CN111967165A - Distributed photovoltaic system output estimation method - Google Patents

Distributed photovoltaic system output estimation method Download PDF

Info

Publication number
CN111967165A
CN111967165A CN202010886789.0A CN202010886789A CN111967165A CN 111967165 A CN111967165 A CN 111967165A CN 202010886789 A CN202010886789 A CN 202010886789A CN 111967165 A CN111967165 A CN 111967165A
Authority
CN
China
Prior art keywords
photovoltaic
power
load
actual load
distributed
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
Application number
CN202010886789.0A
Other languages
Chinese (zh)
Other versions
CN111967165B (en
Inventor
王飞
户霖
李康平
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
North China Electric Power University
Original Assignee
North China Electric Power University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by North China Electric Power University filed Critical North China Electric Power University
Priority to CN202010886789.0A priority Critical patent/CN111967165B/en
Publication of CN111967165A publication Critical patent/CN111967165A/en
Application granted granted Critical
Publication of CN111967165B publication Critical patent/CN111967165B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/04—Power grid distribution networks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/50—Photovoltaic [PV] energy
    • Y02E10/56—Power conversion systems, e.g. maximum power point trackers
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • General Health & Medical Sciences (AREA)
  • Operations Research (AREA)
  • Water Supply & Treatment (AREA)
  • Public Health (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Primary Health Care (AREA)
  • Quality & Reliability (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Supply And Distribution Of Alternating Current (AREA)
  • Photovoltaic Devices (AREA)

Abstract

The invention discloses a distributed photovoltaic system output estimation method, which comprises the following steps: establishing a photovoltaic-load decoupling model; constructing a photovoltaic characteristic vector from the generated power of the observable photovoltaic; identifying a distributed photovoltaic system in a power distribution network, and constructing an actual load characteristic vector by using the identified net load power of a user without the distributed photovoltaic system; learning the photovoltaic characteristic vector and the actual load characteristic vector, and updating the photovoltaic characteristic vector and the actual load characteristic vector; the invention provides a method for estimating real-time output of a distributed photovoltaic system in a non-invasive mode, which can accurately estimate the output power of the unmonitored distributed photovoltaic system, is beneficial to improving the net load prediction precision under the penetration distributed photovoltaic and ensures the safe and stable operation of a power system.

