CN109638830A - A kind of electric load model building method, device and equipment - Google Patents
A kind of electric load model building method, device and equipment Download PDFInfo
- Publication number
- CN109638830A CN109638830A CN201910049509.8A CN201910049509A CN109638830A CN 109638830 A CN109638830 A CN 109638830A CN 201910049509 A CN201910049509 A CN 201910049509A CN 109638830 A CN109638830 A CN 109638830A
- Authority
- CN
- China
- Prior art keywords
- signal
- load model
- electric load
- low
- 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.)
- Granted
Links
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
-
- 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
This application discloses a kind of electric load model building methods, device and equipment, compressed sensing sampling is carried out to electric load signal using random Gaussian calculation matrix, acquisition data volume can be greatly reduced, to save storage and communication resources, accelerate transmission speed, after receiving the low-dimensional signal after compressed sensing sampling, belong to the accurate recovery that steady-state signal or transient signal carry out signal with Fourier's base sparse matrix recovery algorithms and wavelet basis sparse matrix recovery algorithms respectively further according to low-dimensional signal, Static Load model and Dynamic Load Model are constructed respectively according to the signal parameter after recovery, have the advantages that small storage quickly transmission and precisely modeling, solve it is existing lack it is a kind of have it is small store quickly transmit accurate electric load model the technical issues of.
Description
Technical field
This application involves electric load technical field more particularly to a kind of electric load model building method, device and set
It is standby.
Background technique
Electric load model be describe load ports power or electric current with its port voltage and frequency variation characteristics number
Equation and corresponding parameter are learned, load model has important influence to Power System Analysis and calculating, and load model is not
Accurately have become an important factor for restricting electric system simulation computational accuracy, is obtained because of load model inaccuracy excessively happy
It sees or the network analysis of pessimism is as a result, huge loss may be brought to planning, the operation of electric system, due to electric load sheet
Body has randomness, time variation and complexity, so that Electric Load Modeling fails to be well solved always.
With the fast development of network technology, the data volume of electric load is also increasing, and system is to electric load information
Transmission speed and memory capacity also proposed increasingly higher demands, therefore, how to design one kind have it is small storage quickly pass
Defeated accurate electric load model is those skilled in the art's technical problem urgently to be resolved.
Summary of the invention
The embodiment of the present application provides a kind of electric load model building method, device and equipment, solves existing lack
It is a kind of less to have the technical issues of small accurate electric load model for storing and quickly transmitting.
In view of this, the application first aspect provides a kind of electric load model building method, comprising:
101, Real Time Compression perception sampling is carried out to the electric load signal got according to random Gaussian calculation matrix, obtained
To the low-dimensional signal comprising whole signal messages and be transmitted to receiving end；
If the low-dimensional signal 102, received is steady-state signal, step 103 and step 104 are executed, if receive
The low-dimensional signal is transient signal, thens follow the steps 105 and step 106；
103, before the low-dimensional signal being reverted to compressed sensing sampling according to Fourier's base sparse matrix recovery algorithms
First higher-dimension sampled signal；
104, signal electrical quantity parameter, the signal electrical quantity parameter packet are obtained according to the first higher-dimension sampled signal
Include: signal voltage virtual value, signal code virtual value, signal active power, signal reactive power, frequency and harmonic wave are established negative
Lotus static models；
105, the low-dimensional signal is reverted to the before compressed sensing sampling according to wavelet basis sparse matrix recovery algorithms
Two higher-dimension sampled signals；
106, transient signal voltage and transient signal electric current are obtained according to the second higher-dimension sampled signal, it is dynamic establishes load
States model.
Preferably, before step 101 further include:
100, the electric load signal of power feeder is acquired according to high frequency acquisition device in real time and carries out AD signal conversion,
In, frequency acquisition meets the requirement of Shannon-Nyquist theorem.
Preferably, after step 105 and step 106, further includes:
107, the Static Load model and/or the Dynamic Load Model are stored in load model database, and in real time
It updates.
Preferably, the Static Load model is ZIP multinomial model.
The preferably described Dynamic Load Model is three rank induction machine parallel connection ZIP models.
Preferably, the random Gaussian calculation matrix are as follows:
Wherein,For calculation matrix, N is that mean value is the Gauss normal distribution that 0 variance is 1/m, and m is through compressed sensing pressure
Data point number after contracting.
Preferably, the ZIP multinomial model are as follows:
Wherein, P is active power, and Q is reactive power, and V is voltage value, V0For voltage base value, P0For active power base value,
Q0For reactive power base value, apFor the percentage that the active power of constant-impedance accounts in total active power, bpFor having for constant current
The percentage that function power accounts in total active power, cpIt is accounted in total active power for the active power of constant power load model
Percentage, ap+bp+cp=1；aQFor the percentage that the reactive power of constant-impedance accounts in total reactive power, bQFor constant current
The percentage that reactive power accounts in total reactive power, cQIt is accounted in total reactive power for the reactive power of constant power load model
Percentage, aQ+bQ+cQ=1.
Preferably, the three ranks induction machine parallel connection ZIP model are as follows:
State equation:
Output equation:
Wherein,X=Xs+Xm, X '=Xs+Xm//Xr, A ω2+ B ω+C=1, E 'd、E′qIt is respectively equivalent
Motor d axis transient state built-in potential and q axis transient state built-in potential, Id、Iq、Ud、UqThe d axis of respectively equivalent motor and the electric current of q axis
And component of voltage；ω is the revolving speed of equivalent motor, RsFor the stator resistance of equivalent motor, parameter A, B, C are equivalent electronic
The coefficient of machine load torque, H are the inertia time constant of equivalent motor.
