CN108197744B - Method and system for determining photovoltaic power generation power - Google Patents
Method and system for determining photovoltaic power generation power Download PDFInfo
- Publication number
- CN108197744B CN108197744B CN201810001865.8A CN201810001865A CN108197744B CN 108197744 B CN108197744 B CN 108197744B CN 201810001865 A CN201810001865 A CN 201810001865A CN 108197744 B CN108197744 B CN 108197744B
- Authority
- CN
- China
- Prior art keywords
- historical
- power generation
- determining
- photovoltaic power
- day
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
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
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
Abstract
The invention discloses a method and a system for determining photovoltaic power generation power. The method optimizes the radial basis function neural network model based on the improved thinking evolution algorithm, obtains the historical similar days of the prediction days on the basis of considering the weight of each influence factor, and has more rational data on the selection of meteorological factors (namely influence factors); the similar day selection algorithm is more effective; and (3) taking the similar day data and the predicted day weather data as input, predicting the predicted solar photovoltaic power generation power by adopting an improved thinking evolution algorithm optimized radial basis function neural network model, and determining the photovoltaic power generation power of the day to be measured. By adopting the determination method and the determination system provided by the invention, the error in determining the photovoltaic power generation power can be reduced.
Description
Technical Field
The invention relates to the field of power systems, in particular to a method and a system for determining photovoltaic power generation power.
Background
In recent years, China becomes the country with the fastest global photovoltaic power generation installation device growth, and the domestic photovoltaic power generation market is developing from an independent power generation system to a grid-connected power generation system. The photovoltaic power generation power is intermittent due to day and night alternation, and has volatility and randomness under the influence of factors such as weather, so that the photovoltaic power generation power is accurately determined in advance, and the stable and reliable operation of a power grid in the photovoltaic power generation grid-connection process is important for the development of the photovoltaic power generation technology.
At present, the research on the short-term advance determination of the photovoltaic power generation generally needs numerical weather forecast including key meteorological factors and irradiance, and prediction is carried out by adopting a photovoltaic power generation power prediction algorithm such as a neural network, classification regression, a time sequence, wavelet analysis and the like. However, many literatures lack theoretical analysis for selecting meteorological factors, and the weight problem of each meteorological factor (namely, influence factor) is not considered in the selection process of similar days, so that in the prior art, each meteorological factor is analyzed and selected, and the obtained similar days similar to the meteorological data of the day to be measured are many, so that the photovoltaic power generation power error determined in advance is large.
Disclosure of Invention
The invention aims to provide a method and a system for determining photovoltaic power generation power, which aim to solve the problem that the photovoltaic power generation power determined in advance in the prior art has large error.
In order to achieve the purpose, the invention provides the following scheme:
a method of determining photovoltaic power generation, comprising:
acquiring historical meteorological data and historical photovoltaic power generation power; the historical meteorological data comprises temperature, humidity, irradiance, wind speed and wind direction;
determining an influence factor according to the historical meteorological data and the historical photovoltaic power generation power; the influence factors comprise temperature, humidity and irradiance;
determining the weight of the influence factor according to the historical photovoltaic power generation power;
acquiring weather data of a day to be detected;
determining a historical day with the similarity of the weather data of the day to be detected higher than a similarity threshold according to the weight;
acquiring historical solar weather data and historical solar photovoltaic power generation power of the historical days;
optimizing the initial weight and the threshold of the radial basis function neural network to obtain an optimized radial basis function neural network;
establishing an improved thinking evolution algorithm optimized radial basis function neural network model by taking the historical solar meteorological data as the input of the optimized radial basis function neural network and taking the historical solar photovoltaic power generation power as the output of the optimized radial basis function neural network;
and inputting the meteorological data of the day to be detected into the improved thought evolution algorithm to optimize the radial basis function neural network model, and outputting the photovoltaic power generation power of the day to be detected.
Optionally, the determining an influence factor according to the historical meteorological data and the historical photovoltaic power generation power specifically includes:
normalizing the historical meteorological data and the historical photovoltaic power generation power to obtain a relationship between the processed historical meteorological data and the historical photovoltaic power generation power;
and determining an influence factor according to the relation between the historical meteorological data and the historical photovoltaic power generation power.
Optionally, the determining the weight of the impact factor according to the historical photovoltaic power generation power specifically includes:
carrying out correlation analysis on the influence factors and the historical photovoltaic power generation power to obtain correlation coefficients between the historical photovoltaic power generation power and the historical meteorological data;
determining the influence factor by adopting an average influence value algorithm according to the correlation coefficient;
determining the weight of the influence factor according to an entropy weight method.
Optionally, the determining, according to the weight, a historical day for which the similarity with the weather data of the day to be measured is higher than a similarity threshold specifically includes:
and determining the historical days with the similarity higher than the similarity threshold value with the weather data of the day to be detected by adopting an optimal similarity coefficient method according to the weight.