Description

Distributed photovoltaic system output estimation method
Technical Field
The invention relates to the technical field of photovoltaics, in particular to a distributed photovoltaic system output estimation method.
Background
Under the severe situation of increasing shortage of non-renewable fossil energy, energy and environmental sustainability has become a focus of worldwide attention. At present, photovoltaic technology, which is focused on solar technology, has undergone tremendous development over the last few years; in 2019, the newly increased installed capacity of photovoltaic power generation in China reaches 3011 ten thousand kilowatts, the newly increased installed capacity of solar photovoltaic power generation in the United states reaches 13.3GW, and the distributed photovoltaic power generation technology is rapidly developed in the global scope.
Since distributed photovoltaics are installed behind the electricity meter and are invisible to grid dispatchers and power retailers, this invisibility presents a challenge to grid operation: after the distributed photovoltaic power generation system is installed, the photovoltaic output power with random fluctuation characteristics and the actual load with random uncertainty are coupled together, so that the predictability of the net load of the power grid is reduced, and the difficulty of predicting the net load of the power grid is increased.
In order to solve the above problems, the real-time output of the distributed photovoltaic needs to be monitored; generally, a special measuring instrument can be installed for each distributed photovoltaic to monitor the distributed photovoltaic output in real time; however, due to the large number of distributed photovoltaic systems installed, the cost of this approach is high; there is a need for a method of estimating distributed photovoltaic real-time output in a non-intrusive manner.
Disclosure of Invention
The invention mainly aims to provide a distributed photovoltaic system output estimation method, aiming at estimating the real-time output of distributed photovoltaic in a non-intrusive mode.
The invention provides a distributed photovoltaic system output estimation method, which comprises the following steps:
establishing a photovoltaic-load decoupling model;
acquiring the power generation power of observable photovoltaic of a plurality of users in the power distribution network; constructing a photovoltaic characteristic vector from the power generation power of the observable photovoltaic;
identifying a distributed photovoltaic system in the power distribution network through the photovoltaic-load decoupling model, and constructing an actual load characteristic vector from the identified net load power of a user without the distributed photovoltaic system;
learning the photovoltaic feature vector and the actual load feature vector through a grid search updating algorithm, and updating the photovoltaic feature vector and the actual load feature vector at each time point;
and estimating the photovoltaic power and the actual load of the power distribution network according to the updated photovoltaic eigenvector and actual load eigenvector and the photovoltaic-load decoupling model.
Preferably, the establishing a photovoltaic-load decoupling model includes:
establishing a photovoltaic-load decoupling model:
Figure BDA0002655798450000021
wherein | · | purple sweet2Is represented by2Norm, which takes the minimized sum of squares of estimation errors as an objective function, comprises two parts: one part is the measured value p of the photovoltaic powerpvWith an estimated value X of the photovoltaic powerpvθpvAnother part is the measured value p of the load powerlWith an estimated value X of the load powerlθlSum of squares of errors of, XpvRepresenting the photovoltaic feature vector, XlRepresenting said actual load characteristic vector, θpvA scaling factor, θ, representing the photovoltaic eigenvector with respect to the actual valuelθlA proportionality coefficient representing said actual load eigenvector with respect to an actual value, a constraint condition representing that the sum of said photovoltaic power and said load power should equal a net load power, a measured value of said photovoltaic power being negative, a measured value of said load power being positive, pnetRepresenting the net load of all users in the distribution network;
introducing a Lagrange operator lambda to construct a Lagrange function L (p)pv,pl,θpv,θl,λ)：
Figure BDA0002655798450000022
Respectively corresponding to the variable p according to the Lagrange extreme value conditionpv、pl、θpvAnd thetalCalculating the partial derivative and making it equal to 0, we can obtain:
Figure BDA0002655798450000023
the following can be obtained in a simultaneous manner:
Figure BDA0002655798450000031
wherein θ ═ θpv,θl]，X＝[Xpv,Xl]；
Obtaining the optimal solution of the proportionality coefficient after solving
Figure BDA0002655798450000032
Namely, it is
Figure BDA0002655798450000033
Optimal solution according to variable theta
Figure BDA0002655798450000034
The photovoltaic power component and the actual load power component can be separated out from the approximate estimation of the net load power:
Figure BDA0002655798450000035
wherein the content of the first and second substances,
Figure BDA0002655798450000036
for the photovoltaic power estimate, a value is determined,
Figure BDA0002655798450000037
the estimated value of the actual load power is obtained;
acquiring the power generation power of observable photovoltaic of a plurality of users in the power distribution network; and constructing a photovoltaic feature vector from the generated power of the observable photovoltaic, comprising:
acquiring the power generation power of observable photovoltaic of a plurality of users in the power distribution network through an ammeter;
adding the power generation power of observable photovoltaic of a plurality of users in the power distribution network to obtain a photovoltaic characteristic vector;
the identifying whether the distributed photovoltaic system is installed in each user distributed photovoltaic system or not through the photovoltaic-load decoupling model identification, and adding the net loads of the identified users without the distributed photovoltaic systems to construct an actual load characteristic vector comprises the following steps:
identifying users in the power distribution network, which are provided with the distributed photovoltaic systems, so as to obtain the users without the distributed photovoltaic systems in the power distribution network;
summing net load powers of users of the non-distributed photovoltaic system within the power distribution network to construct the actual load eigenvector.
Preferably, the identifying the user in the power distribution network who uses the distributed photovoltaic system to obtain the user without the distributed photovoltaic system in the power distribution network includes:
initialization: t ← t 01, t is 1 day, and one user I is selected from known I users using a distributed photovoltaic systemiThen there is Ii∈I；
Decomposing a photovoltaic-load optimal decoupling model: will IiPhotovoltaic power generation power of user as IiPhotovoltaic feature vector of user
Figure BDA0002655798450000038
Will IiThe actual load power of the user is taken as IiUser's actual load feature vector
Figure BDA0002655798450000039
Net load p of all users in the power distribution network through the photovoltaic-load decoupling modelnetDecomposing to obtain IiLight of a userProportional coefficient corresponding to volt sign vector
Figure BDA0002655798450000041
Figure BDA0002655798450000042
Judgment of IiWhether the scaling factor for a user is less than or equal to 0;
if yes, then illustrate IiThe method comprises the steps that a user does not install the distributed photovoltaic system, and J users without the distributed photovoltaic system in the power distribution network are obtained in total;
the voting mechanism comprises the following steps: if IiIf the user does not install the distributed photovoltaic power generation system, voting is 0, otherwise voting is 1;
moving the time window: t ← t +1, with a time interval of 1 day. Repeating the steps from the decomposition of the photovoltaic-load optimal decoupling model to a voting mechanism;
judging whether the number of votes voted for 0 is more than the number of votes voted for 1 in the votes of all the time windows;
if so, adding IiA user determines that the photovoltaic system is not distributed;
if not, the step IiThe user determines that a distributed photovoltaic system is employed.
Preferably, the learning the photovoltaic feature vector and the actual load feature vector through a grid search update algorithm, and updating the photovoltaic feature vector and the actual load feature vector at each time point includes:
initialization: establishing a candidate sample, and setting an initial value of the candidate sample as the photovoltaic power generation power of a known observable photovoltaic user
Figure BDA0002655798450000043
And the actual load of the identified users without distributed photovoltaic systems
Figure BDA0002655798450000044
Photovoltaic hairThe composite sample of the electric power is the sum of the photovoltaic power generation power of the known user, and the composite sample of the actual load is the sum of the actual load of the known user;
distributing the initial value of the photovoltaic characteristic vector as a photovoltaic composite sample, and distributing the actual load characteristic vector as an actual load composite sample, namely:
Figure BDA0002655798450000045
decomposing using the composite sample: mixing XpvAnd XlSubstituting the model into the photovoltaic-load decoupling model, and calculating the estimated net load through the decomposed photovoltaic power and the actual power as follows:
Figure BDA0002655798450000046
wherein the content of the first and second substances,
Figure BDA0002655798450000047
is the estimated payload;
the decomposition error of the composite sample is calculated from the measured and estimated payload as follows:
Figure BDA0002655798450000048
constructing a refined sample through a grid search algorithm: carrying out network search on the I photovoltaic power generation power candidate samples and the J actual load candidate samples to obtain I multiplied by J search samples;
i multiplied by J search samples are used as characteristic vectors and are respectively put into the photovoltaic-load decoupling model, and the photovoltaic power generation power obtained by decomposition is recorded as
Figure BDA0002655798450000051
The actual load obtained by the decomposition is recorded as
Figure BDA0002655798450000052
Estimated payload is noted
Figure BDA0002655798450000053
As follows:
Figure BDA0002655798450000054
the decomposition residual is shown as follows:
Figure BDA0002655798450000055
wherein the content of the first and second substances,
Figure BDA0002655798450000056
representative of the use of I × J of said search samples
Figure BDA0002655798450000057
The result of decomposing a search sample is better than that of decomposing a composite sample,
Figure BDA0002655798450000058
Figure BDA0002655798450000059
the photovoltaic composite sample
Figure BDA00026557984500000510
Namely the photovoltaic carefully-selected sample is obtained,
Figure BDA00026557984500000511
the actual load composite sample of
Figure BDA00026557984500000512
Namely, the actual load fine selection sample;
and (3) updating the composite sample: weighted averaging of photovoltaic concentration samples
Figure BDA00026557984500000513
And XpvAre added to makeFor renewed photovoltaic compound sample, i.e.
Figure BDA00026557984500000514
Actual load refinement sample weighted average
Figure BDA00026557984500000515
And XlAdded as updated real load composite samples, i.e.
Figure BDA00026557984500000516
And returning to use the composite sample for decomposition.
Through above-mentioned technical scheme, can realize following beneficial effect:
the output estimation method of the distributed photovoltaic system provided by the invention provides a method for estimating the real-time output of the distributed photovoltaic system in a non-intrusive mode, can accurately estimate the output power of the unmonitored distributed photovoltaic system, and can help government departments to monitor whether users install the distributed photovoltaic system privately, so that the accuracy of the estimation of the net load when the distributed photovoltaic system exists is improved, and the safe and stable operation of a power system is guaranteed.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the structures shown in the drawings without creative efforts.
Fig. 1 is a flowchart of a first embodiment of a distributed photovoltaic system output estimation method according to the present invention;
fig. 2 is a distributed photovoltaic power generation identification result at different permeabilities of a fourth embodiment of the distributed photovoltaic system output estimation method according to the present invention;
fig. 3 is a photovoltaic power generation recognition result of a distributed photovoltaic power generation recognition model according to a fourth embodiment of the method for estimating the output of the distributed photovoltaic system according to the present invention;
fig. 4 is a sixth embodiment of the method for estimating the output of the distributed photovoltaic system according to the present invention, where the historical data is 10 months and the number of users for installing the distributed photovoltaic system is known to be 20, the error results of the photovoltaic power and the actual load at different permeabilities are obtained;
fig. 5 is a graph illustrating error results of photovoltaic power and actual load at different permeabilities when historical data is 10 months and a known user for installing a distributed photovoltaic system is 20 in a sixth embodiment of the method for estimating output of a distributed photovoltaic system according to the present invention;
fig. 6 shows an error result of the photovoltaic power and the actual load under different historical data when it is known that the number of users installing the distributed photovoltaic system is 20 and the photovoltaic permeability is 40% in a sixth embodiment of the method for estimating the output of the distributed photovoltaic system according to the present invention.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The invention provides a distributed photovoltaic system output estimation method.
As shown in fig. 1, in a first embodiment of the method for estimating the output of the distributed photovoltaic system according to the present invention, the method includes the following steps:
step S110: and establishing a photovoltaic-load decoupling model.
Step S120: acquiring the power generation power of observable photovoltaic of a plurality of users in the power distribution network; and constructing a photovoltaic characteristic vector from the power generation power of the observable photovoltaic.
Step S130: and identifying the distributed photovoltaic system in the power distribution network through the photovoltaic-load decoupling model, and constructing an actual load characteristic vector from the identified net load power of the user without the distributed photovoltaic system.
Step S140: learning the photovoltaic feature vector and the actual load feature vector through a grid search update algorithm, and updating the photovoltaic feature vector and the actual load feature vector at each time point.
Step S150: and estimating the photovoltaic power and the actual load of the power distribution network according to the updated photovoltaic eigenvector, the updated actual load eigenvector and the photovoltaic-load decoupling model.
The output estimation method of the distributed photovoltaic system provided by the invention provides a method for estimating the real-time output of the distributed photovoltaic system in a non-intrusive mode, can accurately estimate the output power of the unmonitored distributed photovoltaic system, and can help government departments to monitor whether users install the distributed photovoltaic system privately, so that the accuracy of the estimation of the net load when the distributed photovoltaic system exists is improved, and the safe and stable operation of a power system is guaranteed.
In addition, the method and the device can also adopt few known conditions as far as possible on the premise of avoiding installation of expensive metering facilities, and separate the output power and the net load data of the unmonitored distributed photovoltaic system, so that the accuracy of net load prediction in the presence of the distributed photovoltaic system is improved.
In a second embodiment of the method for estimating the output of the distributed photovoltaic system, based on the first embodiment, step S110 includes the following steps:
step S210: establishing a photovoltaic-load decoupling model:
Figure BDA0002655798450000071
wherein | · | purple sweet2Is represented by2Norm, which takes the minimized sum of squares of estimation errors as an objective function, comprises two parts: one part is the measured value p of the photovoltaic powerpvWith an estimated value X of the photovoltaic powerpvθpvAnother part is the measured value p of the load powerlWith an estimated value X of the load powerlθlSum of squares of errors of, XpvRepresenting the photovoltaic feature vector, XlRepresenting said actual load characteristic vector, θpvA scaling factor, θ, representing the photovoltaic eigenvector with respect to the actual valuelA proportionality coefficient representing said actual load eigenvector with respect to an actual value, a constraint condition representing that the sum of said photovoltaic power and said load power should equal a net load power, a measured value of said photovoltaic power being negative, a measured value of said load power being positive, pnetRepresenting the net load of all users in the distribution network.
Step S220: introducing a Lagrange operator lambda to construct a Lagrange function L (p)pv,pl,θpv,θl,λ)：
Figure BDA0002655798450000072
Step S230: respectively corresponding to the variable p according to the Lagrange extreme value conditionpv、pl、θpvAnd thetalCalculating the partial derivative and making it equal to 0, we can obtain:
Figure BDA0002655798450000081
the following can be obtained in a simultaneous manner:
Figure BDA0002655798450000082
wherein θ ═ θpv,θl]，X＝[Xpv,Xl]。
