CN113205228B - 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
CN113205228B
CN113205228B CN202110592303.7A CN202110592303A CN113205228B CN 113205228 B CN113205228 B CN 113205228B CN 202110592303 A CN202110592303 A CN 202110592303A CN 113205228 B CN113205228 B CN 113205228B
Authority
CN
China
Prior art keywords
butterfly
value
wind power
vector machine
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
CN202110592303.7A
Other languages
Chinese (zh)
Other versions
CN113205228A (en
Inventor
李玲玲
宁楠
王海宇
任琦瑛
任心雨
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 CN202110592303.7A priority Critical patent/CN113205228B/en
Publication of CN113205228A publication Critical patent/CN113205228A/en
Application granted granted Critical
Publication of CN113205228B publication Critical patent/CN113205228B/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
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • 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)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Human Resources & Organizations (AREA)
  • General Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Marketing (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computing Systems (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Biomedical Technology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Evolutionary Biology (AREA)
  • Game Theory and Decision Science (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Medical Informatics (AREA)
  • Development Economics (AREA)
  • Molecular Biology (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • Primary Health Care (AREA)
  • Wind Motors (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (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 of: dividing wind power generation data into training data and prediction data, and carrying out normalization processing on the data; setting and initializing parameters of an improved butterfly optimization algorithm and a support vector machine model; training a support vector machine model using the training data; calculating a butterfly position with an optimal fitness value by using an improved butterfly optimization algorithm; substituting the butterfly position with the optimal fitness value into a support vector machine model, and predicting the wind power generation output power by using the support vector machine model; and outputting a prediction result. According to the method, the output power of the short-term wind power generation is predicted by adopting an improved butterfly optimization algorithm-support vector machine model, so that the accuracy of predicting the output power of the short-term wind power generation with randomness and volatility can be improved, the utilization rate of new energy power generation is further improved, and the stability of a power grid is improved.

