Method and device for determining cleaning period of photovoltaic panel

Abstract

The invention discloses a photovoltaic panel cleaning period determination method and device, and relates to the field of photovoltaic cleaning, and the method comprises the steps: obtaining photovoltaic panel characteristic parameters of a target photovoltaic power station, photovoltaic panel operation data and historical meteorological data of a region where the target photovoltaic power station is located; constructing a photovoltaic panel pollution prediction model based on the photovoltaic panel characteristic parameters, the photovoltaic panel operation data and the historical meteorological data, and obtaining a photovoltaic panel pollution time sequence characteristic set according to the photovoltaic panel pollution prediction model; constructing a photovoltaic panel pollution-power generation relation model according to the photovoltaic panel operation data; substituting the photovoltaic panel pollution time sequence feature set and the single cleaning cost of the photovoltaic panel into the photovoltaic panel pollution-power generation relation model to solve and obtain a first cleaning period; and obtaining a first scene according to the historical meteorological data and the region where the target photovoltaic power station is located, and correcting the first cleaning period according to the first scene to obtain a second cleaning period. By implementing the method, the cleaning period of the photovoltaic panel can be accurately determined.

Description

Technical Field

[0001] The present application relates to the field of photovoltaic cleaning, and in particular to a method and device for determining a photovoltaic panel cleaning cycle. Background Art

[0002] Photovoltaic panels, as devices that can convert solar energy into electrical energy, are usually used in areas with sufficient sunlight to convert solar energy into electrical energy to supply human life. Whether the photovoltaic panels are clean or not is closely related to the power generation efficiency of the photovoltaic panels. How to balance the cleaning cost of the photovoltaic panels and the power generation income of the photovoltaic panels is an important factor in determining the cleaning cycle of the photovoltaic panels. In the prior art, when determining or predicting the cleaning cycle of photovoltaic panels, usually only the relationship between the cleaning cost of the photovoltaic panels and the power generation income is considered, while factors such as the self-cleaning ability of the photovoltaic panels, regional climate differences, and differences in the operating conditions of photovoltaic power stations are ignored. Therefore, the determination of the cleaning cycle of the photovoltaic panels is not considered from multiple factors, resulting in deviations in the determination of the cleaning cycle of the photovoltaic panels, thereby affecting the power generation efficiency and power generation income of the photovoltaic panels. Therefore, how to accurately determine the cleaning cycle of photovoltaic panels is still a problem that needs to be solved in the prior art. Summary of the invention

[0003] The present application provides a method and device for determining a photovoltaic panel cleaning cycle to solve the technical problem that the photovoltaic panel cleaning cycle is distorted due to the failure to consider multiple factors in the prior art.

[0004] According to a first aspect of an embodiment of the present application, a method for determining a photovoltaic panel cleaning cycle is provided, comprising:

[0005] Obtain the photovoltaic panel characteristic parameters, photovoltaic panel operation data and historical meteorological data of the target photovoltaic power station; wherein the photovoltaic panel characteristic parameters include photovoltaic panel material, photovoltaic panel inclination and photovoltaic panel self-cleaning performance; the photovoltaic panel operation data includes power generation efficiency data and multiple photovoltaic panel pollution images; the historical meteorological data includes light intensity, rainfall and wind speed;

[0006] Based on the photovoltaic panel characteristic parameters, the photovoltaic panel operation data and the historical meteorological data, a photovoltaic panel pollution prediction model is constructed, and according to the photovoltaic panel pollution prediction model, a photovoltaic panel pollution time series feature set is obtained;

[0007] Constructing a photovoltaic panel pollution-power generation relationship model based on the photovoltaic panel operation data;

[0008] Substituting the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost into the photovoltaic panel pollution-power generation relationship model to solve and obtain the first cleaning cycle;

[0009] A first scenario is obtained based on the historical meteorological data and the region where the target photovoltaic power station is located, and the first cleaning cycle is corrected based on the first scenario to obtain a second cleaning cycle, so as to complete the cleaning of the photovoltaic panels of the target photovoltaic power station according to the second cleaning cycle.

[0010] The present application first obtains the photovoltaic panel characteristic parameters, photovoltaic panel operation data and historical meteorological data of the target photovoltaic power station in the area, and then constructs a photovoltaic panel pollution prediction model based on the photovoltaic panel characteristic parameters and photovoltaic panel operation data, taking into account various factors including photovoltaic panel material, inclination angle, and self-cleaning performance, which can improve the accuracy of the prediction model and obtain a more accurate photovoltaic panel pollution time series feature set, thereby improving the accuracy in determining the photovoltaic panel cleaning cycle; then, a photovoltaic panel pollution-power generation relationship model is constructed based on the photovoltaic panel operation data, and then the photovoltaic panel pollution time series feature set is substituted into the photovoltaic panel pollution-power generation model and solved to obtain the first cleaning cycle, ensuring the basic solution of the cleaning cycle, and then determining the first scenario based on the historical meteorological data and the area where the target photovoltaic power station is located, and correcting the first cleaning cycle according to the first scenario to obtain the second cleaning cycle, taking into account the differences in operating scenarios of photovoltaic power stations, which can improve the accuracy in determining the cleaning cycle of the photovoltaic panel.