Step S240: obtaining the optimal solution of the variable theta after solving
Figure BDA0002655798450000083
And has the following components:
Figure BDA0002655798450000084
step S250: optimal solution according to variable theta
Figure BDA0002655798450000085
Separating the photovoltaic power component and the actual load power component from the approximate estimated payload power:
Figure BDA0002655798450000086
wherein the content of the first and second substances,
Figure BDA0002655798450000087
for the photovoltaic power estimate, a value is determined,
Figure BDA0002655798450000088
is the actual load power estimation value.
Specifically, to achieve accurate decomposition, a suitable feature vector needs to be found. The higher the linear correlation degree of the characteristic vector and the value to be estimated is, the more accurate the estimation result is. Ideally, if a set of eigenvectors of the photovoltaic power and the actual power can be found respectively, so that the eigenvectors are in a linear multiple relation with the actual photovoltaic power and the actual power of the user, the photovoltaic power and the actual power can be estimated completely and accurately in the step S250.
Step S120, including the following steps:
step S260: the method comprises the steps of obtaining the power generation power of observable photovoltaic of a plurality of users in the power distribution network through an ammeter.
Step S270: and adding the power generation power of the observable photovoltaic of a plurality of users in the power distribution network to obtain a photovoltaic characteristic vector.
Specifically, it is assumed herein that a plurality of observable photovoltaic points exist in the power distribution network, and the generated power thereof can be obtained by measuring with an electric meter, and since the photovoltaic power generation curves at different inclination angles and different azimuth angles exhibit a highly linear correlation relationship, the generated power of the plurality of observable photovoltaics in the power distribution network is added to be used as a photovoltaic feature vector.
Step S130, including the steps of:
step S280: and identifying the users adopting the distributed photovoltaic systems in the power distribution network to obtain the users without the distributed photovoltaic systems in the power distribution network.
Step S290: summing net load powers of users of the non-distributed photovoltaic system within the power distribution network to construct the actual load eigenvector.
In a third embodiment of the method for estimating the output of the distributed photovoltaic system, based on the second embodiment, step S280 includes the following steps:
step S310: initialization: t ← t0The time interval of t is 1 day. Selecting a user I from known I users adopting a distributed photovoltaic systemiThen there is Ii∈I。
Step S320: decomposing a photovoltaic-load optimal decoupling model: will I1Photovoltaic power generation power of user as IiPhotovoltaic feature vector of user
Figure BDA0002655798450000091
Will IiThe actual load power of the user is taken as IiUser's actual load feature vector
Figure BDA0002655798450000092
Step S330: net load p of all users in the power distribution network through the photovoltaic-load decoupling modelnetDecomposing to obtain IiProportionality coefficient corresponding to photovoltaic characteristic vector of user
Figure BDA0002655798450000093
Figure BDA0002655798450000094
Step S340: judgment of IiWhether the scaling factor for a user is less than or equal to 0.
If yes, go to step S350: then explain IiAnd the users do not install the distributed photovoltaic system and obtain J users without installing the distributed photovoltaic system in the power distribution network.
Step S360: the voting mechanism comprises the following steps: if IiAnd if the user does not install the distributed photovoltaic power generation system, voting is 0, otherwise voting is 1.
Step S370: moving the time window: t ← t +1, with a time interval of 1 day. The steps S320 to S360 are repeated.
Step S380: it is determined whether the number of votes voted 0 is greater than the number of votes voted 1 in all the time windows.
If yes, go to step S381: will IiThe user determines that there is no distributed photovoltaic system.
If not, go to step S382: will IiThe user determines that a distributed photovoltaic system is employed.
In a fourth embodiment of the method for estimating output of a distributed photovoltaic system according to the present invention, based on the third embodiment, step S382, the following steps are further included:
step S410: adopting a confusion matrix as distributed photovoltaic identification, wherein the expression of the confusion matrix is as follows:
Figure BDA0002655798450000101
where the confusion matrix is a standard format for representing accuracy evaluation, the confusion matrix is typically used to evaluate the classification accuracy of classification models, which contains all information about actual classes and predicted classes, mijIs the number of objects that belong to the ith class but are classified into the jth class, where n is the total number of objects.
Step S420: three metrics are defined to describe the classification accuracy: the 3 classification accuracies are respectively: product's Accuracy (PA), User's Accuracy (UA), and Overall Accuracy (OA):
Figure BDA0002655798450000102
Figure BDA0002655798450000103
Figure BDA0002655798450000104
the user precision PA is used for describing the accuracy of the classification of an actual object of a certain class; the product precision UA is the accuracy of a certain class of classification results output by the model; the overall accuracy OA describes the accuracy of all classification results.
Specifically, in one example, the historical data is 10 months, the average accuracy of the corresponding recognition results under different permeabilities is shown in fig. 2, and the permeability in this patent example is defined as the ratio of the total photovoltaic power generation amount to the total power consumption amount of the power distribution network.
