CN106484994B - A kind of dynamic load model modelling approach using support vector machines linear kernel - Google Patents
A kind of dynamic load model modelling approach using support vector machines linear kernel Download PDFInfo
- Publication number
- CN106484994B CN106484994B CN201610878488.7A CN201610878488A CN106484994B CN 106484994 B CN106484994 B CN 106484994B CN 201610878488 A CN201610878488 A CN 201610878488A CN 106484994 B CN106484994 B CN 106484994B
- Authority
- CN
- China
- Prior art keywords
- model
- load
- power
- support vector
- vector machines
- 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
- 238000012706 support-vector machine Methods 0.000 title claims abstract description 23
- 238000013459 approach Methods 0.000 title claims abstract description 6
- 239000013598 vector Substances 0.000 claims abstract description 30
- 238000000034 method Methods 0.000 claims description 19
- 238000012549 training Methods 0.000 claims description 16
- 238000012546 transfer Methods 0.000 claims description 5
- 238000002790 cross-validation Methods 0.000 claims description 3
- 230000005611 electricity Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 230000017260 vegetative to reproductive phase transition of meristem Effects 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 2
- 230000004044 response Effects 0.000 abstract description 3
- 238000005259 measurement Methods 0.000 description 7
- 238000013528 artificial neural network Methods 0.000 description 3
- 239000002131 composite material Substances 0.000 description 3
- 238000007796 conventional method Methods 0.000 description 2
- 238000013499 data model Methods 0.000 description 2
- 238000012850 discrimination method Methods 0.000 description 2
- 230000002068 genetic effect Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- Computer Hardware Design (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Geometry (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The present invention relates to a kind of dynamic load model modelling approach using support vector machines, comprising: the electric network data acquired in real time collects the electric network fault data of certain power grid particular power branch;Cover sheet is trained load module data using support vector machines, obtains supporting vector machine model, and the structure of setting model with respect to earth fault, the data of phase fault and three phase short circuit fault;According to model equivalency principle, the active power load differential equation model of the electrical branch is obtained;According to the structure of linear nuclear model and unidentified system.The present invention comprehensively considers empirical risk minimization and structural risk minimization, and the electrical load model made has better generalization ability, can more reflect the dynamic of system, the more response characteristic of approaching to reality electric system.
Description
Technical field
The present invention relates to Electrical Power System Dynamic load modeling fields, are related specifically to electric system load module structure
Established and system parameter identification.
Background technique
In field of power system, load module is mainly used for Power System Analysis and emulation, for the safety of electric system
Stability analysis plays an important role, and load module is largely divided into three classes at present: one kind is static load model, and main representative is
ZIP model (Z represents constant-impedance, and I represents constant current, and P represents invariable power);Second class is dynamic load model, mainly includes 1)
Input and output transfer function model;2) composite load model, composite load model are an induction motor model and ZIP model
Parallel connection obtains.Dynamic load model is the needs for studying Electrical Power System Dynamic, and the accuracy of power system dynamic model is to electric power
The stability influence of system is very big.At present there are mainly two types of the method for load modeling, method based on partial model and based on surveying
The method for measuring data.Method based on measurement data is to establish load module using the data of in-site measurement, so that measurement data
Preset model structure, such as transfer function model or composite load model can be fitted.It is general based on measurement number
According to dynamic load model modelling approach be that Levenberg- can be used to nonlinear model using least square method
Marquardt least square method.But no matter which kind of model its inherently depict active power, reactive power and power grid
Voltage, the variation relation between mains frequency.Therefore, new development has gone out the intelligentized load modeling side based on measurement data
Method mainly has fuzzy reasoning method, neural network, genetic algorithm at present.