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
The demand of economic development on energy is increasing day by day, which leads to the increasingly outstanding contradiction between resource exhaustion, environmental pollution and climate change and practical requirements, so that the development and utilization of wind energy are greatly concerned by various countries. Because wind power resources have the outstanding advantages of zero pollution and low cost, the large-scale development and utilization of wind energy can greatly relieve the problems of environmental pollution and safe power supply to power loads, but wind power output power is intermittent and random due to the characteristics of wind energy, wind power integration can influence the voltage and frequency of a power grid, and can also make a reasonable scheduling plan for safe operation of the power grid and an electric power department to cause huge influence, so that the accurate prediction of the wind power output power has important significance for promoting the development and utilization of clean energy and the safe and economic operation of the power grid.
In terms of the time scale of prediction, wind power prediction can be mainly divided into long-term prediction, short-term prediction and ultra-short-term prediction. The long-term wind power prediction refers to the prediction of wind energy over 10 days in the future, can be divided into monthly, seasonal and annual predictions and the like, and is generally used for providing a reference basis for the formulation of a long-term production plan of a wind power enterprise; the ultra-short term prediction has no specific standard, and the wind power prediction within 30min can be regarded as the ultra-short term prediction generally, and is mainly used for generator control and supplement and correction of the short-term wind power prediction. By researching and accumulating daily change and hour change rules of the output power of the wind power plant, daily operation modes are formulated for power departments, and power prediction of 1h-2h in the future is provided when wind power is in grid-connected operation.
The short-term prediction refers to the prediction of the wind power within 3 days (72 hours) in the future, and can make a reasonable scheduling plan for the power department and provide scientific guidance and technical support for the safety production of the wind power plant. The support vector machine obtains good effect by means of unique advantages of the support vector machine in solving nonlinear, small sample and high-dimensional pattern learning, and the basic idea of the support vector machine is to transform low-dimensional input space nonlinearity into a high-dimensional space by introducing an inner product kernel function, and then find the nonlinear relation between an input variable and an output variable from the high-dimensional space, so that the regression problem of small sample, nonlinearity and high-dimensional can be solved. How to accurately predict the short-term wind power generation output power by the support vector machine 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 provide a method for predicting the output power of short-term wind power generation.
The invention solves the technical problems in the prior art by adopting the following technical scheme:
a method for predicting short-term wind power generation output power comprises the following steps:
step 1, dividing wind power generation data into training data and prediction data, and carrying out normalization processing on the data;
step 2, setting and initializing parameters of an improved butterfly optimization algorithm and a support vector machine model;
step 3, training a support vector machine model by using training data;
step 4, calculating the butterfly position with the optimal fitness value by using an improved butterfly optimization algorithm;
step 5, substituting the butterfly position with the optimal fitness value into a support vector machine model, and predicting the output power of the wind power generation by using the support vector machine model;
and 6, outputting a prediction result.
Further, the training data and the prediction data both comprise input data and output data, the input data comprises wind speed and wind direction, and the output data is the output power of wind power generation; the data normalization processing method comprises the following steps:
Figure BDA0003089707100000021
in the formula, P scale,i Representing a normalized data value, P, of the wind power output i Representing the true value, P, of the wind power output min Representing the minimum value, P, of the wind power output max The maximum value of the wind power output is shown.
Further, the specific implementation method of step 2 is as follows: setting the iteration times of an improved butterfly optimization algorithm, the quantity of butterflies in the population and the dimensionality of the population, and initializing the positions of the butterflies; the support vector machine model sets the range of a penalty factor C and a Gaussian kernel function key parameter eta; the range of the butterfly position in the butterfly optimization algorithm is improved to be the range of a penalty factor C of a support vector machine model and a key parameter eta of a Gaussian kernel function, and the Gaussian kernel function is expressed as follows;
Figure BDA0003089707100000022
in the formula, K (x) i ,x j ) Representing a Gaussian kernel function, and eta represents a key parameter of the Gaussian kernel function;
the method for initializing the butterfly position comprises the following steps:
Figure BDA0003089707100000023
in the formula (I), the compound is shown in the specification,
Figure BDA0003089707100000024
indicating the initial position of each butterfly, rand denotes a random number belonging between 0 and 1,
Figure BDA0003089707100000025
represents a vector formed by a penalty factor C and an upper bound of a key parameter eta of a Gaussian kernel function in a support vector machine model,
Figure BDA0003089707100000026
and representing a vector consisting of a penalty factor C and a lower bound of a key parameter eta of the Gaussian kernel function in the support vector machine model.
Further, the specific implementation method of step 4 is as follows:
step 4.1, calculating the fitness value of each butterfly in the butterfly optimization algorithm and the fragrance intensity generated by the butterfly, and recording the optimal fitness value and the corresponding butterfly position in the butterfly optimization algorithm;
step 4.2, updating the position of the butterfly;
step 4.3, calculating the fitness value of each butterfly again, comparing the optimal fitness value bF obtained by calculation with the optimal fitness value obtained by calculation in the step 4.1, assigning the minimum value to the global optimal fitness value BF, and recording the position information of the butterfly;
step 4.4, executing a Levy flight strategy;
and 4.5, updating the population individuals according to the fitness value until the set maximum iteration times are reached, and obtaining the butterfly position with the optimal fitness value.
Further, in the step 4.1, a fitness function calculation formula is used to calculate the fitness value of each butterfly, and the fitness function selects a standard mean square error NRMSE, which is expressed as follows:
Figure BDA0003089707100000027
in the formula, P N The rated power of the generator set, N is the number of wind power generation output power in the data to be predicted, P i Is the true value, Y, of the wind power generation output power in the output data i Predicting values of a group of wind power generation output powers output by the support vector machine model in the step 3;
the NRMSE calculation result is a fitness value, the current fitness values S (i) are arranged from large to small, the optimal fitness value is recorded as bF, and the position of the butterfly individual is stored;
the position update of the butterfly is related to the fragrance intensity generated by the butterfly, and the calculation formula of the fragrance intensity is as follows:
f＝bI α
where f is the perceived intensity of the fragrance, b is the sensory modality of the butterfly, I is the stimulus intensity, and α is a modality-dependent power exponent.
Further, the specific implementation method of the step 4.2 is as follows:
first, a switching probability is generated according to a dynamic switching probability formula as follows:
θ＝θ max -(θ max -θ min )×(T max -t)/T max
where θ is the probability of switching, θ max To switch the probability maximum, θ min For minimum value of switching probability, T max The maximum iteration number of the algorithm is represented, and t is the current iteration number of the algorithm;
then, comparing the switching probability theta with a generated random number r with the value range of 0 to 1, and if the value of the random number r is greater than that of the switching probability theta, performing global position updating according to the following formula:
Figure BDA0003089707100000031
in the formula (I), the compound is shown in the specification,
Figure BDA0003089707100000032
is the position of the ith butterfly in the t iteration, g * Represents the optimal solution among all solutions of the current iteration, f i Represents the amount of fragrance emitted by the ith butterfly, and w is a weighting factor; the formula for the weight factor w is:
Figure BDA0003089707100000033
in the formula, T max The maximum iteration number of the algorithm is represented, and t is the current iteration number of the algorithm;
if the value of the random number r is smaller than the value of the switching probability θ, local location update is performed according to the following formula:
Figure BDA0003089707100000034
in the formula (I), the compound is shown in the specification,
Figure BDA0003089707100000035
and
Figure BDA0003089707100000036
indicating the position of the jth and kth butterflies in the tth iterative solution space.
Further, the specific implementation method of step 4.4 is as follows:
the flight step length of the Levy flight strategy is set as follows:
Figure BDA0003089707100000037
wherein s is the step length of Levy flight, levy (beta), and u and v are both subject to normal distribution, wherein:
Figure BDA0003089707100000038
combining the Levy flight step length with the optimal butterfly position obtained in the step 4.3, wherein the updated new search position is as follows:
Figure BDA0003089707100000041
in the formula (I), the compound is shown in the specification,
Figure BDA0003089707100000042
for the position of the optimal butterfly, dim is the dimension of the population,
Figure BDA0003089707100000043
the position of the optimal butterfly after the Levy flight;
will be provided with
Figure BDA0003089707100000044
Inputting the position information into a support vector machine model, calculating the fitness value again after the step 3 is executed, recording the optimal fitness value, and keeping the position information of the butterfly; and (4) comparing the optimal fitness value with the BF value obtained in the step (4.3), if the optimal fitness value is smaller than the BF value, updating the BF to the current optimal fitness value, storing the position information of the butterfly with the current optimal fitness value, and otherwise, keeping the original BF value and the butterfly position information unchanged.
Further, the specific implementation method of step 5 is as follows: the position of a butterfly with the optimal fitness value corresponds to a penalty factor C and a Gaussian kernel function key parameter eta of the support vector machine model, and the penalty factor C and the optimal value of the Gaussian kernel function key parameter eta of the support vector machine model at the moment are brought into the support vector machine model; and taking the wind speed and the wind direction in the prediction data as the input of the support vector machine model, taking the output power of wind power generation in the prediction data as the output of the support vector machine model, and outputting a group of wind power generation output power prediction results.
Further, when the predicted result is output in step 6, the predicted result needs to be subjected to inverse normalization processing, where the inverse normalization processing adopts the following formula:
P i ＝P scale,i ×(P max -P min )+P min
in the formula, P i Representing true value, P, of wind power output min Represents the minimum value, P, of the wind power output max Representing the maximum value, P, of the wind power output scale,i And the data value after the output power of the wind power generation is normalized is represented.
The invention has the advantages and positive effects that:
1. according to the method, the output power of the short-term wind power generation is predicted by adopting an improved butterfly optimization algorithm-support vector machine model (IBOA-SVM), the accuracy of prediction of the output power of the short-term wind power generation with randomness and volatility can be improved, the utilization rate of new energy power generation is further improved, the stability of a power grid is improved, the problem of impact on safe and stable operation of the power grid when the wind power generation with volatility and randomness is connected into the power grid is solved, and the method has important significance on safe connection of a wind power generation system into the power grid and economic operation of a power system.
2. The improved butterfly optimization algorithm used by the invention improves the position updating formula of the butterfly individual and the judging mode of global search or local search, thereby improving the searching capability of the algorithm.
Drawings
FIG. 1 is a flow chart of short term wind power generation output power prediction of the present invention;
FIG. 2 is a comparison of predicted results obtained using different prediction models.
Detailed Description
The embodiments of the present invention will be described in detail with reference to the accompanying drawings.
A method for predicting the output power of short-term wind power generation is disclosed, as shown in FIG. 1, and comprises the following steps:
step 1, dividing wind power generation data into training data and prediction data, and carrying out normalization processing on the data.
In this step, the training data and the prediction data of the wind power generation output power prediction both include input data including wind speed and wind direction and output data being output power of the wind power generation. And normalizing the data using equation (1):
Figure BDA0003089707100000051
p in formula (1) scale,i Representing a normalized data value, P, of the wind power output i Representing the true value, P, of the wind power output min Representing the minimum value, P, of the wind power output max Represents the maximum value of the wind power output.
And 2, setting and initializing parameters of the butterfly optimization algorithm-support vector machine model.
In this step, the iteration times of the improved butterfly optimization algorithm, the number of butterflies in the population and the dimensionality of the population need to be set, and the butterfly positions are initialized; the support vector machine model sets the range of a penalty factor C and a Gaussian kernel function key parameter eta, namely an upper bound and a lower bound of two parameters to be optimized; the range of the butterfly position in the improved butterfly optimization algorithm is the range of a penalty factor C of the support vector machine model and a key parameter eta of a Gaussian kernel function;
Figure BDA0003089707100000052
k (x) in the formula (2) i ,x j ) Representing a Gaussian kernel function, wherein eta represents the width of the kernel function and is also a key parameter of the Gaussian kernel function;
the initialized position formula of the butterfly position is shown as formula (3):
Figure BDA0003089707100000053
in the formula (3), the reaction mixture is,
Figure BDA0003089707100000054
indicating the initial position of each butterfly, rand denotes a random number belonging between 0 and 1,
