CN113033904B - Wind power prediction error analysis and classification method based on S transformation - Google Patents

Wind power prediction error analysis and classification method based on S transformation Download PDF

Info

Publication number
CN113033904B
CN113033904B CN202110360165.XA CN202110360165A CN113033904B CN 113033904 B CN113033904 B CN 113033904B CN 202110360165 A CN202110360165 A CN 202110360165A CN 113033904 B CN113033904 B CN 113033904B
Authority
CN
China
Prior art keywords
prediction error
wind power
formula
error
err
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.)
Active
Application number
CN202110360165.XA
Other languages
Chinese (zh)
Other versions
CN113033904A (en
Inventor
齐先军
张付华
张晶晶
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.)
Hefei University of Technology
Original Assignee
Hefei University of Technology
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 Hefei University of Technology filed Critical Hefei University of Technology
Priority to CN202110360165.XA priority Critical patent/CN113033904B/en
Publication of CN113033904A publication Critical patent/CN113033904A/en
Application granted granted Critical
Publication of CN113033904B publication Critical patent/CN113033904B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • 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—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • 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)
  • Economics (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Primary Health Care (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention discloses a wind power prediction error analysis and classification method based on S transformation, which comprises the following steps: 1. calculating a wind power prediction error sequence; 2. calculating a time frequency spectrum of the wind power prediction error by using S transformation; 3. establishing and calculating a prediction error analysis index based on a time-frequency domain; 4. and classifying the wind power prediction error. The method can show the distribution condition of the wind power prediction error in the time-frequency domain, extracts the characteristics of the wind power prediction error from the time-frequency angle, further provides the indexes of the transverse prediction error and the longitudinal prediction error of the wind power, and classifies the wind power prediction error by utilizing the provided time-frequency indexes. The method can provide important reference for analyzing and evaluating the prediction error of the wind power and evaluating and comparing the prediction effects of different prediction models.