The application second aspect provides a kind of electric load model construction device, comprising:
Compressed sensing sampling unit, it is real for being carried out according to random Gaussian calculation matrix to the electric load signal got
When compressed sensing sample, the obtained low-dimensional signal comprising whole signal messages is simultaneously transmitted to receiving end；
Judging unit triggers stable state recovery unit and static state if the low-dimensional signal for receiving is steady-state signal
Modeling unit triggers transient recovery unit and dynamic modeling unit if the low-dimensional signal received is transient signal；
Stable state recovery unit, for the low-dimensional signal to be reverted to compression according to Fourier's base sparse matrix recovery algorithms
The first higher-dimension sampled signal before perception sampling；
Static modelling unit, for obtaining signal electrical quantity parameter, the signal according to the first higher-dimension sampled signal
Electrical quantity parameter include: signal voltage virtual value, signal code virtual value, signal active power, signal reactive power, frequency and
Harmonic wave establishes Static Load model；
Transient recovery unit, for the low-dimensional signal to be reverted to compression sense according to wavelet basis sparse matrix recovery algorithms
The second higher-dimension sampled signal before knowing sampling；
Dynamic modeling unit, for obtaining transient signal voltage and transient signal electricity according to the second higher-dimension sampled signal
Stream, establishes Dynamic Load Model.
The application third aspect provides a kind of electric load model construction equipment, and the equipment includes processor and deposits
Reservoir；
Said program code is transferred to the processor for storing program code by the memory；
The processor is used for the electric load model according to the instruction execution first aspect in said program code
Construction method.
As can be seen from the above technical solutions, the embodiment of the present application has the advantage that
In the application, a kind of electric load model building method is provided, comprising:
101, Real Time Compression perception sampling is carried out to the electric load signal got according to random Gaussian calculation matrix, obtained
To the low-dimensional signal comprising whole signal messages and be transmitted to receiving end；If the low-dimensional signal 102, received is stable state letter
Number, step 103 and step 104 are executed, if the low-dimensional signal received is transient signal, thens follow the steps 105 and step 106；
103, the first higher-dimension that low-dimensional signal reverts to before compressed sensing sampling is sampled according to Fourier's base sparse matrix recovery algorithms
Signal；104, signal electrical quantity parameter is obtained according to the first higher-dimension sampled signal, signal electrical quantity parameter includes: that signal voltage has
Valid value, signal code virtual value, signal active power, signal reactive power, frequency and harmonic wave, establish Static Load model；
105, low-dimensional signal is reverted to the second higher-dimension sampling letter before compressed sensing sampling according to wavelet basis sparse matrix recovery algorithms
Number；106, transient signal voltage and transient signal electric current are obtained according to the second higher-dimension sampled signal, establishes Dynamic Load Model.This
The electric load model building method provided in application carries out compression sense to electric load signal using random Gaussian calculation matrix
Know sampling, acquisition data volume can be greatly reduced, to save storage and communication resources, accelerate transmission speed, is receiving
It low-dimensional signal after to compressed sensing sampling and then steady-state signal or transient signal is belonged to according to low-dimensional signal uses respectively
Fourier's base sparse matrix recovery algorithms and wavelet basis sparse matrix recovery algorithms carry out the accurate recovery of signal, after recovery
Signal parameter construct Static Load model and Dynamic Load Model respectively, have it is small storage quickly transmission and precisely model it is excellent
Point, solve it is existing lack it is a kind of have it is small store quickly transmit accurate electric load model the technical issues of.
Detailed description of the invention
Fig. 1 is the flow diagram of one of the embodiment of the present application electric load model building method；
Fig. 2 is one of the embodiment of the present application another flow diagram of electric load model building method；
Fig. 3 is the structural schematic diagram of one of the embodiment of the present application electric load model construction device；
Fig. 4 is a kind of original signal figure of electric load model building method in the embodiment of the present application；
Fig. 5 is to carry out compressed sensing to the original signal of Fig. 4 to sample schematic diagram；
Fig. 6 is to carry out compressed signal schematic representation to the sampled signal in Fig. 5；
Fig. 7 is the recovering signal schematic diagram after restoring with recovery algorithms to the signal in Fig. 6
Fig. 8 is in the embodiment of the present application with pair for restoring signal and original power load signal after compressed sensing sampling algorithm
Than figure.
Specific embodiment
In order to make those skilled in the art more fully understand application scheme, below in conjunction in the embodiment of the present application
Attached drawing, the technical scheme in the embodiment of the application is clearly and completely described, it is clear that described embodiment is only this
Apply for a part of the embodiment, instead of all the embodiments.Based on the embodiment in the application, those of ordinary skill in the art exist
Every other embodiment obtained under the premise of creative work is not made, shall fall in the protection scope of this application.
In order to make it easy to understand, referring to Fig. 1, a kind of electric load model building method provided in the embodiment of the present application,
Include:
Step 101 adopts the electric load signal progress Real Time Compression perception got according to random Gaussian calculation matrix
Sample, the obtained low-dimensional signal comprising whole signal messages are simultaneously transmitted to receiving end.
