Photovoltaic power generation power prediction method, device, equipment and medium

Abstract

The invention discloses a photovoltaic power generation power prediction method and device, equipment and a medium. The method comprises the following steps: determining a cleaning interval duration from a to-be-predicted day to a previous cleaning operation based on the to-be-predicted day and a cleaning operation plan of a photovoltaic module in a photovoltaic power station; under the condition that the cleaning interval duration indicates that the photovoltaic module is in a clean state, determining a first photovoltaic power generation power prediction value through a first prediction model based on meteorological parameters corresponding to the to-be-predicted day; and under the condition that the cleaning interval duration indicates that the photovoltaic module is in a dust deposition state, determining a second photovoltaic power generation power prediction value through a second prediction model based on the meteorological parameters corresponding to the to-be-predicted day and the cleaning interval duration. When the photovoltaic power generation power is predicted, different prediction models are selected according to whether the photovoltaic module is subjected to dust deposition or not, and the influence of meteorological factors and dust deposition factors of the photovoltaic module on the predicted photovoltaic power generation power is considered, so that the prediction result is more reasonable and accurate.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of photovoltaic power generation technology, and in particular to a photovoltaic power generation power prediction method, device, equipment and medium. Background Art

[0002] Photovoltaic power generation prediction is an important part of smart grid construction. By predicting photovoltaic power generation, it can help the power dispatching system to adjust the power of wind power, thermal power, and hydropower based on the predicted photovoltaic power generation, which is beneficial to the stability of the overall operation of the power grid.

[0003] At present, the prediction of photovoltaic power generation can be achieved through a prediction model based on similar days, that is, the photovoltaic power generation of historical days with the same or similar climate conditions as the day to be predicted is used as a reference to determine the predicted photovoltaic power generation. The prediction of photovoltaic power generation in this way only considers the impact of climate conditions on photovoltaic power generation, resulting in low prediction accuracy. Summary of the invention

[0004] The present invention provides a photovoltaic power generation power prediction method, device, equipment and medium, which can improve the accuracy of photovoltaic power generation power prediction.

[0005] In a first aspect, an embodiment of the present invention provides a photovoltaic power generation power prediction method, comprising:

[0006] Determine the cleaning interval between the predicted day and the last cleaning operation based on the cleaning operation plan of the photovoltaic components in the photovoltaic power station;

[0007] When the cleaning interval indicates that the photovoltaic assembly is in a clean state, determining a first photovoltaic power generation power prediction value based on the meteorological parameters corresponding to the day to be predicted by using a first prediction model;

[0008] When the cleaning interval duration indicates that the photovoltaic component is in a dust accumulation state, a second photovoltaic power generation power prediction value is determined by a second prediction model based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration.

[0009] In a second aspect, an embodiment of the present invention provides a photovoltaic power generation power prediction device, comprising:

[0010] A first determination module is used to determine the cleaning interval time between the to-be-predicted day and the last cleaning operation based on the to-be-predicted day and the cleaning operation plan of the photovoltaic components in the photovoltaic power station;

[0011] A second determination module is used to determine a first photovoltaic power generation prediction value based on the meteorological parameters corresponding to the day to be predicted by using a first prediction model when the cleaning interval duration indicates that the photovoltaic assembly is in a clean state;

[0012] The third determination module is used to determine a second photovoltaic power generation prediction value through a second prediction model based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration when the cleaning interval duration indicates that the photovoltaic component is in a dust accumulation state.

[0013] In a third aspect, an embodiment of the present invention provides an electronic device, including:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the method according to the first aspect.

[0017] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0018] The technical solution of the embodiment of the present invention determines the cleaning interval between the predicted day and the last cleaning operation through the cleaning operation plan of the photovoltaic modules in the photovoltaic power station, determines whether the photovoltaic modules are in a clean state or a dusty state through the cleaning interval, predicts the photovoltaic power generation power based on the meteorological parameters of the predicted day through the first prediction model in the clean state, and predicts the photovoltaic power generation power based on the meteorological parameters and the cleaning interval through the second prediction model in the dusty state. When predicting the photovoltaic power generation power, this solution selects different prediction models based on whether the photovoltaic modules are dusty, comprehensively considers the influence of meteorological factors and dust accumulation factors of the photovoltaic modules on the predicted photovoltaic power generation power, and makes the prediction result more reasonable and accurate.

[0019] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 is a flow chart of a photovoltaic power generation power prediction method provided according to the first embodiment of the present invention;

[0022] Figure 2 is a flow chart of a photovoltaic power generation power prediction method provided according to the second embodiment of the present invention;

[0023] Figure 3 is a flow chart of a photovoltaic power generation power prediction method provided according to Embodiment 3 of the present invention;

[0024] Figure 4 is a schematic diagram of the structure of a photovoltaic power prediction device provided according to a fourth embodiment of the present invention;

[0025] Figure 5 It is a schematic diagram of the structure of an electronic device implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0027] It should be noted that the terms "first", "second", etc. in the present invention are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] Embodiment 1

[0029] Figure 1This is a flow chart of a photovoltaic power prediction method provided according to Embodiment 1 of the present invention. This embodiment is applicable to the case of predicting photovoltaic power. The method can be performed by a photovoltaic power prediction device, which can be implemented in the form of software and / or hardware and integrated in an electronic device. Furthermore, the electronic device includes but is not limited to: computers, laptops, servers, etc.

[0030] like Figure 1 As shown, the method includes:

[0031] S110: Based on the day to be predicted and the cleaning operation plan of the photovoltaic components in the photovoltaic power station, determine the cleaning interval time between the day to be predicted and the last cleaning operation.

[0032] In practical applications, photovoltaic power stations usually include a large number of photovoltaic modules, which are responsible for converting sunlight into electrical energy. The design, performance and maintenance of photovoltaic modules directly affect the power generation efficiency and stability of photovoltaic power stations. The cleanliness of photovoltaic modules has a direct impact on their power generation efficiency. With the accumulation of foreign matter such as dust, the photoelectric conversion efficiency of photovoltaic modules gradually decreases. Therefore, it is necessary to clean the photovoltaic modules in photovoltaic power stations.

[0033] The cleaning operation plan may be a plan for cleaning the photovoltaic modules. The cleaning operation plan may be set according to the operation and maintenance plan of the photovoltaic power station, season, rainfall, frequency of wind and sand, etc., which are not limited here. Through the cleaning operation plan, the cleaning operation cycle of the photovoltaic modules in the photovoltaic power station can be determined.

