CN114123185A - ICA-R-based non-invasive electric vehicle charging load identification method - Google Patents
ICA-R-based non-invasive electric vehicle charging load identification method Download PDFInfo
- Publication number
- CN114123185A CN114123185A CN202111413003.4A CN202111413003A CN114123185A CN 114123185 A CN114123185 A CN 114123185A CN 202111413003 A CN202111413003 A CN 202111413003A CN 114123185 A CN114123185 A CN 114123185A
- Authority
- CN
- China
- Prior art keywords
- charging
- amplitude
- signal
- load
- power
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 239000011159 matrix material Substances 0.000 claims abstract description 38
- 238000012545 processing Methods 0.000 claims abstract description 9
- 244000068645 Carya illinoensis Species 0.000 claims description 5
- 235000009025 Carya illinoensis Nutrition 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 2
- 238000007781 pre-processing Methods 0.000 claims description 2
- 238000000605 extraction Methods 0.000 abstract description 14
- 238000012544 monitoring process Methods 0.000 description 10
- 239000013598 vector Substances 0.000 description 7
- 230000005611 electricity Effects 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 5
- 239000000463 material Substances 0.000 description 5
- 238000011160 research Methods 0.000 description 4
- 238000002474 experimental method Methods 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
-
- 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
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/382—Arrangements for monitoring battery or accumulator variables, e.g. SoC
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/28—Arrangements for balancing of the load in a network by storage of energy
- H02J3/32—Arrangements for balancing of the load in a network by storage of energy using batteries with converting means
- H02J3/322—Arrangements for balancing of the load in a network by storage of energy using batteries with converting means the battery being on-board an electric or hybrid vehicle, e.g. vehicle to grid arrangements [V2G], power aggregation, use of the battery for network load balancing, coordinated or cooperative battery charging
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
Abstract
The invention discloses a non-invasive electric vehicle charging load identification method based on ICA-R, which comprises the following steps: active power data of a user at home are collected; setting a reference signal, and classifying the charging power amplitude of the electric automobile; converting a one-dimensional power signal collected at a user entrance into a two-dimensional signal; processing the two-dimensional signal by applying an ICA-R method to obtain an extracted signal matrix; matching the starting time point and the ending time point of the electric automobile charging to obtain a charging time period; eliminating the load sections of an air conditioner and a dryer; reconstructing the original signal twice by using the charging time period and the charging power amplitude; and the amplitude corresponding to the minimum average error between the original signals in the two reconstructions is the charging load amplitude of the electric automobile. The invention can reduce the influence of the household air conditioner load and the dryer load, improve the accuracy of the electric automobile charging time interval extraction, and simultaneously can more accurately identify the charging power of the electric automobile.
Description
Technical Field
The invention relates to the field of charging load monitoring of electric automobiles, in particular to a non-invasive electric automobile charging load extraction method based on ICA-R.
Background
In recent years, the usage rate of electric vehicles in many countries around the world is increasing, and data display: in 2019, the sales volume of electric automobiles in the world reaches 210 thousands, the holding volume of electric automobiles in the world is 720 thousands, and domestic electric automobiles account for 47 percent in the world. Although the market is still inclined to the latter in the competition of electric automobiles and diesel and gasoline automobiles, the electric automobiles still have great development potential. According to the prediction of Penbo new energy resources, the global keeping quantity of electric automobiles reaches 5.59 hundred million by 2040 years, and the keeping quantity of all types of automobiles accounts for 33 percent, wherein 55 percent of new automobiles are electric automobiles. Although the popularization of electric vehicles is an important help for environmental protection and resource conservation, it has a bad influence on the power grid. For example, the charging period of electric vehicles in residential areas is mainly concentrated after work, which increases the peak demand of loads, causes adverse effects on the distribution network, increases the loads of transformers, and even may cause overload, and increases the possibility of grid breakdown. Therefore, monitoring the charging load of the electric automobile plays an important role in participating in demand response of the electric automobile, making a charging plan by a user, deploying an electric automobile charging pile (station) and monitoring the system load by a system operator.
Hart formally introduced non-intrusive load monitoring (NILM) in 1992 as an example of a software system that could analyze a single point of electrical data at a customer premises to obtain power usage information about individual devices within the customer premises.
