CN110133396A - A kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD - Google Patents
A kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD Download PDFInfo
- Publication number
- CN110133396A CN110133396A CN201910280687.1A CN201910280687A CN110133396A CN 110133396 A CN110133396 A CN 110133396A CN 201910280687 A CN201910280687 A CN 201910280687A CN 110133396 A CN110133396 A CN 110133396A
- Authority
- CN
- China
- Prior art keywords
- resident
- recursion
- points
- svd
- electric switch
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R19/00—Arrangements for measuring currents or voltages or for indicating presence or sign thereof
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R21/00—Arrangements for measuring electric power or power factor
- G01R21/001—Measuring real or reactive component; Measuring apparent energy
- G01R21/002—Measuring real component
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
Abstract
The present invention relates to a kind of resident's electric switch event detecting methods decomposed based on two points of recursion SVD, comprising the following steps: (1) data acquire；(2) it carries out two points of recursion SVD to total active power data of resident's electric appliance to decompose, by calculating singular entropy with the best Decomposition order of determination；(3) switching time of resident's electric appliance is obtained using the filtering of hard -threshold filter function；The present invention is based on two points of recursion SVD resident's electric switch events decomposed to detect, and advantage is that simple to operation using method and calculating speed is fast, reduces missing inspection, false detection rate, improves detection accuracy, and testing result is stablized.
Description
Technical field
The present invention relates to non-intrusion type load monitoring technical fields, are decomposed more particularly, to one kind based on two points of recursion SVD
Resident's electric switch event detecting method.
Background technique
With China's urbanization propulsion and economic restructuring, resident living power utility Liang Zhan whole society total electricity consumption ratio is got over
Come bigger.The electricity consumption data of resident can both help user to adjust electricity consumption behavior, play an active part in demand response, and can help electricity
The scheduling of power company reasonable arrangement, peak load shifting.Load monitoring technology is to obtain the important channel of residential electricity consumption data, can be divided into
Two kinds of " intrusive mood " and " non-intrusion type ", traditional intrusive monitoring technology need for each electrical equipment install acquisition device and
Sensing device consumes a large amount of man power and materials.The 1980s, professor Hart proposed non-intrusion type load monitoring (NILM)
Technology, the technology only acquire total load information at user's house lead in end, can pass through load decomposition without installing a large amount of monitoring devices
Obtain the power consumption information of indoor each electric appliance.With the raising of intelligent level, which gradually becomes studies in China hot spot.
Non-intrusion type load monitoring technology include on-load switch event detection, feature extraction, load identify three links,
It is important prerequisite and the basis for realizing non-intrusion type load monitoring that switch events, which accurately detect,.Resident's electric switch event is examined at present
Survey method mainly has bilateral sliding window CUSUM detection of change-point method, the wavelet-decomposing method of imparametrization.Sliding window CUSUM becomes
Point detecting method can detect that the generation moment of electric switch event by the variation of active power in detection sliding window, but
It is the power sequence sliding window and threshold value that this method is limited to setting, easily there is missing inspection and erroneous detection.Wavelet-decomposing method is logical
It crosses and wavelet decomposition is carried out to power signal, detect the position of catastrophe point in active power signal, but this method is difficult to select
Suitable wavelet basis, and detect position and easily shift, detection accuracy is not high.
Summary of the invention
It is an object of the present invention to overcome the above-mentioned drawbacks of the prior art and provide one kind to be based on two points of recursion
Resident's electric switch event detecting method that SVD is decomposed improves resident's electric switch event in non-intrusion type load monitoring technology
Detection accuracy.
