grid structure comprises grid structure integrity and power transmission efficiency; the power grid equipment comprises five indexes of line load rate, transformer load rate, voltage margin, section flow margin and load change rate; the system state comprises four indexes of frequency safety deviation rate, active standby, network loss rate and n-1 passing rate.
Preferably, in the third step, in the situation understanding stage, the original information provided by the EMS and the WMAS, or the state indirectly obtained through appropriate processing and calculation based on the information is selected as the power grid operation trajectory index system;
eliminating uncertainty in index weight distribution by a fuzzy analytic hierarchy process;
let fuzzy complementary matrix F ═ Fij)n×n，fij∈[0,1]Wherein: n is an index criterion number, generally, n is 6, fijTaking a number scale of 0.1-0.9;
the weight of each index is as shown in formula (8):
in the formula: u. ofZi，uZjRespectively representing the weight decision results of the expert Z to the indexes i and j, and satisfying fij＝logαuZi-logαuZj+0.5, α is the resolving power of the decision maker, and the resolution of the weight assignment scheme is improved by increasing the value of α.
Preferably, in the third step, in the situation understanding stage, clustering analysis is performed on the expert decision vectors by using a k-means algorithm, and the algorithm flow includes the following steps:
step one, assuming that the weight of each index is completed by N experts, inputting N11-dimensional (corresponding to 3 types of 11 indexes) expert decision weight vectors { U) to be classified1,U2,…,UZ,…UNIn which U isZ＝(uZ1,uZ2,…,uZ11) Representing the weight distribution result of the expert Z to 11 indexes, wherein the number of the clusters to be classified is k;
step two, randomly selecting a decision vector distributed by k experts to the index weight as an initial clustering center { p1,p2,…,pi,…pkIn which p isi＝(pi1,pi2,…,pi11) A weight decision vector representing the ith cluster center; selecting a clustering maximum iteration number V; determining a maximum convergence coefficient M of the iteration end;
calculating the Euclidean distance from each decision vector to each cluster, and dividing each decision vector into the clusters with the minimum distance, wherein the calculation formula of the Euclidean distance is shown as a formula (9):
step four, recalculating the central values { p ] of the k clusters1,p2,…,pi,…pkIn which p isilAs shown in formula (10):
wherein:
represents UZTo fall into class piThe decision vector of (1); l is the number of decision vectors that fall within the class.
Step five, checking whether the clustering operation is finished or not, wherein if the iteration times are equal to P, the clustering is finished; otherwise, calculating the convergence distance of each cluster of the iteration, if the convergence distances are smaller than the given parameter M, ending, otherwise, continuing the iteration; the convergence distance calculation formula of the mth iteration is as follows (11):
step six, if the category plIncluding nlAnd (3) sequencing the vectors to obtain the weight of the expert decision vector as shown in the formula (12):
the final weight vector of the index layer is thus obtained as equation (13):
preferably, in the fourth step, in the situation prediction stage, the prediction of the grid situation is completed through a learning framework, and the method includes:
step one, memorizing a unit state value c according to t-1 moment through an LSTM unitt-1Hidden layer t-1 time output value ht-1And the input value x at time ttCalculating the hidden layer output value h at the time tt(ii) a According to the flow direction of the signal, the calculation rule is as follows (14) - (19):
ft＝σ(Wfxxt+Wfhht-1+bf) (14)
it＝σ(Wixxt+Wihht-1+bi) (15)
ct＝ct-1⊙ft+gt⊙it(17)
ot＝σ(Woxxt+Wohht-1+bo) (18)
in formulae (14) to (19): w is a weight matrix at the moment t, b is an offset, sigma is a sigmoid activation function,
⊙ denotes an exclusive nor operation for the tanh activation function;
step two, adopting an attention mechanism and using skiExpressing the influence of the ith sequence point on the kth sequence point, and the calculation formula is shown as formula (20):
ski＝Uk-1tanh(V1hk+V2hi+b) (20)
wherein, Uk-1Is the vector, V, held by the Attention hidden layer update1、V2Is the matrix coefficient, b is the bias coefficient;
then each s iskiInputting the softma layer for normalization to obtain probability distribution skiThus like shadowThe probability corresponding to the sequence points with larger sound is larger so as to restrain some invalid information or noise, and the calculation formula is as follows (21):
then each αkiWeighting and summing to obtain the attention coefficient C of the kth sequence point, and finally comprehensively solving the output value U of the attention layerkAnd updating the hidden layer storage value, wherein the formula is as shown in formula (22) and formula (23):
Uk＝tanh(C,hk) (23)
attention layer output ukAs the input of the next dense all-connected layer, outputting a predicted value through the action of sigmoid activation function
The formula is as in formula (24):
and step three, combining the attention mechanism with the LSTM network to construct a prediction base model.
It is also preferable that the number of neurons in the two LSTM layers in step three be 32 and 64, respectively.
The model-driven and data-driven integrated power grid operation situation perception system comprises a memory and a processor, wherein a computer program capable of running on the processor is stored in the memory, and the processor executes the computer program to realize the method steps of the model-driven and data-driven integrated power grid operation situation perception method.
