Method and system for predicting photovoltaic power generation in coastal areas considering atmospheric humidity

Abstract

The application discloses a kind of considering atmospheric humidity's coastal area photovoltaic power generation power prediction method and system, wherein method includes the following steps: obtaining the first meteorological element real-time data of the environment of current detection period coastal area photovoltaic power generation device, first meteorological element includes: relative humidity and evapotranspiration;Based on short-time photovoltaic power generation power prediction model, combine relative humidity real-time data and evapotranspiration real-time data, calculate the power prediction value of photovoltaic power generation device in next detection period.By introducing and coastal area characteristic highly related evapotranspiration and relative humidity and other meteorological elements, as the main influencing factor of short-term photovoltaic power generation power prediction, the accuracy of coastal area photovoltaic power generation power prediction is significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of photovoltaic power generation, in particular to a photovoltaic power generation power prediction method considering atmospheric humidity in coastal areas. BACKGROUND

[0002] The photovoltaic power generation capacity is associated with weather conditions (such as solar irradiance, temperature and cloud cover), which often leads to the unpredictability and randomness of photovoltaic power generation, causing damage to the stability of the power grid. Correspondingly, the purpose of photovoltaic power generation prediction is to predict the future photovoltaic power generation capacity with the highest accuracy, and accurate photovoltaic power generation power prediction plays a crucial role in improving grid efficiency, reducing carbon emissions and promoting the coordinated development of economy and photovoltaic power generation.

[0003] The surface meteorological system has an important influence on photovoltaic power generation, and different geographical and climatic parameters have different influences on photovoltaic power generation prediction. However, existing research lacks comprehensive coverage of a wide range of geographical and meteorological parameters. It is also a challenge to determine the most suitable meteorological parameters for accurate prediction of photovoltaic power generation capacity, especially in coastal developed areas characterized by high power demand. At the same time, existing photovoltaic power generation power prediction based on meteorological elements usually includes radiation, temperature, cloud cover and other meteorological elements, but does not focus on the influence of humidity, evapotranspiration and other atmospheric elements on photovoltaic power generation capacity, and in coastal areas, the influence of such less focused meteorological elements on photovoltaic power generation prediction cannot be ignored. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide a photovoltaic power generation power prediction method and system considering atmospheric humidity in coastal areas, which significantly improves the accuracy of photovoltaic power generation power prediction in coastal areas by introducing meteorological elements such as evapotranspiration and relative humidity highly related to the characteristics of coastal areas as the main influencing factors for short-term photovoltaic power generation power prediction.

[0005] To solve the above technical problems, the first aspect of the embodiments of the present application provides a photovoltaic power generation power prediction method considering atmospheric humidity in coastal areas, comprising the following steps:

[0006] Obtaining real-time data of first meteorological elements of the environment in which the photovoltaic power generation device in the coastal area is located in the current detection period, the first meteorological elements including relative humidity and evapotranspiration;

[0007] Based on a short-term photovoltaic power generation power prediction model, combining the real-time data of the relative humidity and the real-time data of the evapotranspiration, calculating the power prediction value of the photovoltaic power generation device in the next detection period.

[0008] Further, after obtaining the real-time data of the first meteorological elements of the environment in which the photovoltaic power generation device in the coastal area is located in the current detection period, the method further comprises:

[0009] acquire second meteorological element real-time data of an environment where the photovoltaic power generation device is located, the second meteorological element including radiation, cloud cover, temperature, rainfall, air pressure, wind speed, and / or wind direction;

[0010] In combination with the first meteorological element real-time data and the second meteorological element real-time data, calculate a power generation prediction value of the photovoltaic power generation device in a next detection period based on the short-time photovoltaic power generation power prediction model.

[0011] Further, after acquiring the first meteorological element real-time data of the environment where the photovoltaic power generation device in the coastal area is located in the current detection period, the method further includes:

[0012] normalize the real-time data of the relative humidity and the real-time data of the evapotranspiration; and obtain standard values of the relative humidity and the evapotranspiration data.