[0011] In certain embodiments of the present application, the photovoltaic panel pollution prediction model is constructed based on the photovoltaic panel characteristic parameters, the photovoltaic panel operation data and the historical meteorological data, and the photovoltaic panel pollution time series feature set is obtained according to the photovoltaic panel pollution prediction model, specifically including:

[0012] Characterizing the plurality of photovoltaic panel pollution images to obtain a plurality of first pollution feature vectors, and labeling the plurality of first pollution feature vectors according to a plurality of preset pollution thresholds to obtain a plurality of second pollution feature vectors;

[0013] Characterizing the photovoltaic panel characteristic parameters, the historical meteorological data, and the power generation efficiency data respectively to obtain an installation characteristic vector, a meteorological characteristic tensor, and a power generation efficiency tensor;

[0014] Inputting the installation feature vector, the meteorological feature tensor, the power generation efficiency tensor and the plurality of second pollution feature vectors into a preset initial pollution prediction model for multiple iterative training, optimizing the model parameters of the initial pollution prediction model during iteration, and completing the training when the model parameters of the initial pollution prediction model reach preset values, thereby obtaining a photovoltaic panel pollution prediction model;

[0015] According to the meteorological characteristic tensor and the plurality of second pollution characteristic vectors, based on the photovoltaic panel pollution prediction model, a plurality of third pollution characteristic vectors are predicted;

[0016] A photovoltaic panel pollution time series feature set is obtained according to the multiple second pollution feature vectors and the multiple third pollution feature vectors.

[0017] The present application first characterizes multiple photovoltaic panel pollution images to obtain multiple first pollution feature vectors, and then labels the multiple first pollution feature vectors according to multiple preset pollution thresholds to obtain multiple second pollution feature vectors, which can quantify the pollution labeling value of each photovoltaic panel pollution image, and facilitate subsequent model training; then characterizes the photovoltaic panel characteristic parameters, historical meteorological data and power generation efficiency data, and inputs the characterized installation feature vectors, meteorological feature tensors, power generation efficiency tensors and multiple second pollution feature vectors into the initial pollution prediction model for iterative training to obtain a photovoltaic panel pollution prediction model, which takes into account the influence of various factors on the photovoltaic panel cleaning cycle and can improve the accuracy of the prediction model, and then obtains multiple third pollution feature vectors based on the photovoltaic panel pollution prediction model, and then obtains a photovoltaic panel pollution time series feature set, which can obtain a more accurate photovoltaic panel pollution time series feature set, thereby improving the accuracy in determining the photovoltaic panel cleaning cycle.

[0018] In certain embodiments of the present application, the photovoltaic panel pollution-power generation relationship model is constructed based on the photovoltaic panel operation data, specifically including:

[0019] Analyzing the plurality of photovoltaic panel pollution images according to a plurality of preset photovoltaic panel pollution indicators to obtain a plurality of first pollution data;

[0020] Aligning the power generation efficiency data and the plurality of first pollution data in time sequence, and normalizing them respectively, to obtain power generation efficiency time series data and the plurality of first pollution time series data;

[0021] Linear regression fitting is performed on the power generation efficiency time series data and the plurality of first pollution time series data to obtain a photovoltaic panel pollution-power generation relationship model.

[0022] The present application first analyzes multiple photovoltaic panel pollution images according to multiple preset photovoltaic panel pollution indicators to obtain multiple first pollution data, and can analyze and evaluate the photovoltaic panel pollution images from multiple dimensions, quantify the pollution in each dimension, and then align and normalize the power generation efficiency data and the multiple first pollution data in chronological order, and then perform linear regression fitting to obtain the photovoltaic panel pollution-power generation relationship model, which takes into account the pollution in multiple dimensions when the photovoltaic panels are not cleaned, and can improve the matching degree between the photovoltaic panel pollution-power generation relationship model and the actual situation, thereby improving the accuracy in determining the photovoltaic panel cleaning cycle.

[0023] In certain embodiments of the present application, the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost are substituted into the photovoltaic panel pollution-power generation relationship model to solve the first cleaning cycle, specifically including:

[0024] The photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost are substituted into the photovoltaic panel pollution-power generation relationship model, and a solution is performed based on a heuristic optimization algorithm to obtain a first cleaning cycle.

[0025] This application first substitutes the photovoltaic panel pollution time series feature set and the single cleaning cost of the photovoltaic panel into the photovoltaic panel pollution-power generation relationship model, and solves it based on the heuristic optimization algorithm to obtain the first cleaning cycle. According to the heuristic optimization algorithm, the basic first cleaning cycle can be solved more accurately.

[0026] In certain embodiments of the present application, obtaining a first scenario according to the historical meteorological data and the region where the target photovoltaic power station is located, and correcting the first cleaning cycle according to the first scenario to obtain a second cleaning cycle specifically includes:

[0027] According to the location of the target photovoltaic power station, the geographical location of the target photovoltaic power station is obtained;

[0028] According to the historical meteorological data, obtaining a first maximum value, a first minimum value of the rainfall and a second maximum value of the wind speed;

[0029] According to the first maximum value, the first minimum value, the second maximum value and the geographical location, a first scene is obtained based on a preset scene library matching;

[0030] The first cleaning cycle is modified according to the first scenario to obtain a second cleaning cycle.

[0031] The present application first obtains the geographical location of the target photovoltaic power station according to the area where it is located, and then obtains the first maximum value, the first minimum value of rainfall and the second maximum value of wind speed in the historical meteorological data. The geographical location and extreme values ​​are more conducive to determining the scene of the target photovoltaic power station, and then the first scene is obtained based on the preset scene library matching, and the first cleaning cycle is corrected according to the first scene to obtain the second cleaning cycle, which can make the cleaning cycle more in line with the actual cleaning needs of the photovoltaic panels of the target photovoltaic power station, thereby improving the accuracy in determining the cleaning cycle of the photovoltaic panels.

