CN112766566A - FA-FFCM-based station terminal load prediction method - Google Patents

FA-FFCM-based station terminal load prediction method Download PDF

Info

Publication number
CN112766566A
CN112766566A CN202110054959.3A CN202110054959A CN112766566A CN 112766566 A CN112766566 A CN 112766566A CN 202110054959 A CN202110054959 A CN 202110054959A CN 112766566 A CN112766566 A CN 112766566A
Authority
CN
China
Prior art keywords
firefly
clustering
ffcm
day
fireflies
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110054959.3A
Other languages
Chinese (zh)
Inventor
王二王
孙侃
卜权
丁旸
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.)
Jiayuan Technology Co Ltd
Original Assignee
Jiayuan Technology 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 Jiayuan Technology Co Ltd filed Critical Jiayuan Technology Co Ltd
Priority to CN202110054959.3A priority Critical patent/CN112766566A/en
Publication of CN112766566A publication Critical patent/CN112766566A/en
Pending legal-status Critical Current

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
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • 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
    • G06Q50/10—Services
    • G06Q50/26—Government or public services

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Tourism & Hospitality (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • General Business, Economics & Management (AREA)
  • Development Economics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Marketing (AREA)
  • Quality & Reliability (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Operations Research (AREA)
  • Game Theory and Decision Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Educational Administration (AREA)
  • Probability & Statistics with Applications (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Primary Health Care (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a station terminal load prediction method based on FA-FFCM, comprising the following steps: acquiring historical electricity consumption information data, and obtaining historical load and daily characteristic vectors as clustering samples by combining meteorological data; clustering the clustering samples based on an FA-FFCM (firefly-fast fuzzy C-means clustering) method, determining the optimal clustering number, finding out the class to which the day before the predicted day belongs, and determining the class as a similar day; and step three, taking the loads of the similar days as prediction samples, and predicting the loads at the integral point moment of the prediction day through a support vector machine. The method improves the accuracy of load prediction, has better self-adaptive characteristic, and can be applied to load prediction of the power system.