These methods.It is built upon on the basis of empirical risk minimization, i.e. our usually said error criterion bases
On plinth, it is easy over-fitting in this way, so that the generalization ability of the model recognized is weaker, the discrimination method of neural network
Although there is stronger None-linear approximation ability, network structure is not easy to determine, can only be by empirically determined, and training
When be easily trapped into local minimum, fuzzy logic system discrimination method is then that its fuzzy rule is difficult to determine, genetic algorithm
It is easily trapped into local minimum, or even not convergent situation, and the method for support vector machines then considers not only empirical risk minimization
Change, while considering structural risk minimization, therefore there is stronger generalization ability.Also there is document to attempt using RBF core at present
Support vector machines establish load module, but only system is recognized with support vector machines itself, but without proposing one
The model of the parametrization of a structuring, the method that more parameter is not quantified.
Summary of the invention
The purpose of the present invention is overcoming the above-mentioned deficiency of the prior art, a kind of modeling of electric system support structures is provided
And parameter identification method, comprehensively consider empirical risk minimization and structural risk minimization, the electrical load model tool made
There is better generalization ability, can more reflect the dynamic of system, the more response characteristic of approaching to reality electric system.Technology of the invention
Scheme is as follows:
A kind of dynamic load model modelling approach using support vector machines, including the following steps:
1) according to the electric network data acquired in real time, the electric network fault data of certain power grid particular power branch, three-phase electricity are collected
Pressure amplitude value and phase, three-phase current amplitude and phase, sample frequency is greater than or equal to 2 times of network operation frequencies, and calculates accordingly
Record corresponding positive sequence voltage changes delta u, mains frequency variation delta f and corresponding active power variation delta P and idle
Power variation Δ Q;
2) cover sheet is with respect to earth fault, the data of phase fault and three phase short circuit fault, using support vector machines pair
Load module data are trained, and obtain supporting vector machine model, and the structure of setting model is that such as flowering structure, model order are set
It is set to n rank,
For active power
Δ P (k)=d1ΔP(k-1)+d2ΔP(k-2)+d3ΔP(k-3)+…+dnΔP(k-n)
+e1Δu(k)+e2Δu(k-1)+e3Δu(k-2)+e4Δu(k-3)+…+en+1Δu(k-n)
+g1Δf(k)+g2Δf(k-1)+g3Δf(k-2)+g4Δf(k-3)+…+gn+1Δf(k-n)
P (k)=P0+ΔP(k) (1)
Wherein, Δ P (k) is the active power variable quantity at k moment, and Δ u (k) is the power grid positive sequence voltage variable quantity at k moment,
Δ f (k) is the mains frequency variable quantity at k moment, P0It is initial active power, θ=(d1,d2,d3,…,dn,e1,e2,e3,…,
en+1,g1,g2,g3,…,gn+1) it is active load model undetermined parameter vector;
d1,d2,d3,…,dn,e1,e2,e3,…,en+1,g1,g2,g3,…,gn+1It is that active load model undetermined is
Number;
As a same reason, for reactive power
Δ Q (k)=d '1ΔQ(k-1)+d′2ΔQ(k-2)+d′3ΔQ(k-3)+…+d′nΔQ(k-n)
+e′1Δu(k)+e′2Δu(k-1)+e′3Δu(k-2)+e′4Δu(k-3)+…+e′n+1Δu(k-n)
+g′1Δf(k)+g′2Δf(k-1)+g′3Δf(k-2)+g′4Δf(k-3)+…+g′n+1Δf(k-n)
Q (k)=Q0+ΔQ(k) (2)
Wherein, Δ Q (k) is the active power variable quantity at k moment, Q0It is initial active power, θ '=(d '1,d′2,d
′3,…,d′n,e′1,e′2,e′3,…,e′n+1,g′1,g′2,g′3,…,g′n+1) it is reactive load model undetermined parameter vector;
Wherein, d '1,d′2,d′3,…,d′n,e′1,e′2,e′3,…,e′n+1,g′1,g′2,g′3,…,g′n+1It is undetermined
The coefficient of reactive load model;
3) respectively with
X (l)=(Δ P (k-1), Δ P (k-2), Δ P (k-3) ..., Δ P (k-n), Δ u (k), Δ u (k-1), Δ u (k-
2), Δ u (k-3) ..., Δ u (k-n), Δ f (k), Δ f (k-1), Δ f (k-2), Δ f (k-3) ..., Δ f (k-n)) and
X ' (l)=(Δ Q (k-1), Δ Q (k-2), Δ Q (k-3) ..., Δ Q (k-n), Δ u (k), Δ u (k-1), Δ u (k-
