CN111832838B - Method for predicting short-term wind power generation output power - Google Patents

Method for predicting short-term wind power generation output power Download PDF

Info

Publication number
CN111832838B
CN111832838B CN202010723186.9A CN202010723186A CN111832838B CN 111832838 B CN111832838 B CN 111832838B CN 202010723186 A CN202010723186 A CN 202010723186A CN 111832838 B CN111832838 B CN 111832838B
Authority
CN
China
Prior art keywords
optimal
foraging
vector machine
machine model
support vector
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.)
Expired - Fee Related
Application number
CN202010723186.9A
Other languages
Chinese (zh)
Other versions
CN111832838A (en
Inventor
李玲玲
刘佳琪
韩新同
陈文泉
刘汉民
常云彪
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.)
Hebei University of Technology
Original Assignee
Hebei 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 Hebei University of Technology filed Critical Hebei University of Technology
Priority to CN202010723186.9A priority Critical patent/CN111832838B/en
Publication of CN111832838A publication Critical patent/CN111832838A/en
Application granted granted Critical
Publication of CN111832838B publication Critical patent/CN111832838B/en
Expired - Fee Related 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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
    • 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
    • G06Q50/06—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)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • General Business, Economics & Management (AREA)
  • Evolutionary Computation (AREA)
  • Tourism & Hospitality (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Water Supply & Treatment (AREA)
  • Molecular Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Public Health (AREA)
  • Medical Informatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Primary Health Care (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 relates to a method for predicting short-term wind power generation output power, which is technically characterized by comprising the following steps: acquiring input data and output data of wind power generation, and carrying out normalization processing on the data; setting parameters of an improved optimal foraging algorithm and a support vector machine model; running an improved optimal foraging algorithm to obtain an optimal penalty factor in the support vector machine model and an optimal parameter of a kernel function in the support vector machine model; the optimized optimal parameters are brought into a support vector machine model, and the support vector machine model optimized by the optimized optimal foraging algorithm is trained; and inputting the prediction data into a support vector machine model optimized by the improved optimal foraging algorithm to obtain a prediction result, and performing reverse normalization on the prediction result. The method realizes the reliable and high-precision prediction function of the short-term wind power generation output power, effectively processes the hidden trouble of the wind power generation connected to the power grid for operation, and overcomes the defect of low prediction precision of the conventional short-term wind power generation output power prediction method.