Description

Wind power prediction error analysis and classification method based on S transformation
Technical Field
The invention relates to the technical field of prediction error analysis and classification, in particular to a time-frequency analysis and classification method for wind power prediction errors.
Background
An accurate wind power prediction technology is a key for reducing adverse effects of wind power fluctuation on power grid dispatching and guaranteeing safe and reliable operation of a power system. At present, various wind power prediction models such as a time series prediction model, a neural network prediction model, a support vector machine prediction model and the like exist, the models predict future wind power by using historical data from different angles, and prediction errors are different. Analyzing and classifying wind power prediction errors is an important content of wind power prediction technology research. The prediction errors are analyzed and classified, so that the operation condition of the prediction model can be known, the change rule of the prediction errors is explored, and the distribution characteristics of the prediction errors of different prediction models are identified. Different prediction methods and prediction models are compared, analyzed and classified, and the improvement of the prediction models and the error estimation models is facilitated, so that the prediction precision is improved, and the prediction results are better utilized to serve the actual production.
The uncertainty of the wind speed causes the wind power output by the wind turbine generator to show strong randomness, fluctuation and intermittence. Wind power is a non-stationary signal containing various frequency components, and a prediction error signal generated along with the non-stationary signal is often accompanied with certain non-stationarity. However, the analysis of the prediction error is limited to the statistical indexes in the time domain, such as the average error, the average relative error, the average absolute error, the root mean square error, etc., which are statistical analysis of the prediction error in a period of time, and the non-stationary signal with time-varying characteristics cannot be well analyzed, and the index calculation result can only reflect the whole error level of the prediction error in the time domain, but cannot reflect the time-varying spectrum distribution characteristics of the non-stationary prediction error signal, and cannot deeply know the time-varying rules of the prediction error in the amplitude error and the phase error, so that the wind power prediction error cannot be reasonably classified, the data characteristics of the prediction error cannot be accurately mastered, a targeted error evaluation model cannot be established, and the prediction accuracy cannot be improved.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides the wind power prediction error analysis and classification method based on the S transformation, so that the wind power prediction error can be more effectively analyzed and classified, a more accurate prediction error model can be established, and the accuracy of the wind power prediction model can be improved.
In order to achieve the purpose, the invention adopts the technical scheme that:
the invention relates to a wind power prediction error analysis and classification method based on S transformation, which is characterized by comprising the following steps of:
step 1: collecting actually-measured wind power sequence { p) of certain wind power plant act (j) 0,1,2, …, N-1 }; according to the historical wind power sequence of the wind power plant, a prediction wind power sequence { p is obtained by using a prediction model pre (j) 0,1,2, …, N-1 }; wherein j represents the jth sample point; n represents the total number of sampling points and is an even number; p is a radical of act (j) Represents the measured wind power, p, of the jth sampling point pre (j) Representing the predicted wind power of the jth sampling point;
calculating wind power prediction error sequence p of j sampling point by using formula (1) err (j)；
p err (j)＝p act (j)-p pre (j)， j＝0,1,2,…,N-1 (1)
Step 2, predicting an error sequence { p) of the wind power by using the formula (2) err (j) Making discrete S transformation of 0,1,2, …, N-1 to obtain time frequency spectrum S of prediction error err [j,n]；
Figure BDA0003005227220000021
In formula (2): m is the frequency translation amount, and m is 0,1,2, …, N-1; i is an imaginary unit; exp (·) represents an exponential function with a natural constant e as base; p err [·]Predicting an error sequence { p } for wind power err (j) A discrete fourier transform of 0,1,2, …, N-1; j is the jth sampling point, j is 0,1,2, …, N-1; n is the nth frequency point, N is 0,1,2, …, N/2+ 1;
and step 3: calculating a longitudinal prediction error sequence A of the jth sampling point by using the formula (3) err (j)：
Figure BDA0003005227220000022
In formula (3): | · | represents a modulo operation;
and 4, step 4: calculating mean index m of longitudinal prediction error sequence by using formula (4) A ：
Figure BDA0003005227220000023
And 5: calculating a fluctuation index sigma of a longitudinal prediction error sequence by using the formula (5) A ：
Figure BDA0003005227220000024
Step 6: calculating the transverse prediction error sequence gamma of the j sampling point by using the formula (6) err (j)：
Figure BDA0003005227220000031
In formula (6): im (-) represents an imaginary part, Re (-) represents an actual part, and arctan (-) is an arctangent function;
and 7: calculating the mean index m of the transverse prediction error sequence by using the formula (7) Γ ：
Figure BDA0003005227220000032
And 8: calculating fluctuation index sigma of transverse prediction error sequence by using equation (8) Γ ：
Figure BDA0003005227220000033
And step 9: the prediction error pair e at the j time is established by using the formula (9) j ；
e j ＝(A err (j),Γ err (j)),j＝0,1,2,…,N-1 (9)
Step 10: the prediction error at the j time is paired with e j The classification is K:
step 10-1: a set of prediction error pairs E is established using equation (10):
E＝{e j |j＝0,1,2,…,N-1} (10)
step 10-2: randomly extracting K different numbers in the set {0,1,2, …, N-1}, and recording the numbers as i 1 ,i 2 ,…,i k ,…,i K (ii) a Wherein i k Representing the k-th number, and predicting the i-th number in the error pair set E k An error pair