[0013] Further, acquiring the real-time data of the relative humidity of the environment where the photovoltaic power generation device in the coastal area is located in the current detection period includes:

[0014] acquire first relative humidity values of the environment where the photovoltaic power generation device is located based on a plurality of ground meteorological stations;

[0015] acquire second relative humidity values of the environment where the photovoltaic power generation device is located based on a plurality of capacitive humidity sensors arranged on the photovoltaic power generation device;

[0016] calculate the real-time data of the relative humidity of the environment where the photovoltaic power generation device is located according to the plurality of first relative humidity values and the plurality of second relative humidity values.

[0017] Further, a calculation formula of the real-time data of the relative humidity of the environment where the photovoltaic power generation device is located is:

[0018]

[0019] wherein m is the number of the ground meteorological stations, T 1i is a time attenuation factor of the i-th ground meteorological station, C 1i is a confidence coefficient of the i-th ground meteorological station, S 1i is a spatial weight factor of the i-th ground meteorological station, D 1i is a distance value between the i-th ground meteorological station and the photovoltaic power generation device, RH 1i is a humidity monitoring value of the i-th ground meteorological station, n is the number of the capacitive humidity sensors, T 2j is a time attenuation factor of the j-th capacitive humidity sensor, C 2j is a confidence coefficient of the j-th capacitive humidity sensor, S 2j is a spatial weight factor of the j-th capacitive humidity sensor, D2j is a distance value of the jth capacitive humidity sensor from the photovoltaic power generation device, and ρ ij is a correlation coefficient of the ith ground meteorological station and the jth capacitive humidity sensor, and RH 2j is a humidity value detected by the jth capacitive humidity sensor.

[0020] Further, a third relative humidity value of an environment in which the photovoltaic power generation device is located is obtained based on satellite remote sensing data.

[0021] A fourth relative humidity value of the environment in which the photovoltaic power generation device is located is obtained based on a numerical weather prediction model.

[0022] The real-time relative humidity data is calibrated by the third relative humidity value and the fourth relative humidity value.

[0023] Further, the short-time photovoltaic power generation power prediction model is a double-layer LSTM multi-step prediction model established under a Tensorflow architecture.

[0024] The double-layer LSTM multi-step prediction model comprises an input layer, a first LSTM unit and a second LSTM unit.

[0025] The input layer receives corresponding real-time meteorological data of meteorological elements and sends the real-time meteorological data to the first LSTM unit.

[0026] The first LSTM unit extracts features from the real-time meteorological data and sends first data hidden state information obtained after feature extraction to the second LSTM unit, the first data hidden state information comprising a real-time meteorological data time-varying trend of meteorological elements, a correlation between a plurality of meteorological elements and a correlation between meteorological elements and a photovoltaic power generation device output power value.

[0027] The second LSTM unit receives the first data hidden state information and extracts features again to obtain second data hidden state information.

[0028] The double-layer LSTM multi-step prediction model calculates a photovoltaic power generation device output power value based on the second data hidden state information.

[0029] The activation functions of the first LSTM unit and the second LSTM unit are both hyperbolic tangent functions.

[0030] Further, the data training set of the short-time photovoltaic power generation power prediction model comprises ECMWF Reanalysis v5 meteorological reanalysis data.

[0031] Correspondingly, a second aspect of the embodiment of the present application provides a coastal area photovoltaic power generation power prediction system considering atmospheric humidity, comprising:

[0032] a data acquisition module configured to acquire real-time data of first meteorological elements of an environment in which a photovoltaic power generation device in a coastal area is located in a current detection period, the first meteorological elements including relative humidity and evapotranspiration;

[0033] a power prediction module configured to calculate a power prediction value of the photovoltaic power generation device in a next detection period based on a short-time photovoltaic power generation power prediction model in combination with the real-time data of the relative humidity and the real-time data of the evapotranspiration.

[0034] Correspondingly, a third aspect of the embodiment of the present application also provides an electronic device, comprising at least one processor and a memory connected with the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the above-mentioned method for predicting photovoltaic power generation power in a coastal area considering atmospheric humidity.

