Dust retention power prediction method and device for photovoltaic power generation and electronic equipment

Abstract

The invention discloses a dust retention power prediction method and device for photovoltaic power generation and electronic equipment. The method comprises the following steps: acquiring a historical photovoltaic data set, and dividing the historical photovoltaic data set into a historical dust-free data set and a historical dust accumulation data set; thirdly, constructing a dust-free power prediction model, and generating a historical loss data set according to the historical dust accumulation data set and the dust-free power prediction model; and finally, constructing a dust retention power loss prediction model, and determining target dust retention prediction power corresponding to the to-be-detected meteorological data based on the dust-free power prediction model and the dust retention power loss prediction model. By introducing a dust accumulation loss compensation mechanism, dust-free power prediction in a dust-free state and loss prediction caused by dust accumulation are innovatively combined, the problem of power generation efficiency attenuation caused by dust accumulation on the surface of a photovoltaic panel in actual operation is effectively solved, prediction errors are greatly reduced under the condition of dust accumulation, and the power generation efficiency is improved. And reliable support is provided for fine management and power grid dispatching of the photovoltaic power station.

Description

Technical Field

[0001] This application relates to the field of photovoltaic power generation technology, and in particular to a method for predicting the dust accumulation power of photovoltaic power generation. Background Technology

[0002] Photovoltaic power generation forecasting is a core means of balancing grid security and economy, and its importance is increasingly prominent in the context of building a new power system dominated by new energy sources. Since desert and Gobi regions have become the main battleground for new energy development, this need is particularly urgent in large-scale photovoltaic (PV) base scenarios in these areas. Existing PV power forecasting mainly relies on physical models and traditional statistical methods. Its forecasting accuracy is severely limited by the accuracy of weather forecasts and the quality and quantity of data. Furthermore, in desert and Gobi regions, dust accumulation on the surface of PV panels causes a decrease in power generation efficiency, resulting in large errors in PV power generation forecasting. This fails to meet the accuracy requirements for dust-accumulated power forecasting and cannot ensure the safe, stable, and economical operation of a high-proportion new energy power system. Summary of the Invention

[0003] This application provides a method, apparatus, and electronic device for predicting the dust accumulation power of photovoltaic power generation, the technical solution of which is as follows:

[0004] In a first aspect, embodiments of this application provide a method for predicting the dust accumulation power of photovoltaic power generation, the method comprising:

[0005] A historical photovoltaic dataset is obtained and divided into a historical dust-free dataset and a historical dust-accumulated dataset. The historical photovoltaic dataset includes historical meteorological data and historical operational data corresponding to each historical meteorological data. The historical meteorological data includes temperature, ambient humidity, wind speed, irradiance, atmospheric pressure, and rainfall. The historical operational data includes historical actual photovoltaic power generation and dust pollution ratio.

[0006] A cleanroom power prediction model is constructed based on the CEEMDAN-CNN-LSTM network and the historical cleanroom dataset, and a historical loss dataset is generated based on the historical dust accumulation dataset and the cleanroom power prediction model.

[0007] A dust accumulation power loss prediction model is constructed based on the historical loss dataset, and the target dust accumulation prediction power corresponding to the meteorological data to be measured is determined based on the dust-free power prediction model and the dust accumulation power loss prediction model.

[0008] Secondly, a dust accumulation power prediction device for photovoltaic power generation is provided, the device comprising:

[0009] The acquisition module is used to acquire historical photovoltaic datasets and divide the historical photovoltaic datasets into historical dust-free datasets and historical dust-accumulated datasets. The historical photovoltaic datasets include historical meteorological data and historical operating data corresponding to each historical meteorological data. The historical meteorological data include temperature, ambient humidity, wind speed, irradiance, atmospheric pressure, and rainfall. The historical operating data includes historical actual photovoltaic power generation and dust pollution ratio.

[0010] The module is used to build a dust-free power prediction model based on the CEEMDAN-CNN-LSTM network and the historical dust-free dataset, and to generate a historical loss dataset based on the historical dust accumulation dataset and the dust-free power prediction model.

[0011] The determination module is used to construct a dust accumulation power loss prediction model based on the historical loss dataset, and to determine the target dust accumulation prediction power corresponding to the meteorological data to be measured based on the dust-free power prediction model and the dust accumulation power loss prediction model.

[0012] Thirdly, an electronic device is provided, including a device processor and a memory;

[0013] The device processor is connected to the memory;

[0014] The memory is used to store executable program code;

[0015] The device processor runs a program corresponding to the executable program code stored in the memory to perform the steps of the method provided as in the first aspect or any possible implementation thereof.