[0034] The day to be predicted can be the date on which the photovoltaic power generation prediction is to be performed, that is, the photovoltaic power generation on the day to be predicted is predicted. The cleaning interval can be the interval between the day to be predicted and the last cleaning operation, and the last cleaning operation is the most recent cleaning operation before the day to be predicted.

[0035] In this step, the cleaning operation plan of the photovoltaic components can be used to determine the date of the most recent cleaning operation before the predicted date, and the cleaning interval can be determined by the time difference between the predicted date and the date of the most recent cleaning operation before the predicted date.

[0036] S120: When the cleaning interval duration indicates that the photovoltaic assembly is in a clean state, determine a first photovoltaic power generation power prediction value through a first prediction model based on meteorological parameters corresponding to the day to be predicted.

[0037] In the embodiment of the present invention, the judgment condition for whether the photovoltaic module is in a clean state is not limited and can be determined according to the actual application scenario. For example, when the cleaning interval is zero, it indicates that the photovoltaic module is cleaned on the day to be predicted, and the photovoltaic module is in a clean state; for another example, when the cleaning interval is one day, it indicates that the photovoltaic module is cleaned one day before the day to be predicted, and the photovoltaic module is in a clean state.

[0038] The first prediction model can be a model for predicting photovoltaic power generation based on meteorological parameters corresponding to the day to be predicted. The model structure of the first prediction model is not limited, as long as it can predict photovoltaic power generation based on meteorological parameters. The first photovoltaic power generation prediction value can be the photovoltaic power generation corresponding to the day to be predicted predicted by the first prediction model.

[0039] The meteorological parameters corresponding to the day to be predicted may be parameters indicating the weather conditions of the day to be predicted. Optionally, the meteorological parameters may include, but are not limited to: maximum, minimum and mean values ​​of temperature; maximum, minimum and mean values ​​of humidity; maximum, minimum and mean values ​​of irradiance.

[0040] In this step, when it is determined that the photovoltaic module is in a clean state, the meteorological parameters corresponding to the day to be predicted are input into the first prediction model. The first prediction model can determine a historical day with the same or similar meteorological parameters as the day to be predicted from multiple historical days based on the meteorological parameters corresponding to the day to be predicted as the first similar day to the day to be predicted, and determine the photovoltaic power generation power of the first similar day as the first photovoltaic power generation power prediction value. Among them, the historical day is the date before the day to be predicted, and the historical day and its number are not limited here. The first similar day is a historical day with the same or similar meteorological parameters as the day to be predicted.

[0041] S130. When the cleaning interval duration indicates that the photovoltaic component is in a dust accumulation state, determine a second photovoltaic power generation power prediction value through a second prediction model based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration.

[0042] In the embodiment of the present invention, the judgment condition of whether the photovoltaic module is in the dust accumulation state is not limited, and can be determined according to the actual application scenario. For example, when the cleaning interval is longer than one day, the photovoltaic module can be considered to be in the dust accumulation state.

[0043] The second prediction model may be a model for predicting photovoltaic power generation based on meteorological parameters and cleaning interval duration corresponding to the day to be predicted. The model structure of the second prediction model is not limited, as long as it can predict photovoltaic power generation based on meteorological parameters and cleaning interval duration. The second photovoltaic power generation prediction value may be the photovoltaic power generation corresponding to the day to be predicted predicted based on the second prediction model.

[0044] In this step, when it is determined that the photovoltaic module is in a dust accumulation state, the meteorological parameters and cleaning interval duration corresponding to the day to be predicted are input into the second prediction model. The second prediction model can determine from multiple historical days the meteorological parameters that are the same or similar to the meteorological parameters of the day to be predicted, and the number of days from the previous cleaning operation is the same or similar to the aforementioned cleaning interval duration as the second similar day to the day to be predicted, and the photovoltaic power generation power of the second similar day is determined as the second photovoltaic power generation power prediction value. Among them, the second similar day is a historical day with the same or similar meteorological parameters as the meteorological parameters of the day to be predicted, and the number of days from the previous cleaning operation is the same or similar to the cleaning interval duration.

[0045] It should be noted that, in the embodiment of the present invention, after S110, the execution order of S120 and S130 is not limited. S120 is executed when the photovoltaic module is in a clean state, and S130 is executed when the photovoltaic module is in a dusty state.

[0046] The technical solution of the embodiment of the present invention determines the cleaning interval between the predicted day and the last cleaning operation through the cleaning operation plan of the photovoltaic modules in the photovoltaic power station, determines whether the photovoltaic modules are in a clean state or a dusty state through the cleaning interval, predicts the photovoltaic power generation power based on the meteorological parameters of the predicted day through the first prediction model in the clean state, and predicts the photovoltaic power generation power based on the meteorological parameters and the cleaning interval through the second prediction model in the dusty state. When predicting the photovoltaic power generation power, this solution selects different prediction models based on whether the photovoltaic modules are dusty, comprehensively considers the influence of meteorological factors and dust accumulation factors of the photovoltaic modules on the predicted photovoltaic power generation power, and makes the prediction result more reasonable and accurate.

[0047] Embodiment 2

[0048] Figure 2 It is a flow chart of a photovoltaic power prediction method provided according to the second embodiment of the present invention. This embodiment is based on the above-mentioned first embodiment and further refines the first photovoltaic power prediction value determined by the first prediction model based on the meteorological parameters corresponding to the predicted day.

[0049] like Figure 2 As shown, the method includes:

[0050] S110: Based on the day to be predicted and the cleaning operation plan of the photovoltaic components in the photovoltaic power station, determine the cleaning interval time between the day to be predicted and the last cleaning operation.

[0051] S121. When the cleaning interval duration indicates that the photovoltaic component is in a clean state, cluster the data in the first photovoltaic power generation data set through a first prediction model to obtain multiple first clustering results, and determine the weather type corresponding to each first clustering result, wherein the first photovoltaic power generation power data set includes the first historical photovoltaic power generation power of each historical day in multiple historical days.

[0052] The first photovoltaic power generation data set may be a collection of multiple first historical photovoltaic power generation powers. The first historical photovoltaic power generation power may be the photovoltaic power generation power corresponding to any one of multiple historical days. Multiple historical days may be dates on which historical photovoltaic power generation power is collected when the first photovoltaic power generation data set is constructed.