Figure BDA0003089707100000055
representing a vector consisting of a penalty factor C and an upper bound of a gaussian kernel function key parameter η in a support vector machine model,
Figure BDA0003089707100000056
and representing a vector formed by a penalty factor C and a lower bound of a key parameter eta of the Gaussian kernel function in the support vector machine model.
And 3, training a support vector machine model by using the training data.
In the step, wind power and wind direction in training data are used as input of a support vector machine model, and output power of wind power generation in the training data is used as output of the support vector machine model and used for training the support vector machine model and outputting a group of wind power generation power prediction results; the butterfly position initialized or updated in the butterfly optimization algorithm is a vector formed by a penalty factor C supporting a vector machine model and a Gaussian kernel function key parameter eta.
Step 4, calculating the butterfly position with the optimal fitness value by using an improved butterfly optimization algorithm, and comprising the following steps:
step 4.1, calculating the fitness value of each butterfly in the butterfly optimization algorithm and the fragrance intensity generated by the butterfly, and recording the optimal fitness value and the corresponding butterfly position in the butterfly optimization algorithm;
in this step, a fitness function calculation formula is used to calculate the fitness value of each butterfly, and the fitness function selects a standard mean square error NRMSE, whose expression is as follows:
Figure BDA0003089707100000057
in the formula (4), P N The rated power of the generator set, N is the number of wind power generation output power in the data to be predicted, P i For the true value, Y, of the wind power output in the output data i Predicting the output power of the group of wind power generation output in the step 3; the NRMSE calculation result is a fitness value, the smaller the fitness value is, the better the fitness value is, the current fitness value S (i) is arranged from large to small, wherein the optimal fitness value is recorded as bF, and the position of the butterfly individual is stored.
Since in the improved butterfly optimization algorithm, the position update of each butterfly is related to the intensity of the fragrance generated by it, the intensity of the fragrance of each butterfly needs to be calculated, and the calculation formula of the intensity of the fragrance is shown in formula (5):
f＝bI α (5)
in equation (5), f is the perceived intensity of the fragrance, i.e., the intensity at which the fragrance can be perceived by other butterflies, b is the sensory modality of the butterfly, I is the stimulus intensity, the value of which is related to the fitness value of the butterfly, which in the present invention is otherwise equal to the fitness value, and α is a modal-dependent power index, which accounts for the absorption of different degrees of fragrance.
Step 4.2, updating the position of the butterfly;
the updating of the butterfly position is carried out according to the principle of improving the butterfly optimization algorithm, the fragrance emitted by each butterfly is calculated according to a formula (5), the butterfly emitting the most fragrance is recorded, the position of the butterfly is stored, and the butterfly position is updated according to the following process;
first, the switching probability is generated according to the dynamic switching probability formula of formula (6), which is expressed as follows:
θ＝θ max -(θ max -θ min )×(T max -t)/T max (6)
in the formula (6), θ is the switching probability, θ max For maximum value of switching probability, theta min For minimum value of switching probability, T max The maximum iteration number of the algorithm is represented by t, and the current iteration number of the algorithm is represented by t; comparing the switching probability theta calculated according to the formula (6) with a generated random number r with the value range of 0 to 1, wherein the comparison result determines whether the butterfly carries out position updating according to global search or local search, the original global search position updating formula is improved, and the position updating formulas before and after the improvement are respectively shown as a formula (7) and a formula (10):
before the improvement:
Figure BDA0003089707100000061
equation (7) is a global search location update formula before improvement, in which,
Figure BDA0003089707100000062
is the position of the ith butterfly in the t iteration, g * Represents the optimal solution among all solutions of the current iteration, i.e. the position of the butterfly with the minimum fitness value calculated according to equation (4), f i Indicates the amount of fragrance emitted by the ith butterfly, and r is [0,1 ]]A random number in between;
Figure BDA0003089707100000063
equation (8) is a local search location update equation,
Figure BDA0003089707100000064
and
Figure BDA0003089707100000065
representing the positions of j and k butterflies in the t-th iteration solution space, and the definition of the other variables is consistent with that of the formula (7);
after improvement:
in order to further improve the optimizing capability of the algorithm, when the weight factor is larger, the global optimizing capability of the algorithm is stronger, and the search agent is favorable for carrying out global search; when the weight factor is smaller, the local optimization capability of the algorithm is stronger, the fast convergence can be realized, and the result precision is ensured, so the weight factor w is introduced, and the calculation formula is as follows:
Figure BDA0003089707100000066
wherein w is a weight factor, T max The maximum iteration number of the algorithm is represented, and t is the current iteration number of the algorithm;
Figure BDA0003089707100000067
in the formula (I), the compound is shown in the specification,
Figure BDA0003089707100000068
is the position of the ith butterfly in the t iteration, g * Represents the optimal solution among all solutions of the current iteration, i.e. the position of the butterfly with the minimum fitness value calculated according to equation (4), f i Indicates the amount of fragrance emitted by the ith butterfly, and r is [0,1 ]]A random number in between; w is an introduced weight factor;
comparing the switching probability theta calculated by the formula (6) with the size of a random number r generated within the value range of 0-1, and if the value of the random number r is greater than that of the switching probability theta, performing global position updating according to the formula (10); if the value of the random number r is smaller than the value of the switching probability theta, local position updating is carried out according to the formula (8);
step 4.3, calculating the fitness value of each butterfly again, comparing the optimal fitness value bF obtained by calculation with the optimal fitness value obtained by calculation in the step 4.1, assigning the minimum value to the global optimal fitness value BF, and recording the position information of the butterfly;
step 4.4, executing a Levy flight strategy;
because the original butterfly optimization algorithm lacks random factors, the algorithm is easy to fall into local optimum in the searching process, and in order to enhance the randomness of the optimized position of the algorithm, the invention adds a Levy flight strategy to the original butterfly optimization algorithm, and the expression is as follows:
Figure BDA0003089707100000071
where s is the step length of Levy flight, i.e., levy (β), and u and v both follow a normal distribution, where:
Figure BDA0003089707100000072
combining the Levy flight with the optimal butterfly position found in the step 4.3, wherein the updated new search position is as follows:
Figure BDA0003089707100000073
in the formula (I), the compound is shown in the specification,
Figure BDA0003089707100000074