The purpose of the present invention can be achieved through the following technical solutions:
A kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD, comprising the following steps:
Step 1: data acquisition: electric power register one's residence acquisition residential households power information data and obtain residential households
Total active power data；
Step 2: decomposing total active power data using two points of recursion SVD: to resident's electric switch collected in step 1
The total active power data of event carry out Hankel matrixing, then carry out two points of recursion SVD and decompose, by calculating singular entropy to obtain
To the optimal Decomposition number of plies；
Step 3: the switching time of resident's electric appliance is obtained using hard -threshold filter function: using hard -threshold filter function to step
Detail signal under the optimal Decomposition number of plies obtained in rapid 2 is filtered, and realizes the accurate of catastrophe point position in total active power
Detection, the i.e. switching time of resident's electric appliance.
Preferably, the power information data in the step 1 include the total current of resident's electric appliance, voltage data, the residence
Total active power data in home front yard are obtained by MATLAB and as the target objects of switch events detection.
Preferably, the step 2 include it is following step by step:
Step 21: construction is directed to the two-dimentional Hankel matrix of total active power data；
Step 22: using two points of recursion SVD decompose after two dimension Hankel matrixes by inverse transformation obtain approximate signal A with carefully
Save signal D；
Step 23: repeating step 22 for original total active power data and be decomposed into a series of approximate and detail signal；
Step 24: calculating unusual entropy production for a series of detail signals obtained in step 23.
Preferably, the step 3 include it is following step by step:
Step 31: taking optimal value: if the unusual entropy production of judgement is greater than the set value, new Hankel is constructed using approximate signal
Matrix repeats to decompose, and until different entropy production is less than setting value, Decomposition order is optimal；
Step 32: resident's electric appliance is obtained after being filtered using hard threshold function to the detail signal under the optimal Decomposition number of plies
Switching time.
Preferably, the two-dimentional Hankel matrix H in the step 21 are as follows:
In formula, [x1x2x3…xn] it is the corresponding one-dimensional discrete signal of the total active power data of resident's electric appliance, H ∈ R2×n-1。
Preferably,
Preferably, the two-dimentional Hankel matrix H in the step 22 after two points of recursion SVD are decomposed, describes formula are as follows:
H=USVT
In formula, two-dimensional quadrature matrix U=(u1, u2), U ∈ R2×2；V=(v1,v2,…vn-1), n-1 ties up orthogonal matrix V ∈ R(n-1)×(n-1)；Diagonal matrix S=(diag (λ1,λ2), O), S ∈ R2×(n-1), λ1>>λ2, λ1And λ2Respectively first and second is unusual
Value.
Preferably, the hard threshold function in the step 32 are as follows:
In formula, d is hard threshold function, dkFor variable.
Compared with prior art, the invention has the following advantages that
One, two points of recursion SVD Orthogonal Decompositions are carried out on different levels space, and best point is determined by calculating singular entropy
The number of plies is solved, the mutation point feature in original signal has been embodied with the detail signal in different levels space.
Two, the small noise that detail signal is filtered out using hard threshold function, realizes the position of catastrophe point in total active power
Accurate detection is set, and this method is simple to operation and calculating speed is fast, the small noise in detail signal can be effectively eliminated.
Three, based on laboratory data and REDD data under different noise conditions, with existing wavelet-decomposing method and nonparametric
After the bilateral sliding CUSUM detection of change-point method changed compares, it was demonstrated that the technology of the present invention missing inspection, false detection rate are low, by noise shadow
Sound is small, and detection accuracy is high.
