CN109061299A - A kind of trend of harmonic detection method of power based on radial basis function neural network - Google Patents
A kind of trend of harmonic detection method of power based on radial basis function neural network Download PDFInfo
- Publication number
- CN109061299A CN109061299A CN201810627698.8A CN201810627698A CN109061299A CN 109061299 A CN109061299 A CN 109061299A CN 201810627698 A CN201810627698 A CN 201810627698A CN 109061299 A CN109061299 A CN 109061299A
- Authority
- CN
- China
- Prior art keywords
- output
- layer
- hidden layer
- weight vector
- neural network
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R23/00—Arrangements for measuring frequencies; Arrangements for analysing frequency spectra
- G01R23/16—Spectrum analysis; Fourier analysis
Abstract
The invention discloses a kind of trend of harmonic detection method of power based on radial basis function neural network, the following steps are included: (1) acquires current signal to be detected, obtain RBF neural expectation input signal, determine neural network input layer, hidden layer and output layer number of nodes, weight vector, the central value of Gaussian function and width are initialized, determines learning efficiency and dynamic factor；(2): calculating hidden layer output；(3) output layer reality output is calculated；(4) calculation of performance indicators；(5) the more new formula of weight vector, the central value of Gaussian function and width is obtained according to performance indicator gradient descent method；(6) weight vector, the central value of Gaussian function and width are updated；(7) judge whether performance indicator meets area requirement, do not meet such as and execute step (2)；Such as meet and executes step (8)；(8) amplitude, frequency and the phase of harmonic wave are calculated.The present invention have the advantages that can more rapidly, more accurately detect harmonic current.
Description
Technical field
It is especially a kind of based on radial basis function neural network the present invention relates to power quality dynamic detection technology field
Trend of harmonic detection method of power.
Background technique
In electric system, theoretic voltage and current waveform is the sine wave under power frequency, but actual waveform always has not
Same non-sine distortion.Generally acknowledged definition of harmonic in the world are as follows: " harmonic wave is the component sine waves of a cycle electrical quantity, frequency
Rate is the integral multiple of fundamental wave ".In the power system, it is fundamental frequency integral multiple that our usually said harmonic waves, which are primarily referred to as frequency,
Sine wave, also commonly referred to as higher hamonic wave.
Harmonic wave mainly has the following aspects to the harm of utility network and other systems: (1) harmonic wave makes in utility network
Equipment generate additional power loss, reduce the efficiency of power generation, transmission of electricity and electrical equipment；(2) various electrical equipments are influenced
It works normally；(3) harmonic wave will lead to relay protection and the malfunction of automatic control device or tripping, and make the meter of electrical measuring instrument
Amount inaccuracy；(4) harmonic wave can generate interference to neighbouring communication system, and less serious case generates noise, reduces communication quality；Severe one causes
Letter is lost, and can not work normally communication system；(5) harmonic wave can cause parallel resonance local in utility network and connect humorous
Vibration, to make Harmonics amplification, the harm of several aspects is greatly increased before this just makes, or even causes major accident.
Increasingly extensive situation is endangered in face of electric harmonic, governing problem just seems very necessary, the premise of governing problem
It is to find the problem, that is, detects harmonic wave.
Traditional harmonic detecting has Fourier transformation, Short Time Fourier Transform, ip-iqMethod, but Fourier transformation detects
There are some problems, such as non-synchronous sampling, spectral leakage and fence effect etc., i for harmonic wavep-iqMethod can also occur in the detection process
Lag and error problem.Since artificial neural network has certain self study and adaptive ability, even if occurring unknown
Information, as long as unknown message is similar with training set, neural network equally has very strong analysis processing capacity for this unknown message,
It thus needs to invent a kind of harmonic wave rapid detection method based on radial base (RBF) neural network.
Summary of the invention
The object of the present invention is to provide one kind can more rapidly, more accurately detect harmonic current based on radial base letter
The trend of harmonic detection method of power of number neural network.