Description

FA-FFCM-based station terminal load prediction method
Technical Field
The invention relates to the technical field of station terminal load prediction, in particular to a FA-FFCM-based station terminal load prediction method.
Background
The station terminal load prediction usually adopts a clustering algorithm, including a fuzzy clustering algorithm, a K-Means algorithm and the like; the classical fuzzy clustering algorithm (fuzzy C-means clustering algorithm, FCM) minimizes the objective function (C-means function), and obtains the best classification of the data set by iteratively correcting the clustering membership matrix and the clustering center. According to the maximum membership criterion, attributing the data to the class with the maximum membership; and updating the cluster center position according to the membership degree of the data samples in each class. And repeating the iterative loop to divide the samples with larger similar lines in the data set into the same class, and divide the samples with larger differences into different classes. The FCM algorithm is simple to implement, has strong local search capability, effectively solves uncertain problems, but has defects, and mainly comprises the following points: the clustering effect of the FCM algorithm is greatly influenced by the initial value; the FCM algorithm adopts a local search method to solve the fuzzy clustering problem, and is easy to fall into local optimization; the size of the weighting index has important influence on the clustering process and the result; the calculation amount of the algorithm is large, when the data set is large, the calculation cost is greatly increased along with the iteration times, and the execution efficiency is poor.
The method results in lower accuracy of the load prediction result and low calculation speed.
Disclosure of Invention
In order to solve the defects in the prior art, the invention provides a station terminal load prediction method based on FA-FFCM, which solves the problems of low accuracy and low calculation speed of the existing load prediction method.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows: a terminal load prediction method of a station based on FA-FFCM comprises the following steps:
acquiring historical electricity consumption information data, and obtaining historical load and daily characteristic vectors as clustering samples by combining meteorological data;
clustering the clustering samples based on an FA-FFCM (firefly-fast fuzzy C-means clustering) clustering method, determining the optimal clustering number, finding out the class to which the day before the predicted day belongs, and determining the class as a similar day;
and step three, taking the loads of the similar days as prediction samples, and predicting the loads at the integral point moment of the prediction day through a support vector machine.
Further, the day eigenvector includes a date type, a lowest temperature, a highest temperature, rainfall, and season.
Further, the daily feature vector is represented by Z, and Z is (F, T)L，THR, S), wherein,
1) f is a date type, the date types are the same, and the similarity is 1; the same is weekday/weekend but the date types are different, and the similarity is 0.6; the date type and weekday/weekend were both different by 0.3.
2)TLThe temperature is the lowest temperature, 1 is the temperature below 10 ℃,2 is the temperature between 10 ℃ and 20 ℃, and 3 is the temperature above 20 ℃;
3)THthe highest temperature is 1, the temperature is below 25 ℃,2 is 25-30 ℃ and 3 is above 30 ℃;
4) r is rainfall, and 0 is no rain; l is light rain; 2, medium rain; 3, heavy rain;
5) s is season, 0 is spring; 1 is summer; 2 is autumn; and 3, in winter.
Further, the second step includes the steps of:
step 1, initializing parameters of a firefly algorithm, wherein firefly individuals comprise parameters needing to be optimized by an FFCM (flexible flat cm) clustering algorithm, randomly distributing fireflies, determining the movement direction of the fireflies, updating the positions of the fireflies, calculating the brightness of the fireflies, outputting a global extreme point and an optimal individual value if iteration times are met, taking the position of the optimal individual value as an initial clustering center, and otherwise, re-determining the movement direction of the fireflies;
step 2, calculating and updating a membership matrix and a clustering center of the FFCM clustering algorithm until the error is smaller than a set error value;
step 3, outputting a clustering center, and determining the optimal clustering number;
and 4, finding out the class to which the day before the predicted day belongs, and determining the day as a similar day.
Further, the step 1 includes the steps of:
1) setting the number of firefliesIs n, maximum attractive force beta0The light intensity absorption coefficient gamma, the step factor a and the maximum iteration number MaxGeneration;
2) randomly initializing the position of firefly, and calculating the target function value of firefly as the respective maximum fluorescence brightness I0；
3) Calculating the relative brightness and attraction of the fireflies in the population according to the formulas (1) and (2), and determining the movement direction of the fireflies according to the relative brightness;
relative luminance I of firefly I to firefly jijComprises the following steps:
Figure BDA0002900279830000031
attraction beta of firefly i to firefly jijComprises the following steps:
Figure BDA0002900279830000032
4) calculating the difference value between the maximum fluorescence brightness and the average fluorescence brightness of each particle, taking the fluorescence brightness difference value of the firefly as the input of an optimization model, adaptively adjusting the position of the firefly, and updating the position of the firefly to obtain a progeny firefly population;
5) dynamically adjusting the offspring firefly individuals to generate a new generation firefly population;
6) updating the space position of the firefly according to the formula (3), and randomly moving the firefly at the optimal position;
assuming that the jth firefly is attracted by the ith firefly so that the jth firefly carries out position updating, the updating formula is as follows:
xj(t+1)＝xj(t)+βij(rij)(xi(t)-xj(t))+aξj (3)
where t is expressed as the number of iterations of the algorithm, xj(t) is the position of the jth iteration of the j firefly, ξjIs a random number vector obtained by uniformly distributing the jth firefly, and a is a step factor;
7) recalculating the brightness of the firefly according to the updated position of the firefly;
8) when the maximum searching times is reached, turning to 9); otherwise, increasing the searching times by 1, turning to the step 3), and performing the next searching;
9) and outputting a global extreme point and an optimal individual value, wherein the position of the optimal individual is the initial clustering center.
Further, the individual of the firefly is dynamically adjusted, including introducing immigration operators, filtering similar individuals and dynamically supplementing new filial generation individuals.
Further, calculating a membership matrix and a clustering center of the updated FFCM clustering algorithm, including:
let sample data set Y ═ Y1,y2,...,yi,...,yMM denotes the number of samples in the dataset, each sample Y on YiThere are G characteristic parameters, (i 1, 2.. M), i.e. yi＝{yi1,yi2,...,yig,...,yiGIn which y isigDenotes yiThe g-th eigenvalue of (c); assuming that the data set is divided into c clusters, the cluster center V of the c clusters is ═ V1,v2,...,vcLet uijRepresenting a sample yiMembership to the jth cluster centre vjThe degree of (d); membership matrix U ═ Uij}c×MWherein u isijThe following conditions are satisfied:
Figure BDA0002900279830000051
the objective function of the FCM cluster is defined as:
Figure BDA0002900279830000052
wherein m is a blurring factor; dij(yi,vj) Represents the Euclidean distance y between the ith sample and the jth cluster centeri-vj||；
Degree of membership and centre of clusteringUpdating function uijAnd vi:
Figure BDA0002900279830000053
Figure BDA0002900279830000054
Wherein j is more than or equal to 1 and less than or equal to c, and i is more than or equal to 1 and less than or equal to M.
Taking the historical load and the day characteristic vector as a sample of cluster analysis, carrying out cluster analysis through an FA-FFCM algorithm to obtain a cluster result, finding a class similar to the situation of a predicted day as a similar day, and then carrying out support vector machine prediction on the load at the integral point moment of the predicted day by taking similar day data as sample data;
has the advantages that:
the improved FA algorithm has the advantages of accelerating convergence speed, increasing group diversity, overcoming the defect that the FFCM algorithm is sensitive to an initial clustering center, improving the accuracy of load prediction, having better self-adaptive characteristic and being applicable to the practice of load prediction of a power system.
The method can accurately predict the load value.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
Detailed Description
The present invention is further described in detail below with reference to the drawings and examples so that those skilled in the art can practice the invention with reference to the description.
As shown in fig. 1, a method for predicting terminal load of a station based on FA-FFCM includes the steps of:
acquiring historical electricity consumption information data, and obtaining historical load and daily characteristic vectors as clustering samples by combining meteorological data;
the daily feature vector is represented by Z, where Z is (F, T)L，THR, S), wherein,
1) f is a date type, the date types are the same, and the similarity is 1; the same is weekday/weekend but the date types are different, and the similarity is 0.6; the date type and weekday/weekend were both different by 0.3.
2)TLThe temperature is the lowest temperature, 1 is the temperature below 10 ℃,2 is the temperature between 10 ℃ and 20 ℃, and 3 is the temperature above 20 ℃;
3)THthe highest temperature is 1, the temperature is below 25 ℃,2 is 25-30 ℃ and 3 is above 30 ℃;
4) r is rainfall, and 0 is no rain; l is light rain; 2, medium rain; 3, heavy rain;
5) s is season, 0 is spring; 1 is summer; 2 is autumn; and 3, in winter.
Clustering the clustering samples based on an FA-FFCM (firefly-fast fuzzy C-means clustering) clustering method, determining the optimal clustering number, finding out the class to which the day before the predicted day belongs, and determining the class as a similar day;
the method specifically comprises the following steps:
step 1, initializing parameters of a firefly algorithm, wherein firefly individuals comprise parameters needing to be optimized by an FFCM (flexible flat cm) clustering algorithm, randomly distributing fireflies, determining the movement direction of the fireflies, updating the positions of the fireflies, calculating the brightness of the fireflies, outputting a global extreme point and an optimal individual value if iteration times are met, taking the position of the optimal individual value as an initial clustering center, and otherwise, re-determining the movement direction of the fireflies;
the method specifically comprises the following steps:
1) setting the number of fireflies as n and the maximum attraction beta0The light intensity absorption coefficient gamma, the step factor a and the maximum iteration number MaxGeneration;
2) randomly initializing the position of firefly, and calculating the target function value of firefly as the respective maximum fluorescence brightness I0；
3) Calculating the relative brightness and attraction of the fireflies in the population according to the formulas (1) and (2), and determining the movement direction of the fireflies according to the relative brightness;
relative luminance I of firefly I to firefly jijComprises the following steps:
Figure BDA0002900279830000071
attraction beta of firefly i to firefly jijComprises the following steps:
Figure BDA0002900279830000072
4) calculating the difference value between the maximum fluorescence brightness and the average fluorescence brightness of each particle, taking the fluorescence brightness difference value of the firefly as the input of an optimization model, adaptively adjusting the position of the firefly, and updating the position of the firefly to obtain a progeny firefly population;
5) dynamically adjusting the offspring firefly individuals to generate a new generation firefly population;
6) updating the space position of the firefly according to the formula (3), and randomly moving the firefly at the optimal position;
assuming that the jth firefly is attracted by the ith firefly so that the jth firefly carries out position updating, the updating formula is as follows:
xj(t+1)＝xj(t)+βij(rij)(xi(t)-xj(t))+aξj (3)
where t is expressed as the number of iterations of the algorithm, xj(t) is the position of the jth iteration of the j firefly, ξjIs a random number vector obtained by uniformly distributing the jth firefly, a is a step factor, and is usually a constant a ∈ [0,1 ∈ [ ]]. Obviously, the second term of the position formula depends on the attraction force, and the third term is a random term.
7) Recalculating the brightness of the firefly according to the updated position of the firefly;
8) when the maximum searching times is reached, turning to 9); otherwise, increasing the searching times by 1, turning to the step 3), and performing the next searching;
9) outputting a global extreme point and an optimal individual value, wherein the position of the optimal individual is an initial clustering center;
