CN110765703A - Wind power plant aggregation characteristic modeling method - Google Patents

Wind power plant aggregation characteristic modeling method Download PDF

Info

Publication number
CN110765703A
CN110765703A CN201911066604.5A CN201911066604A CN110765703A CN 110765703 A CN110765703 A CN 110765703A CN 201911066604 A CN201911066604 A CN 201911066604A CN 110765703 A CN110765703 A CN 110765703A
Authority
CN
China
Prior art keywords
wind
power
output
wind turbine
layer
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
CN201911066604.5A
Other languages
Chinese (zh)
Other versions
CN110765703B (en
Inventor
尹柏清
丛雨
阎洁
高晨
王琪
高鑫哲
米夏
刘永前
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
Inner Mongolia Electric Power Research Institute of Inner Mongolia Power Group Co Ltd
Original Assignee
North China Electric Power University
Inner Mongolia Electric Power Research Institute of Inner Mongolia Power Group Co Ltd
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, Inner Mongolia Electric Power Research Institute of Inner Mongolia Power Group Co Ltd filed Critical North China Electric Power University
Priority to CN201911066604.5A priority Critical patent/CN110765703B/en
Publication of CN110765703A publication Critical patent/CN110765703A/en
Application granted granted Critical
Publication of CN110765703B publication Critical patent/CN110765703B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of 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
    • 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)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Wind Motors (AREA)

Abstract

The invention discloses a wind power plant aggregation characteristic modeling method, which comprises the following steps: the method comprises the following steps: collecting actual measurement operation data of a plurality of wind turbine generators of a wind power plant, and cleaning and normalizing the data; step two: establishing a wind power fluctuation measurement index; step three: establishing a smoothing effect measurement index of the aggregate output of the wind power plant, and obtaining a relational expression of the smoothing effect measurement index, the number N of the polymer units and the correlation coefficient of the power sequence among the units; step four: establishing a mapping model of multi-position point wind condition information and inter-unit power sequence correlation based on a convolutional neural network; step five: forming a model training sample, training a neural network model by using a root mean square error function index, and outputting a unit output correlation mapping result; step six: and realizing the wind power plant aggregation characteristic modeling based on the convolutional neural network.