Figure BDA0003005227220000034
Put into the kth cluster C k In the formula (11), K is more than or equal to 1 and less than or equal to K, and the kth cluster C is initialized according to the formula (11) k Cluster-like center of
Figure BDA0003005227220000035
Figure BDA0003005227220000036
In formula (11):
Figure BDA0003005227220000037
respectively represent the kth class cluster C k Cluster-like center of
Figure BDA0003005227220000038
Longitudinal error and lateral error of (d); a. the err (i k ),Γ err (i k ) Respectively representing the ith in the set E of prediction error pairs k An error pair
Figure BDA0003005227220000039
Longitudinal error and lateral error of (d); and is provided with
Figure BDA00030052272200000310
1≤k≤K；
Step 10-3: the prediction error pair e at the j-th time is calculated by equation (12) j With the kth cluster center
Figure BDA00030052272200000311
Of Euclidean distance d jk ：
Figure BDA00030052272200000312
Step 10-4: the nearest prediction error pairs are classified into corresponding class clusters using equation (13):
Figure BDA0003005227220000041
in formula (13): u denotes union operation of sets, λ j Representing the prediction error pair e with the j-th time j Cluster number closest to the cluster, and λ j ＝argmin k＝1,2,…,K (d jk ) And argmin represents the Euclidean distance d jk The minimum value of k;
step 10-5: updating the cluster center using equation (14);
Figure BDA0003005227220000042
in formula (14):
Figure BDA0003005227220000043
representing updated class C k Center of (e), e δ Represents the kth class cluster C k Of (1), M represents the kth class cluster C k The number of included prediction error pairs;
step 10-6: repeating the steps 10-3 to 10-5 until the kth cluster class C k K is more than or equal to 1 and less than or equal to K until the center of the key is not changed;
step 10-7: output the kth class C k ，1≤k≤K。
Compared with the prior art, the invention has the beneficial effects that:
the method solves the problems of prediction error classification and time-varying rule analysis of the time domain index. The method comprises the steps of mapping a one-dimensional time domain signal to a two-dimensional time-frequency domain by performing discrete S transformation on a prediction error of wind power, comprehensively considering amplitude and phase information of the wind power prediction error in the time-frequency domain, and establishing and calculating a longitudinal error index and a transverse error index based on the time-frequency domain, wherein the longitudinal index mainly reflects amplitude deviation of the prediction error, the transverse index mainly reflects phase deviation of the prediction error, and classifying the wind power prediction error according to the longitudinal and transverse indexes. The concrete effects are shown in the following aspects:
1. the invention adopts the step 2 to calculate the time frequency spectrum S of the wind power prediction error err [j,n]The wind power time sequence signal is mapped to the time-frequency domain matrix, so that the distribution condition of the wind power prediction error on the time-frequency domain can be accurately described, and comprehensive and specific information is provided for error analysis and classification;
2. the invention adopts A shown in step 3 err (j) Defining a longitudinal error of the wind power prediction model at the j moment so as to obtain a longitudinal error sequence, comprehensively considering amplitude deviations of the wind power prediction error at different frequencies, and reflecting the change rule of the longitudinal prediction error along with time; using σ as shown in step 5 A Defining a fluctuation evaluation index of the longitudinal prediction error, wherein the index can reflect the fluctuation condition of the longitudinal prediction error;
3. the invention adopts the Gamma shown in step 6 err (j) Defining a lateral error of the wind power prediction model at the j moment so as to obtain a lateral error sequence, comprehensively considering phase deviations of the wind power prediction error at different frequencies, and reflecting a rule that the lateral prediction error changes along with time; using σ as shown in step 8 Γ Defining a fluctuation evaluation index of the transverse prediction error, wherein the index can reflect the fluctuation condition of the transverse prediction error;
4. the prediction error classification method shown in step 10 is adopted, so that the prediction errors can be reasonably classified according to the amplitude and phase attributes of the prediction errors, and the class of the prediction errors at each moment is determined.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Detailed Description
In this embodiment, as shown in fig. 1, an evaluation method of a wind power prediction error is performed according to the following steps:
step 1: collecting actually-measured wind power sequence { p) of certain wind power plant act (j) 0,1,2, …, N-1 }; according to the historical wind power sequence of the wind power plant, a prediction wind power sequence { p is obtained by using a prediction model pre (j) 0,1,2, …, N-1 }; wherein j represents the jth sample point; n represents the total number of sampling points and is an even number; p is a radical of act (j) Represents the measured wind power, p, of the jth sampling point pre (j) Representing the predicted wind power of the jth sampling point;
calculating wind power prediction error sequence p of j sampling point by using formula (1) err (j)；
p err (j)＝p act (j)-p pre (j)， j＝0,1,2,…,N-1 (1)
Step 2, predicting an error sequence { p) of the wind power by using the formula (2) err (j) Making discrete S transformation of 0,1,2, …, N-1 to obtain time frequency spectrum S of prediction error err [j,n]；
Figure BDA0003005227220000051
In formula (2): m is the frequency translation amount, and m is 0,1,2, …, N-1; i is an imaginary unit; exp (·) represents an exponential function with a natural constant e as base; p err [·]Predicting an error sequence { p ] for wind power err (j) A discrete fourier transform of 0,1,2, …, N-1 }; j is the jth sampling point, j is 0,1,2, …, N-1; n is the nth frequency point, N is 0,1,2, …, N/2+ 1;