It should be noted that the selection of calculation matrix should meet limited equidistant property (Restricted Isometry
Property, RIP), for the vector θ that any one has K non-zero entry, all there is a 0 < ε < 1, so that perception square
Battle array ACSMeet:
Due to perceiving matrixWork as calculation matrixWhen irrelevant with sparse matrix φ, then matrix A is perceivedCSGreatly
Probability meets RIP condition, when random gaussian matrix is selected asWhen, random gaussian matrix and most orthogonal group moment
Battle array is all irrelevant, whereinFor calculation matrix, N is that mean value is the Gauss normal distribution that 0 variance is 1/m, and m is through compressed sensing
Compressed data point number.M, which is generally required, to be met condition m >=cKlog (n/K), and c is a constant, generally takes 1~10, the K to be
The degree of rarefication of sparse vector θ, n are the sampled point number before compression.This characteristic can make random gaussian matrix be suitable as
Calculation matrix come using.In the embodiment of the present application, the calculation matrix selected is random Gaussian calculation matrix, to the electric power got
Load signal carries out Real Time Compression perception sampling and refers to whenever ADC converts out a digital signal, then uses calculation matrixCorresponding columnIt multiplies it by, and is superimposed to compressed sensing sampled result y=[y1,
y2,…ym]TIn, signal after m compression so then need to be only stored, without storing n original sampled signal, because of m < <
N, therefore greatly saved memory space.Low-dimensional signal is transmitted to receiving end can be sent out by wirelessly or non-wirelessly communication modes
It send, because of m < < n, has greatly saved transmission data volume and sending time.Wire communication mode includes Ethernet, nothing
Line communication modes include WI-FI, ZigBee, GPRS and bluetooth etc..
If step 102, the low-dimensional signal received are steady-state signal, step 103 and step 104 are executed, if receive
Low-dimensional signal is transient signal, thens follow the steps 105 and step 106.
It should be noted that steady-state signal can be anti-in general, electric load signal is divided into steady-state signal and transient signal
The characteristics of signals under normality is answered, steady-state signal can react the characteristics of signals under instantaneous state, therefore, in the embodiment of the present application,
Low-dimensional signal after compressed sensing is sampled is divided into steady-state signal and transient signal, is located accordingly to both signals respectively
Reason, to reach accuracy requirement.
Step 103, according to Fourier's base sparse matrix recovery algorithms by low-dimensional signal revert to compressed sensing sampling before
First higher-dimension sampled signal.
It should be noted that low-dimensional signal is carried out restoring to be converted into solve following optimal problem:
min||θ||0S.t.y=ACSθ,
Wherein, For random Gaussian calculation matrix, φ is sparse matrix, for steady state voltage, current signal,
φ is selected as Fourier's basic matrix, the expression formula of Fourier's basic matrix are as follows: FJ, k=e2πijk/n=ωJ, k。
l0Greedy algorithm and l including MP and OMP can be used in the solution of optimization problem1Convex relaxed algorithm.The application is implemented
OMP algorithm is selected in example, algorithm flow is as follows:
Input: perception matrix ACS, vector of samples x, degree of rarefication K；
Output: K- sparse θ；
Initialization: residual error r0=y, indexed setT=1；
1. finding out and residual error r and perception matrix ACSColumn vector Aj CSThe corresponding footnote of middle inner product maximum value,
I.e.
2. updating indexed set Λt=Λt-1∪{λt, record the reconstruction atom set in the perception matrix found
3. obtaining θ by least square methodt=argminθ||y-Φtθ||2
4. updating residual error rt=y- Φtθt, t=t+1.
5. judging whether to meet, if satisfied, then stopping iteration；Otherwise it repeats to 1..
Circulation execute 1. -5..
Step 104 obtains signal electrical quantity parameter according to the first higher-dimension sampled signal, and signal electrical quantity parameter includes: letter
Number voltage effective value, signal code virtual value, signal active power, signal reactive power, frequency and harmonic wave, establish Static Load
Model.
It should be noted that the frequency and harmonic wave of signal can be obtained after restoring to complete the first higher-dimension sampled signal
Parameter seeks signal voltage virtual value, voltage effective value formula using voltage effective value formula according to the first higher-dimension sampled signal
Are as follows:Wherein, VrmsIndicate voltage effective value, VnIndicate n-th of discrete voltage instantaneous value,
Δ T indicates the interval sampled every time, and T indicates total sampling time, and N indicates that sampling obtains total discrete voltage instantaneous value number.Benefit
Signal code virtual value, current effective value formula are sought with current effective value formula are as follows:Wherein, IrmsIt indicates
Current effective value, InIndicate that n-th of stray currents instantaneous value, N indicate that sampling obtains total stray currents instantaneous value number.It utilizes
Active power calculates formula and seeks signal active power, and active power calculates formula are as follows:Its
In, P indicates active power, VnIndicate n-th of discrete voltage instantaneous value, InIndicate that n-th of stray currents instantaneous value, N indicate sampling
Obtain total separated data number.Formula is calculated using reactive power and seeks signal reactive power, and reactive power calculates formula are as follows:
Q=P × tan (α), wherein P is active power, and α is power factor angle.Apparent energy is S=Vrms×Irms, wherein VrmsTable
Show voltage effective value, IrmsIndicate current effective value.Power factor (PF) is
Static load model be able to reflect load active power, reactive power it is slowly varying with voltage and frequency and change
Rule, general type are as follows:
In formula, P, Q are respectively active power and reactive power, and U, f are respectively the voltage magnitude and frequency of power feeder, Fp
() and Fq() is linearly or nonlinearly function.