[0035] In addition, a fourth aspect of the embodiment of the present application also provides a computer readable storage medium having computer instructions stored thereon, and the instructions are executed by a processor to implement the above-mentioned method for predicting photovoltaic power generation power in a coastal area considering atmospheric humidity.

[0036] The above technical solutions of the embodiment of the present application have the following beneficial technical effects:

[0037] 1. The historical meteorological reanalysis data is used to select evapotranspiration and relative humidity and other atmospheric temperature parameters as main influencing parameters, and other meteorological parameters are supplemented to perform high-precision prediction on photovoltaic power generation power in a coastal area.

[0038] 2. The double-layer LSTM deep learning model is used to improve the prediction effect of the photovoltaic power generation power, and the photovoltaic power generation power within 72 hours in a real application scenario is more accurately predicted, so that high-precision prediction of the power generation power is realized, and technical support and reference are provided for photovoltaic power generation grid connection technology and power grid decision-making. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 is a flow chart of a method for predicting photovoltaic power generation power in a coastal area considering atmospheric humidity provided by the embodiment of the present application;

[0040] Figure 2 is a schematic diagram of a double-layer LSTM multi-step prediction model of a Tensorflow architecture provided by the embodiment of the present application;

[0041] Figure 3is the short-time photovoltaic power prediction model before and after considering atmospheric humidity provided by the embodiment of the present application, and an evaluation schematic diagram;

[0042] Figure 4 is the photovoltaic power prediction precision comparison chart before and after considering atmospheric humidity provided by the embodiment of the present application;

[0043] Figure 5 is the block diagram of the photovoltaic power prediction system in the coastal area considering atmospheric humidity provided by the embodiment of the present application.

[0044] Reference signs:

[0045] 1, data acquisition module, 2, power prediction module. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical scheme and advantages of the present application more clear and obvious, the present application will be further described in detail below in combination with specific embodiments and with reference to the drawings. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present application. In addition, in the following description, the description of the known structures and technologies is omitted to avoid unnecessary confusion of the concept of the present application.

[0047] Please refer to Figure 1 The first aspect of the embodiment of the present application provides a photovoltaic power prediction method in the coastal area considering atmospheric humidity, comprising the following steps:

[0048] Step S100, acquiring real-time data of first meteorological elements of the environment in which the photovoltaic power generation device in the coastal area is located in the current detection period, the first meteorological elements including relative humidity and evapotranspiration.

[0049] Step S300, based on the short-time photovoltaic power prediction model, combining the real-time data of relative humidity and the real-time data of evapotranspiration, calculating the power prediction value of the photovoltaic power generation device in the next detection period.

[0050] For the coastal area, the meteorological elements closely related to the special topography and climate of the coastal area should be considered and taken as the main factor of prediction, such as evapotranspiration and relative humidity, so as to improve the accuracy of photovoltaic power prediction in the coastal area. Then, the classic time series deep learning prediction model LSTM is selected to perform short-term photovoltaic power prediction, so as to realize high-precision power prediction of large photovoltaic power stations.

[0051] In one embodiment of the present application, after the step S100 of acquiring the real-time data of the first meteorological elements of the environment in which the photovoltaic power generation device in the coastal area is located in the current detection period, it further comprises:

[0052] Step S210, obtaining real-time data of second meteorological elements of the environment where the photovoltaic power generation device is located, the second meteorological elements including radiation, cloud cover, temperature, rainfall, air pressure, wind speed and / or wind direction.

[0053] Step S220, combining the real-time data of the first meteorological elements and the real-time data of the second meteorological elements, calculating the predicted value of the power generation of the photovoltaic power generation device in the next detection period based on the short-term photovoltaic power generation power prediction model.

[0054] By comprehensively considering two first meteorological elements and several second meteorological elements, the fluctuation rule of the power generation in different time scales, such as the daily variation and the seasonal variation, can be better understood, and how various meteorological elements interact to cause the power generation to change; and the change of different meteorological scenes can be adapted, and the influence of single factor error can be reduced.