To achieve the above object, present invention employs following technical solutions: described one kind is based on Radial Basis Function neural
The trend of harmonic detection method of power of network, comprising the following steps:
Step 1: acquiring current signal to be detected, obtains the expectation input signal of RBF neural, defeated according to it is expected
Enter signal and determine the input layer number, node in hidden layer and output layer number of nodes of RBF neural, at the same initialize power to
Amount, the central value of Gaussian function and width parameter, and determine learning efficiency and dynamic factor；
Step 2: Gaussian bases are determined as to the radial basis function of hidden layer in RBF neural, and according to Gauss
Function calculates the output of hidden layer；
Step 3: the reality output of output layer is calculated according to the output of hidden layer and weight vector；
Step 4: performance indicator is calculated according to the reality output of output layer and desired output；
Step 5: according to the central value of the more new formula of performance indicator gradient descent method acquisition weight vector, Gaussian function
The more new formula of the width of more new formula and Gaussian function；
Step 6: the more new formula of the weight vector obtained according to step 5, the central value of Gaussian function more new formula, with
And the more new formula of the width of Gaussian function updates weight vector, the central value of Gaussian function and width parameter；
Step 7: whether the performance indicator that judgment step four obtains meets area requirement, does not meet area requirement such as, then jumps
Turn to execute step 2；Such as meet area requirement, thens follow the steps eight；
Step 8: updated weight vector, the central value of Gaussian function and the width parameter obtained according to step 6 calculates
The amplitude of harmonic wave, frequency and phase in current signal out.
Further, a kind of trend of harmonic detection method of power based on radial basis function neural network above-mentioned, in which:
In step 2, the specific formula for calculation of the output of hidden layer are as follows:
H=[h1 h2...hj hn]T, wherein
Wherein, x indicates the current input signal of acquisition, | | | | indicate Euclid norm, cjIndicate hidden layer to mind
Vector distance through center, bjIt is neuron node base width parameter.
Further, a kind of trend of harmonic detection method of power based on radial basis function neural network above-mentioned, in which:
In step 3, the specific formula for calculation of the reality output of output layer are as follows:
ym(k)=wh=w1h1+w2h2+...+wmhm
Wherein, w indicates weight vector of the hidden layer to output layer, the output of h expression hidden layer.
Further, a kind of trend of harmonic detection method of power based on radial basis function neural network above-mentioned, in which:
In step 4, the specific formula for calculation of performance indicator are as follows:
Wherein, y (k) indicates that the desired output of output layer, ym (k) indicate the reality output of output layer.
Further, a kind of trend of harmonic detection method of power based on radial basis function neural network above-mentioned, in which:
In step 5, the more new formula of weight vector are as follows:
wj(k)=wj(k-1)+η(y(k)-ym(k))hj+α(wj(k-1)+wj(k-2))
Wherein, y (k) indicate output layer desired output, ym (k) indicate output layer reality output, α indicate dynamic because
Son, η indicate learning efficiency, hjIndicate the output of hidden layer.
Further, a kind of trend of harmonic detection method of power based on radial basis function neural network above-mentioned, in which:
In step 5, the more new formula of the central value of Gaussian function are as follows:
Wherein, y (k) indicates the desired output of output layer, ym(k) indicate output layer reality output, α indicate dynamic because
Son, η indicate learning efficiency, wjIndicate weight vector of the hidden layer to output layer, hjIndicate the output of hidden layer, xjIndicate input letter
Number, bjIt is neuron node base width parameter.
Further, a kind of trend of harmonic detection method of power based on radial basis function neural network above-mentioned, in which:
In step 5, the more new formula of the width of Gaussian function are as follows:
Wherein, y (k) indicates the desired output of output layer, ym(k) indicate output layer reality output, α indicate dynamic because
Son, η indicate learning efficiency, wjIndicate weight vector of the hidden layer to output layer, hjIndicate the output of hidden layer, X indicates input letter
Number, cjIndicate vector distance of the hidden layer to nerve center, bjIt is neuron node base width parameter.
Through the implementation of the above technical solution, the beneficial effects of the present invention are: (1) RBF neural is before one kind to type
Neural network, signal are to enter neural network by input layer, enter hidden layer by signal after input layer, and input layer is to implicit
Layer is a kind of Nonlinear Mapping, then overcomes detection present in conventional harmonic detection method to output layer by Linear Mapping
Occur lag in the process and the problems such as error, can more rapidly, more accurately detect harmonic current, for the analysis of subsequent harmonic wave
Improvement provides effective foundation；(2) calculation amount of weight vector more new formula is small, and tracing property is good.