At present, the following two main aspects of research on monitoring of charging load of an electric vehicle are provided: firstly, establishing a statistical model of the charging load of the electric automobile; secondly, the algorithm is improved to improve the identification accuracy. A large number of charging load monitoring algorithms of the existing electric automobile comprise principal component analysis, random forests, deep generation models, neural networks and the like. The invention combines an important method ICA-R in the field of signal analysis with non-invasive monitoring so as to more accurately extract the charging time period of the electric automobile and improve the charging power amplitude estimation algorithm, thereby more accurately identifying the charging power amplitude of the electric automobile.
Disclosure of Invention
In view of the defects in the prior art, in order to further improve the accuracy of the charging load extraction of the electric automobile, the invention provides a non-invasive electric automobile charging load identification method based on ICA-R in combination with a non-invasive load monitoring technology, and aims to accurately identify the charging load of a single household electric automobile. The method can reduce the influence of the load of the household air conditioner and the load of the dryer, improve the accuracy of the extraction of the charging time interval of the electric automobile, and simultaneously can more accurately identify the charging power of the electric automobile.
In order to solve the technical problem, the invention provides a non-invasive electric vehicle charging load identification method based on ICA-R, which comprises the following steps:
step 1: active power data of a user at home are collected, a reference signal is set, and charging power amplitude values of the electric automobile are classified;
step 2: data preprocessing, namely converting the one-dimensional power signals acquired in the step 1 into two-dimensional signals;
and step 3: on the basis of taking a sliding window for the two-dimensional signal in the step 2, processing the two-dimensional signal obtained in the step 2 by applying an ICA-R method to obtain an extracted signal matrix Zm；
And 4, step 4: processing the matrix Z obtained in step 3mMatching the starting time point and the ending time point of the electric automobile charging event to obtain a starting time matrix and a stopping time matrix of electric automobile charging;
and 5: eliminating load sections with different load characteristics from the electric automobile;
step 6: estimating the charging power amplitude, reconstructing the original signal by using the charging power amplitude of the electric automobile and the corresponding charging time period in the starting and stopping time matrix, and calculating the average error between the once-reconstructed charging section and the original signal;
and 7: respectively extracting the charging time periods of the charging power amplitude corresponding to the minimum average error in the step 6 and the charging power amplitude in the same class as the charging power amplitude, performing secondary estimation to obtain a secondary reconstruction signal after the optimization estimation, and calculating the average error between the charging section after the secondary reconstruction and the original signal;
and 8: and (3) selecting the minimum value of the average error between the charging section subjected to the primary reconstruction in the step (6) and the original signal and the average error between the charging section subjected to the secondary reconstruction in the step (7) and the original signal, wherein the charging power amplitude corresponding to the minimum value is the charging power amplitude of the electric automobile, and the corresponding charging time period is the charging time period of the electric automobile.
Further, the non-invasive electric vehicle charging load extraction method based on ICA-R of the invention is characterized in that:
in the step 1, the reference signal is set to be a rectangular wave signal with the charging power amplitude as the height and 10 minutes as the time length; according to the public data set provided by Pecan Street, wherein the amplitude of the electric vehicle charging power is divided into five grades, 3.3kW, 3.7kW, 3.9kW, 6.5kW and 7.2kW, the classification of the amplitude of the electric vehicle charging power is as follows: a power amplitude of 3, xkW is taken as a first class (category 1) and a power amplitude of more than 6kW is taken as a second class (category 2).
In step 3, the length of a sliding window taken for the two-dimensional signal in step 2 is set to be N-10, and an ICA-R method is applied to each window to perform two-dimensional signal processing.
In step 4, before the time point when each electric vehicle charging event starts and ends, the matrix ZmThe number of non-zero data in each column increases from 1 to 10, and after the starting and ending time points, the matrix ZmThe number of the non-zero data in each row is distributed in a descending way from 10 to 1, and a starting and stopping time matrix of the electric automobile charging is obtained according to the rule.
In step 5, the load section with different load characteristics from the electric vehicle is as follows: a load section that is similar to an electric vehicle in amplitude or power curve waveform and satisfies the following condition a) or condition b); a) the interval between two adjacent peaks in the power curve waveform is less than 5 minutes; b) one power consumption period lasts no more than 30 minutes, and two adjacent power consumption periods are separated by no more than 18 minutes.