Detailed description of the invention
Fig. 1 is a kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD provided by the invention
Flow chart；
Fig. 2 is approximate matrix H in the present invention1Vector sum detail matrices H2Polar plot, wherein Fig. 2 (a) is approximate square
Battle array H1Polar plot, Fig. 2 (b) be detail matrices H2Polar plot；
Fig. 3 is the multiple dimensioned SVD exploded view of active power in the present invention；
Fig. 4 is single linear invariable power electric switch event detection outcome comparison diagram in the present invention, wherein Fig. 4 (a) is drink
Water dispenser active power curves figure, Fig. 4 (b) are 0.86s~0.98s of the bilateral CUSUM detection of change-point method of sliding of imparametrization
Testing result figure, Fig. 4 (c) is the detection knot of 3.18s~3.3s of the bilateral CUSUM detection of change-point method of sliding of imparametrization
Fruit figure, Fig. 4 (d) are the testing result figure of 0.82s~0.98s of wavelet-decomposing method, and Fig. 4 (e) is wavelet-decomposing method
The testing result figure of 3.16s~3.32s, Fig. 4 (f) are the testing result figure of the 0.82s~0.98s of the method for the present invention, Fig. 4 (g)
For the testing result figure of the 3.14s~3.3s of the method for the present invention；
Fig. 5 is single non-linear Variable power electric switch event detection outcome comparison diagram in the present invention, wherein Fig. 5 (a) is
Micro-wave oven active power curves figure, Fig. 5 (b) are the testing result figure of the bilateral CUSUM detection of change-point method of sliding of imparametrization,
Fig. 5 (c) is the testing result figure of wavelet-decomposing method, and Fig. 5 (d) is the testing result figure of the method for the present invention；
Fig. 6 is multiple electric switch event detection outcome comparison diagrams in the experiment in the embodiment of the present invention, wherein Fig. 6 (a)
For multiple electrical total active power curves figures, Fig. 6 (b) is the detection of the bilateral CUSUM detection of change-point method of sliding of imparametrization
Result figure, Fig. 6 (c) are the testing result figure of wavelet-decomposing method, and Fig. 6 (d) is the testing result figure of the method for the present invention；
Fig. 7 is REDD data set switch events testing result comparison diagram in the embodiment of the present invention, wherein Fig. 7 (a) is REDD
Data power curve graph, Fig. 7 (b) are the testing result figure of the bilateral CUSUM detection of change-point method of sliding of imparametrization, Fig. 7 (c)
For the testing result figure of wavelet-decomposing method, Fig. 7 (d) is the testing result figure of the method for the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiment is a part of the embodiments of the present invention, rather than whole embodiments.Based on this hair
Embodiment in bright, those of ordinary skill in the art's every other reality obtained without making creative work
Example is applied, all should belong to the scope of protection of the invention.
Embodiment
The flow chart of the overall resident's electric switch event detecting method decomposed based on two points of recursion SVD is as shown in Figure 1.
As shown in Figure 1, a kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD, including it is following
Step:
1) data acquire: the total current of acquisition resident's electric appliance, voltage data calculate total active power of resident's electric appliance；
2) two points of recursion SVD decompose the total active power data of resident's electric appliance；
21) construct two dimension Hankel matrix: to the total active power data of resident's electric appliance collected in the step 1) into
Row Hankel matrixing constructs two dimension Hankel matrix H.
Total active power data of resident's electric appliance are one-dimensional discrete signal X=[x1x2x3…xn], become by Hankel matrix
Change the space matrix that one line number of construction is 2:
H∈R2×n-1, the characteristics of H-matrix is that the second row data only lag a data point, therefore H-matrix than the first row
Two row data it is highly relevant.H-matrix has and after SVD is decomposed only there are two singular value, feature be first singular value compared with
Greatly, and second singular value is smaller.
H=USVT (2)
In formula, U=(u1,u2), U ∈ R2×2；V=(v1,v2,…vn-1), V ∈ R(n-1)×(n-1)；S=(diag (λ1,λ2),
O), S ∈ R2×(n-1), λ1>>λ2。
22) SVD decomposing H matrix: according to formulaAcquire approximate matrix H1With detail matrices H2。
In formula, ui∈R2×1, vi∈R(n-1)×1, i=1,2.If Hi=λiuivi T, then Hi∈R2×(n-1), HiHave with H similar
Matrix structure.H1Corresponding is larger singular value, referred to as approximate matrix；H2Corresponding is smaller singular value, referred to as details square
Battle array.H1And H2It is known as approximate signal A and detail signal D by the one-dimensional vector that Hankel inverse transformation acquires, embodies original letter respectively
Number main body general picture and minutia.