Description

Wind power plant aggregation characteristic modeling method
Technical Field
The invention relates to the technical field of new energy power generation, in particular to a wind power plant aggregation characteristic modeling method.
Background
The large-scale wind power access to the power grid is one of the prominent features of the future development of the power system. With the annual increase of the power supply duty ratio of wind power in various countries in the world, the grid-connected wind power has the characteristics of cluster development, weak power grid access and long-distance delivery, the inherent randomness and volatility of the wind power bring a series of problems of power balance, reverse peak regulation, voltage stability and frequency stability to a power system, and the difficulty of power grid planning and scheduling is increased. The smoothing effect is one of common characteristics of large-scale wind power, and is expressed in that the wind power aggregate output is smoother and less fluctuating than the individual output, and the fluctuation of the wind power output is smoothed by a method of regional diversification, so that the key basis for solving the problems is to know and master the wind power aggregate characteristics, particularly the fluctuation rule of the wind power aggregate output, and model the fluctuation rule.
The difference of wind resource distribution is a key factor for generating a smoothing effect, and in essence, due to the spatial dispersion of the wind generation sets in the wind power plant, the wind resources at different wind generation set point positions in the wind power plant show the difference, and the output of each wind generation set changes along with the wind speed and then counteracts the fluctuation of the other wind generation set through the synergistic decoupling effect under certain conditions, so that the fluctuation of the wind power is reduced on the whole. Therefore, the fluctuation and the smooth effect of the aggregated power of the wind power plant are closely related to factors such as the regional scale of the wind power plant, the number of wind power plants of the wind power plant, the wind condition of the wind power plant, the output correlation among the wind power plants of the wind power plant and the like, and the factors influence each other. Most of the existing researches at home and abroad of wind turbine generators describe and qualitatively analyze the fluctuation and smooth effect rules of wind power polymerization, and mainly draw the following conclusions: 1. the wind power aggregation volatility can be reduced and the smoothing effect can be enhanced by expanding the wind power scale (the regional scale and the number of the units). 2. The stronger the correlation of the output force among the wind turbine generator sets, the poorer the complementarity and the poorer the smoothing effect, and otherwise, the better the smoothing effect. A few existing quantitative researches on wind power aggregation volatility are only based on observed historical data, and simple linear or nonlinear fitting is carried out on functional relations between the wind power aggregation volatility and various influence factors. However, the power aggregation characteristic of the wind power plant is a complex space-time coupling process influenced by multivariable factors such as the terrain in the wind power plant, the unit arrangement, the wind conditions at the point positions of multiple units and the like, the specific mapping relation is complex nonlinear mapping, and the relation between each influencing factor and the power aggregation characteristic of the wind power plant is difficult to accurately describe through simple linear or nonlinear fitting.
Therefore, a modeling method for the aggregation characteristic of the wind power plant is expected to solve the problems in the prior art.
Disclosure of Invention
The invention discloses a modeling method for aggregation characteristics of a wind power plant, which comprises the following steps:
the method comprises the following steps: collecting actual measurement operation data of a plurality of wind turbine generators of a wind power plant, and cleaning and normalizing the data;
step two: establishing a wind power fluctuation measurement index;
step three: establishing a smoothing effect measurement index of the aggregate output of the wind power plant, and obtaining a relational expression of the smoothing effect measurement index, the number N of the polymer units and the correlation coefficient of the power sequence among the units by a mathematical statistic analysis method;
step four: establishing a mapping model of multi-position point wind condition information and inter-unit power sequence correlation based on a convolutional neural network, and modeling the number of units according to the time number in a specified time period and the aggregation characteristic of the wind power plant to set corresponding parameters of the convolutional neural network;
step five: taking actual measurement wind speed sequence data and actual measurement wind direction sequence data of multiple unit point positions with a specified time scale as model input, taking output correlation coefficients between every two units as output to form a model training sample, training a neural network model by using root mean square error function indexes, and outputting unit output correlation mapping results;