by predicting the wind power errorTime frequency spectrum S of prediction error obtained by line dispersion S transformation calculation err [j,n]The distribution situation of the prediction errors on the time-frequency domain can be obtained, the characteristic indexes of the prediction errors are favorably defined from the angle of the time-frequency domain, and the longitudinal and transverse error characteristics of the prediction model are extracted, so that the distribution characteristics of the prediction errors can be more deeply mastered, a more accurate error correction model can be favorably established, and the accuracy of the wind power prediction model is improved;
and step 3: calculating a longitudinal prediction error sequence A of the jth sampling point by using the formula (3) err (j)：
Figure BDA0003005227220000061
In formula (3): | · | represents a modulo operation;
in order to master the change rule of longitudinal prediction error of the wind power prediction model along with time, A is adopted err (j) A longitudinal error sequence of the wind power prediction model is defined, the sequence comprehensively considers amplitude information of the prediction error under different frequencies at the jth moment, can reflect the change rule of the longitudinal error of the prediction model along with time, and is favorable for establishing a longitudinal prediction error correction model of the wind power so as to improve the prediction precision;
and 4, step 4: calculating mean index m of longitudinal prediction error sequence by using formula (4) A ：
Figure BDA0003005227220000062
And 5: calculating a fluctuation index sigma of a longitudinal prediction error sequence by using the formula (5) A ：
Figure BDA0003005227220000063
To grasp the fluctuation of the longitudinal prediction error sequence, sigma is adopted A Defining a fluctuation index of a longitudinal error sequence of the wind power prediction model, wherein the fluctuation index can reflect longitudinal prediction errors(ii) a fluctuating situation;
step 6: calculating the transverse prediction error sequence gamma of the j sampling point by using the formula (6) err (j)：
Figure BDA0003005227220000064
In formula (6): im (-) represents an imaginary part, Re (-) represents an actual part, and arctan (-) is an arctangent function;
in order to master the time-varying rule of the transverse prediction error of the wind power prediction model, gamma is adopted err (j) Defining a transverse error sequence of the wind power prediction model, wherein the transverse error sequence can reflect the phase deviation of the predicted wind power sequence compared with the actually-measured wind power sequence at each moment, and is favorable for establishing a transverse prediction error correction model of the wind power and improving the prediction precision;
and 7: calculating the mean index m of the transverse prediction error sequence by using the formula (7) Γ ：
Figure BDA0003005227220000071
And 8: calculating a fluctuation index sigma of a transverse prediction error sequence by using equation (8) Γ ：
Figure BDA0003005227220000072
In order to master the transverse prediction error fluctuation condition of the wind power prediction model, sigma is adopted Γ Defining a fluctuation index of a transverse error of the wind power prediction model, wherein the fluctuation index can reflect the fluctuation condition of the transverse error of the prediction model and provide reference for power grid scheduling personnel to arrange a real-time scheduling plan;
and step 9: based on the above mentioned vertical prediction error sequence { A } err (j) I j 0,1,2, …, N-1 and lateral prediction error sequence { Γ | err (j) I j is 0,1,2, …, N-1, and the prediction error pair x at the j-th time is established by equation (9) j ；
x j ＝(A err (j),Γ err (j)),j＝0,1,2,…,N-1 (9)
Step 10: the prediction error at the j time is paired with e j Classified into K types:
step 10-1: a set of prediction error pairs E is established using equation (10):
E＝{e j |j＝0,1,2,…,N-1} (10)
step 10-2: randomly extracting K different numbers in the set {0,1,2, …, N-1}, and recording the numbers as i 1 ,i 2 ,…,i k ,…,i K (ii) a Wherein i k Representing the k-th number, and predicting the i-th number in the error pair set E k An error pair
Figure BDA0003005227220000073
Put into the kth cluster C k In the formula (11), K is more than or equal to 1 and less than or equal to K, and the kth cluster center is initialized according to the formula (11)
Figure BDA0003005227220000074
Figure BDA0003005227220000075
In formula (11):
Figure BDA0003005227220000076
respectively represent the kth cluster center
Figure BDA0003005227220000077
Longitudinal error and lateral error of (d); a. the err (i k ),Γ err (i k ) Respectively representing the ith in the set E of prediction error pairs k An error pair
Figure BDA0003005227220000078
Longitudinal error and lateral error of (d); and is
Figure BDA0003005227220000079
1≤k≤K；
Step 10-3: the prediction error pair e at the j-th time is calculated by equation (12) j With the kth cluster center
Figure BDA0003005227220000081
Of Euclidean distance d jk ：
Figure BDA0003005227220000082
Step 10-4: the nearest prediction error pair is divided into corresponding class clusters using equation (13):
Figure BDA0003005227220000083
in formula (13): u denotes union operation of sets, λ j Represents a prediction error pair e with the j-th time j Cluster number of closest class, and λ j ＝argmin k＝1,2,…,K (d jk ) And argmin represents the Euclidean distance d jk The minimum value of k;
step 10-5: updating the cluster center using equation (14);
Figure BDA0003005227220000084
in formula (14):
Figure BDA0003005227220000085
representing updated class C k Center of (e), e δ Represents the kth class cluster C k Of (1), M represents the kth class cluster C k The number of included prediction error pairs;
step 10-6: repeating the steps 10-3 to 10-5 until the kth cluster class C k K is more than or equal to 1 and less than or equal to K;
step 10-7: output the kth class C k ，1≤k≤K。
In order to master the distribution characteristics of the wind power prediction errors, the prediction errors are classified according to the provided longitudinal and transverse time-frequency analysis indexes, the classification result can comprehensively consider the amplitude and phase information of the prediction errors in a time-frequency domain, the prediction errors are classified more reasonably, a more accurate prediction error evaluation model is favorably established, and the reliability of wind power prediction is improved.