2), Δ u (k-3) ..., Δ u (k-n), Δ f (k), Δ f (k-1), Δ f (k-2), Δ f (k-3) ..., Δ f (k-n)) as defeated
Incoming vector, using T (l)=Δ P (k) and T (l)=Δ Q (k) as objective function, using linear kernel support vector machines to the load
Model is trained identification, obtains supporting vector machine model expression formula active power load module expression formula and isReactive power load module expression formula is
Wherein ypIt (k) is that the supporting vector machine model at k moment has power output, yq(k) be the k moment support vector machines mould
The output of type inactivity, t, t ' are supporting vector number, xi,x′iIt is supporting vector, x, x ' are input vectors, and b, b ' are that training obtains
Offset constant, σi,σ′iIt is the model constants vector that support vector machines training obtains;The packet of the used support vector machines of training
The constant including ε is included, can be chosen in the training process by cross validation;
4) according to model equivalency principle
Enable yp(k)=Δ P (k), yq(k)=Δ Q (k)
And handle
It is write as
So as to obtain system parameter undetermined,
If system is without biasing, usually
B=0, b '=0
The active power load differential equation model for obtaining the electrical branch as a result, is shown in formula (1) and reactive power load
Difference equation model is shown in formula (2);According to the structure of linear nuclear model and unidentified system, obtain system undetermined parameter θ and
θ ' is shown in formula (3), (4).
The present invention acquires electric power network electricity characteristic data by PMU (synchronized phase measurement device), considers the dynamic of model
Step response determines the structure and order N of model, then the input and output difference equation of constructing variable.On this basis, with this
Training data of the autoregressive inputoutput data as support vector machines carries out regression training using linear kernel, thus obtains
The input and output support vector machines system data model of the data model, then according to system model equivalence principle, according to this hair
Bright method obtains load module system parameter undetermined, thus the electrical load model completely parameterized.It is basic herein
On can also carry out frequency-domain transform according to the model and sampling period and become the transfer function model in the domain Z and the domain S.Using the party
The effect and advantage of method are that the disadvantage that can effectively avoid conventional method generalization ability weak, conventional method are often changing system
When initial value or the service condition of uniting, it is easy to appear model mismatch.The method (such as neural network) of other intelligence simultaneously
Often after identification system, since obtained model structure complexity and function are non-linear, it is difficult to which the model obtained from training pushes away
The parameter of guiding system.
Specific embodiment
The present invention will be described below
1, the electric network fault data of certain power grid particular power branch, three-phase are collected using PMU (synchronized phase measurement device)
Voltage magnitude and phase, three-phase current amplitude and phase;Calculating records corresponding positive sequence voltage changes delta u, mains frequency variation
The changes delta Q of Δ f and corresponding active power Δ P and reactive power;3) data of the electrical branch are divided into three classes, it is single
Opposite earth fault, phase fault, three phase short circuit fault;It 4) is data X (l) by the data summarization that section occurs for three classes failure
It is data amount check with data T (l), l.
2, and the structure of setting model is that such as flowering structure, model order are set as n rank.
For active power
Δ P (k)=d1ΔP(k-1)+d2ΔP(k-2)+d3ΔP(k-3)+…+dnΔP(k-n)
+e1Δu(k)+e2Δu(k-1)+e3Δu(k-2)+e4Δu(k-3)+…+en+1Δu(k-n)
+g1Δf(k)+g2Δf(k-1)+g3Δf(k-2)+g4Δf(k-3)+…+gn+1Δf(k-n)
P (k)=P0+ΔP(k) (1)
P0It is initial active power, θ=(d1,d2,d3,…,dn,e1,e2,e3,…,en+1,g1,g2,g3,…,gn+1) it is to have
Function load module undetermined parameter vector.