And 6, estimating the charging power amplitude by adopting a method of trying to reconstruct total power data by using the electric vehicle charging load extracted from each amplitude, solving an average error between a reconstructed signal and an original signal, wherein the amplitude corresponding to the minimum error is the amplitude of the charging power of the electric vehicle.
Compared with the prior art, the invention has the beneficial effects that:
according to the invention, the non-invasive load monitoring technical method is applied to the charging load extraction of the electric automobile, the ICA-R method is applied to the non-invasive load monitoring, and the non-invasive charging load extraction method of the electric automobile based on the ICA-R is established, so that the charging time period of the electric automobile can be extracted and the charging power of the electric automobile can be estimated under the condition of not invading the interior of a user, and the accuracy is higher. The method can help users reasonably make charging plans, planning personnel to deploy electric vehicle charging piles (stations), system load prediction and the like.
Drawings
FIG. 1 is a flow chart of the ICA-R based non-intrusive electric vehicle charging load identification of the present invention;
FIG. 2 is a comparison of load curves of an electric vehicle, an air conditioner and a dryer;
FIG. 3 shows the result of the charging load extraction of an electric vehicle for residence No. 661 in 4 months and 3 days in the research materials;
FIG. 4 shows the result of the charging load extraction of an electric vehicle for residence No. 661 in 4 months and 18 days in the research materials;
fig. 5 shows the result of the charge load extraction of the electric vehicle for residence No. 6139 at 4 months and 18 days in the study materials.
Detailed Description
The invention will be further described with reference to the following figures and specific examples, which are not intended to limit the invention in any way.
The invention provides a non-invasive electric vehicle charging load identification method based on ICA-R, a flow chart of which is shown in figure 1, and the method mainly comprises the following steps:
step 1: and acquiring an active power sequence of a user at the house to obtain an input signal x. And classifying the charging power amplitude of the electric automobile by setting a reference signal. And meanwhile, collecting an actual value of the charging power of the electric automobile for verifying the accuracy of the algorithm subsequently.
And counting the charging power amplitude of the electric automobile, and classifying the approximate amplitude into one class. Setting the reference signal as a rectangular wave signal with each charging power amplitude as height, namely defining the vector of the reference signal as rm＝[r1m，r2m，...，rNm]The reference signal matrix is R ═ R1，r2，...，rM]Wherein M is 1, …, M represents the mth reference signal, M is the total number of reference signals, i.e. the number of charging power amplitudes, N is the length of the reference signal vector, and N is 10.
Through the analysis discovery to the public data set that Pecan Street provided, wherein electric automobile charging power's amplitude is divided into five grades, is 3.3kW, 3.7kW, 3.9kW, 6.5kW and 7.2kW respectively, classifies electric automobile charging power amplitude and is: a power amplitude of 3, xkW is taken as a first class (category 1) and a power amplitude of more than 6kW is taken as a second class (category 2).
Step 2: and (3) converting the one-dimensional active power signal acquired in the step (1) into a two-dimensional signal. In extracting reference signal vector rmBefore the corresponding electric automobile is charged with load, all the originally observed one-dimensional active power signals x are greater than rmSubtracting a certain value from the observation point of the amplitude to obtain a new observation sequence
Combining to obtain a power signal matrix
And step 3: on the basis of taking a sliding window for the two-dimensional signal in the step 2, processing the two-dimensional signal obtained in the step 2 by applying an ICA-R methodThen, an extracted signal matrix Z is obtainedm。
For each reference signal rmThe window length is set to N-10 and the entire power signal matrix X is covered by the shiftmExtracting R in each window by applying ICA-R methodmCorresponding source signal to obtain solution vector zn＝[zn1，...，znT]Wherein N is 1, …, N. And each time the window is moved to the observed signal XmAt the last bit of the initial position, the window of the initial position is shifted to the right by one bit, i.e. at the nth cycle, the data of the first window is shifted from the power signal matrix XmThe nth number of (a). After repeating the above steps ten times, the obtained window can cover XmAll ten consecutive points in the sequence, thereby obtaining an NxT extraction signal matrix, which is marked as Zm＝[z1，...，zN]。
And 4, step 4: processing the extracted signal matrix Z obtained in step 3mAnd matching the starting time point and the ending time point of the electric automobile charging event to obtain a starting time matrix and a stopping time matrix of electric automobile charging.
