Numerical weather forecast data updating method and system

Abstract

The invention relates to the technical field of new energy photovoltaic power stations, and provides a numerical weather forecast data updating method and system, and the method comprises the steps: obtaining multi-source weather data and power data; carrying out downscaling processing on the meteorological satellite cloud picture data, and carrying out initial weather prediction to obtain initial weather forecast data; carrying out secondary fusion processing on the initialized meteorological forecast data, the meteorological station data and the sky imager monitoring data, and then carrying out secondary meteorological prediction to obtain multi-source meteorological fusion data; performing three-time fusion processing on the multi-source meteorological fusion data and the power data, and then performing three-time meteorological prediction to obtain historical available meteorological forecast data; sky imager monitoring data and meteorological station data in the next four hours are obtained in real time, historical available weather forecast data are corrected through the obtained data, and real-time available weather forecast data are obtained. According to the invention, high-precision numerical weather forecast data can be obtained, so that the power prediction accuracy of the photovoltaic power station is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy photovoltaic power stations, and in particular to a method and system for updating numerical weather forecast data. Background Art

[0002] With the gradual implementation of the new energy power consumption target under the "dual carbon" goal, in order to ensure the safe and stable operation of the power grid, various regions have increasingly stringently assessed the power forecast of new energy power stations, and put forward higher requirements for the power forecast accuracy, power forecast duration and power forecast data category of new energy power generation. The assessment penalty mechanism has been further clarified and strengthened. The accuracy of power forecast will directly affect the operation and profitability of the power station, and the accuracy and update frequency of numerical meteorological forecast data directly determine the accuracy of power forecast. In the context of power marketization, the customization of electricity prices depends on accurate production forecasts. Accurate power forecasts will become an important bargaining chip to help power generation groups obtain satisfactory electricity prices. As the proportion of new energy in the power grid increases, the fluctuation range of electricity prices has gradually increased, which has brought uncontrollable effects on power market participants. Accurate forecasts of quantity and price can enable new energy power generation groups and new energy power stations participating in power transactions to understand future power generation data and reasonably allocate electricity in short-term spot and medium- and long-term power transactions.

[0003] However, the current traditional solutions have a single data input source, so they cannot effectively improve the accuracy and update frequency of numerical meteorological forecast data, resulting in low accuracy in photovoltaic power station power generation prediction and prediction duration. Summary of the invention

[0004] The purpose of the present invention is to solve at least one technical problem in the background technology and to provide a method and system for updating numerical weather forecast data.

[0005] To achieve the above object, the present invention provides a method for updating numerical weather forecast data, comprising:

[0006] Acquire and store multi-source meteorological data and power data in the photovoltaic power station area through the ground-satellite joint data management cloud platform;

[0007] Preprocess multi-source meteorological data and power data;

[0008] Downscaling the meteorological satellite cloud image data in the multi-source meteorological observation data, and using a model forecasting method to perform an initial meteorological forecast to obtain initial meteorological forecast data;

[0009] Performing secondary fusion processing on the initialized meteorological forecast data and the meteorological observation station data and sky imager monitoring data in the multi-source meteorological data, and then performing secondary meteorological forecasting to obtain multi-source meteorological fusion data;

[0010] The multi-source meteorological fusion data and power data are subjected to three fusion processes, and then three meteorological forecasts are performed to obtain historically available meteorological forecast data;

[0011] The sky imager monitoring data and the meteorological observation station data for the next 4 hours are acquired in real time, and the historical available meteorological forecast data are corrected by the acquired data to obtain the real-time available meteorological forecast data.

[0012] According to one aspect of the present invention, the ground-satellite joint data management cloud platform acquires and stores multi-source meteorological data and power data in the photovoltaic power station area, including:

[0013] An operating system and a relational database are installed in the Earth-Satellite Joint Data Management Cloud Platform to establish a fixed channel for storing multi-source meteorological observation data and power data;

[0014] The ground-to-satellite joint data management cloud platform establishes data transmission channels with meteorological satellites, optical power prediction systems, meteorological observation stations, and sky imagers to regularly obtain multi-source meteorological data and power data from various data sources.

[0015] According to one aspect of the present invention, the preprocessing of multi-source meteorological data and power data includes:

[0016] Read the meteorological satellite cloud image data from the multi-source meteorological data in the photovoltaic power station area through the meteorological satellite dedicated channel, perform integrity and rationality verification, bad data removal, missing and abnormal data smoothing and supplementation preprocessing on the meteorological satellite cloud image data, and store the preprocessed meteorological satellite cloud image data in the relational database of the ground-satellite joint data management cloud platform;

[0017] The meteorological observation station data and sky imager monitoring data are preprocessed to check integrity and rationality, remove bad data, and smooth and supplement missing and abnormal data, and the preprocessed meteorological observation station data and sky imager monitoring data are stored in the relational database of the ground-satellite joint data management cloud platform;

[0018] The power data set is preprocessed by checking its integrity and rationality, removing bad data, and smoothing and supplementing missing and abnormal data. The preprocessed power data set is stored in the relational database of the Earth-Satellite Joint Data Management Cloud Platform.