Optionally, the optimizing the initial weight and the threshold of the radial basis function neural network to obtain an optimized radial basis function neural network specifically includes:
and optimizing the initial weight and the threshold of the radial basis function neural network by adopting an improved thought evolution algorithm to obtain the optimized radial basis function neural network.
A photovoltaic power generation power determination system, comprising:
the first acquisition module is used for acquiring historical meteorological data and historical photovoltaic power generation power; the historical meteorological data comprises temperature, humidity, irradiance, wind speed and wind direction;
the influence factor determining module is used for determining an influence factor according to the historical meteorological data and the historical photovoltaic power generation power; the influence factors comprise temperature, humidity and irradiance;
the weight determining module is used for determining the weight of the influence factor according to the historical photovoltaic power generation power;
the to-be-detected solar weather data acquisition module is used for acquiring to-be-detected solar weather data of a to-be-detected day;
the historical day determining module is used for determining a historical day with the similarity higher than a similarity threshold value with the weather data of the day to be detected according to the weight;
the second acquisition module is used for acquiring historical solar weather data and historical solar photovoltaic power generation power of the historical days;
the optimization module is used for optimizing the initial weight and the threshold of the radial basis function neural network to obtain an optimized radial basis function neural network;
the model establishing module is used for establishing an improved thinking evolution algorithm optimized radial basis function neural network model by taking the historical solar meteorological data as the input of the optimized radial basis function neural network and taking the historical solar photovoltaic power generation power as the output of the optimized radial basis function neural network;
and the photovoltaic power generation power determining module is used for inputting the meteorological data of the day to be detected into the improved thinking evolution algorithm optimized radial basis function neural network model and outputting the photovoltaic power generation power of the day to be detected.
Optionally, the influence factor determining module specifically includes:
the normalization processing unit is used for performing normalization processing on the historical meteorological data and the historical photovoltaic power generation power to obtain a processed historical meteorological data-historical photovoltaic power generation power relation;
and the first determining unit of the influence factor is used for determining the influence factor according to the relation between the historical meteorological data and the historical photovoltaic power generation power.
Optionally, the weight determining module specifically includes:
the correlation analysis unit is used for carrying out correlation analysis on the influence factors and the historical photovoltaic power generation power to obtain a correlation coefficient between the historical photovoltaic power generation power and the historical meteorological data;
the second determining unit of the influence factor is used for determining the influence factor by adopting an average influence value algorithm according to the correlation coefficient;
and the second weight determining unit is used for determining the weight of the influence factor according to an entropy weight method.
Optionally, the historical date determining module specifically includes:
and the historical day determining unit is used for determining the historical day with the similarity higher than the similarity threshold value with the weather data of the day to be detected by adopting an optimal similarity coefficient method according to the weight.
Optionally, the optimization module specifically includes:
and the optimization unit is used for optimizing the initial weight and the threshold of the radial basis function neural network by adopting an improved thought evolution algorithm to obtain the optimized radial basis function neural network.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects: the invention provides a method and a system for determining photovoltaic power generation power, wherein influence factors are determined according to historical meteorological data and historical photovoltaic power generation power, and all historical days with higher similarity to the meteorological data of days to be detected are selected selectively according to the weight of the influence factors, so that the photovoltaic power generation power of the days to be detected can be determined more accurately, and errors are reduced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a flow chart of a method for determining photovoltaic power generation provided by the present invention;
FIG. 2 is a comparative graph of photovoltaic power generation power obtained in 4 ways of 6 months, 30 days and 4 days provided by the invention;
fig. 3 is a diagram of a system for determining photovoltaic power generation provided by the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a method and a system for determining photovoltaic power generation power, which can reduce the error of determining the photovoltaic power generation power of a day to be measured and improve the prediction precision of the photovoltaic power generation power.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
The method aims at the problem that photovoltaic power generation power is influenced by meteorological factors and has volatility and randomness, short-term photovoltaic power generation power is predicted based on a mode of optimizing a Radial Basis Function (RBF) by an improved thought evolution algorithm (IMEA), corresponding meteorological factors are selected as input indexes by utilizing correlation analysis and Mean Impact Value (MIV) algorithm, similar days of predicted days are obtained by computing by considering an optimal similarity coefficient method of weight, similar day data and predicted day meteorological data are used as input, and a radial basis neural network model is optimized by the improved thought evolution algorithm to predict the predicted solar power generation power, so that the prediction accuracy of the solar power generation power is improved.
Fig. 1 is a flowchart of a method for determining photovoltaic power generation provided by the present invention, and as shown in fig. 1, a method for determining photovoltaic power generation includes:
step 101: acquiring historical meteorological data and historical photovoltaic power generation power; the historical meteorological data comprises temperature, humidity, irradiance, wind speed and wind direction.
Step 102: determining an influence factor according to the historical meteorological data and the historical photovoltaic power generation power; the influence factors include temperature, humidity, irradiance.