[0032] According to a second aspect of the embodiments of the present application, there is provided a photovoltaic panel cleaning cycle determination device, comprising a data acquisition module, a prediction processing module, a model building module, a cycle solving module and a cycle correction module;

[0033] The data acquisition module is used to acquire the photovoltaic panel characteristic parameters, photovoltaic panel operation data and historical meteorological data of the target photovoltaic power station; wherein the photovoltaic panel characteristic parameters include photovoltaic panel material, photovoltaic panel inclination and photovoltaic panel self-cleaning performance; the photovoltaic panel operation data includes power generation efficiency data and multiple photovoltaic panel pollution images; the historical meteorological data includes light intensity, rainfall and wind speed;

[0034] The prediction processing module is used to construct a photovoltaic panel pollution prediction model based on the photovoltaic panel characteristic parameters, the photovoltaic panel operation data and the historical meteorological data, and obtain a photovoltaic panel pollution time series feature set according to the photovoltaic panel pollution prediction model;

[0035] The model building module is used to build a photovoltaic panel pollution-power generation relationship model based on the photovoltaic panel operation data;

[0036] The cycle solving module is used to substitute the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost into the photovoltaic panel pollution-power generation relationship model to solve and obtain the first cleaning cycle;

[0037] The cycle correction module is used to obtain a first scenario based on the historical meteorological data and the area where the target photovoltaic power station is located, and to correct the first cleaning cycle based on the first scenario to obtain a second cleaning cycle, so as to complete the cleaning of the photovoltaic panels of the target photovoltaic power station according to the second cleaning cycle.

[0038] In certain embodiments of the present application, the prediction processing module includes a first characterization unit, a second characterization unit, a model training unit, a model prediction unit, and a data acquisition unit;

[0039] The first characterization unit is used to characterize the multiple photovoltaic panel pollution images to obtain multiple first pollution feature vectors, and mark the multiple first pollution feature vectors according to multiple preset pollution thresholds to obtain multiple second pollution feature vectors;

[0040] The second characterization unit is used to characterize the photovoltaic panel characteristic parameters, the historical meteorological data and the power generation efficiency data respectively to obtain an installation characteristic vector, a meteorological characteristic tensor and a power generation efficiency tensor;

[0041] The model training unit is used to input the installation feature vector, the meteorological feature tensor, the power generation efficiency tensor and the multiple second pollution feature vectors into a preset initial pollution prediction model for multiple iterative training, optimize the model parameters of the initial pollution prediction model during iteration, and complete the training when the model parameters of the initial pollution prediction model reach a preset value to obtain a photovoltaic panel pollution prediction model;

[0042] The model prediction unit is used to predict a plurality of third pollution feature vectors based on the photovoltaic panel pollution prediction model according to the meteorological feature tensor and the plurality of second pollution feature vectors;

[0043] The data acquisition unit is used to obtain a photovoltaic panel pollution time series feature set according to the multiple second pollution feature vectors and the multiple third pollution feature vectors.

[0044] In certain embodiments of the present application, the model building module includes a pollution analysis unit, a normalization processing unit, and a model building unit;

[0045] The pollution analysis unit is used to analyze the plurality of photovoltaic panel pollution images according to a plurality of preset photovoltaic panel pollution indicators to obtain a plurality of first pollution data;

[0046] The normalization processing unit is used to align the power generation efficiency data and the plurality of first pollution data in time sequence, and normalize them respectively to obtain power generation efficiency time series data and the plurality of first pollution time series data;

[0047] The model building unit is used to perform linear regression fitting on the power generation efficiency time series data and the multiple first pollution time series data to obtain a photovoltaic panel pollution-power generation relationship model.

[0048] In certain embodiments of the present application, the period solving module includes a period solving unit;

[0049] The cycle solving unit is used to substitute the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost into the photovoltaic panel pollution-power generation relationship model, solve it based on the heuristic optimization algorithm, and obtain the first cleaning cycle.

[0050] In certain embodiments of the present application, the period correction module includes a position acquisition unit, an extreme value acquisition unit, a scene matching unit, and a period correction unit;

[0051] The location acquisition unit is used to acquire the geographical location of the target photovoltaic power station according to the location of the target photovoltaic power station;

[0052] The extreme value acquisition unit is used to acquire the first maximum value, the first minimum value of the rainfall and the second maximum value of the wind speed according to the historical meteorological data;

[0053] The scene matching unit is used to obtain a first scene based on a preset scene library matching according to the first maximum value, the first minimum value, the second maximum value and the geographical location;

[0054] The cycle correction unit is used to correct the first cleaning cycle according to the first scenario to obtain a second cleaning cycle.

[0055] The present application first obtains the photovoltaic panel characteristic parameters, photovoltaic panel operation data and historical meteorological data of the target photovoltaic power station in the area, and then constructs a photovoltaic panel pollution prediction model based on the photovoltaic panel characteristic parameters and photovoltaic panel operation data, taking into account various factors including photovoltaic panel material, inclination angle, and self-cleaning performance, which can improve the accuracy of the prediction model and obtain a more accurate photovoltaic panel pollution time series feature set, thereby improving the accuracy in determining the photovoltaic panel cleaning cycle; then, a photovoltaic panel pollution-power generation relationship model is constructed based on the photovoltaic panel operation data, and then the photovoltaic panel pollution time series feature set is substituted into the photovoltaic panel pollution-power generation model and solved to obtain the first cleaning cycle, ensuring the basic solution of the cleaning cycle, and then determining the first scenario based on the historical meteorological data and the area where the target photovoltaic power station is located, and correcting the first cleaning cycle according to the first scenario to obtain the second cleaning cycle, taking into account the differences in operating scenarios of photovoltaic power stations, which can improve the accuracy in determining the cleaning cycle of the photovoltaic panel. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 : A schematic flow chart of a method for determining a photovoltaic panel cleaning cycle according to certain embodiments of the present application;

[0057] Figure 2 : A module structure diagram of a photovoltaic panel cleaning cycle determination device shown in certain embodiments of the present application. DETAILED DESCRIPTION

[0058] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals represent the same or similar elements or elements having the same or similar functions from beginning to end. The embodiments described below in conjunction with the accompanying drawings are exemplary and are only used to explain some embodiments of the present application, and should not be construed as limitations on the embodiments of the present application. Based on the embodiments shown in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0059] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, unless otherwise clearly and specifically defined, "multiple" and "several" mean two or more.