[0053] In one embodiment, the process of constructing the first photovoltaic power generation data set includes:

[0054] For each of the plurality of historical days, a reference photovoltaic power generation power of a photovoltaic power station model corresponding to the photovoltaic power station is monitored, and photovoltaic components of the photovoltaic power station model are cleaned daily;

[0055] Determining the ideal photovoltaic power generation power of the photovoltaic power station based on the reference photovoltaic power generation power and the ratio of the photovoltaic power station model to the photovoltaic component area of ​​the photovoltaic power station;

[0056] Correcting the ideal photovoltaic power generation power by the actual photovoltaic power generation power of the photovoltaic power station to obtain the first historical photovoltaic power generation power of each historical day;

[0057] The set of the first historical photovoltaic power generation powers is determined as the first photovoltaic power generation power data set.

[0058] In an embodiment of the present invention, the photovoltaic power station model can be a model that simulates the operation of an actual photovoltaic power station. Specifically, a preset location can be selected, which is as close as possible to the geographical environment of the actual photovoltaic power station, including topography, climate conditions, etc., to ensure the representativeness of the photovoltaic power station model; the photovoltaic power station model is set at the preset location, and the scale of the photovoltaic power station model can be smaller than the scale of the actual photovoltaic power station, but it must be able to simulate the operation of the actual photovoltaic power station, which is not limited here; the photovoltaic components of the photovoltaic power station model are cleaned every day to keep the photovoltaic components of the photovoltaic power station model clean at all times; the photovoltaic power generation power of the photovoltaic power station model is monitored on each historical day, that is, the reference photovoltaic power generation power corresponding to the historical day is obtained. Among them, the reference photovoltaic power generation power can be the photovoltaic power generation power of the photovoltaic power station model.

[0059] Based on the reference photovoltaic power generation, the ideal photovoltaic power generation corresponding to the photovoltaic power station on the historical day can be determined by using the known photovoltaic power station model and the photovoltaic module area ratio of the photovoltaic power station. The photovoltaic module area ratio can be the ratio of the total area of ​​the photovoltaic modules used in the photovoltaic power station model to the total area of ​​the photovoltaic modules in the actual photovoltaic power station. The ideal photovoltaic power generation can be the photovoltaic power generation that should be achieved if the photovoltaic modules of the photovoltaic power station are always kept clean.

[0060] Taking into account the many limitations of actual photovoltaic power stations, the actual photovoltaic power generation of photovoltaic power stations is often lower than the ideal state. Therefore, it is necessary to correct the ideal photovoltaic power generation power through the actual photovoltaic power generation power of the photovoltaic power station on that historical day, such as by using linear regression or least squares sinusoidal fitting. After correction, the first historical photovoltaic power generation power corresponding to that historical day can be obtained.

[0061] A set of first historical photovoltaic power generation powers corresponding to each historical day is determined as a first photovoltaic power generation power data set.

[0062] Through the first prediction model, multiple first historical photovoltaic power generation powers in the first photovoltaic power generation data set are analyzed, and based on the statistical characteristics of each first historical photovoltaic power generation power, a clustering algorithm is used to cluster multiple first historical photovoltaic power generation powers to obtain multiple first clustering results; the multiple clustering results are combined with the weather types released by the Meteorological Bureau, that is, the corresponding weather type is divided for each first clustering result. Among them, the statistical characteristics may include but are not limited to: daily power generation maximum value, daily power generation total power, daily power generation average power, daily power generation variance, daily power generation standard deviation, daily power generation coefficient of variation, etc. Weather types may include but are not limited to: sunny, sunny to cloudy, cloudy, rainy, and hazy.

[0063] S122. Determine a first target weather type corresponding to the day to be predicted based on the meteorological parameter corresponding to the day to be predicted and the Euclidean distance between the meteorological parameter corresponding to the meteorological cluster center of each first clustering result.

[0064] In this step, the meteorological parameters corresponding to the day to be predicted may be meteorological parameters that are highly correlated with the multiple first historical photovoltaic power generation powers in the first photovoltaic power generation data set, selected from the meteorological parameters of the day to be predicted. The meteorological cluster center may be a cluster center related to meteorological factors in the first clustering result, such as the meteorological parameters corresponding to each first historical photovoltaic power generation power in the first clustering result, and the center point position of the cluster may be determined as the meteorological cluster center corresponding to the first clustering result.

[0065] In this step, by calculating the Euclidean distance between the meteorological parameters corresponding to the day to be predicted and the meteorological parameters corresponding to each meteorological cluster center, the weather type corresponding to the first clustering result corresponding to the smallest Euclidean distance among multiple Euclidean distances is determined as the weather type corresponding to the day to be predicted, that is, the first target weather type.

[0066] The calculation formula of Euclidean distance is as follows:

[0067]

[0068] Among them, D i represents the Euclidean distance, i represents a first clustering result, that is, i represents a weather type; j represents the jth meteorological parameter, J represents the total number of meteorological parameters; x0(j) represents the meteorological parameter corresponding to the day to be predicted; x i (j) represents the meteorological parameters corresponding to the meteorological cluster center of the i-th first clustering result.

[0069] S123: Based on the meteorological parameters corresponding to the day to be predicted, determine a first similar day to the day to be predicted from the first target clustering result corresponding to the first target weather type.

[0070] In this step, a historical day with meteorological parameters identical or similar to those of the day to be predicted can be determined from multiple historical days corresponding to the first target clustering result corresponding to the first target weather type as the first similar day to the day to be predicted. The first target clustering result can be the first clustering result corresponding to the first target weather type.

[0071] In one embodiment, based on the meteorological parameters corresponding to the day to be predicted, determining a first similar day of the day to be predicted from the first target clustering result corresponding to the first target weather type includes:

[0072] Determining a plurality of first target historical days included in the first target clustering result;

[0073] For any first target historical day, determine a first characteristic correlation coefficient mean value between the meteorological parameter corresponding to the to-be-predicted day and the meteorological parameter corresponding to any first target historical day;

[0074] The first target historical day corresponding to the maximum value among the mean values ​​of the plurality of first characteristic correlation coefficients is determined as the first similar day of the day to be predicted.

[0075] Among them, the first target historical day may be a historical day included in the first target clustering result, and the first target clustering result includes multiple first target historical days, and the multiple first target historical days all correspond to the first target weather type.

[0076] For any first target historical day, the corresponding meteorological parameters include multiple ones, and the meteorological parameters corresponding to the day to be predicted also include multiple ones. The meteorological parameters corresponding to the day to be predicted are taken as the reference sequence, and the meteorological parameters corresponding to any first target historical day are taken as the subsequence, and the parameters of the two sequences are normalized to be dimensionless; the characteristic correlation coefficient of each meteorological parameter in the reference sequence and each meteorological parameter in the subsequence is calculated, and the calculated multiple characteristic correlation coefficients are averaged to obtain the first characteristic correlation coefficient mean, and the characteristic correlation coefficient is a coefficient indicating the degree of correlation between the two meteorological parameters.

[0077] Through the above operation, the first characteristic correlation coefficient mean of the meteorological parameters corresponding to the day to be predicted and the meteorological parameters corresponding to each first target historical day can be obtained, and the first target historical day corresponding to the maximum value among multiple first characteristic correlation coefficient means is determined as the first similar day of the day to be predicted.

[0078] S124: Determine the first historical photovoltaic power generation power of the first similar day as the first photovoltaic power generation power prediction value.

[0079] In this step, the first photovoltaic power generation data set is concentrated, and the first historical photovoltaic power generation power of the first similar day is determined as the first photovoltaic power generation power prediction value.

[0080] S130. When the cleaning interval duration indicates that the photovoltaic component is in a dust accumulation state, determine a second photovoltaic power generation power prediction value through a second prediction model based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration.

[0081] The technical solution of the embodiment of the present invention, when the photovoltaic module is in a clean state, determines the first photovoltaic power generation prediction value according to the meteorological parameters corresponding to the day to be predicted through the first prediction model. The model makes full use of the potential laws of multiple first historical photovoltaic power generation, and realizes the refined prediction of photovoltaic power generation under different weather types through cluster analysis, which is of great significance for improving the intelligent management level of photovoltaic power stations.

[0082] Embodiment 3

[0083] Figure 3 It is a flow chart of a photovoltaic power prediction method provided according to Example 3 of the present invention. This embodiment is based on the above-mentioned Example 1 and further refines the second photovoltaic power prediction value determined by the second prediction model based on the meteorological parameters corresponding to the predicted day and the cleaning interval duration.

[0084] like Figure 3 As shown, the method includes:

[0085] S110: Based on the day to be predicted and the cleaning operation plan of the photovoltaic components in the photovoltaic power station, determine the cleaning interval time between the day to be predicted and the last cleaning operation.

[0086] S120: When the cleaning interval duration indicates that the photovoltaic assembly is in a clean state, determine a first photovoltaic power generation power prediction value through a first prediction model based on meteorological parameters corresponding to the day to be predicted.

[0087] S131. When the cleaning interval indicates that the photovoltaic component is in a dust accumulation state, cluster the data in the second photovoltaic power generation data set through a second prediction model to obtain multiple second clustering results, and determine the weather type corresponding to each second clustering result, wherein the second photovoltaic power generation power data set includes the second historical photovoltaic power generation power of each historical day in multiple historical days.

[0088] The second photovoltaic power generation data set may be a collection of multiple second historical photovoltaic power generation powers. The second historical photovoltaic power generation power may be the photovoltaic power generation power corresponding to any one of multiple historical days. The multiple historical days may be the dates on which the historical photovoltaic power generation power is collected when the second photovoltaic power generation data set is constructed.

[0089] Optionally, in actual applications, the actual photovoltaic power generation power of the photovoltaic power station can be monitored on each of multiple historical days as the second historical photovoltaic power generation power corresponding to the historical day, and the set of the second historical photovoltaic power generation powers corresponding to each historical day can be determined as the second photovoltaic power generation power data set.

[0090] Through the second prediction model, multiple second historical photovoltaic power generation powers in the second photovoltaic power generation data are analyzed, and based on the statistical characteristics of each second historical photovoltaic power generation power, a clustering algorithm is used to cluster multiple second historical photovoltaic power generation powers to obtain multiple second clustering results; the multiple clustering results are combined with the weather types released by the Meteorological Bureau, that is, the corresponding weather type is divided for each second clustering result. Among them, the statistical characteristics may include but are not limited to: daily power generation maximum value, daily power generation total power, daily power generation average power, daily power generation variance, daily power generation standard deviation, daily power generation coefficient of variation, etc. Weather types may include but are not limited to: sunny, sunny to cloudy, cloudy, rainy, and hazy.

[0091] S132. Determine a second target weather type corresponding to the day to be predicted based on the meteorological parameter corresponding to the day to be predicted and the Euclidean distance between the meteorological parameter corresponding to the meteorological cluster center of each second clustering result.

[0092] In this step, the meteorological parameters corresponding to the day to be predicted may be meteorological parameters that are highly correlated with the multiple second historical photovoltaic power generation powers in the second photovoltaic power generation data set, selected from the meteorological parameters of the day to be predicted. The meteorological cluster center may be a cluster center related to meteorological factors in the second clustering result, such as the meteorological parameters corresponding to each second historical photovoltaic power generation power in the second clustering result, and the center point position of the cluster may be determined as the meteorological cluster center corresponding to the second clustering result.

[0093] In this step, by calculating the Euclidean distance between the meteorological parameters corresponding to the day to be predicted and the meteorological parameters corresponding to each meteorological cluster center, the weather type corresponding to the second clustering result corresponding to the smallest Euclidean distance among multiple Euclidean distances is determined as the weather type corresponding to the day to be predicted, that is, the second target weather type.

[0094] The calculation formula of Euclidean distance is as follows:

[0095]

[0096] Among them, D a represents the Euclidean distance, a represents a second clustering result, that is, a represents a weather type; b represents the bth meteorological parameter, B represents the total number of meteorological parameters; y0(b) represents the meteorological parameter corresponding to the day to be predicted; y a (b) The meteorological parameters corresponding to the meteorological cluster center of the a-th second clustering result.

[0097] S133: Based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration, determine a second similar day to the day to be predicted from the second target clustering results corresponding to the second target weather type.

[0098] In this step, from among the multiple historical days corresponding to the second target clustering result corresponding to the second target weather type, a historical day having the same or similar meteorological parameters as the meteorological parameters of the day to be predicted and the same or similar number of days from the last cleaning operation and the cleaning interval duration can be determined as the second similar day to the day to be predicted. The second target clustering result can be the second clustering result corresponding to the second target weather type.

[0099] In one embodiment, based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration, determining a second similar day of the day to be predicted from the second target clustering result corresponding to the second target weather type includes:

[0100] determining a plurality of second target historical days included in the second target clustering result;

[0101] For any second target historical day, determining a second characteristic correlation coefficient mean value between the meteorological parameter corresponding to the to-be-predicted day and the meteorological parameter corresponding to any second target historical day;

[0102] For any second target historical day, determine the average of the correlation coefficients between the cleaning interval duration and the historical cleaning interval duration from the second target historical day to the previous cleaning operation;

[0103] Taking a weighted average of the second feature correlation coefficient mean and the duration correlation coefficient mean to obtain a comprehensive mean corresponding to any second target historical day;

[0104] The second target historical day corresponding to the maximum value among the multiple comprehensive averages is determined as the second similar day of the day to be predicted.

[0105] Among them, the second target historical day may be a historical day included in the second target clustering result, and the second target clustering result includes multiple second target historical days, and the multiple second target historical days all correspond to the second target weather type.

[0106] For any second target historical day, the corresponding meteorological parameters include multiple ones, and the meteorological parameters corresponding to the day to be predicted also include multiple ones. The meteorological parameters corresponding to the day to be predicted are taken as the reference sequence, and the meteorological parameters corresponding to any second target historical day are taken as the first subsequence. The parameters of the reference sequence and the first subsequence are normalized to be dimensionless; the characteristic correlation coefficient of each meteorological parameter in the reference sequence and each meteorological parameter in the first subsequence is calculated, and the calculated multiple characteristic correlation coefficients are averaged to obtain the second characteristic correlation coefficient mean, and the characteristic correlation coefficient is a coefficient indicating the degree of correlation between the two meteorological parameters.

[0107] There may be multiple historical cleaning interval durations corresponding to any second target historical day. For example, for different photovoltaic modules in a photovoltaic power station, the historical cleaning interval durations between the second target historical day and the previous cleaning operation may be different. Similarly, there may be multiple cleaning interval durations corresponding to the day to be predicted. The cleaning interval duration corresponding to the day to be predicted is taken as the duration reference sequence, and the historical cleaning interval duration corresponding to any second target historical day is taken as the second subsequence. The parameters of the duration reference sequence and the second subsequence are normalized to be dimensionless. The duration correlation coefficient of each duration in the duration reference sequence and each duration in the second subsequence is calculated, and the multiple calculated duration correlation coefficients are averaged to obtain the mean of the duration correlation coefficients. The duration correlation coefficient is a coefficient indicating the degree of correlation between the two durations.

[0108] The first weight is assigned to the mean value of the second characteristic correlation coefficient, and the second weight is assigned to the mean value of the duration correlation coefficient. Then, the comprehensive mean value corresponding to any second target historical day = the mean value of the second characteristic correlation coefficient × the first weight + the mean value of the duration correlation coefficient × the second weight. The second target historical day corresponding to the maximum value among the multiple comprehensive means is determined as the second similar day of the predicted day. The first weight and the second weight can be determined based on actual application needs.

[0109] S134. Determine the second historical photovoltaic power generation power of the second similar day as the second photovoltaic power generation power prediction value.

[0110] In this step, the second photovoltaic power generation data is concentrated, and the second historical photovoltaic power generation power of the second similar day is determined as the second photovoltaic power generation power prediction value.

[0111] The technical solution of the embodiment of the present invention, when the photovoltaic components are in a dusty state, determines the second photovoltaic power generation prediction value through a second prediction model based on the meteorological parameters and the cleaning interval duration corresponding to the predicted day, and comprehensively considers the influence of meteorological factors and dust accumulation factors of the photovoltaic components on the predicted photovoltaic power generation power, so that the prediction result is more reasonable and accurate.

[0112] In one embodiment, the second historical photovoltaic power generation power included in the second photovoltaic power generation power data set is achieved by monitoring the photovoltaic power generation power of the photovoltaic power station on the corresponding historical day and performing data anomaly processing on the monitored photovoltaic power generation power.

[0113] During the second historical photovoltaic power generation collection process, errors in any link such as the working status of the sensor, data transmission, etc. will cause the collected data to be abnormal, such as causing the collected value to be too large, too small or missing relative to the true value.

[0114] The photovoltaic power generation collected on historical days has many adverse effects on the learning process, training process and prediction results of the second prediction model. For example, it will destroy the trend and smoothness of the power curve, interfere with the training process of the second prediction model, and cause the learning error of the second prediction model to be large, making it difficult to meet the set size. For another example, it may also cause misjudgment when testing the prediction performance of the second prediction model. Therefore, if the second prediction model is expected to obtain a higher prediction accuracy, it is necessary to first perform a series of preprocessing on the photovoltaic power generation collected on historical days to ensure the smoothness and order of magnitude consistency of the power data in the time series, and then use the preprocessed power data as the second historical photovoltaic power generation for the learning, training and prediction process of the second prediction model.

[0115] Abnormal data can be divided into the following two situations: 1) the photovoltaic power generation data collected on the historical day is missing or displayed as zero value; 2) the photovoltaic power generation data collected on the historical day suddenly increases or decreases compared with the data collected on the adjacent historical day.

[0116] For case 1), delete the zero value and complete the part where the zero value is deleted and the part where the data is missing. Specifically:

[0117] The zero-value data in the photovoltaic power generation data collected on historical days are deleted, and the part with deleted zero values ​​and the part with missing data are regarded as missing data; several values ​​closest to the missing data are selected, and the missing power values ​​are filled in by weighted average using the following formula.

[0118] p(d)=w1p(d1)+w2p(d2)+w3p(d3)+…+wmp(dm)

[0119] Wherein, p(d) is the missing power data of day d; p(d1), p(d2), p(d3), ..., p(dm) are the power data of day d1, day d2, day d3, ..., day dm selected closest to day d; w1, w2, w3, ..., wm are the weight values ​​of the power data of day d1, day d2, day d3, ..., day dm. Optionally, the aforementioned weight values ​​can be set by calculating the Euclidean distance.

[0120] For situation 2), the collected data is processed by callback, specifically:

[0121] The photovoltaic power generation data collected in the embodiment of the present invention may have obvious seasonal and periodic characteristics. According to the seasonal characteristics, the data of one year is divided into four seasons. At the same time, the photovoltaic power generation data in each season has periodicity.

[0122] Based on the above characteristics, the historically collected photovoltaic power generation is set to p(h,n), where h represents a day in a cycle, h=1,2,…,15; and n represents the nth cycle in a season, n=1,2…,N.

[0123] First, calculate the mean power E(h) and mean square error V(h) of the dates in the same cycle. The calculation formula is as follows:

[0124]

[0125] Then, the mutation rate δ(h,n) of the photovoltaic power generation under the longitudinal coordinate is calculated as follows:

[0126]

[0127] When the mutation rate δ(h,n) is greater than or equal to the set first power generation mutation rate threshold, if n = 1, then If n = N, then If 1<n<N, then There is no limit on the first power generation power mutation rate threshold. The value that replaces p(h,n) after callback.

[0128] When the mutation rate δ(h,n) is less than the first power generation mutation rate threshold,

[0129] It should be noted that the degree of callback can be changed by adjusting the size of the first power generation mutation rate threshold. The larger the first power generation mutation rate threshold, the lower the degree of callback. If the mutation rate δ(h,n) is greater than or equal to the first power generation mutation rate threshold, it means that the point is an abnormal point and the value needs to be adjusted. Replace p(h,n); otherwise, if the mutation rate δ(h,n) is less than the first power generation power mutation rate threshold, no callback processing is required.

[0130] In addition, it should be noted that, since photovoltaic power generation is periodic, after the above callback processing, the photovoltaic power generation data sequence within a cycle should also be guaranteed to have a certain smoothness, that is, it needs to be preprocessed under the horizontal coordinate. The specific processing process is as follows:

[0131] First, set a period of photovoltaic power generation data sequence p(n) = [p(1), p(2), ..., p(15)], and then calculate the median of nine points based on the sequence to generate the first power generation sequence p 1 (n), and then calculate the median of the five points of the new power generation sequence to obtain the second power generation sequence p 2 (n), as follows:

[0132]

[0133] Next, the smoothed power callback sequence p is obtained. p (n), as follows:

[0134] p p (n) = 0.2p 2 (n-1)+0.6p 2 (n)+0.2p 2 (n+1)

[0135] Then calculate the mutation rate ζ(n) in the horizontal coordinate after the power generation is called back, as follows:

[0136]

[0137] If the mutation rate ζ(n) is greater than or equal to the second power generation mutation rate threshold, it means that the point is an abnormal point and the value needs to be called back. The power generation callback sequence p p On the contrary, if the mutation rate ζ(n) is less than the second power generation power mutation rate threshold, then callback processing is not required. The second power generation power mutation rate threshold is not limited.

[0138] The above technical scheme, by processing the missing data or zero-value data in the photovoltaic power generation power collected on historical days, and performing callback processing on the power data that increases or decreases abnormally in the photovoltaic power generation power data, can make the photovoltaic power generation power data have better smoothness and consistency of order of magnitude in the time series, and effectively ensure that the second prediction model can obtain higher prediction accuracy.

[0139] In one embodiment, the second prediction model is trained based on a training sample set and a test sample set; wherein the training sample set includes at least the second photovoltaic power generation data set, the meteorological parameters and historical cleaning interval duration of each historical day corresponding to the second photovoltaic power generation data set, and the meteorological parameters, cleaning interval duration to be trained and photovoltaic power generation power to be trained corresponding to the day to be trained, wherein each historical day includes a third similar day of the day to be trained.

[0140] The training sample set may be a data set for training the second prediction model. The second photovoltaic power generation data set, the meteorological parameters of each historical day corresponding to the second photovoltaic power generation data set, and the historical cleaning interval duration may be collected in advance. A day to be trained may also be set. The day to be trained may be a day after each historical day. The meteorological parameters, the cleaning interval duration to be trained, and the photovoltaic power generation power to be trained corresponding to the day to be trained may be considered as known parameters for model training. Among them, the cleaning interval duration to be trained is the time between the day to be trained and the last cleaning operation; the photovoltaic power generation power to be trained is the photovoltaic power generation power corresponding to the day to be trained.

[0141] The test sample set may be a data set for testing the second prediction model, and may be set based on the training sample set, which is not limited here. The output effect of the second prediction model may be tested through the test sample set, so as to iteratively optimize the second prediction model.

[0142] Based on the second photovoltaic power generation data set, the meteorological parameters and historical cleaning interval durations of each historical day corresponding to the second photovoltaic power generation data set, and the meteorological parameters and cleaning interval durations corresponding to the day to be trained, the third similar day corresponding to the day to be trained is determined. The meteorological parameters and historical cleaning interval durations of the third similar day, and the meteorological parameters and cleaning interval durations corresponding to the day to be trained are used as inputs of the second prediction model, and the second historical photovoltaic power generation power of the third similar day is used as the output of the second prediction model for model training.

[0143] It should be noted that the model parameters can be optimized by the ant colony algorithm during the model training process. Specifically, the initial values ​​of the model weights and thresholds are randomly set as the input of the ant colony algorithm; assuming that there are m ants in the ant colony, each ant iteratively updates the pheromone during the foraging process; determine whether the path of the i-th ant is due to the path of the i-1th ant, if it is determined to be yes, then return to the operation of iteratively updating the pheromone, if it is determined to be no, retain the i-1th group of weights; determine whether all the m ants have walked, if it is determined to be yes, then use the i-1th group of weights as the initial weights of the neural network model, if it is determined to be no, return to the operation of iteratively updating the pheromone; determine whether the error between the predicted value output by the model and the actual value (i.e., the photovoltaic power generation to be trained) is less than the set maximum allowable error, if not, the training is completed; if it is exceeded, the error is fed back to the input layer by back propagation; iteratively update the weights and thresholds of each layer until the error between the predicted value and the actual value does not exceed the maximum allowable error, that is, the second prediction model with the optimal parameters is obtained.

[0144] It should be noted that due to the obvious differences in photovoltaic power generation under different weather conditions, the method of the present invention can effectively predict the short-term power generation of photovoltaic power stations under different meteorological conditions, and use the ant colony algorithm to select the initial weights and thresholds, which can effectively avoid the problem of traditional neural network algorithms falling into local minima when predicting photovoltaic power.

[0145] Embodiment 4

[0146] Figure 4 is a schematic diagram of the structure of a photovoltaic power prediction device provided according to the fourth embodiment of the present invention. This embodiment can be applied to the situation of predicting photovoltaic power, such as Figure 4 As shown, the specific structure of the device includes:

[0147] A first determination module 41 is used to determine the cleaning interval time between the predicted day and the last cleaning operation based on the predicted day and the cleaning operation plan of the photovoltaic components in the photovoltaic power station;

[0148] A second determination module 42 is configured to determine a first photovoltaic power prediction value based on the meteorological parameters corresponding to the day to be predicted by using a first prediction model when the cleaning interval duration indicates that the photovoltaic assembly is in a clean state;

[0149] The third determination module 43 is used to determine a second photovoltaic power generation prediction value based on the meteorological parameters corresponding to the predicted day and the cleaning interval duration through a second prediction model when the cleaning interval duration indicates that the photovoltaic component is in a dust accumulation state.

[0150] The photovoltaic power prediction device provided in this embodiment determines the cleaning interval duration from the to-be-predicted day to the last cleaning operation based on the cleaning operation plan of the photovoltaic components in the photovoltaic power station through the first determination module; determines the first photovoltaic power prediction value through the first prediction model based on the meteorological parameters corresponding to the to-be-predicted day when the cleaning interval duration indicates that the photovoltaic component is in a clean state through the second determination module; determines the second photovoltaic power prediction value through the second prediction model based on the meteorological parameters corresponding to the to-be-predicted day and the cleaning interval duration when the cleaning interval duration indicates that the photovoltaic component is in a dusty state through the third determination module. When predicting the photovoltaic power generation, this solution selects different prediction models based on whether the photovoltaic component is dusty, comprehensively considering the influence of meteorological factors and dust accumulation factors of the photovoltaic component on the predicted photovoltaic power generation, making the prediction result more reasonable and accurate.

[0151] Furthermore, the second determining module 42 is specifically configured to:

[0152] Clustering data in a first photovoltaic power generation data set to obtain a plurality of first clustering results through a first prediction model, and determining a weather type corresponding to each first clustering result, wherein the first photovoltaic power generation data set includes a first historical photovoltaic power generation of each historical day in a plurality of historical days;

[0153] Determine a first target weather type corresponding to the day to be predicted based on the meteorological parameter corresponding to the day to be predicted and the Euclidean distance between the meteorological parameter corresponding to the meteorological cluster center of each first clustering result;

[0154] Based on the meteorological parameters corresponding to the day to be predicted, determining a first similar day to the day to be predicted from the first target clustering result corresponding to the first target weather type;

[0155] The first historical photovoltaic power generation power of the first similar day is determined as the first photovoltaic power generation power prediction value.

[0156] Furthermore, the process of constructing the first photovoltaic power generation data set includes:

[0157] For each of the plurality of historical days, a reference photovoltaic power generation power of a photovoltaic power station model corresponding to the photovoltaic power station is monitored, and photovoltaic components of the photovoltaic power station model are cleaned daily;

[0158] Determining the ideal photovoltaic power generation power of the photovoltaic power station based on the reference photovoltaic power generation power and the ratio of the photovoltaic power station model to the photovoltaic component area of ​​the photovoltaic power station;

[0159] Correcting the ideal photovoltaic power generation power by the actual photovoltaic power generation power of the photovoltaic power station to obtain the first historical photovoltaic power generation power of each historical day;

[0160] The set of the first historical photovoltaic power generation powers is determined as the first photovoltaic power generation power dataset.

[0161] Furthermore, the second determining module 42 is specifically configured to:

[0162] Determining a plurality of first target historical days included in the first target clustering result;

[0163] For any first target historical day, determine a first characteristic correlation coefficient mean value between the meteorological parameter corresponding to the to-be-predicted day and the meteorological parameter corresponding to any first target historical day;

[0164] The first target historical day corresponding to the maximum value among the mean values ​​of the plurality of first characteristic correlation coefficients is determined as the first similar day of the day to be predicted.

[0165] Furthermore, the third determining module 43 is specifically configured to:

[0166] Clustering the data in the second photovoltaic power generation data set to obtain a plurality of second clustering results by using a second prediction model, and determining a weather type corresponding to each second clustering result, wherein the second photovoltaic power generation data set includes a second historical photovoltaic power generation power of each historical day in a plurality of historical days;

[0167] Determine the second target weather type corresponding to the day to be predicted based on the meteorological parameter corresponding to the day to be predicted and the Euclidean distance between the meteorological parameter corresponding to the meteorological cluster center of each second clustering result;

[0168] Based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration, determining a second similar day to the day to be predicted from the second target clustering results corresponding to the second target weather type;

[0169] The second historical photovoltaic power generation power of the second similar day is determined as the second photovoltaic power generation power prediction value.

[0170] Furthermore, the third determining module 43 is specifically configured to:

[0171] determining a plurality of second target historical days included in the second target clustering result;

[0172] For any second target historical day, determining a second characteristic correlation coefficient mean value between the meteorological parameter corresponding to the to-be-predicted day and the meteorological parameter corresponding to any second target historical day;

[0173] For any second target historical day, determine the average of the correlation coefficients between the cleaning interval duration and the historical cleaning interval duration from the second target historical day to the previous cleaning operation;

[0174] Taking a weighted average of the second feature correlation coefficient mean and the duration correlation coefficient mean to obtain a comprehensive mean corresponding to any second target historical day;

[0175] The second target historical day corresponding to the maximum value among the multiple comprehensive averages is determined as the second similar day of the day to be predicted.

[0176] Furthermore, the second prediction model is trained based on a training sample set and a test sample set;

[0177] The training sample set includes at least the second photovoltaic power generation data set, the meteorological parameters and historical cleaning interval duration of each historical day corresponding to the second photovoltaic power generation data set, and the meteorological parameters, cleaning interval duration to be trained and photovoltaic power generation to be trained corresponding to the day to be trained, and each of the historical days includes the third similar day of the day to be trained;

[0178] The second historical photovoltaic power generation power included in the second photovoltaic power generation power data set is achieved by monitoring the photovoltaic power generation power of the photovoltaic power station on the corresponding historical day and performing data anomaly processing on the monitored photovoltaic power generation power.

[0179] The photovoltaic power prediction device provided in the embodiment of the present invention can execute the photovoltaic power prediction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0180] Embodiment 5

[0181] Figure 5Schematic diagram of the structure of the electronic device implementing the embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0182] like Figure 5 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0183] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0184] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a photovoltaic power generation power prediction method.

[0185] In some embodiments, the photovoltaic power generation power prediction method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the photovoltaic power generation power prediction method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the photovoltaic power generation power prediction method in any other appropriate manner (e.g., by means of firmware).

[0186] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0187] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0188] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0189] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback; and the input from the user can be received in any form.

[0190] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0191] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0192] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0193] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A photovoltaic power generation power prediction method, characterized in that: include: Determine the cleaning interval between the predicted day and the last cleaning operation based on the cleaning operation plan of the photovoltaic components in the photovoltaic power station; When the cleaning interval indicates that the photovoltaic assembly is in a clean state, determining a first photovoltaic power generation power prediction value based on the meteorological parameters corresponding to the day to be predicted by using a first prediction model; When the cleaning interval duration indicates that the photovoltaic component is in a dust accumulation state, a second photovoltaic power generation power prediction value is determined by a second prediction model based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration.

2. The method according to claim 1, characterized in that Determining a first photovoltaic power generation power prediction value by a first prediction model based on meteorological parameters corresponding to the day to be predicted includes: Clustering data in a first photovoltaic power generation data set to obtain a plurality of first clustering results through a first prediction model, and determining a weather type corresponding to each first clustering result, wherein the first photovoltaic power generation data set includes a first historical photovoltaic power generation of each historical day in a plurality of historical days; Determine a first target weather type corresponding to the day to be predicted based on the meteorological parameter corresponding to the day to be predicted and the Euclidean distance between the meteorological parameter corresponding to the meteorological cluster center of each first clustering result; Based on the meteorological parameters corresponding to the day to be predicted, determining a first similar day to the day to be predicted from the first target clustering result corresponding to the first target weather type; The first historical photovoltaic power generation power of the first similar day is determined as the first photovoltaic power generation power prediction value.

3. The method according to claim 2, characterized in that The process of constructing the first photovoltaic power generation data set includes: For each of the plurality of historical days, a reference photovoltaic power generation power of a photovoltaic power station model corresponding to the photovoltaic power station is monitored, and photovoltaic components of the photovoltaic power station model are cleaned daily; Determining the ideal photovoltaic power generation power of the photovoltaic power station based on the reference photovoltaic power generation power and the ratio of the photovoltaic power station model to the photovoltaic component area of ​​the photovoltaic power station; Correcting the ideal photovoltaic power generation power by the actual photovoltaic power generation power of the photovoltaic power station to obtain the first historical photovoltaic power generation power of each historical day; The set of the first historical photovoltaic power generation powers is determined as the first photovoltaic power generation power dataset.

4. The method according to claim 2, characterized in that: Based on the meteorological parameters corresponding to the day to be predicted, determining a first similar day to the day to be predicted from the first target clustering result corresponding to the first target weather type, comprising: Determining a plurality of first target historical days included in the first target clustering result; For any first target historical day, determine a first characteristic correlation coefficient mean value between the meteorological parameter corresponding to the to-be-predicted day and the meteorological parameter corresponding to any first target historical day; The first target historical day corresponding to the maximum value among the mean values ​​of the plurality of first characteristic correlation coefficients is determined as the first similar day of the day to be predicted.

5. The method according to claim 1, characterized in that Determining a second photovoltaic power generation power prediction value by a second prediction model based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration includes: Clustering the data in the second photovoltaic power generation data set to obtain a plurality of second clustering results by using a second prediction model, and determining a weather type corresponding to each second clustering result, wherein the second photovoltaic power generation data set includes a second historical photovoltaic power generation power of each historical day in a plurality of historical days; Determine the second target weather type corresponding to the day to be predicted based on the meteorological parameter corresponding to the day to be predicted and the Euclidean distance between the meteorological parameter corresponding to the meteorological cluster center of each second clustering result; Based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration, determining a second similar day to the day to be predicted from the second target clustering results corresponding to the second target weather type; The second historical photovoltaic power generation power of the second similar day is determined as the second photovoltaic power generation power prediction value.

6. The method according to claim 5, characterized in that Based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration, determining a second similar day of the day to be predicted from the second target clustering result corresponding to the second target weather type, including: determining a plurality of second target historical days included in the second target clustering result; For any second target historical day, determining a second characteristic correlation coefficient mean value between the meteorological parameter corresponding to the to-be-predicted day and the meteorological parameter corresponding to any second target historical day; For any second target historical day, determine the average of the correlation coefficients between the cleaning interval duration and the historical cleaning interval duration from the second target historical day to the previous cleaning operation; Taking a weighted average of the second feature correlation coefficient mean and the duration correlation coefficient mean to obtain a comprehensive mean corresponding to any second target historical day; The second target historical day corresponding to the maximum value among the multiple comprehensive averages is determined as the second similar day of the day to be predicted.

7. The method according to claim 5, characterized in that The second prediction model is trained based on a training sample set and a test sample set; The training sample set includes at least the second photovoltaic power generation data set, the meteorological parameters and historical cleaning interval duration of each historical day corresponding to the second photovoltaic power generation data set, and the meteorological parameters, cleaning interval duration to be trained and photovoltaic power generation to be trained corresponding to the day to be trained, and each of the historical days includes the third similar day of the day to be trained; The second historical photovoltaic power generation power included in the second photovoltaic power generation power data set is achieved by monitoring the photovoltaic power generation power of the photovoltaic power station on the corresponding historical day and performing data anomaly processing on the monitored photovoltaic power generation power.

8. A photovoltaic power generation power prediction device, characterized in that: include: A first determination module is used to determine the cleaning interval time between the to-be-predicted day and the last cleaning operation based on the to-be-predicted day and the cleaning operation plan of the photovoltaic components in the photovoltaic power station; A second determination module is used to determine a first photovoltaic power generation prediction value based on the meteorological parameters corresponding to the day to be predicted by using a first prediction model when the cleaning interval duration indicates that the photovoltaic assembly is in a clean state; The third determination module is used to determine a second photovoltaic power generation prediction value through a second prediction model based on the meteorological parameters corresponding to the day to be predicted and the cleaning interval duration when the cleaning interval duration indicates that the photovoltaic component is in a dust accumulation state.

9. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images