As a same reason, for reactive power
Δ Q (k)=d '1ΔQ(k-1)+d′2ΔQ(k-2)+d′3ΔQ(k-3)+…+d′nΔQ(k-n)
+e′1Δu(k)+e′2Δu(k-1)+e′3Δu(k-2)+e′4Δu(k-3)+…+e′n+1Δu(k-n)
+g′1Δf(k)+g′2Δf(k-1)+g′3Δf(k-2)+g′4Δf(k-3)+…+g′n+1Δf(k-n)
Q (k)=Q0+ΔQ(k) (2)
Q0It is initial active power, θ '=(d '1,d′2,d′3,…,d′n,e′1,e′2,e′3,…,e′n+1,g′1,g′2,g
′3,…,g′n+1) it is reactive load model undetermined parameter vector.
3, respectively with
X (l)=(Δ P (k-1), Δ P (k-2), Δ P (k-3) ..., Δ P (k-n), Δ u (k), Δ u (k-1), Δ u (k-
2),Δu(k-3),…,Δu(k-n),Δf(k),Δf(k-1),Δf(k-2),Δf(k-3),…,,Δf(k-n))
X ' (l)=(Δ Q (k-1), Δ Q (k-2), Δ Q (k-3) ..., Δ Q (k-n), Δ u (k), Δ u (k-1), Δ u (k-
2), Δ u (k-3) ..., Δ u (k-n), Δ f (k), Δ f (k-1), Δ f (k-2), Δ f (k-3) ..., Δ f (k-n)) as defeated
Incoming vector, using T (l)=Δ P (k) and T (l)=Δ Q (k) as objective function, using linear kernel support vector machines to the load
Model is trained identification, obtains supporting vector machine model expression formula active power load module and is
Reactive power load module isMiddle t, t ' are supporting vector number, xi,x′iIt is supporting vector, x, x '
It is input vector, b, b ' are the offset constant that training obtains, σi,σ′iIt is the model constants vector that support vector machines training obtains.
The used other constants such as ε of support vector machines etc. of training, can be chosen by cross validation in the training process.
4, according to model equivalency principle, load module parameter is obtained
Enable yp(k)=P (k), yq(k)=Q (k)
3 I, handle
It is write as
So as to obtain system parameter undetermined,
If system is without biasing, usually
B=0, b '=0
The active power load differential equation model of our the available electrical branch as a result, is shown in formula (1) and idle
Power termination difference equation model is shown in formula (2).According to the structure of linear nuclear model and unidentified system, system undetermined is derived
The parameter θ and θ ' of system are shown in formula (3), (4).
5, it obtains the difference equation model of system, while by frequency-domain transform, difference equation model is passed through transform and double
The difference equation load module is converted into transfer function model by linear transformation.
Z domain model is
Δ P (z)=G1(z)Δu(z)+G2(z) Δ f (z) Δ Q (z)=G3(z)Δu(z)+G4(z)Δf(z)
S domain model is
Δ P (s)=G1(s)Δu(s)+G2(s) Δ f (s) Δ Q (s)=G3(s)Δu(s)+G4(s)Δf(s)。
Claims (1)
1. a kind of dynamic load model modelling approach using support vector machines, including the following steps:
1) according to the electric network data acquired in real time, the electric network fault data of certain power grid particular power branch, three-phase electricity pressure amplitude are collected
Value and phase, three-phase current amplitude and phase, sample frequency are greater than or equal to 2 times of network operation frequencies, and calculate record accordingly
Corresponding positive sequence voltage changes delta u, mains frequency variation delta f and corresponding active power variation delta P and reactive power
Variation delta Q;
2) cover sheet is with respect to earth fault, the data of phase fault and three phase short circuit fault, using support vector machines to load
Model data is trained, and obtains supporting vector machine model, and the structure of setting model is that such as flowering structure, model order are set as
N rank,
For active power
Δ P (k)=d1ΔP(k-1)+d2ΔP(k-2)+d3ΔP(k-3)+…+dnΔP(k-n)+e1Δu(k)+e2Δu(k-1)+
e3Δu(k-2)+e4Δu(k-3)+…+en+1Δu(k-n)+g1Δf(k)+g2Δf(k-1)+g3Δf(k-2)+g4Δf(k-3)
+…+gn+1Δf(k-n)
P (k)=P0+ΔP(k) (1)
Wherein, Δ P (k) is the active power variable quantity at k moment, and Δ u (k) is the power grid positive sequence voltage variable quantity at k moment, Δ f