Detailed description of the invention
Fig. 1 is a kind of stream of the trend of harmonic detection method of power based on radial basis function neural network of the present invention
Cheng Tu.
Fig. 2 is in a kind of trend of harmonic detection method of power based on radial basis function neural network of the present invention
The topological diagram of RBF neural.
Specific embodiment
Technical solution of the present invention is described in further detail with reference to the accompanying drawing:
As shown in Figure 1 and Figure 2, a kind of trend of harmonic detection method of power based on radial basis function neural network,
The following steps are included:
Step 1: acquiring current signal to be detected, obtains the expectation input signal of RBF neural, defeated according to it is expected
Enter signal and determine the input layer number, node in hidden layer and output layer number of nodes of RBF neural, at the same initialize power to
Amount, the central value of Gaussian function and width parameter, and determine learning efficiency and dynamic factor；Wherein, initialization weight vector is w
=[w1 w2…wj wm]T, the central value c of Gaussian functionijWith width parameter bj> 0, j=1,2 ... m；
Step 2: Gaussian bases are determined as to the radial basis function of hidden layer in RBF neural, and according to Gauss
Function calculates the output of hidden layer；Wherein, the specific formula for calculation of the output of hidden layer are as follows:
H=[h1 h2...hj hn]T, wherein
Wherein, x indicates input signal, | | | | indicate Euclid norm, cjArrow of the expression hidden layer to nerve center
Span is from bjIt is neuron node base width parameter；
Wherein, cj=[c1j c2j...cij...cnj]T, wherein i=1,2 ... n, j=1,2 ... m；RBF neural
Sound stage width phasor be set as b, expression formula is b=[b1 b2...bj bm]T, bj> 0, j=1,2 ..., m；
Step 3: the reality output of output layer is calculated according to the output of hidden layer and weight vector；Wherein, the reality of output layer
The specific formula for calculation of border output are as follows:
ym(k)=wh=w1h1+w2h2+...+wmhm
Wherein, w indicates weight vector of the hidden layer to output layer, hjIndicate the output of hidden layer；
Step 4: performance indicator is calculated according to the reality output of output layer and desired output；Wherein, the tool of performance indicator
Body calculation formula are as follows:
Wherein, y (k) indicates the desired output of output layer, ym(k) reality output of output layer is indicated；
Step 5: according to the central value of the more new formula of performance indicator gradient descent method acquisition weight vector, Gaussian function
The more new formula of the width of more new formula and Gaussian function；
Wherein, the more new formula of weight vector are as follows:
wj(k)=wj(k-1)+η(y(k)-ym(k))hj+α(wj(k-1)+wj(k-2))
Wherein, y (k) indicates the desired output of output layer, ym(k) indicate output layer reality output, α indicate dynamic because
Son, η indicate learning efficiency, hjIndicate the output of hidden layer；
In the present embodiment, the specific derivation process of the more new formula of weight vector is as follows:
Firstly, more new algorithm, that is, passage capacity index E (k) of weight vector is to weight vector wj(k) local derviation is sought
WhereinCan simplify for,
It is available by formula (1), (2)
Weight w is corrected using above formula local derviationj(k), Δ w is obtainedj(k) expression formula:
Wherein, η is learning rate, and to prevent RBF from falling into Local Minimum, dynamic factor component is added in weight vector update
α(wj(k-1)-wj(k-2)), by Δ wj(k) and dynamic factor component derives weight vector wj(k) more new formula:
wj(k)=wj(k-1)+η(y(k)-ym(k))hj+α(wj(k-1)+wj(k-2))
Wherein, the more new formula of the central value of Gaussian function are as follows:
Wherein, wherein y (k) indicates the desired output of output layer, ym(k) reality output of output layer is indicated, α indicates dynamic
The state factor, η indicate learning efficiency, wjIndicate weight vector of the hidden layer to output layer, hjIndicate the output of hidden layer, xjIndicate defeated
Enter signal, bjIt is neuron node base width parameter；
Wherein, the more new formula of the width of Gaussian function are as follows:
Wherein, y (k) indicates the desired output of output layer, ym(k) indicate output layer reality output, α indicate dynamic because
Son, η indicate learning efficiency, wjIndicate weight vector of the hidden layer to output layer, hjIndicate the output of hidden layer, X indicates input letter
Number, cjIndicate vector distance of the hidden layer to nerve center, bjIt is neuron node base width parameter；