step six: and calculating a smooth effect measurement index s representing the aggregation characteristic of the wind power plant according to the unit output correlation mapping result in the step five and a functional relation between the smooth effect measurement index in the step three and the number N of the aggregation units and the correlation coefficient of the power sequences among the units, and realizing the modeling of the aggregation characteristic of the wind power plant based on the convolutional neural network.
Preferably, the step of measuring the operation data includes: measured wind speed data, measured wind direction data and measured power data.
Preferably, the aggregate output of the wind turbine generator cluster is the sum of the outputs of the single machines, as shown in formula (1):
Figure BDA0002259564710000031
in the formula, P∑(t) is the total power of the wind turbine generator cluster at the moment t; i is the number of the wind turbine generator; n is the number of the wind turbine generators; pi(t) is the power of the No. i wind turbine generator at the moment t;
in the second step, the standard deviation of the power sequence is expressed by formulas (2) and (3) to measure the fluctuation of the output of a single wind turbine generator and the aggregate output of the wind turbine generators:
Figure BDA0002259564710000032
Figure BDA0002259564710000033
in the formula, σi、σ∑Are respectively No. i windThe standard deviation of the power and the standard deviation of the aggregate power of the motor set; t is a statistical time scale;
Figure BDA0002259564710000034
respectively taking the mean values of the sampling values of the power and the aggregate power of the No. i wind turbine generator under the corresponding time statistical scale; the standard deviation of the power sequence reflects the degree of power fluctuation of the time sequence, and the smaller the numerical value is, the smaller the power fluctuation of the time sequence is, and the better the stationarity is.
Preferably, the ratio of the per unit value of the aggregate output standard deviation of the wind turbine generator with the installed capacity as a base value to the per unit value of the output standard deviation of the single unit is formula (4):
Figure BDA0002259564710000035
wherein, PRFor the rated power, sigma, of the wind turbinei、σ∑The power standard deviation and the polymerization power standard deviation of the No. i wind turbine generator are respectively shown, N represents the number of the polymerization generator sets, and the smaller the coefficient s is, the stronger the complementarity among the output forces of the wind turbine generators is, and the more obvious the smoothing effect is;
and (3) obtaining a relational formula (5) of the standard deviation of the aggregate power and the standard deviation of the single machine power of the wind turbine generator through the three mathematical statistics:
Figure BDA0002259564710000036
in the formula, ri,jThe correlation coefficient of the power sequences of the No. i wind turbine generator and the No. j wind turbine generator is obtained;
if the output standard deviations of the units are the same and are regarded as the average value of the output standard deviations of the single unit, the correlation coefficient relation of the smooth effect measurement index, the number N of the polymerization units and the power sequence among the units is obtained by the formulas (4) and (5) and is a formula (6):
Figure BDA0002259564710000041
preferably, the convolutional neural network of step four comprises two convolutional layers, two pooling layers, a flattening layer and a full-link layer; the method comprises the steps that actually measured wind speed sequence data and actually measured wind direction sequence data of multiple sets pass through a first convolution layer and a ReLU activation layer to generate a set of characteristic graphs, then non-overlapping maximum pooling is conducted to carry out down-sampling, then the second convolution layer and the ReLU activation layer are used for generating another set of characteristic graphs, the set of characteristic graphs are connected with a flattening layer to enable a multi-dimensional tensor to be one-dimensional, then the characteristic graphs are connected with a full connection layer, and finally the characteristic graphs are activated through a sigmoid function to serve as output of a convolutional neural network.
Preferably, in the fourth step, a normalization layer is added after the partial-layer output by using a batch normalization method, the output is normalized to the same distribution according to the feature values of the same batch, and a convolution operation process is represented by a formula (7):
Figure BDA0002259564710000042
wherein the content of the first and second substances,
Figure BDA0002259564710000043
the j-th characteristic diagram in the L-th convolutional layer, namely the wind speed distribution characteristic extracted by the convolutional layer,
Figure BDA0002259564710000044
is the feature map set of the previous layer, which is also the input of this layer,for the jth convolution kernel in the lth convolution layer, a convolution operation is defined, b represents an additive bias value, and f (x) represents an activation function which maps a value in a real number domain into a finite field;
the operation process of the pooling layer is expressed by formula (8):
Figure BDA0002259564710000046
wherein the content of the first and second substances,