Description

Method for predicting short-term wind power generation output power
Technical Field
The invention belongs to the technical field of wind power generation, and particularly relates to a method for predicting short-term wind power generation output power.
Background
In order to achieve the ambitious goal of sustainable energy development, the development and utilization of new energy sources are becoming more and more important in the energy industry today. Wind energy is one of the most attractive renewable energy sources, and has the characteristics of environmental protection, no need of transportation, low cost and inexhaustible energy. Therefore, wind power generation technology is considered as one of the important solutions to meet today's energy demand.
In recent decades, the wind power industry has been rapidly developing worldwide, and numerous wind power plants have been built in succession around the world, with the proportion of the number of wind power plants in the power grid increasing. Although wind power generation has the advantages of cleanness, environmental protection, low price and the like, the wind power generation still has some defects and limitations in the utilization process. Due to wind uncertainty, wind farms are a typical fluctuating and intermittent source of power. In addition, the output of the wind power generation output power is influenced by climate, weather and other factors, which also greatly increases the difficulty of accessing the wind power generation into the power grid and the complexity of power grid dispatching.
When unstable wind power generation is connected to a power grid, a large amount of harmonic waves are brought to the power grid, so that the power quality is reduced, and the reliability of power grid operation and the economy of power grid dispatching are reduced. When large-scale unstable wind power generation is connected to a power grid, huge impact is generated on the operation of the power grid, so that the voltage fluctuation of the power grid and the frequency change of the power grid are caused, the complexity of power grid scheduling is increased to a great extent, huge hidden dangers are brought to safe and reliable operation of the power grid, power equipment in the power grid can be damaged if the power grid is serious, and serious loss is brought to the power grid. Therefore, the method for predicting the wind power generation output power is an effective method for solving the problems of fluctuation and uncertain power of wind power generation and improving the power quality and the operation reliability of a power grid.
The prediction of the wind power generation output power is an important foundation for realizing the access of large-scale wind power generation to a power grid. The short-term wind power generation output power prediction is important for the arrangement of a daily power generation plan of a power grid and the realization of efficient operation and economic dispatching of the power grid. The short-term wind power generation output power prediction means prediction of wind power generation output power within 3 days (72 hours) in the future. The method has the advantages that the reliable and high-precision prediction of the short-term wind power generation output power is realized, the impact of wind power generation fluctuation on a power grid is effectively reduced, the occurrence rate of voltage fluctuation and frequency change of the power grid is reduced, the real-time adjustment of a power grid dispatching plan is facilitated, the power supply quality of the power grid is improved, effective power generation time intervals and the overhaul and maintenance of a fan can be reasonably arranged according to the prediction result of the short-term wind power generation output power, the operation cost of a power generation enterprise is reduced, and the economic benefit of wind power generation is improved.
In conclusion, how to accurately predict the short-term wind power generation output power and ensure that the impact on the operation of the power grid is reduced and the power grid is operated safely and reliably when the wind power generation is connected to the power grid is a problem which needs to be solved urgently at present.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a method for predicting the output power of short-term wind power generation.
The technical problem to be solved by the invention is realized by adopting the following technical scheme:
a method for predicting short-term wind power generation output power comprises the following steps:
step 1, acquiring input data and output data of wind power generation, and carrying out normalization processing on the data;
step 2, setting parameters of an improved optimal foraging algorithm and a support vector machine model;
step 3, running an improved optimal foraging algorithm to obtain an optimal penalty factor in the support vector machine model and optimal parameters of a kernel function in the support vector machine model;
step 4, bringing the optimized optimal parameters into a support vector machine model, and training the support vector machine model optimized by the improved optimal foraging algorithm;
and 5, inputting the prediction data into a support vector machine model optimized by the improved optimal foraging algorithm to obtain a prediction result, and performing reverse normalization on the prediction result.
The input data of the wind power generation is meteorological parameters comprising wind speed and wind direction, and the output data refers to wind power generation output power.
Further, in the step 1, the input data and the output data are divided into training data and prediction data, the training data include wind power generation output power, wind speed and wind direction sine-cosine values, and the prediction data include wind speed and wind direction sine-cosine values.
Further, the parameters set in step 2 include: the method comprises the steps of improving the population quantity in the optimal foraging algorithm, improving the maximum search times of the optimal foraging algorithm, supporting the search range of a penalty factor C in a vector machine model, supporting the range of a kernel function parameter g of the vector machine model and improving the population dimension in the optimal foraging algorithm.
Further, the specific implementation method of step 3 includes the following steps:
step 3.1, initializing foraging positions of the improved optimal foraging algorithm population, and calculating an objective function value of each animal individual;
step 3.2, sequencing the foraging position and the objective function value, and recording the current optimal foraging position and the optimal objective function value;