Claims (1)

1. A wind power prediction error analysis and classification method based on S transformation is characterized by comprising the following steps:
step 1: collecting actually-measured wind power sequence { p) of certain wind power plant act (j) 0,1,2, …, N-1 }; according to the historical wind power sequence of the wind power plant, a prediction wind power sequence { p is obtained by using a prediction model pre (j) 0,1,2, …, N-1 }; wherein j represents the jth sample point; n represents the total number of sampling points and is an even number; p is a radical of act (j) Represents the measured wind power p of the jth sampling point pre (j) Representing the predicted wind power of the jth sampling point;
calculating wind power prediction error sequence p of j sampling point by using formula (1) err (j)；
p err (j)＝p act (j)-p pre (j)，j＝0,1,2,…,N-1 (1)
Step 2, predicting an error sequence { p) of the wind power by using the formula (2) err (j) Making discrete S transformation of 0,1,2, …, N-1 | j ═ to obtain time frequency spectrum S of prediction error err [j,n]；
Figure FDA0003005227210000011
In formula (2): m is the frequency translation amount, and m is 0,1,2, …, N-1; i is an imaginary unit; exp (·) represents an exponential function with a natural constant e as base; p err [·]Predicting an error sequence { p ] for wind power err (j) A discrete fourier transform of 0,1,2, …, N-1; j is the jth sampling point, j is 0,1,2, …, N-1; n is the nth frequency point, n is 0,1,2,…,N/2+1；
and step 3: calculating a longitudinal prediction error sequence A of the jth sampling point by using the formula (3) err (j)：
Figure FDA0003005227210000012
In formula (3): | cndot | represents a modulo operation;
and 4, step 4: calculating mean index m of longitudinal prediction error sequence by using formula (4) A ：
Figure FDA0003005227210000013
And 5: calculating fluctuation index sigma of longitudinal prediction error sequence by using formula (5) A ：
Figure FDA0003005227210000014
Step 6: calculating the transverse prediction error sequence gamma of the j sampling point by using the formula (6) err (j)：
Figure FDA0003005227210000021
In formula (6): im (·) represents an imaginary part, Re (·) represents an actual part, and arctan (·) is an arctangent function;
and 7: calculating the mean index m of the transverse prediction error sequence by using the formula (7) Γ ：
Figure FDA0003005227210000022
And 8: calculating a fluctuation index sigma of a transverse prediction error sequence by using equation (8) Γ ：
Figure FDA0003005227210000023
And step 9: the prediction error pair e at the j time is established by using the formula (9) j ；
e j ＝(A err (j),Γ err (j)),j＝0,1,2,…,N-1 (9)
Step 10: the prediction error at the j time is paired with e j Classified into K types:
step 10-1: a set of prediction error pairs E is established using equation (10):
E＝{e j |j＝0,1,2,…,N-1} (10)
step 10-2: randomly extracting K different numbers in the set {0,1,2, …, N-1}, and recording the numbers as i 1 ,i 2 ,…,i k ,…,i K (ii) a Wherein i k Representing the k-th number, and predicting the i-th number in the error pair set E k An error pair
Figure FDA0003005227210000024
Put into the kth cluster C k In the formula (11), K is more than or equal to 1 and less than or equal to K, and the kth cluster C is initialized according to the formula (11) k Cluster-like center of
Figure FDA0003005227210000025
Figure FDA0003005227210000026
In formula (11):
Figure FDA0003005227210000027
respectively represent the kth class cluster C k Cluster-like center of
Figure FDA0003005227210000028
Longitudinal error and lateral error of (a); a. the err (i k ),Γ err (i k ) Respectively representing the ith in the set E of prediction error pairs k An error pair
Figure FDA0003005227210000029
Longitudinal error and lateral error of (d); and is
Figure FDA00030052272100000210
Step 10-3: the prediction error pair e at the j-th time is calculated by equation (12) j With the kth cluster center
Figure FDA00030052272100000211
Of Euclidean distance d jk ：
Figure FDA00030052272100000212
Step 10-4: the nearest prediction error pair is divided into corresponding class clusters using equation (13):
Figure FDA0003005227210000031
in formula (13): u denotes union operation of sets, λ j Represents a prediction error pair e with the j-th time j Cluster number of closest class, and λ j ＝argmin k＝1,2,…,K (d jk ) And argmin represents the Euclidean distance d jk A minimum value of k;
step 10-5: updating the cluster center using equation (14);
Figure FDA0003005227210000032
in formula (14):
Figure FDA0003005227210000033
representing updated class C k Center of (e), e δ Represents the kth class cluster C k Of (1), M represents the kth class cluster C k The number of included prediction error pairs;
step 10-6: repeating the steps 10-3 to 10-5 until the kth cluster class C k K is more than or equal to 1 and less than or equal to K until the center of the key is not changed;
step 10-7: output the kth class C k ，1≤k≤K。
CN202110360165.XA 2021-04-02 2021-04-02 Wind power prediction error analysis and classification method based on S transformation Active CN113033904B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110360165.XA CN113033904B (en) 2021-04-02 2021-04-02 Wind power prediction error analysis and classification method based on S transformation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110360165.XA CN113033904B (en) 2021-04-02 2021-04-02 Wind power prediction error analysis and classification method based on S transformation