Figure BDA0002259564710000047
denotes the jth feature map in the lth pooling layer, f (x) denotes the activation function,
Figure BDA0002259564710000048
a multiplicative bias representing a profile, down (x) is a down-sampling function,
Figure BDA0002259564710000049
showing the additive bias of the signature.
Preferably, the calculation formula of the output correlation coefficient between every two units in the step five is formula (9):
Figure BDA00022595647100000410
wherein sigmaxyRefers to the covariance, σ, of the two data sequences X and Yx、σyThe respective variances of X and Y,
Figure BDA0002259564710000051
respectively, mean values, X, of the sequence X, Yi、YiRespectively, the ith number of the sequence X, Y.
The invention provides a wind power plant aggregation characteristic modeling method, which is based on a convolutional neural network, is derived from a deep learning method, is combined with a mathematical model obtained by an existing analysis method through mathematical statistics, can adapt to the training of a large data volume sample, and has stronger generalization capability; compared with the traditional simple linear or nonlinear fitting method, the wind power plant power aggregation characteristic is a complex time-space coupling process influenced by multivariable factors such as the shape in a wind power plant, the unit arrangement, the wind conditions at the point positions of multiple units and the like, and the specific mapping relation is complex nonlinear mapping.
Drawings
FIG. 1 is a flow chart of a wind farm aggregate characteristic modeling method
FIG. 2 is a schematic structural diagram of a wind turbine generator output correlation mapping model based on a convolutional neural network.
Detailed Description
In order to make the implementation objects, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be described in more detail below with reference to the accompanying drawings in the embodiments of the present invention. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are only some, but not all embodiments of the invention. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The method takes the actually measured wind speed data, actually measured wind direction data and actually measured power data of 5 wind power generating sets in a certain wind power plant in the north of China as a test example, the example collects the actually measured wind speed data, the actually measured wind direction data and the actually measured power data from 2016 to 2017 for 11 months, and the data time resolution is 10 minutes. And selecting a set of 70% of data before each month as a training sample, and a set of 30% of data after each month as a test sample, and modeling the wind power aggregation characteristic with the time scale of 4 hours.
As shown in FIG. 1, the modeling method for the aggregation characteristics of the wind power plant comprises the following steps:
the method comprises the following steps: collecting actual measurement operation data of a plurality of wind turbine generators in a wind power plant, and cleaning and preprocessing the data; the actual measurement operation data comprises actual measurement wind speed data, actual measurement wind direction data and actual measurement power data, and the preprocessing comprises data cleaning and normalization processing;
step two: and establishing a wind power fluctuation measurement index. The aggregated output of the wind turbine generator cluster is the sum of the outputs of the single machines, and is shown in a formula (1):
Figure BDA0002259564710000061
in the formula, P∑(t) is the total power of the wind turbine generator cluster at the moment t; i is the number of the wind turbine generator; n is the number of the wind turbine generators; pi(t) is the power of the No. i wind turbine generator at the moment t;
the standard deviation of the power sequence is expressed by formulas (2) and (3) to measure the fluctuation of the output of a single wind turbine generator and the aggregate output of the wind turbine generators:
Figure BDA0002259564710000062
in the formula, σi、σ∑Respectively is the power standard deviation and the aggregate power standard deviation of the No. i wind turbine generator; t is a statistical time scale;
Figure BDA0002259564710000064
respectively taking the mean values of the sampling values of the power and the aggregate power of the No. i wind turbine generator under the corresponding time statistical scale; the standard deviation of the power sequence reflects the degree of power fluctuation of the time sequence, and the smaller the numerical value is, the smaller the power fluctuation of the time sequence is, and the better the stationarity is.
Step three: establishing a wind power plant polymerization output smooth effect measurement index, and defining a smooth effect coefficient as a ratio of a wind turbine generator polymerization output standard deviation per unit value to a single unit output standard deviation per unit value by taking installed capacity as a base value, namely a formula (4):
Figure BDA0002259564710000065
wherein, PRFor the rated power, sigma, of the wind turbinei、σ∑The power standard deviation and the polymerization power standard deviation of the No. i wind turbine generator are respectively shown, N represents the number of the polymerization generator sets, and the smaller the coefficient s is, the stronger the complementarity among the output forces of the wind turbine generators is, and the more obvious the smoothing effect is;
