CN114266416A - Photovoltaic power generation power short-term prediction method and device based on similar days and storage medium - Google Patents
Photovoltaic power generation power short-term prediction method and device based on similar days and storage medium Download PDFInfo
- Publication number
- CN114266416A CN114266416A CN202111654839.3A CN202111654839A CN114266416A CN 114266416 A CN114266416 A CN 114266416A CN 202111654839 A CN202111654839 A CN 202111654839A CN 114266416 A CN114266416 A CN 114266416A
- Authority
- CN
- China
- Prior art keywords
- power generation
- day
- short
- historical
- photovoltaic power
- 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
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a photovoltaic power generation short-term prediction method, a photovoltaic power generation short-term prediction device and a storage medium based on similar days. The method can effectively predict the short-term power of the photovoltaic power generation under different meteorological conditions, and provides a basis for the development and research work of the photovoltaic power generation grid connection and the stable operation of the power station.
Description
Technical Field
The invention relates to the technical field of photovoltaic power generation power prediction, in particular to a photovoltaic power generation power short-term prediction method and device based on similar days and a storage medium.
Background
After the new century, environmental protection and energy crisis become more severe, and solar power generation is widely applied in the global scope by virtue of the advantages of cleanness, renewability, easiness in distributed popularization and the like. With the proposal of carbon peak reaching and carbon neutralization strategies, the photovoltaic power generation industry in China is about to develop a new good opportunity, and the proportion of the photovoltaic power generation to the total power generation is obviously increased. However, the photovoltaic power generation power is closely related to meteorological factors, and has strong intermittence, fluctuation and randomness, and large-scale photovoltaic power generation grid connection not only causes strong impact on a power grid but also influences the quality of electric energy, and also causes adverse effects on stable operation and effective scheduling of a power system. The reasonable prediction of the photovoltaic power generation power can effectively reduce the impact on a power grid caused by grid connection, improve the stability of the operation of a power station and improve the photovoltaic receiving capacity of a main grid. Therefore, the photovoltaic power generation power can be effectively predicted, and the method has important significance for promoting the solar power generation application and the safe, efficient and economic operation of the whole power grid.
The existing photovoltaic power generation power prediction methods are mainly divided into a physical method and a statistical method. The physical method is widely applied to prediction of photovoltaic power generation power due to simple model and easily-obtained input information, but parameters of photovoltaic power generation equipment need to be considered, so that the stability, reliability and anti-interference capability of the method are poor. Traditional statistical methods, while simpler and more accurate than physical methods, are limited by variable feature assumptions and nonlinear models. In recent years, a deep learning neural network algorithm with nonlinear processing capability is widely applied to photovoltaic power generation power prediction, but the prediction capability of the algorithm under the condition of sudden change is poor, and the algorithm is easy to fall into local optimization. Therefore, short-term accurate photovoltaic power generation power prediction techniques are also few.
Disclosure of Invention
The invention provides a photovoltaic power generation short-term prediction method, a photovoltaic power generation short-term prediction device and a storage medium based on similar days, and aims to solve the problem that the conventional photovoltaic power generation short-term prediction method is low in estimation result precision.
In a first aspect, a photovoltaic power generation power short-term prediction method based on similar days is provided, and includes:
acquiring weather information of a forecast day, and determining the weather type of the weather information;
selecting A pieces of historical power generation day data with the maximum comprehensive similarity from a historical power generation day database according to the predicted weather information and weather types to construct a similar day sample set, wherein A is a preset value;
decomposing historical photovoltaic power generation power data of each historical power generation day in a similar day sample set, and superposing IMF components which are obtained by decomposition and represent short-term rules as short-term prediction intermediate frequency information; short-term prediction intermediate frequency information corresponding to the historical power generation day with the highest comprehensive similarity is used as short-term prediction intermediate frequency information of the prediction day;
carrying out normalization processing on meteorological information, short-term predicted intermediate frequency information and historical photovoltaic power generation power data of a sample set on a similar day;
taking meteorological information and short-term prediction intermediate frequency information of each historical power generation day in the similar day sample set after normalization processing as input of a neural network, taking corresponding historical photovoltaic power generation power data as output of the neural network, and training the neural network to obtain a photovoltaic power generation power short-term prediction model;
performing normalization processing on meteorological information corresponding to the forecast day and short-term forecast intermediate frequency information of the meteorological information, inputting the meteorological information and the short-term forecast intermediate frequency information into a photovoltaic power generation power short-term forecast model, and outputting normalized photovoltaic power generation forecast power;
and performing inverse normalization processing on the photovoltaic power generation predicted power to obtain the predicted solar photovoltaic power generation short-term predicted power.