In the same example, the permeability is 40%, and the average accuracy of the corresponding recognition results under different historical data is shown in the following fig. 3.
In a fifth embodiment of the method for estimating output of a distributed photovoltaic system according to the present invention, based on the third embodiment, step S140 includes the following steps:
step S510: initialization: establishing a candidate sample, and setting the initial value of the candidate sample plate as the photovoltaic power generation power of a known user
Figure BDA0002655798450000105
And the actual load of the identified users without distributed photovoltaic systems
Figure BDA0002655798450000106
The composite sample of the photovoltaic power generation power is the sum of the photovoltaic power generation power of the known user, and the composite sample of the actual load is the sum of the actual load of the known user.
Step S520: distributing the initial value of the photovoltaic characteristic vector as a photovoltaic composite sample, and distributing the actual load characteristic vector as an actual load composite sample, namely:
Figure BDA0002655798450000111
step S530: decomposing using the composite sample: mixing XpvAnd XlSubstituting the model into the photovoltaic-load decoupling model, and calculating the estimated net load through the decomposed photovoltaic power and the actual power as follows:
Figure BDA0002655798450000112
wherein the content of the first and second substances,
Figure BDA0002655798450000113
is the estimated payload.
Step S540: the decomposition error of the composite sample is calculated from the measured and estimated payload as follows:
Figure BDA0002655798450000114
step S550: constructing a refined sample through a grid search algorithm: and carrying out network search on the I photovoltaic power generation power candidate samples and the J actual load candidate samples to obtain I multiplied by J search samples.
Step S560: i multiplied by J search samples are used as characteristic vectors and are respectively put into the photovoltaic-load decoupling model, and the photovoltaic power generation power obtained by decomposition is recorded as
Figure BDA0002655798450000115
The actual load obtained by the decomposition is recorded as
Figure BDA0002655798450000116
Estimated payload is noted
Figure BDA0002655798450000117
As follows:
Figure BDA0002655798450000118
step S570: the decomposition residual is shown as follows:
Figure BDA0002655798450000119
wherein the content of the first and second substances,
Figure BDA00026557984500001110
representative of the use of I × J of said search samples
Figure BDA00026557984500001111
The result of decomposing a search sample is better than that of decomposing a composite sample,
Figure BDA00026557984500001112
Figure BDA00026557984500001113
the photovoltaic composite sample
Figure BDA00026557984500001114
Namely the photovoltaic carefully-selected sample is obtained,
Figure BDA00026557984500001115
the actual load composite sample of
Figure BDA00026557984500001116
Namely the actual load refining sample.
Step S580: updating the composite sample for decomposition: weighted averaging of photovoltaic concentration samples
Figure BDA00026557984500001117
And XpvAdded as a renewed photovoltaic composite sample, i.e.
Figure BDA00026557984500001118
Actual load refinement sample weighted average
Figure BDA00026557984500001119
And XlAdded as updated real load composite samples, i.e.
Figure BDA00026557984500001120
Step S590: returning to step S530.
In a sixth embodiment of the method for estimating the output of the distributed photovoltaic system, based on the second embodiment, in any of the embodiments, distributed photovoltaic decomposition is performed under different numbers of photovoltaic users, different permeabilities, and different historical data, and a Mean Absolute Percentage Error (MAPE) is used to evaluate the accuracy of a decomposition result.
Specifically, in order to illustrate the effectiveness of the method, the distributed photovoltaic decomposition is performed by adopting the method for estimating the output of the distributed photovoltaic system provided by the invention under different photovoltaic user numbers, different permeabilities and different historical data, and the accuracy of the decomposition result is evaluated by using Mean Absolute Percentage Error (MAPE).
As shown in the following formula:
Figure BDA0002655798450000121
wherein, MAPEPV(%) is the decomposition error structure of photovoltaic power without grid search update mechanism, MAPELoad(%) is the result of the decomposition error of the actual load without the grid search update mechanism,
Figure BDA0002655798450000122
for the decomposition error result of the photovoltaic power in the case of the grid search update mechanism,
Figure BDA0002655798450000123
decomposing error results for actual load with grid search update mechanism, wherein ytIn the form of an actual value of the value,
Figure BDA0002655798450000124
for the estimation, N is the data length.
When the historical data is 10 months and the photovoltaic permeability is 40%, under the condition of different known users for installing the distributed photovoltaic system, the error results of the photovoltaic power and the actual load are shown in the attached figure 4.
When the historical data is 10 months and the user of the installed distributed photovoltaic system is known to be 20, the error results of the photovoltaic power and the actual load at different permeabilities are shown in fig. 5.
When the user of the installed distributed photovoltaic system is known to be 20 and the photovoltaic permeability is known to be 40%, the error results of the photovoltaic power and the actual load under different historical data are shown in fig. 6.
As can be seen from the attached drawings 4-6, through practical tests, the output estimation method for the distributed photovoltaic system provided by the invention has an ideal effect and higher estimation precision.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, wherein the software product is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk), and includes instructions for enabling a terminal (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.