(k) be the k moment mains frequency variable quantity, P0It is initial active power, θ=(d1,d2,d3,…,dn,e1,e2,e3,…,en+1,
g1,g2,g3,…,gn+1) it is active load model undetermined parameter vector;
d1,d2,d3,…,dn,e1,e2,e3,…,en+1,g1,g2,g3,…,gn+1It is the coefficient of active load model undetermined;
As a same reason, for reactive power
Δ Q (k)=d '1ΔQ(k-1)+d′2ΔQ(k-2)+d′3ΔQ(k-3)+…+d′nΔQ(k-n)+e′1Δu(k)+e′2Δu
(k-1)+e′3Δu(k-2)+e′4Δu(k-3)+…+e′n+1Δu(k-n)+g′1Δf(k)+g′2Δf(k-1)+g′3Δf(k-2)
+g′4Δf(k-3)+…+g′n+1Δf(k-n)
Q (k)=Q0+ΔQ(k) (2)
Wherein, Δ Q (k) is the active power variable quantity at k moment, Q0It is initial active power, θ '=(d '1,d′2,d′3,…,d
′n,e′1,e′2,e′3,…,e′n+1,g′1,g′2,g′3,…,g′n+1) it is reactive load model undetermined parameter vector;
Wherein, d '1,d′2,d′3,…,d′n,e′1,e′2,e′3,…,e′n+1,g′1,g′2,g′3,…,g′n+1It is undetermined idle negative
Carry the coefficient of model;
3) respectively with
X (l)=(Δ P (k-1), Δ P (k-2), Δ P (k-3) ..., Δ P (k-n), Δ u (k), Δ u (k-1), Δ u (k-2), Δ
U (k-3) ..., Δ u (k-n), Δ f (k), Δ f (k-1), Δ f (k-2), Δ f (k-3) ..., Δ f (k-n)) and
X ' (l)=(Δ Q (k-1), Δ Q (k-2), Δ Q (k-3) ..., Δ Q (k-n), Δ u (k), Δ u (k-1), Δ u (k-2),
Δ u (k-3) ..., Δ u (k-n), Δ f (k), Δ f (k-1), Δ f (k-2), Δ f (k-3) ..., Δ f (k-n)) as input
Vector, using T (l)=Δ P (k) and T (l)=Δ Q (k) as objective function, using linear kernel support vector machines to the load mould
Type is trained identification, obtains supporting vector machine model expression formula active power load module expression formula and isReactive power load module expression formula is
Wherein, ypIt (k) is that the supporting vector machine model at k moment has power output, yq(k) be the k moment supporting vector machine model without
Power output, t, t ' are supporting vector number, xi,x′iIt is supporting vector, x, x ' are input vectors, and it is inclined that b, b ' are that training obtains
Set constant, σi,σ′iIt is the model constants vector that support vector machines training obtains;The used support vector machines of training includes ε
Constant inside can be chosen by cross validation in the training process;
4) according to model equivalency principle
Enable yp(k)=Δ P (k), yq(k)=Δ Q (k)
And handle
It is write as
So as to obtain system parameter undetermined,
If system without biasing,
B=0, b '=0
The active power load differential equation model for obtaining the electrical branch as a result, is shown in formula (1) and reactive power load differential
Equation model is shown in formula (2);According to the structure of linear nuclear model and unidentified system, the parameter θ and θ ' of system undetermined are obtained,
See formula (3), (4);
5) difference equation model of system is obtained, while by frequency-domain transform, difference equation model is passed through transform and bilinearity
The difference equation load module, is converted into transfer function model by transformation,
Z domain model is
Δ P (z)=G1(z)Δu(z)+G2(z)Δf(z)
Δ Q (z)=G3(z)Δu(z)+G4(z)Δf(z)
S domain model is
Δ P (s)=G1(s)Δu(s)+G2(s)Δf(s)
Δ Q (s)=G3(s)Δu(s)+G4(s)Δf(s)。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610878488.7A CN106484994B (en) | 2016-09-30 | 2016-09-30 | A kind of dynamic load model modelling approach using support vector machines linear kernel |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610878488.7A CN106484994B (en) | 2016-09-30 | 2016-09-30 | A kind of dynamic load model modelling approach using support vector machines linear kernel |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106484994A CN106484994A (en) | 2017-03-08 |
CN106484994B true CN106484994B (en) | 2019-09-10 |
Family
ID=58268640