Further, the method for constructing the similar day sample set by selecting the A pieces of historical power generation day data with the maximum comprehensive similarity from the historical power generation day database according to the predicted weather information and weather types specifically comprises the following steps:
screening historical power generation day data of a weather type system from a historical power generation day database according to the predicted day weather type to construct a similar day initial sample set;
and calculating the comprehensive similarity between the meteorological information characteristic vector constructed based on the weather information of the predicted days and the meteorological information characteristic vector of each historical power generation day in the initial sample set of the similar days, and selecting A pieces of historical power generation day data with the maximum comprehensive similarity to construct the sample set of the similar days.
Further, the comprehensive similarity is calculated by the following formula:
Si＝αFi+(1-α)Dcosi
wherein S isiRepresenting the comprehensive similarity of the ith historical generation day and the prediction day; alpha represents a value of [0,1]Empirical weighting coefficients between; fiIndicating a gray degree of correlation between weather information on the ith historical generation day and the predicted day, DcosiExpressing the cosine similarity between the meteorological information of the ith historical generating day and the forecasting day;
wherein, FiThe calculation formula is as follows:
wherein M is the total number of components contained in the meteorological information; omegai(k) And expressing a gray correlation coefficient between the ith historical power generation day and the kth meteorological information component of the prediction day, wherein the calculation formula is as follows:
Dcosithe calculation formula is as follows:
Further, the decomposing processing is respectively performed on the historical photovoltaic power generation power data of each historical power generation day in the similar day sample set, and the IMF components representing the short-term law obtained through the decomposing processing are respectively superposed to be used as the short-term prediction intermediate frequency information, and the method specifically includes:
for each historical power generation day in the similar day sample set, decomposing historical photovoltaic power generation power data by using an EEMD algorithm to obtain a plurality of IMF components and a residual error component;
checking and calculating the maximum run length and the run number of each IMF component obtained by decomposing historical photovoltaic power generation power data by using a run length checking method, and screening out the intermediate-frequency IMF components according to the maximum run length and the run number;
and superposing the screened intermediate frequency IMF components as short-term predicted intermediate frequency information of the historical power generation day.
Further, before training the neural network, the method further includes:
determining the node numbers n, m and l of an input layer, an output layer and a hidden layer of the neural network, giving a training function, a node transfer function and a network learning function of the neural network, optimizing an initial weight and a threshold of the neural network by using a wolf optimization algorithm, and assigning the optimized values to the neural network.
Further, the neural network adopts a BP neural network, and the hidden layer and the output layer of the BP neural network can be calculated by the following formula:
hidden layer:
an output layer:
wherein, bjFor the jth node of the hidden layer, XiFor the i-th component of the input vector, wijIs the weight value between the ith node of the input layer and the jth node of the hidden layer, thetajThreshold for the jth node of the hidden layer, YkFor the kth node of the output layer, vjkIs the weight between the ith node of the hidden layer and the kth node of the output layer, gammakA threshold value of the kth node of the output layer; f. of1(. is a hidden layer node function, f2(. h) is an output layer node function;
the network error is calculated by the following formula:
Ek＝yk-Yk
wherein E iskError of k node of output layer, ykIs the true value, Y, of the kth node of the output layerkA predicted value of the kth node of an output layer;
the updating method of the weight and the threshold of the hidden layer and the output layer is as follows:
hidden layer:
an output layer:
wjk＝wjk+λbjEk
γk＝γk+Ek
wherein λ represents a network iteration speed;
in the iterative training processDetermining the network error E and the setting error epsilon1If E > ε1If yes, updating the weights and thresholds of the hidden layer and the output layer and returning to the computation of the hidden layer and the output layer; otherwise, terminating the iteration and outputting a predicted value.
Further, the node numbers of the input layer and the output layer of the BP neural network are determined according to input and output data during training, the node number of the hidden layer is jointly determined by the node numbers of the output layer and the input layer, and the optimal node number of the hidden layer is obtained by one of the following modes:
l＝n-1
in the formula, C is a regulating constant, and the value range is [1,10 ].
Further, the meteorological information includes daily average solar irradiance, maximum irradiance, average temperature, maximum temperature, minimum temperature, average humidity, and average wind speed.