[0016] Fourthly, a computer-readable storage medium is provided having a computer program stored thereon, the computer-readable storage medium storing instructions that, when executed on a computer or device processor, cause the computer or device processor to perform the method provided as in the first aspect or any possible implementation thereof.

[0017] The beneficial effects of the technical solutions provided in some embodiments of this application include at least the following:

[0018] In one or more embodiments of this application, a historical photovoltaic dataset is acquired and divided into a historical cleanliness dataset and a historical dust accumulation dataset. Next, a cleanliness power prediction model is constructed based on a hybrid model of Complete Ensemble Empirical Mode Decomposition with Adaptive Noise-Convolutional Neural Network - Long Short-Term Memory Network (CEEMDAN-CNN-LSTM) and the historical cleanliness dataset. A historical loss dataset is then generated based on the historical dust accumulation dataset and the cleanliness power prediction model. Finally, a dust accumulation power loss prediction model is constructed based on the historical loss dataset, and the target dust accumulation prediction power corresponding to the meteorological data to be measured is determined based on the cleanliness power prediction model and the dust accumulation power loss prediction model. By constructing a dust accumulation power loss prediction model and introducing a dust accumulation loss compensation mechanism, the model innovatively combines dust-free power prediction under dust-free conditions with loss prediction caused by dust accumulation. This enhances the adaptability of the prediction model under dust accumulation conditions, effectively solves the problem of power generation efficiency degradation caused by dust accumulation on the surface of photovoltaic panels in actual operation, and significantly reduces prediction errors under dust accumulation conditions, providing reliable support for the refined management of photovoltaic power plants and grid dispatch. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A flowchart illustrating a method for predicting the dust accumulation power of photovoltaic power generation, provided in an embodiment of this application;

[0021] Figure 2 A schematic diagram of the structure of a photovoltaic power generation dust accumulation power prediction device provided in an embodiment of this application;

[0022] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0023] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0024] The terms "first," "second," "third," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0025] The following description provides examples and does not limit the scope, applicability, or examples set forth in the claims. Changes may be made to the function and arrangement of the described elements without departing from the scope of this application. Various processes or components may be appropriately omitted, substituted, or added to the examples. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described with respect to some examples may be combined into other examples.

[0026] Please see Figure 1 , Figure 1 The diagram shows an overall flowchart of a method for predicting the dust accumulation power of photovoltaic power generation according to an embodiment of this application.

[0027] like Figure 1 As shown, the method for predicting the dust accumulation power of photovoltaic power generation may include at least the following steps:

[0028] Step 101: Obtain the historical photovoltaic dataset and divide the historical photovoltaic dataset into the historical dust-free dataset and the historical dust-accumulated dataset.

[0029] The historical photovoltaic dataset includes historical meteorological data and corresponding historical operational data. The historical meteorological data includes temperature, ambient humidity, wind speed, irradiance, atmospheric pressure, and rainfall. The historical operational data includes historical actual photovoltaic power generation and dust accumulation ratio.

[0030] In this embodiment, to construct a predictive model that can directly obtain dust accumulation power from any meteorological data to be measured, it is necessary to first acquire a large amount of historical photovoltaic datasets corresponding to various meteorological environments. This historical photovoltaic dataset may include, but is not limited to, a large amount of historical meteorological data and historical operational data corresponding to each historical meteorological data. The historical meteorological data may include, but is not limited to, temperature, ambient humidity, wind speed, irradiance, atmospheric pressure, and rainfall for each photovoltaic power generation scenario. The historical operational data includes the historical actual photovoltaic power generation power and dust accumulation pollution ratio for each photovoltaic power generation scenario. Specifically, since the historical photovoltaic dataset includes both dust-free and dusty power generation environments, and dusty environments can cause dust accumulation on the surface of photovoltaic panels, leading to a decrease in power generation efficiency and drastic fluctuations in photovoltaic power generation, inaccurate prediction of dust accumulation power will pose a serious challenge to the safety, supply, and consumption of the power system. Therefore, to construct an accurate dust accumulation power prediction model, it is necessary to first divide the historical photovoltaic dataset into historical dust-free datasets and historical dust-accumulated datasets according to the specific data acquisition environment, providing basic data for power prediction models under different environments.

[0031] In one possible implementation, dividing the historical photovoltaic dataset into a historical dust-free dataset and a historical dust-accumulated dataset includes:

[0032] Anomalies are removed from the historical photovoltaic dataset based on the rainfall data to obtain a qualified historical dataset.

[0033] Based on the dust contamination ratio, the historical qualified dataset is divided into a historical dust-free dataset and a historical dust-accumulated dataset.