[0055] Further, after obtaining the real-time data of the first meteorological elements of the environment where the photovoltaic power generation device in the coastal area is located in the current detection period in step S100, it further includes:

[0056] Step S110, normalizing the real-time data of relative humidity and the real-time data of evapotranspiration; obtaining the standard value of the relative humidity and evapotranspiration data.

[0057] In the photovoltaic power generation power prediction model, the data unit and magnitude of the relative humidity and evapotranspiration are usually different from other meteorological elements that can be used for prediction (such as temperature, radiation intensity, etc.). The relative humidity is a proportional value, ranging from 0 to 100%, while the evapotranspiration can be measured in different units (such as millimeters per day, etc.). If normalization is not performed, these different magnitudes of data will cause the characteristics with larger values to have an excessive influence on the results in the model training and calculation process, while the characteristics with smaller values may be hidden. After normalization, all data are converted to a relatively unified range, so that the relative humidity and evapotranspiration data can be operated and compared with other meteorological element data on the same scale.

[0058] In addition, normalization helps to improve the generalization ability of the model, that is, the prediction performance of the model on unseen data. When the model faces new relative humidity and evapotranspiration data, if the distribution of these data is too different from the training data (due to dimensional factors, etc.), the model may produce inaccurate predictions. Normalization processing makes the model more robust to the distribution of data, and can better adapt to data changes in different environments, so that the photovoltaic power generation power can be more stably predicted in different coastal areas or different seasons, etc.

[0059] Further, the obtaining of the real-time data of the relative humidity of the environment where the photovoltaic power generation device in the coastal area is located in the current detection period in step S100 includes:

[0060] Step S101, respectively acquiring first relative humidity values of the environment where the photovoltaic power generation device is located based on a plurality of ground meteorological stations.

[0061] Step S102, acquiring second relative humidity values of the environment where the photovoltaic power generation device is located based on a plurality of capacitive humidity sensors arranged on the photovoltaic power generation device.

[0062] Step S103, calculating real-time data of the relative humidity of the environment where the photovoltaic power generation device is located according to the plurality of first relative humidity values and the plurality of second relative humidity values.

[0063] The relative humidity values are acquired by both ground meteorological stations and capacitive humidity sensors, and multiple data are acquired by each method, which can effectively reduce the influence of errors in a single data source on the final result. The measurement of the ground meteorological station may be affected by factors such as changes in the surrounding environment, instrument calibration errors, or data transmission interference. For example, the construction of a new building nearby may change the airflow around the meteorological station, causing the measured relative humidity to deviate. Similarly, the capacitive humidity sensor installed on the photovoltaic power generation device may produce errors due to factors such as local heat dissipation, dust accumulation, or performance fluctuations. When there are multiple data sources, the error of individual data points will not greatly interfere with the final real-time relative humidity data, because it can be corrected and balanced by other data.

[0064] In the long-term monitoring process, a single data source may fail or have missing data. For example, the ground meteorological station may temporarily be unable to provide data due to equipment maintenance, power supply problems, or communication failures. If only a single ground meteorological station data is relied on, the monitoring of relative humidity will be interrupted once this happens. By using multiple methods and acquiring multiple values, even if some data is missing or unavailable, the real-time relative humidity data can still be calculated using other available data, ensuring the continuity and reliability of the data. This redundant data acquisition method can provide more stable data support for photovoltaic power generation power prediction in complex coastal environments.