in order to prevent premature convergence, the firefly individuals can be dynamically adjusted, including introducing immigration operators, filtering similar individuals, and dynamically supplementing new filial individuals. Immigration operators are a good way to avoid premature. In the process of immigration, poor individuals can be eliminated in an accelerated way, and the diversity of solutions is increased. The immigration operator is an operation of eliminating the worst individual with a certain elimination rate (generally 15-20%) in the evolution process of each generation and then replacing the worst individual with the generated new individual. In order to accelerate the convergence rate, the operation of filtering similar individuals can be adopted, and the gene uniqueness is reduced. The filtering operation to delete similar individuals is: sorting the offspring individuals according to the fluorescence brightness, and sequentially calculating the generalized Hamming distance between similar individuals of which the fluorescence brightness difference is smaller than the threshold delta (the number of different corresponding positions in two character strings with the same length and based on e is called the generalized Hamming distance between the two character strings). If the difference value of the fluorescence brightness is smaller than the threshold delta and the generalized Hamming distance is smaller than the threshold eta, the individuals with smaller fluorescence brightness are filtered. After the filtering operation, new individuals are generated by the change from the excellent parent individuals. And (3) randomly carrying out a plurality of variations on the Q individuals with higher fluorescence brightness in the parent to generate new individuals, and adding the new individuals into the offspring. The new individuals inherit the pattern segments of the parent-generation superior individuals and generate new patterns, and the new patterns are easy to combine with other individuals to generate new superior child individuals. And the number of new individuals added is related to the number of filter operations deleted. If the population gene unicity increases, the number of similar individuals filtered out increases, and the number of new individuals supplemented increases; otherwise, only a small amount of similar individuals are filtered, even the similar individuals are not filtered, and the number of the supplemented new individuals is reduced. Thus, the problem that the groups fall into local solutions due to lack of diversity is solved dynamically.
Step 2, calculating and updating a membership matrix and a clustering center of the FFCM clustering algorithm until the error is smaller than a set error value;
let sample data set Y ═ Y1,y2,...,yi,...,yMM denotes the number of samples in the dataset, each sample Y on YiThere are G characteristic parameters, (i 1, 2.. M), i.e. yi＝{yi1,yi2,...,yig,...,yiGIn which y isigDenotes yiTog characteristic values; assuming that the data set is divided into c clusters, the cluster center V of the c clusters is ═ V1,v2,...,vcLet uijRepresenting a sample yiMembership to the jth cluster centre vjTo the extent of (c). Membership matrix U ═ Uij}c×MWherein u isijThe following conditions are satisfied:
Figure BDA0002900279830000091
the objective function of the FCM cluster is defined as:
Figure BDA0002900279830000092
where m is a blurring factor whose magnitude affects the degree of blurring of the clustering result, typically m ∈ [1, ∞ [ ]]The optimal value of m in practical application is [1.5.2.5 ]]In the range, typically m is 2; dij(yi,vj) Represents the Euclidean distance y between the ith sample and the jth cluster centeri-vj||。
Updating function u of membership degree and clustering centerijAnd vi:
Figure BDA0002900279830000101
Figure BDA0002900279830000102
Wherein j is more than or equal to 1 and less than or equal to c, and i is more than or equal to 1 and less than or equal to M.
And 3, outputting the clustering centers and determining the optimal clustering number.
And 4, finding out the class to which the day before the predicted day belongs according to the maximum membership rule, and determining the class as a similar day. The load change rule of the current day influences the load rule of the next day, and the class with the shortest distance to the current day can be found from the clustering result and used as the similar day;
and step three, taking the loads of the similar days as prediction samples, and predicting the loads at the integral point moment of the prediction day through a support vector machine.
The method has the advantages that the improved FA algorithm is high in convergence speed and increased in group diversity, the defect that the FFCM algorithm is sensitive to an initial clustering center is overcome, the load prediction precision is improved, the method has good self-adaptive characteristics, and the method can be applied to power system load prediction practice.
While embodiments of the invention have been described above, it is not limited to the applications set forth in the description and the embodiments, which are fully applicable in various fields of endeavor to which the invention pertains, and further modifications may readily be made by those skilled in the art, it being understood that the invention is not limited to the details shown and described herein without departing from the general concept defined by the appended claims and their equivalents.