[0019] According to one aspect of the present invention, downscaling is performed on the meteorological satellite cloud image data in the multi-source meteorological observation data, and a model forecasting method is used to perform an initial meteorological forecast to obtain initial meteorological forecast data, including:

[0020] The original numerical weather forecast NWP data were obtained from the US Global Weather Forecast System (GFS) and the European Centre for Medium-Range Weather Forecasts (ECMWF). The two sets of numerical weather forecast NWP data were first checked for rationality according to the local historical measured data, and the abnormal data were removed. Then the average of the two sets of numerical weather forecast NWP data was taken.

[0021] The average value data is downscaled using the Uformer model and smoothing interpolation algorithm in the Transformer model to obtain the initial meteorological forecast data with a forecast time scale of 15 minutes.

[0022] According to one aspect of the present invention, the initialized meteorological forecast data and the meteorological observation station data and the sky imager monitoring data in the multi-source meteorological data are subjected to secondary fusion processing, and then secondary meteorological prediction is performed to obtain multi-source meteorological fusion data, including:

[0023] Build data fusion prediction model;

[0024] Input the initialized meteorological forecast data, meteorological observation station data, and sky imager monitoring data into the data fusion prediction model, and output the original multi-source meteorological fusion data;

[0025] With reference to the data from meteorological observation stations, the original multi-source meteorological fusion data is corrected for correlation, the large-scale deviation is corrected linearly, and the multi-source meteorological fusion data is output.

[0026] According to one aspect of the present invention, the multi-source meteorological fusion data and power data are fused three times, and then meteorological forecasts are performed three times to obtain historically available meteorological forecast data, including:

[0027] The power data in the photovoltaic power station area pre-processed in the relational database of the joint data management cloud platform of the Earth and the Star is integrated and processed and predicted with the multi-source fusion meteorological data input data fusion prediction model to obtain the initial available meteorological forecast data;

[0028] The initial available weather forecast data set is smoothed and abnormal data is cleaned to obtain historical available weather forecast data.

[0029] According to one aspect of the present invention, sky imager monitoring data and meteorological observation station data for the next 4 hours are acquired in real time, and the historical available meteorological forecast data are corrected by the acquired data to obtain real-time available meteorological forecast data, including:

[0030] Every hour, the sky imager monitoring data and meteorological observation station data of the current time and the past 24 hours are used to predict the sky imager monitoring data and meteorological observation station data of the next 4 hours using the convolutional long short-term memory network model to obtain the sky imager monitoring data and meteorological observation station data of the next 4 hours;

[0031] Then, using the weather observation station data and sky imager monitoring data for the next 4 hours, the BP neural network model is used to correct the historical available weather forecast data to obtain real-time available weather forecast data, including:

[0032] The BP neural network model was obtained by training the available weather forecast data and sky imager monitoring data of the past three months as the input data of the BP neural network, and the weather observation station data of the past three months as the output data of the BP neural network.

[0033] Every hour, the sky imager monitoring data for the next 4 hours and the historical available weather forecast data are used as the input data of the BP neural network model, and the original numerical weather forecast NWP data output by the model is obtained through the BP neural network;

[0034] The original numerical weather forecast NWP data and the weather observation station data for the next 4 hours are weighted to obtain real-time available weather forecast data, where the weighting coefficient is dynamically adjusted according to the deviation rate between the original numerical weather forecast NWP data and the real-time weather observation station data.

[0035] To achieve the above object, the present invention further provides a numerical weather forecast data updating system, comprising:

[0036] The data acquisition module acquires and stores multi-source meteorological data and power data in the photovoltaic power station area through the ground-satellite joint data management cloud platform;

[0037] Data preprocessing module, which preprocesses multi-source meteorological data and power data;

[0038] An initial meteorological forecast data prediction module performs downscaling processing on the meteorological satellite cloud image data in the multi-source meteorological observation data, and uses a model forecasting method to perform an initial meteorological forecast to obtain initial meteorological forecast data;

[0039] A multi-source meteorological fusion data prediction module performs secondary fusion processing on the initialized meteorological forecast data and the meteorological observation station data and sky imager monitoring data in the multi-source meteorological data, and then performs secondary meteorological prediction to obtain multi-source meteorological fusion data;

[0040] A historically available meteorological forecast data prediction module performs three fusion processes on the multi-source meteorological fusion data and power data, and then performs three meteorological forecasts to obtain historically available meteorological forecast data;

[0041] The real-time available weather forecast data update module obtains the sky imager monitoring data and weather observation station data for the next 4 hours in real time, and corrects the historical available weather forecast data through the acquired data to obtain the real-time available weather forecast data.