Publications (2)

Publication Number Publication Date
CN113033904A CN113033904A (en) 2021-06-25
CN113033904B true CN113033904B (en) 2022-09-13

Family

ID=76454125

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110360165.XA Active CN113033904B (en) 2021-04-02 2021-04-02 Wind power prediction error analysis and classification method based on S transformation

Country Status (1)

Country Link
CN (1) CN113033904B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103440529A (en) * 2013-08-12 2013-12-11 新疆大学 Lateral error translation modification method of wind electricity power forecast system
CN106846173A (en) * 2016-12-30 2017-06-13 国网新疆电力公司电力科学研究院 Short-term wind power forecast method based on EWT ESN
CN107230977A (en) * 2017-05-05 2017-10-03 浙江工商大学 Wind power forecasting method based on error correction and Lifting Wavelet combination forecasting
CN111697621A (en) * 2020-06-23 2020-09-22 石家庄铁道大学 Short-term wind power prediction method based on EWT-PDBN combination
CN112035783A (en) * 2020-09-02 2020-12-04 合肥工业大学 Wind power characteristic evaluation method based on time-frequency analysis
CN112330042A (en) * 2020-11-17 2021-02-05 合肥工业大学 Power distribution network reconstruction method based on self-adaptive fuzzy C-means clustering scene division
CN112580900A (en) * 2021-02-23 2021-03-30 国能日新科技股份有限公司 Short-term power prediction method and system based on single fan modeling