During solving approximate signal A, H1There are two row vectors, respectively λ1u1,1v1 TAnd λ1u1,2v1 T, wherein u1,1With
u1,2For u1Two coordinates.If La1And La2It is the sub-vector of the two row vectors respectively, all represents the number in approximate signal A
According to a1,2,a1,3,…a1,n-1, as shown in Fig. 2, Fig. 2 (a) is approximate matrix H1Vector, Fig. 2 (b) be detail matrices H2Vector.
23) approximate signal A and detail signal D:L is solveda1And La2It is not completely equivalent, to obtain the complete approximate letter of information
Number, according to formula A=(a1,1, (La1+La2)/2,a1,n) and formula D=(d1,1,(Ld2+Ld2)/2,d1,n) by H1、H2In the same data
Coherent element average, constitute approximate signal A and detail signal D.The result that jth time is decomposed is AjAnd Dj, after decomposing
Approximate signal AjConstruct new matrix Hj+1, repeat the above steps, original signal be decomposed into a series of approximate and details and is believed
Number.The SVD of total active power decomposes as shown in Figure 3.
24) singular entropy increment Delta E is solvedj: the detail signal that two points of recursion SVD are decomposed contains total active power
Catastrophe point information.With the increase of Decomposition order, the characteristic information for including in detail signal increases, but the increment of characteristic information
It reduces, and increases the time required to detecting catastrophe point.With the complexity of singular entropy characteristic feature information in the present invention, according to formula (4)
Singular entropy increment Delta E under the Decomposition order is calculated with formula (5)j。
3) switching time of resident's electric appliance is obtained using hard -threshold filter function
31) optimal value is taken: if judgement Δ EjIt is greater than the set value ε, new matrix H is constructed using approximate signal A, is repeated
It decomposes, until Δ EjLess than setting value ε, Decomposition order is optimal.
32) resident's electric switch moment: two points of recursion SVD are carried out to total active power data of resident's electric appliance and are decomposed, are obtained
To the detail signal containing the smaller noise of amplitude.Accurately to detect that exact time occurs for event, using hard threshold function
Detail signal D under the optimal Decomposition number of plies is filtered, in the curve after filtering when characteristic peak zero cross point
Carve the switching time of as resident's electric appliance.
4) experimental verification: building test platform in laboratory, using 8860 electric power wave-recording instrument of Zhi company acquisition water dispenser,
Micro-wave oven, electric heater, 4 kinds of resident's electric switch events of fan voltage and current data, sample frequency 2000Hz.
It can be to single linear invariable power electric switch event, single non-linear Variable power electric switch event, multiple electric appliances
Switch events, the bilateral CUSUM detection of change-point method of sliding, wavelet-decomposing method, the present invention that imparametrization is respectively adopted are mentioned
And two points of recursion SVD decomposition methods carry out detection verifying.
Fig. 4 is single linear invariable power electric switch event detection outcome comparison
Fig. 5 is single non-linear Variable power electric switch event detection outcome comparison
Fig. 6 is the comparison of multiple electric switch event detection outcomes
Fig. 4 (a) is water dispenser active power curves, and Fig. 5 (a) is micro-wave oven active power curves, and Fig. 6 (a) is multiple electricity
The total active power curves of device.In Fig. 4 (b) and (c), (d) and (e), (f) and (g), (b) (c) (d) is respectively corresponded in Fig. 5, Fig. 6
Two points of recursion SVD mentioned by the bilateral CUSUM detection of change-point method of sliding, wavelet-decomposing method, the present invention for imparametrization
Testing result figure under decomposition method.
Three experimental result pictures prove the bilateral CUSUM detection of change-point method of sliding, and there are sliding window size and threshold value h are difficult
The shortcomings that set, missing inspection, false detection rate are high.Obvious offset occurs for wavelet-decomposing method testing result, and under different decomposition level
It is different.For institute's inventive method compared to other two methods, omission factor is low, and erroneous detection number is few, and detection position stabilization does not occur
Offset, detection accuracy are high.