In the present embodiment, the tool of the more new formula of the width of the more new formula and Gaussian function of the central value of Gaussian function
Body derivation process is as follows:
Passage capacity index E (k) first is to base width parameter bjLocal derviation is sought, derivation process is as follows:
WhereinIt can be with abbreviation are as follows:
WhereinCan simplify for,
It is available by formula (6), (7) and (8)
Utilize local derviationTo correct width parameter bj(k), Δ b is obtainedj(k) expression formula
In width parameter bj(k) factor of momentum component α (b is added in renewal processj(k-1)-bj(k-2)), by Δ bjWith it is dynamic
Amount factor component derives the width parameter b of Gaussian functionj(k) more new formula:
Further according to performance indicator E (k) to the central value c of Gaussian functionijLocal derviation is sought, derivation process is as follows,
In formula,It can abbreviation are as follows:
By formula (11), (12), that (13) can obtain local derviation is as follows:
Center c is corrected using local derviationij, obtain Δ cijExpression formula are as follows:
In the central value c of Gaussian functionijDynamic factor component α (c is inserted into renewal processij(k-1)-cij(k-2)), then
By Δ cijThe central value more new formula of Gaussian function is acquired with dynamic factor component are as follows:
Step 6: the more new formula of the weight vector obtained according to step 5, the central value of Gaussian function more new formula, with
And the more new formula of the width of Gaussian function updates weight vector, the central value of Gaussian function and width parameter；
Step 7: whether the performance indicator that judgment step four obtains meets area requirement, does not meet area requirement such as, then jumps
Turn to execute step 2；Such as meet area requirement, thens follow the steps eight；
Step 8: updated weight vector, the central value of Gaussian function and the width parameter obtained according to step 6 calculates
The amplitude of harmonic wave, frequency and phase in current signal out.
The invention has the advantages that (1) RBF neural is a kind of Feed-forward neural networks, signal be by input layer into
Enter neural network, hidden layer is entered by signal after input layer, input layer to hidden layer is a kind of Nonlinear Mapping, then passes through line
Property be mapped to output layer, overcome in detection process present in conventional harmonic detection method and lag occur and the problems such as error,
Can more rapidly, more accurately detect harmonic current, for subsequent harmonic wave analysis improvement provide effective foundation；(2) weigh to
The calculation amount for measuring more new formula is small, and tracing property is good.
Claims (7)
1. a kind of trend of harmonic detection method of power based on radial basis function neural network, which is characterized in that including following step
It is rapid:
Step 1: acquiring current signal to be detected, obtains the expectation input signal of RBF neural, inputs letter according to expectation
Number determine RBF neural input layer number, node in hidden layer and output layer number of nodes, while initialize weight vector,
The central value and width parameter of Gaussian function, and determine learning efficiency and dynamic factor；
Step 2: Gaussian bases are determined as to the radial basis function of hidden layer in RBF neural, and according to Gaussian function
Calculate the output of hidden layer；
Step 3: the reality output of output layer is calculated according to the output of hidden layer and weight vector；
Step 4: performance indicator is calculated according to the reality output of output layer and desired output；
Step 5: the update of the more new formula of weight vector, the central value of Gaussian function is obtained according to performance indicator gradient descent method
The more new formula of the width of formula and Gaussian function；
Step 6: more new formula, the Yi Jigao of the more new formula of the weight vector obtained according to step 5, the central value of Gaussian function
The more new formula of the width of this function updates weight vector, the central value of Gaussian function and width parameter；
Step 7: whether the performance indicator that judgment step four obtains meets area requirement, does not meet area requirement such as, then jumps and hold
Row step 2；Such as meet area requirement, thens follow the steps eight；
Step 8: updated weight vector, the central value of Gaussian function and the width parameter obtained according to step 6 calculates electricity
Flow amplitude, frequency and the phase of harmonic wave in signal.