The invention has the beneficial effects that:
the model-driven and data-driven integrated power grid operation situation perception method establishes a basic framework for evaluating the power grid operation situation, and divides the power grid situation perception into: the method comprises three stages of key element mining, situation understanding and situation prediction. In the key element mining stage, removing the factor of excessive overlapping of operation index information by using a principal component analysis method based on model driving, and constructing a power grid operation situation evaluation system for representing the operation track of a power grid; in a situation understanding stage, evaluating the current running state by using a model-driven fuzzy analytic hierarchy process, and providing a sample for situation prediction; in the situation prediction stage, the perception of the power grid safety situation is finished by training and learning through sample data by utilizing a data-driven LSTM-attention mechanism; the result proves that the power grid operation situation perception method integrating model driving and data driving can be used as a scientific, reasonable and comprehensive power grid regulation and control situation perception system, can be accurately suitable for perceiving the current operation state of a power system, and further has higher accuracy and applicability when being used as a situation prediction model of an LSTM neural network based on an Attention mechanism.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the invention and, together with the description, serve to explain the invention and not to limit the invention. In the drawings:
fig. 1 is a situation awareness flow chart of a power grid operation situation awareness method based on model driving and data driving fusion.
FIG. 2 is a schematic diagram of basic units of an LSTM network of the model-driven and data-driven integrated power grid operation situation awareness method.
Fig. 3 is a schematic structural diagram of an Attention module of an Attention mechanism of the power grid operation situation awareness method based on model driving and data driving fusion.
FIG. 4 is a schematic diagram of an Attention-LSTM network model of the model-driven and data-driven integrated power grid operation situation awareness method of the present invention.
Detailed Description
The present invention will be described in detail below with reference to the embodiments with reference to the attached drawings. It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict.
The following detailed description is exemplary in nature and is intended to provide further details of the invention. Unless otherwise defined, all technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention.
Example 1
A power grid operation situation perception method integrating model driving and data driving, as shown in figure 1, comprises the following steps,
step one, digging key elements: removing the factor of excessive overlapping of operation index information by using a principal component analysis method based on model driving, and constructing a power grid operation situation evaluation system for representing the operation track of a power grid;
and secondly, situation understanding: evaluating the current running state by using a model-driven based fuzzy analytic hierarchy process;
thirdly, situation prediction: the perception of the power grid safety situation is completed by training and learning through sample data by utilizing an LSTM-attention mechanism based on data driving; wherein the content of the first and second substances,
the first step, the excavation of key elements,
firstly, removing the factor of excessive overlapping of operation index information by using a principal component analysis method, and mining key elements capable of reflecting the situation operation track of the power grid to avoid excessive calculation amount; the method comprises the following specific steps:
step one, setting n attribute indexes of an index system, and evaluating by adopting m samples to form an index matrix B:
step two, preprocessing the index matrix B: the selected index types are different and comprise a positive index (the larger the attribute value is, the higher the energy efficiency level) and a negative index (the smaller the attribute value is, the higher the energy efficiency level) and the orientation of the comprehensive evaluation function cannot be judged, so the index types are unified. If the index types are unified into the negative-direction index, the reciprocal of the positive-direction index is calculated. In addition, in order to eliminate inconsistency of the evaluation of the level of effectiveness due to the difference in the dimension and magnitude of the index, the index value needs to be standardized. The index data is normalized by the following formula:
in the formula: z is a radical ofijFor the purpose of the index data after the normalization,
the j-th index value is the mean value and the mean square error respectively. All index data after normalization form a matrix of Z ═ Z (Z)ij)m×n。
Step three, establishing Z ═ Zij)m×nIs given by the correlation coefficient matrix R ═ (rij)m×nElement r in the matrixijReflecting index ZiAnd ZjThe degree of correlation of (c).
In the formula: cov (Z)i,Zj) Is ZiAnd ZjThe covariance of (a); d (Z)i)、D(Zj) Are each ZiAnd ZjThe variance of (a);
step four, solving q characteristic roots of the correlation matrix R:
λ1≥λ2≥λ3≥…≥λq≥0 (4)
its corresponding feature vector: e.g. of the typej＝(l1j,l2j,…,lnj),j＝1,…q (5)
Step five, calculating the variance contribution rate
And to wjIn descending order, the cumulative variance contribution rate of the first l (l ≦ q) principal components
Meanwhile, the main component is considered to comprehensively represent n indexes;
step six, calculating the principal components of the m samples: zi,j＝Zm,n×[e1e2…eq]′ (6)
Step seven, selecting the first main components, and calculating to obtain indexes:
through principal component analysis, a power grid operation situation evaluation index system is constructed from three aspects of a power grid structure, power grid equipment and a system state. The power grid structure comprises grid structure integrity and power transmission efficiency; the power grid equipment comprises five indexes of line load rate, transformer load rate, voltage margin, section flow margin and load change rate; the system state comprises four indexes of frequency safety deviation rate, active standby, network loss rate and n-1 passing rate
Secondly, situation understanding is carried out, and original information which can be provided by EMS and WMAS or a state which is indirectly obtained through proper processing and calculation based on the information is selected as a power grid operation track index system; however, the running state of the power grid needs to be comprehensively evaluated, the relative importance degree of each index is different, and how to set the weight value among the indexes is the key of the comprehensive evaluation;
the fuzzy analytic hierarchy process is a multi-index weight distribution process combining fuzzy mathematics and analytic hierarchy process, which overcomes the problems that the consistency of judgment matrix of analytic hierarchy process is difficult to check, and the like, and can eliminate the uncertainty problem in index weight distribution;
let fuzzy complementary matrix F ═ Fij)n×n(fij∈[0,1]Wherein: n is index criterion number, generally, n is 6), fijTaking a number scale of 0.1-0.9；
The index weights are:
in the formula: u. ofZi，uZjRespectively representing the weight decision results of the expert Z to the indexes i and j, and satisfying fij＝logαuZi-logαuZjα is the resolving power of the decision maker, the resolution of the weight allocation scheme can be improved by increasing the value of α;
in order to extract a uniform decision weight vector from decision results of a plurality of experts and reduce the subjectivity of decision, a k-means algorithm is introduced to carry out cluster analysis on the expert decision vectors; the algorithm flow is as follows:
step one, assuming that the weight of each index is completed by N experts, inputting N11-dimensional (corresponding to 3 types of 11 indexes) expert decision weight vectors { U) to be classified1,U2,…,UZ,…UNIn which U isZ＝(uZ1,uZ2,…,uZ11) Representing the weight distribution result of the expert Z to 11 indexes, wherein the number of the clusters to be classified is k;
step two, randomly selecting a decision vector distributed by k experts to the index weight as an initial clustering center { p1,p2,…,pi,…pkIn which p isi＝(pi1,pi2,…,pi11) A weight decision vector representing the ith cluster center; selecting a clustering maximum iteration number V; determining a maximum convergence coefficient M of the iteration end;
step three, calculating the Euclidean distance from each decision vector to each cluster, and dividing each decision vector into the cluster with the minimum distance, wherein the calculation formula of the Euclidean distance is as follows:
step four, recalculating the central values { p ] of the k clusters1,p2,…,pi,…pkIn which p isilComprises the following steps:
wherein:
represents UZTo fall into class piThe decision vector of (1); l is the number of decision vectors that fall within the class.
Step five, checking whether the clustering operation is finished or not, wherein if the iteration times are equal to P, the clustering is finished; otherwise, calculating the convergence distance of each cluster of the iteration, if the convergence distances are smaller than the given parameter M, ending, otherwise, continuing the iteration. The convergence distance calculation formula of the mth iteration is as follows:
step six, if the category plIncluding nlAnd (3) the weights of the expert decision vectors are obtained as follows:
the final weight vector for the index layer is thus:
and thirdly, predicting the operation trend of the power grid, wherein due to the randomness and uncertainty of the operation situation of the power grid, the situation result is often influenced by various factors, and a more accurate situation predicted value is difficult to obtain by a traditional prediction model. The method completes the prediction of the power grid situation through a deep learning framework;
wherein, based on the long-time memory network model principle of the attention mechanism, the LSTM principle is as follows: the LSTM network is a modified time-cycled neural network (RNN) for processing timing signals. LSTM basic unit masterThe system is composed of an input gate, an output gate and a forgetting gate, as shown in fig. 2. In the process of processing information by the basic unit, the most important thing is the transfer of the state of the unit, i.e. from c above in fig. 1t-1To ctA horizontal line that passes information from the previous cell to the next cell.
The LSTM cell memorizes the cell state value c according to the t-1 timet-1Hidden layer t-1 time output value ht-1And the input value x at time ttCalculating the hidden layer output value h at the time tt. According to the flow direction of the signal, the specific calculation rule is as follows:
ft＝σ(Wfxxt+Wfhht-1+bf) (14)
it＝σ(Wixxt+Wihht-1+bi) (15)
ct＝ct-1⊙ft+gt⊙it(17)
ot＝σ(Woxxt+Wohht-1+bo) (18)
in the formula: w is a weight matrix at the moment t, b is an offset, sigma is a sigmoid activation function,
for the tanh activation function, "⊙" represents an "exclusive nor" multiplication.
Among them, Attention mechanism (Attention Mode) mechanism: the Attention Model is a mechanism Model simulating human brain Attention, and aims to enable a neural network to selectively pay Attention to input features, save learned feature weights and assign the learned feature weights to input vectors of the next time step, and allocate Attention by using a weight matrix, so that the influence of key input features on prediction is highlighted.
The principle of the attention mechanism used is shown in figure 3. Output values h of the second-level LSTM nodes of the model shown in FIG. 31,h2,…,htRepresenting an input characteristic sequence of the Attention structure, and also serving as a state value in an Attention first-layer hidden layer; u shapekThe vector representing the sequence point weighted by the Attention coefficient is the vector stored in the last hidden layer of the Attention, and is also the output of the Attention layer.