[0042] To achieve the above objectives, the present invention also provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the numerical weather forecast data updating method as described above.

[0043] To achieve the above object, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the numerical meteorological forecast data updating method as described above is implemented.

[0044] According to the solution of the present invention, the present invention integrates multiple sources of meteorological satellite cloud map data, sky imager data, meteorological observation station data, and real-time power data of photovoltaic power stations, and adopts a ground-satellite joint approach. High-precision numerical meteorological forecast data assists in the accurate prediction of photovoltaic power station power. The accurate power prediction data in turn reversely corrects the numerical meteorological forecast data, solving the current problem of a single input source and low accuracy of numerical meteorological forecast data, and can greatly improve the accuracy of photovoltaic power station power prediction.

[0045] The present invention is applicable to photovoltaic power stations, and can improve the prediction accuracy and update frequency of numerical meteorological forecast data, thereby improving the accuracy of photovoltaic power prediction, helping the power system dispatching department to coordinate the coordination of conventional energy and photovoltaic power generation, timely adjust the dispatching plan, and reasonably arrange the operation mode of the power grid. On the one hand, it effectively reduces the impact of photovoltaic access on the power grid and improves the safety and stability of power grid operation. On the other hand, it reduces the rotating reserve and operating costs of the power system, so as to make full use of solar energy resources and obtain greater economic and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A flowchart schematically showing a method for updating numerical weather forecast data according to an embodiment of the present invention;

[0047] Figure 2 A diagram schematically shows a data fusion prediction model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The present invention will now be discussed with reference to exemplary embodiments. It should be understood that the embodiments discussed are only to enable those skilled in the art to better understand and thus implement the present invention, rather than implying any limitation on the scope of the present invention.

[0049] As used herein, the term "including" and variations thereof are to be interpreted as open-ended terms meaning "including, but not limited to." The term "based on" is to be interpreted as "based, at least in part, on." The terms "one embodiment" and "an embodiment" are to be interpreted as "at least one embodiment."

[0050] Figure 1 The flowchart schematically shows a method for updating numerical weather forecast data according to an embodiment of the present invention. Figure 1 As shown, in this embodiment, the numerical weather forecast data updating method includes:

[0051] Acquire and store multi-source meteorological data and power data in the photovoltaic power station area through the ground-satellite joint data management cloud platform;

[0052] Preprocess multi-source meteorological data and power data;

[0053] The meteorological satellite cloud image data in the multi-source meteorological observation data are downscaled, and the model forecast method is used to make the initial meteorological forecast to obtain the initial meteorological forecast data;

[0054] Perform secondary fusion processing on the initialization meteorological forecast data and the meteorological observation station data and sky imager monitoring data in the multi-source meteorological data, and then perform secondary meteorological forecast to obtain multi-source meteorological fusion data;

[0055] The multi-source meteorological fusion data and power data are fused three times, and then three meteorological forecasts are performed to obtain historically available meteorological forecast data;

[0056] The sky imager monitoring data and meteorological observation station data for the next 4 hours are acquired in real time, and the historically available meteorological forecast data are corrected by the acquired data to obtain real-time available meteorological forecast data.

[0057] In this embodiment, the multi-source meteorological data includes meteorological satellite cloud image data, meteorological observation station data, and sky imager monitoring data.

[0058] Further, according to an embodiment of the present invention, the ground-satellite joint data management cloud platform acquires and stores multi-source meteorological data and power data in the photovoltaic power station area, including:

[0059] An operating system and a relational database are installed in the Earth-Satellite Joint Data Management Cloud Platform to establish a fixed channel for storing multi-source meteorological observation data and power data;

[0060] The ground-to-satellite joint data management cloud platform establishes data transmission channels with meteorological satellites, optical power prediction systems, meteorological observation stations, and sky imagers to regularly obtain multi-source meteorological data and power data from various data sources.

[0061] Further, according to an embodiment of the present invention, preprocessing of multi-source meteorological data and power data includes:

[0062] The ground-satellite joint data management cloud platform reads the meteorological satellite cloud image data from the multi-source meteorological data in the photovoltaic power station area through the meteorological satellite dedicated channel, performs integrity and rationality verification, bad data removal, missing and abnormal data smoothing and supplementation preprocessing on the meteorological satellite cloud image data, and stores the preprocessed meteorological satellite cloud image data in the relational database of the ground-satellite joint data management cloud platform;

[0063] The ground-satellite joint data management cloud platform establishes a communication channel with the meteorological observation station and the sky imager to obtain the meteorological observation station data and the sky imager monitoring data, and then performs integrity and rationality verification, bad data removal, missing and abnormal data smoothing and supplementation preprocessing on the meteorological observation station data and the sky imager monitoring data, and stores the preprocessed meteorological observation station data and the sky imager monitoring data in the relational database of the ground-satellite joint data management cloud platform;

[0064] The Earth-Satellite Joint Data Management Cloud Platform establishes a communication channel with the power prediction system to obtain the power data of the photovoltaic power station, and then performs integrity and rationality verification, bad data elimination, missing and abnormal data smoothing and supplementation preprocessing on the power data set, and stores the preprocessed power data set in the relational database of the Earth-Satellite Joint Data Management Cloud Platform.