In a second aspect, a photovoltaic power generation short-term prediction device based on similar days is provided, which includes:
the data acquisition module is used for acquiring meteorological information of a predicted day and determining the weather type of the meteorological information;
the similar day sample set construction module is used for selecting A pieces of historical power generation day data with the maximum comprehensive similarity from the historical power generation day database according to the forecast weather information and the weather type to construct a similar day sample set, wherein A is a preset value;
the intermediate frequency information acquisition module is used for respectively decomposing historical photovoltaic power generation power data of each historical power generation day in the similar day sample set and superposing IMF components which are obtained by decomposition and represent short-term rules as short-term prediction intermediate frequency information; short-term prediction intermediate frequency information corresponding to the historical power generation day with the highest comprehensive similarity is used as short-term prediction intermediate frequency information of the prediction day;
the normalization module is used for performing normalization processing on the meteorological information, the short-term predicted intermediate frequency information and the historical photovoltaic power generation power data of the similar day sample set;
the model training module is used for taking the meteorological information and the short-term prediction intermediate frequency information of each historical power generation day in the similar day sample set after normalization processing as the input of a neural network, taking the corresponding historical photovoltaic power generation power data as the output of the neural network, and training the neural network to obtain a photovoltaic power generation power short-term prediction model;
the photovoltaic power generation prediction module is used for normalizing the meteorological information corresponding to the prediction day and the short-term prediction intermediate frequency information of the meteorological information, inputting the meteorological information into the photovoltaic power generation power short-term prediction model and outputting normalized photovoltaic power generation prediction power;
and the inverse normalization module is used for carrying out inverse normalization processing on the photovoltaic power generation predicted power to obtain the predicted short-term solar photovoltaic power generation predicted power.
In a third aspect, a computer readable storage medium is provided, which stores a computer program that when executed by a processor implements a similar day based short term prediction method of photovoltaic generated power as described above.
Advantageous effects
The invention provides a photovoltaic power generation short-term prediction method, a photovoltaic power generation short-term prediction device and a storage medium based on similar days. The method can effectively predict the short-term power of the photovoltaic power generation under different meteorological conditions, and provides a basis for the development and research work of the photovoltaic power generation grid connection and the stable operation of the power station.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of a short-term photovoltaic power generation power prediction method according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail below. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the examples given herein without any inventive step, are within the scope of the present invention.
Example 1
As shown in fig. 1, the present embodiment provides a short-term photovoltaic power generation power prediction method based on similar days, including:
s1: and acquiring weather information of the predicted day, and determining the type of the weather.
During specific implementation, the temperature T, the wind speed W, the humidity H and the solar irradiation intensity I of a power station to be predicted can be measured and recorded by utilizing devices such as a thermocouple, a wind speed tester, a hygrometer and an irradiation meter, once relevant data can be collected and recorded according to a preset time interval, and the preset time interval is set according to actual needs, such as 5 minutes, 10 minutes, half an hour and the like. Daily average solar irradiance I can be extracted based on collected dataaMaximum irradiance ImaxAverage temperature TaMaximum temperature TmaxMinimum temperature TminAverage humidityDegree HaAnd the average wind speed WaMethod for constructing meteorological information feature vector X by using meteorological information*＝[Ia,Imax,Ta,Tmax,Tmin,Ha,Wa]Meanwhile, according to the generalized weather classification principle, the weather is classified into one of three categories, namely sunny weather, rainy weather and cloudy weather.
S2: and selecting A pieces of historical power generation day data with the maximum comprehensive similarity from the historical power generation day database according to the predicted weather information and weather types to construct a similar day sample set, wherein A is a preset value.
Wherein, the historical power generation day database can be established in advance: the method comprises the steps of constructing a meteorological information characteristic vector of each historical day by collecting a large amount of historical day meteorological data of a power station to be predicted and generating power data of each historical day in advance, determining a weather type of each historical day, and establishing a historical generating day database based on the meteorological characteristic vector, the weather type and the generating power data corresponding to each historical day.
Based on this, the specific process of step S2 includes:
s21: screening historical power generation day data of a weather type system from a historical power generation day database according to the predicted day weather type to construct a similar day initial sample set;
s22: and calculating the comprehensive similarity between the weather information characteristic vector constructed based on the weather information of the forecast day and the weather information characteristic vector of each historical power generation day in the initial sample set of the similar days, selecting A pieces of historical power generation day data with the maximum comprehensive similarity to construct the sample set of the similar days, and selecting the value of A according to actual needs, such as 10, 15, 20 and the like.