S in FIG. 3kiThe influence of the ith sequence point on the kth sequence point is shown, and the calculation formula is as follows:
ski＝Uk-1tanh(V1hk+V2hi+b) (20)
Uk-1is the vector, V, held by the Attention hidden layer update1、V2Is the matrix coefficient and b is the bias coefficient. Then each s iskiInputting the softma layer for normalization to obtain probability distribution skiTherefore, the probability corresponding to the sequence point with larger influence is larger, and some invalid information or noise can be suppressed. The calculation formula is as follows:
then each αkiWeighting and summing to obtain the attention coefficient C of the kth sequence point, and finally comprehensively solving the output value U of the attention layerkAnd updating the hidden layer storage value, wherein the formula is as follows:
Uk＝tanh(C,hk) (23)
in the model shown in FIG. 3, the attribute layer outputs ukAs the input of the next dense all-connected layer, outputting a predicted value through the action of sigmoid activation function
The formula is as follows:
wherein, the situation prediction based on the LSTM-attention mechanism is as follows: the attention mechanism is combined with the LSTM network to construct a prediction basis model, and the network structure is shown in fig. 4, in which the numbers of neurons of two LSTM layers are 32 and 64, respectively.
Example 2
In the model-driven and data-driven integrated power grid operation situation sensing method of the embodiment 1, power grid operation data of 2018, 7, month 1 and 12, month 31 in a certain eastern region are used as samples, sampling is performed according to a minute unit, and 1440 sections are provided every day. The data section information comprises active and reactive power output of a generator, line tide, transformer power, node voltage and active and reactive power of load;
step one in the attached figure 1 illustrates that input sample data is collected, the method uses power grid operation data of 2018, 7, month 1 and 12, month 31 in a certain eastern region as samples, sampling is carried out according to the unit of minutes, and 1440 sections are arranged every day. The data section information comprises active and reactive power output of the generator, line tide, transformer power, node voltage and active and reactive power of load.
The second step in fig. 1 is to analyze 128 indexes of the existing power grid operation by using a principal component analysis method, eliminate the factor of excessive overlapping of operation index information, and mine a key index capable of reflecting the power grid situation operation track.
Step three in the attached figure 1 describes that on the basis of step two, indexes which are complex in calculation and can not be obtained through direct or simple operation of section data are removed, and a power grid operation situation evaluation index system in three aspects of a power grid structure, power grid equipment and a system state is obtained. The power grid structure comprises grid structure integrity and power transmission efficiency; the power grid equipment comprises five indexes of line load rate, transformer load rate, voltage margin, section flow margin and load change rate; the system state comprises four indexes of frequency safety deviation rate, active standby, network loss rate and n-1 passing rate. .
Step four 4 in the attached figure 1 describes that each index weight coefficient is obtained by using a fuzzy analytic hierarchy process, and each index reflecting different aspects of the power grid operation is integrated into a parameter capable of comprehensively evaluating the current state.
The fifth step in the attached figure 1 describes that an LSTM-Attention network model is constructed, an LSTM layer in the model provided by the method comprises 64 neurons, an Attention layer serves as an interface of the LSTM layer, a memory unit is solved by reasonably distributing Attention through a whole atrium, and finally a prediction result is output through a whole connection Dense layer.
Step six of fig. 1 illustrates that the training set data is input into the LSTM-Attention network training model, the method sets the parameter batch _ size (the number of samples selected in one training) to 200, epoch to 100, and the training can be stopped in advance after the model converges, thereby avoiding overfitting of the model.
Step seven in the attached figure 1 illustrates that prediction of the future power grid operation situation can be obtained by inputting certain section information according to the training result.
Wherein, when a complete data set passes through the neural network once and returns once, the process is called an epoch; that is, all training samples are propagated in the neural network in a forward direction and a backward direction, that is, an Epoch is a process of training all training samples once. LSTM is a Long Short-term memory network (Long Short-term memory).
In another embodiment of the present invention, a model-driven and data-driven converged power grid operation situation awareness system is further provided, which includes a memory and a processor, where the memory stores a computer program executable on the processor, and the processor implements the method steps of the model-driven and data-driven converged power grid operation situation awareness method when executing the computer program.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting the same, and although the present invention is described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications and equivalents may be made to the embodiments of the invention without departing from the spirit and scope of the invention, which is to be covered by the claims.
Claims (9)
1. The model-driven and data-driven integrated power grid operation situation perception method is characterized by comprising the following steps of,
the method comprises the following steps of firstly, establishing a basic framework for evaluating the operation situation of the power grid, and dividing the perception of the situation of the power grid into: a key element mining stage, a situation understanding stage and a situation prediction stage;
secondly, in a key element mining stage, removing the factor of excessive overlapping of operation index information by using a principal component analysis method based on model driving, and constructing a power grid operation situation evaluation system for representing the operation track of a power grid;
thirdly, in a situation understanding stage, evaluating the current operation state by using a model-driven fuzzy analytic hierarchy process;
and fourthly, in the situation prediction stage, training and learning through sample data by using a data-driven LSTM-attention mechanism, and finishing the perception of the safety situation of the power grid.
2. The model-driven and data-driven integrated power grid operation situation perception method as claimed in claim 1, wherein in the second step, in a key element mining stage, a factor of excessive overlapping of operation index information is removed by using a principal component analysis method, and key elements capable of reflecting a power grid situation operation track are mined.