[0065] In one specific embodiment of the present application, the calculation formula of the real-time data of the relative humidity of the environment where the photovoltaic power generation device is located is:

[0066]

[0067] where m is the number of ground meteorological stations, T 1i is the time decay factor of the i-th ground meteorological station, C 1i is the confidence coefficient of the i-th ground meteorological station, S 1i is the spatial weight factor of the i-th ground meteorological station, and D 1iLet RH be the distance between the i-th ground weather station and the photovoltaic power generation device. 1i Let T be the humidity monitoring value of the i-th ground weather station, n be the number of capacitive humidity sensors, and T be the humidity value of the i-th ground weather station. 2j Let C be the time decay factor of the j-th capacitive humidity sensor. 2j Let S be the confidence coefficient of the j-th capacitive humidity sensor. 2j Let D be the spatial weighting factor for the j-th capacitive humidity sensor. 2j Let ρ be the distance between the j-th capacitive humidity sensor and the photovoltaic power generation device. ij Let RH be the correlation coefficient between the i-th ground weather station and the j-th capacitive humidity sensor. 2j The humidity value detected by the j-th capacitive humidity sensor.

[0068] In addition, after step S102, this prediction method also includes:

[0069] Step S102a: Obtain the third relative humidity value of the environment where the photovoltaic power generation device is located based on satellite remote sensing data.

[0070] Satellite remote sensing data offers the advantage of wide-area coverage, enabling the acquisition of atmospheric humidity information over large areas. In coastal regions, it can capture macroscopic features such as the transition of humidity between the ocean and land, and the impact of large weather systems on humidity distribution. In contrast, humidity sensors at ground-based meteorological stations and photovoltaic devices primarily focus on data collection from localized areas. By incorporating satellite remote sensing data, this macroscopic humidity distribution information can be combined with localized measurements, allowing real-time relative humidity data to better reflect the true humidity status of the entire region. For example, when sea fog occurs in coastal areas, satellite remote sensing can promptly detect large areas of abnormal humidity, thereby correcting local measurements, supplementing potentially missing information, and improving the accuracy of humidity values.

[0071] Step S102b: Obtain the fourth relative humidity value of the environment where the photovoltaic power generation device is located based on the numerical weather prediction model.

[0072] Numerical weather prediction models can calculate humidity changes over a future period based on atmospheric physical processes and initial meteorological conditions. They comprehensively consider the interactions between various meteorological factors, such as the effects of temperature, air pressure, wind speed, and water vapor transport on humidity. In the calibration of real-time relative humidity data, this predictive humidity information can help correct biases in current measurements. For example, if a numerical weather prediction model predicts a sharp drop in humidity due to an approaching cold front, and current ground measurements and satellite remote sensing data have not yet reflected this change, then the real-time humidity data can be appropriately adjusted based on the model's prediction, making the data more forward-looking and accurate.

[0073] Step S102c, calibrating the relative humidity real-time data by the third relative humidity value and the fourth relative humidity value.

[0074] Please refer to Figure 2 The prediction method can obviously improve the short-term photovoltaic power prediction in the coastal area by introducing meteorological elements such as evapotranspiration and relative humidity which are highly related to the characteristics of the region as the main influencing factors of the short-term photovoltaic power prediction in the coastal area, and inputting the prediction model as input parameters. The accuracy comparison is shown in Figure 2 .

[0075] In an embodiment of the present application, please refer to Figure 3 The short-term photovoltaic power prediction model is a double-layer LSTM multi-step prediction model established under the Tensorflow architecture. The learning rate, batchsize, epochs and other parameter settings are reasonably adjusted by using the Adam optimizer, and the LSTM model is iteratively trained to ensure that it can accurately capture the dynamic changes and potential laws in the data, and realize the short-term prediction of photovoltaic power (within 72 hours) by using evaporation and other meteorological elements. Finally, the test set is also used for accuracy evaluation, and R 2 , RMSE and other indicators are used to quantitatively evaluate the prediction accuracy and error size.

[0076] The double-layer LSTM multi-step prediction model includes an input layer, a first LSTM unit and a second LSTM unit. The input layer receives the corresponding real-time meteorological data of the meteorological elements and sends them to the first LSTM unit. The first LSTM unit extracts features from the real-time meteorological data and sends the first data hidden state information after feature extraction to the second LSTM unit. The first data hidden state information includes the time-varying trend of the real-time meteorological data of the meteorological elements, the correlation between the meteorological elements, and the correlation between the meteorological elements and the output power value of the photovoltaic power generation device. The second LSTM unit receives the first data hidden state information and extracts features again to obtain second data hidden state information. The double-layer LSTM multi-step prediction model calculates the output power value of the photovoltaic power generation device based on the second data hidden state information.