2. a kind of trend of harmonic detection method of power based on radial basis function neural network according to claim 1,
It is characterized in that: in step 2, the specific formula for calculation of the output of hidden layer are as follows: h=[h1 h2...hj hn]T, wherein
Wherein, x indicates the current input signal of acquisition, | | | | indicate Euclid norm, cjIndicate hidden layer to nerve center
Vector distance, bjIt is neuron node base width parameter.
3. a kind of trend of harmonic detection method of power based on radial basis function neural network according to claim 1,
It is characterized in that: in step 3, the specific formula for calculation of the reality output of output layer are as follows:
ym(k)=wh=w1h1+w2h2+...+wmhm
Wherein, w indicates weight vector of the hidden layer to output layer, the output of h expression hidden layer.
4. a kind of trend of harmonic detection method of power based on radial basis function neural network according to claim 1,
It is characterized in that: in step 4, the specific formula for calculation of performance indicator are as follows:
Wherein, y (k) indicates the desired output of output layer, ym(k) reality output of output layer is indicated.
5. a kind of trend of harmonic detection method of power based on radial basis function neural network according to claim 1,
It is characterized in that: in step 5, the more new formula of weight vector are as follows:
wj(k)=wj(k-1)+η(y(k)-ym(k))hj+α(wj(k-1)+wj(k-2))
Wherein, y (k) indicates the desired output of output layer, ym(k) reality output of output layer is indicated, α indicates dynamic factor, η table
Show learning efficiency, hjIndicate the output of hidden layer.
6. a kind of trend of harmonic detection method of power based on radial basis function neural network according to claim 1,
It is characterized in that: in step 5, the more new formula of the central value of Gaussian function are as follows:
Wherein, y (k) indicates the desired output of output layer, ym(k) reality output of output layer is indicated, α indicates dynamic factor, η table
Show learning efficiency, wjIndicate weight vector of the hidden layer to output layer, hjIndicate the output of hidden layer, xjIndicate input signal, bjIt is
Neuron node base width parameter.
7. a kind of trend of harmonic detection method of power based on radial basis function neural network according to claim 1,
It is characterized in that: in step 5, the more new formula of the width of Gaussian function are as follows:
Wherein, y (k) indicates the desired output of output layer, ym(k) reality output of output layer is indicated, α indicates dynamic factor, η table
Show learning efficiency, wjIndicate weight vector of the hidden layer to output layer, hjIndicate the output of hidden layer, X indicates input signal, cjTable
Show vector distance of the hidden layer to nerve center, bjIt is neuron node base width parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810627698.8A CN109061299A (en) | 2018-06-15 | 2018-06-15 | A kind of trend of harmonic detection method of power based on radial basis function neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810627698.8A CN109061299A (en) | 2018-06-15 | 2018-06-15 | A kind of trend of harmonic detection method of power based on radial basis function neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109061299A true CN109061299A (en) | 2018-12-21 |
Family
ID=64820443
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810627698.8A Pending CN109061299A (en) | 2018-06-15 | 2018-06-15 | A kind of trend of harmonic detection method of power based on radial basis function neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109061299A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110222743A (en) * | 2019-05-23 | 2019-09-10 | 山西金晖隆开关有限公司 | The method for generating plan of power grid accident |
| CN114157191A (en) * | 2021-11-17 | 2022-03-08 | 国网湖北省电力有限公司电力科学研究院 | Permanent magnet synchronous motor optimal efficiency control method based on online gradient descent method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006098287A (en) * | 2004-09-30 | 2006-04-13 | Yokogawa Electric Corp | Harmonic component measuring apparatus |
| CN102243272A (en) * | 2011-04-01 | 2011-11-16 | 重庆大学 | High-precision harmonic analysis method of sampling data synchronization |
| CN103066603A (en) * | 2013-01-25 | 2013-04-24 | 河海大学常州校区 | Self-adaptive radial basis function (RBF) neural sliding mode control method for three-phase shunt active power filter |
| CN105610162A (en) * | 2016-01-04 | 2016-05-25 | 河海大学常州校区 | Adaptive fuzzy sliding mode RBF neural network control method for active power filter |