3.3, calculating a new foraging position of the individual animal, updating the foraging position of the individual animal, and calculating a new objective function value;
3.4, introducing the Cauchy variation to the position of the animal individual for updating;
3.5, introducing a differential variation strategy, continuously searching and updating foraging positions in population individuals, and judging whether the maximum searching times is reached; if the maximum search times are reached, acquiring an optimal punishment factor C in the support vector machine model and an optimal parameter g of a kernel function in the support vector machine model; if the maximum number of searches has not been reached, then the process returns to and continues to execute step 3.2.
The specific implementation method of the step 3.1 is as follows:
in the improved optimal foraging algorithm, the foraging position of an individual animal is a d-dimensional vector [ x ]1,…xi,…xd]T,xi∈[xL,xU],xLAnd xUAre respectively a variable xiAnd (3) searching the optimal position near the foraging position of the current improved optimal foraging algorithm on the upper and lower boundaries:
Figure BDA0002600756930000021
in the above formula, the first and second carbon atoms are,
Figure BDA0002600756930000022
the foraging position of the ith animal individual after the t search,
Figure BDA0002600756930000023
is an animal individual
Figure BDA0002600756930000024
New foraging position after updating, k is a scale factor, r1iAnd r2iIs uniformly distributed in [0,1 ]]A random value in between, and a random value,
Figure BDA0002600756930000025
is an animal individual
Figure BDA0002600756930000026
Updating the position increment of the foraging position;
increasing ith foraging position of jth animal individual in tth search
Figure BDA0002600756930000027
Comprises the following steps:
Figure BDA0002600756930000028
in the above formula, the first and second carbon atoms are,
Figure BDA0002600756930000031
and
Figure BDA0002600756930000032
respectively at the ith searching time, the ith foraging position where the b-th animal individual is located, the ith foraging position where the jth animal individual is located and the worst foraging position in the animal individual in the current searching;
Figure BDA0002600756930000033
and
Figure BDA0002600756930000034
are respectively as
Figure BDA0002600756930000035
And
Figure BDA0002600756930000036
the corresponding objective function value; the new foraging position of the individual animal can appear at the current individual animal position
Figure BDA0002600756930000037
At an arbitrary position in the vicinity, i.e.
Figure BDA0002600756930000038
To appear in
Figure BDA0002600756930000039
At an arbitrary position nearby, two random numbers r1iAnd r2iCan make it possible to
Figure BDA00026007569300000310
Plus, minus, or not plus or minus position increments
Figure BDA00026007569300000311
When r is1i＞r2iThen, then
Figure BDA00026007569300000312
When r is1i＜r2iThen, then
Figure BDA00026007569300000313
When r is1i＝r2iThen, then
Figure BDA00026007569300000314
Combining the two formulas to obtain the ith foraging position update of the jth individual animal in the tth search as:
Figure BDA00026007569300000315
the specific implementation method of the step 3.3 comprises the following steps: and the improved optimal foraging algorithm judges whether the updated position is better than the original position according to the objective function value of the individual position of the animal at the moment, and determines whether the updated position is used in the subsequent searching process, wherein the objective function is described as follows:
Figure BDA00026007569300000316
in the above formula, the first and second carbon atoms are,
Figure BDA00026007569300000317
is [0,1 ]]The random number of (2);
the step 3.4 is realized by adopting the following algorithm:
Figure BDA00026007569300000318
in the above formula, the first and second carbon atoms are,
Figure BDA00026007569300000319
is a vector linearly decreasing from 1 to 0, and gamma is uniformly distributed in [0,1 ]]A random value in between;
the specific implementation method of the step 3.5 is as follows:
the method comprises the following steps of carrying out mutation on vectors of population individuals by adopting a DE/rand/1 strategy in a differential mutation strategy, adding the differential mutation strategy at the later stage of each search, and describing as follows:
Figure BDA00026007569300000320
in the above formula, p1≠p2≠p3，
Figure BDA00026007569300000321
Is a difference vector, F ∈ [0.1,0.9 ∈ [ ]]As a scaling factor, hi,tObtaining a variation vector for the ith position in the t-th search, and then performing a crossover operation by:
Figure BDA00026007569300000322
in the above formula, vi,tA cross variable for the ith search location; j0 is a random value in the dimension, each crossover operation only involves one dimension of the individual, pCR ∈ [0,1 ∈]Is the cross probability;
and carrying out selection operation, reserving the optimal vector of the objective function value as a next generation individual, and expressing the selection operation as follows:
Figure BDA0002600756930000041
continuously searching and updating foraging positions in population individuals according to the formula, and judging whether the maximum searching times is reached; if the maximum search times are reached, acquiring an optimal punishment factor C in the support vector machine model and an optimal parameter g of a kernel function in the support vector machine model; if the maximum number of searches has not been reached, then the process returns to and continues to execute step 3.2.
Further, the specific implementation method of step 4 is as follows: inputting the optimal punishment factor C in the support vector machine model obtained in the step 3 and the optimal parameter g of the kernel function in the support vector machine model into the support vector machine model to form the support vector machine model optimized by the improved optimal foraging algorithm, and training the support vector machine model by using the optimal punishment factor C in the optimized support vector machine model and the optimal parameter g of the kernel function in the support vector machine model.
Further, the method for performing the inverse normalization processing in the step 5 comprises the following steps: selecting the root mean square error as an objective function RMSE of the support vector machine model, wherein the objective function is expressed as follows:
Figure BDA0002600756930000042
in the above formula, N is the number of wind power output power in the data to be predicted, PiIs the true value of the wind power output power, YiIs the predicted value of the output power.