[0034] In this embodiment, since the photovoltaic power generation data corresponding to rainfall is not of reference value during actual photovoltaic power generation and can be determined as abnormal and accidental data, in order to improve the accuracy of subsequent training models, anomalies can be removed from the historical photovoltaic dataset based on rainfall. That is, if the rainfall on a certain day is determined to be greater than zero, the photovoltaic data corresponding to that day needs to be removed. This process is repeated to complete all anomaly removal, resulting in a qualified historical dataset. Next, since each historical operating data in the historical photovoltaic dataset may include a dust accumulation ratio, which is used to characterize the ratio of actual photovoltaic power generation to theoretical photovoltaic power generation, the qualified historical dataset can be further divided based on the dust accumulation ratio. All qualified data with a dust accumulation ratio of 100% are determined as the historical dust-free dataset, and the remaining data are determined as the historical dust-accumulated dataset.

[0035] In one possible implementation, after dividing the historical photovoltaic dataset into a historical dust-free dataset and a historical dust-accumulated dataset, the method further includes:

[0036] For any historical meteorological data in the historical photovoltaic dataset, the historical meteorological data is normalized based on the Min-Max method to obtain standard meteorological data.

[0037] In this embodiment, since any historical meteorological data in the historical photovoltaic dataset includes data in different formats and dimensions such as temperature, ambient humidity, wind speed, irradiance, atmospheric pressure, and rainfall, in order to facilitate unified data access for subsequent models and improve subsequent data processing efficiency, after dividing the historical photovoltaic dataset into a historical clean dataset and a historical dusty dataset, the Min-Max method can be used to normalize any historical meteorological data in the historical photovoltaic dataset, mapping all of them to... The interval is used to obtain standard meteorological data. Specifically, the Min-Max normalization formula is as follows:

[0038]

[0039] in, For the i-th meteorological factor on the j-th meteorological day, and These are the maximum and minimum values ​​for the i-th meteorological event, respectively.

[0040] Step 103: Construct a cleanroom power prediction model based on the CEEMDAN-CNN-LSTM network and the historical cleanroom dataset, and generate a historical loss dataset based on the historical dust accumulation dataset and the cleanroom power prediction model.

[0041] In this embodiment, to predict the dust accumulation power for any given weather day, it is necessary to first determine the dust-free prediction power under ideal dust-free conditions. Subsequently, only the influence of dust accumulation power needs to be compensated for to obtain the accurate dust accumulation power. When constructing the dust-free power prediction model, since the CEEMDAN method can extract multi-frequency features of the irradiance sequence, and the CNN-LSTM network can uncover the nonlinear relationship between meteorological factors and power generation, a dust-free power prediction model can be constructed based on the CEEMDAN-CNN-LSTM network and historical dust-free datasets. Specifically, the CEEMDAN method decomposes the multi-frequency solar irradiance into simple frequency characteristic signals, and normalizes meteorological data such as temperature, ambient humidity, wind speed, and atmospheric pressure. Then, a CNN network is used for deep feature extraction, and the parameters in the LSTM network are optimized using a particle swarm optimization algorithm. These parameters specifically include the number of neurons. Learning rate and abandonment probability The fitness function is set to root mean square error. The model is iterated until the maximum number of iterations is reached to obtain the optimal parameter population and construct a cleanroom power prediction model.

[0042] Next, to determine the impact of dust accumulation on power, it is necessary to determine the power loss under dust accumulation conditions relative to an ideal dust-free state. Therefore, a dust accumulation power loss model can be constructed to predict the dust accumulation power loss under various meteorological conditions. However, a large amount of standard loss data is needed for model training when constructing the model. Therefore, the dust accumulation power loss corresponding to different meteorological conditions can be determined based on the historical dust accumulation dataset and the dust-free power prediction model, thereby generating a historical loss dataset.

[0043] In one possible implementation, the step of constructing a cleanroom power prediction model based on the CEEMDAN-CNN-LSTM network and the historical cleanroom dataset includes:

[0044] A CEEMDAN-CNN-LSTM network is constructed based on the CEEMDAN method and the CNN-LSTM network.

[0045] The CEEMDAN-CNN-LSTM network is trained based on the historical cleanliness dataset to obtain a cleanliness power prediction model. The historical cleanliness dataset includes first historical meteorological data as input to the cleanliness power prediction model and first historical power as output of the cleanliness power prediction model.

[0046] In this embodiment, the cleanroom power prediction model can be constructed by first building a CEEMDAN-CNN-LSTM network based on the CEEMDAN method and a CNN-LSTM network. Specifically, the CEEMDAN method includes the following steps: first, adding K times Gaussian white noise with a mean of 0 to the signal to be decomposed to construct K sequences to be decomposed, as shown in the following formula:

[0047]

[0048] in, For the nth sequence to be decomposed, The signal to be decomposed These are noise weighting coefficients. The noise is Gaussian white noise, and K is the total number of sequences to be decomposed.