Claims (7)

1. A method for predicting terminal load of a station based on FA-FFCM (flexible flat cable), which is characterized by comprising the following steps:
acquiring historical electricity consumption information data, and obtaining historical load and daily characteristic vectors as clustering samples by combining meteorological data;
clustering the clustering samples based on an FA-FFCM clustering method, determining the optimal clustering number, finding out the class to which the day before the predicted day belongs, and determining the class as a similar day;
and step three, taking the loads of the similar days as prediction samples, and predicting the loads at the integral point moment of the prediction day through a support vector machine.
2. The FA-FFCM-based substation terminal load forecasting method as recited in claim 1, wherein the day eigenvector includes a date type, a minimum air temperature, a maximum air temperature, rainfall and season.
3. The FA-FFCM-based station terminal load prediction method as claimed in claim 2, wherein the daily eigenvector is denoted by Z, and Z ═ is (F, T)L，THR, S), wherein,
1) f is a date type, the date types are the same, and the similarity is 1; the same is weekday/weekend but the date types are different, and the similarity is 0.6; the date type and weekday/weekend were both different by 0.3.
2)TLThe temperature is the lowest temperature, 1 is the temperature below 10 ℃,2 is the temperature between 10 ℃ and 20 ℃, and 3 is the temperature above 20 ℃;
3)THthe highest temperature is 1, the temperature is below 25 ℃,2 is 25-30 ℃ and 3 is above 30 ℃;
4) r is rainfall, and 0 is no rain; l is light rain; 2, medium rain; 3, heavy rain;
5) s is season, 0 is spring; 1 is summer; 2 is autumn; and 3, in winter.
4. The FA-FFCM-based station terminal load prediction method as claimed in claim 1, wherein the second step comprises the steps of:
step 1, initializing parameters of a firefly algorithm, wherein firefly individuals comprise parameters needing to be optimized by an FFCM (flexible flat cm) clustering algorithm, randomly distributing fireflies, determining the movement direction of the fireflies, updating the positions of the fireflies, calculating the brightness of the fireflies, outputting a global extreme point and an optimal individual value if iteration times are met, taking the position of the optimal individual value as an initial clustering center, and otherwise, re-determining the movement direction of the fireflies;
step 2, calculating and updating a membership matrix and a clustering center of the FFCM clustering algorithm until the error is smaller than a set error value;
step 3, outputting a clustering center, and determining the optimal clustering number;
and 4, finding out the class to which the day before the predicted day belongs, and determining the day as a similar day.
5. The FA-FFCM-based station terminal load prediction method as claimed in claim 4, wherein the step 1 comprises the steps of:
1) setting the number of fireflies as n and the maximum attraction beta0Light intensity absorption coefficient γ, stepA long factor a, maximum iteration number MaxGeneration;
2) randomly initializing the position of firefly, and calculating the target function value of firefly as the respective maximum fluorescence brightness I0；
3) Calculating the relative brightness and attraction of the fireflies in the population according to the formulas (1) and (2), and determining the movement direction of the fireflies according to the relative brightness;
relative luminance I of firefly I to firefly jijComprises the following steps:
Figure FDA0002900279820000021
attraction beta of firefly i to firefly jijComprises the following steps:
Figure FDA0002900279820000022
4) calculating the difference value between the maximum fluorescence brightness and the average fluorescence brightness of each particle, taking the fluorescence brightness difference value of the firefly as the input of an optimization model, adaptively adjusting the position of the firefly, and updating the position of the firefly to obtain a progeny firefly population;
5) dynamically adjusting the offspring firefly individuals to generate a new generation firefly population;
6) updating the space position of the firefly according to the formula (3), and randomly moving the firefly at the optimal position;
assuming that the jth firefly is attracted by the ith firefly so that the jth firefly carries out position updating, the updating formula is as follows:
xj(t+1)＝xj(t)+βij(rij)(xi(t)-xj(t))+aξj (3)
where t is expressed as the number of iterations of the algorithm, xj(t) is the position of the jth iteration of the j firefly, ξjIs a random number vector obtained by uniformly distributing the jth firefly, and a is a step factor;
7) recalculating the brightness of the firefly according to the updated position of the firefly;
8) when the maximum searching times is reached, turning to 9); otherwise, increasing the searching times by 1, turning to the step 3), and performing the next searching;
9) and outputting a global extreme point and an optimal individual value, wherein the position of the optimal individual is the initial clustering center.
6. The FA-FFCM-based substation terminal load prediction method as claimed in claim 5, wherein the dynamic adjustment of firefly individuals includes introducing immigration operators, filtering similar individuals, and dynamically supplementing new children.
7. The FA-FFCM-based station terminal load prediction method as claimed in claim 1, wherein calculating the membership matrix and the clustering center of the updated FFCM clustering algorithm comprises:
let sample data set Y ═ Y1,y2,...,yi,...,yMM denotes the number of samples in the dataset, each sample Y on YiThere are G characteristic parameters, (i 1, 2.. M), i.e. yi＝{yi1,yi2,...,yig,...,yiGIn which y isigDenotes yiThe g-th eigenvalue of (c); assuming that the data set is divided into c clusters, the cluster center V of the c clusters is ═ V1,v2,...,vcLet uijRepresenting a sample yiMembership to the jth cluster centre vjThe degree of (d); membership matrix U ═ Uij}c×MWherein u isijThe following conditions are satisfied:
Figure FDA0002900279820000041
the objective function of the FCM cluster is defined as:
Figure FDA0002900279820000042
wherein m is a blurring factor; dij(yi,vj) Represents the Euclidean distance y between the ith sample and the jth cluster centeri-vj||；
Updating function u of membership degree and clustering centerijAnd vi:
Figure FDA0002900279820000043
Figure FDA0002900279820000044
Wherein j is more than or equal to 1 and less than or equal to c, and i is more than or equal to 1 and less than or equal to M.
CN202110054959.3A 2021-01-15 2021-01-15 FA-FFCM-based station terminal load prediction method Pending CN112766566A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110054959.3A CN112766566A (en) 2021-01-15 2021-01-15 FA-FFCM-based station terminal load prediction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110054959.3A CN112766566A (en) 2021-01-15 2021-01-15 FA-FFCM-based station terminal load prediction method