and (3) obtaining a relational formula (5) of the standard deviation of the aggregate power and the standard deviation of the single machine power of the wind turbine generator through the three mathematical statistics:
Figure BDA0002259564710000071
in the formula, ri,jThe correlation coefficient of the power sequences of the No. i wind turbine generator and the No. j wind turbine generator is obtained;
if the output standard deviations of the units are the same and are regarded as the average value of the output standard deviations of the single unit, the correlation coefficient relation of the smooth effect measurement index, the number N of the polymerization units and the power sequence among the units is obtained by the formulas (4) and (5) and is a formula (6):
Figure BDA0002259564710000072
step four: establishing a mapping model of multi-position point wind condition information and inter-unit power sequence correlation based on a convolutional neural network, wherein the convolutional neural network comprises two convolutional layers, two pooling layers, a flattening layer and a full connection layer, and the structure of the convolutional neural network is shown in the attached figure 2. The method comprises the steps that actually measured wind speed sequence data and actually measured wind direction sequence data of 5 units pass through a first convolution layer and a ReLU activation layer to generate a group of feature maps, are subjected to non-overlapping maximum pooling for down-sampling, then pass through a second convolution layer and the ReLU activation layer to generate another group of feature maps, the group of feature maps are connected with a flattening layer to enable a multi-dimensional tensor to be one-dimensional, then are connected with a full connection layer, and finally are activated through a sigmoid function to serve as output of a convolutional neural network, aiming at the problems of uneven distribution and gradient dispersion of each layer, a Batch Normalization method (Batch Normalization-BN) is adopted to reduce the influence, namely a Normalization layer is added after partial layer output, and output is normalized to the same distribution according to feature values of the same Batch;
the convolution operation process is expressed by equation (7):
Figure BDA0002259564710000073
wherein the content of the first and second substances,
Figure BDA0002259564710000074
the j-th characteristic diagram in the L-th convolutional layer, namely the wind speed distribution characteristic extracted by the convolutional layer,
Figure BDA0002259564710000075
is the feature map set of the previous layer, which is also the input of this layer,for the jth convolution kernel in the lth convolution layer, a convolution operation is defined, b represents an additive bias value, and f (x) represents an activation function which maps a value in a real number domain into a finite field;
the operation process of the pooling layer is expressed by formula (8):
Figure BDA0002259564710000081
wherein the content of the first and second substances,
Figure BDA0002259564710000082
denotes the jth feature map in the lth pooling layer, f (x) denotes the activation function,
Figure BDA0002259564710000083
a multiplicative bias representing a profile, down (x) is a down-sampling function,
Figure BDA0002259564710000084
showing the additive bias of the signature.
The structural parameters of the convolutional neural network-based output correlation mapping model are shown in table 1:
TABLE 1 convolution neural network-based output correlation mapping model network structure parameters
Figure BDA0002259564710000085
Step five: the method comprises the steps of inputting actual measurement wind speed sequence data and actual measurement wind direction sequence data of multiple unit point positions in a specified time scale (such as 4 hours) as a model, outputting an output correlation coefficient between every two units to form a model training sample, training a neural network model by using a root mean square error function index, and outputting a mapping result. The calculation formula of the output correlation coefficient between every two units is a formula (9):
Figure BDA0002259564710000091
wherein sigmaxyRefers to the covariance, σ, of the two data sequences X and Yx、σyThe respective variances of X and Y,
Figure BDA0002259564710000092
respectively, mean values, X, of the sequence X, Yi、YiRespectively, the ith number of the sequence X, Y.
After training and prediction, the error of the model mapping result and the real unit output correlation coefficient is shown in table 2.
TABLE 2 error of correlation coefficient between model mapping result and real unit output
Figure BDA0002259564710000093
Step six: and calculating a smoothing effect measurement index s representing the aggregation characteristic of the wind power plant according to the output correlation mapping result of the set obtained in the step five and the formula (6), and realizing wind power plant aggregation characteristic modeling based on the convolutional neural network, wherein the average absolute error and the root mean square error of a smoothing effect system s and a real smoothing effect coefficient calculated according to the mapping result are 0.1386 and 0.1781 respectively.
Finally, it should be pointed out that: the above examples are only for illustrating the technical solutions of the present invention, and are not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims (7)