[0049] Next, EMD decomposition is performed on the above K sequences to be decomposed sequentially to obtain the first mode component (IMF), and the mean is calculated as the first IMF of CEEMDAN. The specific formula is as follows:

[0050]

[0051] in, For Ceemdan's first IMF, Let be the first IMF of the i-th sequence, and K be the total number of sequences to be decomposed.

[0052] Furthermore, after adding specific noise to the residual signal obtained after decomposition, EMD decomposition is performed again, as shown in the following formula:

[0053]

[0054] in, This is a margin signal. These are noise weighting coefficients. Let be the Gaussian white noise of the i-th sequence to be decomposed. If the stopping condition is met, the iteration stops; otherwise, return to the previous step and continue iterating.

[0055] The convolutional neural network in the CEEMDAN-CNN-LSTM network consists of three one-dimensional convolutional layers, one max-pooling layer, and one flattening layer. The one-dimensional convolutional layer reads the input sequence and projects the result onto the feature map, amplifying some potentially significant features. In this embodiment, the Rectified Linear Activation Function (ReLU) is used, and the number of filters is set to 16, 32, and 64, respectively. The max-pooling layer reduces the feature dimension by saving the maximum signal value and retains important feature information. The flattening layer converts the multi-dimensional data output by the max-pooling layer into one-dimensional data, making it conform to the input data of the LSTM model. The Long Short-Term Memory (LSTM) network in the CEEMDAN-CNN-LSTM model consists of two LSTM hidden layers, a dropout layer, and a fully connected layer. LSTM introduces a forget gate, an input gate, and an output gate. The forget gate removes unnecessary information from the previous time step of the previous hidden layer. The input gate determines the information to be added to the current input. The output gate determines the information to be passed to the next time step and the information to be output. The dropout layer prevents the model from overfitting. The fully connected layer outputs the final prediction result.

[0056] The particle swarm optimization algorithm is used to optimize the parameters in the LSTM network. The population particles are initialized, and the position of the target particle i is set to... The feature sequences extracted from the convolutional neural network depth are input into the LSTM network for training. RMSE is used as the fitness function, calculated according to the following formula:

[0057]

[0058] Where N is the size of the training set. For the predicted photovoltaic power value of the k-th training data, This represents the actual measured value of photovoltaic power for the kth training data point.

[0059] Iterate through the particles using the following formula, update their position and velocity, find the optimal positions of the particles and the particle swarm, and then update the positions of all optimal particles and the particle swarm:

[0060]

[0061]

[0062]

[0063] in, Let be the position of particle i at time t. Let be the velocity of particle i at time t. Let be the position of particle i at time t+1. X is the velocity of particle i at time t+1. Let be the historical best position of particle i at time t. Let `rand()` be the historical best position of particle swarm i at time `t`, and its output be located at... interval, As the inertia weighting factor, This is the initial inertia weighting factor. This is the final inertia weighting factor. and Here, t is the learning factor, t is the number of iterations, and G is the maximum number of iterations. In this embodiment... and Take 2 for all. and We set the values ​​to 0.9 and 0.5 respectively, and G to 200. Then we update the parameters according to the fitness function, and determine whether the maximum number of iterations has been reached. If it has been reached, we update the parameters and output the optimal parameters. If it has not been reached, we continue to update the particles.

[0064] Finally, the first historical meteorological data in the historical cleanliness dataset is used as the model input, and the actual historical photovoltaic power generation in the historical cleanliness dataset, i.e. the first historical power, is used as the model output. The constructed CEEMDAN-CNN-LSTM network is then trained to obtain the cleanliness power prediction model.

[0065] In one possible implementation, generating the historical loss dataset based on the historical dust accumulation dataset and the cleanroom power prediction model includes:

[0066] The data in the historical dust accumulation dataset are determined as each second historical meteorological data and the second historical power corresponding to each second historical meteorological data;

[0067] Based on the cleanroom power prediction model, determine the historical cleanroom prediction power corresponding to each of the second historical meteorological data.

[0068] A historical loss dataset is constructed based on each second historical meteorological data and the difference between each historical dust-free predicted power and each second historical power corresponding to each second historical meteorological data.