[0060] At present, when determining or predicting the cleaning cycle of photovoltaic panels, generally only the relationship between the cleaning cost of photovoltaic panels and the power generation income is considered, and the self-cleaning ability of photovoltaic panels, regional climate differences, and differences in the operating scenarios of photovoltaic power stations are not considered. The determination of the cleaning cycle of photovoltaic panels fails to take multiple factors into consideration, resulting in certain deviations in determining the cleaning cycle of photovoltaic panels. Therefore, how to accurately determine the cleaning cycle of photovoltaic panels is still a problem that needs to be solved urgently in the existing technology.

[0061] Based on the above technical background, please refer to Figure 1 The embodiment of the present application provides a method for determining a photovoltaic panel cleaning cycle, including steps S101 to S105, each of which is specifically as follows:

[0062] Step S101: Acquire the photovoltaic panel characteristic parameters, photovoltaic panel operation data and historical meteorological data of the target photovoltaic power station; wherein the photovoltaic panel characteristic parameters include photovoltaic panel material, photovoltaic panel inclination and photovoltaic panel self-cleaning performance; the photovoltaic panel operation data includes power generation efficiency data and multiple photovoltaic panel pollution images; the historical meteorological data includes light intensity, rainfall and wind speed.

[0063] Step S102: constructing a photovoltaic panel pollution prediction model based on the photovoltaic panel characteristic parameters, the photovoltaic panel operation data and the historical meteorological data, and obtaining a photovoltaic panel pollution time series feature set according to the photovoltaic panel pollution prediction model.

[0064] In certain embodiments of the present application, the photovoltaic panel pollution prediction model is constructed based on the photovoltaic panel characteristic parameters, the photovoltaic panel operation data and the historical meteorological data, and the photovoltaic panel pollution time series feature set is obtained according to the photovoltaic panel pollution prediction model, specifically including:

[0065] Characterizing the plurality of photovoltaic panel pollution images to obtain a plurality of first pollution feature vectors, and labeling the plurality of first pollution feature vectors according to a plurality of preset pollution thresholds to obtain a plurality of second pollution feature vectors;

[0066] Characterizing the photovoltaic panel characteristic parameters, the historical meteorological data, and the power generation efficiency data respectively to obtain an installation characteristic vector, a meteorological characteristic tensor, and a power generation efficiency tensor;

[0067] Inputting the installation feature vector, the meteorological feature tensor, the power generation efficiency tensor and the plurality of second pollution feature vectors into a preset initial pollution prediction model for multiple iterative training, optimizing the model parameters of the initial pollution prediction model during iteration, and completing the training when the model parameters of the initial pollution prediction model reach preset values, thereby obtaining a photovoltaic panel pollution prediction model;

[0068] According to the meteorological characteristic tensor and the plurality of second pollution characteristic vectors, based on the photovoltaic panel pollution prediction model, a plurality of third pollution characteristic vectors are predicted;

[0069] A photovoltaic panel pollution time series feature set is obtained according to the multiple second pollution feature vectors and the multiple third pollution feature vectors.

[0070] Considering vectorizing the characteristic parameters of photovoltaic panels and tensorizing the historical meteorological data and power generation efficiency data is because the characteristic parameters of photovoltaic panels are essentially a one-dimensional combination of the characteristic parameters of photovoltaic panels, while the historical meteorological data is a multidimensional data set of time-light intensity-rainfall-wind speed-photovoltaic panel number, and the power generation efficiency data is a multidimensional data set of at least three dimensions of time-photovoltaic panel number-power generation efficiency. Vectorizing or tensorizing them separately is conducive to representing the actual data set and facilitating the model's learning of the data set during subsequent model training.

[0071] In certain embodiments of the present application, preferred implementations of the initial pollution prediction model include but are not limited to LSTM models, GRU models, Transformer models and ARI MA models.

[0072] In certain embodiments of the present application, obtaining a photovoltaic panel pollution time series feature set according to the plurality of second pollution feature vectors and the plurality of third pollution feature vectors specifically includes:

[0073] The plurality of second pollution feature vectors are concatenated with the plurality of third pollution feature vectors to obtain a photovoltaic panel pollution time series feature set.

[0074] The present application first characterizes multiple photovoltaic panel pollution images to obtain multiple first pollution feature vectors, and then labels the multiple first pollution feature vectors according to multiple preset pollution thresholds to obtain multiple second pollution feature vectors, which can quantify the pollution labeling value of each photovoltaic panel pollution image, and facilitate subsequent model training; then characterizes the photovoltaic panel characteristic parameters, historical meteorological data and power generation efficiency data, and inputs the characterized installation feature vectors, meteorological feature tensors, power generation efficiency tensors and multiple second pollution feature vectors into the initial pollution prediction model for iterative training to obtain a photovoltaic panel pollution prediction model, which takes into account the influence of various factors on the photovoltaic panel cleaning cycle and can improve the accuracy of the prediction model, and then obtains multiple third pollution feature vectors based on the photovoltaic panel pollution prediction model, and then obtains a photovoltaic panel pollution time series feature set, which can obtain a more accurate photovoltaic panel pollution time series feature set, thereby improving the accuracy in determining the photovoltaic panel cleaning cycle.

[0075] Step S103: constructing a photovoltaic panel pollution-power generation relationship model based on the photovoltaic panel operation data.