| CN106779053A (en) * | 2016-12-15 | 2017-05-31 | 福州瑞芯微电子股份有限公司 | The knowledge point of a kind of allowed for influencing factors and neutral net is known the real situation method |
-
2018
- 2018-06-15 CN CN201810627698.8A patent/CN109061299A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006098287A (en) * | 2004-09-30 | 2006-04-13 | Yokogawa Electric Corp | Harmonic component measuring apparatus |
| CN102243272A (en) * | 2011-04-01 | 2011-11-16 | 重庆大学 | High-precision harmonic analysis method of sampling data synchronization |
| CN103066603A (en) * | 2013-01-25 | 2013-04-24 | 河海大学常州校区 | Self-adaptive radial basis function (RBF) neural sliding mode control method for three-phase shunt active power filter |
| CN105610162A (en) * | 2016-01-04 | 2016-05-25 | 河海大学常州校区 | Adaptive fuzzy sliding mode RBF neural network control method for active power filter |
| CN106779053A (en) * | 2016-12-15 | 2017-05-31 | 福州瑞芯微电子股份有限公司 | The knowledge point of a kind of allowed for influencing factors and neutral net is known the real situation method |
Non-Patent Citations (3)
| Title |
|---|
| 伍凯 等: "三种RBF网络函数逼近性能对比及应用研究", 《火力与指挥控制》 * |
| 杨磊: "基于神经网络的电力系统谐波检测", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
| 葛明臣 等: "基于RBF神经网络的谐波电流检测方法", 《煤矿机械》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110222743A (en) * | 2019-05-23 | 2019-09-10 | 山西金晖隆开关有限公司 | The method for generating plan of power grid accident |
| CN114157191A (en) * | 2021-11-17 | 2022-03-08 | 国网湖北省电力有限公司电力科学研究院 | Permanent magnet synchronous motor optimal efficiency control method based on online gradient descent method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Jain et al. | Low-order dominant harmonic estimation using adaptive wavelet neural network | |
| Cobreces et al. | Grid impedance monitoring system for distributed power generation electronic interfaces | |
| CN107016236B (en) | Power grid false data injection attack detection method based on nonlinear measurement equation | |
| Kanna et al. | Distributed widely linear Kalman filtering for frequency estimation in power networks | |
| Khoshkhoo et al. | On-line dynamic voltage instability prediction based on decision tree supported by a wide-area measurement system | |
| CN103941163A (en) | Resonant earthed system fault line selection method utilizing fuzzy K-means clustering | |
| Reddy et al. | Fast tracking of power quality disturbance signals using an optimized unscented filter | |
| Dash et al. | Dynamic phasor and frequency estimation of time-varying power system signals | |
| CN104616061B (en) | Island detection method based on wavelet packet logarithmic energy entropy and genetic algorithm optimization | |
| CN110687399B (en) | Method for judging waveform fault starting time of power distribution network fault indicator | |
| Zhang et al. | Cyber-attack detection for photovoltaic farms based on power-electronics-enabled harmonic state space modeling | |
| CN108918931A (en) | Power signal adaptive filter method in a kind of load decomposition | |
| Zhao et al. | Recent development of frequency estimation methods for future smart grid | |
| CN109061299A (en) | A kind of trend of harmonic detection method of power based on radial basis function neural network | |
| Shair et al. | Monitoring of subsynchronous oscillation in a series‐compensated wind power system using an adaptive extended Kalman filter | |
| CN114167134B (en) | Characteristic signal intensity detection method under condition of data loss of low-voltage distribution network | |
| CN109635430B (en) | Power grid transmission line transient signal monitoring method and system | |
| CN106019043B (en) | Power grid fault diagnosis method based on fuzzy integral multi-source information fusion theory | |
| Sun et al. | High-refresh-rate robust state estimation based on recursive correction for large-scale power systems | |
| CN113162002A (en) | Direct current traveling wave protection method and system considering broadband measurement link | |
| CN116699319A (en) | Fault positioning method for grid type submarine observation network direct current power supply system | |
| Sedghi et al. | On the conditional mutual information in the Gaussian–Markov structured grids | |
| CN103618309A (en) | Method and system for adjusting error parameters in power grid | |
| Wang et al. | Stockwell‐transform and random‐forest based double‐terminal fault diagnosis method for offshore wind farm transmission line | |
| CN108918929A (en) | Power signal adaptive filter method in a kind of load decomposition |
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: 20181221 |
|
| RJ01 | Rejection of invention patent application after publication |