5) the bilateral CUSUM detection of change-point method of sliding, the wavelet decomposition of imparametrization REDD database authentication: is respectively adopted
Two points of recursion SVD decomposition methods mentioned by method, the present invention examine the REDD data set load power data of 1 hour
Survey, true switch events have occurred 45 times, three kinds of method statistic results as shown in fig. 7,
Fig. 7 (a) is REDD data power curve, in Fig. 7 (b), (c), (d) be respectively imparametrization sliding it is bilateral
CUSUM detection of change-point method, wavelet-decomposing method, the present invention mentioned by two points of recursion SVD decomposition methods testing result figure.
Detection method result is judged using three indexs, respectively (1) erroneous detection: detecting switch events,
It is practically without event；(2) missing inspection: switch events are not detected, actually event occurs；(3) recall ratio:
The switch events number correctly detected accounts for the ratio of all real events；(4) mean absolute error: the switching time that detects with
Mean absolute error between the true moment.Statistical result is as shown in table 1, the results show that the bilateral CUSUM detection of change-point of sliding window
Missing inspection value be 2, erroneous detection value be 26, recall ratio 95%；The missing inspection value value of wavelet-decomposing method is 3, and erroneous detection value value is 6, is looked into
Full rate is 93%；The missing inspection value value of two points of recursion SVD decomposition methods is 0, and erroneous detection value value is 2, and recall ratio reaches 100%.It compares
In other two methods, the method that the present invention is mentioned effectively increases resident's electric switch event detection precision.
Table 1
| Detection switch number | Missing inspection number | Erroneous detection number | It detects accuracy (%) | Mean absolute error (s) | |
| CUSUM | 43 | 2 | 26 | 95 | 1.29 |
| Wavelet decomposition | 42 | 3 | 6 | 93 | 1.43 |
| SVD is decomposed | 45 | 0 | 2 | 100 | 0.90 |
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can readily occur in various equivalent modifications or replace
It changes, these modifications or substitutions should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be with right
It is required that protection scope subject to.
Claims (8)
1. a kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD, which is characterized in that including following step
It is rapid:
Step 1: data acquisition: electric power register one's residence acquisition residential households power information data and obtain always having for residential households
Function power data；
Step 2: decomposing total active power data using two points of recursion SVD: to resident's electric switch event collected in step 1
Total active power data carry out Hankel matrixing, then carry out two points of recursion SVD and decompose, by calculating singular entropy to obtain most
Excellent Decomposition order；
Step 3: the switching time of resident's electric appliance is obtained using hard -threshold filter function: using hard -threshold filter function to step 2
Detail signal under the optimal Decomposition number of plies of middle acquisition is filtered, and realizes the accurate inspection of catastrophe point position in total active power
It surveys, i.e. the switching time of resident's electric appliance.
2. a kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD according to claim 1,
It is characterized in that, the power information data in the step 1 include the total current of resident's electric appliance, voltage data, the residential households
Total active power data obtained by MATLAB and as switch events detection target object.
3. a kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD according to claim 1,
Be characterized in that, the step 2 include it is following step by step:
Step 21: construction is directed to the two-dimentional Hankel matrix of total active power data；
Step 22: approximate signal A being obtained by inverse transformation after decomposing two dimension Hankel matrix using two points of recursion SVD and details is believed
Number D；
Step 23: repeating step 22 for original total active power data and be decomposed into a series of approximate and detail signal；
Step 24: calculating unusual entropy production for a series of detail signals obtained in step 23.
4. a kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD according to claim 1,
Be characterized in that, the step 3 include it is following step by step:
Step 31: taking optimal value: if the unusual entropy production of judgement is greater than the set value, new Hankel matrix is constructed using approximate signal
Repeat to decompose, until different entropy production is less than setting value, Decomposition order is optimal；
Step 32: opening for resident's electric appliance is obtained after being filtered using hard threshold function to the detail signal under the optimal Decomposition number of plies
Close the moment.