Step 105, low-dimensional signal reverted to according to wavelet basis sparse matrix recovery algorithms before compressed sensing sampling the
Two higher-dimension sampled signals.
It should be noted that being transient signal for low-dimensional signal, utilizing the sparse square of wavelet basis in the embodiment of the present application
Battle array recovery algorithms are restored, the data before being reduced into compression.Based on wavelet basis sparse matrix recovery algorithms process and it is based on Fu
In phyllopodium sparse matrix recovery algorithms it is consistent, unique difference is that sparse matrix is changed to small echo basic matrix by Fourier's basic matrix.
Fourier transformation is suitable for analyzing stable cosine and sine signal, but it does not have localization analysis ability, cannot analyze non-stationary
Signal, and the partial transformation of space (time) and frequency may be implemented in wavelet transformation, therefore is suitble to decompose fast-changing signal.
Step 106 obtains transient signal voltage and transient signal electric current according to the second higher-dimension sampled signal, and it is dynamic to establish load
States model.
It should be noted that the Static Load model in the case where voltage and frequency change greatly, be not enough to describe be
Therefore the part throttle characteristics of system in the embodiment of the present application, has also set up Dynamic Load Model corresponding with transient signal, load is dynamic
States model mainly describes behavioral trait of the aggregate power load when system voltage and frequency quickly change.Load dynamic analog
Type indicates that the electric power of a certain moment load active power and reactive power and preceding several moment (also typically including current time) is presented
Functional relation between line voltage amplitude and frequency, general type are as follows:
In formula, P, Q are respectively active power and reactive power, and U, f are respectively the voltage magnitude and frequency of power feeder, t
For time, Fp() and Fq() is linearly or nonlinearly function.
The electric load model building method provided in the embodiment of the present application, using random Gaussian calculation matrix to power load
Lotus signal carries out compressed sensing sampling, can greatly reduce acquisition data volume, to save storage and communication resources, accelerate
Transmission speed, low-dimensional signal after receiving compressed sensing sampling and then according to low-dimensional signal be to belong to steady-state signal also
It is the standard that transient signal carries out signal with Fourier's base sparse matrix recovery algorithms and wavelet basis sparse matrix recovery algorithms respectively
Really restore, Static Load model and Dynamic Load Model are constructed according to the signal parameter after recovery respectively, has small storage quick
The advantages of transmission and precisely modeling, solve it is existing lack it is a kind of have small store the accurate electric load model that quickly transmits
The technical issues of.
In order to make it easy to understand, referring to Fig. 2, another electric load model building method in the embodiment of the present application, comprising:
Step 201, the electric load signal for acquiring power feeder in real time according to high frequency acquisition device simultaneously carry out AD signal turn
It changes, wherein frequency acquisition meets the requirement of Shannon-Nyquist theorem.
It should be noted that high frequency acquisition device can be data acquisition analysis system, Wide Area Measurement System, failure record
Wave monitors system or other high-frequency intelligent acquisition devices.In the embodiment of the present application, high frequency acquisition is installed first on power feeder
Device acquires electric load voltage, current signal, and it is fixed that the sample frequency of high frequency acquisition device should meet Shannon-Nyquist
Reason requires, i.e., sample frequency cannot be below 2 times of highest frequency in analog signal frequency spectrum, such as considers to obtain preceding 30 subharmonic, then
Sample frequency needs to reach 3000Hz, and otherwise data will appear distortion, influences the accuracy of result.Wherein, high frequency acquisition device
The power feeder installed can be the keys such as substation exit bus, important distribution bifurcated line, to load purulence accuracy
More demanding place.
After collecting electric load signal, digital signal is converted by AD, AD conversion process should be with sample rate
Matching, and guarantee certain precision, at least select 12 or more conversion accuracies.
Step 202 adopts the electric load signal progress Real Time Compression perception got according to random Gaussian calculation matrix
Sample, the obtained low-dimensional signal comprising whole signal messages are simultaneously transmitted to receiving end.
Further, random Gaussian calculation matrix are as follows:
Wherein,For calculation matrix, N is that mean value is the Gauss normal distribution that 0 variance is 1/m, and m is through compressed sensing pressure
Data point number after contracting.
It should be noted that the step 202 in the embodiment of the present application is consistent with the step 101 in a upper embodiment, herein
No longer it is described in detail.
The expression formula of random Gaussian calculation matrix in the embodiment of the present application isIn the embodiment of the present application
Selection sampled signal length is n=1024, signal degree of rarefication K=6, according to data length term of reference m >=cKlog after compression
(n/K), comprehensively consider reduction precision and transmission cost, the value of constant c is unsuitable too small and excessive, generally requires in conjunction with priori
To determine, value is optimal when being 3 for experiment, and when m value is 45, reduction degree can be optimal effect, therefore, in practical survey
It selects to compress electric number in amount to be 45 data points, random gaussian matrix obeys distribution
If step 203, the low-dimensional signal received are steady-state signal, step 204 and step 205 are executed, if receive
Low-dimensional signal is transient signal, thens follow the steps 206 and step 207.