The step 102 specifically includes: normalizing the historical meteorological data and the historical photovoltaic power generation power to obtain a relationship between the processed historical meteorological data and the historical photovoltaic power generation power; and determining an influence factor according to the relation between the historical meteorological data and the historical photovoltaic power generation power.
Step 103: and determining the weight of the influence factor according to the historical photovoltaic power generation power.
The step 103 specifically includes: carrying out correlation analysis on the influence factors and the historical photovoltaic power generation power to obtain correlation coefficients between the historical photovoltaic power generation power and the historical meteorological data; determining the influence factor by adopting an average influence value algorithm according to the correlation coefficient; determining the weight of the influence factor according to an entropy weight method.
Selection of influencing factors
And after the data are normalized, carrying out correlation analysis on the photovoltaic power generation power and 5 meteorological factors including temperature, humidity, irradiance, wind speed and wind direction to obtain correlation coefficients of the photovoltaic power generation power and the corresponding meteorological factors every day. Wind direction can be removed through calculation of the correlation coefficient, the correlation coefficient of wind speed and power is large in some time periods, and the wind speed, which is a meteorological factor, is not easy to directly ignore only through correlation analysis, so that a Mean Impact Value algorithm (MIV) is adopted to select a subsequent model input meteorological factor through a BP neural network. The specific application steps are as follows:
1) 4 meteorological factors are combined into a training sample S, S ═ X1；X2；X3；X4]，X1To X4Respectively representing 4 meteorological factor vectors of temperature, humidity, irradiance and wind speed, and outputting the photovoltaic power generation power P as a network. After the network training is finished, each meteorological factor in the training sample S is respectively added or subtracted by 10 percent on the basis of the original value to form a new training sample S1，S2。
2) Will S1，S2Respectively used as simulation samples to carry out simulation by using the established network to obtain two simulation results F1、F2Calculating F1、F2The difference value is the pair after the independent variable is changedAnd outputting the generated influence change Value (IV), and finally averaging the IV according to the number of observation cases to obtain the MIV of the meteorological factor corresponding to the dependent variable network output.
3) After calculating the MIV values corresponding to the 4 meteorological factors, the bit number of relative importance of the 4 meteorological factors to the network output is obtained after the 4 meteorological factors are sequenced according to the MIV absolute value.
According to the steps, the MIV values of 4 meteorological factors of the temperature, the humidity, the irradiance and the wind speed are respectively 0.324, 0.128, 0.408 and 0.003. Wherein the MIV value of irradiance is the highest, the MIV value of temperature and humidity is the second order, and the MIV value of wind speed is not in an order of magnitude with the first three meteorological factors. Therefore, through the MIV index of the BP neural network, the temperature, the humidity and the irradiance are selected as the input meteorological factors of the subsequent model, and the wind speed is not used as the input meteorological factor.
Step 104: and acquiring weather data of the day to be detected.
Step 105: and determining the historical days with the similarity higher than the similarity threshold value with the weather data of the day to be detected according to the weight.
The step 105 specifically includes: and determining the historical days with the similarity higher than the similarity threshold value with the weather data of the day to be detected by adopting an optimal similarity coefficient method according to the weight.
And determining the weight of each meteorological factor by adopting an optimal similarity coefficient method and combining an entropy weight method to select the similar day of the day to be predicted. Record the weather feature vector of every day as Xi＝[Xi1，Xi2，Xi3]TI represents day i, wherein Xi1＝[Xi1(1)，Xi1(2)...Xi(n)]TRepresents a daily temperature vector; xi2＝[Xi2(1)，Xi2(2)...Xi2(n)]TRepresents the humidity vector of each day, where T is the transpose of the matrix; xi3＝[Xi3(1)，Xi3(2)...Xi3(n)]TRepresenting the daily solar irradiance vector. The meteorological feature vector of the day to be predicted is X0＝[X01，X02，X03]T. Calculating three meteorology features by entropy weight methodThe weight occupied by the eigenvector is mainly divided into the following two steps:
1) calculating information entropy of each feature vector
Wherein, YijRepresents Xi(j) An evaluation value of an i-th meteorological feature vector (i-1, 2,3) under an index (j-1, 2, …, n);
wherein, PijIs the proportion of the index value of the ith meteorological feature vector under the jth index;
if p isijWhen 0, then
2) Determining the weights of the indexes
According to the calculation formula of the information entropy, calculating the information entropy of each meteorological feature vector to be E1,E2,...,Ek. Calculating the weight of each index through the information entropy:
the optimal similarity coefficient method mainly comprises the following steps:
1) shape coefficient of optimal similarity coefficient
2) Value coefficient of optimal similarity coefficient
Vijk＝e-Dijk (5)
3) Combining the shape coefficient and the value coefficient to obtain the optimal similarity coefficient
BFVijk＝Fijk·Vijk(7)
4) And synthesizing the optimal similarity coefficients of all the influence factors, wherein the similarity between the day to be predicted and the historical day i is
At this time, j represents 3 meteorological feature vectors
5) According to the similarity value lambdaiThe sizes of the data are sorted from big to small to obtain a similar day sequence D of the days to be predictedi＝[d1，d2...di]Selecting a part with the maximum similarity value as a similar day, wherein diIndicating the day number.