[0077] The activation functions of the first LSTM unit and the second LSTM unit are both hyperbolic tangent functions.

[0078] Long Short-Term Memory (LSTM) is a recursive neural network (RNN) algorithm for deep learning, which has better effect on processing time series data. LSTM inherits the recursive characteristics of RNN, fully utilizes time series data, and makes up for the shortcomings of RNN such as gradient disappearance and gradient explosion, and insufficient long-term memory ability.

[0079] The basic structure mainly includes an input gate, a forget gate, an output gate and a memory unit. Among them, x t represents the input, y t represents the output, h t and c t can be regarded as short-term state and long-term state, g t is a candidate value. The input gate accepts input information and updates the state of the memory unit according to different conditions. The forget gate determines the information to be discarded according to the specific condition, and the output gate determines the output content according to the input information and the memory unit. The working process of the LSTM unit is as follows: at each time, the forget gate receives the current state x t and the hidden layer state h t-1 of the previous moment, and the output value of the forget gate is mapped to the interval [0, 1] through the activation function σ. When the output of the forget gate is 0, it means that the information of the previous state is completely discarded; when the output is 1, the information of the previous state is completely retained. The input of the input gate is transformed through a nonlinear function, and is superimposed with the output of the forget gate to obtain the updated memory unit c t . Finally, the output gate controls the output h t (y t ) of the LSTM according to c t according to the operation of the nonlinear function.

[0080] i t =σ(W xi x t +W hi h t-1 +W ci c t-1 +b i );

[0081] f t =σ(W xf x t +W hf h t-1 +W cf c t-1 +b f );

[0082] o t =σ(W xo x t +W ho h t-1 +W co c t +b o );

[0083] gt = tanh(W xc x t + W hc h t-1 + b c );

[0084] c(t) = f t c t-1 + i t g t ;

[0085] h t = y t = o t tanh(c t ).

[0086] In weather forecasting, LSTM can use past weather data to build a model and predict future weather conditions through training. Weather landscape short-term forecasting refers to predicting weather changes in the next few hours. Compared with traditional weather forecasting methods, LSTM-based short-term forecasting can more accurately capture weather change trends, especially in cases where weather changes are more dramatic in a short period of time, which can better improve the accuracy of forecasts.

[0087] Further, the data training set of the short-term photovoltaic power generation prediction model comprises ECMWF Reanalysis v5 meteorological reanalysis data.

[0088] Please refer to Figure 4 , the short-term photovoltaic power generation prediction model selects ECMWF Reanalysis v5 (ERA5) meteorological reanalysis data, and uses historical reanalysis data as input parameters. Through preliminary screening, 11 meteorological elements including evaporation, relative humidity, radiation, cloud cover, temperature, etc. are selected as input parameters of the prediction model. The accuracy improvement effect after supplementing meteorological elements such as evaporation and relative humidity is shown in Figure 4 , and the output data is the 15min resolution power generation data of a power plant in the coastal area.

[0089] In addition, in order to exclude the influence of different seasons, the data of the previous 21 days of each month is selected for training, and the data of the remaining days is selected for testing, to ensure that the model has reliability and accuracy in predicting changes in surface shortwave radiation parameters.

[0090] Correspondingly, please refer to Figure 5 , the second aspect of the embodiment of the present application provides a photovoltaic power generation power prediction system considering atmospheric humidity in a coastal area, comprising:

[0091] a data acquisition module 1 configured to acquire real-time data of first meteorological elements of an environment in which a photovoltaic power generation device in a coastal area is located in a current detection period, the first meteorological elements including relative humidity and evapotranspiration;

[0092] a power prediction module 2 configured to calculate a predicted value of power generation of the photovoltaic power generation device in a next detection period based on a short-time photovoltaic power generation power prediction model in combination with the real-time data of the relative humidity and the real-time data of the evapotranspiration.