[0076] In certain embodiments of the present application, the photovoltaic panel pollution-power generation relationship model is constructed based on the photovoltaic panel operation data, specifically including:

[0077] Analyzing the plurality of photovoltaic panel pollution images according to a plurality of preset photovoltaic panel pollution indicators to obtain a plurality of first pollution data;

[0078] Aligning the power generation efficiency data and the plurality of first pollution data in time sequence, and normalizing them respectively, to obtain power generation efficiency time series data and the plurality of first pollution time series data;

[0079] Linear regression fitting is performed on the power generation efficiency time series data and the plurality of first pollution time series data to obtain a photovoltaic panel pollution-power generation relationship model.

[0080] In certain embodiments of the present application, preferred implementations of the plurality of preset photovoltaic panel pollution indicators are photovoltaic panel pollution area, photovoltaic panel pollution type and photovoltaic panel pollution duration.

[0081] In certain embodiments of the present application, a preferred implementation of the linear regression fitting is multiple linear regression.

[0082] The present application first analyzes multiple photovoltaic panel pollution images according to multiple preset photovoltaic panel pollution indicators to obtain multiple first pollution data, and can analyze and evaluate the photovoltaic panel pollution images from multiple dimensions, quantify the pollution in each dimension, and then align and normalize the power generation efficiency data and the multiple first pollution data in chronological order, and then perform linear regression fitting to obtain the photovoltaic panel pollution-power generation relationship model, which takes into account the pollution in multiple dimensions when the photovoltaic panels are not cleaned, and can improve the matching degree between the photovoltaic panel pollution-power generation relationship model and the actual situation, thereby improving the accuracy in determining the photovoltaic panel cleaning cycle.

[0083] Step S104: Substituting the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost into the photovoltaic panel pollution-power generation relationship model to solve and obtain the first cleaning cycle.

[0084] In certain embodiments of the present application, the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost are substituted into the photovoltaic panel pollution-power generation relationship model to solve the first cleaning cycle, specifically including:

[0085] The photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost are substituted into the photovoltaic panel pollution-power generation relationship model, and a solution is performed based on a heuristic optimization algorithm to obtain a first cleaning cycle.

[0086] In certain embodiments of the present application, the heuristic optimization algorithm includes but is not limited to a genetic algorithm, a simulated annealing algorithm or a particle swarm optimization algorithm, preferably a genetic algorithm.

[0087] This application first substitutes the photovoltaic panel pollution time series feature set and the single cleaning cost of the photovoltaic panel into the photovoltaic panel pollution-power generation relationship model, and solves it based on the heuristic optimization algorithm to obtain the first cleaning cycle. According to the heuristic optimization algorithm, the basic first cleaning cycle can be solved more accurately.

[0088] Step S105: obtaining a first scenario based on the historical meteorological data and the region where the target photovoltaic power station is located, and correcting the first cleaning cycle based on the first scenario to obtain a second cleaning cycle, so as to complete the cleaning of the photovoltaic panels of the target photovoltaic power station according to the second cleaning cycle.

[0089] In certain embodiments of the present application, obtaining a first scenario according to the historical meteorological data and the region where the target photovoltaic power station is located, and correcting the first cleaning cycle according to the first scenario to obtain a second cleaning cycle specifically includes:

[0090] According to the location of the target photovoltaic power station, the geographical location of the target photovoltaic power station is obtained;

[0091] According to the historical meteorological data, obtaining a first maximum value, a first minimum value of the rainfall and a second maximum value of the wind speed;

[0092] According to the first maximum value, the first minimum value, the second maximum value and the geographical location, a first scene is obtained based on a preset scene library matching;

[0093] The first cleaning cycle is modified according to the first scenario to obtain a second cleaning cycle.

[0094] In certain embodiments of the present application, the geographical location of the target photovoltaic power station is actually the longitude and latitude coordinates corresponding to the center point of the photovoltaic panel power generation site in the target photovoltaic power station.

[0095] In certain embodiments of the present application, the scenes in the scene library include but are not limited to desert, next door and plain scenes.

[0096] In certain embodiments of the present application, the first cleaning cycle is modified according to the first scenario to obtain a second cleaning cycle, specifically including:

[0097] Obtaining a scene correction coefficient of the first scene in the scene library;

[0098] The first cleaning cycle is corrected according to the scene correction coefficient to obtain a second cleaning cycle.

[0099] In certain embodiments of the present application, after the first cleaning cycle is modified according to the first scenario to obtain the second cleaning cycle, the method further includes:

[0100] Substituting the second cleaning cycle back into the photovoltaic panel pollution-power generation relationship model, and inversely solving to obtain the single cleaning cost of the second photovoltaic panel;

[0101] When the first ratio of the single cleaning cost of the second photovoltaic panel to the single cleaning cost of the photovoltaic panel is greater than the preset cleaning cost ratio threshold, the scene correction coefficient is adjusted according to the first ratio and the second cleaning cycle is iteratively solved until the first ratio corresponding to the second cleaning cycle is less than the preset cleaning cost ratio threshold.

[0102] The present application first obtains the geographical location of the target photovoltaic power station according to the area where it is located, and then obtains the first maximum value, the first minimum value of rainfall and the second maximum value of wind speed in the historical meteorological data. The geographical location and extreme values ​​are more conducive to determining the scene of the target photovoltaic power station, and then the first scene is obtained based on the preset scene library matching, and the first cleaning cycle is corrected according to the first scene to obtain the second cleaning cycle, which can make the cleaning cycle more in line with the actual cleaning needs of the photovoltaic panels of the target photovoltaic power station, thereby improving the accuracy in determining the cleaning cycle of the photovoltaic panels.