[0069] In this embodiment, after obtaining the trained cleanroom power prediction model, in order to generate a historical loss dataset to provide a data foundation for the subsequent dust accumulation power loss prediction model, each data point in the historical dust accumulation dataset can first be identified as a second historical meteorological data point and the corresponding historical photovoltaic power generation actual power, i.e., the second historical power. Next, the second historical meteorological data is used as the model input for the cleanroom power prediction model to obtain the historical cleanroom predicted power output by the model under ideal cleanroom conditions. Further, the difference between this historical cleanroom predicted power and the historical photovoltaic power generation actual power in its original corresponding historical dust accumulation dataset is calculated to obtain the power difference, i.e., the power loss caused by dust accumulation. Each second historical meteorological data point, and the corresponding historical cleanroom predicted power and the power difference between each second historical power are then paired one-to-one to generate a historical loss dataset. For each sample in the historical dust accumulation dataset, the formula for calculating its power loss L is as follows:

[0070]

[0071] in, This represents the actual historical photovoltaic power generation. This is the historical cleanroom power prediction output by the cleanroom power prediction model.

[0072] Step 105: Construct a dust accumulation power loss prediction model based on the historical loss dataset, and determine the target dust accumulation prediction power corresponding to the meteorological data to be measured based on the dust-free power prediction model and the dust accumulation power loss prediction model.

[0073] In this embodiment, after obtaining the historical loss dataset, it can be used as training data to construct a dust accumulation power loss prediction model under dusty meteorological conditions. That is, inputting any dusty meteorological data into this model will yield the corresponding dust accumulation power loss. Next, the constructed dust-free power prediction model and the dust accumulation power loss prediction model are fused together to obtain a fused prediction model. Directly inputting any meteorological data under any meteorological condition into this model will yield the target dust accumulation predicted power that takes into account dust accumulation power loss compensation. This improves the accuracy of photovoltaic power generation prediction in desert and Gobi areas and ensures the safe, stable, and economical operation of a high-proportion new energy power system.

[0074] In one possible implementation, constructing a dust accumulation power loss prediction model based on the historical loss dataset includes:

[0075] An initial dust accumulation power loss prediction model was constructed based on a BO-LSTM network.

[0076] The initial dust accumulation power loss prediction model is trained based on the historical loss dataset to obtain the dust accumulation power loss prediction model.

[0077] In this embodiment, when constructing a dust accumulation power loss prediction model based on a historical loss dataset, an initial dust accumulation power loss prediction model can be constructed first based on a BO-LSTM network. Specifically, the BO-LSTM network consists of an input layer, an LSTM hidden layer, a Dropout layer, and a fully connected layer. The input layer receives the feature sequence of the loss dataset, and the time step is set to T, which is determined according to the dust accumulation characteristics, for example, T=5. The LSTM hidden layer adopts a multi-layer LSTM structure, and its gating mechanism models the time dependency of the loss. The forget gate is used in the gating mechanism, and its update formula is:

[0078]

[0079] in, It is the sigmoid activation function. and These are weights and biases, respectively. This is the hidden state from the previous moment. This is the current input.

[0080] The Dropout layer is used to prevent overfitting, the dropout rate is determined by Bayesian optimization, and the fully connected layer is used to output the predicted dust accumulation loss value.

[0081] Next, the historical meteorological data in the historical loss dataset is used as the model input, and the corresponding historical dust accumulation power loss is used as the model output. The initial dust accumulation power loss prediction model is trained. The forward propagation is used to calculate the prediction loss, and the backpropagation is used to update the weights through the adaptive moment estimation (Adam) optimizer. The training stops when the validation set error does not decrease for 5 consecutive rounds. After training, the dust accumulation power loss prediction model is obtained.

[0082] In one possible implementation, determining the target dust accumulation prediction power corresponding to the meteorological data to be measured based on the dust-free power prediction model and the dust accumulation power loss prediction model includes:

[0083] The target dust-free predicted power corresponding to the meteorological data to be measured is determined based on the dust-free power prediction model.

[0084] The target dust loss power corresponding to the meteorological data to be measured is determined based on the dust power loss prediction model.

[0085] The target dust accumulation prediction power is determined based on the target dust-free prediction power and the target dust accumulation loss power.

[0086] In this embodiment, when determining the target dust accumulation prediction power corresponding to any meteorological data to be measured, it can first be input into a trained dust-free power prediction model to obtain the target dust-free prediction power corresponding to the meteorological data to be measured under theoretical dust-free conditions. Then, it can be input into a trained dust accumulation power loss prediction model to obtain the target dust accumulation loss power corresponding to the meteorological data to be measured. Furthermore, by calculating the difference between the target dust-free prediction power and the target dust accumulation loss power, the target dust accumulation prediction power considering dust accumulation loss compensation can be accurately obtained.

[0087] The foregoing has described specific embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0088] Please refer to the following. Figure 2 , Figure 2 A schematic diagram of a photovoltaic power generation dust accumulation power prediction device provided in an embodiment of this application is shown. It should be noted that... Figure 2 The photovoltaic power generation dust accumulation power prediction device shown is used to perform the functions described in this application. Figure 1 The methods shown in the embodiments are for illustrative purposes only, illustrating the parts relevant to the embodiments of this application. For specific technical details not disclosed, please refer to this application. Figure 1 The example shown.