[0065] Further, according to an embodiment of the present invention, downscaling is performed on meteorological satellite cloud image data in multi-source meteorological observation data, and a model forecasting method is used to perform an initial meteorological forecast to obtain initial meteorological forecast data, including:

[0066] The original numerical weather forecast NWP data were obtained from the US Global Weather Forecast System (GFS) and the European Centre for Medium-Range Weather Forecasts (ECMWF). The two sets of numerical weather forecast NWP data were first checked for rationality according to the local historical measured data, and the abnormal data were removed. Then the average of the two sets of numerical weather forecast NWP data was taken.

[0067] The average value data is downscaled using the Uformer model and smoothing interpolation algorithm in the Transformer model to obtain the initial meteorological forecast data with a forecast time scale of 15 minutes.

[0068] Further, according to an embodiment of the present invention, the initialized meteorological forecast data and the meteorological observation station data and the sky imager monitoring data in the multi-source meteorological data are subjected to secondary fusion processing, and then secondary meteorological prediction is performed to obtain multi-source meteorological fusion data, including:

[0069] Build data fusion prediction model;

[0070] Input the initialized meteorological forecast data, meteorological observation station data, and sky imager monitoring data into the data fusion prediction model, and output the original multi-source meteorological fusion data;

[0071] With reference to the data from meteorological observation stations, the original multi-source meteorological fusion data is corrected for correlation, the large-scale deviation is corrected linearly, and the multi-source meteorological fusion data is output.

[0072] In this embodiment, a data fusion prediction model is constructed as follows: Figure 2 As shown, the specific principles are as follows:

[0073] Sum: y′=(W1*X1+W2*X2+W3*X3+b);

[0074] Activation function: provides nonlinear fitting capability y=fy′;

[0075] in:

[0076] (X1, X2, X3) represents the external input signal, which refers to multiple attributes of a training data sample, that is, the total radiation, diffuse radiation, direct radiation and other data corresponding to the initial weather forecast data. b is the offset value, which enables the straight line to move up and down along the Y axis.

[0077] (W1, W2, W3) is the weight of each input signal. During the model training process, different weight values ​​are assigned according to the different synaptic strengths;

[0078] Input: simulates the input from other neurons in biological neural networks;

[0079] Weight: To simulate the different synaptic strengths received by each neuron in a biological neural network, the input received from the outside world is multiplied by a certain weight;

[0080] Sum: simulate the accumulation of signals received by neurons in biological neural networks;

[0081] Activation function: simulates the action potential generated by the accumulation of signals to a certain extent in biological neural networks. When the accumulation reaches a certain level, the action potential will be "activated". This method uses the Sigmoid function;

[0082] Output: simulates the new signals released by neurons in biological neural networks.

[0083] The training process is to continuously update the weight W and offset b until stable W and b are found to minimize the overall error of the model.

[0084] In this embodiment, the model training input is the total radiation, diffuse radiation, direct radiation, temperature, wind speed, wind direction, humidity, and cloud cover in the initial weather forecast data, and the cloud condition changes, cloud cover trends, and sunshine duration ratio in the sky imager monitoring data; the output is the total radiation, diffuse radiation, direct radiation, temperature, wind speed, wind direction, humidity, and cloud cover in the meteorological observation station data, thereby obtaining a data fusion prediction model.

[0085] Further, according to an embodiment of the present invention, the multi-source meteorological fusion data and power data are fused three times, and then meteorological forecasts are performed three times to obtain historically available meteorological forecast data, including:

[0086] The power data in the photovoltaic power station area pre-processed in the relational database of the joint data management cloud platform of the Earth and the Star is integrated and processed and predicted with the multi-source fusion meteorological data input data fusion prediction model to obtain the initial available meteorological forecast data;

[0087] The initial available weather forecast data set is smoothed and abnormal data is cleaned to obtain historical available weather forecast data.

[0088] Further, according to an embodiment of the present invention, sky imager monitoring data and meteorological observation station data for the next 4 hours are acquired in real time, and the historical available meteorological forecast data are corrected by the acquired data to obtain real-time available meteorological forecast data, including:

[0089] Every hour, the sky imager monitoring data and meteorological observation station data of the current time and the past 24 hours are used to predict the sky imager monitoring data and meteorological observation station data of the next 4 hours using the convolutional long short-term memory network model to obtain the sky imager monitoring data and meteorological observation station data of the next 4 hours;

[0090] Then, using the weather observation station data and sky imager monitoring data for the next 4 hours, the BP neural network model is used to correct the historical available weather forecast data to obtain real-time available weather forecast data, including:

[0091] The BP neural network model was obtained by training the available weather forecast data and sky imager monitoring data of the past three months as the input data of the BP neural network, and the weather observation station data of the past three months as the output data of the BP neural network.