[0103] Compared with the prior art, the present application first obtains the photovoltaic panel characteristic parameters, photovoltaic panel operation data and historical meteorological data of the target photovoltaic power station in the area, and then constructs a photovoltaic panel pollution prediction model based on the photovoltaic panel characteristic parameters and photovoltaic panel operation data, taking into account various factors including photovoltaic panel material, inclination angle, and self-cleaning performance, which can improve the accuracy of the prediction model and obtain a more accurate photovoltaic panel pollution time series feature set, thereby improving the accuracy in determining the photovoltaic panel cleaning cycle; then, a photovoltaic panel pollution-power generation relationship model is constructed based on the photovoltaic panel operation data, and then the photovoltaic panel pollution time series feature set is substituted into the photovoltaic panel pollution-power generation model and solved to obtain the first cleaning cycle, ensuring the basic solution of the cleaning cycle, and then determining the first scenario based on the historical meteorological data and the area where the target photovoltaic power station is located, and correcting the first cleaning cycle based on the first scenario to obtain the second cleaning cycle, taking into account the differences in operating scenarios of photovoltaic power stations, which can improve the accuracy in determining the cleaning cycle of the photovoltaic panel.

[0104] Corresponding to the previous method, see Figure 2 , the embodiment of the present application provides a photovoltaic panel cleaning cycle determination device, including a data acquisition module 210, a prediction processing module 220, a model building module 230, a cycle solving module 240 and a cycle correction module 250;

[0105] The data acquisition module 210 is used to acquire the photovoltaic panel characteristic parameters, photovoltaic panel operation data and historical meteorological data of the target photovoltaic power station; wherein the photovoltaic panel characteristic parameters include photovoltaic panel material, photovoltaic panel inclination and photovoltaic panel self-cleaning performance; the photovoltaic panel operation data includes power generation efficiency data and multiple photovoltaic panel pollution images; the historical meteorological data includes light intensity, rainfall and wind speed;

[0106] The prediction processing module 220 is used to construct a photovoltaic panel pollution prediction model based on the photovoltaic panel characteristic parameters, the photovoltaic panel operation data and the historical meteorological data, and obtain a photovoltaic panel pollution time series feature set according to the photovoltaic panel pollution prediction model;

[0107] The model building module 230 is used to build a photovoltaic panel pollution-power generation relationship model based on the photovoltaic panel operation data;

[0108] The cycle solving module 240 is used to substitute the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost into the photovoltaic panel pollution-power generation relationship model to solve and obtain the first cleaning cycle;

[0109] The cycle correction module 250 is used to obtain a first scenario based on the historical meteorological data and the area where the target photovoltaic power station is located, and to correct the first cleaning cycle based on the first scenario to obtain a second cleaning cycle, so as to complete the cleaning of the photovoltaic panels of the target photovoltaic power station according to the second cleaning cycle.

[0110] In certain embodiments of the present application, the prediction processing module 220 includes a first characterization unit 221, a second characterization unit 222, a model training unit 223, a model prediction unit 224, and a data acquisition unit 225;

[0111] The first characterization unit 221 is used to characterize the multiple photovoltaic panel pollution images to obtain multiple first pollution feature vectors, and mark the multiple first pollution feature vectors according to multiple preset pollution thresholds to obtain multiple second pollution feature vectors;

[0112] The second characterization unit 222 is used to characterize the photovoltaic panel characteristic parameters, the historical meteorological data and the power generation efficiency data respectively to obtain an installation characteristic vector, a meteorological characteristic tensor and a power generation efficiency tensor;

[0113] The model training unit 223 is used to input the installation feature vector, the meteorological feature tensor, the power generation efficiency tensor and the multiple second pollution feature vectors into a preset initial pollution prediction model for multiple iterative training, optimize the model parameters of the initial pollution prediction model during iteration, and complete the training when the model parameters of the initial pollution prediction model reach a preset value to obtain a photovoltaic panel pollution prediction model;

[0114] The model prediction unit 224 is used to predict a plurality of third pollution feature vectors based on the photovoltaic panel pollution prediction model according to the meteorological feature tensor and the plurality of second pollution feature vectors;

[0115] The data acquisition unit 225 is used to obtain a photovoltaic panel pollution time series feature set according to the multiple second pollution feature vectors and the multiple third pollution feature vectors.

[0116] In certain embodiments of the present application, the model building module 230 includes a pollution analysis unit 231, a normalization processing unit 232, and a model building unit 233;

[0117] The pollution analysis unit 231 is used to analyze the plurality of photovoltaic panel pollution images according to a plurality of preset photovoltaic panel pollution indicators to obtain a plurality of first pollution data;

[0118] The normalization processing unit 232 is used to align the power generation efficiency data and the plurality of first pollution data in time sequence, and normalize them respectively to obtain power generation efficiency time series data and the plurality of first pollution time series data;

[0119] The model building unit 233 is used to perform linear regression fitting on the power generation efficiency time series data and the plurality of first pollution time series data to obtain a photovoltaic panel pollution-power generation relationship model.

[0120] In certain embodiments of the present application, the period solving module 240 includes a period solving unit 241;

[0121] The cycle solving unit 241 is used to substitute the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost into the photovoltaic panel pollution-power generation relationship model, and solve it based on the heuristic optimization algorithm to obtain the first cleaning cycle.

[0122] In certain embodiments of the present application, the period correction module 250 includes a position acquisition unit 251, an extreme value acquisition unit 252, a scene matching unit 253, and a period correction unit 254;

[0123] The location acquisition unit 251 is used to acquire the geographical location of the target photovoltaic power station according to the location of the target photovoltaic power station;

[0124] The extreme value acquisition unit 252 is used to acquire the first maximum value, the first minimum value of the rainfall and the second maximum value of the wind speed according to the historical meteorological data;

[0125] The scene matching unit 253 is used to obtain a first scene based on a preset scene library matching according to the first maximum value, the first minimum value, the second maximum value and the geographical location;

[0126] The cycle correction unit 254 is used to correct the first cleaning cycle according to the first scenario to obtain a second cleaning cycle.