Publications (1)

Publication Number Publication Date
CN112766566A true CN112766566A (en) 2021-05-07

Family

ID=75701830

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110054959.3A Pending CN112766566A (en) 2021-01-15 2021-01-15 FA-FFCM-based station terminal load prediction method

Country Status (1)

Country Link
CN (1) CN112766566A (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103123665A (en) * 2012-07-31 2013-05-29 上海交通大学 Short-term power load forecasting method based on fuzzy clustering similar day
CN103390200A (en) * 2013-07-18 2013-11-13 国家电网公司 Photovoltaic power station electricity generation output power forecasting method based on similar days
CN111934718A (en) * 2020-08-04 2020-11-13 佳源科技有限公司 Method for applying improved minimum spanning tree algorithm to electric power broadband carrier
CN112132351A (en) * 2020-09-28 2020-12-25 华北电力大学 Electric vehicle charging load clustering prediction method and device

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103123665A (en) * 2012-07-31 2013-05-29 上海交通大学 Short-term power load forecasting method based on fuzzy clustering similar day
CN103390200A (en) * 2013-07-18 2013-11-13 国家电网公司 Photovoltaic power station electricity generation output power forecasting method based on similar days
CN111934718A (en) * 2020-08-04 2020-11-13 佳源科技有限公司 Method for applying improved minimum spanning tree algorithm to electric power broadband carrier
CN112132351A (en) * 2020-09-28 2020-12-25 华北电力大学 Electric vehicle charging load clustering prediction method and device

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
吴亚雄;高崇;曹华珍;陈吕鹏;唐俊熙;李浩;: "基于灰狼优化聚类算法的日负荷曲线聚类分析", 电力系统保护与控制, no. 06 *
李娜: "基于相似日选取的马尔科夫短期负荷预测方法", 电工技术, pages 1 - 3 *
林睦纲 等: "一种基于萤火虫算法的模糊聚类方法", 《计算机工程与应用》, pages 2 - 4 *
骆东松;李雄伟;赵小强;: "基于人工萤火虫的模糊聚类算法研究", 工业仪表与自动化装置, no. 02 *

Similar Documents

Publication Publication Date Title
CN109389608B (en) There is the fuzzy clustering image partition method of noise immunity using plane as cluster centre
CN108564192B (en) Short-term photovoltaic power prediction method based on meteorological factor weight similarity day
CN109766950B (en) Industrial user short-term load prediction method based on morphological clustering and LightGBM
CN106971091B (en) Tumor identification method based on deterministic particle swarm optimization and support vector machine
CN112069310A (en) Text classification method and system based on active learning strategy
CN112685504B (en) Production process-oriented distributed migration chart learning method
CN109002858B (en) Evidence reasoning-based integrated clustering method for user behavior analysis
CN111583031A (en) Application scoring card model building method based on ensemble learning
CN116503676B (en) Picture classification method and system based on knowledge distillation small sample increment learning
CN108875933A (en) A kind of transfinite learning machine classification method and the system of unsupervised Sparse parameter study
CN110287985B (en) Depth neural network image identification method based on variable topology structure with variation particle swarm optimization
CN112700060A (en) Station terminal load prediction method and prediction device
CN106528705A (en) Repeated record detection method and system based on RBF neural network
CN110895772A (en) Electricity sales amount prediction method based on combination of grey correlation analysis and SA-PSO-Elman algorithm
CN108596204B (en) Improved SCDAE-based semi-supervised modulation mode classification model method
WO2023124342A1 (en) Low-cost automatic neural architecture search method for image classification
CN115358305A (en) Incremental learning robustness improving method based on iterative generation of boundary samples
CN117277279A (en) Deep learning short-term load prediction method based on particle swarm optimization
CN114882335A (en) Intelligent image sensing device for sensing and computing cloud integration based on federal learning framework
CN109919374A (en) Prediction of Stock Price method based on APSO-BP neural network
CN112766566A (en) FA-FFCM-based station terminal load prediction method
CN112446435A (en) City data classification method and system
CN110059871B (en) Photovoltaic power generation power prediction method
CN116862043A (en) Load prediction method and system based on multilayer feedforward network
CN111461237A (en) QPSO-based ABAC model for optimizing K-Means

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