[0089] like Figure 2 As shown, the photovoltaic power generation dust accumulation power prediction device may include at least:

[0090] The acquisition module 201 is used to acquire historical photovoltaic datasets and divide the historical photovoltaic datasets into historical dust-free datasets and historical dust-accumulated datasets. The historical photovoltaic datasets include historical meteorological data and historical operating data corresponding to each historical meteorological data. The historical meteorological data include temperature, ambient humidity, wind speed, irradiance, atmospheric pressure and rainfall. The historical operating data includes historical photovoltaic power generation actual power and dust pollution ratio.

[0091] Module 202 is used to build a dust-free power prediction model based on the CEEMDAN-CNN-LSTM network and the historical dust-free dataset, and to generate a historical loss dataset based on the historical dust accumulation dataset and the dust-free power prediction model.

[0092] The determination module 203 is used to construct a dust accumulation power loss prediction model based on the historical loss dataset, and to determine the target dust accumulation prediction power corresponding to the meteorological data to be measured based on the dust-free power prediction model and the dust accumulation power loss prediction model.

[0093] In one possible implementation, the acquisition module 201 is specifically used for:

[0094] Anomalies are removed from the historical photovoltaic dataset based on the rainfall data to obtain a qualified historical dataset.

[0095] Based on the dust contamination ratio, the historical qualified dataset is divided into a historical dust-free dataset and a historical dust-accumulated dataset.

[0096] In one possible implementation, the acquisition module 201 is further configured to:

[0097] For any historical meteorological data in the historical photovoltaic dataset, the historical meteorological data is normalized based on the Min-Max method to obtain standard meteorological data.

[0098] In one possible implementation, the construction module 202 is specifically used for:

[0099] A CEEMDAN-CNN-LSTM network is constructed based on the CEEMDAN method and the CNN-LSTM network.

[0100] The CEEMDAN-CNN-LSTM network is trained based on the historical cleanliness dataset to obtain a cleanliness power prediction model. The historical cleanliness dataset includes first historical meteorological data as input to the cleanliness power prediction model and first historical power as output of the cleanliness power prediction model.

[0101] In one possible implementation, the construction module 202 is further configured to:

[0102] The data in the historical dust accumulation dataset are determined as each second historical meteorological data and the second historical power corresponding to each second historical meteorological data;

[0103] Based on the cleanroom power prediction model, determine the historical cleanroom prediction power corresponding to each of the second historical meteorological data.

[0104] A historical loss dataset is constructed based on each second historical meteorological data and the difference between each historical dust-free predicted power and each second historical power corresponding to each second historical meteorological data.

[0105] In one possible implementation, the determining module 203 is specifically used for:

[0106] An initial dust accumulation power loss prediction model was constructed based on a BO-LSTM network.

[0107] The initial dust accumulation power loss prediction model is trained based on the historical loss dataset to obtain the dust accumulation power loss prediction model.

[0108] In one possible implementation, the determining module 203 is further configured to:

[0109] The target dust-free predicted power corresponding to the meteorological data to be measured is determined based on the dust-free power prediction model.

[0110] The target dust loss power corresponding to the meteorological data to be measured is determined based on the dust power loss prediction model.

[0111] The target dust accumulation prediction power is determined based on the target dust-free prediction power and the target dust accumulation loss power.

[0112] Those skilled in the art will clearly understand that the technical solutions of the embodiments of this application can be implemented by means of software and / or hardware. In this application, "unit" and "module" refer to software and / or hardware that can independently complete or cooperate with other components to complete a specific function, wherein the hardware may be, for example, a field-programmable gate array (FPGA), an integrated circuit (IC), etc.

[0113] Each processing unit and / or module in the embodiments of this application can be implemented by an analog circuit that implements the functions described in the embodiments of this application, or by software that executes the functions described in the embodiments of this application.

[0114] Please refer to the following. Figure 3 , Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown.

[0115] like Figure 3 As shown, the electronic device 300 may include: at least one device processor 301, at least one network interface 304, user interface 303, memory 305, and at least one communication bus 302.

[0116] The communication bus 302 can be used to realize the connection and communication of the above components.

[0117] The user interface 303 may include buttons, and the optional user interface may also include a standard wired interface or a wireless interface.

[0118] The network interface 304 may include, but is not limited to, Bluetooth modules, NFC modules, Wi-Fi modules, etc.