the position of the optimal butterfly recorded in step 4.3, dim is the dimension of the population,
Figure BDA0003089707100000075
the position of the optimal butterfly in the step 4.3 after the Levy flight; will be provided with
Figure BDA0003089707100000076
Inputting the position information into a support vector machine model, calculating a fitness value according to the formula (4) again after the step 3 is executed, recording the optimal fitness value, and keeping the position information of the butterfly; comparing the optimal fitness value with the BF value obtained in the step 4.3, if the optimal fitness value is smaller than the BF value, updating the BF value to the current optimal fitness value, storing the position information of the current optimal fitness value butterfly, and otherwise, keeping the original BF value and the butterfly position information unchanged;
and 4.5, updating the population individuals according to the fitness value until the set maximum iteration times are reached, and obtaining the butterfly position with the optimal fitness value.
In the step, whether the optimization process of the improved butterfly optimization algorithm is finished or not is judged according to the maximum iteration number of the improved butterfly optimization algorithm set in the step 2, and if the iteration number is smaller than the set iteration number, the steps 4.2, 4.3 and 4.4 are carried out again for iteration; if the iteration number at this time is larger than the set iteration number, the butterfly position with the optimal fitness value is obtained, and the step 4.2 is carried out.
And 5, substituting the butterfly position with the optimal fitness value into a support vector machine model, and predicting the output power of the wind power generation by using the support vector machine model.
In this step, the position of the butterfly with the optimal fitness value corresponds to a vector formed by the penalty factor C of the support vector machine model and the gaussian kernel function key parameter η, and the optimal values of the penalty factor C and the gaussian kernel function key parameter η of the support vector machine model at this time are brought into the support vector machine model.
In this step, the wind speed and the wind direction in the prediction data are taken as the input of the support vector machine model, the output power of the wind power generation in the prediction data is taken as the output of the support vector machine model, and a set of wind power generation output power prediction results is output.
And 6, outputting a prediction result, and performing inverse normalization on the prediction result.
In this step, the prediction result obtained in step 9 is subjected to inverse normalization, and the inverse normalization formula is shown as formula (14):
P i ＝P scale,i ×(P max -P min )+P min (14)
the reverse normalization is carried out according to a normalized calculation rule, and an upper and lower bound range which is the same as the data before normalization is obtained after the reverse normalization, so that the comparison of prediction results is facilitated;
fig. 1 is a process implemented by a computer program according to the above-described method.
The present invention will be explained below with reference to a specific example. In this embodiment, a PC is used as a platform for model building, where the CPU is i5-8300H 2.3ghz, the installed memory is 8G, the operating system is Windows 10-64 bits, and the MATLAB R2016b version is used. The specific prediction process is as follows:
in step 1, the training data and the prediction data of the wind power generation output power prediction both comprise input data and output data, wherein the input data comprise wind power and wind direction, and the output data are the output power of the wind power generation; data were normalized using equation (1); in formula (1), P scale,i Representing a normalized data value, P, of the wind power output i Representing true value, P, of wind power output min Representing the minimum value, P, of the wind power output max The maximum value of the wind power output is shown.
During the test, the data are from a La Haute Borne wind power plant in the big east area of northeast of famous France; according to the method, data of a first generator set in the wind power plant in 2017 are predicted, the rated power generation power of the generator is 2050kW, 8778 groups of data in the whole year comprise the output power of a fan and environmental factors such as wind speed, wind direction, temperature, humidity and the like in the same period, the wind power data in spring are selected for prediction, in a prediction experiment, 120 groups of data in 5 consecutive days are selected for model training, 48 groups of data in 2 consecutive days are used as prediction data, and the proportion of training samples to the prediction samples is set as 5: and 2, taking the wind speed and the wind direction as the input of the prediction model, and taking the wind power generation output power as the output of the prediction model.
In step 2, parameters of the improved butterfly optimization algorithm and the support vector machine model are set and initialized. Wherein the number of butterflies of the improved butterfly optimization algorithm is 30, the maximum iteration times of the improved butterfly optimization algorithm is 100, and the search range of the penalty factor C is [0.1,1200]The kernel function parameter η is [0.01,100']The algorithm only optimizes two parameters so that the population dimension is 2, and other parameters of the improved butterfly optimization algorithm and the support vector machine model are default values; the vector formed by the penalty factor C of the support vector machine model and the key parameter eta of the Gaussian kernel function is the position of a butterfly initialized or updated by a butterfly optimization algorithm; the Gaussian kernel function is shown as formula (2); k (x) i ,x j ) The representation of the gaussian kernel function is shown,eta represents the width of the kernel function and is also a key parameter of the Gaussian kernel function; initializing the butterfly position according to an initialization formula (3) of the butterfly position; in the formula (3), the reaction mixture is,
Figure BDA0003089707100000081
indicating the initial position of each butterfly, rand denotes a random number belonging between 0 and 1,
Figure BDA0003089707100000082
representing a vector consisting of a penalty factor C and an upper bound of a gaussian kernel function key parameter η in a support vector machine model,
Figure BDA0003089707100000083
Figure BDA0003089707100000084
representing a vector consisting of a penalty factor C and a lower bound of a gaussian kernel function key parameter η in a support vector machine model,
Figure BDA0003089707100000085
in step 3, the support vector machine model is trained using the training data. The method comprises the steps of taking wind power and wind direction in training data as input of a support vector machine model, taking output power of wind power generation in the training data as output of the support vector machine model, training the support vector machine model, and outputting a group of wind power generation power prediction results; the vector formed by the penalty factor C of the support vector machine model and the key parameter eta of the Gaussian kernel function is the position of the butterfly initialized or updated by the butterfly optimization algorithm.
In step 4, calculating the fitness value of each butterfly by using a fitness function calculation formula, selecting a standard mean square error NRMSE (mean square error) formula (4) as the fitness function, wherein in the formula (4), P N The rated power of the generator set is 2050kW, N is the number of the wind power generation output power in the data to be predicted is 48, P is the number of the wind power generation output power in the data to be predicted i For the true value, Y, of the wind power output in the output data i A predicted value of a group of wind power generation output power output for each iteration;the NRMSE calculation result is a fitness value, the smaller the fitness value is, the better the fitness value is, the fitness value of each butterfly is recorded, and the position of the butterfly with the optimal fitness value is stored.