Before the time point of starting and ending of each electric vehicle charging event, the signal matrix Z is extractedmThe number of non-zero data in each column increases from 1 to 10, and after the starting and ending time points, the matrix ZmThe number of the non-zero data in each row is distributed in a descending way from 10 to 1, and a starting and stopping time matrix of the electric automobile charging is obtained according to the rule. Defining a variable p, let p equal 1, …, N/2, representing ZmThe number of non-zero elements in each column of the matrix. If the number of non-zero elements in the t-th column is not equal to p, cptThe value of (d) is equal to 0, calculated as follows:
for each p, a solution vector can be obtained
All solution vectors form a matrix Cm，
Wherein p is 1, …, N/2.
CmIn the matrix, data are mostly distributed in a main diagonal line before the starting or ending time point, and data are mostly distributed in a secondary diagonal line after the time point, and the time point at which the data in the fourth row and the data in the fifth row are distributed in the main diagonal line is taken as a judgment mark for starting and ending. However, when some electricity usage periods are close, the situation of time point mark disappearance may be encountered, so when the last element of the main diagonal and the first element of the secondary diagonal are separated by more than 15 sampling periods, the time point is supplemented appropriately therein. After finding the mountain time point mark, matching the starting time and the ending time. If there is
Then the first charging period starts from 1 and ends at the first time mark point, and then the time points are paired in pairs; if (2) is not satisfied, pair-by-pair pairing is started from the first time marker point. Finally, a charging start-stop time matrix of the electric automobile can be obtained
Where E represents the total number of electric vehicle charging periods,
indicating the start time of the e-th charging period,
indicates the end time of the E-th charging period, E-1, …, E.
And 5: and eliminating the load sections with different load characteristics from the electric automobile.
Through the analysis of the data set, it was found that: the charging power curve of the electric automobile is similar to a square wave, the amplitude is high, and the amplitude or the waveform of the air conditioner and the dryer is similar to that of the electric automobile to a certain extent. From fig. 2, it can be seen that the difference between the air conditioner, the dryer electrical load and the electric vehicle charging load is small, the dryer is operated for a short time and has a high peak frequency, and the interval between two adjacent peaks is less than 5 minutes. The duration of the electricity utilization period of the air conditioner is short, two adjacent electricity utilization periods are close to each other, generally, the duration of one electricity utilization period of the air conditioner is not more than 30 minutes, the interval time between two adjacent periods is not more than 18 minutes, the charging time of the electric automobile is more than 30 minutes, and the interval time is several hours. The load influence of the air conditioner and the dryer is eliminated. Since the charging loads of the air conditioner, the dryer and the electric vehicle have similar power characteristics and can be mistakenly identified as the electric vehicle, the load sections are removed by utilizing the load characteristics of the air conditioner, the dryer and the electric vehicle, which are different from those of the electric vehicle.
Therefore, in the present invention, the load section having a load characteristic different from that of the electric vehicle is: a load section that is similar to an electric vehicle in amplitude or power curve waveform and satisfies the following condition a) or condition b);
a) the interval between two adjacent peaks in the power curve waveform is less than 5 minutes;
b) one power consumption period lasts no more than 30 minutes, and two adjacent power consumption periods are separated by no more than 18 minutes.
In the present invention, the interval time between two charging sections is defined as
One charging segment has a duration of
If the interval between a plurality of consecutive electricity usage periods is less than 18 and at least one of the adjacent dual-purpose electricity periods is less than 120, then:
then these consecutive charging segments are discarded and D is updatedmAnd (4) matrix. This can eliminate the influence of the dryer and part of the air conditioner.
If a charging period duration is less than 30, that is:
then the charging section is removed and D is updatedmAnd (4) matrix. This can eliminate the influence of the air conditioner.
Step 6: estimating the charging power amplitude, reconstructing the original signal by using the charging power amplitude of the electric automobile and the corresponding charging time period in the start-stop time matrix, and calculating the average error between the first reconstructed signal and the original signal.