Claims (4)

1. A distributed photovoltaic system output estimation method is characterized by comprising the following steps:
establishing a photovoltaic-load decoupling model;
acquiring the power generation power of observable photovoltaic of a plurality of users in the power distribution network; constructing a photovoltaic characteristic vector from the power generation power of the observable photovoltaic;
identifying a distributed photovoltaic system in the power distribution network through the photovoltaic-load decoupling model, and constructing an actual load characteristic vector from the identified net load power of a user without the distributed photovoltaic system;
learning the photovoltaic feature vector and the actual load feature vector through a grid search updating algorithm, and updating the photovoltaic feature vector and the actual load feature vector at each time point;
and estimating the photovoltaic power and the actual load of the power distribution network according to the updated photovoltaic eigenvector and actual load eigenvector and the photovoltaic-load decoupling model.
2. The distributed photovoltaic system contribution estimation method of claim 1, wherein said building a photovoltaic-load decoupling model comprises:
establishing a photovoltaic-load decoupling model:
Figure FDA0002655798440000011
wherein | · | purple sweet2Is represented by2Norm, which takes the minimized sum of squares of estimation errors as an objective function, comprises two parts: one part is the measured value p of the photovoltaic powerpvWith an estimated value X of the photovoltaic powerpvθpvAnother part is the measured value p of the load powerlWith an estimated value X of the load powerlθlSum of squares of errors of, XpvRepresents the photovoltaicEigenvectors, XlRepresenting said actual load characteristic vector, θpvA scaling factor, θ, representing the photovoltaic eigenvector with respect to the actual valuelθlA proportionality coefficient representing said actual load eigenvector with respect to an actual value, a constraint condition representing that the sum of said photovoltaic power and said load power should equal a net load power, a measured value of said photovoltaic power being negative, a measured value of said load power being positive, pnetRepresenting the net load of all users in the distribution network;
introducing a Lagrange operator lambda to construct a Lagrange function L (p)pv,pl,θpv,θl,λ)：
Figure FDA0002655798440000021
Respectively corresponding to the variable p according to the Lagrange extreme value conditionpv、pl、θpvAnd thetalCalculating the partial derivative and making it equal to 0, we can obtain:
Figure FDA0002655798440000022
the following can be obtained in a simultaneous manner:
Figure FDA0002655798440000023
wherein θ ═ θpv,θl]，X＝[Xpv,Xl]；
Obtaining the optimal solution of the proportionality coefficient after solving
Figure FDA0002655798440000024
Namely, it is
Figure FDA0002655798440000025
Optimal solution according to variable theta
Figure FDA0002655798440000026
Separating the photovoltaic power component and the actual load power component from the approximate estimated payload power:
Figure FDA0002655798440000027
wherein the content of the first and second substances,
Figure FDA0002655798440000028
for the photovoltaic power estimate, a value is determined,
Figure FDA0002655798440000029
the estimated value of the actual load power is obtained;
acquiring the power generation power of observable photovoltaic of a plurality of users in the power distribution network; and constructing a photovoltaic feature vector from the generated power of the observable photovoltaic, comprising:
acquiring the power generation power of observable photovoltaic of a plurality of users in the power distribution network through an ammeter;
adding the power generation power of observable photovoltaic of a plurality of users in the power distribution network to obtain a photovoltaic characteristic vector;
the identifying whether the distributed photovoltaic system is installed in each user distributed photovoltaic system or not through the photovoltaic-load decoupling model identification, and adding the net loads of the identified users without the distributed photovoltaic systems to construct an actual load characteristic vector comprises the following steps:
identifying users in the power distribution network, which are provided with the distributed photovoltaic systems, so as to obtain the users without the distributed photovoltaic systems in the power distribution network;
summing net load powers of users of the non-distributed photovoltaic system within the power distribution network to construct the actual load eigenvector.
3. The method of claim 2, wherein identifying the users within the distribution network who use the distributed photovoltaic system to obtain the users without the distributed photovoltaic system within the distribution network comprises:
initialization: t ← t01, t is 1 day, and one user I is selected from known I users using a distributed photovoltaic systemiThen there is Ii∈I；
Decomposing a photovoltaic-load optimal decoupling model: will IiPhotovoltaic power generation power of user as IiPhotovoltaic feature vector of user
Figure FDA0002655798440000031
Will IiThe actual load power of the user is taken as IiUser's actual load feature vector
Figure FDA0002655798440000032
Net load p of all users in the power distribution network through the photovoltaic-load decoupling modelnetDecomposing to obtain IiProportionality coefficient corresponding to photovoltaic characteristic vector of user
Figure FDA0002655798440000033
Figure FDA0002655798440000034
Judgment of IiWhether the scaling factor for a user is less than or equal to 0;
if yes, then illustrate IiThe method comprises the steps that a user does not install the distributed photovoltaic system, and J users without the distributed photovoltaic system in the power distribution network are obtained in total;
the voting mechanism comprises the following steps: if IiIf the user does not install the distributed photovoltaic power generation system, voting is 0, otherwise voting is 1;
moving the time window: t ← t +1, with a time interval of 1 day. Repeating the steps from the decomposition of the photovoltaic-load optimal decoupling model to a voting mechanism;
judging whether the number of votes voted for 0 is more than the number of votes voted for 1 in the votes of all the time windows;
if so, adding IiA user determines that the photovoltaic system is not distributed;
if not, the step IiThe user determines that a distributed photovoltaic system is employed.
4. The distributed photovoltaic system output estimation method according to claim 3, wherein the learning the photovoltaic feature vector and the actual load feature vector through a grid search update algorithm, and updating the photovoltaic feature vector and the actual load feature vector at each time point comprises:
initialization: establishing a candidate sample, and setting an initial value of the candidate sample as the photovoltaic power generation power of a known observable photovoltaic user
Figure FDA0002655798440000035
And the actual load of the identified users without distributed photovoltaic systems
Figure FDA0002655798440000036
The composite sample of the photovoltaic power generation power is the sum of the photovoltaic power generation power of the known user, and the composite sample of the actual load is the sum of the actual load of the known user;
distributing the initial value of the photovoltaic characteristic vector as a photovoltaic composite sample, and distributing the actual load characteristic vector as an actual load composite sample, namely:
Figure FDA0002655798440000041
decomposing using the composite sample: mixing XpvAnd XlSubstituting the model into the photovoltaic-load decoupling model, and calculating the estimated net load through the decomposed photovoltaic power and the actual power as follows:
Figure FDA0002655798440000042
wherein the content of the first and second substances,
Figure FDA0002655798440000043
is the estimated payload;
the decomposition error of the composite sample is calculated from the measured and estimated payload as follows:
Figure FDA0002655798440000044
constructing a refined sample through a grid search algorithm: carrying out network search on the I photovoltaic power generation power candidate samples and the J actual load candidate samples to obtain I multiplied by J search samples;
i multiplied by J search samples are used as characteristic vectors and are respectively put into the photovoltaic-load decoupling model, and the photovoltaic power generation power obtained by decomposition is recorded as
Figure FDA0002655798440000045
The actual load obtained by the decomposition is recorded as
Figure FDA0002655798440000046
Estimated payload is noted
Figure FDA0002655798440000047
As follows:
Figure FDA0002655798440000048
the decomposition residual is shown as follows:
Figure FDA0002655798440000049
wherein the content of the first and second substances,
Figure FDA00026557984400000410
representative of the use of I × J of said search samples
Figure FDA00026557984400000411
The result of decomposing a search sample is better than that of decomposing a composite sample,
Figure FDA00026557984400000412
Figure FDA00026557984400000413
the photovoltaic composite sample
Figure FDA00026557984400000414
Namely the photovoltaic carefully-selected sample is obtained,
Figure FDA00026557984400000415
the actual load composite sample of
Figure FDA00026557984400000416
Namely, the actual load fine selection sample;
and (3) updating the composite sample: weighted averaging of photovoltaic concentration samples
Figure FDA00026557984400000417
And XpvAdded as a renewed photovoltaic composite sample, i.e.
Figure FDA00026557984400000418
Actual load refinement sample weighted average
Figure FDA00026557984400000419
And XlAdded as updated real load composite samples, i.e.
Figure FDA00026557984400000420
And returning to the decomposition by using the composite sample.
CN202010886789.0A 2020-08-28 2020-08-28 Output estimation method of distributed photovoltaic system Active CN111967165B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010886789.0A CN111967165B (en) 2020-08-28 2020-08-28 Output estimation method of distributed photovoltaic system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010886789.0A CN111967165B (en) 2020-08-28 2020-08-28 Output estimation method of distributed photovoltaic system