It should be noted that the step 203 in the embodiment of the present application is consistent with the step 102 in a upper embodiment, herein
No longer it is described in detail.
Step 204, according to Fourier's base sparse matrix recovery algorithms by low-dimensional signal revert to compressed sensing sampling before
First higher-dimension sampled signal.
It should be noted that the step 204 in the embodiment of the present application is consistent with the step 103 in a upper embodiment, herein
No longer it is described in detail.
Step 205 obtains signal electrical quantity parameter according to the first higher-dimension sampled signal, and signal electrical quantity parameter includes: letter
Number voltage effective value, signal code virtual value, signal active power, signal reactive power, frequency and harmonic wave, establish Static Load
Model.
It should be noted that the step 205 in the embodiment of the present application is consistent with the step 104 in a upper embodiment, herein
No longer it is described in detail.
Further, Static Load model is ZIP multinomial model.
It should be noted that selecting ZIP multinomial model as Static Load model in the embodiment of the present application.
Further, ZIP multinomial model are as follows:
Wherein, P is active power, and Q is reactive power, and V is voltage value, V0For voltage base value, P0For active power base value,
Q0For reactive power base value, apFor the percentage that the active power of constant-impedance accounts in total active power, bpFor having for constant current
The percentage that function power accounts in total active power, cpIt is accounted in total active power for the active power of constant power load model
Percentage, ap+bp+cp=1；aQFor the percentage that the reactive power of constant-impedance accounts in total reactive power, bQFor constant current
The percentage that reactive power accounts in total reactive power, cQIt is accounted in total reactive power for the reactive power of constant power load model
Percentage, aQ+bQ+cQ=1.
It should be noted that the electrical quantity supplemental characteristic based on steady-state signal, carries out parameter identification using least square method,
Including { ap,bp,cp,aQ,bQ,cQ, Static Load model can be obtained.
Step 206, low-dimensional signal reverted to according to wavelet basis sparse matrix recovery algorithms before compressed sensing sampling the
Two higher-dimension sampled signals.
It should be noted that be transient signal for low-dimensional signal after receiving end receives low-dimensional signal, using being based on
Wavelet basis sparse matrix recovery algorithms restore signal, the data before that compression of reduction planning, the expression of wavelet matrix
Formula are as follows:
In formula, W is wavelet matrix, and ψ (t) is wavelet mother function, and a is scale factor, and b is shift factor.
Step 207, low-dimensional signal reverted to according to wavelet basis sparse matrix recovery algorithms before compressed sensing sampling the
Two higher-dimension sampled signals.
It should be noted that the step 207 in the embodiment of the present application is consistent with the step 106 in a upper embodiment, herein
No longer it is described in detail.
Further, Dynamic Load Model is three rank induction machine parallel connection ZIP models.
It should be noted that selecting three rank induction machine parallel connection ZIP models as load dynamic analog in the embodiment of the present application
Type.
Further, three rank induction machine parallel connection ZIP model are as follows:
State equation:
Output equation:
Wherein,X=Xs+Xm, X '=Xs+Xm//Xr,Aω2+ B ω+C=1, E 'd、E′qIt is respectively equivalent
Motor d axis transient state built-in potential and q axis transient state built-in potential, Id、Iq、Ud、UqThe d axis of respectively equivalent motor and the electric current of q axis
And component of voltage；ω is the revolving speed of equivalent motor, RsFor the stator resistance of equivalent motor, parameter A, B, C are equivalent electronic
The coefficient of machine load torque, H are the inertia time constant of equivalent motor.
It should be noted that voltage, current data based on transient signal, carry out parameter identification using least square method,
Including { ap,bp,cp,aQ,bQ,cQ,Rs,Xs,Xm,Rr,Xr, H, A, B }, Dynamic Load Model can be obtained.
Static Load model and/or Dynamic Load Model are stored in load model database, and real-time update by step 208.
It should be noted that after obtaining Static Load model, Dynamic Load Model, by the Static Load model of generation,
Dynamic Load Model is stored into load model database, and load model database constantly can carry out model according to the data received
Foundation, preservation and update, renewal frequency can reach minute grade.
In order to more specifically be illustrated to the embodiment of the present application, Fig. 2, Fig. 4 are please referred to Fig. 8, Fig. 4 to Fig. 7 is using imitative
Genuine form simulation compressed sensing is as follows from the entire overall process for sampling signal and restoring:
1) taking initial signal fundamental wave is power frequency component, and carries a series of harmonic, and highest number is 24 times；
2) initial signal is sampled, sample frequency 3200Hz, when sampling a length of 0.32s, raw 1024 numbers of common property
Word signaling point；
3) initial signal is compressed using random Gaussian calculation matrix, raw 45 digital signaling points of common property only need
It is stored in and transmits this 45 data points；
4) 45 data points are based in receiving end and recover 1024 digital signaling points again using recovery algorithms.
Fig. 8 is illustrated sampled using compressed sensing after the error of the signal restored and initial signal compare, count as seen from the figure
Accurately ideal according to recovery situation, error is minimum.