1. A wind power plant aggregation characteristic modeling method is characterized by comprising the following steps:
the method comprises the following steps: collecting actual measurement operation data of a plurality of wind turbine generators of a wind power plant, and cleaning and normalizing the data;
step two: establishing a wind power fluctuation measurement index;
step three: establishing a smoothing effect measurement index of the aggregate output of the wind power plant, and obtaining a relational expression of the smoothing effect measurement index, the number N of the polymer units and the correlation coefficient of the power sequence among the units by a mathematical statistic analysis method;
step four: establishing a mapping model of multi-position point wind condition information and inter-unit power sequence correlation based on a convolutional neural network, and modeling the number of units according to the time number in a specified time period and the aggregation characteristic of the wind power plant to set corresponding parameters of the convolutional neural network;
step five: taking actual measurement wind speed sequence data and actual measurement wind direction sequence data of multiple unit point positions with a specified time scale as model input, taking output correlation coefficients between every two units as output to form a model training sample, training a neural network model by using root mean square error function indexes, and outputting unit output correlation mapping results;
step six: and calculating a smooth effect measurement index s representing the aggregation characteristic of the wind power plant according to the unit output correlation mapping result in the step five and a functional relation between the smooth effect measurement index in the step three and the number N of the aggregation units and the correlation coefficient of the power sequences among the units, and realizing the modeling of the aggregation characteristic of the wind power plant based on the convolutional neural network.
2. The wind farm aggregate characteristic modeling method of claim 1, characterized in that: the step of measuring the operation data includes: measured wind speed data, measured wind direction data and measured power data.
3. The wind farm aggregate characteristic modeling method of claim 1, characterized in that: the aggregated output of the wind turbine generator cluster is the sum of the outputs of the single machines, and is shown in a formula (1):
in the formula, P∑(t) is the total power of the wind turbine generator cluster at the moment t; i is the number of the wind turbine generator; n is the number of the wind turbine generators; pi(t) is the power of the No. i wind turbine generator at the moment t;
in the second step, the standard deviation of the power sequence is expressed by formulas (2) and (3) to measure the fluctuation of the output of a single wind turbine generator and the aggregate output of the wind turbine generators:
Figure FDA0002259564700000021
Figure FDA0002259564700000022
in the formula, σi、σ∑Respectively is the power standard deviation and the aggregate power standard deviation of the No. i wind turbine generator; t is a statistical time scale;
Figure FDA0002259564700000023
respectively taking the mean values of the sampling values of the power and the aggregate power of the No. i wind turbine generator under the corresponding time statistical scale; power sequence standard deviation reflectionThe degree of time series power fluctuation is reduced, and the smaller the numerical value is, the smaller the time series power fluctuation is, and the better the stationarity is.
4. The wind farm aggregate characteristic modeling method of claim 1, characterized in that: the smooth effect coefficient is the ratio of the wind turbine generator output standard deviation per unit value taking installed capacity as a base value to the wind turbine generator output standard deviation per unit value as a formula (4):
Figure FDA0002259564700000024
wherein, PRFor the rated power, sigma, of the wind turbinei、σ∑The power standard deviation and the polymerization power standard deviation of the No. i wind turbine generator are respectively shown, N represents the number of the polymerization generator sets, and the smaller the coefficient s is, the stronger the complementarity among the output forces of the wind turbine generators is, and the more obvious the smoothing effect is;
and (3) obtaining a relational formula (5) of the standard deviation of the aggregate power and the standard deviation of the single machine power of the wind turbine generator through the three mathematical statistics:
Figure FDA0002259564700000025
in the formula, ri,jThe correlation coefficient of the power sequences of the No. i wind turbine generator and the No. j wind turbine generator is obtained;
if the output standard deviations of the units are the same and are regarded as the average value of the output standard deviations of the single unit, the correlation coefficient relation of the smooth effect measurement index, the number N of the polymerization units and the power sequence among the units is obtained by the formulas (4) and (5) and is a formula (6):
Figure FDA0002259564700000031
5. the wind farm aggregate characteristic modeling method of claim 1, characterized in that: the convolutional neural network of the fourth step comprises two convolutional layers, two pooling layers, a flattening layer and a full-connection layer; the method comprises the steps that actually measured wind speed sequence data and actually measured wind direction sequence data of multiple sets pass through a first convolution layer and a ReLU activation layer to generate a set of characteristic graphs, then non-overlapping maximum pooling is conducted to carry out down-sampling, then the second convolution layer and the ReLU activation layer are used for generating another set of characteristic graphs, the set of characteristic graphs are connected with a flattening layer to enable a multi-dimensional tensor to be one-dimensional, then the characteristic graphs are connected with a full connection layer, and finally the characteristic graphs are activated through a sigmoid function to serve as output of a convolutional neural network.
6. The wind farm aggregate characteristic modeling method of claim 5, characterized in that: step four, a batch normalization method is adopted to add a normalization layer after partial layer output, output is normalized to the same distribution according to the characteristic values of the same batch, and the convolution operation process is represented by a formula (7):
Figure FDA0002259564700000032
wherein the content of the first and second substances,
Figure FDA0002259564700000033
the j-th characteristic diagram in the L-th convolutional layer, namely the wind speed distribution characteristic extracted by the convolutional layer,
Figure FDA0002259564700000034
is the feature map set of the previous layer, which is also the input of this layer,
Figure FDA0002259564700000035
for the jth convolution kernel in the lth convolution layer, a convolution operation is defined, b represents an additive bias value, and f (x) represents an activation function which maps a value in a real number domain into a finite field;
the operation process of the pooling layer is expressed by formula (8):
Figure FDA0002259564700000036
wherein the content of the first and second substances,
Figure FDA0002259564700000037
denotes the jth feature map in the lth pooling layer, f (x) denotes the activation function,
Figure FDA0002259564700000038
a multiplicative bias representing a profile, down (x) is a down-sampling function,showing the additive bias of the signature.
7. The wind farm aggregate characteristic modeling method of claim 1, characterized in that: in the fifth step, the calculation formula of the output correlation coefficient between every two units is a formula (9):
wherein sigmaxyRefers to the covariance, σ, of the two data sequences X and Yx、σyThe respective variances of X and Y,
Figure FDA0002259564700000042
respectively, mean values, X, of the sequence X, Yi、YiRespectively, the ith number of the sequence X, Y.
CN201911066604.5A 2019-11-04 2019-11-04 Wind power plant aggregation characteristic modeling method Active CN110765703B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911066604.5A CN110765703B (en) 2019-11-04 2019-11-04 Wind power plant aggregation characteristic modeling method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911066604.5A CN110765703B (en) 2019-11-04 2019-11-04 Wind power plant aggregation characteristic modeling method