Step 106: historical solar weather data and historical solar photovoltaic power generation power of the historical days are obtained.
Step 107: and optimizing the initial weight and the threshold of the radial basis function neural network to obtain the optimized radial basis function neural network.
The step 107 specifically includes: and optimizing the initial weight and the threshold of the radial basis function neural network by adopting an improved thought evolution algorithm to obtain the optimized radial basis function neural network.
The initial weight is the weight from the hidden layer of the radial basis function neural network to the output layer; the threshold is a parameter used for adjusting the sensitivity of the neuron in the radial basis function neural network.
Step 108: and establishing an improved thinking evolution algorithm optimized radial basis function neural network model by taking the historical solar meteorological data as the input of the optimized radial basis function neural network and taking the historical solar photovoltaic power generation power as the output of the optimized radial basis function neural network.
Improved thought evolution algorithm
The evolutionary thinking algorithm respectively carries out two rounds of individual scattering in the population initialization process, wherein a plurality of individuals with high scores in the first round of scattering form a winner sub-population, and a plurality of individuals with high scores in the second round of scattering form a temporary sub-population. The individual and the sub-group post the own serial number, action and score information on the local bulletin board and the global bulletin board respectively. In the course of evolution, within a sub-population, the process of competition of individuals in order to become a winner is called convergence; in the entire solution space, each sub-population continuously probes for new points in the competition process, and the process of replacing the winning sub-population with a temporary sub-population that scores more than the winning sub-population is called dissimilarity.
The convergence and differentiation strategies in the evolutionary thinking algorithm are very important, and improved convergence and differentiation strategies are respectively provided.
1) Dynamic convergence strategy
The convergence process is to distribute new generation individuals around the winner of the previous generation of the sub-population according to normal distribution, start a new round of individual score calculation, and generate a new winner sub-population. Wherein, the sub-population scale distributed near the winner is dynamically obtained according to the score of the winning individual, the new generation of individuals are distributed more near the winning individual with higher score of the previous generation, the calculation steps are as follows:
calculating individual x in winner sub-group containing n individualsiA score of s for (i ═ 1, 2.., n)i(i＝1,2,...,n)；
XiAs a center, obeys X to N (mu, sigma)2) Spreading MiNew individuals, where mu is the mathematical expectation, sigma is the standard deviation, L ≦ MiH or less, wherein L is the upper variable limit, H is the lower variable limit, MiIs a new individual; then:
and combining the new individuals with the new generation of sub-population.
The variance in normal distribution is dynamically obtained according to the distance between adjacent two evolutionary winning individuals and the difference of scores, and when the distance between the two evolutionary winning individuals is short, the score difference is obtainedWhen the value is large, the variance is reduced, the optimal point is searched for in a fine mode, otherwise, the variance is increased, and the optimal point is searched for in a rough mode. Let the win individuals of two adjacent evolutions in the region omega be respectively scored as sm kX ofm kAnd a score of sn k+1X ofn k+1The dynamic variance is calculated as follows:
calculating the distance per unit value of adjacent two evolutionary dominant individuals
Wherein, i, j respectively represent i individuals and j individuals obtained by two adjacent evolutions, m, n are serial numbers of the dominant individuals in the two evolutions;
calculating per unit value of the difference between the two adjacent evolutionary scores
Obtaining dynamic variance
σk+1＝σk·(d1/d2) (12)
2) Simplex diversification strategy
Let the winner have n +1, xiIs a simplex of n +1 vertices. i is 0,1, …, n
(1) reflection
Worst sub-population score
Wherein h is an abbreviation for high, i.e.: x satisfying this equation with the largest function value being the worstiIs marked as xh。
Sub-differential population score
Where s represents between the high and low sub-populations.
Optimal sub-population score
l is an abbreviation for low, i.e.: the smallest function value is optimal.
Calculating out xhCentroid of all posterior vertices
Calculating xhReflection point x ofr
xr＝2·xc-xh (17)
② expanding
If f (x)r)＜flThen give an order
xe＝xc+α(xc-xh)，(α＞1) (18)
Alpha is an expansion coefficient, is used for amplification, and can be a value larger than 1, and is generally 2.
If f (x)e)＜f(xr) By xeSubstitution of xhReturn to 1 after forming a new winner sub-population, otherwise, use xrIn place of xhReturn to 1 after forming a new winner sub-population);
if fl≤f(xr)＜fsBy xrIn place of xhReturn to 1 after constructing a new winner sub-population).