In the invention, the method for estimating the charging power amplitude is to try to extract the charging load of the electric automobile from the mountain by using each amplitude to reconstruct total power data, solve the average error between a reconstructed signal and an original signal, and the amplitude corresponding to the minimum error is the amplitude of the charging power of the electric automobile. For better estimation, two variables are defined here
And
data are shown for the first 5 to 15 minutes and the last 5 to 15 minutes, respectively, of the e-th charging session. The specific method comprises the following steps:
1) firstly, calculating fundamental wave power of each section of electric automobile charging section of the extracted mountain
Whether or not they contain integrity when calculated
Or
As a classification criterion, i.e. whether there is complete data for the first 5 to 15 minutes or the last 5 to 15 minutes,
is intact
Or
Minimum value of (1).
2) Calculating the data in the e-th charging section in the original observation signal x as xeEstimating x by using each section of fundamental wave power to obtain a first reconstructed signal ym：
ym＝x
3) Computing a reconstructed signal ymAnd the average error between the original signal x:
for matrix CmWherein p 1, N2 each
Repeating the above steps to calculate all ME0 of mountainp，mThen minimum mean error of first reconstruction
ME1m＝min(ME0p，m) (7)
Repeat steps 2 through 6 for each M, M1.., M, resulting in a vector of errors ME1, ME1 ═ ME11，...，ME1M]. Thus, a reference signal corresponding to the estimated signal with the minimum average error can be obtained
m*The amplitude of the corresponding reference signal is the charging power with the minimum error.
And 7: and (4) respectively extracting the charging power amplitude corresponding to the minimum average error in the step (6) and the charging power amplitude belonging to the same class as the minimum average error, carrying out secondary estimation on the charging time period and the reconstructed signal in the step (6), thus obtaining a secondary reconstructed signal after optimized estimation, and calculating the average error between the secondary reconstructed signal and the original signal. The specific process is as follows:
m is to be*And the power amplitudes belonging to the same class are all recorded as m*And re-estimated. For each m*Respectively extracting their corresponding time matrixes
And reconstructing the signal in step 6
For further reconstruction
For each m*E, reconstructing its estimated signal to obtain
Wherein B2eAt the e-thIn the charging section, the average error of the signal and the acquired active power sequence is reconstructed in step 6. The quadratic reconstructed signal and the mean error ME2 of the original signal are then solved in a similar way as in step 6.
And 8: and (3) selecting the minimum value of the average error between the charging section subjected to the primary reconstruction in the step (6) and the original signal and the average error between the charging section subjected to the secondary reconstruction in the step (7) and the original signal, wherein the charging power amplitude corresponding to the minimum value is the charging power amplitude of the electric automobile, and the corresponding charging time period is the charging time period of the electric automobile. The specific process is as follows:
reference signal r corresponding to optimal reconstructed signalmWherein
The value is obtained by the following formula:
the final result is obtained
Charging time period of rmCorresponding to
Matrix, charging power amplitude of rm，1。
Study materials:
experiments were conducted using the data of year 2018, residence number 661, 4, 18, and 4, 18, days 6139, residence number 4, 18 in the Pecan Street database, and the data were obtained as shown in fig. 3, 4, and 5The result of the extraction of the charging load of the electric automobile shows that the method of the invention has more accurate extraction of the charging time period and small amplitude estimation error. Definition F1scoreTo evaluate the accuracy of the algorithm [ A.A.Munshi and Y.A.I.Mohamed.Upperviced Noninistributive Extraction of electric Vehicle steering Load Patterns.]The above three sets of experimental results F1scoreThe values of (A) were 95.74%, 98.52%, 97.46%, respectively.
Other sets of data from homes 1642, 661 and 1639 from Pecan Street were used in the study for experiments and the RMSE was calculated to give table 1.
Residential test results in tables 1661, 1642 and 6139
According to the experimental results of the research materials, the method can extract the charging time period of the electric automobile and estimate the charging power of the electric automobile without invading the interior of a user, and has high accuracy. The method can help users reasonably make charging plans, planning personnel to deploy electric vehicle charging piles (stations), system load prediction and the like.
While the present invention has been described with reference to the accompanying drawings, the present invention is not limited to the above-described embodiments, which are illustrative only and not restrictive, and various modifications which do not depart from the spirit of the present invention and which are intended to be covered by the claims of the present invention may be made by those skilled in the art.