[0092] Every hour, the sky imager monitoring data for the next 4 hours and the historical available weather forecast data are used as the input data of the BP neural network model, and the original numerical weather forecast NWP data output by the model is obtained through the BP neural network;

[0093] The original numerical weather forecast NWP data and the weather observation station data for the next 4 hours are weighted to obtain real-time available weather forecast data, where the weighting coefficient is dynamically adjusted according to the deviation rate between the original numerical weather forecast NWP data and the real-time weather observation station data.

[0094] According to the above scheme of the present invention, the present invention integrates multiple sources of meteorological satellite cloud map data, sky imager data, meteorological observation station data, and real-time power data of photovoltaic power stations, and adopts a ground-satellite joint approach. High-precision numerical meteorological forecast data assists in the accurate prediction of photovoltaic power station power. The accurate power prediction data in turn reversely corrects the numerical meteorological forecast data, solving the current problem of a single input source and low accuracy of numerical meteorological forecast data, and can greatly improve the accuracy of photovoltaic power station power prediction.

[0095] The present invention is applicable to photovoltaic power stations, and can improve the prediction accuracy and update frequency of numerical meteorological forecast data, thereby improving the accuracy of photovoltaic power prediction, helping the power system dispatching department to coordinate the coordination of conventional energy and photovoltaic power generation, timely adjust the dispatching plan, and reasonably arrange the operation mode of the power grid. On the one hand, it effectively reduces the impact of photovoltaic access on the power grid and improves the safety and stability of power grid operation. On the other hand, it reduces the rotating reserve and operating costs of the power system, so as to make full use of solar energy resources and obtain greater economic and social benefits.

[0096] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a numerical weather forecast data updating system, comprising:

[0097] The data acquisition module acquires and stores multi-source meteorological data and power data in the photovoltaic power station area through the ground-satellite joint data management cloud platform;

[0098] Data preprocessing module, which preprocesses multi-source meteorological data and power data;

[0099] An initial meteorological forecast data prediction module performs downscaling processing on the meteorological satellite cloud image data in the multi-source meteorological observation data, and uses a model forecasting method to perform an initial meteorological forecast to obtain initial meteorological forecast data;

[0100] A multi-source meteorological fusion data prediction module performs secondary fusion processing on the initialized meteorological forecast data and the meteorological observation station data and sky imager monitoring data in the multi-source meteorological data, and then performs secondary meteorological prediction to obtain multi-source meteorological fusion data;

[0101] A historically available meteorological forecast data prediction module performs three fusion processes on the multi-source meteorological fusion data and power data, and then performs three meteorological forecasts to obtain historically available meteorological forecast data;

[0102] The real-time available weather forecast data update module obtains the sky imager monitoring data and weather observation station data for the next 4 hours in real time, and corrects the historical available weather forecast data through the acquired data to obtain the real-time available weather forecast data.

[0103] In this embodiment, the multi-source meteorological data includes meteorological satellite cloud image data, meteorological observation station data, and sky imager monitoring data.

[0104] Further, according to an embodiment of the present invention, the ground-satellite joint data management cloud platform acquires and stores multi-source meteorological data and power data in the photovoltaic power station area, including:

[0105] An operating system and a relational database are installed in the Earth-Satellite Joint Data Management Cloud Platform to establish a fixed channel for storing multi-source meteorological observation data and power data;

[0106] The ground-to-satellite joint data management cloud platform establishes data transmission channels with meteorological satellites, optical power prediction systems, meteorological observation stations, and sky imagers to regularly obtain multi-source meteorological data and power data from various data sources.

[0107] Further, according to an embodiment of the present invention, preprocessing of multi-source meteorological data and power data includes:

[0108] The ground-satellite joint data management cloud platform reads the meteorological satellite cloud image data from the multi-source meteorological data in the photovoltaic power station area through the meteorological satellite dedicated channel, performs integrity and rationality verification, bad data removal, missing and abnormal data smoothing and supplementation preprocessing on the meteorological satellite cloud image data, and stores the preprocessed meteorological satellite cloud image data in the relational database of the ground-satellite joint data management cloud platform;

[0109] The ground-satellite joint data management cloud platform establishes a communication channel with the meteorological observation station and the sky imager to obtain the meteorological observation station data and the sky imager monitoring data, and then performs integrity and rationality verification, bad data removal, missing and abnormal data smoothing and supplementation preprocessing on the meteorological observation station data and the sky imager monitoring data, and stores the preprocessed meteorological observation station data and the sky imager monitoring data in the relational database of the ground-satellite joint data management cloud platform;

[0110] The Earth-Satellite Joint Data Management Cloud Platform establishes a communication channel with the power prediction system to obtain the power data of the photovoltaic power station, and then performs integrity and rationality verification, bad data elimination, missing and abnormal data smoothing and supplementation preprocessing on the power data set, and stores the preprocessed power data set in the relational database of the Earth-Satellite Joint Data Management Cloud Platform.