(iii) shrinkage
If fs≤f(xr)＜fhThen give an order
xp＝xc+β(xr-xc)，(0＜β＜1) (19)
Beta is a contraction coefficient, generally 1/2.
If f (x)r)≥fhThen, thenOrder to
xp＝xc+β(xh-xc)，(0＜β＜1) (20)
If f (x)p)＜fhBy xpIn place of xhReturn to 1 after forming a new winner sub-population);
otherwise, let xi＝(xi+xl) N returns to 1 after constituting a new winner sub-population).
The winner of the new temporary sub-population can be obtained by the simplex optimization method.
(4) Improved thought evolution algorithm for optimizing radial basis function parameters
The method comprises the following steps of performing parameter optimization on a radial basis function by adopting an improved thought evolution algorithm, establishing an improved thought evolution algorithm optimized radial basis function neural network model for parameter photovoltaic power generation power prediction, and performing optimization steps as follows:
1) mapping from a solution space to a coding space is realized according to the topological structure of the RBF, and the MEA coding length is determined to be
L＝L1*L2+L2*L2+L2*L3+L2+L3 (23)
L1For RBF input layer node number, L2For hiding the number of layer nodes, L3The number of output layer nodes.
2) Selecting reciprocal of mean square error of training set as score function of each individual and population, and expression is
xobs,iRepresenting the true value, x, of the ith sampleobs,iRepresenting the predicted value of the ith sample.
3) Initializing the group to obtain a winner sub-group and a temporary sub-group, and calculating a global optimal individual and a score thereof through convergence and differentiation operations;
4) and substituting the optimal parameters obtained by optimizing the RBF by the IMEA algorithm into the RBF for continuous training.
Step 109: and inputting the meteorological data of the day to be detected into the improved thought evolution algorithm to optimize the radial basis function neural network model, and outputting the photovoltaic power generation power of the day to be detected.
In order to verify the effectiveness of the photovoltaic power generation power determination method provided by the invention, a certain photovoltaic power generation station in australia is taken as an example for example to carry out example analysis.
Meteorological data and photovoltaic power generation data of every half hour from 5/month 1/day to 6/month 30/day 7: 00-17: 00 in 2017 are used as sample data, wherein 5/month 1/day to 6/month 29/day is used as a historical day, and 6/month 30/day is used as a prediction day.
Predicting the photovoltaic power generation power of every 30min of the prediction day, respectively establishing RBF (radial basis function) and GA-RBF (genetic radial basis function) prediction models and PSO-RBF (particle swarm optimization radial basis function) prediction models for verifying the optimization effect of the IMEA algorithm, and comparing the prediction models with actual values measured by the determination method of the invention, wherein FIG. 2 is a comparison graph of the photovoltaic power generation power obtained in 4 modes of 6 months, 30 days and 4 days provided by the invention, and is shown in FIG. 2.
It can be seen from fig. 2 that the photovoltaic power generation power curve predicted by the 4 prediction models is substantially consistent with the actual power generation power curve, and the deviation of the power curve predicted by the RBF prediction model from the actual power curve is significantly larger than that of the other 3 models. The predicted effect of the 4 models was evaluated using Root-Mean-Square Error (RMSE) and Mean Absolute Percentage Error (MAPE). Comparing the RMSE corresponding to the four prediction models with the MAPE, the RBF after GA optimization is known, the RMSE is reduced from 30.62KW to 21.09KW, the MAPE is reduced from 43.26% to 29.57%, the RBF after PSO optimization is reduced from 21.84KW to 28.35%, the MAPE is reduced to 28.35%, the RBF after IMEA optimization has the best prediction effect in the four models, and the RMSE and the MAPE are reduced by 14.73KW and 20.51% respectively compared with the RBF before optimization. The analysis shows that the RFBNN has the worst prediction effect, the GA-RBF and PSO-RBF prediction accuracy is improved, and the IMEA-RNFNN prediction result is the best, so that the optimization of the RBF parameters based on the IMEA algorithm of similar days is effective, and the model prediction capability is greatly improved.
The method for determining the photovoltaic power generation power provided by the invention can achieve the following effects:
(1) according to the method, after the correlation coefficient of each meteorological factor and the photovoltaic power generation power is calculated, the influence factor is further selected by adopting an average influence value algorithm, so that the selection of the meteorological factors is more reasonable;
(2) obtaining historical similar days of the prediction days by adopting an optimal similarity coefficient method on the basis of considering the weight of each influence factor, so that a similar day selection algorithm is more effective;
(3) the improved thought evolution algorithm is used for optimizing parameters of the radial basis function neural network, the similar daily data and the predicted daily meteorological data are used as the input of the improved thought evolution algorithm optimized radial basis function neural network model to predict the photovoltaic power generation power, the prediction precision of the predicted photovoltaic power generation power is improved, and the error is reduced.