Wherein, the comprehensive similarity is calculated by the following formula:
Si＝αFi+(1-α)Dcosi
wherein S isiRepresenting the comprehensive similarity of the ith historical generation day and the prediction day; alpha represents a value of [0,1]The empirical weight coefficient between the weather conditions is combined with the specific weather conditions of the forecast day during calculation, and when the meteorological information such as irradiance, relative humidity, temperature and the like fluctuates severely, the weather conditions occurWhen the change is obvious, the value is close to 0, otherwise, the value is close to 1; fiRepresenting the gray degree of correlation between the weather information characteristic vectors of the ith historical generation day and the predicted day, DcosiExpressing the cosine similarity between meteorological information characteristic vectors of the ith historical generating day and the forecast day;
wherein, FiThe calculation formula is as follows:
wherein M is a meteorological information feature vector dimension, which is 7 in this embodiment; omegai(k) The grey correlation coefficient between the ith historical generation day and the kth meteorological information characteristic vector component of the prediction day is represented, and the calculation formula is as follows:
Dcosithe calculation formula is as follows:
wherein the content of the first and second substances,
a k-th component representing a feature vector of the predicted solar weather information,
the kth component of the ith historical generation solar weather information characteristic vector is represented, and rho represents the value of [0,1 ]]The resolution factor therebetween, ρ is 0.5 in this embodiment.
Comparing the association degree between the meteorological information characteristic vectors of the historical generating day and the predicted day by adopting a grey association degree analysis method, calculating the similarity of the change trend between the meteorological information characteristic vectors of the historical generating day and the predicted day by utilizing cosine similarity, wherein the closer the value is to 1, the more the value is, the more the meteorological information characteristic is indicatedThe smaller the included angle between the vectors is, the more similar the two meteorological information characteristic vectors are, and the more consistent the change trends (directions) of the two meteorological information characteristic vectors are. Finally, the cosine similarity D is calculatedcosiDegree of correlation with Gray FiWeighting to obtain a similarity index SiFurther reflecting the overall similarity of the meteorological information characteristic vectors of the predicted day and the historical generation day in two aspects of change trend and numerical value, SiThe closer the value is to 1, the more similar.
S3: decomposing historical photovoltaic power generation power data of each historical power generation day in a similar day sample set, and superposing IMF components which are obtained by decomposition and represent short-term rules as short-term prediction intermediate frequency information; and the short-term prediction intermediate frequency information corresponding to the historical power generation day with the highest comprehensive similarity is used as the short-term prediction intermediate frequency information of the prediction day.
More specifically, the method comprises the following steps:
s31: and decomposing the historical photovoltaic power generation power data of each historical power generation day in the similar day sample set by using an EEMD (ensemble empirical mode decomposition) algorithm to obtain a plurality of IMF components and a residual component. The specific process is as follows:
white Gaussian noise ei(t) adding the sequence into a historical photovoltaic power generation power data information sequence P (t), wherein the new sequence after adding the Gaussian white noise is as follows:
Pi(t)＝P(t)+ei(t)
applying empirical mode decomposition technique to new sequence Pi(t) decomposing to obtain N IMF components cj(t) and a residual component r0(t)：
Wherein, i represents that Gaussian white noise is added to the original signal for the ith time;
repeating the process H, adding new white Gaussian noise to the original historical photovoltaic power generation power data information sequence P (t) each time, and solving corresponding IMF components;
carrying out averaging processing on the IMF components obtained by calculation, so that the IMF components of the original historical photovoltaic power generation power data information sequence P (t) can be represented by the following equation:
and finally, decomposing the original historical photovoltaic power generation power data information sequence P (t) into:
s32: and (4) checking and calculating the maximum run length and the run number of each IMF component obtained by decomposing historical photovoltaic power generation data by using a run length checking method, and screening out the intermediate-frequency IMF component according to the maximum run length and the run number. The specific process is as follows:
for each IMF component cj(t), which are all a time series, are assumed to be represented by the series { G (t) } (t ═ 1,2, …, p), with the average value GaThen the timing symbol is defined as:
wherein i is 1,2, …, p;
consecutive 1 s or 0 s each represent a run, the number of 1 s or 0 s within a run representing the run length, the larger the maximum run length representing the more stable the data, the lower the frequency. And then, the intermediate frequency IMF component can be screened out according to the comparison between the maximum run length and the run number and a preset threshold value.