[0111] Further, according to an embodiment of the present invention, downscaling is performed on meteorological satellite cloud image data in multi-source meteorological observation data, and a model forecasting method is used to perform an initial meteorological forecast to obtain initial meteorological forecast data, including:

[0112] The original numerical weather forecast NWP data were obtained from the US Global Weather Forecast System (GFS) and the European Centre for Medium-Range Weather Forecasts (ECMWF). The two sets of numerical weather forecast NWP data were first checked for rationality according to the local historical measured data, and the abnormal data were removed. Then the average of the two sets of numerical weather forecast NWP data was taken.

[0113] The average value data is downscaled using the Uformer model and smoothing interpolation algorithm in the Transformer model to obtain the initial meteorological forecast data with a forecast time scale of 15 minutes.

[0114] Further, according to an embodiment of the present invention, the initialized meteorological forecast data and the meteorological observation station data and the sky imager monitoring data in the multi-source meteorological data are subjected to secondary fusion processing, and then secondary meteorological prediction is performed to obtain multi-source meteorological fusion data, including:

[0115] Build data fusion prediction model;

[0116] Input the initialized meteorological forecast data, meteorological observation station data, and sky imager monitoring data into the data fusion prediction model, and output the original multi-source meteorological fusion data;

[0117] With reference to the data from meteorological observation stations, the original multi-source meteorological fusion data is corrected for correlation, the large-scale deviation is corrected linearly, and the multi-source meteorological fusion data is output.

[0118] In this embodiment, a data fusion prediction model is constructed as follows: Figure 2 As shown, the specific principles are as follows:

[0119] Sum: y′=(W1*X1+W2*X2+W3*X3+b);

[0120] Activation function: provides nonlinear fitting capability y=fy′;

[0121] in:

[0122] (X1, X2, X3) represents the external input signal, which refers to multiple attributes of a training data sample, that is, the total radiation, diffuse radiation, direct radiation and other data corresponding to the initial weather forecast data. b is the offset value, which enables the straight line to move up and down along the Y axis.

[0123] (W1, W2, W3) is the weight of each input signal. During the model training process, different weight values ​​are assigned according to the different synaptic strengths;

[0124] Input: simulates the input from other neurons in biological neural networks;

[0125] Weight: To simulate the different synaptic strengths received by each neuron in a biological neural network, the input received from the outside world is multiplied by a certain weight;

[0126] Sum: simulate the accumulation of signals received by neurons in biological neural networks;

[0127] Activation function: simulates the action potential generated by the accumulation of signals to a certain extent in biological neural networks. When the accumulation reaches a certain level, the action potential will be "activated". This method uses the Sigmoid function;

[0128] Output: simulates the new signals released by neurons in biological neural networks.

[0129] The training process is to continuously update the weight W and offset b until stable W and b are found to minimize the overall error of the model.

[0130] In this embodiment, the model training input is the total radiation, diffuse radiation, direct radiation, temperature, wind speed, wind direction, humidity, and cloud cover in the initial weather forecast data, and the cloud condition changes, cloud cover trends, and sunshine duration ratio in the sky imager monitoring data; the output is the total radiation, diffuse radiation, direct radiation, temperature, wind speed, wind direction, humidity, and cloud cover in the meteorological observation station data, thereby obtaining a data fusion prediction model.

[0131] Further, according to an embodiment of the present invention, the multi-source meteorological fusion data and power data are fused three times, and then meteorological forecasts are performed three times to obtain historically available meteorological forecast data, including:

[0132] The power data in the photovoltaic power station area pre-processed in the relational database of the joint data management cloud platform of the Earth and the Star is integrated and processed and predicted with the multi-source fusion meteorological data input data fusion prediction model to obtain the initial available meteorological forecast data;

[0133] The initial available weather forecast data set is smoothed and abnormal data is cleaned to obtain historical available weather forecast data.

[0134] Further, according to an embodiment of the present invention, sky imager monitoring data and meteorological observation station data for the next 4 hours are acquired in real time, and the historical available meteorological forecast data are corrected by the acquired data to obtain real-time available meteorological forecast data, including:

[0135] Every hour, the sky imager monitoring data and meteorological observation station data of the current time and the past 24 hours are used to predict the sky imager monitoring data and meteorological observation station data of the next 4 hours using the convolutional long short-term memory network model to obtain the sky imager monitoring data and meteorological observation station data of the next 4 hours;

[0136] Then, using the weather observation station data and sky imager monitoring data for the next 4 hours, the BP neural network model is used to correct the historical available weather forecast data to obtain real-time available weather forecast data, including:

[0137] The available weather forecast data and sky imager monitoring data of the past three months are used as the input data of the BP neural network, and the weather observation station data of the past three months are used as the output data of the BP neural network to train the BP neural network model.