Publications (2)

Publication Number Publication Date
CN111967165A true CN111967165A (en) 2020-11-20
CN111967165B CN111967165B (en) 2023-07-21

Family

ID=73400903

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010886789.0A Active CN111967165B (en) 2020-08-28 2020-08-28 Output estimation method of distributed photovoltaic system

Country Status (1)

Country Link
CN (1) CN111967165B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109698512A (en) * 2019-02-19 2019-04-30 武汉大学 A kind of photovoltaic concentration grid-connected system asynchronous oscillation Center shift characteristic analysis method
CN110610475A (en) * 2019-07-07 2019-12-24 河北工业大学 Visual defect detection method of deep convolutional neural network
JP2021125910A (en) * 2020-02-03 2021-08-30 株式会社日立製作所 Actual load calculation device, actual load calculation method, and physical quantity calculation method of power system

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109698512A (en) * 2019-02-19 2019-04-30 武汉大学 A kind of photovoltaic concentration grid-connected system asynchronous oscillation Center shift characteristic analysis method
CN110610475A (en) * 2019-07-07 2019-12-24 河北工业大学 Visual defect detection method of deep convolutional neural network
JP2021125910A (en) * 2020-02-03 2021-08-30 株式会社日立製作所 Actual load calculation device, actual load calculation method, and physical quantity calculation method of power system

Non-Patent Citations (6)

* Cited by examiner, † Cited by third party
Title
ZHIMING XUAN 等: "PV-Load Decoupling Based Demand Response Baseline Load Estimation Approach for Residential Customer with Distributed PV System" *
党克;王民涛;: "光伏微电网供电在线负荷频率控制" *
党克;王民涛;: "光伏微电网供电在线负荷频率控制", 计算机仿真 *
冯昌 等: "基于模型解耦的协调控制系统研究" *
王飞;米增强;杨奇逊;赵洪山;: "基于神经网络与关联数据的光伏电站发电功率预测方法" *
王飞;米增强;杨奇逊;赵洪山;: "基于神经网络与关联数据的光伏电站发电功率预测方法", 太阳能学报 *

Also Published As

Publication number Publication date
CN111967165B (en) 2023-07-21

Similar Documents

Publication Publication Date Title
Sun et al. Using Bayesian deep learning to capture uncertainty for residential net load forecasting
Sheng et al. Short-term solar power forecasting based on weighted Gaussian process regression
CN109142171B (en) Urban PM10 concentration prediction method based on feature expansion and fusing with neural network
Pan et al. Performance degradation prediction of proton exchange membrane fuel cell using a hybrid prognostic approach
Shin et al. Heterogeneous mixture distributions for modeling wind speed, application to the UAE
CN102185735B (en) Network security situation prediction method
Khan et al. Assessing hydrologic impact of climate change with uncertainty estimates: Bayesian neural network approach
CN103488869A (en) Wind power generation short-term load forecast method of least squares support vector machine
Zhang et al. Accelerated evaluation of autonomous vehicles in the lane change scenario based on subset simulation technique
CN104951834A (en) LSSVM (least squares support vector machine) wind speed forecasting method based on integration of GA (genetic algorithm) and PSO (particle swarm optimization)
CN109241493A (en) Key Performance Indicator flexible measurement method based on Markov random field and EM algorithm
CN113283491A (en) Fault diagnosis method of electric vehicle alternating current charging pile based on optimized deep confidence network
Straka et al. Analysis of energy consumption at slow charging infrastructure for electric vehicles
CN109978236B (en) Small hydropower station short-term power refined prediction method based on feature combination
CN115545333A (en) Method for predicting load curve of multi-load daily-type power distribution network
CN115099511A (en) Photovoltaic power probability estimation method and system based on optimized copula
Zheng et al. Improved iterative prediction for multiple stop arrival time using a support vector machine
CN114139940A (en) Generalized demand side resource network load interaction level assessment method based on combined empowerment-cloud model
CN113886454A (en) Cloud resource prediction method based on LSTM-RBF
CN109543922B (en) Time-period borrowing and returning amount prediction method for single-vehicle station group shared by piles
CN111967165A (en) Distributed photovoltaic system output estimation method
CN116167465A (en) Solar irradiance prediction method based on multivariate time series ensemble learning
Valgaev et al. Building power demand forecasting using k-nearest neighbors model-initial approach
Oesting et al. Sampling from a Max‐Stable Process Conditional on a Homogeneous Functional with an Application for Downscaling Climate Data
CN116663889A (en) Novel power system risk assessment method based on improved Gaussian model

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