Fig. 3 is a structural diagram of a system for determining photovoltaic power generation provided by the present invention, and as shown in fig. 3, a system for determining photovoltaic power generation includes:
the first acquisition module 301 is used for acquiring historical meteorological data and historical photovoltaic power generation power; the historical meteorological data comprises temperature, humidity, irradiance, wind speed and wind direction.
An influence factor determination module 302, configured to determine an influence factor according to the historical meteorological data and the historical photovoltaic power generation power; the influence factors include temperature, humidity, irradiance.
The influence factor determining module 302 specifically includes: the normalization processing unit is used for performing normalization processing on the historical meteorological data and the historical photovoltaic power generation power to obtain a processed historical meteorological data-historical photovoltaic power generation power relation; and the first determining unit of the influence factor is used for determining the influence factor according to the relation between the historical meteorological data and the historical photovoltaic power generation power.
A weight determining module 303, configured to determine a weight of the impact factor according to the historical photovoltaic power generation power.
The weight determining module 303 specifically includes:
the correlation analysis unit is used for carrying out correlation analysis on the influence factors and the historical photovoltaic power generation power to obtain a correlation coefficient between the historical photovoltaic power generation power and the historical meteorological data;
the second determining unit of the influence factor is used for determining the influence factor by adopting an average influence value algorithm according to the correlation coefficient;
and the second weight determining unit is used for determining the weight of the influence factor according to an entropy weight method.
The to-be-detected day meteorological data acquisition module 304 is used for acquiring to-be-detected day meteorological data of a to-be-detected day;
and a historical day determining module 305, configured to determine, according to the weight, a historical day for which the similarity with the weather data of the day to be detected is higher than a similarity threshold.
The historical date determination module specifically comprises: and the historical day determining unit is used for determining the historical day with the similarity higher than the similarity threshold value with the weather data of the day to be detected by adopting an optimal similarity coefficient method according to the weight.
A second obtaining module 306 for obtaining historical solar weather data and historical solar photovoltaic power generation power for the historical day.
And an optimizing module 307, configured to optimize the initial weight and the threshold of the radial basis function neural network, to obtain an optimized radial basis function neural network.
The optimization module 307 specifically includes: and the initialization processing unit is used for optimizing the initial weight and the threshold of the radial basis function neural network by adopting an improved thought evolution algorithm to obtain the optimized radial basis function neural network.
And the model establishing module 308 is used for establishing an improved thinking evolution algorithm optimized radial basis function neural network model by taking the historical solar meteorological data as the input of the optimized radial basis function neural network and taking the historical solar photovoltaic power generation power as the output of the optimized radial basis function neural network.
And the photovoltaic power generation power determining module 309 is configured to input the daily meteorological data to be detected into the improved thought evolution algorithm optimized radial basis function neural network model, and output the photovoltaic power generation power of the day to be detected.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (4)
1. A method for determining photovoltaic power generation, comprising:
acquiring historical meteorological data and historical photovoltaic power generation power; the historical meteorological data comprises temperature, humidity, irradiance, wind speed and wind direction;
determining an influence factor according to the historical meteorological data and the historical photovoltaic power generation power; the influence factors comprise temperature, humidity and irradiance;
the determining an influence factor according to the historical meteorological data and the historical photovoltaic power generation power specifically comprises:
normalizing the historical meteorological data and the historical photovoltaic power generation power to obtain a relationship between the processed historical meteorological data and the historical photovoltaic power generation power;
determining an influence factor according to the relation between the historical meteorological data and the historical photovoltaic power generation power;
determining the weight of the influence factor according to the historical photovoltaic power generation power;
the method specifically comprises the following steps:
carrying out correlation analysis on the influence factors and the historical photovoltaic power generation power to obtain correlation coefficients between the historical photovoltaic power generation power and the historical meteorological data;
determining the influence factor by adopting an average influence value algorithm according to the correlation coefficient;
determining the weight of the influence factor according to an entropy weight method;
acquiring weather data of a day to be detected;
determining the historical day with the similarity higher than the similarity threshold value with the weather data of the day to be detected according to the weight, and specifically comprising the following steps: determining a historical day with the similarity of the meteorological data of the day to be detected higher than a similarity threshold value by adopting an optimal similarity coefficient method according to the weight of the influence factors; 1) determining a shape coefficient of the optimal similarity coefficient, 2) determining a value coefficient of the optimal similarity coefficient, 3) determining the optimal similarity coefficient by combining the shape coefficient and the value coefficient, and 4) integrating the optimal similarity coefficients of all the influence factors and calculating the similarity between the day to be predicted and the historical day;
acquiring historical solar weather data and historical solar photovoltaic power generation power of the historical days;
optimizing the initial weight and the threshold of the radial basis function neural network to obtain an optimized radial basis function neural network;
establishing an improved thinking evolution algorithm optimized radial basis function neural network model by taking the historical solar meteorological data as the input of the optimized radial basis function neural network and taking the historical solar photovoltaic power generation power as the output of the optimized radial basis function neural network;
and inputting the meteorological data of the day to be detected into the improved thought evolution algorithm to optimize the radial basis function neural network model, and outputting the photovoltaic power generation power of the day to be detected.
2. The determination method according to claim 1, wherein the optimizing the initial weight and the threshold of the radial basis function neural network to obtain the optimized radial basis function neural network specifically includes:
and optimizing the initial weight and the threshold of the radial basis function neural network by adopting an improved thought evolution algorithm to obtain the optimized radial basis function neural network.
3. A system for determining photovoltaic power generation, comprising:
the first acquisition module is used for acquiring historical meteorological data and historical photovoltaic power generation power; the historical meteorological data comprises temperature, humidity, irradiance, wind speed and wind direction;
the influence factor determining module is used for determining an influence factor according to the historical meteorological data and the historical photovoltaic power generation power; the influence factors comprise temperature, humidity and irradiance;
the influence factor determination module specifically includes:
the normalization processing unit is used for performing normalization processing on the historical meteorological data and the historical photovoltaic power generation power to obtain a processed historical meteorological data-historical photovoltaic power generation power relation;
the first determining unit of the influence factor is used for determining the influence factor according to the relation between the historical meteorological data and the historical photovoltaic power generation power;
the weight determining module is used for determining the weight of the influence factor according to the historical photovoltaic power generation power;
the weight determination module specifically includes:
the correlation analysis unit is used for carrying out correlation analysis on the influence factors and the historical photovoltaic power generation power to obtain a correlation coefficient between the historical photovoltaic power generation power and the historical meteorological data;
the second determining unit of the influence factor is used for determining the influence factor by adopting an average influence value algorithm according to the correlation coefficient;
a second weight determination unit, configured to determine a weight of the impact factor according to an entropy weight method;
the to-be-detected solar weather data acquisition module is used for acquiring to-be-detected solar weather data of a to-be-detected day;
the historical day determining module is used for determining a historical day with the similarity higher than a similarity threshold value with the weather data of the day to be detected according to the weight; the historical date determination module specifically comprises: the historical day determining unit is used for determining the historical day with the similarity higher than a similarity threshold value with the weather data of the day to be detected by adopting an optimal similarity coefficient method according to the weight; 1) determining a shape coefficient of the optimal similarity coefficient, 2) determining a value coefficient of the optimal similarity coefficient, 3) determining the optimal similarity coefficient by combining the shape coefficient and the value coefficient, and 4) integrating the optimal similarity coefficients of all the influence factors and calculating the similarity between the day to be predicted and the historical day;
the second acquisition module is used for acquiring historical solar weather data and historical solar photovoltaic power generation power of the historical days;
the optimization module is used for optimizing the initial weight and the threshold of the radial basis function neural network to obtain an optimized radial basis function neural network;
the model establishing module is used for establishing an improved thinking evolution algorithm optimized radial basis function neural network model by taking the historical solar meteorological data as the input of the optimized radial basis function neural network and taking the historical solar photovoltaic power generation power as the output of the optimized radial basis function neural network;
and the photovoltaic power generation power determining module is used for inputting the meteorological data of the day to be detected into the improved thinking evolution algorithm optimized radial basis function neural network model and outputting the photovoltaic power generation power of the day to be detected.
4. The determination system according to claim 3, wherein the optimization module specifically comprises:
and the optimization unit is used for optimizing the initial weight and the threshold of the radial basis function neural network by adopting an improved thought evolution algorithm to obtain the optimized radial basis function neural network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810001865.8A CN108197744B (en) | 2018-01-02 | 2018-01-02 | Method and system for determining photovoltaic power generation power |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810001865.8A CN108197744B (en) | 2018-01-02 | 2018-01-02 | Method and system for determining photovoltaic power generation power |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108197744A CN108197744A (en) | 2018-06-22 |
| CN108197744B true CN108197744B (en) | 2022-03-08 |
Family
ID=62588079
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810001865.8A Active CN108197744B (en) | 2018-01-02 | 2018-01-02 | Method and system for determining photovoltaic power generation power |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108197744B (en) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109165792B (en) * | 2018-09-14 | 2021-09-21 | 国网天津市电力公司 | Photovoltaic short-term output power prediction method based on SOA-WNN |
| CN109376863A (en) * | 2018-11-02 | 2019-02-22 | 国网浙江省电力有限公司宁波供电公司 | Photovoltaic power based on MIV-BP neural network is classified prediction technique in short term |
| CN111323847B (en) * | 2018-12-13 | 2023-05-23 | 北京金风慧能技术有限公司 | Method and apparatus for determining weight ratios for analog integration algorithms |
| CN110070226B (en) * | 2019-04-24 | 2020-06-16 | 河海大学 | Photovoltaic power prediction method and system based on convolutional neural network and meta-learning |
| CN110147908A (en) * | 2019-05-22 | 2019-08-20 | 东莞理工学院 | A kind of wind power forecasting method based on three-dimensional optimal similarity and improvement cuckoo algorithm |
| CN110675278A (en) * | 2019-09-18 | 2020-01-10 | 上海电机学院 | Photovoltaic power short-term prediction method based on RBF neural network |
| CN110929953A (en) * | 2019-12-04 | 2020-03-27 | 国网山东省电力公司电力科学研究院 | Photovoltaic power station ultra-short term output prediction method based on cluster analysis |
| CN111260126B (en) * | 2020-01-13 | 2022-12-09 | 燕山大学 | Short-term photovoltaic power generation prediction method considering correlation degree of weather and meteorological factors |
| CN111612648B (en) * | 2020-05-19 | 2024-01-19 | 广东电网有限责任公司电力调度控制中心 | Training method and device for photovoltaic power generation prediction model and computer equipment |
| CN112364477B (en) * | 2020-09-29 | 2022-12-06 | 中国电器科学研究院股份有限公司 | Outdoor empirical prediction model library generation method and system |
| CN113151842B (en) * | 2021-01-29 | 2023-07-11 | 河北建投新能源有限公司 | Method and device for determining conversion efficiency of wind-solar complementary water electrolysis hydrogen production |
| CN113379143A (en) * | 2021-06-23 | 2021-09-10 | 阳光电源股份有限公司 | Typical meteorological year construction method, power generation amount prediction method and related device |
| CN113761023A (en) * | 2021-08-24 | 2021-12-07 | 国网甘肃省电力公司 | Photovoltaic power generation short-term power prediction method based on improved generalized neural network |
| CN116381480A (en) * | 2023-03-30 | 2023-07-04 | 湖南雪墨电气科技有限公司 | Monitoring method, system and medium of intelligent desulfurization, denitrification and dedusting equipment |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104700156B (en) * | 2015-01-29 | 2018-07-03 | 广东电网有限责任公司电力科学研究院 | A kind of wind power forecasting method based on support vector machines selection variables |
| CN105469163A (en) * | 2015-12-08 | 2016-04-06 | 国家电网公司 | Similar day selection method used for photovoltaic power station power prediction |
| CN105631517A (en) * | 2015-12-17 | 2016-06-01 | 河海大学 | Photovoltaic power generation power short term prediction method based on mind evolution Elman neural network |
-
2018
- 2018-01-02 CN CN201810001865.8A patent/CN108197744B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108197744A (en) | 2018-06-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108197744B (en) | Method and system for determining photovoltaic power generation power | |
| CN108280552B (en) | Power load prediction method and system based on deep learning and storage medium | |
| CN110766200A (en) | Method for predicting generating power of wind turbine generator based on K-means mean clustering | |
| CN114970362B (en) | Power grid load scheduling prediction method and system under multi-energy structure | |
| CN114462718A (en) | CNN-GRU wind power prediction method based on time sliding window | |
| CN114444378A (en) | Short-term power prediction method for regional wind power cluster | |
| CN113344288A (en) | Method and device for predicting water level of cascade hydropower station group and computer readable storage medium | |
| CN113837499A (en) | Ultra-short-term wind power prediction method and system | |
| CN115545333A (en) | Method for predicting load curve of multi-load daily-type power distribution network | |
| CN115186923A (en) | Photovoltaic power generation power prediction method and device and electronic equipment | |
| CN115759389A (en) | Day-ahead photovoltaic power prediction method based on weather type similar day combination strategy | |
| CN116702937A (en) | Photovoltaic output day-ahead prediction method based on K-means mean value clustering and BP neural network optimization | |
| CN109116300B (en) | Extreme learning positioning method based on insufficient fingerprint information | |
| CN114357670A (en) | Power distribution network power consumption data abnormity early warning method based on BLS and self-encoder | |
| CN113344279A (en) | Resident load prediction method based on LSTM-SAM model and pooling | |
| CN112307672A (en) | BP neural network short-term wind power prediction method based on cuckoo algorithm optimization | |
| CN113487064A (en) | Photovoltaic power prediction method and system based on principal component analysis and improved LSTM | |
| CN116307139A (en) | Wind power ultra-short-term prediction method for optimizing and improving extreme learning machine | |
| CN110163437A (en) | Day-ahead photovoltaic power generation power prediction method based on DPK-means | |
| CN115907131A (en) | Method and system for building electric heating load prediction model in northern area | |
| CN113762591B (en) | Short-term electric quantity prediction method and system based on GRU and multi-core SVM countermeasure learning | |
| CN114971090A (en) | Electric heating load prediction method, system, equipment and medium | |
| CN114897204A (en) | Method and device for predicting short-term wind speed of offshore wind farm | |
| CN115481788A (en) | Load prediction method and system for phase change energy storage system | |
| CN110991743B (en) | Wind power short-term combination prediction method based on cluster analysis and neural network optimization |
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 |