3. The model-driven and data-driven integrated power grid operation situation sensing method as claimed in claim 2, wherein in the second step, in a key element mining stage, a factor of excessive overlapping of operation index information is removed by using a principal component analysis method, and a key element capable of reflecting a power grid situation operation track is mined, and the specific steps are as follows:
step one, setting n attribute indexes of an index system, and evaluating by adopting m samples to form an index matrix B as formula (1):
step two, preprocessing the index matrix B: the selected index types are different and comprise positive indexes and negative indexes, and the index types need to be unified; if the index types are unified into the negative-direction index, taking the reciprocal of the positive-direction index for calculation; the index data is normalized by the following equation (2):
in the formula: z is a radical ofijFor the purpose of the index data after the normalization,
respectively obtaining the mean value and the mean square error of the jth index value; all index data after normalization form a matrix of Z ═ Z (Z)ij)m×n；
Step three, establishing Z ═ Z by formula (3)ij)m×nIs given by the correlation coefficient matrix R ═ (rij)m×nElement r in the matrixijReflecting index ZiAnd ZjThe degree of correlation of (a) with (b),
in formula (3): cov (Z)i,Zj) Is ZiAnd ZjThe covariance of (a); d (Z)i)、D(Zj) Are each ZiAnd ZjThe variance of (a);
step four, solving q characteristic roots of the correlation matrix R according to the formula (4):
λ1≥λ2≥λ3≥…≥λq≥0 (4)
the corresponding feature vector is as shown in formula (5):
ej＝(l1j,l2j,…,lnj),j＝1,…q (5)
step five, calculating the variance contribution rate
And to wjIn descending order, if the cumulative variance contribution of the first one principal component
Meanwhile, the main component is considered to comprehensively embody n indexes; l is less than or equal to q;
step six, calculating the principal components of the m samples according to the formula (6):
Zi,j＝Zm,n×[e1e2…eq]′ (6)
step seven, selecting the first main components, and calculating by the formula (7) to obtain indexes:
4. the model-driven and data-driven integrated power grid operation situation perception method according to claim 3, characterized in that in step seven, a power grid operation situation evaluation index system is constructed from three aspects of a power grid structure, power grid equipment and a system state through principal component analysis; the power grid structure comprises grid structure integrity and power transmission efficiency; the power grid equipment comprises five indexes of line load rate, transformer load rate, voltage margin, section flow margin and load change rate; the system state comprises four indexes of frequency safety deviation rate, active standby, network loss rate and n-1 passing rate.
5. The model-driven and data-driven integrated power grid operation situation awareness method as claimed in claim 1, wherein in the third step, in the situation understanding stage, the original information provided by EMS and WMAS, or the state indirectly obtained through processing and calculation based on the information, is selected as the power grid operation track index system;
eliminating uncertainty in index weight distribution by a fuzzy analytic hierarchy process;
let fuzzy complementary matrix F ═ Fij)n×n，fij∈[0,1]Wherein: n is an index criterion number, generally, n is 6, fijTaking a number scale of 0.1-0.9;
the weight of each index is as shown in formula (8):
in the formula: u. ofZi，uZjRespectively representing the weight decision results of the expert Z to the indexes i and j, and satisfying fij＝logαuZi-logαuZj+0.5, α is the resolving power of the decision maker.
6. The model-driven and data-driven integrated power grid operation situation awareness method according to claim 5, characterized in that in the third step, in a situation understanding phase, an expert decision vector is subjected to clustering analysis through a k-means algorithm, and the algorithm flow comprises the following steps:
step one, assuming that the weight of each index is completed by N experts, inputting N11-dimensional expert decision weight vectors { U) to be classified1,U2,…,UZ,…UNIn which U isZ＝(uZ1,uZ2,…,uZ11) Representing the weight distribution result of the expert Z to 11 indexes, wherein the number of the clusters to be classified is k;
step two, randomly selecting a decision vector distributed by k experts to the index weight as an initial clustering center { p1,p2,…,pi,…pkIn which p isi＝(pi1,pi2,…,pi11) A weight decision vector representing the ith cluster center; selecting a clustering maximum iteration number V; determining a maximum convergence coefficient M of the iteration end;
calculating the Euclidean distance from each decision vector to each cluster, and dividing each decision vector into the clusters with the minimum distance, wherein the calculation formula of the Euclidean distance is shown as a formula (9):
step four, recalculating the central values { p ] of the k clusters1,p2,…,pi,…pkIn which p isilAs shown in formula (10):
wherein:
represents UZTo fall into class piThe decision vector of (1); l is the number of decision vectors falling into the class;
step five, checking whether the clustering operation is finished or not, wherein if the iteration times are equal to P, the clustering is finished; otherwise, calculating the convergence distance of each cluster of the iteration, if the convergence distances are smaller than the given parameter M, ending, otherwise, continuing the iteration; the convergence distance calculation formula of the mth iteration is as follows (11):
step six, if the category plIncluding nlAnd (3) sequencing the vectors to obtain the weight of the expert decision vector as shown in the formula (12):
the final weight vector of the index layer is thus obtained as equation (13):
7. the model-driven and data-driven integrated power grid operation situation awareness method according to claim 1, wherein in the fourth step, in the situation prediction stage, the prediction of the power grid situation is completed through a learning framework, and the method comprises the following steps:
step one, memorizing a unit state value c according to t-1 moment through an LSTM unitt-1Hidden layer t-1 time output value ht-1And the input value x at time ttCalculating the hidden layer output value h at the time tt(ii) a According to the flow direction of the signal, the calculation rule is as follows (14) - (19):
ft＝σ(Wfxxt+Wfhht-1+bf) (14)
it＝σ(Wixxt+Wihht-1+bi) (15)
ct＝ct-1⊙ft+gt⊙it(17)
ot＝σ(Woxxt+Wohht-1+bo) (18)
in formulae (14) to (19): w is a weight matrix at the moment t, b is an offset, sigma is a sigmoid activation function,
⊙ denotes an exclusive nor operation for the tanh activation function;
step two, adopting an attention mechanism and using skiExpressing the influence of the ith sequence point on the kth sequence point, and the calculation formula is shown as formula (20):
ski＝Uk-1tanh(V1hk+V2hi+b) (20)
wherein, Uk-1Is the vector, V, held by the Attention hidden layer update1、V2Is the matrix coefficient, b is the bias coefficient;
then each s iskiInputting the softma layer for normalization to obtain probability distribution skiThe probability corresponding to the sequence points with larger influence is larger so as to suppress some invalid information or noise, and the calculation formula is as follows (21):
then each αkiWeighting and summing to obtain the attention coefficient C of the kth sequence point, and finally comprehensively solving the output value U of the attention layerkAnd updating the hidden layer storage value, wherein the formula is as shown in formula (22) and formula (23):
Uk＝tanh(C,hk) (23)
attention layer output ukAs the input of the next dense all-connected layer, outputting a predicted value through the action of sigmoid activation function
The formula is as in formula (24):
and step three, combining the attention mechanism with the LSTM network to construct a prediction base model.