5. a kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD according to claim 3,
It is characterized in that, the two-dimentional Hankel matrix H in the step 21 are as follows:
In formula, [x1 x2 x3…xn] it is the corresponding one-dimensional discrete signal of the total active power data of resident's electric appliance, H ∈ R2×n-1。
6. a kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD according to claim 5,
It is characterized in that, the two-dimentional Hankel matrix H in the step 22 after two points of recursion SVD are decomposed describes formula are as follows:
H=USVT
In formula, two-dimensional quadrature matrix U=(u1, u2), U ∈ R2×2；V=(v1,v2,…vn-1), n-1 ties up orthogonal matrix V ∈ R(n -1)×(n-1)；Diagonal matrix S=(diag (λ1,λ2), O), S ∈ R2×(n-1), λ1>>λ2, λ1And λ2Respectively first and second is unusual
Value.
7. a kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD according to claim 6,
It is characterized in that, the calculation formula of the unusual entropy production in the step 24 are as follows:
In formula, λpAnd λqSingular value respectively on diagonal matrix, r are order.
8. a kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD according to claim 4, institute
State the hard threshold function in step 32 are as follows:
In formula, d is hard threshold function, dkFor variable.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910280687.1A CN110133396B (en) | 2019-04-09 | 2019-04-09 | Method for detecting household electrical appliance switch event based on binary recursive SVD decomposition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910280687.1A CN110133396B (en) | 2019-04-09 | 2019-04-09 | Method for detecting household electrical appliance switch event based on binary recursive SVD decomposition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110133396A true CN110133396A (en) | 2019-08-16 |
| CN110133396B CN110133396B (en) | 2021-07-20 |
Family
ID=67569327
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910280687.1A Active CN110133396B (en) | 2019-04-09 | 2019-04-09 | Method for detecting household electrical appliance switch event based on binary recursive SVD decomposition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110133396B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101876546A (en) * | 2009-11-27 | 2010-11-03 | 北京航空航天大学 | MEMS (Micro Electronic Mechanical System) gyro data processing method based on wavelet threshold de-noising and FAR (Finite Automaton Recognizable) model |
| CN104748961A (en) * | 2015-03-30 | 2015-07-01 | 中国矿业大学 | Gear fault diagnosis method based on SVD decomposition and noise reduction and correlation EEMD entropy features |
| CN105224872A (en) * | 2015-09-30 | 2016-01-06 | 河南科技大学 | A kind of user's anomaly detection method based on neural network clustering |
| CN106706285A (en) * | 2016-11-24 | 2017-05-24 | 重庆工业职业技术学院 | Brake disc inherent frequency online detection method |
| CN107133195A (en) * | 2017-04-14 | 2017-09-05 | 大连理工大学 | A kind of Methodology of The Determination of The Order of Model of engineering structure modal idenlification |
| CN108506742A (en) * | 2018-03-10 | 2018-09-07 | 西安电子科技大学 | A kind of adaptive useful signal judgement fluid line leakage locating method |
| CN109101465A (en) * | 2018-07-13 | 2018-12-28 | 广东石油化工学院 | A kind of electric switch event detecting method and system |
| CN110443104A (en) * | 2019-03-29 | 2019-11-12 | 杭州电子科技大学 | A kind of data filtering methods applied to ultrasonic partial discharge detection |
-
2019
- 2019-04-09 CN CN201910280687.1A patent/CN110133396B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101876546A (en) * | 2009-11-27 | 2010-11-03 | 北京航空航天大学 | MEMS (Micro Electronic Mechanical System) gyro data processing method based on wavelet threshold de-noising and FAR (Finite Automaton Recognizable) model |
| CN104748961A (en) * | 2015-03-30 | 2015-07-01 | 中国矿业大学 | Gear fault diagnosis method based on SVD decomposition and noise reduction and correlation EEMD entropy features |