[0119] The device processor 301 may include one or more processing cores. The device processor 301 connects to various parts within the electronic device 300 using various interfaces and lines. It executes various functions and processes data of the electronic device 300 by running or executing instructions, programs, code sets, or instruction sets stored in the memory 305, and by calling data stored in the memory 305. Optionally, the device processor 301 may be implemented using at least one hardware form of DSP, FPGA, or PLA. The device processor 301 may integrate one or more of the following: CPU, GPU, and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the device processor 301 and may be implemented as a separate chip.

[0120] The memory 305 may include RAM or ROM. Optionally, the memory 305 may include a non-transitory computer-readable medium. The memory 305 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 305 may also be at least one storage device located remotely from the aforementioned device processor 301. Figure 3 As shown, the memory 305, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and program instructions.

[0121] Specifically, the device processor 301 can be used to call the photovoltaic power generation dust accumulation prediction application stored in the memory 305, and specifically perform the following operations:

[0122] A historical photovoltaic dataset is obtained and divided into a historical dust-free dataset and a historical dust-accumulated dataset. The historical photovoltaic dataset includes historical meteorological data and historical operational data corresponding to each historical meteorological data. The historical meteorological data includes temperature, ambient humidity, wind speed, irradiance, atmospheric pressure, and rainfall. The historical operational data includes historical actual photovoltaic power generation and dust pollution ratio.

[0123] A cleanroom power prediction model is constructed based on the CEEMDAN-CNN-LSTM network and the historical cleanroom dataset, and a historical loss dataset is generated based on the historical dust accumulation dataset and the cleanroom power prediction model.

[0124] A dust accumulation power loss prediction model is constructed based on the historical loss dataset, and the target dust accumulation prediction power corresponding to the meteorological data to be measured is determined based on the dust-free power prediction model and the dust accumulation power loss prediction model.

[0125] As an optional embodiment of this application, dividing the historical photovoltaic dataset into a historical dust-free dataset and a historical dust-accumulated dataset includes:

[0126] Anomalies are removed from the historical photovoltaic dataset based on the rainfall data to obtain a qualified historical dataset.

[0127] Based on the dust contamination ratio, the historical qualified dataset is divided into a historical dust-free dataset and a historical dust-accumulated dataset.

[0128] As an optional embodiment of this application, after dividing the historical photovoltaic dataset into a historical dust-free dataset and a historical dust-accumulated dataset, the method further includes:

[0129] For any historical meteorological data in the historical photovoltaic dataset, the historical meteorological data is normalized based on the Min-Max method to obtain standard meteorological data.

[0130] As an optional embodiment of this application, the step of constructing a cleanroom power prediction model based on the CEEMDAN-CNN-LSTM network and the historical cleanroom dataset includes:

[0131] A CEEMDAN-CNN-LSTM network is constructed based on the CEEMDAN method and the CNN-LSTM network.

[0132] The CEEMDAN-CNN-LSTM network is trained based on the historical cleanliness dataset to obtain a cleanliness power prediction model. The historical cleanliness dataset includes first historical meteorological data as input to the cleanliness power prediction model and first historical power as output of the cleanliness power prediction model.

[0133] As an optional embodiment of this application, the step of generating a historical loss dataset based on the historical dust accumulation dataset and the cleanroom power prediction model includes:

[0134] The data in the historical dust accumulation dataset are determined as each second historical meteorological data and the second historical power corresponding to each second historical meteorological data;

[0135] Based on the cleanroom power prediction model, determine the historical cleanroom prediction power corresponding to each of the second historical meteorological data.

[0136] A historical loss dataset is constructed based on each second historical meteorological data and the difference between each historical dust-free predicted power and each second historical power corresponding to each second historical meteorological data.

[0137] As an optional embodiment of this application, the step of constructing a dust accumulation power loss prediction model based on the historical loss dataset includes:

[0138] An initial dust accumulation power loss prediction model was constructed based on a BO-LSTM network.

[0139] The initial dust accumulation power loss prediction model is trained based on the historical loss dataset to obtain the dust accumulation power loss prediction model.

[0140] As an optional embodiment of this application, determining the target dust accumulation prediction power corresponding to the meteorological data to be measured based on the cleanroom power prediction model and the dust accumulation power loss prediction model includes:

[0141] The target dust-free predicted power corresponding to the meteorological data to be measured is determined based on the dust-free power prediction model.

[0142] The target dust loss power corresponding to the meteorological data to be measured is determined based on the dust power loss prediction model.

[0143] The target dust accumulation prediction power is determined based on the target dust-free prediction power and the target dust accumulation loss power.