Claims (6)
1. A non-invasive electric vehicle charging load identification method based on ICA-R is characterized by comprising the following steps:
step 1: active power data of a user at home are collected, a reference signal is set, and charging power amplitude values of the electric automobile are classified;
step 2: data preprocessing, namely converting the one-dimensional power signals acquired in the step 1 into two-dimensional signals;
and step 3: on the basis of taking a sliding window for the two-dimensional signal in the step 2, processing the two-dimensional signal obtained in the step 2 by applying an ICA-R method to obtain an extracted signal matrix Zm；
And 4, step 4: processing the matrix Z obtained in step 3mMatching the starting time point and the ending time point of the electric automobile charging event to obtain a starting time matrix and a stopping time matrix of electric automobile charging;
and 5: eliminating load sections with different load characteristics from the electric automobile;
step 6: estimating the charging power amplitude, reconstructing the original signal by using the charging power amplitude of the electric automobile and the corresponding charging time period in the starting and stopping time matrix, and calculating the average error between the once-reconstructed charging section and the original signal;
and 7: respectively extracting the charging time periods of the charging power amplitude corresponding to the minimum average error in the step 6 and the charging power amplitude in the same class as the charging power amplitude, performing secondary estimation to obtain a secondary reconstruction signal after the optimization estimation, and calculating the average error between the charging section after the secondary reconstruction and the original signal;
and 8: and (3) selecting the minimum value of the average error between the charging section subjected to the primary reconstruction in the step (6) and the original signal and the average error between the charging section subjected to the secondary reconstruction in the step (7) and the original signal, wherein the charging power amplitude corresponding to the minimum value is the charging power amplitude of the electric automobile, and the corresponding charging time period is the charging time period of the electric automobile.
2. The ICA-R based non-invasive method for extracting charging load of an electric vehicle according to claim 1, wherein in step 1, the reference signal is set as a rectangular wave signal with charging power amplitude as height and time length of 10 minutes; according to the public data set provided by Pecan Street, wherein the amplitude of the electric vehicle charging power is divided into five grades, 3.3kW, 3.7kW, 3.9kW, 6.5kW and 7.2kW, the classification of the amplitude of the electric vehicle charging power is as follows: a power amplitude of 3, xkW is taken as a first class (category 1) and a power amplitude of more than 6kW is taken as a second class (category 2).
3. The ICA-R based non-invasive charging load recognition method of claim 1, wherein in step 3, the length of a sliding window taken for the two-dimensional signal of step 2 is set to be N-10, and the ICA-R method is applied to each window for two-dimensional signal processing.
4. The ICA-R based non-intrusive electric vehicle charging load identification method of claim 1, wherein in step 4, the matrix Z is prior to the time point when each electric vehicle charging event starts and endsmThe number of non-zero data in each column increases from 1 to 10, and after the starting and ending time points, the matrix ZmThe number of the non-zero data in each row is distributed in a descending way from 10 to 1, and a starting and stopping time matrix of the electric automobile charging is obtained according to the rule.
5. The ICA-R based non-intrusive electric vehicle charging load identification method according to claim 1, wherein in the step 5, the load segment with different load characteristics from the electric vehicle is: a load section that is similar to an electric vehicle in amplitude or power curve waveform and satisfies the following condition a) or condition b);
a) the interval between two adjacent peaks in the power curve waveform is less than 5 minutes;
b) one power consumption period lasts no more than 30 minutes, and two adjacent power consumption periods are separated by no more than 18 minutes.