[0138] Every hour, the sky imager monitoring data for the next 4 hours and the historical available weather forecast data are used as the input data of the BP neural network model, and the original numerical weather forecast NWP data output by the model is obtained through the BP neural network;

[0139] The original numerical weather forecast NWP data and the weather observation station data for the next 4 hours are weighted to obtain real-time available weather forecast data, where the weighting coefficient is dynamically adjusted according to the deviation rate between the original numerical weather forecast NWP data and the real-time weather observation station data.

[0140] According to the above scheme of the present invention, the present invention integrates multiple sources of meteorological satellite cloud map data, sky imager data, meteorological observation station data, and real-time power data of photovoltaic power stations, and adopts a ground-satellite joint approach. High-precision numerical meteorological forecast data assists in the accurate prediction of photovoltaic power station power. The accurate power prediction data in turn reversely corrects the numerical meteorological forecast data, solving the current problem of a single input source and low accuracy of numerical meteorological forecast data, and can greatly improve the accuracy of photovoltaic power station power prediction.

[0141] The present invention is applicable to photovoltaic power stations, and can improve the prediction accuracy and update frequency of numerical meteorological forecast data, thereby improving the accuracy of photovoltaic power prediction, helping the power system dispatching department to coordinate the coordination of conventional energy and photovoltaic power generation, timely adjust the dispatching plan, and reasonably arrange the operation mode of the power grid. On the one hand, it effectively reduces the impact of photovoltaic access on the power grid and improves the safety and stability of power grid operation. On the other hand, it reduces the rotating reserve and operating costs of the power system, so as to make full use of solar energy resources and obtain greater economic and social benefits.

[0142] Furthermore, to achieve the above-mentioned purpose, the present invention also provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the numerical meteorological forecast data updating method as described above is implemented.

[0143] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the numerical meteorological forecast data updating method as described above is implemented.

[0144] Those skilled in the art will appreciate that the modules and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0145] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and equipment can refer to the corresponding processes in the aforementioned method implementation methods, and will not be repeated here.

[0146] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0147] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiments of the present invention.

[0148] In addition, each functional module in the embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0149] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the energy-saving signal sending / receiving method of each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical disks.

[0150] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other.

[0151] It should be understood that the size of the serial numbers of each step in the content of the invention and the implementation methods of the present invention does not absolutely mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the implementation methods of the present invention.

Claims

1. A method for updating numerical weather forecast data, characterized in that: include: Acquire and store multi-source meteorological data and power data in the photovoltaic power station area through the ground-satellite joint data management cloud platform; Preprocess multi-source meteorological data and power data; Downscaling the meteorological satellite cloud image data in the multi-source meteorological observation data, and using a model forecasting method to perform an initial meteorological forecast to obtain initial meteorological forecast data; Performing secondary fusion processing on the initialized meteorological forecast data and the meteorological observation station data and sky imager monitoring data in the multi-source meteorological data, and then performing secondary meteorological forecasting to obtain multi-source meteorological fusion data; The multi-source meteorological fusion data and power data are subjected to three fusion processes, and then three meteorological forecasts are performed to obtain historically available meteorological forecast data; The sky imager monitoring data and the meteorological observation station data for the next 4 hours are acquired in real time, and the historical available meteorological forecast data are corrected by the acquired data to obtain the real-time available meteorological forecast data.

2. The method for updating numerical weather forecast data according to claim 1, characterized in that: The ground-satellite joint data management cloud platform acquires and stores multi-source meteorological data and power data in the photovoltaic power station area, including: An operating system and a relational database are installed in the Earth-Satellite Joint Data Management Cloud Platform to establish a fixed channel for storing multi-source meteorological observation data and power data; The ground-to-satellite joint data management cloud platform establishes data transmission channels with meteorological satellites, optical power prediction systems, meteorological observation stations, and sky imagers to regularly obtain multi-source meteorological data and power data from various data sources.

3. The method for updating numerical weather forecast data according to claim 2, characterized in that: The preprocessing of multi-source meteorological data and power data includes: Read the meteorological satellite cloud image data from the multi-source meteorological data in the photovoltaic power station area through the meteorological satellite dedicated channel, perform integrity and rationality verification, bad data removal, missing and abnormal data smoothing and supplementation preprocessing on the meteorological satellite cloud image data, and store the preprocessed meteorological satellite cloud image data in the relational database of the ground-satellite joint data management cloud platform; The meteorological observation station data and sky imager monitoring data are preprocessed to check integrity and rationality, remove bad data, and smooth and supplement missing and abnormal data, and the preprocessed meteorological observation station data and sky imager monitoring data are stored in the relational database of the ground-satellite joint data management cloud platform; The power data set is preprocessed by checking its integrity and rationality, removing bad data, and smoothing and supplementing missing and abnormal data. The preprocessed power data set is stored in the relational database of the Earth-Satellite Joint Data Management Cloud Platform.