8. The model-driven and data-driven fused power grid operation situation awareness method according to claim 7, wherein in the third step, the number of neurons in the two LSTM layers is 32 and 64 respectively.
9. Model-driven and data-driven converged power grid operation situation awareness system, characterized by comprising a memory, a processor, a computer program being stored in the memory and being executable on the processor, the processor implementing the method steps of the model-driven and data-driven converged power grid operation situation awareness method according to any one of claims 1 to 8 when executing the computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010364427.5A CN111582571A (en) | 2020-04-30 | 2020-04-30 | Power grid operation situation sensing method and system with model driving and data driving integrated |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010364427.5A CN111582571A (en) | 2020-04-30 | 2020-04-30 | Power grid operation situation sensing method and system with model driving and data driving integrated |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111582571A true CN111582571A (en) | 2020-08-25 |
Family
ID=72124625
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010364427.5A Pending CN111582571A (en) | 2020-04-30 | 2020-04-30 | Power grid operation situation sensing method and system with model driving and data driving integrated |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111582571A (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112101673A (en) * | 2020-09-22 | 2020-12-18 | 华北电力大学 | Power grid development trend prediction method and system based on hidden Markov model |
| CN112199421A (en) * | 2020-12-04 | 2021-01-08 | 中国电力科学研究院有限公司 | Multi-source heterogeneous data fusion and measurement data multi-source mutual verification method and system |
| CN112329868A (en) * | 2020-11-10 | 2021-02-05 | 西安电子科技大学 | CLARA clustering-based manufacturing and processing equipment group energy efficiency state evaluation method |
| CN112488417A (en) * | 2020-12-14 | 2021-03-12 | 国网江苏省电力有限公司苏州供电分公司 | Power grid operation characteristic sensing method and system based on LBP and neural network |
| CN112505551A (en) * | 2020-12-16 | 2021-03-16 | 北方工业大学 | Screening method for retired power battery |
| CN112866262A (en) * | 2021-01-25 | 2021-05-28 | 东方电气自动控制工程有限公司 | Power plant safety I area situation perception platform based on neural network |
| CN113052469A (en) * | 2021-03-30 | 2021-06-29 | 贵州电网有限责任公司 | Method for calculating wind-solar-water-load complementary characteristic of small hydropower area lacking measurement runoff |
| CN113393102A (en) * | 2021-06-02 | 2021-09-14 | 重庆大学 | Distribution transformer operation state trend prediction method based on data driving |
| CN113688910A (en) * | 2021-08-26 | 2021-11-23 | 国网江苏省电力有限公司常州供电分公司 | Power grid operation multi-source information fusion method based on k-BP algorithm |
| CN113762716A (en) * | 2021-07-30 | 2021-12-07 | 国网山东省电力公司营销服务中心(计量中心) | Method and system for evaluating running state of transformer area based on deep learning and attention |
| CN113850481A (en) * | 2021-09-07 | 2021-12-28 | 华南理工大学 | Power system scheduling service assistant decision method, system, device and storage medium |
| CN114021449A (en) * | 2021-10-29 | 2022-02-08 | 江苏方天电力技术有限公司 | Prediction method for coal mill safety evaluation |
| CN114418194A (en) * | 2021-12-29 | 2022-04-29 | 广东电网有限责任公司 | Tower damage prediction method and device based on data driving and model driving |
| CN114626449A (en) * | 2022-03-03 | 2022-06-14 | 南京航空航天大学 | Target recognition network degradation analysis method under countermeasure environment based on least square weight determination method |
| CN114938385A (en) * | 2022-04-29 | 2022-08-23 | 广东电网有限责任公司 | Power physical network security situation sensing method, device, equipment and medium |
| CN115022194A (en) * | 2022-05-24 | 2022-09-06 | 桂林电子科技大学 | Network security situation prediction method based on SA-GRU |
| CN115022187A (en) * | 2022-01-20 | 2022-09-06 | 长沙理工大学 | Data-driven electric-gas integrated energy system situation perception method and model considering time sequence correlation |
| CN115130387A (en) * | 2022-07-14 | 2022-09-30 | 北京中泰瑞通科技有限公司 | Intelligent situation sensing system for wind power generation |
| CN115481702A (en) * | 2022-10-28 | 2022-12-16 | 中国人民解放军国防科技大学 | Predictive comparison characterization method for multi-element time series data processing |