[0144] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method. The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, as well as magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0145] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0146] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0147] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some service interface; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0148] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0149] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0150] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0151] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0152] The foregoing has described specific embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for predicting the dust accumulation power of photovoltaic power generation, characterized in that, The method includes: A historical photovoltaic dataset is obtained and divided into a historical dust-free dataset and a historical dust-accumulated dataset. The historical photovoltaic dataset includes historical meteorological data and historical operational data corresponding to each historical meteorological data. The historical meteorological data includes temperature, ambient humidity, wind speed, irradiance, atmospheric pressure, and rainfall. The historical operational data includes historical actual photovoltaic power generation and dust pollution ratio. A cleanroom power prediction model is constructed based on the CEEMDAN-CNN-LSTM network and the historical cleanroom dataset, and a historical loss dataset is generated based on the historical dust accumulation dataset and the cleanroom power prediction model. A dust accumulation power loss prediction model is constructed based on the historical loss dataset, and the target dust accumulation prediction power corresponding to the meteorological data to be measured is determined based on the dust-free power prediction model and the dust accumulation power loss prediction model.

2. The method according to claim 1, characterized in that, The step of dividing the historical photovoltaic dataset into a historical dust-free dataset and a historical dust-accumulated dataset includes: Anomalies are removed from the historical photovoltaic dataset based on the rainfall data to obtain a qualified historical dataset. Based on the dust contamination ratio, the historical qualified dataset is divided into a historical dust-free dataset and a historical dust-accumulated dataset.

3. The method according to claim 1, characterized in that, After dividing the historical photovoltaic dataset into historical dust-free dataset and historical dust-accumulated dataset, the method further includes: For any historical meteorological data in the historical photovoltaic dataset, the historical meteorological data is normalized based on the Min-Max method to obtain standard meteorological data.

4. The method according to claim 1, characterized in that, The cleanroom power prediction model constructed based on the CEEMDAN-CNN-LSTM network and the historical cleanroom dataset includes: A CEEMDAN-CNN-LSTM network is constructed based on the CEEMDAN method and the CNN-LSTM network. The CEEMDAN-CNN-LSTM network is trained based on the historical cleanliness dataset to obtain a cleanliness power prediction model. The historical cleanliness dataset includes first historical meteorological data as input to the cleanliness power prediction model and first historical power as output of the cleanliness power prediction model.

5. The method according to claim 1, characterized in that, The step of generating a historical loss dataset based on the historical dust accumulation dataset and the cleanroom power prediction model includes: The data in the historical dust accumulation dataset are determined as each second historical meteorological data and the second historical power corresponding to each second historical meteorological data; Based on the cleanroom power prediction model, determine the historical cleanroom prediction power corresponding to each of the second historical meteorological data. A historical loss dataset is constructed based on each second historical meteorological data and the difference between each historical dust-free predicted power and each second historical power corresponding to each second historical meteorological data.

6. The method according to claim 1, characterized in that, The step of constructing a dust accumulation power loss prediction model based on the historical loss dataset includes: An initial dust accumulation power loss prediction model was constructed based on a BO-LSTM network. The initial dust accumulation power loss prediction model is trained based on the historical loss dataset to obtain the dust accumulation power loss prediction model.

7. The method according to claim 1, characterized in that, The step of determining the target dust accumulation prediction power corresponding to the meteorological data to be measured based on the dust-free power prediction model and the dust accumulation power loss prediction model includes: The target dust-free predicted power corresponding to the meteorological data to be measured is determined based on the dust-free power prediction model. The target dust loss power corresponding to the meteorological data to be measured is determined based on the dust power loss prediction model. The target dust accumulation prediction power is determined based on the target dust-free prediction power and the target dust accumulation loss power.

8. A device for predicting the dust accumulation power of photovoltaic power generation, characterized in that, The device includes: The acquisition module is used to acquire historical photovoltaic datasets and divide the historical photovoltaic datasets into historical dust-free datasets and historical dust-accumulated datasets. The historical photovoltaic datasets include historical meteorological data and historical operating data corresponding to each historical meteorological data. The historical meteorological data include temperature, ambient humidity, wind speed, irradiance, atmospheric pressure, and rainfall. The historical operating data includes historical actual photovoltaic power generation and dust pollution ratio. The module is used to build a dust-free power prediction model based on the CEEMDAN-CNN-LSTM network and the historical dust-free dataset, and to generate a historical loss dataset based on the historical dust accumulation dataset and the dust-free power prediction model. The determination module is used to construct a dust accumulation power loss prediction model based on the historical loss dataset, and to determine the target dust accumulation prediction power corresponding to the meteorological data to be measured based on the dust-free power prediction model and the dust accumulation power loss prediction model.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer-readable storage medium stores instructions that, when executed on a computer or processor, cause the computer or processor to perform the steps of the method as described in any one of claims 1-7.

Images