The electric load model building method provided in the embodiment of the present application passes through high frequency acquisition device collection voltages, electricity
Stream signal data obtains compressed low-dimensional signal data, is sent to receiving end, receiving end by the compressed sensing method of sampling
It can be cloud server, after receiving data, restore former higher-dimension signal data with recovery algorithms, and establish using the data
Static Load model and Dynamic Load Model, store into load model database, load model database constantly receive data,
For more new model to ensure that the real-time and accuracy of load model, it is steady to send out tidal current analysis, transient stability, voltage to electric system
The accuracy that the electric system simulations such as fixed, low frequency oscillation calculate is of great advantage.
In order to make it easy to understand, referring to Fig. 3, the embodiment of the present application provides a kind of electric load model construction device, packet
It includes:
Compressed sensing sampling unit 301, for according to random Gaussian calculation matrix to the electric load signal got into
The perception sampling of row Real Time Compression, the obtained low-dimensional signal comprising whole signal messages are simultaneously transmitted to receiving end.
Judging unit 302 triggers stable state recovery unit 303 and quiet if low-dimensional signal for receiving is steady-state signal
State modeling unit 304 triggers transient recovery unit 305 and dynamic modeling list if the low-dimensional signal received is transient signal
Member 306.
Stable state recovery unit 303, for low-dimensional signal to be reverted to compression according to Fourier's base sparse matrix recovery algorithms
The first higher-dimension sampled signal before perception sampling.
Static modelling unit 304, for obtaining signal electrical quantity parameter, signal electrical quantity according to the first higher-dimension sampled signal
Parameter includes: signal voltage virtual value, signal code virtual value, signal active power, signal reactive power, frequency and harmonic wave,
Establish Static Load model.
Transient recovery unit 305, for low-dimensional signal to be reverted to compression sense according to wavelet basis sparse matrix recovery algorithms
The second higher-dimension sampled signal before knowing sampling.
Dynamic modeling unit 306, for obtaining transient signal voltage and transient signal electricity according to the second higher-dimension sampled signal
Stream, establishes Dynamic Load Model.
Further, further includes:
Acquisition unit 300, for acquiring the electric load signal of power feeder in real time according to high frequency acquisition device and carrying out
AD signal conversion, wherein frequency acquisition meets the requirement of Shannon-Nyquist theorem.
Database Unit 307, for Static Load model and/or Dynamic Load Model to be stored in load model database,
And real-time update.
Further, Static Load model is ZIP multinomial model.
Further, Dynamic Load Model is three rank induction machine parallel connection ZIP models.
Further, random Gaussian calculation matrix are as follows:
Wherein,For calculation matrix, N is that mean value is the Gauss normal distribution that 0 variance is 1/m, and m is through compressed sensing pressure
Data point number after contracting.
Further, ZIP multinomial model are as follows:
Wherein, P is active power, and Q is reactive power, and V is voltage value, V0For voltage base value, P0For active power base value,
Q0For reactive power base value, apFor the percentage that the active power of constant-impedance accounts in total active power, bpFor having for constant current
The percentage that function power accounts in total active power, cpIt is accounted in total active power for the active power of constant power load model
Percentage, ap+bp+cp=1；aQFor the percentage that the reactive power of constant-impedance accounts in total reactive power, bQFor constant current
The percentage that reactive power accounts in total reactive power, cQIt is accounted in total reactive power for the reactive power of constant power load model
Percentage, aQ+bQ+cQ=1.
Further, three rank induction machine parallel connection ZIP model are as follows:
State equation:
Output equation:
Wherein,X=Xs+Xm, X '=Xs+Xm//Xr, A ω2+ B ω+C=1, E 'd、E′qIt is respectively equivalent
Motor d axis transient state built-in potential and q axis transient state built-in potential, Id、Iq、Ud、UqThe d axis of respectively equivalent motor and the electric current of q axis
And component of voltage；ω is the revolving speed of equivalent motor, RsFor the stator resistance of equivalent motor, parameter A, B, C are equivalent electronic
The coefficient of machine load torque, H are the inertia time constant of equivalent motor.
A kind of electric load model construction equipment is provided in the embodiment of the present application, equipment includes processor and storage
Device:
Program code is transferred to processor for storing program code by memory；
Processor is used for according to the electric load model construction side in the instruction execution embodiment above-mentioned in program code
Method.
The description of the present application and term " first " in above-mentioned attached drawing, " second ", " third ", " the 4th " etc. are (if deposited
) it is to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that use in this way
Data are interchangeable under appropriate circumstances, so that embodiments herein described herein for example can be in addition to illustrating herein
Or the sequence other than those of description is implemented.In addition, term " includes " and " having " and their any deformation, it is intended that
Covering non-exclusive includes to be not necessarily limited to clearly for example, containing the process, method of a series of steps or units, product or equipment
Those of list to Chu step or unit, but may include be not clearly listed or for these process, methods, product or
The intrinsic other step or units of equipment.
In several embodiments provided herein, it should be understood that disclosed system, device and method can be with
It realizes by another way.For example, the apparatus embodiments described above are merely exemplary, for example, the unit
It divides, only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components
It can be combined or can be integrated into another system, or some features can be ignored or not executed.Another point, it is shown or
The mutual coupling, direct-coupling or communication connection discussed can be through some interfaces, the indirect coupling of device or unit
It closes or communicates to connect, can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, each functional unit in each embodiment of the application can integrate in one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, the technical solution of the application is substantially
The all or part of the part that contributes to existing technology or the technical solution can be in the form of software products in other words
It embodies, which is stored in a storage medium, including some instructions are used so that a computer
Equipment (can be personal computer, server or the network equipment etc.) executes the complete of each embodiment the method for the application
Portion or part steps.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (full name in English: Read-Only
Memory, english abbreviation: ROM), random access memory (full name in English: Random Access Memory, english abbreviation:
RAM), the various media that can store program code such as magnetic or disk.