6. The ICA-R based non-invasive electric vehicle charging load identification method according to claim 1, wherein for step 6, the charging power amplitude is estimated by trying to reconstruct total power data by using the electric vehicle charging load extracted from each amplitude, and solving an average error between the reconstructed signal and the original signal, wherein the amplitude corresponding to the minimum error is the amplitude of the electric vehicle charging power.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111413003.4A CN114123185A (en) | 2021-11-25 | 2021-11-25 | ICA-R-based non-invasive electric vehicle charging load identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111413003.4A CN114123185A (en) | 2021-11-25 | 2021-11-25 | ICA-R-based non-invasive electric vehicle charging load identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114123185A true CN114123185A (en) | 2022-03-01 |
Family
ID=80372991
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111413003.4A Pending CN114123185A (en) | 2021-11-25 | 2021-11-25 | ICA-R-based non-invasive electric vehicle charging load identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114123185A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117370873A (en) * | 2023-12-08 | 2024-01-09 | 四川大学 | Non-invasive load extraction method for charging load of residential electric automobile |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105759113A (en) * | 2016-02-29 | 2016-07-13 | 北京工业大学 | Non-intrusive load monitoring and decomposition method for electric vehicle charging |
| CN107832705A (en) * | 2017-11-08 | 2018-03-23 | 天津大学 | A kind of non-intrusion type load recognition methods based on canonical correlation analysis |
| CN112394220A (en) * | 2020-11-13 | 2021-02-23 | 四川大学 | Non-invasive electric vehicle charging load mode extraction method |
-
2021
- 2021-11-25 CN CN202111413003.4A patent/CN114123185A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105759113A (en) * | 2016-02-29 | 2016-07-13 | 北京工业大学 | Non-intrusive load monitoring and decomposition method for electric vehicle charging |
| CN107832705A (en) * | 2017-11-08 | 2018-03-23 | 天津大学 | A kind of non-intrusion type load recognition methods based on canonical correlation analysis |
| CN112394220A (en) * | 2020-11-13 | 2021-02-23 | 四川大学 | Non-invasive electric vehicle charging load mode extraction method |
Non-Patent Citations (1)
| Title |
|---|
| YI WANG等: "Non-intrusive Electric Vehicle Charging Load Disaggregation based on Independent Component Analysis with Reference", 《PREPRINTS OF THE 5TH IEEE CONFERENCE ON ENERGY INTERNET AND ENERGY SYSTEM INTEGRATION》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117370873A (en) * | 2023-12-08 | 2024-01-09 | 四川大学 | Non-invasive load extraction method for charging load of residential electric automobile |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112041693B (en) | Power distribution network fault positioning system based on mixed wave recording | |
| CN108107324B (en) | Power distribution network fault positioning method based on deep convolutional neural network | |
| CN101576443B (en) | Life prediction method of accelerated life test based on grey RBF neural network | |
| Abeysinghe et al. | Topological properties of medium voltage electricity distribution networks | |
| CN103810637A (en) | Motor vehicle insurance fraud detecting method and system | |
| KR20190070838A (en) | Battery classification method and system | |
| Zhu et al. | Grid impact analysis of heavy-duty electric vehicle charging stations | |
| Majidpour et al. | Modified pattern sequence-based forecasting for electric vehicle charging stations | |
| CN108806248B (en) | Vehicle travel track division method for RFID electronic license plate data | |
| CN111080105A (en) | Transformer area user-to-user relationship identification method and system based on voltage time sequence data | |
| CN108717579B (en) | Short-term wind power interval prediction method | |
| CN104966161A (en) | Electric energy quality recording data calculating analysis method based on Gaussian mixture model | |
| Li et al. | A hybrid model for residential loads in a distribution system with high PV penetration | |
| CN106874950A (en) | A kind of method for identifying and classifying of transient power quality recorder data | |
| CN114123185A (en) | ICA-R-based non-invasive electric vehicle charging load identification method | |
| Giasemidis et al. | A genetic algorithm approach for modelling low voltage network demands | |
| CN112819649A (en) | Method and device for determining station area subscriber change relationship | |
| CN111193260A (en) | Power system transient stability automatic evaluation method of self-adaptive expansion data | |
| CN105550450B (en) | Electric energy quality interference source characteristic harmonic modeling method | |
| Zhu et al. | DWT-based aggregated load modeling and evaluation for quasi-static time-series simulation on distribution feeders | |
| CN116953554A (en) | Multi-fragment data-based method and device for estimating SOH of lithium battery of energy storage power station | |
| Li et al. | Preliminary statistical study of low voltage distribution feeders under a representative HV network in Western Australia | |
| Dang et al. | Privacy preservation needed for smart meter system: A methodology to recognize electric vehicle (EV) models | |
| CN114707292A (en) | Voltage stability analysis method for power distribution network containing electric automobile | |
| Li et al. | A statistical study on topological features of high voltage distribution networks in Western Australia |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20220301 |