| CN116881704A (en) * | 2023-09-06 | 2023-10-13 | 北京新亚盛创电气技术有限公司 | Early warning method and system for power grid running state |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104794534A (en) * | 2015-04-16 | 2015-07-22 | 国网山东省电力公司临沂供电公司 | Power grid security situation predicting method based on improved deep learning model |
| CN105956757A (en) * | 2016-04-27 | 2016-09-21 | 上海交通大学 | Comprehensive evaluation method for sustainable development of smart power grid based on AHP-PCA algorithm |
-
2020
- 2020-04-30 CN CN202010364427.5A patent/CN111582571A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104794534A (en) * | 2015-04-16 | 2015-07-22 | 国网山东省电力公司临沂供电公司 | Power grid security situation predicting method based on improved deep learning model |
| CN105956757A (en) * | 2016-04-27 | 2016-09-21 | 上海交通大学 | Comprehensive evaluation method for sustainable development of smart power grid based on AHP-PCA algorithm |
Non-Patent Citations (2)
| Title |
|---|
| ANAPARTHI, K.K. ET AL.: "Coherency identification in power systems through principal component analysis", 《IEEE TRANSACTIONS ON POWER SYSTEMS》 * |
| 林静怀 等: "模型驱动和数据驱动融合的电网运行态势感知方法", 《电力信息与通信技术》 * |
Cited By (34)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112101673B (en) * | 2020-09-22 | 2024-01-16 | 华北电力大学 | Power grid development trend prediction method and system based on hidden Markov model |
| CN112101673A (en) * | 2020-09-22 | 2020-12-18 | 华北电力大学 | Power grid development trend prediction method and system based on hidden Markov model |
| CN112329868A (en) * | 2020-11-10 | 2021-02-05 | 西安电子科技大学 | CLARA clustering-based manufacturing and processing equipment group energy efficiency state evaluation method |
| CN112329868B (en) * | 2020-11-10 | 2023-08-01 | 西安电子科技大学 | CLARA clustering-based manufacturing and processing equipment group energy efficiency state evaluation method |
| CN112199421A (en) * | 2020-12-04 | 2021-01-08 | 中国电力科学研究院有限公司 | Multi-source heterogeneous data fusion and measurement data multi-source mutual verification method and system |
| CN112199421B (en) * | 2020-12-04 | 2021-03-09 | 中国电力科学研究院有限公司 | Multi-source heterogeneous data fusion and measurement data multi-source mutual verification method and system |
| CN112488417B (en) * | 2020-12-14 | 2022-06-21 | 国网江苏省电力有限公司苏州供电分公司 | Power grid operation characteristic sensing method and system based on LBP and neural network |
| CN112488417A (en) * | 2020-12-14 | 2021-03-12 | 国网江苏省电力有限公司苏州供电分公司 | Power grid operation characteristic sensing method and system based on LBP and neural network |
| CN112505551B (en) * | 2020-12-16 | 2023-01-10 | 北方工业大学 | Screening method for retired power battery |
| CN112505551A (en) * | 2020-12-16 | 2021-03-16 | 北方工业大学 | Screening method for retired power battery |
| CN112866262A (en) * | 2021-01-25 | 2021-05-28 | 东方电气自动控制工程有限公司 | Power plant safety I area situation perception platform based on neural network |
| CN113052469B (en) * | 2021-03-30 | 2022-04-22 | 贵州电网有限责任公司 | Method for calculating wind-solar-water-load complementary characteristic of small hydropower area lacking measurement runoff |
| CN113052469A (en) * | 2021-03-30 | 2021-06-29 | 贵州电网有限责任公司 | Method for calculating wind-solar-water-load complementary characteristic of small hydropower area lacking measurement runoff |
| CN113393102B (en) * | 2021-06-02 | 2022-08-12 | 重庆大学 | Distribution transformer operation state trend prediction method based on data driving |
| CN113393102A (en) * | 2021-06-02 | 2021-09-14 | 重庆大学 | Distribution transformer operation state trend prediction method based on data driving |
| CN113762716A (en) * | 2021-07-30 | 2021-12-07 | 国网山东省电力公司营销服务中心(计量中心) | Method and system for evaluating running state of transformer area based on deep learning and attention |
| CN113688910A (en) * | 2021-08-26 | 2021-11-23 | 国网江苏省电力有限公司常州供电分公司 | Power grid operation multi-source information fusion method based on k-BP algorithm |
| CN113850481A (en) * | 2021-09-07 | 2021-12-28 | 华南理工大学 | Power system scheduling service assistant decision method, system, device and storage medium |
| CN114021449A (en) * | 2021-10-29 | 2022-02-08 | 江苏方天电力技术有限公司 | Prediction method for coal mill safety evaluation |
| CN114418194B (en) * | 2021-12-29 | 2022-10-11 | 广东电网有限责任公司 | Tower damage prediction method and device based on data driving and model driving |
| CN114418194A (en) * | 2021-12-29 | 2022-04-29 | 广东电网有限责任公司 | Tower damage prediction method and device based on data driving and model driving |
| CN115022187A (en) * | 2022-01-20 | 2022-09-06 | 长沙理工大学 | Data-driven electric-gas integrated energy system situation perception method and model considering time sequence correlation |
| CN115022187B (en) * | 2022-01-20 | 2023-12-08 | 长沙理工大学 | Situation awareness method and device for electric-gas comprehensive energy system |