4. The method for updating numerical weather forecast data according to claim 1, characterized in that: The meteorological satellite cloud image data in the multi-source meteorological observation data is downscaled, and the model forecast method is used to perform the initial meteorological forecast to obtain the initial meteorological forecast data, including: The original numerical weather forecast NWP data were obtained from the US Global Weather Forecast System (GFS) and the European Centre for Medium-Range Weather Forecasts (ECMWF). The two sets of numerical weather forecast NWP data were first checked for rationality according to the local historical measured data, and the abnormal data were removed. Then the average of the two sets of numerical weather forecast NWP data was taken. The average value data is downscaled using the Uformer model and smoothing interpolation algorithm in the Transformer model to obtain the initial meteorological forecast data with a forecast time scale of 15 minutes.

5. The method for updating numerical weather forecast data according to claim 1, characterized in that: The initialized meteorological forecast data and the meteorological observation station data and the sky imager monitoring data in the multi-source meteorological data are subjected to secondary fusion processing, and then secondary meteorological forecasting is performed to obtain multi-source meteorological fusion data, including: Build data fusion prediction model; Input the initialized meteorological forecast data, meteorological observation station data, and sky imager monitoring data into the data fusion prediction model, and output the original multi-source meteorological fusion data; With reference to the data from meteorological observation stations, the original multi-source meteorological fusion data is corrected for correlation, the large-scale deviation is corrected linearly, and the multi-source meteorological fusion data is output.

6. The method for updating numerical weather forecast data according to claim 2, characterized in that: The multi-source meteorological fusion data and power data are fused three times, and then meteorological forecasts are performed three times to obtain historically available meteorological forecast data, including: The power data in the photovoltaic power station area pre-processed in the relational database of the joint data management cloud platform of the Earth and the Star is integrated and processed and predicted with the multi-source fusion meteorological data input data fusion prediction model to obtain the initial available meteorological forecast data; The initial available weather forecast data set is smoothed and abnormal data is cleaned to obtain historical available weather forecast data.

7. The method for updating numerical weather forecast data according to any one of claims 1 to 6, characterized in that: Acquire the sky imager monitoring data and meteorological observation station data for the next 4 hours in real time, and correct the historical available meteorological forecast data by using the acquired data to obtain real-time available meteorological forecast data, including: Every hour, the sky imager monitoring data and meteorological observation station data of the current time and the past 24 hours are used to predict the sky imager monitoring data and meteorological observation station data of the next 4 hours using the convolutional long short-term memory network model to obtain the sky imager monitoring data and meteorological observation station data of the next 4 hours; Then, using the weather observation station data and sky imager monitoring data for the next 4 hours, the BP neural network model is used to correct the historical available weather forecast data to obtain real-time available weather forecast data, including: The BP neural network model was obtained by training the available weather forecast data and sky imager monitoring data of the past three months as the input data of the BP neural network, and the weather observation station data of the past three months as the output data of the BP neural network. Every hour, the sky imager monitoring data for the next 4 hours and the historical available weather forecast data are used as the input data of the BP neural network model, and the original numerical weather forecast NWP data output by the model is obtained through the BP neural network; The original numerical weather forecast NWP data and the weather observation station data for the next 4 hours are weighted to obtain real-time available weather forecast data, where the weighting coefficient is dynamically adjusted according to the deviation rate between the original numerical weather forecast NWP data and the real-time weather observation station data.

8. Numerical weather forecast data updating system, characterized in that: include: The data acquisition module acquires and stores multi-source meteorological data and power data in the photovoltaic power station area through the ground-satellite joint data management cloud platform; Data preprocessing module, which preprocesses multi-source meteorological data and power data; An initial meteorological forecast data prediction module performs downscaling processing on the meteorological satellite cloud image data in the multi-source meteorological observation data, and uses a model forecasting method to perform an initial meteorological forecast to obtain initial meteorological forecast data; A multi-source meteorological fusion data prediction module performs secondary fusion processing on the initialized meteorological forecast data and the meteorological observation station data and sky imager monitoring data in the multi-source meteorological data, and then performs secondary meteorological prediction to obtain multi-source meteorological fusion data; A historically available meteorological forecast data prediction module performs three fusion processes on the multi-source meteorological fusion data and power data, and then performs three meteorological forecasts to obtain historically available meteorological forecast data; The real-time available weather forecast data update module obtains the sky imager monitoring data and weather observation station data for the next 4 hours in real time, and corrects the historical available weather forecast data through the acquired data to obtain the real-time available weather forecast data.

9. An electronic device, characterized in that The method comprises a processor, a memory and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the method for updating numerical meteorological forecast data according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the numerical meteorological forecast data updating method according to any one of claims 1 to 7 is implemented.

Images