[0127] Compared with the prior art, the present application first obtains the photovoltaic panel characteristic parameters, photovoltaic panel operation data and historical meteorological data of the target photovoltaic power station in the area, and then constructs a photovoltaic panel pollution prediction model based on the photovoltaic panel characteristic parameters and photovoltaic panel operation data, taking into account various factors including photovoltaic panel material, inclination angle, and self-cleaning performance, which can improve the accuracy of the prediction model and obtain a more accurate photovoltaic panel pollution time series feature set, thereby improving the accuracy in determining the photovoltaic panel cleaning cycle; then, a photovoltaic panel pollution-power generation relationship model is constructed based on the photovoltaic panel operation data, and then the photovoltaic panel pollution time series feature set is substituted into the photovoltaic panel pollution-power generation model and solved to obtain the first cleaning cycle, ensuring the basic solution of the cleaning cycle, and then determining the first scenario based on the historical meteorological data and the area where the target photovoltaic power station is located, and correcting the first cleaning cycle based on the first scenario to obtain the second cleaning cycle, taking into account the differences in operating scenarios of photovoltaic power stations, which can improve the accuracy in determining the cleaning cycle of the photovoltaic panel.

[0128] It should be understood that the device provided in the embodiment of the present application corresponds to the aforementioned method, and a photovoltaic panel cleaning cycle determination device provided in the embodiment of the present application can implement a photovoltaic panel cleaning cycle determination method provided in any embodiment of the present application.

[0129] Adaptively, the embodiments of the present application also provide a computer device and a computer-readable storage medium.

[0130] The computer device comprises: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor;

[0131] Wherein, when the processor executes the computer program, a method for determining a photovoltaic panel cleaning cycle of the present application is implemented.

[0132] The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute a photovoltaic panel cleaning cycle determination method of the present application.

[0133] The above are some implementation methods of the present application, and the purpose, technical scheme and beneficial effects of the present application are further described in detail. It should be clear that the above-mentioned partial implementation methods of the present application cannot be understood as limiting the present application. It is particularly pointed out that for those skilled in the art, any changes, modifications, equivalent substitutions and variations made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for determining a photovoltaic panel cleaning cycle, characterized in that: include: Obtain the photovoltaic panel characteristic parameters, photovoltaic panel operation data and historical meteorological data of the target photovoltaic power station; wherein the photovoltaic panel characteristic parameters include photovoltaic panel material, photovoltaic panel inclination and photovoltaic panel self-cleaning performance; the photovoltaic panel operation data includes power generation efficiency data and multiple photovoltaic panel pollution images; the historical meteorological data includes light intensity, rainfall and wind speed; Based on the photovoltaic panel characteristic parameters, the photovoltaic panel operation data and the historical meteorological data, a photovoltaic panel pollution prediction model is constructed, and according to the photovoltaic panel pollution prediction model, a photovoltaic panel pollution time series feature set is obtained; Constructing a photovoltaic panel pollution-power generation relationship model based on the photovoltaic panel operation data; Substituting the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost into the photovoltaic panel pollution-power generation relationship model to solve and obtain the first cleaning cycle; A first scenario is obtained based on the historical meteorological data and the region where the target photovoltaic power station is located, and the first cleaning cycle is corrected based on the first scenario to obtain a second cleaning cycle, so as to complete the cleaning of the photovoltaic panels of the target photovoltaic power station according to the second cleaning cycle.

2. A method for determining a photovoltaic panel cleaning cycle according to claim 1, characterized in that: The photovoltaic panel pollution prediction model is constructed based on the photovoltaic panel characteristic parameters, the photovoltaic panel operation data and the historical meteorological data, and the photovoltaic panel pollution time series feature set is obtained according to the photovoltaic panel pollution prediction model, specifically including: Characterizing the plurality of photovoltaic panel pollution images to obtain a plurality of first pollution feature vectors, and labeling the plurality of first pollution feature vectors according to a plurality of preset pollution thresholds to obtain a plurality of second pollution feature vectors; Characterizing the photovoltaic panel characteristic parameters, the historical meteorological data, and the power generation efficiency data respectively to obtain an installation characteristic vector, a meteorological characteristic tensor, and a power generation efficiency tensor; Inputting the installation feature vector, the meteorological feature tensor, the power generation efficiency tensor and the plurality of second pollution feature vectors into a preset initial pollution prediction model for multiple iterative training, optimizing the model parameters of the initial pollution prediction model during iteration, and completing the training when the model parameters of the initial pollution prediction model reach preset values, thereby obtaining a photovoltaic panel pollution prediction model; According to the meteorological characteristic tensor and the plurality of second pollution characteristic vectors, based on the photovoltaic panel pollution prediction model, a plurality of third pollution characteristic vectors are predicted; A photovoltaic panel pollution time series feature set is obtained according to the multiple second pollution feature vectors and the multiple third pollution feature vectors.

3. A method for determining a photovoltaic panel cleaning cycle according to claim 1, characterized in that: The step of constructing a photovoltaic panel pollution-power generation relationship model based on the photovoltaic panel operation data specifically includes: Analyzing the plurality of photovoltaic panel pollution images according to a plurality of preset photovoltaic panel pollution indicators to obtain a plurality of first pollution data; Aligning the power generation efficiency data and the plurality of first pollution data in time sequence, and normalizing them respectively, to obtain power generation efficiency time series data and the plurality of first pollution time series data; Linear regression fitting is performed on the power generation efficiency time series data and the plurality of first pollution time series data to obtain a photovoltaic panel pollution-power generation relationship model.