The invention has the advantages and positive effects that:
1. the method adopts a combined artificial intelligence method, constructs a support vector machine model optimized by an improved optimal foraging algorithm to predict the output power of the short-term wind power generation, has higher-precision prediction capability on the output power of the short-term wind power generation with randomness, volatility and uncertainty, realizes the reliable and high-precision prediction function on the output power of the short-term wind power generation, is beneficial to reasonably arranging the economic dispatching of a power grid, can effectively treat the hidden danger of the operation of the wind power generation connected to the power grid, and simultaneously makes up the defect that the prediction precision of the existing short-term prediction method of the output power of the wind power generation is not high enough.
2. In the selection of the optimization algorithm, an improved optimal foraging algorithm is adopted for optimizing the support vector machine model, and the improved optimal foraging algorithm has good performance in the process of acquiring a penalty factor C and a kernel function parameter g of the support vector machine model in the support vector machine model; the Cauchy variation is used, so that large disturbance is generated near the current foraging position of the population individual, the searching position is made to have larger randomness, the searching capability of the algorithm is improved, and local optima are skipped out in time; a differential variation strategy is introduced in the searching process, so that the diversity of population individuals can be increased, and the premature phenomenon in the searching process is avoided.
3. The method can be used for predicting the output power of the short-term wind power generation, and can also be expanded to be used for predicting other fields.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a graph comparing predicted results of the support vector machine model optimized for the unmodified optimal foraging algorithm of the present invention;
fig. 3 is a histogram comparing predicted results of the support vector machine model optimized by the present invention and an unmodified optimal foraging algorithm.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
The design idea of the invention is as follows:
in the short-term prediction of the wind power generation output power, the selection of a short-term wind power generation output power prediction method is very critical. In recent years, with the continuous and deep research of the artificial intelligence method, people find that the artificial intelligence method has strong self-adaptive capacity, does not need to solve a complex mathematical expression, and can well reflect the nonlinear relation of input and output by adopting the artificial intelligence method to establish a nonlinear model between input data and output power of wind power generation. In addition, in addition to the single artificial intelligence method prediction model, the research of the combined artificial intelligence method prediction model further enriches the wind power prediction technology, and compared with the single artificial intelligence method prediction model, the combined artificial intelligence method prediction model can improve the short-term prediction precision of the wind power generation output power and effectively reduce the short-term wind power generation output power prediction error. Therefore, the method adopts a combined artificial intelligence method to construct the support vector machine model optimized by the improved optimal foraging algorithm to predict the output power of the short-term wind power generation, and solves the problems of impact on the operation of the power grid and huge hidden danger in the safe and reliable operation of the power grid when the wind power generation is connected into the power grid.
Based on the above design concept, the invention provides a short-term wind power generation output power prediction method, as shown in fig. 1, comprising the following steps:
step 1, acquiring input data and output data of wind power generation, and normalizing the data.
The input data of the wind power generation refers to meteorological parameters influencing the output power of the wind power generation, and the meteorological parameters mainly influencing the output power of the wind power generation are meteorological parameters such as wind speed and wind direction; the output data refers to the output power of wind power generation.
In the step, through meteorological data acquisition of a wind power plant, on-site wind power generation output power data acquisition is carried out to obtain data, a sine value and a cosine value corresponding to wind speed and wind direction are recorded as input data, output power at a corresponding moment is recorded as output data, and the data are divided into training data and prediction data. The training data comprises wind power generation output power, wind speed and wind direction sine and cosine values, and the prediction data comprises the wind speed and wind direction sine and cosine values. The wind power generation output power, the wind speed and the wind direction sine and cosine values in the training data and the wind speed and wind direction sine and cosine values in the prediction data are normalized, and the normalization principle by taking the wind power generation output power as an example is as follows:
Figure BDA0002600756930000051
in the formula (1), Pscale,iFor normalized wind power output power, PiIs the true value of the wind power output, PmaxAnd PminRespectively the maximum value and the minimum value in the original wind power.
In this embodiment, the data set is collected and selected from operation data of a first generator set 2017 in autumn of a La Haute born wind farm in france, the rated power of the generator set is 2050kW, the sampling interval of the data set is 1 hour, data of the first five days in the autumn of 2017 are selected as original training data, data of the last two days are used as data to be predicted, sine values and cosine values corresponding to wind speeds and wind directions are used as input data, and output power at corresponding moments is used as output data.
And 2, setting parameters of an improved optimal foraging algorithm and a support vector machine model.
The method comprises the steps of setting the number of population in the optimal foraging algorithm, the maximum search times of the optimal foraging algorithm, the search range of a penalty factor C in a support vector machine model, the range of a kernel function parameter g of the support vector machine model, the number of population dimensions in the optimal foraging algorithm and the like.