The above, above embodiments are only to illustrate the technical solution of the application, rather than its limitations；Although referring to before
Embodiment is stated the application is described in detail, those skilled in the art should understand that: it still can be to preceding
Technical solution documented by each embodiment is stated to modify or equivalent replacement of some of the technical features；And these
It modifies or replaces, the spirit and scope of each embodiment technical solution of the application that it does not separate the essence of the corresponding technical solution.
Claims (10)
1. a kind of electric load model building method characterized by comprising
101, Real Time Compression perception sampling is carried out to the electric load signal got according to random Gaussian calculation matrix, obtained
Low-dimensional signal comprising whole signal messages is simultaneously transmitted to receiving end；
If the low-dimensional signal 102, received is steady-state signal, step 103 and step 104 are executed, if what is received is described
Low-dimensional signal is transient signal, thens follow the steps 105 and step 106；
103, the low-dimensional signal is reverted to first before compressed sensing sampling according to Fourier's base sparse matrix recovery algorithms
Higher-dimension sampled signal；
104, signal electrical quantity parameter is obtained according to the first higher-dimension sampled signal, the signal electrical quantity parameter includes: letter
Number voltage effective value, signal code virtual value, signal active power, signal reactive power, frequency and harmonic wave, establish Static Load
Model；
105, second before the low-dimensional signal to be reverted to compressed sensing sampling according to wavelet basis sparse matrix recovery algorithms is high
Tie up sampled signal；
106, transient signal voltage and transient signal electric current are obtained according to the second higher-dimension sampled signal, establishes load dynamic analog
Type.
2. electric load model building method according to claim 1, which is characterized in that before step 101 further include:
100, the electric load signal of power feeder is acquired according to high frequency acquisition device in real time and carries out AD signal conversion, wherein
Frequency acquisition meets the requirement of Shannon-Nyquist theorem.
3. electric load model building method according to claim 2, which is characterized in that step 105 and step 106 it
Afterwards, further includes:
107, the Static Load model and/or the Dynamic Load Model are stored in load model database, and real-time update.
4. electric load model building method according to claim 1, which is characterized in that the Static Load model is
ZIP multinomial model.
5. electric load model building method according to claim 1, which is characterized in that the Dynamic Load Model is three
Rank induction machine parallel connection ZIP model.
6. electric load model building method according to claim 1, which is characterized in that the random Gaussian calculation matrix
Are as follows:
Wherein,For calculation matrix, N is that mean value is the Gauss normal distribution that 0 variance is 1/m, and m is compressed through compressed sensing
Data point number.
7. electric load model building method according to claim 4, which is characterized in that the ZIP multinomial model are as follows:
Wherein, P is active power, and Q is reactive power, and V is voltage value, V0For voltage base value, P0For active power base value, Q0For
Reactive power base value, apFor the percentage that the active power of constant-impedance accounts in total active power, bpFor the wattful power of constant current
The percentage that rate accounts in total active power, cpThe percentage accounted in total active power for the active power of constant power load model
Than ap+bp+cp=1；aQFor the percentage that the reactive power of constant-impedance accounts in total reactive power, bQFor the idle of constant current
The percentage that power accounts in total reactive power, cQFor constant power load model reactive power accounted in total reactive power hundred
Divide ratio, aQ+bQ+cQ=1.
8. electric load model building method according to claim 5, which is characterized in that the three ranks induction machine is in parallel
ZIP model are as follows:
State equation:
Output equation:
Wherein,X=Xs+Xm, X '=Xs+Xm//Xr,Aω2+ B ω+C=1, E 'd、E′qIt is respectively equivalent electronic
Machine d axis transient state built-in potential and q axis transient state built-in potential, Id、Iq、Ud、UqThe d axis of respectively equivalent motor and the electric current and electricity of q axis
Press component；ω is the revolving speed of equivalent motor, RsFor the stator resistance of equivalent motor, parameter A, B, C are that equivalent motor is negative
The coefficient of load forces square, H are the inertia time constant of equivalent motor.
9. a kind of electric load model construction device characterized by comprising
Compressed sensing sampling unit, for being pressed in real time according to random Gaussian calculation matrix the electric load signal got
Contracting perception sampling, the obtained low-dimensional signal comprising whole signal messages are simultaneously transmitted to receiving end；
Judging unit triggers stable state recovery unit and static modelling if the low-dimensional signal for receiving is steady-state signal
Unit triggers transient recovery unit and dynamic modeling unit if the low-dimensional signal received is transient signal；
Stable state recovery unit, for the low-dimensional signal to be reverted to compressed sensing according to Fourier's base sparse matrix recovery algorithms
The first higher-dimension sampled signal before sampling；
Static modelling unit, for obtaining signal electrical quantity parameter according to the first higher-dimension sampled signal, the signal is electrical
Amount parameter includes: signal voltage virtual value, signal code virtual value, signal active power, signal reactive power, frequency harmony
Wave establishes Static Load model；
Transient recovery unit is adopted for the low-dimensional signal to be reverted to compressed sensing according to wavelet basis sparse matrix recovery algorithms
The second higher-dimension sampled signal before sample；
Dynamic modeling unit, for obtaining transient signal voltage and transient signal electric current according to the second higher-dimension sampled signal,
Establish Dynamic Load Model.