Publications (2)

Publication Number Publication Date
CN110765703A true CN110765703A (en) 2020-02-07
CN110765703B CN110765703B (en) 2022-05-27

Family

ID=69336090

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911066604.5A Active CN110765703B (en) 2019-11-04 2019-11-04 Wind power plant aggregation characteristic modeling method

Country Status (1)

Country Link
CN (1) CN110765703B (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112149903A (en) * 2020-09-22 2020-12-29 内蒙古电力(集团)有限责任公司内蒙古电力科学研究院分公司 Primary frequency modulation analysis and optimization method of thermal power generating unit based on BP neural network algorithm
CN112395812A (en) * 2020-11-26 2021-02-23 华北电力大学 Method for evaluating wind speed time shifting performance
CN112507604A (en) * 2020-11-04 2021-03-16 清华大学 Data driving voltage frequency response characteristic modeling method of renewable power supply
CN114583767A (en) * 2022-03-10 2022-06-03 中国电力科学研究院有限公司 Data-driven wind power plant frequency modulation response characteristic modeling method and system

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102570449A (en) * 2011-12-21 2012-07-11 西安交通大学 Regional wind power prediction method based on spatial and temporal distribution characteristics
CN103268366A (en) * 2013-03-06 2013-08-28 辽宁省电力有限公司电力科学研究院 Combined wind power prediction method suitable for distributed wind power plant
CN103401236A (en) * 2013-07-30 2013-11-20 华北电力大学 Wind power farm generator unit grouping method based on flow correlation of wind power farm
CN104978608A (en) * 2015-06-24 2015-10-14 云南电网有限责任公司电力科学研究院 Wind power prediction apparatus and prediction method
US20170300788A1 (en) * 2014-01-30 2017-10-19 Hrl Laboratories, Llc Method for object detection in digital image and video using spiking neural networks
US20180188714A1 (en) * 2016-05-09 2018-07-05 Strong Force Iot Portfolio 2016, Llc Methods and systems for the industrial internet of things
CN109086527A (en) * 2018-08-03 2018-12-25 合肥工业大学 A kind of practical equivalent modeling method based on running of wind generating set state
CN109670213A (en) * 2018-11-29 2019-04-23 国电南瑞科技股份有限公司 A kind of wind power plant multimachine equivalent modeling method and device
CN109802430A (en) * 2018-12-29 2019-05-24 上海电力学院 A kind of wind-powered electricity generation power grid control method based on LSTM-Attention network
CN110061524A (en) * 2019-05-06 2019-07-26 中国科学院电工研究所 A kind of distributed generation resource virtual plant active power dispatch equivalence polymerization and system based on deep neural network

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102570449A (en) * 2011-12-21 2012-07-11 西安交通大学 Regional wind power prediction method based on spatial and temporal distribution characteristics
CN103268366A (en) * 2013-03-06 2013-08-28 辽宁省电力有限公司电力科学研究院 Combined wind power prediction method suitable for distributed wind power plant
CN103401236A (en) * 2013-07-30 2013-11-20 华北电力大学 Wind power farm generator unit grouping method based on flow correlation of wind power farm
US20170300788A1 (en) * 2014-01-30 2017-10-19 Hrl Laboratories, Llc Method for object detection in digital image and video using spiking neural networks
CN104978608A (en) * 2015-06-24 2015-10-14 云南电网有限责任公司电力科学研究院 Wind power prediction apparatus and prediction method
US20180188714A1 (en) * 2016-05-09 2018-07-05 Strong Force Iot Portfolio 2016, Llc Methods and systems for the industrial internet of things
CN109086527A (en) * 2018-08-03 2018-12-25 合肥工业大学 A kind of practical equivalent modeling method based on running of wind generating set state
CN109670213A (en) * 2018-11-29 2019-04-23 国电南瑞科技股份有限公司 A kind of wind power plant multimachine equivalent modeling method and device
CN109802430A (en) * 2018-12-29 2019-05-24 上海电力学院 A kind of wind-powered electricity generation power grid control method based on LSTM-Attention network
CN110061524A (en) * 2019-05-06 2019-07-26 中国科学院电工研究所 A kind of distributed generation resource virtual plant active power dispatch equivalence polymerization and system based on deep neural network