| CN114626449A (en) * | 2022-03-03 | 2022-06-14 | 南京航空航天大学 | Target recognition network degradation analysis method under countermeasure environment based on least square weight determination method |
| CN114938385A (en) * | 2022-04-29 | 2022-08-23 | 广东电网有限责任公司 | Power physical network security situation sensing method, device, equipment and medium |
| CN114938385B (en) * | 2022-04-29 | 2023-10-24 | 广东电网有限责任公司 | Electric power physical network security situation awareness method, device, equipment and medium |
| CN115022194A (en) * | 2022-05-24 | 2022-09-06 | 桂林电子科技大学 | Network security situation prediction method based on SA-GRU |
| CN115022194B (en) * | 2022-05-24 | 2023-09-26 | 桂林电子科技大学 | Network security situation prediction method based on SA-GRU |
| CN115130387A (en) * | 2022-07-14 | 2022-09-30 | 北京中泰瑞通科技有限公司 | Intelligent situation sensing system for wind power generation |
| CN115130387B (en) * | 2022-07-14 | 2024-04-30 | 北京中泰瑞通科技有限公司 | Intelligent situation sensing system for wind power generation |
| CN115481702A (en) * | 2022-10-28 | 2022-12-16 | 中国人民解放军国防科技大学 | Predictive comparison characterization method for multi-element time series data processing |
| US11882299B1 (en) | 2022-10-28 | 2024-01-23 | National University Of Defense Technology | Predictive contrastive representation method for multivariate time-series data processing |
| CN116881704B (en) * | 2023-09-06 | 2023-11-14 | 北京新亚盛创电气技术有限公司 | Early warning method and system for power grid running state |
| CN116881704A (en) * | 2023-09-06 | 2023-10-13 | 北京新亚盛创电气技术有限公司 | Early warning method and system for power grid running state |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111582571A (en) | Power grid operation situation sensing method and system with model driving and data driving integrated | |
| Groumpos | Fuzzy cognitive maps: basic theories and their application to complex systems | |
| CN112381320A (en) | Community personnel flow prediction method, system, storage medium and equipment | |
| Chandiok et al. | Machine learning techniques for cognitive decision making | |
| Wang et al. | A remaining useful life prediction model based on hybrid long-short sequences for engines | |
| Alizadeh et al. | Using data mining for learning and clustering FCM | |
| Buragohain | Adaptive network based fuzzy inference system (ANFIS) as a tool for system identification with special emphasis on training data minimization | |
| Mohammadazadeh et al. | Neural Networks and Learning Algorithms in MATLAB | |
| Marques-Quinteiro et al. | Improvisation: A taste of chaos in the middle of order and a taste of order in the middle of chaos | |
| Iraji et al. | Students classification with adaptive neuro fuzzy | |
| Kravets et al. | Dynamic coordination of strategies for multi-agent systems | |
| Milad et al. | Expanded neo-fuzzy adaptive decayed brain emotional learning network for online time series predication | |
| Gonzalez et al. | Parallelization of an evolving artificial neural networks system to forecast time series using openmp and mpi | |
| Yusof et al. | A concise fuzzy rule base to reason student performance based on rough-fuzzy approach | |
| Vasylkiv et al. | Estimation Method of Information System Functioning Quality Based on the Fuzzy Logic. | |
| Napitupulu | Artificial neural network application in gross domestic product forecasting: an Indonesia case | |
| Ilunga | Short-term hourly load forecasting in south africa using neural networks | |
| Dragomir et al. | Adaptive Neuro–Fuzzy Inference Systems–An Alternative Forecasting Tool for Prosumers | |
| Kaplan et al. | Fuzzy Reliability Theory in the Decision-Making Process | |
| CN115456073B (en) | Long-short-term memory-based modeling analysis method for generating type countermeasure network model | |
| Tokede et al. | Mapping relational efficiency in neuro-fuzzy hybrid cost models | |
| Babaali et al. | The Comparison of Single-Layer and Two-Layer MLP Neural Networks with the LM Learning Method and ANFIS Network in Determining the Stability Factor of Earth Dams | |
| Loseva et al. | Eensembles of neural networks with application of multi-objective self-configuring genetic programming in forecasting problems | |
| Komal et al. | Two phase approach for performance analysis and optimisation of industrial systems using uncertain data | |
| Müller | Automatic Model Generation in Process Modelling |
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 |
Application publication date: 20200825 |
|
| RJ01 | Rejection of invention patent application after publication |