Because in the improved butterfly optimization algorithm, the position update of each butterfly is related to the strength of the fragrance generated by the butterfly, the strength of the fragrance of each butterfly needs to be calculated, and the calculation formula of the strength of the fragrance is shown as formula (5), wherein f is the perceived strength of the fragrance, and b is the sensory mode of the butterfly, and the value of the sensory mode of the butterfly is 0.01; i is the stimulus intensity, the value of which is related to the fitness value of the butterfly, and in the invention, the value of which is equal to the fitness value; α is a modality-dependent power exponent with a value of 0.1;
in step 4, the location of the butterfly is updated. The butterfly position is updated according to the principle of improving the butterfly optimization algorithm, the fragrance emitted by each butterfly is calculated according to a formula (5), the butterfly emitting the most fragrance is recorded, and the position of the butterfly is stored, and then the butterfly updates the position according to the following process; first, a value of a switching probability θ is calculated as equation (6), where θ max Has a value of 0.9, [ theta ] min Has a value of 0.1,T max The maximum iteration times of the algorithm is 100, and t is the current iteration times of the algorithm; then, comparing the switching probability theta calculated according to the formula (6) with a generated random number r with the value range of 0-1, and if the value of the random number r is greater than that of the switching probability theta, performing global position updating according to the formula (10); if the value of the random number r is smaller than the value of the switching probability θ, local position updating is performed according to equation (8).
In step 4, the fitness value of each butterfly is calculated again, the optimal fitness value bF obtained by calculation is compared with the optimal fitness value obtained in step 4, the minimum value is assigned to the global optimal fitness value BF, and the position information of the butterfly is recorded.
In step 4, combining the Levy flight with the optimal butterfly position found in step 4.3, and updating the search position according to the formula (13); wherein the content of the first and second substances,
Figure BDA0003089707100000091
the position of the optimal butterfly recorded in step 4.3, dim is the dimension 2 of the population,
Figure BDA0003089707100000092
the position of the optimal butterfly in the step 4.3 after the Levy flight; will be provided with
Figure BDA0003089707100000093
Inputting the position information into a support vector machine model, calculating a fitness value according to the formula (4) again after the step 3 is executed, recording an optimal fitness value, retaining the position information of the butterfly, comparing the optimal fitness value with the global optimal fitness value BF obtained in the step 4.3, updating the BF to a current optimal fitness value if the value is smaller than the BF, storing the position information of the current optimal fitness value, and keeping the original BF value and the original butterfly position information unchanged if the value is not smaller than the BF.
In step 4, whether the optimization process of the improved butterfly optimization algorithm is finished is judged according to the maximum iteration number of the improved butterfly optimization algorithm set in the step 2, and if the iteration number is smaller than the set iteration number of 100, iteration is performed again; if the number of iterations at this time is greater than the set number of iterations 100, step 5 is performed.
In step 4, the butterfly position with the optimal fitness value is output and is brought into the support vector machine model. The position of the butterfly with the optimal fitness value corresponds to a vector formed by the penalty factor C of the support vector machine model and the key parameter eta of the Gaussian kernel function, and the penalty factor C of the support vector machine model and the optimal value of the key parameter eta of the Gaussian kernel function at the moment are brought into the support vector machine model.
In step 5, the wind speed and the wind direction in the prediction data are used as the input of the support vector machine model, the output power of the wind power generation in the prediction data is used as the output of the support vector machine model, and a group of wind power generation output power prediction results are output.
In step 6, the prediction result obtained in step 5 is denormalized, and the prediction result is denormalized according to equation (14).
Through the steps, the function of predicting the short-term wind power generation output power can be completed.
In order to better show the performance of the support vector machine model optimized by the improved butterfly optimization algorithm, MATLAB software is used for displaying on a display screen of a computer, and the actual value, the improved butterfly optimization algorithm-support vector machine model (IBOA-SVM), the butterfly optimization algorithm optimization support vector machine model (BOA-SVM), the particle swarm optimization algorithm optimization support vector machine model (PSO-SVM), the genetic algorithm optimization support vector machine model (GA-SVM) and the prediction effect of a BP neural network (BP) are compared; and displaying a comparison graph of prediction results of the output feedforward neural network, the extreme learning machine, the support vector machine and the optimized correlation vector machine model of the improved slime optimization algorithm on a display screen of the computer, wherein the abscissa is the prediction time and the ordinate is the output power of the wind power generation, and the comparison graph is shown in fig. 2.
In order to further verify the prediction effect of the improved butterfly optimization algorithm-support vector machine model, mean Absolute Error (MAE), root-Mean-square Error (RMSE) and coefficient of determination (R) are selected 2 ) As an evaluation index for the predicted effect.
Under the same conditions and parameters, selecting indexes MAE, RMSE and R 2 Evaluating the prediction results of an improved butterfly optimization algorithm-support vector machine model (IBOA-SVM), a butterfly optimization algorithm optimization support vector machine model (BOA-SVM), a particle swarm optimization algorithm optimization support vector machine model (PSO-SVM), a genetic algorithm optimization support vector machine model (GA-SVM) and a BP neural network (BP), wherein the evaluation results are shown in a table 1;
table 1 photovoltaic power generation output power prediction result evaluation
Figure BDA0003089707100000101
As can be seen from Table 1, the support vector machine model optimized by the improved butterfly optimization algorithm is superior to other comparison models in terms of average absolute error, root mean square error and decision coefficient.
In the above embodiments, the butterfly optimization algorithm and the support vector machine model are well known in the art and are well known to those skilled in the art; the wind power output, wind speed and wind direction used for the prediction are well known to those skilled in the art; a method of inputting the acquired training data and prediction data of the wind power generation output power into a computer is a known method; the computer, display and MATLAB computer software were all commercially available.
It should be emphasized that the embodiments described herein are illustrative and not restrictive, and thus the present invention includes, but is not limited to, the embodiments described in the detailed description, as well as other embodiments that can be derived by one skilled in the art from the teachings herein.