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016136237A1 (en) * 2015-02-23 2016-09-01 Okinawa Institute of Science and Technology Graduate University System and method of determining forecast error for renewable energy fluctuations

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103440529A (en) * 2013-08-12 2013-12-11 新疆大学 Lateral error translation modification method of wind electricity power forecast system
CN106846173A (en) * 2016-12-30 2017-06-13 国网新疆电力公司电力科学研究院 Short-term wind power forecast method based on EWT ESN
CN107230977A (en) * 2017-05-05 2017-10-03 浙江工商大学 Wind power forecasting method based on error correction and Lifting Wavelet combination forecasting
CN111697621A (en) * 2020-06-23 2020-09-22 石家庄铁道大学 Short-term wind power prediction method based on EWT-PDBN combination
CN112035783A (en) * 2020-09-02 2020-12-04 合肥工业大学 Wind power characteristic evaluation method based on time-frequency analysis
CN112330042A (en) * 2020-11-17 2021-02-05 合肥工业大学 Power distribution network reconstruction method based on self-adaptive fuzzy C-means clustering scene division
CN112580900A (en) * 2021-02-23 2021-03-30 国能日新科技股份有限公司 Short-term power prediction method and system based on single fan modeling

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
Short term wind power forecasting using time series neural networks;Mohammadsaleh Zakerinia;《Proceedings of the 2011 Emerging M&S Applications in Industry and Academia Symposium》;20110403;全文 *
以预测区间边界为约束的风电功率场景预测方法;张海波等;《电网与清洁能源》;20181225(第12期);全文 *
基于云模型和k-means聚类的风电场储能容量优化配置方法;吴杰 等;《电力系统自动化》;20181225;第42卷(第24期);全文 *
考虑功率转换系统停运的风电场储能容量优化;金志君 齐先军;《电器与能效管理技术》;20160229(第3期);全文 *

Also Published As

Publication number Publication date
CN113033904A (en) 2021-06-25

Similar Documents

Publication Publication Date Title
CN111830408B (en) Motor fault diagnosis system and method based on edge calculation and deep learning
CN110991725B (en) RBF ultra-short-term wind power prediction method based on wind speed frequency division and weight matching
CN110059845B (en) Metering device clock error trend prediction method based on time sequence evolution gene model
CN112818604A (en) Wind turbine generator risk degree assessment method based on wind power prediction
CN112507479B (en) Oil drilling machine health state assessment method based on manifold learning and softmax
Li et al. A two-tier wind power time series model considering day-to-day weather transition and intraday wind power fluctuations
CN110533007A (en) A kind of vehicle-mounted strain of bridge influences the identification of line feature intelligent and extracting method
CN111695724A (en) Wind speed prediction method based on hybrid neural network model
CN116380445A (en) Equipment state diagnosis method and related device based on vibration waveform
CN108830405B (en) Real-time power load prediction system and method based on multi-index dynamic matching
CN114298132A (en) Wind power prediction method and device and electronic equipment
CN110852492A (en) Photovoltaic power ultra-short-term prediction method for finding similarity based on Mahalanobis distance
CN112800540B (en) Aircraft engine load spectrum task segment modeling method based on Gaussian process regression
CN113033904B (en) Wind power prediction error analysis and classification method based on S transformation
CN113361782A (en) Photovoltaic power generation power short-term rolling prediction method based on improved MKPLS
CN116070669B (en) Workshop energy consumption prediction method and management system based on improved deep belief network
CN115795341A (en) Two-dimensional piston pump health state assessment method based on variable rotating speed
CN116070458A (en) New wind power plant scene generation method based on RAC-GAN
CN115081487A (en) Cross-time radiation source individual identification method based on multi-domain transfer learning
CN103678953A (en) Biological fermentation yield on-line forecasting method based on Bayes combination neural network
CN113449920A (en) Wind power prediction method, system and computer readable medium
Liu et al. Wind-speed forecasting model based on DBN-Elman combined with improved PSO-HHT
CN113256018B (en) Wind power ultra-short term probability prediction method based on conditional quantile regression model
CN117312895A (en) Structural moving part performance degradation trend analysis and fault prediction method based on clustering algorithm
CN113569951B (en) Near infrared quantitative analysis model construction method based on generation countermeasure network

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