In this embodiment, it is noted that the population number in the improved optimal foraging algorithm is 30, the maximum search frequency of the improved optimal foraging algorithm is 300, the search range of the penalty factor C in the support vector machine model is [0.1,1200], the range of the kernel function parameter g of the support vector machine model is [0.01,100], and the population dimension in the improved optimal foraging algorithm is 2; the rest parameters in the improved optimal foraging algorithm are default values.
And 3, running an improved optimal foraging algorithm to obtain an optimal penalty factor C in the support vector machine model and an optimal parameter g of a kernel function in the support vector machine model.
In the step, the input data and the output data of the wind power generation, which are obtained in the step 1, and the parameters of the improved optimal foraging algorithm and the support vector machine, which are set in the step 2, are input into a computer, and the following processing of the data is realized by programming by means of MATLAB computer software, so that an optimal punishment factor C in a support vector machine model and an optimal parameter g of a kernel function in the support vector machine model are obtained;
and 3.1, initializing the foraging position of the improved optimal foraging algorithm population, and calculating the objective function value of each animal individual.
In the step, in the improved optimal foraging algorithm, the foraging position of each animal is a d-dimensional vector [ x ]1,…xi,…xd]T(xi∈[xL,xU],xLAnd xUAre respectively a variable xiUpper and lower boundaries), the optimal position is searched near the foraging position of the current improved optimal foraging algorithm
Figure BDA0002600756930000061
In the formula (2), the reaction mixture is,
Figure BDA0002600756930000062
the foraging position of the ith animal individual after the t search,
Figure BDA0002600756930000063
is an animal individual
Figure BDA0002600756930000064
New foraging position after updating, k is a scale factor, r1iAnd r2iIs uniformly distributed in [0,1 ]]A random value in between, and a random value,
Figure BDA0002600756930000065
is an animal individual
Figure BDA0002600756930000066
The position increment of the foraging position is updated.
In order to avoid the animal individual from being trapped in the local optimal solution and continuously search the potential optimal solution, the optimal foraging algorithm enlarges the search space by recruiting other individuals so as to solve the problem that the animal individual is trapped in the local optimal solution, and the ith foraging position increment of the jth animal individual in the tth search
Figure BDA0002600756930000067
Comprises the following steps:
Figure BDA0002600756930000068
in the formula (3), the reaction mixture is,
Figure BDA0002600756930000069
and
Figure BDA00026007569300000610
respectively at the ith searching time, the ith foraging position where the b-th animal individual is located, the ith foraging position where the jth animal individual is located and the worst foraging position in the animal individual in the current searching;
Figure BDA00026007569300000611
and
Figure BDA00026007569300000612
are respectively as
Figure BDA00026007569300000613
And
Figure BDA00026007569300000614
the corresponding objective function value; the new foraging position of the individual animal can appear at the current individual animal position
Figure BDA00026007569300000615
At an arbitrary position in the vicinity, i.e.
Figure BDA00026007569300000616
To appear in
Figure BDA00026007569300000617
Two random numbers r in the formula (2) at arbitrary positions in the vicinity1iAnd r2iCan make it possible to
Figure BDA00026007569300000618
Plus, minus, or not plus or minus position increments
Figure BDA00026007569300000619
When r is1i＞r2iThen, then
Figure BDA00026007569300000620
When r is1i＜r2iThen, then
Figure BDA00026007569300000621
When r is1i＝r2iThen, then
Figure BDA00026007569300000622
Combining the formula (3) with the formula (2) to obtain the result that the ith foraging position of the jth animal individual in the tth search is updated to be
Figure BDA00026007569300000623
In the improved optimal foraging algorithm, a new location for individual animal j may be calculated by equation (4), by which it may be determined when and how the individual animal leaves the current location, except for the individual animal being in the optimal location, the remaining individual animals tend to find a better foraging location.
And 3.2, sequencing the foraging position and the objective function value, and recording the current optimal foraging position and the optimal objective function value.
In this step, the foraging positions and the objective function values of the individual animals obtained in step 3.1 are sorted, and the current optimal foraging position and the optimal objective function value can be obtained through sorting.
And 3.3, calculating a new foraging position of the individual animal, updating the foraging position of the individual animal, and calculating a new objective function value.
After steps 3.1 and 3.2 are completed, updating of foraging positions of all animal individuals in the animal population is completed, so that the animal population is to be subjected to next search, the improved optimal foraging algorithm judges whether the updated position is better than the original position according to the objective function value of the position of the animal individual at the time, and determines whether the updated position is used in the following search process, which can be described as follows:
Figure BDA0002600756930000071
in the formula (5), the reaction mixture is,
Figure BDA0002600756930000072
is [0,1 ]]The random number of (2).
And 3.4, introducing the Cauchy variation to the position of the individual animal.
The addition of the Cauchy variation can enable the current animal individual to generate larger disturbance near the foraging position, the variation is towards a wider range during searching, the searching foraging position is enabled to obtain larger randomness, the searching capability of an optimal foraging algorithm is improved, local optima is timely jumped out, and the principle is described as
Figure BDA0002600756930000073
In the formula (6), the reaction mixture is,
Figure BDA0002600756930000074
is a vector linearly decreasing from 1 to 0, and gamma is uniformly distributed in [0,1 ]]A random value in between.
And 3.5, introducing a differential mutation strategy to increase the diversity of the population.
The method comprises the steps of adopting a DE/rand/1 strategy in a differential variation strategy to perform variation on vectors of population individuals, adding the differential variation strategy in the later stage of each search, and describing as
Figure BDA0002600756930000078
In the formula (7), p1≠p2≠p3，
Figure BDA0002600756930000075
Is a difference vector, F ∈ [0.1,0.9 ∈ [ ]]As a scaling factor, hi,tObtaining the variation vector of the ith position in the t-th search, and performing cross operation as
Figure BDA0002600756930000076
In the formula (8), vi,tA cross variable for the ith search location; j0 is a random value in the dimension, each crossover operation only involves one dimension of the individual, pCR ∈ [0,1 ∈]Is the cross probability;
then, selecting operation is carried out, the optimal vector of the objective function value is reserved as the next generation of individuals, and the selecting operation is expressed as
Figure BDA0002600756930000077
Continuously searching and updating foraging positions in population individuals according to the formula, and judging whether the maximum searching times is reached; if the maximum search times are reached, acquiring an optimal punishment factor C in the support vector machine model and an optimal parameter g of a kernel function in the support vector machine model; if the maximum number of searches has not been reached, then the process returns to and continues to execute step 3.2.
And 4, inputting the optimized optimal parameters into the support vector machine model, and training the support vector machine model optimized by the improved optimal foraging algorithm. The specific implementation method comprises the following steps:
and (3) recording the optimal punishment factor C in the support vector machine model obtained in the step (3) and the optimal parameter g of the kernel function in the support vector machine model, inputting the optimal punishment factor C and the optimal parameter g of the kernel function in the support vector machine model into the support vector machine model to form the support vector machine optimized by the improved optimal foraging algorithm, and training the support vector machine model by using the optimal punishment factor C in the optimized support vector machine model and the optimal parameter g of the kernel function in the support vector machine.
And 5, inputting the prediction data into a support vector machine model optimized by the improved optimal foraging algorithm, obtaining a prediction result through operation, and performing reverse normalization on the prediction result.
In this step, in order to better evaluate the wind power prediction result, a Root Mean Square Error (RMSE) is selected as an objective function of the support vector machine model, which can be described as:
Figure BDA0002600756930000081
in the above formula, N is the number of wind power output power in the data to be predicted, PiIs the true value of the wind power output power, YiIs the predicted value of the output power.
Through the steps, the function of predicting the short-term wind power generation output power is completed.
In the following, in order to verify the quality of the support vector machine model optimized by the improved optimal foraging algorithm, an average Absolute Error (MAE), a standard Absolute value average Error (NMAE), and a standard Mean Square Error (NRMSE) are selected as Error magnitude rating indexes for measuring the prediction model result, and are respectively described as:
mean Absolute Error (Mean Absolute Error, MAE)
Figure BDA0002600756930000082
Mean Error of Absolute standard (Normalized Mean Absolute Error, NMAE)
Figure BDA0002600756930000083
Standard Mean Square Error (Normalized Root Mean Square Error, NRMSE)
Figure BDA0002600756930000084
In the above formula, PNThe rated power of the generator set.
Under the same conditions and parameters, indexes MAE, NMAE and NRMSE are selected to evaluate a support vector machine optimized by an improved optimal foraging algorithm, in order to better show the performance of the support vector machine optimized by the improved optimal foraging algorithm, a support vector machine optimized by the unmodified optimal foraging algorithm and a support vector machine optimized by the improved optimal foraging algorithm are selected in the embodiment to be compared, a comparison graph of the results of outputting the support vector machine optimized by the unmodified optimal foraging algorithm and the support vector machine optimized by the improved optimal foraging algorithm is displayed on a display screen of a computer, as shown in FIGS. 2 and 3, the comparison of prediction results shows that the MAE value of the support vector machine optimized by the improved optimal foraging algorithm is 2.56%, the MAE value of the support vector machine optimized by the unmodified optimal foraging algorithm is reduced by 0.49%, the NMAE value of the support vector machine optimized by the improved optimal foraging algorithm is 3.83%, compared with the NMAE value of the support vector machine model optimized by the optimal foraging algorithm which is not improved by 0.53%, the NRMSE value of the support vector machine model optimized by the optimal foraging algorithm which is improved by 3.31% and the NRMSE value of the support vector machine model optimized by the optimal foraging algorithm which is not improved by 0.50%, the data show that the deviation between the predicted value and the real value of the support vector machine model optimized by the optimal foraging algorithm which is improved by the data is smaller, the fitting effect of the predicted value and the real value is better, and the data show that the prediction accuracy and the accuracy of the short-term wind power generation output power are really improved by the support vector machine model optimized by the optimal foraging algorithm, so that the method has important significance for economic dispatching of a power grid and stable operation of the power grid.
In all the above embodiments, the optimal foraging algorithm and the support vector machine model are the prior art and are well known to those skilled in the art; the method of inputting the acquired training data and test data of the wind power generation output power into the computer is a well-known method; the computer, display and MATLAB computer software were all commercially available.
It should be emphasized that the embodiments described herein are illustrative rather than restrictive, and thus the present invention is not limited to the embodiments described in the detailed description, but other embodiments derived from the technical solutions of the present invention by those skilled in the art are also within the scope of the present invention.