Claims (4)

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, dividing wind power generation data into training data and prediction data, and carrying out normalization processing on the data;
step 2, setting and initializing parameters of an improved butterfly optimization algorithm and a support vector machine model;
step 3, training a support vector machine model by using training data;
step 4, calculating the butterfly position with the optimal fitness value by using an improved butterfly optimization algorithm;
step 5, substituting the butterfly position with the optimal fitness value into a support vector machine model, and predicting the output power of the wind power generation by using the support vector machine model;
step 6, outputting a prediction result;
the specific implementation method of the step 2 comprises the following steps: setting the iteration times of an improved butterfly optimization algorithm, the quantity of butterflies in the population and the dimensionality of the population, and initializing the positions of the butterflies; the support vector machine model sets the range of a penalty factor C and a Gaussian kernel function key parameter eta; the range of the butterfly position in the butterfly optimization algorithm is improved to be the range of a penalty factor C of a support vector machine model and a key parameter eta of a Gaussian kernel function, and the Gaussian kernel function is expressed as follows;
Figure FDA0003727571450000011
in the formula, K (x) i ,x j ) Representing a Gaussian kernel function, wherein eta represents a key parameter of the Gaussian kernel function;
the method for initializing the butterfly position comprises the following steps:
Figure FDA0003727571450000012
in the formula (I), the compound is shown in the specification,
Figure FDA0003727571450000013
indicating the initial position of each butterfly, rand denotes a random number belonging between 0 and 1,
Figure FDA0003727571450000014
represents a vector formed by a penalty factor C and an upper bound of a key parameter eta of a Gaussian kernel function in a support vector machine model,
Figure FDA0003727571450000015
representing a vector formed by a penalty factor C and a lower bound of a Gaussian kernel function key parameter eta in a support vector machine model;
the specific implementation method of the step 4 comprises the following steps:
step 4.1, calculating the fitness value of each butterfly in the butterfly optimization algorithm and the fragrance intensity generated by the butterfly, and recording the optimal fitness value and the corresponding butterfly position in the butterfly optimization algorithm;
step 4.2, updating the position of the butterfly;
step 4.3, calculating the fitness value of each butterfly again, comparing the optimal fitness value bF obtained by calculation with the optimal fitness value obtained by calculation in the step 4.1, assigning the minimum value to the global optimal fitness value BF, and recording the position information of the butterfly;
step 4.4, executing a Levy flight strategy;
step 4.5, updating population individuals according to the fitness value until the set maximum iteration times are reached, and obtaining a butterfly position with the optimal fitness value;
the specific implementation method of the step 4.2 is as follows:
first, a switching probability is generated according to a dynamic switching probability formula as follows:
θ＝θ max -(θ max -θ min )×(T max -t)/T max
where θ is the probability of switching, θ max To switch the probability maximum, θ min To switch the minimum probability, T max The maximum iteration number of the algorithm is represented, and t is the current iteration number of the algorithm;
then, comparing the switching probability theta with a generated random number r with the value range of 0 to 1, and if the value of the random number r is greater than that of the switching probability theta, updating the global position according to the following formula:
Figure FDA0003727571450000021
in the formula (I), the compound is shown in the specification,
Figure FDA0003727571450000022
is the position of the ith butterfly in the t iteration, g * Represents the optimal solution among all solutions of the current iteration, f i Represents the amount of fragrance emitted by the ith butterfly, and w is a weighting factor; the calculation formula of the weight factor w is as follows:
Figure FDA0003727571450000023
in the formula, T max The maximum iteration number of the algorithm is represented, and t is the current iteration number of the algorithm;
if the value of the random number r is smaller than the value of the switching probability θ, local location update is performed according to the following formula:
Figure FDA0003727571450000024
in the formula (I), the compound is shown in the specification,
Figure FDA0003727571450000025
and
Figure FDA0003727571450000026
representing the positions of the jth and kth butterflies in the tth iterative solution space;
the specific implementation method of the step 4.4 comprises the following steps:
the flight step length of the Levy flight strategy is set as follows:
Figure FDA0003727571450000027
wherein s is the step length of Levy flight, levy (beta), and u and v are both subject to normal distribution, wherein:
Figure FDA0003727571450000028
combining the step length of the Levy flight with the optimal butterfly position obtained in the step 4.3, wherein the updated new search position is as follows:
Figure FDA0003727571450000029
in the formula (I), the compound is shown in the specification,
Figure FDA00037275714500000210
for the position of the optimal butterfly, dim is the dimension of the population,
Figure FDA00037275714500000211
to pass through Levy the position of the optimal butterfly after flight;
will be provided with
Figure FDA00037275714500000212
Inputting the position information into a support vector machine model, calculating the fitness value again after the step 3 is executed, recording the optimal fitness value, and keeping the position information of the butterfly; comparing the optimal fitness value with the BF value obtained in the step 4.3, if the optimal fitness value is smaller than the BF value, updating the BF value to the current optimal fitness value, storing the position information of the current optimal fitness value butterfly, and otherwise, keeping the original BF value and the butterfly position information unchanged;
the specific implementation method of the step 5 is as follows: the position of a butterfly with the optimal fitness value corresponds to a penalty factor C and a Gaussian kernel function key parameter eta of the support vector machine model, and the penalty factor C and the optimal value of the Gaussian kernel function key parameter eta of the support vector machine model at the moment are brought into the support vector machine model; and taking the wind speed and the wind direction in the predicted data as the input of the support vector machine model, taking the output power of wind power generation in the predicted data as the output of the support vector machine model, and outputting a group of wind power generation output power prediction results.
2. The method of claim 1, wherein the method of predicting short term wind power generation output power comprises: the training data and the prediction data comprise input data and output data, the input data comprise wind speed and wind direction, and the output data are output power of wind power generation; the data normalization processing method comprises the following steps:
Figure FDA0003727571450000031
in the formula, P scale,i Representing a normalized data value, P, of the wind power output i Representing the true value, P, of the wind power output min Representing the minimum value, P, of the wind power output max Represents the maximum value of the wind power output.
3. The method for predicting the short-term wind power generation output power of claim 1, wherein: step 4.1 is to use a fitness function calculation formula to calculate the fitness value of each butterfly, wherein the fitness function selects a standard mean square error NRMSE, which is expressed as follows:
Figure FDA0003727571450000032
in the formula, P N The rated power of the generator set, N is the number of wind power generation output power in the data to be predicted, P i For the true value, Y, of the wind power output in the output data i Predicting values of a group of wind power generation output powers output by the support vector machine model in the step 3;
NRMSE calculation result is a fitness value, the current fitness value S (i) is arranged from large to small, the optimal fitness value is recorded as bF, and the position of the butterfly individual is stored;
the position update of the butterfly is related to the fragrance intensity generated by the butterfly, and the calculation formula of the fragrance intensity is as follows:
f＝bI α
where f is the perceived intensity of the fragrance, b is the sensory modality of the butterfly, I is the stimulus intensity, and α is a modality-dependent power exponent.
4. The method for predicting the short-term wind power generation output power of claim 1, wherein: when the prediction result is output in step 6, the prediction result needs to be subjected to inverse normalization processing, wherein the inverse normalization processing adopts the following formula:
P i ＝P scale,i ×(P max -P min )+P min
in the formula, P i Representing true value, P, of wind power output min Representing the minimum value, P, of the wind power output max Representing the maximum value, P, of the wind power output scale,i And the data value after the output power of the wind power generation is normalized is represented.
CN202110592303.7A 2021-05-28 2021-05-28 Method for predicting short-term wind power generation output power Expired - Fee Related CN113205228B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110592303.7A CN113205228B (en) 2021-05-28 2021-05-28 Method for predicting short-term wind power generation output power