| CN105224872A (en) * | 2015-09-30 | 2016-01-06 | 河南科技大学 | A kind of user's anomaly detection method based on neural network clustering |
| CN106706285A (en) * | 2016-11-24 | 2017-05-24 | 重庆工业职业技术学院 | Brake disc inherent frequency online detection method |
| CN107133195A (en) * | 2017-04-14 | 2017-09-05 | 大连理工大学 | A kind of Methodology of The Determination of The Order of Model of engineering structure modal idenlification |
| CN108506742A (en) * | 2018-03-10 | 2018-09-07 | 西安电子科技大学 | A kind of adaptive useful signal judgement fluid line leakage locating method |
| CN109101465A (en) * | 2018-07-13 | 2018-12-28 | 广东石油化工学院 | A kind of electric switch event detecting method and system |
| CN110443104A (en) * | 2019-03-29 | 2019-11-12 | 杭州电子科技大学 | A kind of data filtering methods applied to ultrasonic partial discharge detection |
Non-Patent Citations (2)
| Title |
|---|
| 唐炬，等: "基于Hankel矩阵的复小波_奇异值分解法提取局部放电特征信息", 《中国电机工程学报》 * |
| 林顺富，等: "基于二分递推SVD分解的居民电器开关事件检测方法", 《电网技术》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110133396B (en) | 2021-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lam et al. | A novel method to construct taxonomy electrical appliances based on load signaturesof | |
| Dinesh et al. | Residential appliance identification based on spectral information of low frequency smart meter measurements | |
| Chen et al. | Temporal and spectral feature learning with two-stream convolutional neural networks for appliance recognition in NILM | |
| Liang et al. | Load signature study—Part I: Basic concept, structure, and methodology | |
| Najmeddine et al. | State of art on load monitoring methods | |
| JP5347027B2 (en) | Method and apparatus for non-intrusive determination of facility power consumption by analyzing transients of load | |
| Aquino et al. | Characterization of electric load with information theory quantifiers | |
| CN110035090A (en) | A kind of smart grid false data detection method for injection attack | |
| CN109633301B (en) | Non-invasive electrical appliance load identification method based on quantum genetic optimization | |
| Jia et al. | Non-intrusive load identification using reconstructed voltage–current images | |
| CN107392391A (en) | A kind of electric load electricity consumption decomposition method based on deep learning | |
| CN110490760B (en) | Non-invasive electrical appliance load identification method based on hidden Markov chain | |
| CN107356819B (en) | A kind of intelligent domestic electromagnetic radiation information service system | |
| CN104793124A (en) | Switched circuit fault diagnosis method based on wavelet transformation and ICA (independent component analysis) feature extraction | |
| CN109697548A (en) | Electrical energy consumption analysis server and its electrical energy consumption analysis method | |
| CN109840691A (en) | Non-intrusion type subitem electricity estimation method based on deep neural network | |
| CN106250658B (en) | On-load switch electromagnetic mechanism quick calculation method based on radial basis function | |
| Gu et al. | Identification method for low-voltage arc fault based on the loose combination of wavelet transformation and neural network | |
| Li et al. | Non-intrusive load monitoring based on convolutional neural network mixed residual unit | |
| CN108918928B (en) | Power signal self-adaptive reconstruction method in load decomposition | |
| Bartman et al. | Identification of electrical appliances using their virtual description and data selection for non-intrusive load monitoring | |
| Hoyo-Montaño et al. | Overview of Non-Intrusive Load Monitoring: A way to energy wise consumption | |
| CN110133396A (en) | A kind of resident's electric switch event detecting method decomposed based on two points of recursion SVD | |
| Han et al. | Non‐intrusive load monitoring by voltage–current trajectory enabled asymmetric deep supervised hashing | |
| CN109521296B (en) | Non-invasive power load identification optimization algorithm under steady-state working condition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |