CN111985678A - Photovoltaic power short-term prediction method - Google Patents

Photovoltaic power short-term prediction method Download PDF

Info

Publication number
CN111985678A
CN111985678A CN202010644968.3A CN202010644968A CN111985678A CN 111985678 A CN111985678 A CN 111985678A CN 202010644968 A CN202010644968 A CN 202010644968A CN 111985678 A CN111985678 A CN 111985678A
Authority
CN
China
Prior art keywords
day
historical
value
predicted
ant
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202010644968.3A
Other languages
Chinese (zh)
Inventor
王昕�
姚强
林鹏
张昕
谢震伟
刘春平
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.)
Yanbian Power Supply Co Of State Grid Jilinsheng Electric Power Supply Co
Shanghai Jiaotong University
State Grid Jilin Electric Power Corp
Original Assignee
Yanbian Power Supply Co Of State Grid Jilinsheng Electric Power Supply Co
Shanghai Jiaotong University
State Grid Jilin Electric Power Corp
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 Yanbian Power Supply Co Of State Grid Jilinsheng Electric Power Supply Co, Shanghai Jiaotong University, State Grid Jilin Electric Power Corp filed Critical Yanbian Power Supply Co Of State Grid Jilinsheng Electric Power Supply Co
Priority to CN202010644968.3A priority Critical patent/CN111985678A/en
Publication of CN111985678A publication Critical patent/CN111985678A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • 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
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02S—GENERATION OF ELECTRIC POWER BY CONVERSION OF INFRARED RADIATION, VISIBLE LIGHT OR ULTRAVIOLET LIGHT, e.g. USING PHOTOVOLTAIC [PV] MODULES
    • H02S50/00—Monitoring or testing of PV systems, e.g. load balancing or fault identification
    • 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
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A30/00—Adapting or protecting infrastructure or their operation
    • 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
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/50—Photovoltaic [PV] energy

Landscapes

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

Abstract

The invention provides a short-term photovoltaic power prediction method, which comprises the following steps: s1: collecting historical data of a local photovoltaic power plant as a data sample, the historical data comprising: date, weather information for each historical day within a historical time period, the weather information comprising: weather type, meteorological factors; the meteorological factors include: temperature value, humidity value, radiation value and cloud cluster value at each moment of the day; randomly selecting one day from historical days as a day to be predicted; s2: judging the weather type of the weather of the day to be predicted, and selecting the historical day most similar to the day to be predicted from the data samples of the same weather type by combining the weighted gray correlation degree; s3: and (4) optimizing and predicting the photovoltaic power by adopting the regularization parameter and the kernel parameter of the improved ant lion algorithm optimized least squares support vector machine with the most similar historical day in the step S2 as the background. Compared with the prior art, the method can improve the accuracy of photovoltaic short-term prediction.

Description

Photovoltaic power short-term prediction method
Technical Field
The invention relates to a photovoltaic power generation technology, in particular to a short-term photovoltaic power prediction method.
Background
Solar energy is an ideal renewable clean energy source, is inexhaustible, does not pollute the atmospheric environment, and has wide distribution region and strong exploitability. At present, solar power generation becomes a supplement mode of traditional power generation, the application technology of the solar power generation is more and more mature, and the photovoltaic power generation technology is widely applied. In recent decades, the government of China also focuses on research on new energy technologies on photovoltaic power generation projects, and a plurality of policies are issued in sequence to support wide popularization of photovoltaic grid-connected projects. According to the statistics of the national energy bureau, when the year is 2019 and the month is 6, 18559 ten thousand kilowatts are installed in the national photovoltaic power generation accumulation machine, the increase is 20 percent on the same scale, and 1140 thousand kilowatts are newly added. Photovoltaic power generation brings many problems to the safety of a power grid while relieving the energy crisis. Because the photovoltaic power generation is greatly influenced by meteorological factors, the output randomness and the intermittence of the photovoltaic power generation are high, and in order to avoid the damage of the photovoltaic power generation to a power grid, the photovoltaic power generation power needs to be predicted according to the meteorological factors. The prior art scheme is as follows:
1.1 physical method
Photovoltaic conversion efficiency model: the principle of solar cell power generation is that solar radiation is directly converted into electric energy based on the photovoltaic effect of a photoelectric semiconductor diode, and the output of the direct current power generation capacity of the photovoltaic cell is as follows:
PD(t)＝ηAG (1)
in the formula, PDIs the dc output electrical power (W) of the photovoltaic cell array; η is the efficiency coefficient of photoelectric conversion; a is the effective area (m) of solar radiation2) (ii) a G is the intensity of solar radiation (W/m)2)。
Assuming that a photovoltaic I/V curve is a binary function only containing two unknowns of photovoltaic cell temperature and solar radiation intensity, an ANN algorithm is utilized to calculate the temperature TcAnd the radiation quantity G is used as an input layer, in the training process, in order to reduce the sum of the prediction variances, an optimized reverse L-M propagation algorithm is used, and finally, an objective function, namely a photovoltaic I/V curve generated by MPL (multi-layer sensor), is output; still, japanese scholars propose a three-dimensional simulation calculation method, starting from a photovoltaic cell equivalent circuit, comprehensively considering factors such as shadows and wiring which affect photovoltaic power generation, obtaining an I/V curve of a single photovoltaic cell at a specific time, and further calculating instantaneous photovoltaic power generation power at a certain time according to P ═ VI.
1.2 prediction method based on cloud picture
Cloud is one of important weather factors influencing the irradiation intensity of sunlight, and if the generation, development and disappearance processes of the cloud can be accurately described, the prediction precision of the irradiation quantity of the sunlight can be obviously improved. Hammer, d.heinemann, e.lorenz, et al.short-term modeling of solar radiation: a static adaptive using satellite data [ J ]. Solar Energy, 1999, 67 (01-03): 139-. The main idea of the prediction method is mainly based on the Heliostat algorithm of satellite geometry.
Hammer, d.heinemann, c.hoyer, et al.solar energy analysis using Remote Sensing technologies [ J ]. Remote Sensing of Environment, 2003, 86 (03): 423 + 432, it is described that the researchers use the Heliostat semi-empirical method to determine the motion vector field of the cloud, first give the intensity of the sun radiation in clear sky at a certain place and time, then obtain the cloud index from the weather satellite image after considering the extinction of the cloud, the value of the index is related to the cloud transmission efficiency, as the cloud transmission effect is enhanced, the intensity of the sun radiation will show negative correlation change, thereby deducing the irradiation intensity reaching the ground. The same assumptions are used for the mathematical model of the cloud motion vector field, namely: the resolution/pixel size of each satellite picture is fixed; a constant gradient; each cloud motion vector field is described by a possible function equation; for a given motion vector field, a corresponding probability model is used, and the maximum probability model can be searched.
1.3 numerical weather mode prediction method
The numerical weather prediction is different from the traditional weather forecasting method, when the initial value and the boundary condition are known and determined, numerical operation is carried out through a large computer, so that a mathematical model, namely a hydrodynamics-thermodynamics equation set, for describing the weather evolution process is obtained, and the solar irradiation intensity in a certain period of time in the future can be predicted.
E.lorenz, j.hurka, d.heinemann, et al.irradiance for evaluating for the power prediction of grid-connected photovoltaic systems [ J ]. IEEE Journal of Selected Topics in Applied Earth underpressure and Remote Sensing, 2009, 2 (01): 2-10, recording that relevant researchers predict an hour predicted value of a specific station through a low-resolution European mesoscale prediction center, a coordinate position of a photovoltaic system and some post-processing measures, then obtaining a predicted value of the inclined plane irradiance by taking the photovoltaic system as a guide and combining an inclined plane irradiation model, then establishing a photovoltaic simulation model by considering the physical characteristics of a photovoltaic element, and finally obtaining a solar radiation predicted value from the photovoltaic simulation model to obtain photovoltaic output power. Verified, the root mean square error for a single site reached 37% on the first prediction day and increased to 46% by the third day.
The problems of the prior art are as follows:
2.1 physical method
The physical method carries out prediction based on physical equations such as a solar irradiation transfer equation and a photovoltaic module operation equation, and needs detailed geographic information of a photovoltaic power station and meteorological and solar irradiation data. The physical method prediction accuracy is greatly influenced by the built model, extreme abnormal weather conditions and slow changes of environment and photovoltaic module parameters along with time are difficult to simulate, the model has poor anti-interference capability and low robustness, and therefore the application in actual engineering is less.
2.2 cloud-based prediction method
Cloud-based prediction methods are limited by measurement device and method limitations. Since the shape and speed of the cloud cluster are assumed to be constant in a short time by default in the prediction, the method cannot obtain an accurate prediction result when the shape and moving speed of the cloud cluster are changed drastically. And the space resolution of the satellite cloud picture is low, and a pixel point on the picture corresponds to a large ground area, so that the cloud layer condition in a small range cannot be judged.
2.3 numerical weather mode prediction method
The irradiance obtained through the numerical weather mode may have a large error, and the main error sources are errors of an irradiation conversion model, a cloud amount prediction model and other forecast variables in the numerical weather model. The power value error obtained when performing prediction is large.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a short-term photovoltaic power prediction method. The technical scheme of the invention is as follows:
a photovoltaic power short-term prediction method comprises the following steps:
s1: collecting historical data of a local photovoltaic power plant as a data sample, the historical data comprising: date, weather information for each historical day within a historical time period, the weather information comprising: weather type, meteorological factors; the meteorological factors include: temperature value, humidity value, radiation value and cloud cluster value at each moment of the day; randomly selecting one day from historical days as a day to be predicted;
s2: judging the weather type of the weather of the day to be predicted, and selecting the historical day most similar to the day to be predicted from the data samples of the same weather type by combining the weighted gray correlation degree;
s3: and (4) optimizing and predicting the photovoltaic power by adopting the regularization parameter and the kernel parameter of the improved ant lion algorithm optimized least squares support vector machine with the most similar historical day in the step S2 as the background.
Optionally, the weather type includes sunny days, rainy days, cloudy days.
Optionally, the step S2 further includes:
s21: judging the weather type of the weather of the day to be predicted, and correspondingly listing the information of s historical days in the data sample belonging to the same weather type; s is a positive integer; one history day of the s history days is recorded as the ith history day, i is 1, … …, s;
s22: calculating the grey correlation degree of the day to be predicted and the ith historical day by using the formula (4);
the process is as follows:
the weather information is expressed by the formula (2)
Ti＝[xi1，xi2，xi3，xi4] (2)
In the above equation T represents a data sample meteorological information set,
Figure BDA0002571045670000041
is the value of the temperature at each moment in time,
Figure BDA0002571045670000042
represents the value of the humidity at each moment in time,
Figure BDA0002571045670000043
representing the radiance value at each time instant,
Figure BDA0002571045670000044
representing the cloud cluster magnitude at each moment, n representing the dimension of the meteorological factor;
calculating the gray correlation degree of the kth meteorological factor of the ith historical day and the kth meteorological factor of the day to be predicted as formula (3); k is 1, 2, 3, 4;
Figure BDA0002571045670000045
in the above formula, rho is 0.5; x is the number of0kThe kth meteorological factor, x, representing the day to be predictedikA kth meteorological factor representing an ith historical day;
calculating the grey correlation degrees of temperature, humidity, radiance, cloud amount and power respectively according to the formula; the grey correlation degree of the day to be predicted and the ith day is expressed as formula (4);
Figure BDA0002571045670000046
in the above formula ωkWeight of kth meteorological factor, m number of meteorological factors, thetaiRepresenting the weighted grey relevance of the day to be predicted and the ith day, wherein
Figure BDA0002571045670000047
ωk＞0；m＝4；
S23: according to thetaiAnd arranging the historical days from large to small, and taking the previous five days as the historical day most similar to the day to be measured.
Alternatively, ωkThe calculation formula is as follows:
Figure BDA0002571045670000048
wherein:
Figure BDA0002571045670000049
in the above formula vkCoefficient of variation, σ, representing the kth meteorological factorkIs the standard deviation of the kth meteorological factor,
Figure BDA00025710456700000410
is the average value of the kth meteorological factor, and m represents the number of meteorological factors.
Optionally, the step S3 further includes:
s31: optimizing a prediction model of a least square support vector machine by using an improved ant lion algorithm;
s32: the most similar historical day sample obtained in the step S2 is used as a training set, meteorological information of the training set sample is used as input in the prediction model, photovoltaic power generation power is used as output, and the root mean square error is used as a return value, so that the optimal regularization parameter and the optimal nuclear parameter are obtained;
s33: and predicting the photovoltaic output power of the day to be predicted by taking the meteorological information of the day sample to be predicted as input through the trained model.
Optionally, a Cauchy mutation operator is introduced into the ant lion algorithm, and the formula is
Figure BDA0002571045670000051
XiIs the solution at the current time, Xi+1For the solution after mutation, x is a random number following the Cauchy distribution.
Optionally, step S31 further includes:
s311: setting basic parameters including maximum iteration times, the number of ants and ant lions and variable dimensions;
s312: initializing the positions of ants and the ant lions, calculating the fitness value at the moment according to an LSSVM model, recording the optimal value, and selecting the ant lions with the best fitness value as elite ant lions individuals;
s313: optimized selection of ant lions by roulette searchBy using the formula
Figure BDA0002571045670000052
Formula (II)
Figure BDA0002571045670000053
Updating
Figure BDA0002571045670000054
C in the above formula, realizes the random walk of ants around the selected elite ant liontIs the minimum of all variables in the t-th iteration, dtRepresenting the maximum value that contains all variables in the t-th iteration,
Figure BDA0002571045670000059
is the position of the jth ant lion in the tth iteration;
the ants are slowly close to the ant lion, and then the formula is adopted
Figure BDA0002571045670000055
Position of updated ant, in the above formula
Figure BDA00025710456700000510
Is the ants randomly wandering around the ant lion selected by the roulette in the t-th iteration,
Figure BDA0002571045670000056
ants that randomly walked around the elite ant lion in the t iteration;
s314: after the ants are captured, according to the following formula:
Figure BDA0002571045670000057
comparing with the best Elite lion at the current position, readjusting the position of the best Elite lion to update the optimal value, in the above formula
Figure BDA0002571045670000058
The position of the ith ant at the t iteration is shown;
s315: performing Cauchy variation, calculating the fitness at the moment by using an LSSVM model, comparing the fitness with the fitness before non-variation, and selecting the optimal fitness between the fitness and the fitness;
s316: and judging whether the end permission condition is reached, if the end is reached, returning to the step S313 to continue the circulation if the end permission condition is not reached.
Compared with the prior art, the invention has the following beneficial effects:
the accuracy of photovoltaic power generation power prediction is increased.
The prediction precision is slightly influenced by the built model, extreme abnormal weather conditions, environments and slow changes of photovoltaic module parameters along with time can be simulated, and the model is high in anti-interference capability and robustness;
when the cloud cluster form and the moving speed are changed violently, an accurate prediction result can be obtained.
The power value error obtained when prediction is carried out is small.
Drawings
Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:
fig. 1 is a flow chart of a method for short-term photovoltaic power prediction according to an embodiment of the present invention;
FIG. 2 is a flowchart of step S3 according to an embodiment of the present invention;
FIG. 3 is a flowchart of step S31 according to an embodiment of the present invention;
FIG. 4 is a diagram illustrating predicted results of different prediction models according to an embodiment of the present invention.
Detailed Description
The present invention will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the invention, but are not intended to limit the invention in any way. It should be noted that it would be obvious to those skilled in the art that various changes and modifications can be made without departing from the spirit of the invention. All falling within the scope of the present invention.
Referring to fig. 1 to 3, the present embodiment discloses a short-term photovoltaic power prediction method, including the following steps:
s1: collecting historical data of a local photovoltaic power plant as a data sample, the historical data comprising: date, weather information for each historical day within a historical time period, the weather information comprising: weather type, meteorological factors; the meteorological factors include: temperature value, humidity value, radiation value and cloud cluster value at each moment of the day; randomly selecting one day from historical days as a day to be predicted;
s2: judging the weather type of the weather of the day to be predicted, and selecting the historical day most similar to the day to be predicted from the data samples of the same weather type by combining the weighted gray correlation degree;
s3: and (4) optimizing and predicting the photovoltaic power by adopting the regularization parameter and the kernel parameter of the improved ant lion algorithm optimized least squares support vector machine with the most similar historical day in the step S2 as the background.
In this embodiment, the selected historical time is one year; the historical data can be divided into a training set and a test set, wherein the day to be predicted serves as a test set sample, and the selected historical day most similar to the day to be predicted serves as a training set sample. In this embodiment, weather information is recorded at intervals of 15min from 7:00 am to 19:00 pm on each historical day. Therefore, the term "time" referred to in the step means every 15 minutes.
Wherein the weather types comprise sunny days, rainy days and cloudy days. In this embodiment, the weather type of the historical day and the weather type of the day to be predicted may be found through 2345 weather king or weather forecast, or may be determined in a fuzzy clustering manner.
Wherein the step S2 further includes:
s21: judging the weather type of the weather of the day to be predicted, and correspondingly listing the information of s historical days in the data sample belonging to the same weather type; s is a positive integer; one history day of the s history days is recorded as the ith history day, i is 1, … …, s;
s22: calculating the grey correlation degree of the day to be predicted and the ith historical day by using the formula (4);
the process is as follows:
the weather information is expressed by the formula (2)
Ti＝[xi1，xi2，xi3，xi4] (2)
In the above equation T represents a data sample meteorological information set,
Figure BDA0002571045670000071
is the value of the temperature at each moment in time,
Figure BDA0002571045670000072
represents the value of the humidity at each moment in time,
Figure BDA0002571045670000073
representing the radiance value at each time instant,
Figure BDA0002571045670000074
representing the cloud cluster magnitude at each moment, n representing the dimension of the meteorological factor;
calculating the gray correlation degree of the kth meteorological factor of the ith historical day and the kth meteorological factor of the day to be predicted as formula (3); k is 1, 2, 3, 4;
Figure BDA0002571045670000075
in the above formula, rho is 0.5; x is the number of0kThe kth meteorological factor, x, representing the day to be predictedikIndicating the kth meteorological factor for the ith historical day.
Calculating the grey correlation degrees of temperature, humidity, radiance, cloud amount and power respectively according to the formula; the grey correlation degree of the day to be predicted and the ith day is expressed as formula (4);
Figure BDA0002571045670000076
in the above formula ωkRepresents the k < th >Weight of meteorological factors, m represents the number of meteorological factors, θiRepresenting the weighted grey relevance of the day to be predicted and the ith day, wherein
Figure BDA0002571045670000077
ωk＞0；m＝4；
ωkThe calculation formula is as follows:
Figure BDA0002571045670000081
wherein:
Figure BDA0002571045670000082
in the above formula vkCoefficient of variation, σ, representing the kth meteorological factorkIs the standard deviation of the kth meteorological factor,
Figure BDA0002571045670000083
is the average value of the kth meteorological factor, and m represents the number of meteorological factors.
S23: according to thetaiAnd arranging the historical days from large to small, and taking the previous five days as the historical day most similar to the day to be measured.
Wherein the step S3 further includes:
s31: optimizing a prediction model of a least square support vector machine by using an improved ant lion algorithm;
s32: the most similar historical day sample obtained in the step S2 is used as a training set, meteorological information of the training set sample is used as input in the prediction model, photovoltaic power generation power is used as output, and the root mean square error is used as a return value, so that the optimal regularization parameter and the optimal nuclear parameter are obtained;
s33: and predicting the photovoltaic output power of the day to be predicted by taking the meteorological information of the day sample to be predicted as input through the trained model.
Wherein, the step S31 further includes:
s311: setting basic parameters including maximum iteration times, the number of ants and ant lions and variable dimensions;
s312: initializing the positions of ants and the ant lions, calculating the fitness value at the moment according to an LSSVM model, recording the optimal value, and selecting the ant lions with the best fitness value as elite ant lions individuals;
s313: optimized selection of ant lions by roulette search, exploitation
Figure BDA0002571045670000084
Formula (II)
Figure BDA0002571045670000085
Updating
Figure BDA0002571045670000086
C in the above formula, realizes the random walk of ants around the selected elite ant liontIs the minimum of all variables in the t-th iteration, dtRepresenting the maximum value that contains all variables in the t-th iteration,
Figure BDA00025710456700000810
is the position of the jth ant lion in the tth iteration;
the ants are slowly close to the ant lion, and then the formula is adopted
Figure BDA0002571045670000087
Position of updated ant, in the above formula
Figure BDA00025710456700000811
Is the ants randomly wandering around the ant lion selected by the roulette in the t-th iteration,
Figure BDA0002571045670000088
ants that randomly walked around the elite ant lion in the t iteration;
s314: after the ants are captured, according to the following formula:
Figure BDA0002571045670000089
comparing with the best Elite lion at the current position, readjusting the position of the best Elite lion to update the optimal value, in the above formula
Figure BDA0002571045670000091
The position of the ith ant at the t iteration is shown;
s315: performing Cauchy variation, calculating the fitness at the moment by using an LSSVM model, comparing the fitness with the fitness before non-variation, and selecting the optimal fitness between the fitness and the fitness;
s316: and judging whether the end permission condition is reached, if the end is reached, returning to the step S313 to continue the circulation if the end permission condition is not reached.
The improvement point of improving the ant lion algorithm is as follows: introducing a Cauchy mutation operator into the ant lion algorithm, wherein the formula is
Figure BDA0002571045670000092
XiIs the solution at the current time, Xi+1For the solution after mutation, x is a random number following the Cauchy distribution. If the fitness value is better than that before the Cauchy variation after the Cauchy variation in the algorithm execution process, the original solution is replaced by the solution at the moment.
As shown in fig. 4, the inventor divides prediction samples into sunny days, cloudy days and rainy days, then finds an optimal similar day in each weather type, and predicts the optimal similar day, and compares prediction results with an optimized ant lion algorithm LSSVM model (ALO-LSSVM) which is not improved and an optimized ant lion algorithm LSSVM model (unweighted IALO-LSSVM) which is not weighted to extract the similar day, so as to prove the superiority of the method (IALO-LSSVM).
As shown in table 1, the inventors used the mean absolute error percentage MAPE and the root mean square error RMSE as evaluation indexes, and calculated the results as follows:
TABLE 1 MAPE and RMSE for different prediction models
Figure BDA0002571045670000093
From the above table, it can be seen that the result of prediction by using the IALO-LSSVM model on the similar day extracted by using the weighted gray correlation is the best, and the absolute error percentage and the root mean square error are both very small, thus proving that the method is very superior.
The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes or modifications may be made by one skilled in the art within the scope of the appended claims without departing from the spirit of the invention. The embodiments and features of the embodiments of the present application may be combined with each other arbitrarily without conflict.

Claims (7)

1. A photovoltaic power short-term prediction method is characterized by comprising the following steps:
s1: collecting historical data of a local photovoltaic power plant as a data sample, the historical data comprising: date, weather information for each historical day within a historical time period, the weather information comprising: weather type, meteorological factors; the meteorological factors include: temperature value, humidity value, radiation value and cloud cluster value at each moment of the day; randomly selecting one day from historical days as a day to be predicted;
s2: judging the weather type of the weather of the day to be predicted, and selecting the historical day most similar to the day to be predicted from the data samples of the same weather type by combining the weighted gray correlation degree;
s3: and (4) optimizing and predicting the photovoltaic power by adopting the regularization parameter and the kernel parameter of the improved ant lion algorithm optimized least squares support vector machine with the most similar historical day in the step S2 as the background.
2. The method of claim 1, wherein the weather type comprises sunny, rainy, cloudy.
3. The method of claim 2, wherein the step S2 further comprises:
s21: judging the weather type of the weather of the day to be predicted, and correspondingly listing the information of s historical days in the data sample belonging to the same weather type; s is a positive integer; one history day of the s history days is recorded as the ith history day, i is 1, … …, s;
s22: calculating the grey correlation degree of the day to be predicted and the ith historical day by using the formula (4);
the process is as follows:
the weather information is expressed by the formula (2)
Ti＝[xi1，xi2，xi3，xi4] (2)
In the above equation T represents a data sample meteorological information set,
Figure FDA0002571045660000012
is the value of the temperature at each moment in time,
Figure FDA0002571045660000013
represents the value of the humidity at each moment in time,
Figure FDA0002571045660000014
representing the radiance value at each time instant,
Figure FDA0002571045660000015
representing the cloud cluster magnitude at each moment, n representing the dimension of the meteorological factor;
calculating the gray correlation degree of the kth meteorological factor of the ith historical day and the kth meteorological factor of the day to be predicted as formula (3); k is 1, 2, 3, 4;
Figure FDA0002571045660000011
in the above formula, rho is 0.5; x is the number of0kThe kth meteorological factor, x, representing the day to be predictedikA kth meteorological factor representing an ith historical day;
calculating the grey correlation degrees of temperature, humidity, radiance, cloud amount and power respectively according to the formula; the grey correlation degree of the day to be predicted and the ith day is expressed as formula (4);
Figure FDA0002571045660000021
in the above formula ωkWeight of kth meteorological factor, m number of meteorological factors, thetaiRepresenting the weighted grey relevance of the day to be predicted and the ith day, wherein
Figure FDA0002571045660000022
S23: according to thetaiAnd arranging the historical days from large to small, and taking the previous five days as the historical day most similar to the day to be measured.
4. The method of claim 3, wherein ω is ωkThe calculation formula is as follows:
Figure FDA0002571045660000023
wherein:
Figure FDA0002571045660000024
in the above formula vkCoefficient of variation, σ, representing the kth meteorological factorkIs the standard deviation of the kth meteorological factor,
Figure FDA0002571045660000025
is the average value of the kth meteorological factor, and m represents the number of meteorological factors.
5. The method of claim 1, wherein the step S3 further comprises:
s31: optimizing a prediction model of a least square support vector machine by using an improved ant lion algorithm;
s32: the most similar historical day sample obtained in the step S2 is used as a training set, meteorological information of the training set sample is used as input in the prediction model, photovoltaic power generation power is used as output, and the root mean square error is used as a return value, so that the optimal regularization parameter and the optimal nuclear parameter are obtained;
s33: and predicting the photovoltaic output power of the day to be predicted by taking the meteorological information of the day sample to be predicted as input through the trained model.
6. The method of claim 1,
introducing a Cauchy mutation operator into the ant lion algorithm, wherein the formula is
Figure FDA0002571045660000026
XiIs the solution at the current time, Xi+1For the solution after mutation, x is a random number following the Cauchy distribution.
7. The method of claim 5, wherein step S31 further comprises:
s311: setting basic parameters including maximum iteration times, the number of ants and ant lions and variable dimensions;
s312: initializing the positions of ants and the ant lions, calculating the fitness value at the moment according to an LSSVM model, recording the optimal value, and selecting the ant lions with the best fitness value as elite ant lions individuals;
s313: optimized selection of ant lions by roulette search, exploitation
Figure FDA0002571045660000033
Formula (II)
Figure FDA0002571045660000034
Updating
Figure FDA0002571045660000035
C in the above formula, realizes the random walk of ants around the selected elite ant liont is the minimum value of all variables in the t-th iteration, dt represents the maximum value of all variables in the t-th iteration,
Figure FDA0002571045660000036
is the position of the jth ant lion in the tth iteration;
the ants are slowly close to the ant lion, and then the formula is adopted
Figure FDA0002571045660000031
Position of updated ant, in the above formula
Figure FDA0002571045660000037
Is the ants randomly wandering around the ant lion selected by the roulette in the t-th iteration,
Figure FDA0002571045660000038
ants that randomly walked around the elite ant lion in the t iteration;
s314: after the ants are captured, according to the following formula:
Figure FDA0002571045660000032
comparing with the best Elite lion at the current position, readjusting the position of the best Elite lion to update the optimal value, in the above formula
Figure FDA0002571045660000039
The position of the ith ant at the t iteration is shown;
s315: performing Cauchy variation, calculating the fitness at the moment by using an LSSVM model, comparing the fitness with the fitness before non-variation, and selecting the optimal fitness between the fitness and the fitness;
s316: and judging whether the end permission condition is reached, if the end is reached, returning to the step S313 to continue the circulation if the end permission condition is not reached.
CN202010644968.3A 2020-07-06 2020-07-06 Photovoltaic power short-term prediction method Pending CN111985678A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010644968.3A CN111985678A (en) 2020-07-06 2020-07-06 Photovoltaic power short-term prediction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010644968.3A CN111985678A (en) 2020-07-06 2020-07-06 Photovoltaic power short-term prediction method

Publications (1)

Publication Number Publication Date
CN111985678A true CN111985678A (en) 2020-11-24

Family

ID=73438429

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010644968.3A Pending CN111985678A (en) 2020-07-06 2020-07-06 Photovoltaic power short-term prediction method

Country Status (1)

Country Link
CN (1) CN111985678A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112865075A (en) * 2021-01-12 2021-05-28 许继集团有限公司 Optimization method for alternating current-direct current hybrid micro-grid
CN113504582A (en) * 2021-06-24 2021-10-15 中国气象局公共气象服务中心(国家预警信息发布中心) Photovoltaic power station solar radiation short-term forecasting method based on satellite radiation product
CN114064755A (en) * 2021-11-15 2022-02-18 珠海格力电器股份有限公司 Method, device and equipment for estimating solar radiation quantity and storage medium
CN114362137A (en) * 2021-11-23 2022-04-15 国网浙江省电力有限公司嘉兴供电公司 New energy cluster power prediction method
CN116611004A (en) * 2023-07-20 2023-08-18 江苏中矿安华科技发展有限公司 Coal mine safety disaster prediction model training method, early warning method and early warning system

Citations (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105447509A (en) * 2015-11-10 2016-03-30 天津大学 Short-term power prediction method for photovoltaic power generation system
CN106815655A (en) * 2016-12-26 2017-06-09 浙江工业大学 Photovoltaic based on fuzzy rule is exerted oneself 2D interval prediction methods
JP2017117356A (en) * 2015-12-25 2017-06-29 株式会社東芝 Imbalance price prediction device, method, and program
CN107274038A (en) * 2017-07-31 2017-10-20 上海电力学院 A kind of LSSVM Prediction of annual electricity consumption methods optimized based on ant lion
CN107798199A (en) * 2017-11-09 2018-03-13 华中科技大学 A kind of Hydropower Unit parameter closed-loop identification method
CN108416366A (en) * 2018-02-06 2018-08-17 武汉大学 A kind of power-system short-term load forecasting method of the weighting LS-SVM based on Meteorological Index
CN108564192A (en) * 2017-12-29 2018-09-21 河海大学 A kind of short-term photovoltaic power prediction technique based on meteorological factor weight similar day
CN108932557A (en) * 2018-04-28 2018-12-04 云南电网有限责任公司临沧供电局 A kind of Short-term Load Forecasting Model based on temperature cumulative effect and grey relational grade
CN109165774A (en) * 2018-08-07 2019-01-08 燕山大学 A kind of short-term photovoltaic power prediction technique
CN109685257A (en) * 2018-12-13 2019-04-26 国网青海省电力公司 A kind of photovoltaic power generation power prediction method based on Support vector regression
CN109740052A (en) * 2018-12-26 2019-05-10 武汉大学 The construction method and device of network behavior prediction model, network behavior prediction technique
CN109767036A (en) * 2018-12-28 2019-05-17 北京航空航天大学 Support vector machines failure prediction method based on the optimization of adaptive ant lion
CN109858673A (en) * 2018-12-27 2019-06-07 南京工程学院 A kind of photovoltaic generating system power forecasting method
CN110297851A (en) * 2019-05-31 2019-10-01 国网上海市电力公司 A kind of improvement relevance acquisition methods of electric load
CN110414788A (en) * 2019-06-25 2019-11-05 国网上海市电力公司 A kind of power quality prediction technique based on similar day and improvement LSTM
CN111027759A (en) * 2019-11-28 2020-04-17 湖北工业大学 Ant-lion algorithm-based urban traffic emergency scheduling method
CN111079980A (en) * 2019-11-22 2020-04-28 天合云能源互联网技术(杭州)有限公司 Optical power prediction method based on self-adaptive classification strategy and hybrid optimization SVR
CN111260126A (en) * 2020-01-13 2020-06-09 燕山大学 Short-term photovoltaic power generation prediction method considering correlation degree of weather and meteorological factors
CN111340273A (en) * 2020-02-17 2020-06-26 南京邮电大学 Short-term load prediction method for power system based on GEP parameter optimization XGboost

Patent Citations (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105447509A (en) * 2015-11-10 2016-03-30 天津大学 Short-term power prediction method for photovoltaic power generation system
JP2017117356A (en) * 2015-12-25 2017-06-29 株式会社東芝 Imbalance price prediction device, method, and program
CN106815655A (en) * 2016-12-26 2017-06-09 浙江工业大学 Photovoltaic based on fuzzy rule is exerted oneself 2D interval prediction methods
CN107274038A (en) * 2017-07-31 2017-10-20 上海电力学院 A kind of LSSVM Prediction of annual electricity consumption methods optimized based on ant lion
CN107798199A (en) * 2017-11-09 2018-03-13 华中科技大学 A kind of Hydropower Unit parameter closed-loop identification method
CN108564192A (en) * 2017-12-29 2018-09-21 河海大学 A kind of short-term photovoltaic power prediction technique based on meteorological factor weight similar day
CN108416366A (en) * 2018-02-06 2018-08-17 武汉大学 A kind of power-system short-term load forecasting method of the weighting LS-SVM based on Meteorological Index
CN108932557A (en) * 2018-04-28 2018-12-04 云南电网有限责任公司临沧供电局 A kind of Short-term Load Forecasting Model based on temperature cumulative effect and grey relational grade
CN109165774A (en) * 2018-08-07 2019-01-08 燕山大学 A kind of short-term photovoltaic power prediction technique
CN109685257A (en) * 2018-12-13 2019-04-26 国网青海省电力公司 A kind of photovoltaic power generation power prediction method based on Support vector regression
CN109740052A (en) * 2018-12-26 2019-05-10 武汉大学 The construction method and device of network behavior prediction model, network behavior prediction technique
CN109858673A (en) * 2018-12-27 2019-06-07 南京工程学院 A kind of photovoltaic generating system power forecasting method
CN109767036A (en) * 2018-12-28 2019-05-17 北京航空航天大学 Support vector machines failure prediction method based on the optimization of adaptive ant lion
CN110297851A (en) * 2019-05-31 2019-10-01 国网上海市电力公司 A kind of improvement relevance acquisition methods of electric load
CN110414788A (en) * 2019-06-25 2019-11-05 国网上海市电力公司 A kind of power quality prediction technique based on similar day and improvement LSTM
CN111079980A (en) * 2019-11-22 2020-04-28 天合云能源互联网技术(杭州)有限公司 Optical power prediction method based on self-adaptive classification strategy and hybrid optimization SVR
CN111027759A (en) * 2019-11-28 2020-04-17 湖北工业大学 Ant-lion algorithm-based urban traffic emergency scheduling method
CN111260126A (en) * 2020-01-13 2020-06-09 燕山大学 Short-term photovoltaic power generation prediction method considering correlation degree of weather and meteorological factors
CN111340273A (en) * 2020-02-17 2020-06-26 南京邮电大学 Short-term load prediction method for power system based on GEP parameter optimization XGboost

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
WANG, JP ； WANG, X ；YANG, XW ；YU, SZ: "《The Research of External Characteristic Model Based on Photovoltaic System》", 《CHINA INTERNATIONAL CONFERENCE ON ELECTRICITY DISTRIBUTION》 *
XIAO LI, XIN WANG*, YI HUI ZHENG, LI XUE LI, LI DAN ZHOU, XI KU: "《Short-Term Wind Power Forecasting Based on Least-Square Support Vector Machine (LSSVM)》", 《APPLIED MECHANICS AND MATERIALS》 *
于建芳；刘升；韩斐斐；肖子雅: "《基于柯西变异的蚁狮优化算法》", 《微电子学与计算机》 *
李 霄；王 昕；郑益慧；李立学；生西奎；吴 昊: "《基于改进最小二乘支持向量机和预测误差校正的》", 《电力系统保护与控制》 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112865075A (en) * 2021-01-12 2021-05-28 许继集团有限公司 Optimization method for alternating current-direct current hybrid micro-grid
CN112865075B (en) * 2021-01-12 2023-10-20 许继集团有限公司 AC/DC hybrid micro-grid optimization method
CN113504582A (en) * 2021-06-24 2021-10-15 中国气象局公共气象服务中心(国家预警信息发布中心) Photovoltaic power station solar radiation short-term forecasting method based on satellite radiation product
CN113504582B (en) * 2021-06-24 2021-12-07 中国气象局公共气象服务中心(国家预警信息发布中心) Photovoltaic power station solar radiation short-term forecasting method based on satellite radiation product
CN114064755A (en) * 2021-11-15 2022-02-18 珠海格力电器股份有限公司 Method, device and equipment for estimating solar radiation quantity and storage medium
CN114362137A (en) * 2021-11-23 2022-04-15 国网浙江省电力有限公司嘉兴供电公司 New energy cluster power prediction method
CN116611004A (en) * 2023-07-20 2023-08-18 江苏中矿安华科技发展有限公司 Coal mine safety disaster prediction model training method, early warning method and early warning system
CN116611004B (en) * 2023-07-20 2023-11-10 江苏中矿安华科技发展有限公司 Coal mine safety disaster prediction model training method, early warning method and early warning system

Similar Documents

Publication Publication Date Title
Huang et al. Multiple-input deep convolutional neural network model for short-term photovoltaic power forecasting
Wang et al. Exploring key weather factors from analytical modeling toward improved solar power forecasting
CN111985678A (en) Photovoltaic power short-term prediction method
Liu et al. An improved photovoltaic power forecasting model with the assistance of aerosol index data
Yang et al. Classification and summarization of solar irradiance and power forecasting methods: A thorough review
CN113919232A (en) Photovoltaic power station power prediction method and system based on recurrent neural network
Liu et al. Solar forecasting by K-Nearest Neighbors method with weather classification and physical model
Alomari et al. A predictive model for solar photovoltaic power using the Levenberg-Marquardt and Bayesian regularization algorithms and real-time weather data
CN107562992B (en) Photovoltaic array maximum power tracking method based on SVM and particle swarm algorithm
Hmamou et al. Particle swarm optimization approach to determine all parameters of the photovoltaic cell
Kayri et al. Predicting power production from a photovoltaic panel through artificial neural networks using atmospheric indicators
Aliberti et al. Forecasting Short-term Solar Radiation for Photovoltaic Energy Predictions.
CN111259550A (en) Grid search and improved NM simplex algorithm-based photovoltaic model updating method
Gupta et al. Solar irradiation forecasting technologies: a review
Xiu-Yan et al. Global horizontal irradiance prediction model considering the effect of aerosol optical depth based on the Informer model
Ehsan et al. Artificial neural network predictor for grid-connected solar photovoltaic installations at atmospheric temperature
CN117236720A (en) Photovoltaic power station generated power prediction method utilizing multi-meteorological factor characteristics
Stoliarov Efficient electricity generation forecasting from solar power plants using technology: Integration, benefits and prospects
Zhang et al. Graph neural network based short-term solar irradiance forcasting model considering surrounding meteorological factors
Cahyadi et al. Solar prediction strategy for managing virtual power stations
Guo et al. Short-Term Photovoltaic Power Forcasting Based on PSO-BP Neural Network
Jogunuri et al. Artificial intelligence methods for solar forecasting for optimum sizing of PV systems: A review
Rao et al. Bright sunshine duration index-based prediction of solar PV power using ANN approach
Zhou et al. PV power characteristic modeling based on multi-scale clustering and its application in generation prediction
Matera et al. Time-dependent photovoltaic performance assessment on a global scale using artificial neural networks

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20201124