10. a kind of electric load model construction equipment, which is characterized in that the equipment includes processor and memory:
Said program code is transferred to the processor for storing program code by the memory；
The processor is used for according to the described in any item electric loads of instruction execution claim 1-8 in said program code
Model building method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910049509.8A CN109638830B (en) | 2019-01-18 | 2019-01-18 | Power load model construction method, device and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910049509.8A CN109638830B (en) | 2019-01-18 | 2019-01-18 | Power load model construction method, device and equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109638830A true CN109638830A (en) | 2019-04-16 |
| CN109638830B CN109638830B (en) | 2022-03-22 |
Family
ID=66062147
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910049509.8A Active CN109638830B (en) | 2019-01-18 | 2019-01-18 | Power load model construction method, device and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109638830B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111276978A (en) * | 2020-03-28 | 2020-06-12 | 福建华电万安能源有限公司 | Real-time calculation method for active power shortage of system after fault considering load characteristics |
| CN113949386A (en) * | 2021-09-03 | 2022-01-18 | 国网冀北电力有限公司计量中心 | Electric energy meter compressed sensing dynamic test signal construction method based on symmetric run distribution |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103872678A (en) * | 2014-03-06 | 2014-06-18 | 国家电网公司 | Load model identification method based on transformer substation measurement |
| CN104967114A (en) * | 2015-06-01 | 2015-10-07 | 华南理工大学 | Power grid load real-time digital modeling method and system |
| US20180254662A1 (en) * | 2017-03-01 | 2018-09-06 | University Of Central Florida Research Foundation, Inc. | Adaptive power grid restoration |
| CN108595798A (en) * | 2018-04-11 | 2018-09-28 | 广东电网有限责任公司 | A kind of load model parameters computational methods suitable for electromechanical transient simulation |
-
2019
- 2019-01-18 CN CN201910049509.8A patent/CN109638830B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103872678A (en) * | 2014-03-06 | 2014-06-18 | 国家电网公司 | Load model identification method based on transformer substation measurement |
| CN104967114A (en) * | 2015-06-01 | 2015-10-07 | 华南理工大学 | Power grid load real-time digital modeling method and system |
| US20180254662A1 (en) * | 2017-03-01 | 2018-09-06 | University Of Central Florida Research Foundation, Inc. | Adaptive power grid restoration |
| CN108595798A (en) * | 2018-04-11 | 2018-09-28 | 广东电网有限责任公司 | A kind of load model parameters computational methods suitable for electromechanical transient simulation |
Non-Patent Citations (1)
| Title |
|---|
| 赵茜茜等: "电力系统扰动后稳态频率预测快速算法", 《电力系统保护与控制》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111276978A (en) * | 2020-03-28 | 2020-06-12 | 福建华电万安能源有限公司 | Real-time calculation method for active power shortage of system after fault considering load characteristics |
| CN113949386A (en) * | 2021-09-03 | 2022-01-18 | 国网冀北电力有限公司计量中心 | Electric energy meter compressed sensing dynamic test signal construction method based on symmetric run distribution |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109638830B (en) | 2022-03-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106940679A (en) | Data processing method and device | |
| Li et al. | A comparative analysis of evolutionary and memetic algorithms for community detection from signed social networks | |
| CN111177095A (en) | Log analysis method and device, computer equipment and storage medium | |
| CN105790279B (en) | Reactive voltage partition method based on spectral clustering | |
| Khan et al. | Parallel detrended fluctuation analysis for fast event detection on massive PMU data | |
| CN101841565B (en) | Database cluster system load balancing method and database cluster system | |
| CN103401937B (en) | Daily record data processing method and system | |
| CN109492002A (en) | A kind of storage of smart grid big data and analysis system and processing method | |
| CN104573906B (en) | System and method for analyzing oscillation stability in power transmission system | |
| CN103761309A (en) | Operation data processing method and system | |
| CN107301243A (en) | Switchgear fault signature extracting method based on big data platform | |
| CN104063501B (en) | copy balance method based on HDFS | |
| CN110188825B (en) | Image clustering method, system, device and medium based on discrete multi-view clustering | |
| CN113221475A (en) | Grid self-adaption method for high-precision flow field analysis | |
| CN109638830A (en) | A kind of electric load model building method, device and equipment | |
| US20230066703A1 (en) | Method for estimating structural vibration in real time | |
| CN115833395A (en) | Analysis method and device for running state of power distribution network and online monitoring system | |
| CN115801412A (en) | Method for extracting attack behavior characteristics of power internet of things information network | |
| CN102932847B (en) | Distributed compressed sensing data classification method based on sparse representation classifier | |
| CN109117352B (en) | Server performance prediction method and device | |
| CN108898190A (en) | A kind of environmental monitoring method for early warning and system based on environmentally friendly big data | |
| CN109213742A (en) | Log collection method and device | |
| CN106131238B (en) | The classification method and device of IP address | |
| Liu et al. | A MapReduce based high performance neural network in enabling fast stability assessment of power systems | |
| CN105608638B (en) | A kind of method and its system of the table code data synchronous regime determining intelligent terminal and electric energy meter |
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 |