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610878488.7A Active CN106484994B (en) | 2016-09-30 | 2016-09-30 | A kind of dynamic load model modelling approach using support vector machines linear kernel |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106484994B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109816555A (en) * | 2019-01-30 | 2019-05-28 | 云南电网有限责任公司电力科学研究院 | A kind of load modeling method based on support vector machines |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103605757B (en) * | 2013-11-25 | 2017-02-08 | 国家电网公司 | High-speed rail power quality data sorting method based on SVM (support vector machine) |
EP3011895B1 (en) * | 2014-10-26 | 2021-08-11 | Tata Consultancy Services Limited | Determining cognitive load of a subject from electroencephalography (EEG) signals |
-
2016
- 2016-09-30 CN CN201610878488.7A patent/CN106484994B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN106484994A (en) | 2017-03-08 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Khazraj et al. | A performance comparison between extended Kalman Filter and unscented Kalman Filter in power system dynamic state estimation | |
Tong et al. | Adaptive fuzzy decentralized output stabilization for stochastic nonlinear large-scale systems with unknown control directions | |
Alonge et al. | Extended complex Kalman filter for sensorless control of an induction motor | |
Zhang et al. | Enhanced robust fractional order proportional-plus-integral controller based on neural network for velocity control of permanent magnet synchronous motor | |
Sengamalai et al. | Three phase induction motor drive: a systematic review on dynamic modeling, parameter estimation, and control schemes | |
Zhang et al. | Measurement-based power system dynamic model reductions | |
Qing et al. | Extended set-membership filter for power system dynamic state estimation | |
Ye et al. | Strong robust and optimal chaos control for permanent magnet linear synchronous motor | |
CN106208879A (en) | A kind of method controlling asynchronous motor frequency-changing speed-regulating system based on automatic disturbance rejection controller | |
Kim et al. | Parameter estimation for the composite load model | |
CN106484994B (en) | A kind of dynamic load model modelling approach using support vector machines linear kernel | |
Nagamune et al. | Parameter reduction in estimated model sets for robust control | |
Goos et al. | Estimation of linear parameter-varying affine state space models using synchronized periodic input and scheduling signals | |
Liu et al. | A novel control-performance-oriented data-driven fault classification approach | |
Sun et al. | System identification for nonlinear FOPDT model with input-dependent dead-time | |
Malik | Reduced Order Modeling for Smart Grids’ Simulation and Optimization | |
Rahman et al. | Scalable cellular computational network based WLS state estimator for power systems | |
Pareek et al. | Small-signal stability constrained optimal power flow: A convexification approach | |
Rigatos et al. | Nonlinear optimal control for synchronous reluctance machines | |
Filip et al. | Givens Orthogonal Transformation-Based Estimator Versus RLS Estimator—Case Study for an Induction Generator Model | |
Liu et al. | Fuzzy equivalence relation clustering-based algorithm for coherency identification among generators | |
Oh | Analytical solution to swing equations in power grids | |
Pakzad et al. | Stability map of fractional delay systems in the parametric space of delays and coefficient | |
Jiang et al. | Measurement-based power system dynamic model reduction using ARX equivalents | |
Rasouli et al. | An efficient approach for measurement-based composite load modeling |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | 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 |