CN115186923A - Photovoltaic power generation power prediction method and device and electronic equipment - Google Patents
Photovoltaic power generation power prediction method and device and electronic equipment Download PDFInfo
- Publication number
- CN115186923A CN115186923A CN202210893658.4A CN202210893658A CN115186923A CN 115186923 A CN115186923 A CN 115186923A CN 202210893658 A CN202210893658 A CN 202210893658A CN 115186923 A CN115186923 A CN 115186923A
- Authority
- CN
- China
- Prior art keywords
- historical
- training
- target
- prediction model
- power generation
- 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
- 238000010248 power generation Methods 0.000 title claims abstract description 102
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000012549 training Methods 0.000 claims abstract description 114
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 40
- 238000013528 artificial neural network Methods 0.000 claims abstract description 39
- 210000002569 neuron Anatomy 0.000 claims description 40
- 238000012545 processing Methods 0.000 claims description 24
- 238000011156 evaluation Methods 0.000 claims description 15
- 238000012360 testing method Methods 0.000 claims description 11
- 230000002159 abnormal effect Effects 0.000 claims description 10
- 238000012795 verification Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 238000000638 solvent extraction Methods 0.000 claims 1
- 239000010410 layer Substances 0.000 description 24
- 230000006870 function Effects 0.000 description 11
- 230000008569 process Effects 0.000 description 9
- 230000000694 effects Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000007781 pre-processing Methods 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 239000000428 dust Substances 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/004—Generation forecast, e.g. methods or systems for forecasting future energy generation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Abstract
The application discloses a photovoltaic power generation power prediction method and device and electronic equipment. Wherein, the method comprises the following steps: acquiring meteorological data of an area where the photovoltaic power station is located at the current moment and power generation power of the photovoltaic power station at the current moment, wherein the meteorological data at least comprise weather types; performing characteristic selection on meteorological data and power generation power through a lasso algorithm to obtain a target characteristic value; determining a target prediction model from a plurality of pre-trained prediction models according to weather types, wherein each prediction model corresponds to one weather type, and the prediction model is obtained by training through a deep neural network algorithm; and inputting the target characteristic value into a target prediction model, and outputting the predicted power generation power of the photovoltaic power station at a preset moment. The method and the device solve the technical problems that in the prior art, a photovoltaic power generation power prediction model is low in accuracy and deep characteristic information utilization rate.
Description
Technical Field
The application relates to the field of artificial intelligence and the field of photovoltaic power generation, in particular to a method and a device for predicting photovoltaic power generation power and electronic equipment.
Background
With the development of photovoltaic power generation technology, the number of photovoltaic power generation stations is increasing. The photovoltaic power generation power of the photovoltaic power station is influenced by factors such as meteorological change and day and night periodic change, so that certain intermittence and fluctuation are reflected. Therefore, the method can accurately predict the power generation power of the photovoltaic power station, and becomes a key for improving the operation efficiency and stability of the photovoltaic power station.
At present, in the prior art, a physical model method is usually adopted to predict photovoltaic power generation power of a photovoltaic power station, but the physical model cannot be widely applied to different weather types and needs to meet various assumed conditions, and the power generation power of the photovoltaic power station under different weather types is greatly different, so that the accuracy of the photovoltaic power generation power predicted by the physical model is low.
In view of the above problems, no effective solution has been proposed.
Disclosure of Invention
The embodiment of the application provides a method and a device for predicting photovoltaic power generation power and electronic equipment, and aims to at least solve the technical problems that in the prior art, a photovoltaic power generation power prediction model is low in accuracy and deep characteristic information utilization rate.
According to an aspect of an embodiment of the present application, there is provided a method for predicting photovoltaic power generation power, including: acquiring meteorological data of an area where the photovoltaic power station is located at the current moment and power generation power of the photovoltaic power station at the current moment, wherein the meteorological data at least comprises a weather type; performing characteristic selection on meteorological data and power generation power through a lasso algorithm to obtain a target characteristic value; determining a target prediction model from a plurality of pre-trained prediction models according to weather types, wherein each prediction model corresponds to one weather type, and the prediction model is a model obtained through deep neural network algorithm training; and inputting the target characteristic value into a target prediction model, and outputting the predicted power generation power of the photovoltaic power station at a preset moment.
Further, the photovoltaic power generation power prediction method further comprises the following steps: performing primary characteristic selection on meteorological data and power generation power to obtain a plurality of characteristics; calculating a regression coefficient of each feature through a lasso algorithm, wherein the regression coefficient of the feature is used for representing the influence degree of the feature on the power generation power; determining a target feature from the plurality of features according to the regression coefficients; and determining a target characteristic value corresponding to the target characteristic based on the meteorological data and the generated power.
Further, the prediction model is generated by training through the following method: acquiring historical generating power of a photovoltaic power station at a plurality of historical moments; acquiring historical meteorological data of an area where a photovoltaic power station is located at a plurality of historical moments, wherein the historical meteorological data at least comprise historical weather types; constructing a historical data set according to historical meteorological data and historical generated power at a plurality of historical moments; dividing a historical data set into a plurality of historical subdata sets according to historical weather types, wherein each historical weather type corresponds to one historical subdata set; performing feature selection on each historical subdata set through a lasso algorithm to obtain a training feature value corresponding to each historical subdata set; and training according to the training characteristic values corresponding to the historical sub data sets to obtain prediction models, wherein each prediction model corresponds to one historical sub data set.
Further, the photovoltaic power generation power prediction method further comprises the following steps: abnormal value processing is carried out on the historical data set to obtain a first historical data set; missing value filling processing is carried out on the first historical data set in a linear interpolation mode to obtain a second historical data set; carrying out data normalization processing on the second historical data set to obtain a target historical data set; the target historical data set is divided into a plurality of historical subdata sets according to the historical weather types.
Further, the photovoltaic power generation power prediction method further comprises the following steps: dividing the historical sub data set into a training set, a verification set and a test set according to a preset proportion, wherein the training set is used for training to obtain a prediction model, the verification set is used for carrying out preliminary evaluation on the performance of the prediction model, and the test set is used for carrying out final evaluation on the performance of the prediction model; determining a training characteristic value corresponding to the training set as a target training characteristic value; and training according to the training set and the target training characteristic value to obtain a prediction model.
Further, the photovoltaic power generation power prediction method further comprises the following steps: randomly determining the weight of each neuron in a deep neural network, wherein the deep neural network at least comprises a plurality of input layer neurons, a plurality of hidden layer neurons and at least one output layer neuron; inputting the target training characteristic value into a deep neural network; calculating the output value of the target training characteristic value after passing through each neuron according to the weight of each neuron; and training according to the output value to obtain a prediction model.
Further, the photovoltaic power generation power prediction method further comprises the following steps: determining an output result of the deep neural network according to an output value output by the neuron of the output layer; determining a loss value between an output result and actual historical generated power through a preset loss function; determining a minimum value of a loss function according to a back propagation algorithm and the loss value; and updating the weight of each neuron according to the minimum value to obtain a prediction model.
According to another aspect of the embodiments of the present application, there is also provided a photovoltaic power generation power prediction apparatus, including: the acquisition module is used for acquiring meteorological data of an area where the photovoltaic power station is located at the current moment and the power generation power of the photovoltaic power station at the current moment, wherein the meteorological data at least comprises a weather type; the characteristic selection module is used for carrying out characteristic selection on meteorological data and generated power through a lasso algorithm to obtain a target characteristic value; the system comprises a determining module, a predicting module and a judging module, wherein the determining module is used for determining a target predicting model from a plurality of pre-trained predicting models according to weather types, each predicting model corresponds to one weather type, and the predicting model is obtained by training through a deep neural network algorithm; and the input module is used for inputting the target characteristic value into the target prediction model and outputting the predicted power generation power of the photovoltaic power station at the preset moment.
According to another aspect of the embodiments of the present application, there is also provided a computer-readable storage medium, in which a computer program is stored, where the computer program is configured to execute the above-mentioned method for predicting photovoltaic power generation power when running.
According to another aspect of embodiments of the present application, there is also provided an electronic device, wherein the electronic device includes one or more processors; a storage device for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement a method for operating a program, wherein the program is arranged to perform the above-described method for predicting photovoltaic power generation when operated.
According to the technical scheme, a mode of predicting the predicted power generation power of the photovoltaic power station at the preset time through a target prediction model is adopted, firstly, meteorological data of an area where the photovoltaic power station is located at the current time and the power generation power of the photovoltaic power station at the current time are obtained, wherein the meteorological data at least comprise a weather type, then, the meteorological data and the power generation power are subjected to characteristic selection through a lasso algorithm to obtain a target characteristic value, the target prediction model is determined from a plurality of pre-trained prediction models according to the weather type, finally, the target characteristic value is input into the target prediction model, and the predicted power generation power of the photovoltaic power station at the preset time is output. Each prediction model corresponds to a weather type, and the prediction model is obtained through deep neural network algorithm training.
According to the above, because different weather types have different influences on the photovoltaic power generation power, but the power generation power under the same weather type shows similarity, the corresponding deep neural network prediction model is trained for each weather type, and when the power generation power of the photovoltaic power station is actually predicted, the corresponding prediction model is selected according to the weather type of the area where the photovoltaic power station is located to complete the prediction of the power generation power, so that the power generation power of the photovoltaic power station can be more accurately predicted by combining weather data such as the weather type and the like, and the prediction accuracy of the power generation power is improved.
Therefore, according to the technical scheme, the purpose of improving the prediction accuracy of the power generation power of the photovoltaic power station is achieved, the effects of improving the operation efficiency and stability of the photovoltaic power station are achieved, and the technical problems that in the prior art, a photovoltaic power generation power prediction model is low in accuracy and deep characteristic information utilization rate are solved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
FIG. 1 is a flow chart of an alternative method of predicting photovoltaic power generation according to embodiments of the present application;
FIG. 2 is a flow chart of an alternative predictive model training method according to an embodiment of the present application;
FIG. 3 is a flow chart of an alternative method of photovoltaic power generation prediction according to an embodiment of the present application;
FIG. 4 is a flow chart of an alternative predictive model training process according to an embodiment of the application;
FIG. 5 is a graphical illustration of an alternative photovoltaic power generation prediction according to an embodiment of the present application;
fig. 6 is a schematic diagram of an alternative photovoltaic power generation prediction device according to an embodiment of the present application.
Detailed Description
In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only partial embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and claims of this application and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the application described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Example 1
According to an embodiment of the present application, there is provided an embodiment of a method for predicting photovoltaic power generation, where the steps illustrated in the flowchart of the drawings may be performed in a computer system, such as a set of computer executable instructions, and where a logical order is illustrated in the flowchart, in some cases the steps illustrated or described may be performed in an order different than here.
In addition, it should be further noted that a generated power prediction system may be an execution subject of the photovoltaic generated power prediction method in the embodiment of the present application.
Fig. 1 is a flow chart of an alternative photovoltaic power generation power prediction method according to an embodiment of the present application, as shown in fig. 1, the method includes the following steps:
step S101, acquiring meteorological data of the area of the photovoltaic power station at the current moment and the power generation power of the photovoltaic power station at the current moment.
In step S101, the weather data includes at least a weather type. Specifically, the weather types include, but are not limited to, eight types of weather types, including sunny, cloudy, fog, rain, snow, haze, and dust. In addition, the meteorological data includes at least a variety of weather data such as temperature, humidity, relative humidity, rainfall, wind speed, and the like, in addition to the weather type. The temperature in the meteorological data includes, in addition to the temperature outside the photovoltaic power plant, the temperature inside the photovoltaic power plant, for example, the internal temperature of an inverter in the photovoltaic power plant and the temperature of an inverter module.
And S102, performing characteristic selection on the meteorological data and the generated power through a lasso algorithm to obtain a target characteristic value.
In an optional embodiment, the generated power prediction system firstly performs preliminary feature selection on meteorological data and generated power to obtain a plurality of features, and then calculates a regression coefficient of each feature through a lasso algorithm, wherein the regression coefficient of the features is used for representing the influence degree of the features on the generated power. And then the generated power prediction system determines a target characteristic from the multiple characteristics according to the regression coefficient, and determines a target characteristic value corresponding to the target characteristic based on the meteorological data and the generated power.
Optionally, the basic principle of the lasso algorithm is to minimize the sum of squares of residuals under the constraint condition that the regression coefficient is ensured to be less than or equal to a certain threshold, and compress the regression coefficient of the less-affected feature to zero, so as to exclude the feature having less influence on the generated power. The mathematical expression for minimizing the sum-of-squared residuals problem is as follows:
wherein x is ij Characteristics that affect the power generated; y is i Generating power for the output parameter; beta is a j Regression coefficients for the jth feature; the threshold t is for j A norm penalty of (1), the value range is approximately [0, + ∞]When the value is small, regression coefficients of some features having small influence on the generated power are compressed to be 0, and the features with the regression coefficients compressed to be 0 are discarded, so that more important target features, namely features having large influence on the generated power, can be selected from a plurality of features.
For example, taking the weather type "sunny" as an example, the multiple features preliminarily selected are: the method comprises four characteristics, namely a characteristic A, a characteristic B, a characteristic C and a characteristic D, wherein the characteristic A, the characteristic B and the characteristic C are characteristics related to meteorological data, the characteristic D is the power generation power at the current moment, the regression coefficient of each characteristic is calculated through a lasso algorithm, and then the regression coefficient of the characteristic A is compressed to be 0, namely the condition that the influence of the characteristic A on the power generation power under the weather type is small is explained, so that the characteristic A can be abandoned, and the characteristic B, the characteristic C and the characteristic D are finally selected as input characteristics (namely target characteristics) of a model. Since the target feature is not an actual feature value, it is also necessary to determine a target feature value corresponding to each target feature from the meteorological data and the generated power, for example, for the feature D, the target feature value corresponding to the feature D is an actual generated power value (for example, 8 KW) at the current time, and if the feature B is an air temperature feature, the target feature value corresponding to the feature B is an actual air temperature value (for example, 25 degrees celsius).
By selecting the characteristics through a lasso algorithm, the characteristics with small influence on the generated power can be abandoned, so that the generated power prediction efficiency is improved, the interference of irrelevant characteristics is avoided, and the generated power prediction accuracy can be improved.
Step S103, determining a target prediction model from a plurality of pre-trained prediction models according to the weather type.
In step S103, each prediction model corresponds to a weather type, and the prediction model is a model obtained by training through a deep neural network algorithm.
Optionally, a plurality of pre-trained prediction models are stored in the generated power prediction system, wherein each prediction model is a deep neural network model with a multilayer neural network, and each prediction model is trained on training characteristic values selected by a lasso algorithm. In addition, each prediction model corresponds to one weather type, and assuming that the weather types include eight types of weather types, namely sunny, cloudy, fog, rain, snow, haze and dust, the generated power prediction system stores 8 prediction models in total. And if the weather type of the area where the photovoltaic power station is located at the current moment is the weather type of 'fine', the target prediction model is a prediction model corresponding to the weather type of 'fine'.
And step S104, inputting the target characteristic value into a target prediction model, and outputting the predicted power generation power of the photovoltaic power station at a preset moment.
In step S104, after the target prediction model is determined, the generated power prediction system may input the target characteristic value obtained in step S102 into the target prediction model, then calculate the target characteristic value by using the multilayer neural network inside the target prediction model, and finally output the predicted generated power of the photovoltaic power plant at the preset time by using the output layer of the target prediction model.
It should be noted that the preset time is a time in the future, for example, the preset time may be a future time separated from the current time by a preset time interval, and the preset time interval may be any time length, for example, 1 hour or 2 hours, and this application is not limited in this respect.
It should be noted that, for different weather types, if a single-layer network algorithm such as a BP neural network is used for predicting the power generation power of the photovoltaic power station, the problems of low prediction precision and poor generalization capability are often caused by a single calculation process, and the method adopts a deep neural network model with a multilayer neural network to predict the power generation power of the photovoltaic power station, can give different weights to neurons of each layer of neural network, and uses the multilayer neural network structure to fit a complex nonlinear relationship, so that the prediction precision and the generalization capability of the prediction model are improved, and the prediction accuracy of the power generation power is further improved.
Based on the contents of the steps S101 to S104, in the technical solution of the present application, a manner of predicting the predicted power generation power of the photovoltaic power plant at the preset time by using a target prediction model is adopted, firstly, meteorological data of an area where the photovoltaic power plant is located at the current time and the power generation power of the photovoltaic power plant at the current time are obtained, wherein the meteorological data at least include a weather type, then, feature selection is performed on the meteorological data and the power generation power through a lasso algorithm to obtain a target feature value, the target prediction model is determined from a plurality of pre-trained prediction models according to the weather type, and finally, the target feature value is input into the target prediction model to output the predicted power generation power of the photovoltaic power plant at the preset time. Each prediction model corresponds to a weather type, and the prediction model is obtained through deep neural network algorithm training.
According to the above, because different weather types have different influences on the photovoltaic power generation power, but the power generation power under the same weather type shows similarity, the corresponding deep neural network prediction model is trained for each weather type, and when the power generation power of the photovoltaic power station is actually predicted, the corresponding prediction model is selected according to the weather type of the area where the photovoltaic power station is located to complete the prediction of the power generation power, so that the power generation power of the photovoltaic power station can be more accurately predicted by combining weather data such as the weather type, and the prediction accuracy of the power generation power is improved.
Therefore, according to the technical scheme, the purpose of improving the prediction accuracy of the generated power of the photovoltaic power station is achieved, the effects of improving the operation efficiency and stability of the photovoltaic power station are achieved, and the technical problems that in the prior art, a photovoltaic generated power prediction model is low in accuracy and deep feature information utilization rate are solved.
In an alternative embodiment, fig. 2 is a flowchart illustrating a training method of an alternative predictive model according to an embodiment of the present application, and as shown in fig. 2, the predictive model is generated by training according to the following method:
step S201, a generated power prediction system acquires historical generated power of a photovoltaic power station at a plurality of historical moments;
step S202, a power generation power prediction system acquires historical meteorological data of an area where a photovoltaic power station is located at a plurality of historical moments, wherein the historical meteorological data at least comprises historical weather types;
step S203, the generating power prediction system constructs a historical data set according to historical meteorological data and historical generating power at a plurality of historical moments;
step S204, the generated power prediction system divides the historical data set into a plurality of historical subdata sets according to the historical weather types, wherein each historical weather type corresponds to one historical subdata set;
step S205, the generated power prediction system performs characteristic selection on each historical subdata set through a lasso algorithm to obtain a training characteristic value corresponding to each historical subdata set;
and S206, training the generated power prediction system according to the training characteristic values corresponding to the historical subdata sets to obtain prediction models, wherein each prediction model corresponds to one historical subdata set.
Optionally, in step S201 and step S202, the historical meteorological data and the historical generated power are meteorological data and generated power in a past historical time period, and the historical time period may be a past time period of one year, one month, two years, or two months. The sampling may be performed at a preset time interval within the historical time period to obtain the meteorological data and the generated power at each historical time, for example, the preset time interval may be one hour, the historical time period may be one year (calculated according to 365 days) from the current time, and the sampling may be performed once per hour to obtain 8760 historical meteorological data and 8760 historical generated power.
Optionally, in step S203, as shown in fig. 3, the generated power prediction system constructs a historical data set according to the obtained historical meteorological data and historical generated power, where the historical data set includes a plurality of data samples, each data sample is composed of historical meteorological data and historical generated power corresponding to one historical time, for example, 8760 data samples may be constructed for 8760 historical generated powers and each data sample includes one historical meteorological data and one historical generated power.
Optionally, in step S204, the generated power prediction system first performs a preprocessing operation on the historical data set to obtain a target historical data set, where the preprocessing operation includes an abnormal value processing, a missing value filling processing, and a data normalization processing. The generated power prediction system then divides the target historical data set into a plurality of historical subdata sets according to historical weather types.
Specifically, in the preprocessing process, as shown in fig. 3, the generated power prediction system first performs abnormal value processing on the historical data set to obtain a first historical data set, then performs missing value filling processing on the first historical data set in a linear interpolation manner to obtain a second historical data set, and finally performs data normalization processing on the second historical data set to obtain a target historical data set.
For example, some abnormal data may exist in the historical data set, for example, a data value of a certain historical generated power of the photovoltaic power plant at the evening time is greater than 0, and the data value is the abnormal data. In order to avoid that the abnormal data influence the accuracy of the trained prediction model, the generated power prediction system can remove the abnormal data, namely, abnormal value processing operation is carried out on the historical data set.
In addition, when some missing values exist among data samples in the historical data set, the problem of time discontinuity of the historical generated power data can occur, so that the difficulty of data processing is increased, and therefore, the generated power prediction system can also adopt a linear interpolation mode to supplement the missing values.
In addition, because the dimensions and the magnitude of the historical data set have certain differences, if the data in the historical data set is directly used as input data to perform a training experiment, the data with a higher numerical level has certain influence on the process of establishing the model, and the data with a lower numerical level is ignored in the experiment. Therefore, the generated power prediction system in the application can utilize a Min-Max method to carry out normalization processing on data required by an experiment, eliminate dimensional differences among different data and then carry out modeling training. Wherein the Min-Max method equation is as follows:
wherein, y i In the form of a raw measurement value,
for the normalized measured values, y min 、y max I represents the number of samples for the minimum and maximum values in the measurement.
It should be noted that the generated power prediction system obtains the target historical data set after preprocessing the historical data set, and on this basis, the generated power prediction system also divides the target historical data set into a plurality of historical sub-data sets according to the weather type.
Specifically, as shown in fig. 3, taking eight weather types of sunny, cloudy, fog, rain, snow, haze, and dust as examples, the target historical data set may be divided into 8 historical sub data sets, where each weather type corresponds to one historical sub data set. Because different weather types can cause certain influence on the light resources and the generated power of the photovoltaic power station, the prediction accuracy of the generated power can be improved by respectively training the prediction model for each weather type. It should be noted that since irradiance of photovoltaic power plants is difficult to obtain, typically through a specific instrument or a billing channel of a meteorological office, other data may be used in place of solar irradiance information.
Optionally, in step S205, the generated power prediction system performs feature selection on each historical sub data set by using a lasso algorithm. As shown in fig. 3, taking the weather type "yin" as an example, the preliminarily selected features are: the rainfall, the relative humidity, the air temperature, the wind speed, the temperature inside the inverter, the temperature of the inverter module, the historical generated power in the previous hour and the historical generated power in the previous two hours are 8 characteristics. Since the regression coefficients of the features such as rainfall, wind speed, inverter internal temperature, inverter module temperature and the like are compressed to zero, 4 features including relative humidity, air temperature, historical generated power in the previous hour and historical generated power in the previous two hours are finally selected as input features of the prediction model, and for each input feature, the generated power prediction system can determine a feature value (i.e., a training feature value) corresponding to each input feature according to historical meteorological data and historical generated power.
It should be noted that, when training the prediction model, if the dimensionality of the input features is too high, the computational efficiency of the modeling process is reduced, and the time for training the model is greatly increased, thereby reducing the accuracy of the prediction model to be trained. Therefore, before the prediction model is trained, the characteristics with small influence on the photovoltaic power generation power need to be removed, so that the complexity of the model is reduced, and the modeling efficiency is improved. The characteristics are selected through the lasso algorithm, and the characteristics which have small influence on the photovoltaic power generation power can be just identified, so that the purpose of selecting the characteristics which have large influence on the photovoltaic power generation power as the input characteristics of the training prediction model is achieved, and the training efficiency of the prediction model and the prediction accuracy of the prediction model are improved.
Optionally, in step S206, as shown in fig. 3, when a prediction model is obtained by training according to each historical sub data set, the generated power prediction system first divides the historical sub data set into a training set, a verification set and a test set according to a preset ratio, where the training set is used for training to obtain the prediction model, the verification set is used for performing preliminary evaluation on the performance of the prediction model, and the test set is used for performing final evaluation on the performance of the prediction model. And then the generating power prediction system determines the training characteristic value corresponding to the training set as a target training characteristic value, and obtains a prediction model according to the training set and the target training characteristic value.
Specifically, the preset ratio may be set in a self-defined manner, and in general, for the training process of the neural network model, the preset ratio may be set to 6, for example, if one history sub data set includes 990 data samples in total, the number of data samples in the training set corresponding to the history sub data set is 532, the number of data samples in the verification set is 228, and the number of data samples in the test set is 230.
In addition, as shown in fig. 3, after 8 prediction models are obtained by training according to the training set, the power generation prediction system may perform a first evaluation (i.e., a preliminary evaluation) on each prediction model by using the validation set, and after obtaining a result of the preliminary evaluation, the power generation prediction system may determine a training effect of the prediction model according to the result of the preliminary evaluation, and adjust the prediction model based on the training effect. Further, after obtaining the adjusted prediction model, the generated power prediction system performs a second evaluation (i.e., a final evaluation) on the adjusted prediction model by using the test set, and determines a prediction accuracy of the adjusted prediction model according to a final evaluation result.
Optionally, as shown in fig. 3, after obtaining the training set, the test set, and the verification set, the generated power prediction system determines the training eigenvalue in the training set as a target training eigenvalue, and then trains according to the training set and the target training eigenvalue to obtain the prediction model.
Specifically, when a prediction model is obtained through training of each training set, the power generation prediction system first randomly determines the weight of each neuron in a deep neural network, wherein the deep neural network at least comprises a plurality of input layer neurons, a plurality of hidden layer neurons and at least one output layer neuron. And then the generated power prediction system inputs the target training characteristic value into the deep neural network, calculates the output value of the target training characteristic value after passing through each neuron according to the weight of each neuron, and finally trains the generated power prediction system according to the output value to obtain a prediction model.
Optionally, in order to ensure the accuracy of the prediction model, the generation power prediction system may also adjust the deep neural network according to a loss function. Specifically, the power generation power prediction system determines an output result of the deep neural network according to an output value output by neurons in an output layer, determines a loss value between the output result and actual historical power generation power through a preset loss function, then determines a minimum value of the loss function according to a back propagation algorithm and the loss value, and finally updates the weight of each neuron according to the minimum value to obtain a prediction model.
Specifically, as shown in fig. 4, taking a prediction model corresponding to any one weather type as an example, the training process of the prediction model includes the following steps:
step S401: the network structure composition of the deep neural network DNN is defined. The deep neural network can be divided into an input layer, a hidden layer and an output layer. The number of neurons in the input layer is M; the total number of the hidden layers is N, the number of neurons of the hidden layer of the ith layer is X, and the number of neurons of the output layer is at least one;
step S402: the generated power prediction system randomly initializes the weight of each neuron, then inputs the target training characteristic value into a deep neural network, and the deep neural network sequentially calculates the output value of each neuron in each layer. For example, for the output of the 1 st neuron at layer 3
The calculation formula of (a) is as follows:
where σ (·) is the relu function. The convergence speed of the relu function is high, and the problem of gradient disappearance can be relieved;
step S403: the generated power prediction system measures a loss value between an output result of the deep neural network and a power measurement value (corresponding to actual historical generated power) by using a loss function 'MSE';
step S404: the generating power prediction system obtains a minimum value of a loss function through a back propagation algorithm and the obtained loss value optimization calculation, and determines the latest weight of each neuron according to the minimum value, so that a prediction model is obtained.
In an optional embodiment, when the generation power of the photovoltaic power station needs to be predicted by using the prediction model, the generation power prediction system only needs to perform feature selection on the meteorological data and the generation power of the photovoltaic power station at the current moment through a lasso algorithm to obtain a target feature value, determines a corresponding target prediction model according to the weather type of the area where the photovoltaic power station is located at the current moment, inputs the target feature value into the target prediction model, and the target prediction model can output the predicted generation power of the photovoltaic power station at the preset moment.
Fig. 5 shows the predicted generated power predicted by the prediction model in the present application, wherein the abscissa in fig. 5 represents the preset time and the ordinate represents the power value of the predicted generated power.
As shown in table 1 below, the generated power of the same photovoltaic power plant was predicted using the prediction model (DNN) of the present application, the ELM model of the related art, and the SVR model, and the analysis results were as follows:
TABLE 1
As shown in Table 1, either the root mean square error or the goodness of fit R are compared 2 The prediction model in the application has better prediction effect. Therefore, according to the technical scheme, the purpose of improving the prediction accuracy of the generated power of the photovoltaic power station is achieved, and the photovoltaic power station is improvedThe operation efficiency and stability of the power station are improved, and the technical problems that in the prior art, a photovoltaic power generation power prediction model is low in accuracy and deep characteristic information utilization rate are solved.
Example 2
According to an embodiment of the present application, there is also provided an embodiment of a photovoltaic power generation prediction apparatus, where fig. 6 is a schematic diagram of an alternative photovoltaic power generation prediction apparatus according to the embodiment of the present application, and as shown in fig. 6, the apparatus includes: an obtaining module 601, a feature selecting module 602, a determining module 603 and an inputting module 604.
The acquiring module 601 is configured to acquire meteorological data of an area where the photovoltaic power plant is located at a current moment and power generation power of the photovoltaic power plant at the current moment, where the meteorological data at least includes a weather type; the characteristic selection module 602 is configured to perform characteristic selection on meteorological data and power generation power through a lasso algorithm to obtain a target characteristic value; a determining module 603, configured to determine a target prediction model from multiple pre-trained prediction models according to a weather type, where each prediction model corresponds to one weather type, and the prediction model is a model obtained through deep neural network algorithm training; the input module 604 is configured to input the target feature value into the target prediction model, and output the predicted power generation power of the photovoltaic power plant at a preset time.
It should be noted that the obtaining module 601, the feature selecting module 602, the determining module 603, and the inputting module 604 correspond to steps S101 to S104 in the foregoing embodiment, and the four modules are the same as the corresponding steps in the implementation example and the application scenario, but are not limited to the disclosure in embodiment 1.
Optionally, the feature selecting module further includes: the device comprises a preliminary feature selection unit, a calculation unit, a first determination unit and a second determination unit. Wherein. The preliminary feature selection unit is used for performing preliminary feature selection on the meteorological data and the power generation power to obtain a plurality of features; the calculating unit is used for calculating a regression coefficient of each characteristic through a lasso algorithm, wherein the regression coefficient of the characteristic is used for representing the influence degree of the characteristic on the generated power; a first determining unit configured to determine a target feature from the plurality of features based on the regression coefficient; and the second determining unit is used for determining a target characteristic value corresponding to the target characteristic based on the meteorological data and the generated power.
Optionally, the photovoltaic power generation power prediction apparatus further includes: the device comprises a first acquisition module, a second acquisition module, a construction module, a division module, a first characteristic selection module and a training module. The first acquisition module is used for acquiring historical generating power of the photovoltaic power station at a plurality of historical moments; the second acquisition module is used for acquiring historical meteorological data of an area where the photovoltaic power station is located at a plurality of historical moments, wherein the historical meteorological data at least comprises historical weather types; the building module is used for building a historical data set according to historical meteorological data and historical power generation at a plurality of historical moments; the dividing module is used for dividing the historical data set into a plurality of historical subdata sets according to historical weather types, wherein each historical weather type corresponds to one historical subdata set; the first characteristic selection module is used for performing characteristic selection on each historical subdata set through a lasso algorithm to obtain a training characteristic value corresponding to each historical subdata set; and the training module is used for training according to the training characteristic values corresponding to the historical subdata sets to obtain prediction models, wherein each prediction model corresponds to one historical subdata set.
Optionally, the dividing module further includes: the device comprises a first processing unit, a second processing unit, a third processing unit and a first dividing unit. The first processing unit is used for processing abnormal values of the historical data set to obtain a first historical data set; the second processing unit is used for performing missing value filling processing on the first historical data set in a linear interpolation mode to obtain a second historical data set; the third processing unit is used for carrying out data normalization processing on the second historical data set to obtain a target historical data set; the device comprises a first dividing unit, a second dividing unit and a third dividing unit, wherein the first dividing unit is used for dividing the target historical data set into a plurality of historical sub data sets according to the historical weather types.
Optionally, the training module further includes: the device comprises a second dividing unit, a third determining unit and a first training unit. The second dividing unit is used for dividing the history subdata set into a training set, a verification set and a test set according to a preset proportion, wherein the training set is used for training to obtain a prediction model, the verification set is used for carrying out preliminary evaluation on the performance of the prediction model, and the test set is used for carrying out final evaluation on the performance of the prediction model; a third determining unit, configured to determine a training feature value corresponding to the training set as a target training feature value; and the first training unit is used for training according to the training set and the target training characteristic value to obtain a prediction model.
Optionally, the first training unit further comprises: the device comprises a first determining subunit, an input subunit, a calculating subunit and a first training subunit. The first determining subunit is used for randomly determining the weight of each neuron in the deep neural network, wherein the deep neural network at least comprises a plurality of input layer neurons, a plurality of hidden layer neurons and at least one output layer neuron; the input subunit is used for inputting the target training characteristic value into the deep neural network; the calculating subunit is used for calculating an output value of the target training characteristic value after passing through each neuron according to the weight of each neuron; and the first training subunit is used for training according to the output value to obtain a prediction model.
Optionally, the first training subunit further includes: the device comprises a first determining submodule, a second determining submodule, a third determining submodule and an updating submodule. The first determining submodule is used for determining an output result of the deep neural network according to an output value output by the neuron of the output layer; the second determining submodule is used for determining a loss value between the output result and the actual historical generating power through a preset loss function; the third determining submodule is used for determining the minimum value of the loss function according to the back propagation algorithm and the loss value; and the updating submodule is used for updating the weight of each neuron according to the minimum value to obtain the prediction model.
Example 3
According to an embodiment of the present application, there is also provided a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the method for predicting photovoltaic power generation in embodiment 1.
Example 4
There is also provided, in accordance with an embodiment of the present application, an electronic device, including one or more processors; a storage device configured to store one or more programs, which when executed by the one or more processors, cause the one or more processors to implement a method for operating a program, wherein the program is configured to execute the method for predicting photovoltaic power generation power in embodiment 1 described above when executed.
The above-mentioned serial numbers of the embodiments of the present application are merely for description, and do not represent the advantages and disadvantages of the embodiments.
In the embodiments of the present application, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to the related descriptions of other embodiments.
In the embodiments provided in the present application, it should be understood that the disclosed technical content can be implemented in other manners. The above-described apparatus embodiments are merely illustrative, and for example, a division of a unit may be a logical division, and in actual implementation, there may be another division, for example, multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or may not be executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, units or modules, and may be in an electrical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on a plurality of units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, or portions or all or portions of the technical solutions that contribute to the prior art, may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.
The foregoing is only a preferred embodiment of the present application and it should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present application, and these improvements and modifications should also be considered as the protection scope of the present application.
Claims (10)
1. A method for predicting photovoltaic power generation power is characterized by comprising the following steps:
acquiring meteorological data of an area where a photovoltaic power station is located at the current moment and power generation power of the photovoltaic power station at the current moment, wherein the meteorological data at least comprises a weather type;
performing characteristic selection on the meteorological data and the power generation power through a lasso algorithm to obtain a target characteristic value;
determining a target prediction model from a plurality of pre-trained prediction models according to the weather types, wherein each prediction model corresponds to one weather type, and the prediction model is a model obtained by training through a deep neural network algorithm;
and inputting the target characteristic value into the target prediction model, and outputting the predicted power generation power of the photovoltaic power station at a preset moment.
2. The method of claim 1, wherein the step of performing feature selection on the meteorological data and the power generation power through a lasso algorithm to obtain a target feature value comprises:
performing preliminary feature selection on the meteorological data and the generated power to obtain a plurality of features;
calculating a regression coefficient of each feature through the lasso algorithm, wherein the regression coefficient of the feature is used for representing the influence degree of the feature on the generated power;
determining a target feature from a plurality of said features based on said regression coefficients;
and determining a target characteristic value corresponding to the target characteristic based on the meteorological data and the generated power.
3. The method of claim 1, wherein the predictive model is generated by training:
acquiring historical generating power of the photovoltaic power station at a plurality of historical moments;
acquiring historical meteorological data of an area where the photovoltaic power station is located at the plurality of historical moments, wherein the historical meteorological data at least comprise historical weather types;
constructing a historical data set according to the historical meteorological data and the historical generated power at the plurality of historical moments;
dividing the historical data set into a plurality of historical subdata sets according to the historical weather types, wherein each historical weather type corresponds to one historical subdata set;
performing characteristic selection on each historical subdata set through the lasso algorithm to obtain a training characteristic value corresponding to each historical subdata set;
and training according to the training characteristic values corresponding to the historical sub data sets to obtain the prediction models, wherein each prediction model corresponds to one historical sub data set.
4. The method of claim 3, wherein partitioning the historical data set into a plurality of historical sub data sets according to the historical weather type comprises:
carrying out abnormal value processing on the historical data set to obtain a first historical data set;
filling missing values in the first historical data set in a linear interpolation mode to obtain a second historical data set;
carrying out data normalization processing on the second historical data set to obtain a target historical data set;
and dividing the target historical data set into a plurality of historical subdata sets according to the historical weather types.
5. The method of claim 3, wherein the training of the prediction model according to the training eigenvalues corresponding to the historical sub data sets comprises:
dividing the historical subdata data set into a training set, a verification set and a test set according to a preset proportion, wherein the training set is used for training to obtain the prediction model, the verification set is used for carrying out preliminary evaluation on the performance of the prediction model, and the test set is used for carrying out final evaluation on the performance of the prediction model;
determining a training characteristic value corresponding to the training set as a target training characteristic value;
and training according to the training set and the target training characteristic value to obtain the prediction model.
6. The method of claim 5, wherein training the predictive model based on the training set and the target training feature values comprises:
randomly determining a weight of each neuron in a deep neural network, wherein the deep neural network comprises at least a plurality of input layer neurons, a plurality of hidden layer neurons, and at least one output layer neuron;
inputting the target training feature values into the deep neural network;
calculating an output value of the target training characteristic value after passing through each neuron according to the weight of each neuron;
and training according to the output value to obtain the prediction model.
7. The method of claim 6, wherein training the predictive model based on the output values comprises:
determining an output result of the deep neural network according to the output value output by the neuron of the output layer;
determining a loss value between the output result and the actual historical generated power through a preset loss function;
determining a minimum value of the loss function according to a back propagation algorithm and the loss value;
and updating the weight of each neuron according to the minimum value to obtain the prediction model.
8. A photovoltaic power generation power prediction apparatus, comprising:
the system comprises an acquisition module, a power generation module and a power generation module, wherein the acquisition module is used for acquiring meteorological data of an area where a photovoltaic power station is located at the current moment and power generation power of the photovoltaic power station at the current moment, and the meteorological data at least comprises a weather type;
the characteristic selection module is used for performing characteristic selection on the meteorological data and the power generation power through a lasso algorithm to obtain a target characteristic value;
the determining module is used for determining a target prediction model from a plurality of pre-trained prediction models according to the weather types, wherein each prediction model corresponds to one weather type, and the prediction model is obtained by training through a deep neural network algorithm;
and the input module is used for inputting the target characteristic value into the target prediction model and outputting the predicted power generation power of the photovoltaic power station at a preset moment.
9. A computer-readable storage medium, in which a computer program is stored, wherein the computer program is arranged to execute the method of predicting photovoltaic power generation as claimed in any one of claims 1 to 7 when executed.
10. An electronic device, characterized in that the electronic device comprises one or more processors; storage means for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement a method for operating a program, wherein the program is arranged to, when executed, perform the method of predicting photovoltaic power generation as claimed in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210893658.4A CN115186923A (en) | 2022-07-27 | 2022-07-27 | Photovoltaic power generation power prediction method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210893658.4A CN115186923A (en) | 2022-07-27 | 2022-07-27 | Photovoltaic power generation power prediction method and device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115186923A true CN115186923A (en) | 2022-10-14 |
Family
ID=83521816
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210893658.4A Pending CN115186923A (en) | 2022-07-27 | 2022-07-27 | Photovoltaic power generation power prediction method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115186923A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116307269A (en) * | 2023-05-16 | 2023-06-23 | 华能山东发电有限公司众泰电厂 | Photovoltaic power generation power prediction method and device based on artificial intelligence |
| CN117318024A (en) * | 2023-09-19 | 2023-12-29 | 华能澜沧江水电股份有限公司 | CNN neural network-based photovoltaic power generation power prediction management method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112862630A (en) * | 2021-03-08 | 2021-05-28 | 海南省电力学校(海南省电力技工学校) | Weather type index interval-based photovoltaic power prediction method, terminal and medium |
| WO2021238505A1 (en) * | 2020-05-27 | 2021-12-02 | 华北电力大学 | Federated learning-based regional photovoltaic power probability prediction method, and cooperative regulation and control system |
-
2022
- 2022-07-27 CN CN202210893658.4A patent/CN115186923A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021238505A1 (en) * | 2020-05-27 | 2021-12-02 | 华北电力大学 | Federated learning-based regional photovoltaic power probability prediction method, and cooperative regulation and control system |
| CN112862630A (en) * | 2021-03-08 | 2021-05-28 | 海南省电力学校(海南省电力技工学校) | Weather type index interval-based photovoltaic power prediction method, terminal and medium |
Non-Patent Citations (1)
| Title |
|---|
| 张玉来: "一种改进的DNN瓶颈特征提取方法" * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116307269A (en) * | 2023-05-16 | 2023-06-23 | 华能山东发电有限公司众泰电厂 | Photovoltaic power generation power prediction method and device based on artificial intelligence |
| CN116307269B (en) * | 2023-05-16 | 2023-10-31 | 华能山东泰丰新能源有限公司 | Photovoltaic power generation power prediction method and device based on artificial intelligence |
| CN117318024A (en) * | 2023-09-19 | 2023-12-29 | 华能澜沧江水电股份有限公司 | CNN neural network-based photovoltaic power generation power prediction management method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112365040B (en) | Short-term wind power prediction method based on multi-channel convolution neural network and time convolution network | |
| CN108197744B (en) | Method and system for determining photovoltaic power generation power | |
| CN110909919A (en) | Photovoltaic power prediction method of depth neural network model with attention mechanism fused | |
| CN112149879B (en) | New energy medium-and-long-term electric quantity prediction method considering macroscopic volatility classification | |
| CN115186923A (en) | Photovoltaic power generation power prediction method and device and electronic equipment | |
| CN111401599B (en) | Water level prediction method based on similarity search and LSTM neural network | |
| CN103117546A (en) | Ultrashort-term slide prediction method for wind power | |
| CN110766200A (en) | Method for predicting generating power of wind turbine generator based on K-means mean clustering | |
| CN113505923B (en) | Regional power grid short-term load prediction method and system | |
| CN114792156A (en) | Photovoltaic output power prediction method and system based on curve characteristic index clustering | |
| CN113537582B (en) | Photovoltaic power ultra-short-term prediction method based on short-wave radiation correction | |
| CN113554466A (en) | Short-term power consumption prediction model construction method, prediction method and device | |
| CN107977728A (en) | It is a kind of medium-term and long-term by hour Temperature prediction method based on BP artificial neural networks | |
| CN112686481A (en) | Runoff forecasting method and processor | |
| CN115860177A (en) | Photovoltaic power generation power prediction method based on combined machine learning model and application thereof | |
| CN114330934A (en) | Model parameter self-adaptive GRU new energy short-term power generation power prediction method | |
| CN112365056A (en) | Electrical load joint prediction method and device, terminal and storage medium | |
| CN115759389A (en) | Day-ahead photovoltaic power prediction method based on weather type similar day combination strategy | |
| CN110852492A (en) | Photovoltaic power ultra-short-term prediction method for finding similarity based on Mahalanobis distance | |
| CN111697560B (en) | Method and system for predicting load of power system based on LSTM | |
| CN117526274A (en) | New energy power prediction method, electronic equipment and storage medium in extreme climate | |
| CN117543537A (en) | Agent electricity purchasing user electric quantity prediction method, device and storage medium | |
| CN115907131A (en) | Method and system for building electric heating load prediction model in northern area | |
| CN116760006A (en) | Wind power prediction method, device, equipment and storage medium | |
| CN114971090A (en) | Electric heating load prediction method, system, equipment and 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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20221014 |