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110592303.7A CN113205228B (en) 2021-05-28 2021-05-28 Method for predicting short-term wind power generation output power

Publications (2)

Publication Number Publication Date
CN113205228A CN113205228A (en) 2021-08-03
CN113205228B true CN113205228B (en) 2022-10-14

Family

ID=77023578

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110592303.7A Expired - Fee Related CN113205228B (en) 2021-05-28 2021-05-28 Method for predicting short-term wind power generation output power

Country Status (1)

Country Link
CN (1) CN113205228B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114115431B (en) * 2021-11-30 2023-05-30 浙江佳乐科仪股份有限公司 Photovoltaic power generation maximum power tracking method and system
CN117521534B (en) * 2024-01-08 2024-03-15 中国人民解放军陆军装甲兵学院 Integrated matching design method for hub motor driving device based on data driving

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112670984A (en) * 2020-12-16 2021-04-16 广州大学 Power system economic load distribution method based on improved whale algorithm

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111832838B (en) * 2020-07-24 2022-03-01 河北工业大学 Method for predicting short-term wind power generation output power

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112670984A (en) * 2020-12-16 2021-04-16 广州大学 Power system economic load distribution method based on improved whale algorithm

Also Published As

Publication number Publication date
CN113205228A (en) 2021-08-03

Similar Documents

Publication Publication Date Title
Liu et al. Random forest solar power forecast based on classification optimization
CN108921339B (en) Quantile regression-based photovoltaic power interval prediction method for genetic support vector machine
CN104978613B (en) A kind of photovoltaic output short term prediction method considering component temperature
CN109615146B (en) Ultra-short-term wind power prediction method based on deep learning
CN113205228B (en) Method for predicting short-term wind power generation output power
CN110942205B (en) Short-term photovoltaic power generation power prediction method based on HIMVO-SVM
Ye et al. Combined approach for short-term wind power forecasting based on wave division and Seq2Seq model using deep learning
CN105447509A (en) Short-term power prediction method for photovoltaic power generation system
CN112186761B (en) Wind power scene generation method and system based on probability distribution
CN112669168B (en) Short-term prediction method for wind power
CN113537582B (en) Photovoltaic power ultra-short-term prediction method based on short-wave radiation correction
CN113240193A (en) Short-term power load prediction method based on variational modal decomposition and artificial bee colony optimization
CN111832839B (en) Energy consumption prediction method based on sufficient incremental learning
CN112418495A (en) Building energy consumption prediction method based on longicorn stigma optimization algorithm and neural network
CN115275991A (en) Active power distribution network operation situation prediction method based on IEMD-TA-LSTM model
CN115759415A (en) Power consumption demand prediction method based on LSTM-SVR
CN116402203A (en) Method, system and medium for predicting short-time photovoltaic power generation capacity considering weather conditions
CN112307672A (en) BP neural network short-term wind power prediction method based on cuckoo algorithm optimization
CN115995810A (en) Wind power prediction method and system considering weather fluctuation self-adaptive matching
CN108694475B (en) Short-time-scale photovoltaic cell power generation capacity prediction method based on hybrid model
CN115907131B (en) Method and system for constructing electric heating load prediction model in northern area
CN110991748A (en) Short-term load prediction method for urban power grid
CN117874584A (en) Neural network ultra-short term photovoltaic power prediction method based on meteorological clustering
CN115796327A (en) Wind power interval prediction method based on VMD (vertical vector decomposition) and IWOA-F-GRU (empirical mode decomposition) -based models
CN116167508A (en) Short-term photovoltaic output rapid prediction method and system based on meteorological factor decomposition

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: 20221014

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