[0093] Correspondingly, a third aspect of the embodiment of the present application also provides an electronic device, comprising: at least one processor; and a memory connected with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the above-mentioned method for predicting photovoltaic power generation power in a coastal area considering atmospheric humidity.

[0094] In addition, a fourth aspect of the embodiment of the present application also provides a computer readable storage medium having computer instructions stored thereon, and the instructions are executed by a processor to implement the above-mentioned method for predicting photovoltaic power generation power in a coastal area considering atmospheric humidity.

[0095] The embodiment of the present application aims to protect a method and system for predicting photovoltaic power generation power in a coastal area considering atmospheric humidity, wherein the method comprises the following steps: acquiring real-time data of first meteorological elements of an environment in which a photovoltaic power generation device in a coastal area is located in a current detection period, the first meteorological elements including relative humidity and evapotranspiration; and calculating a predicted value of power generation of the photovoltaic power generation device in a next detection period based on a short-time photovoltaic power generation power prediction model in combination with the real-time data of the relative humidity and the real-time data of the evapotranspiration.

[0096] The above-mentioned technical solution has the following effects:

[0097] 1. The historical meteorological reanalysis data is used to select evapotranspiration and relative humidity and other atmospheric temperature parameters as main influencing parameters, and other meteorological parameters are supplemented to perform high-precision prediction of photovoltaic power generation power in a coastal area;

[0098] 2. The double-layer LSTM deep learning model is used to improve the prediction effect of the photovoltaic power generation power, and the photovoltaic power generation power within 72 hours in a real application scenario is more accurately predicted, so that high-precision prediction of power generation power is realized, and technical support and reference are provided for photovoltaic power generation grid connection technology and power grid decision-making.

[0099] Those skilled in the art will appreciate that embodiments of the application can be devised for a method, a system, or a computer program product. Accordingly, the present application can be embodied in the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0100] The present application is described in reference to the 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 block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can 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 illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0101] These computer program instructions can 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 instructions which implement the functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0102] These computer program instructions can 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 illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams. Figure 1 one or more functions specified in the flowchart illustrations and / or block diagrams.

[0103] Finally, it should be noted that the above-mentioned embodiments are merely intended for describing and illustrating, not limiting, the technical solutions of the present application. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that the specific embodiments of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and any modification or equivalent replacement without departing from the spirit and scope of the present application should be covered in the protection scope of the claims of the present application.

Claims

1. A method for predicting photovoltaic power generation in coastal areas taking into account atmospheric humidity, characterized in that, Includes the following steps: Acquire real-time data of the first meteorological element of the environment in which the photovoltaic power generation device in the coastal area is located during the current detection period. The first meteorological element includes: relative humidity and evapotranspiration. Based on the short-time photovoltaic power generation prediction model, combined with the real-time relative humidity data and the real-time evaporation data, the predicted power generation value of the photovoltaic power generation device in the next detection cycle is calculated. After obtaining the real-time data of the first meteorological element of the environment where the photovoltaic power generation device in the coastal area is located during the current detection period, the method further includes: The real-time data of the second meteorological element of the environment in which the photovoltaic power generation device is located are obtained. The second meteorological element includes: radiation, cloud cover, temperature, rainfall, air pressure, wind speed and / or wind direction. Combining the real-time data of the first meteorological element and the real-time data of the second meteorological element, and based on the short-term photovoltaic power generation prediction model, the predicted power generation value of the photovoltaic power generation device in the next detection cycle is calculated; Obtain real-time relative humidity data for the environment where photovoltaic power generation devices are located in the coastal area during the current monitoring period, including: The first relative humidity value of the environment where the photovoltaic power generation device is located is obtained based on several ground meteorological stations; The second relative humidity value of the environment in which the photovoltaic power generation device is located is obtained based on a number of capacitive humidity sensors installed on the photovoltaic power generation device; Based on several first relative humidity values ​​and several second relative humidity values, calculate the real-time relative humidity data of the environment in which the photovoltaic power generation device is located; The formula for calculating the real-time relative humidity data of the environment in which the photovoltaic power generation device is located is as follows: ； ； in, The number of ground weather stations. Let be the time decay factor for the i-th ground weather station. Let be the confidence coefficient for the i-th ground weather station. Let i be the spatial weighting factor for the i-th surface meteorological station. Let be the distance between the i-th ground weather station and the photovoltaic power generation device. Let be the humidity monitoring value of the i-th ground meteorological station. This refers to the number of capacitive humidity sensors. Let be the time decay factor of the j-th capacitive humidity sensor. Let be the confidence coefficient of the j-th capacitive humidity sensor. Let be the spatial weighting factor for the j-th capacitive humidity sensor. Let j be the distance between the j-th capacitive humidity sensor and the photovoltaic power generation device. Let be the correlation coefficient between the i-th ground weather station and the j-th capacitive humidity sensor. The humidity value detected by the j-th capacitive humidity sensor.

2. The method for predicting photovoltaic power generation in coastal areas taking into account atmospheric humidity as described in claim 1, characterized in that, After obtaining the real-time data of the first meteorological element of the environment where the photovoltaic power generation device in the coastal area is located during the current detection period, the method further includes: The real-time data of relative humidity and the real-time data of evapotranspiration are normalized to obtain standard values ​​for the relative humidity and the evapotranspiration data.

3. The method for predicting photovoltaic power generation in coastal areas taking into account atmospheric humidity as described in claim 1, characterized in that, Also includes: The third relative humidity value of the environment in which the photovoltaic power generation device is located is obtained based on satellite remote sensing data; The fourth relative humidity value of the environment where the photovoltaic power generation device is located is obtained based on the numerical weather prediction model; The real-time relative humidity data is calibrated using the third and fourth relative humidity values.

4. The method for predicting photovoltaic power generation in coastal areas taking into account atmospheric humidity according to any one of claims 1-3, characterized in that, The short-time photovoltaic power generation prediction model is a two-layer LSTM multi-step prediction model built under the Tensorflow architecture. The two-layer LSTM multi-step prediction model includes: an input layer, a first LSTM unit, and a second LSTM unit; The input layer receives real-time meteorological data corresponding to meteorological elements and sends it to the first LSTM unit; The first LSTM unit extracts features from the real-time meteorological data and sends the extracted first data hidden state information to the second LSTM unit. The first data hidden state information includes: the real-time meteorological data of meteorological elements changing over time, the correlation between several meteorological elements, and the correlation between meteorological elements and the output power value of photovoltaic power generation devices. The second LSTM unit receives the first data hiding state information and performs feature extraction again to obtain the second data hiding state information; The dual-layer LSTM multi-step prediction model calculates the output power value of the photovoltaic power generation device based on the second data hidden state information. The activation functions of the first LSTM unit and the second LSTM unit are both hyperbolic non-tangent functions.

5. The method for predicting photovoltaic power generation in coastal areas taking into account atmospheric humidity as described in claim 4, characterized in that, The training set for the short-time photovoltaic power generation prediction model includes: ECMWF Reanalysis v5 meteorological reanalysis data.

6. A photovoltaic power generation prediction system for coastal areas taking into account atmospheric humidity, characterized in that, The method for predicting photovoltaic power generation in coastal areas based on any one of claims 1-5, taking into account atmospheric humidity, is used to predict photovoltaic power generation in coastal areas, including: The data acquisition module is used to acquire real-time data of the first meteorological element of the environment in which the photovoltaic power generation device in the coastal area is located during the current detection period. The first meteorological element includes: relative humidity and evapotranspiration. The power prediction module is used to calculate the predicted power generation value of the photovoltaic power generation device in the next detection cycle based on the short-time photovoltaic power generation prediction model, combined with the real-time relative humidity data and the real-time evapotranspiration data.

7. An electronic device, characterized in that, include: At least one processor; And a memory connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the at least one processor to perform the coastal photovoltaic power generation prediction method taking into account atmospheric humidity as described in any one of claims 1-5.

Images