4. A method for determining a photovoltaic panel cleaning cycle according to claim 1, characterized in that: Substituting the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost into the photovoltaic panel pollution-power generation relationship model to solve the first cleaning cycle specifically includes: The photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost are substituted into the photovoltaic panel pollution-power generation relationship model, and a solution is performed based on a heuristic optimization algorithm to obtain a first cleaning cycle.

5. A method for determining a photovoltaic panel cleaning cycle according to claim 1, characterized in that: The step of obtaining a first scenario according to the historical meteorological data and the region where the target photovoltaic power station is located, and correcting the first cleaning cycle according to the first scenario to obtain a second cleaning cycle specifically includes: According to the location of the target photovoltaic power station, the geographical location of the target photovoltaic power station is obtained; According to the historical meteorological data, obtaining a first maximum value, a first minimum value of the rainfall and a second maximum value of the wind speed; According to the first maximum value, the first minimum value, the second maximum value and the geographical location, a first scene is obtained based on a preset scene library matching; The first cleaning cycle is modified according to the first scenario to obtain a second cleaning cycle.

6. A photovoltaic panel cleaning cycle determination device, characterized in that: It includes a data acquisition module, a prediction processing module, a model building module, a cycle solving module and a cycle correction module; The data acquisition module is used to acquire the photovoltaic panel characteristic parameters, photovoltaic panel operation data and historical meteorological data of the target photovoltaic power station; wherein the photovoltaic panel characteristic parameters include photovoltaic panel material, photovoltaic panel inclination and photovoltaic panel self-cleaning performance; the photovoltaic panel operation data includes power generation efficiency data and multiple photovoltaic panel pollution images; the historical meteorological data includes light intensity, rainfall and wind speed; The prediction processing module is used to construct a photovoltaic panel pollution prediction model based on the photovoltaic panel characteristic parameters, the photovoltaic panel operation data and the historical meteorological data, and obtain a photovoltaic panel pollution time series feature set according to the photovoltaic panel pollution prediction model; The model building module is used to build a photovoltaic panel pollution-power generation relationship model based on the photovoltaic panel operation data; The cycle solving module is used to substitute the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost into the photovoltaic panel pollution-power generation relationship model to solve and obtain the first cleaning cycle; The cycle correction module is used to obtain a first scenario based on the historical meteorological data and the area where the target photovoltaic power station is located, and to correct the first cleaning cycle based on the first scenario to obtain a second cleaning cycle, so as to complete the cleaning of the photovoltaic panels of the target photovoltaic power station according to the second cleaning cycle.

7. A photovoltaic panel cleaning cycle determination device according to claim 6, characterized in that: The prediction processing module includes a first characterization unit, a second characterization unit, a model training unit, a model prediction unit and a data acquisition unit; The first characterization unit is used to characterize the multiple photovoltaic panel pollution images to obtain multiple first pollution feature vectors, and mark the multiple first pollution feature vectors according to multiple preset pollution thresholds to obtain multiple second pollution feature vectors; The second characterization unit is used to characterize the photovoltaic panel characteristic parameters, the historical meteorological data and the power generation efficiency data respectively to obtain an installation characteristic vector, a meteorological characteristic tensor and a power generation efficiency tensor; The model training unit is used to input the installation feature vector, the meteorological feature tensor, the power generation efficiency tensor and the multiple second pollution feature vectors into a preset initial pollution prediction model for multiple iterative training, optimize the model parameters of the initial pollution prediction model during iteration, and complete the training when the model parameters of the initial pollution prediction model reach a preset value to obtain a photovoltaic panel pollution prediction model; The model prediction unit is used to predict a plurality of third pollution feature vectors based on the photovoltaic panel pollution prediction model according to the meteorological feature tensor and the plurality of second pollution feature vectors; The data acquisition unit is used to obtain a photovoltaic panel pollution time series feature set according to the multiple second pollution feature vectors and the multiple third pollution feature vectors.

8. A photovoltaic panel cleaning cycle determination device according to claim 6, characterized in that: The model building module includes a pollution analysis unit, a normalization processing unit and a model building unit; The pollution analysis unit is used to analyze the plurality of photovoltaic panel pollution images according to a plurality of preset photovoltaic panel pollution indicators to obtain a plurality of first pollution data; The normalization processing unit is used to align the power generation efficiency data and the plurality of first pollution data in time sequence, and normalize them respectively to obtain power generation efficiency time series data and the plurality of first pollution time series data; The model building unit is used to perform linear regression fitting on the power generation efficiency time series data and the multiple first pollution time series data to obtain a photovoltaic panel pollution-power generation relationship model.

9. A photovoltaic panel cleaning cycle determination device according to claim 6, characterized in that: The period solving module includes a period solving unit; The cycle solving unit is used to substitute the photovoltaic panel pollution time series feature set and the preset photovoltaic panel single cleaning cost into the photovoltaic panel pollution-power generation relationship model, solve it based on the heuristic optimization algorithm, and obtain the first cleaning cycle.

10. The photovoltaic panel cleaning cycle determination device according to claim 6, characterized in that: The period correction module includes a position acquisition unit, an extreme value acquisition unit, a scene matching unit and a period correction unit; The location acquisition unit is used to acquire the geographical location of the target photovoltaic power station according to the location of the target photovoltaic power station; The extreme value acquisition unit is used to acquire the first maximum value, the first minimum value of the rainfall and the second maximum value of the wind speed according to the historical meteorological data; The scene matching unit is used to obtain a first scene based on a preset scene library matching according to the first maximum value, the first minimum value, the second maximum value and the geographical location; The cycle correction unit is used to correct the first cleaning cycle according to the first scenario to obtain a second cleaning cycle.

Images