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
CHRISTOPHE GUILLE等: "The integration of PHEV aggregations into a power system with wind resources", 《IEEE》 *
苏适等: "光伏发电出力特征提取及区域集群聚合特性", 《云南电力技术》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112149903A (en) * 2020-09-22 2020-12-29 内蒙古电力(集团)有限责任公司内蒙古电力科学研究院分公司 Primary frequency modulation analysis and optimization method of thermal power generating unit based on BP neural network algorithm
CN112507604A (en) * 2020-11-04 2021-03-16 清华大学 Data driving voltage frequency response characteristic modeling method of renewable power supply
CN112395812A (en) * 2020-11-26 2021-02-23 华北电力大学 Method for evaluating wind speed time shifting performance
CN112395812B (en) * 2020-11-26 2024-03-26 华北电力大学 Method for evaluating time shifting property of wind speed
CN114583767A (en) * 2022-03-10 2022-06-03 中国电力科学研究院有限公司 Data-driven wind power plant frequency modulation response characteristic modeling method and system
CN114583767B (en) * 2022-03-10 2023-03-17 中国电力科学研究院有限公司 Data-driven wind power plant frequency modulation response characteristic modeling method and system

Also Published As

Publication number Publication date
CN110765703B (en) 2022-05-27

Similar Documents

Publication Publication Date Title
CN110765703B (en) Wind power plant aggregation characteristic modeling method
CN110991786B (en) 10kV static load model parameter identification method based on similar daily load curve
CN102663513B (en) Utilize the wind power combined prediction modeling method of grey relational grade analysis
CN111722046B (en) Transformer fault diagnosis method based on deep forest model
CN113126019B (en) Remote estimation method, system, terminal and storage medium for error of intelligent ammeter
CN111475909B (en) Wind turbine generator output correlation mapping modeling method based on long-term and short-term memory network
CN110222887A (en) Prediction technique based on VMD and DNN and the application in short-term load forecasting
CN104318482A (en) Comprehensive assessment system and method of smart distribution network
CN113205207A (en) XGboost algorithm-based short-term power consumption load fluctuation prediction method and system
CN108020752A (en) A kind of distribution line loss diagnostic method and system based on multi-source perforation correlation
CN103971175B (en) Short-term load prediction method of multistage substations
CN106980044B (en) A kind of Harmonious Waves in Power Systems current estimation method adapting to wind power integration
CN113141008A (en) Data-driven power distribution network distributed new energy consumption capacity assessment method
CN110458189A (en) Compressed sensing and depth convolutional neural networks Power Quality Disturbance Classification Method
CN113837432A (en) Power system frequency prediction method driven by physics-data combination
CN112651576A (en) Long-term wind power prediction method and device
CN106372440A (en) Method and device for estimating self-adaptive robust state of distribution network through parallel computation
CN109193791B (en) Wind power convergence tendency state-based quantification method based on improved shape value
CN116596129A (en) Electric vehicle charging station short-term load prediction model construction method
CN113705885B (en) VMD, XGBoost and TCN fused power distribution network voltage prediction method and system
CN104573313B (en) The acquisition methods and system of customer response model under Peak-valley TOU power price
Chen et al. Variation-cognizant probabilistic power flow analysis via multi-task learning
CN112230056A (en) Multi-harmonic source contribution calculation method based on OFMMK-Means clustering and composite quantile regression
CN110378541A (en) A kind of short-term multistep forecasting method of wind power and device
CN113722970B (en) Photovoltaic power ultra-short-term online prediction method

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