S33: and superposing the screened intermediate frequency IMF components as short-term predicted intermediate frequency information of the historical power generation day.
S4: and performing normalization processing on the meteorological information, the short-term predicted intermediate frequency information and the historical photovoltaic power generation power data of the similar day sample set, so that the calculation efficiency and the prediction precision of the prediction model are improved.
S5: and taking the meteorological information and short-term prediction intermediate frequency information of each historical power generation day in the similar day sample set after normalization processing as the input of a neural network, taking corresponding historical photovoltaic power generation power data as the output of the neural network, and training the neural network to obtain a photovoltaic power generation power short-term prediction model.
Before training the neural network, the method further comprises the following steps:
determining the node numbers n, m and l of an input layer, an output layer and a hidden layer of the neural network, giving a training function, a node transfer function and a network learning function of the neural network, optimizing an initial weight and a threshold of the neural network by using a wolf optimization algorithm, and assigning the optimized values to the neural network.
In this embodiment, the neural network is a BP neural network, and the hidden layer and the output layer thereof can be calculated by the following formula:
hidden layer:
an output layer:
wherein, bjFor the jth node of the hidden layer, XiFor the i-th component of the input vector, wijIs the weight value between the ith node of the input layer and the jth node of the hidden layer, thetajThreshold for the jth node of the hidden layer, YkFor the kth node of the output layer, vjkIs the weight between the ith node of the hidden layer and the kth node of the output layer, gammakA threshold value of the kth node of the output layer; f. of1(. is a hidden layer node function, f2(. h) is an output layer node function;
the network error is calculated by the following formula:
Ek＝yk-Yk
wherein E iskError of k node of output layer, ykTo be transportedTrue value of the kth node out of the layer, YkA predicted value of the kth node of an output layer;
the updating method of the weight and the threshold of the hidden layer and the output layer is as follows:
hidden layer:
an output layer:
wjk＝wjk+λbjEk
γk＝γk+Ek
wherein λ represents a network iteration speed;
in the iterative training process, the network error E and the set error epsilon are judged1If E > ε1If yes, updating the weights and thresholds of the hidden layer and the output layer and returning to the computation of the hidden layer and the output layer; otherwise, terminating the iteration and outputting a predicted value. And setting another iteration termination condition, namely reaching the maximum iteration training times.
The node number of the input layer and the output layer of the BP neural network is determined according to input and output data during training, the node number of the hidden layer is jointly determined by the node number of the output layer and the node number of the input layer, and the optimal node number of the hidden layer is obtained by one of the following modes:
l＝n-1
in the formula, C is a regulating constant, and the value range is [1,10 ].
S6: the meteorological information corresponding to the forecast day and the short-term forecast intermediate frequency information are normalized, then the normalized meteorological information and the short-term forecast intermediate frequency information are input into a photovoltaic power generation power short-term forecast model, and normalized photovoltaic power generation forecast power P is outputn。
S7: predicting power P for photovoltaic power generationnCarrying out inverse normalization processing to obtain short-term predicted power P of predicted solar photovoltaic power generationa。
Example 2
The embodiment provides a photovoltaic power generation short-term prediction device based on similar days, which comprises:
the data acquisition module is used for acquiring meteorological information of a predicted day and determining the weather type of the meteorological information;
the similar day sample set construction module is used for selecting A pieces of historical power generation day data with the maximum comprehensive similarity from the historical power generation day database according to the forecast weather information and the weather type to construct a similar day sample set, wherein A is a preset value;
the intermediate frequency information acquisition module is used for respectively decomposing historical photovoltaic power generation power data of each historical power generation day in the similar day sample set and superposing IMF components which are obtained by decomposition and represent short-term rules as short-term prediction intermediate frequency information; short-term prediction intermediate frequency information corresponding to the historical power generation day with the highest comprehensive similarity is used as short-term prediction intermediate frequency information of the prediction day;
the normalization module is used for performing normalization processing on the meteorological information, the short-term predicted intermediate frequency information and the historical photovoltaic power generation power data of the similar day sample set;
the model training module is used for taking the meteorological information and the short-term prediction intermediate frequency information of each historical power generation day in the similar day sample set after normalization processing as the input of a neural network, taking the corresponding historical photovoltaic power generation power data as the output of the neural network, and training the neural network to obtain a photovoltaic power generation power short-term prediction model;
the photovoltaic power generation prediction module is used for normalizing the meteorological information corresponding to the prediction day and the short-term prediction intermediate frequency information of the meteorological information, inputting the meteorological information into the photovoltaic power generation power short-term prediction model and outputting normalized photovoltaic power generation prediction power;
and the inverse normalization module is used for carrying out inverse normalization processing on the photovoltaic power generation predicted power to obtain the predicted short-term solar photovoltaic power generation predicted power.
Example 3
The present embodiment provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements a similar day-based photovoltaic generated power short-term prediction method as described in the above embodiments.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It is understood that the same or similar parts in the above embodiments may be mutually referred to, and the same or similar parts in other embodiments may be referred to for the content which is not described in detail in some embodiments.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (10)
1. A photovoltaic power generation power short-term prediction method based on similar days is characterized by comprising the following steps:
acquiring weather information of a forecast day, and determining the weather type of the weather information;
selecting A pieces of historical power generation day data with the maximum comprehensive similarity from a historical power generation day database according to the predicted weather information and weather types to construct a similar day sample set, wherein A is a preset value;
decomposing historical photovoltaic power generation power data of each historical power generation day in a similar day sample set, and superposing IMF components which are obtained by decomposition and represent short-term rules as short-term prediction intermediate frequency information; short-term prediction intermediate frequency information corresponding to the historical power generation day with the highest comprehensive similarity is used as short-term prediction intermediate frequency information of the prediction day;
carrying out normalization processing on meteorological information, short-term predicted intermediate frequency information and historical photovoltaic power generation power data of a sample set on a similar day;
taking meteorological information and short-term prediction intermediate frequency information of each historical power generation day in the similar day sample set after normalization processing as input of a neural network, taking corresponding historical photovoltaic power generation power data as output of the neural network, and training the neural network to obtain a photovoltaic power generation power short-term prediction model;
performing normalization processing on meteorological information corresponding to the forecast day and short-term forecast intermediate frequency information of the meteorological information, inputting the meteorological information and the short-term forecast intermediate frequency information into a photovoltaic power generation power short-term forecast model, and outputting normalized photovoltaic power generation forecast power;
and performing inverse normalization processing on the photovoltaic power generation predicted power to obtain the predicted solar photovoltaic power generation short-term predicted power.
2. The photovoltaic power generation short-term prediction method based on similar days according to claim 1, wherein the method for selecting a pieces of historical power generation day data with the maximum comprehensive similarity from the historical power generation day database according to the predicted daily weather information and weather types to construct a similar day sample set specifically comprises:
screening historical power generation day data of a weather type system from a historical power generation day database according to the predicted day weather type to construct a similar day initial sample set;
and calculating the comprehensive similarity between the meteorological information characteristic vector constructed based on the weather information of the predicted days and the meteorological information characteristic vector of each historical power generation day in the initial sample set of the similar days, and selecting A pieces of historical power generation day data with the maximum comprehensive similarity to construct the sample set of the similar days.
3. The similar-day-based short-term prediction method for photovoltaic power generation power as claimed in claim 1 or 2, wherein the comprehensive similarity is calculated by the following formula:
Si＝αFi+(1-α)Dcosi
wherein S isiRepresenting the comprehensive similarity of the ith historical generation day and the prediction day; alpha represents a value of [0,1]Empirical weighting coefficients between; fiIndicating a gray degree of correlation between weather information on the ith historical generation day and the predicted day, DcosiExpressing the cosine similarity between the meteorological information of the ith historical generating day and the forecasting day;
wherein, FiThe calculation formula is as follows:
wherein M is the total number of components contained in the meteorological information; omegai(k) And expressing a gray correlation coefficient between the ith historical power generation day and the kth meteorological information component of the prediction day, wherein the calculation formula is as follows:
Dcosithe calculation formula is as follows:
wherein the content of the first and second substances,
the kth component representing the predicted weather information,
the kth component of the ith historical generation solar weather information is represented, and rho represents the value of [0, 1%]The resolution factor between.
4. The photovoltaic power generation power short-term prediction method based on similar days as claimed in claim 1, wherein the decomposing process is performed on the historical photovoltaic power generation power data of each historical power generation day in the sample set of similar days, and the IMF components representing the short-term law obtained by the decomposing process are superimposed as the short-term prediction intermediate frequency information, specifically comprising:
for each historical power generation day in the similar day sample set, decomposing historical photovoltaic power generation power data by using an EEMD algorithm to obtain a plurality of IMF components and a residual error component;
checking and calculating the maximum run length and the run number of each IMF component obtained by decomposing historical photovoltaic power generation power data by using a run length checking method, and screening out the intermediate-frequency IMF components according to the maximum run length and the run number;
and superposing the screened intermediate frequency IMF components as short-term predicted intermediate frequency information of the historical power generation day.
5. The similar-day-based photovoltaic power generation short-term prediction method according to claim 1, wherein before training the neural network, the method further comprises:
determining the node numbers n, m and l of an input layer, an output layer and a hidden layer of the neural network, giving a training function, a node transfer function and a network learning function of the neural network, optimizing an initial weight and a threshold of the neural network by using a wolf optimization algorithm, and assigning the optimized values to the neural network.
6. The similar-day-based photovoltaic power generation short-term prediction method according to claim 5, wherein the neural network is a BP neural network, and the hidden layer and the output layer of the BP neural network can be calculated by the following formula:
hidden layer:
an output layer:
wherein, bjFor the jth node of the hidden layer, XiFor the i-th component of the input vector, wijIs the weight value between the ith node of the input layer and the jth node of the hidden layer, thetajThreshold for the jth node of the hidden layer, YkFor the kth node of the output layer, vjkIs the weight between the ith node of the hidden layer and the kth node of the output layer, gammakA threshold value of the kth node of the output layer; f. of1(. is a hidden layer node function, f2(. h) is an output layer node function;
the network error is calculated by the following formula:
Ek＝yk-Yk
wherein E iskError of k node of output layer, ykIs the true value, Y, of the kth node of the output layerkA predicted value of the kth node of an output layer;
the updating method of the weight and the threshold of the hidden layer and the output layer is as follows:
hidden layer:
an output layer:
wjk＝wjk+λbjEk
γk＝γk+Ek
wherein λ represents a network iteration speed;
in the iterative training process, the network error E and the set error epsilon are judged1If E > ε1If yes, updating the weights and thresholds of the hidden layer and the output layer and returning to the computation of the hidden layer and the output layer; otherwise, terminating the iteration and outputting a predicted value.
7. The photovoltaic power generation short-term prediction method based on similar days as claimed in claim 6, wherein the node numbers of the input layer and the output layer of the BP neural network are determined according to input and output data during training, the node number of the hidden layer is determined by the node numbers of the output layer and the input layer, and the optimal node number of the hidden layer is obtained by one of the following modes:
l＝n-1
in the formula, C is a regulating constant, and the value range is [1,10 ].
8. The similar-day-based short-term prediction method for photovoltaic power generation according to claim 1, wherein the meteorological information comprises daily average solar irradiance, maximum irradiance, average temperature, maximum temperature, minimum temperature, average humidity, and average wind speed.
9. A photovoltaic power generation short-term prediction device based on similar days is characterized by comprising the following components:
the data acquisition module is used for acquiring meteorological information of a predicted day and determining the weather type of the meteorological information;
the similar day sample set construction module is used for selecting A pieces of historical power generation day data with the maximum comprehensive similarity from the historical power generation day database according to the forecast weather information and the weather type to construct a similar day sample set, wherein A is a preset value;
the intermediate frequency information acquisition module is used for respectively decomposing historical photovoltaic power generation power data of each historical power generation day in the similar day sample set and superposing IMF components which are obtained by decomposition and represent short-term rules as short-term prediction intermediate frequency information; short-term prediction intermediate frequency information corresponding to the historical power generation day with the highest comprehensive similarity is used as short-term prediction intermediate frequency information of the prediction day;
the normalization module is used for performing normalization processing on the meteorological information, the short-term predicted intermediate frequency information and the historical photovoltaic power generation power data of the similar day sample set;
the model training module is used for taking the meteorological information and the short-term prediction intermediate frequency information of each historical power generation day in the similar day sample set after normalization processing as the input of a neural network, taking the corresponding historical photovoltaic power generation power data as the output of the neural network, and training the neural network to obtain a photovoltaic power generation power short-term prediction model;
the photovoltaic power generation prediction module is used for normalizing the meteorological information corresponding to the prediction day and the short-term prediction intermediate frequency information of the meteorological information, inputting the meteorological information into the photovoltaic power generation power short-term prediction model and outputting normalized photovoltaic power generation prediction power;
and the inverse normalization module is used for carrying out inverse normalization processing on the photovoltaic power generation predicted power to obtain the predicted short-term solar photovoltaic power generation predicted power.
10. A computer-readable storage medium, storing a computer program, wherein the computer program, when executed by a processor, implements a similar day based short term prediction method of photovoltaic generated power as claimed in any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111654839.3A CN114266416A (en) | 2021-12-30 | 2021-12-30 | Photovoltaic power generation power short-term prediction method and device based on similar days and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111654839.3A CN114266416A (en) | 2021-12-30 | 2021-12-30 | Photovoltaic power generation power short-term prediction method and device based on similar days and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114266416A true CN114266416A (en) | 2022-04-01 |
Family
ID=80831836
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111654839.3A Pending CN114266416A (en) | 2021-12-30 | 2021-12-30 | Photovoltaic power generation power short-term prediction method and device based on similar days and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114266416A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115759483A (en) * | 2023-01-06 | 2023-03-07 | 国能日新科技股份有限公司 | Photovoltaic electric field solar irradiance prediction method, electronic device and storage medium |
| CN117117859A (en) * | 2023-10-20 | 2023-11-24 | 华能新能源股份有限公司山西分公司 | Photovoltaic power generation power prediction method and system based on neural network |
| CN118095784A (en) * | 2024-04-17 | 2024-05-28 | 杭州城投建设有限公司 | Project management method, system and storage medium |
-
2021
- 2021-12-30 CN CN202111654839.3A patent/CN114266416A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115759483A (en) * | 2023-01-06 | 2023-03-07 | 国能日新科技股份有限公司 | Photovoltaic electric field solar irradiance prediction method, electronic device and storage medium |
| CN117117859A (en) * | 2023-10-20 | 2023-11-24 | 华能新能源股份有限公司山西分公司 | Photovoltaic power generation power prediction method and system based on neural network |
| CN117117859B (en) * | 2023-10-20 | 2024-01-30 | 华能新能源股份有限公司山西分公司 | Photovoltaic power generation power prediction method and system based on neural network |
| CN118095784A (en) * | 2024-04-17 | 2024-05-28 | 杭州城投建设有限公司 | Project management method, system and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| De Giorgi et al. | Assessment of the benefits of numerical weather predictions in wind power forecasting based on statistical methods | |
| CN110309603B (en) | Short-term wind speed prediction method and system based on wind speed characteristics | |
| CN114266416A (en) | Photovoltaic power generation power short-term prediction method and device based on similar days and storage medium | |
| Niu et al. | Uncertainty modeling for chaotic time series based on optimal multi-input multi-output architecture: Application to offshore wind speed | |
| CN111582551B (en) | Wind power plant short-term wind speed prediction method and system and electronic equipment | |
| CN109886464B (en) | Low-information-loss short-term wind speed prediction method based on optimized singular value decomposition generated feature set | |
| CN104573876A (en) | Wind power plant short-period wind speed prediction method based on time sequence long memory model | |
| CN109583588B (en) | Short-term wind speed prediction method and system | |
| CN114970952B (en) | Photovoltaic output short-term prediction method and system considering environmental factors | |
| CN113554466A (en) | Short-term power consumption prediction model construction method, prediction method and device | |
| CN111553510A (en) | Short-term wind speed prediction method | |
| CN113159102B (en) | Multi-time-scale photovoltaic power prediction method and system | |
| CN112434891A (en) | Method for predicting solar irradiance time sequence based on WCNN-ALSTM | |
| CN113052469A (en) | Method for calculating wind-solar-water-load complementary characteristic of small hydropower area lacking measurement runoff | |
| CN116151464A (en) | Photovoltaic power generation power prediction method, system and storable medium | |
| CN115905904A (en) | Line loss abnormity evaluation method and device for power distribution network feeder line | |
| Sun | Real estate evaluation model based on genetic algorithm optimized neural network | |
| Usman et al. | Data analytics for short term price and load forecasting in smart grids using enhanced recurrent neural network | |
| CN113837434A (en) | Solar photovoltaic power generation prediction method and device, electronic equipment and storage medium | |
| CN104239689B (en) | Short-term wind speed forecasting method based on optimization Method Using Relevance Vector Machine | |
| CN116739172B (en) | Method and device for ultra-short-term prediction of offshore wind power based on climbing identification | |
| CN114282336A (en) | New energy power station output scene generation method and system | |
| CN112132344A (en) | Short-term wind power prediction method based on similar day and FRS-SVM | |
| CN111310974A (en) | Short-term water demand prediction method based on GA-ELM | |
| CN114492944A (en) | TLBO-Elman-based photovoltaic power station short-term power generation power prediction method and device and storage medium |
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 |