Claims (2)

1. A method for predicting short-term wind power generation output power is characterized by comprising the following steps: the method comprises the following steps:
step 1, acquiring input data and output data of wind power generation, and carrying out normalization processing on the data;
step 2, setting parameters of an improved optimal foraging algorithm and a support vector machine model;
step 3, running an improved optimal foraging algorithm to obtain an optimal penalty factor in the support vector machine model and optimal parameters of a kernel function in the support vector machine model;
step 4, bringing the optimized optimal parameters into a support vector machine model, and training the support vector machine model optimized by the improved optimal foraging algorithm;
step 5, inputting the prediction data into a support vector machine model optimized by an improved optimal foraging algorithm to obtain a prediction result, and performing reverse normalization on the prediction result;
the input data of the wind power generation in the step 1 are meteorological parameters including wind speed and wind direction, the output data refer to wind power generation output power, the input data and the output data are divided into training data and prediction data, the training data comprise the wind power generation output power, the wind speed and the wind direction sine and cosine values, and the prediction data comprise the wind speed and the wind direction sine and cosine values;
the parameters set in the step 2 comprise: improving the population quantity in the optimal foraging algorithm, improving the maximum search times of the optimal foraging algorithm, supporting the search range of a penalty factor C in a vector machine model, supporting the range of a kernel function parameter g of the vector machine model and improving the population dimension in the optimal foraging algorithm;
the specific implementation method of the step 3 comprises the following steps:
step 3.1, initializing foraging positions of the improved optimal foraging algorithm population, and calculating an objective function value of each animal individual;
step 3.2, sequencing the foraging position and the objective function value, and recording the current optimal foraging position and the optimal objective function value;
3.3, calculating a new foraging position of the individual animal, updating the foraging position of the individual animal, and calculating a new objective function value;
3.4, introducing the Cauchy variation to the position of the animal individual for updating;
3.5, introducing a differential variation strategy, continuously searching and updating foraging positions in population individuals, and judging whether the maximum searching times is reached; if the maximum search times are reached, acquiring an optimal punishment factor C in the support vector machine model and an optimal parameter g of a kernel function in the support vector machine model; if the maximum searching times is not reached, returning and continuing to execute the step 3.2;
the specific implementation method of the step 3.1 is as follows:
in the improved optimal foraging algorithm, the foraging position of an individual animal is a d-dimensional vector [ x ]1,…xi,…xd]T，xi∈[xL,xU]，xLAnd xUAre respectively a variable xiAnd (3) searching the optimal position near the foraging position of the current improved optimal foraging algorithm on the upper and lower boundaries:
Figure FDA0003462013650000011
in the above formula, the first and second carbon atoms are,
Figure FDA0003462013650000012
the foraging position of the ith animal individual after the t search,
Figure FDA0003462013650000013
is an animal individual
Figure FDA0003462013650000014
New foraging position after updating, k is a scale factor, r1iAnd r2iIs uniformly distributed in [0,1 ]]A random value in between, and a random value,
Figure FDA0003462013650000015
is an animal individual
Figure FDA0003462013650000016
Updating the position increment of the foraging position;
increasing ith foraging position of jth animal individual in tth search
Figure FDA0003462013650000017
Comprises the following steps:
Figure FDA0003462013650000018
in the above formula, the first and second carbon atoms are,
Figure FDA0003462013650000021
and
Figure FDA0003462013650000022
respectively at the ith searching time, the ith foraging position where the b-th animal individual is located, the ith foraging position where the jth animal individual is located and the worst foraging position in the animal individual in the current searching;
Figure FDA0003462013650000023
and
Figure FDA0003462013650000024
are respectively as
Figure FDA0003462013650000025
And
Figure FDA0003462013650000026
the corresponding objective function value; the new foraging position of the individual animal can appear at the current individual animal position
Figure FDA0003462013650000027
At an arbitrary position in the vicinity, i.e.
Figure FDA0003462013650000028
To appear in
Figure FDA0003462013650000029
At an arbitrary position nearby, two random numbers r1iAnd r2iCan make it possible to
Figure FDA00034620136500000210
Plus, minus, or not plus or minus position increments
Figure FDA00034620136500000211
When r is1i＞r2iThen, then
Figure FDA00034620136500000212
When r is1i＜r2iThen, then
Figure FDA00034620136500000213
When r is1i＝r2iThen, then
Figure FDA00034620136500000214
Combining the two formulas to obtain the ith foraging position update of the jth individual animal in the tth search as:
Figure FDA00034620136500000215
the specific implementation method of the step 3.3 comprises the following steps: and the improved optimal foraging algorithm judges whether the updated position is better than the original position according to the objective function value of the individual position of the animal at the moment, and determines whether the updated position is used in the subsequent searching process, wherein the objective function is described as follows:
Figure FDA00034620136500000216
in the above formula, the first and second carbon atoms are,
Figure FDA00034620136500000217
is [0,1 ]]The random number of (2);
the step 3.4 is realized by adopting the following algorithm:
Figure FDA00034620136500000218
in the above formula, the first and second carbon atoms are,
Figure FDA00034620136500000219
is a vector linearly decreasing from 1 to 0, and gamma is uniformly distributed in [0,1 ]]A random value in between;
the specific implementation method of the step 3.5 is as follows:
the method comprises the following steps of carrying out mutation on vectors of population individuals by adopting a DE/rand/1 strategy in a differential mutation strategy, adding the differential mutation strategy at the later stage of each search, and describing as follows:
Figure FDA00034620136500000220
in the above formula, p1≠p2≠p3，
Figure FDA00034620136500000221
Is a difference vector, F ∈ [0.1,0.9 ∈ [ ]]As a scaling factor, hi,tObtaining a variation vector for the ith position in the t-th search, and then performing a crossover operation by:
Figure FDA00034620136500000222
in the above formula, vi,tA cross variable for the ith search location; j0 is a random value in the dimension, each crossover operation only involves one dimension of the individual, pCR ∈ [0,1 ∈]Is the cross probability;
and carrying out selection operation, reserving the optimal vector of the objective function value as a next generation individual, and expressing the selection operation as follows:
Figure FDA0003462013650000031
continuously searching and updating foraging positions in population individuals according to the formula, and judging whether the maximum searching times is reached; if the maximum search times are reached, acquiring an optimal punishment factor C in the support vector machine model and an optimal parameter g of a kernel function in the support vector machine model; if the maximum searching times is not reached, returning and continuing to execute the step 3.2;
the specific implementation method of the step 4 comprises the following steps: inputting the optimal punishment factor C in the support vector machine model obtained in the step 3 and the optimal parameter g of the kernel function in the support vector machine model into the support vector machine model to form the support vector machine model optimized by the improved optimal foraging algorithm, and training the support vector machine model by using the optimal punishment factor C in the optimized support vector machine model and the optimal parameter g of the kernel function in the support vector machine model;
the method for performing the inverse normalization processing in the step 5 comprises the following steps: selecting the root mean square error as an objective function RMSE of the support vector machine model, wherein the objective function is expressed as follows:
Figure FDA0003462013650000032
in the above formula, N is the number of wind power output power in the data to be predicted, PiIs the true value of the wind power output power, YiIs the predicted value of the output power.
2. A method of short term wind power generation output power prediction as claimed in claim 1, characterized by: the wind power generation output power is the wind power generation output power of the wind power plant.
CN202010723186.9A 2020-07-24 2020-07-24 Method for predicting short-term wind power generation output power Expired - Fee Related CN111832838B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010723186.9A CN111832838B (en) 2020-07-24 2020-07-24 Method for predicting short-term wind power generation output power

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010723186.9A CN111832838B (en) 2020-07-24 2020-07-24 Method for predicting short-term wind power generation output power

Publications (2)

Publication Number Publication Date
CN111832838A CN111832838A (en) 2020-10-27
CN111832838B true CN111832838B (en) 2022-03-01

Family

ID=72925373

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010723186.9A Expired - Fee Related CN111832838B (en) 2020-07-24 2020-07-24 Method for predicting short-term wind power generation output power

Country Status (1)

Country Link
CN (1) CN111832838B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113159437B (en) * 2021-04-30 2022-03-25 河北工业大学 Method for predicting short-term photovoltaic power generation output power
CN113139346A (en) * 2021-05-06 2021-07-20 河北工业大学 IOFA-SVM-based short-term wind power prediction method
CN113205228B (en) * 2021-05-28 2022-10-14 河北工业大学 Method for predicting short-term wind power generation output power
CN113570555B (en) * 2021-07-07 2024-02-09 温州大学 Two-dimensional segmentation method of multi-threshold medical image based on improved grasshopper algorithm

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106527176A (en) * 2016-10-21 2017-03-22 哈尔滨理工大学 MFOA (modified fruit fly optimization algorithm)-SVM (support vector machine)-based boiler thermal efficiency and NOX modeling method
CN107330781A (en) * 2017-06-19 2017-11-07 南京信息工程大学 A kind of individual credit risk appraisal procedure based on IFOA SVM
CN108074014A (en) * 2017-12-13 2018-05-25 宁波市镇海区气象局 A kind of method for the phase fitful wind that detects a typhoon
CN108229401A (en) * 2018-01-05 2018-06-29 杭州电子科技大学 A kind of multi-modal Modulation recognition method based on AFSA-SVM
CN109002918A (en) * 2018-07-16 2018-12-14 国网浙江省电力有限公司经济技术研究院 Based on drosophila optimization algorithm-support vector machines electricity sales amount prediction technique
CN109062180A (en) * 2018-07-25 2018-12-21 国网江苏省电力有限公司检修分公司 A kind of oil-immersed electric reactor method for diagnosing faults based on IFOA optimization SVM model
CN110705681A (en) * 2019-09-29 2020-01-17 北京工商大学 Support vector machine parameter optimization method and system

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106527176A (en) * 2016-10-21 2017-03-22 哈尔滨理工大学 MFOA (modified fruit fly optimization algorithm)-SVM (support vector machine)-based boiler thermal efficiency and NOX modeling method
CN107330781A (en) * 2017-06-19 2017-11-07 南京信息工程大学 A kind of individual credit risk appraisal procedure based on IFOA SVM
CN108074014A (en) * 2017-12-13 2018-05-25 宁波市镇海区气象局 A kind of method for the phase fitful wind that detects a typhoon
CN108229401A (en) * 2018-01-05 2018-06-29 杭州电子科技大学 A kind of multi-modal Modulation recognition method based on AFSA-SVM
CN109002918A (en) * 2018-07-16 2018-12-14 国网浙江省电力有限公司经济技术研究院 Based on drosophila optimization algorithm-support vector machines electricity sales amount prediction technique
CN109062180A (en) * 2018-07-25 2018-12-21 国网江苏省电力有限公司检修分公司 A kind of oil-immersed electric reactor method for diagnosing faults based on IFOA optimization SVM model
CN110705681A (en) * 2019-09-29 2020-01-17 北京工商大学 Support vector machine parameter optimization method and system

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
"优化觅食算法改进支持向量机的僵尸网络检测模型研究";朱艳;《中国优秀硕士学位论文全文数据库 信息科技辑》;20181115(第11期);论文全文 *
"基于果绳优化算法的支持向量机故障诊断";张翔 等;《电子设计工程》;20130831;第21卷(第16期);90-93 *

Also Published As

Publication number Publication date
CN111832838A (en) 2020-10-27

Similar Documents

Publication Publication Date Title
CN111832838B (en) Method for predicting short-term wind power generation output power
Wu et al. A data mining approach combining $ K $-means clustering with bagging neural network for short-term wind power forecasting
Ye et al. Combined approach for short-term wind power forecasting based on wave division and Seq2Seq model using deep learning
CN110866633B (en) Micro-grid ultra-short-term load prediction method based on SVR support vector regression
CN114757427B (en) Autoregressive-corrected LSTM intelligent wind power plant ultra-short-term power prediction method
CN111832812A (en) Wind power short-term prediction method based on deep learning
Banik et al. Wind power generation probabilistic modeling using ensemble learning techniques
CN110991737A (en) Ultra-short-term wind power prediction method based on deep belief network
CN112580874A (en) Short-term wind power prediction method based on random forest algorithm and TCN
CN117132132A (en) Photovoltaic power generation power prediction method based on meteorological data
CN112651576A (en) Long-term wind power prediction method and device
CN113052389A (en) Distributed photovoltaic power station ultra-short-term power prediction method and system based on multiple tasks
CN116629416A (en) Photovoltaic power station power prediction method and device
Amer et al. Solar power prediction based on Artificial Neural Network guided by feature selection for Large-scale Solar Photovoltaic Plant
CN117117968A (en) Wind farm climbing event prediction method based on data enhancement
CN114298441A (en) Photovoltaic power prediction method and system
CN114021818A (en) Wind power multistep prediction method considering space-time distribution characteristics
Kader et al. Active power control strategy for wind farms based on power prediction errors distribution considering regional data
CN116167508B (en) Short-term photovoltaic output rapid prediction method and system based on meteorological factor decomposition
Xu et al. NWP feature selection and GCN-based ultra-short-term wind farm cluster power forecasting method
CN116128211A (en) Wind-light-water combined short-term optimization scheduling method based on wind-light uncertainty prediction scene
CN113159437B (en) Method for predicting short-term photovoltaic power generation output power
CN113780643A (en) Photovoltaic power station short-term output prediction method based on case reasoning
Sun et al. Design of a wind power forecasting system based on deep learning
Li et al. Ultra-short-term wind power forecasting based on the strategy of “dynamic matching and online modeling”

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
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20220301

CF01 Termination of patent right due to non-payment of annual fee