Photovoltaic user identification method and device, storage medium and computer equipment

Abstract

The invention discloses a photovoltaic user identification method and device, a storage medium and computer equipment, relates to the technical field of user classification, and mainly aims to solve the problems of difficulty in data acquisition and low identification accuracy in an existing photovoltaic user identification method. Comprising the following steps: respectively acquiring historical net loads of each sampling user under different irradiances in a sampling period; constructing a graph data set based on the historical net load; training a classification model by adopting the graph data set to obtain a trained user identification model; obtaining to-be-analyzed net loads of a to-be-identified user under different irradiances, and converting the to-be-analyzed net loads into to-be-analyzed graph data corresponding to the to-be-identified user; and performing classification processing on the to-be-analyzed graph data based on the user identification model to obtain a user type corresponding to the to-be-identified user.

Description

Technical Field

[0001] The present invention relates to the technical field of user classification, and in particular to a photovoltaic user identification method and device, a storage medium, and a computer device. Background Art

[0002] With the severe energy situation in the world and the increasing awareness of environmental protection, the proportion of photovoltaic power generation in the total power generation in the world and my country continues to increase, and the application of distributed photovoltaics is gradually becoming popular. Since the net load data that can be read on the user's meter side is the coupling result of the user's load consumption minus the photovoltaic output, the access of distributed photovoltaics will cause a significant change in the user's load curve shape and affect the estimation of the user's demand response baseline load. Therefore, it is necessary to identify whether the user has installed distributed photovoltaics, and then take further estimation measures for the baseline of such users.

[0003] Existing photovoltaic user identification methods can be classified by identifying photovoltaic physical devices in satellite images and extracting features that can reflect different types of users based on historical load data. The first method of identifying through physical devices requires additional satellite image data, which is not only affected by environmental factors, but also has the problem of data sampling difficulties. The second method of extracting user features is easier to obtain data than the first method, but still has the problem of insufficient recognition accuracy of photovoltaic users. Summary of the invention

[0004] In view of this, the present invention provides a photovoltaic user identification method and device, storage medium, and computer equipment, the main purpose of which is to solve the problems of difficult data collection and low identification accuracy in existing photovoltaic user identification methods.

[0005] According to one aspect of the present invention, a method for identifying a photovoltaic user is provided, comprising:

[0006] Obtain the historical net load of each sampled user under different irradiances during the sampling period respectively; and construct a graph data set based on the historical net load;

[0007] Using the graph data set to train a classification model to obtain a trained user identification model;

[0008] Obtaining the net load to be analyzed of the user to be identified under different irradiances, and converting the net load to be analyzed into graph data to be analyzed corresponding to the user to be identified;

[0009] The graph data to be analyzed is classified based on the user identification model to obtain a user type corresponding to the user to be identified.

[0010] Furthermore, the sampled users are divided into photovoltaic users and non-photovoltaic users; and the historical net load of each sampled user under different irradiances within the sampling period is obtained respectively, including:

[0011] Obtain the historical net load of the photovoltaic user under different irradiances during the sampling period;

[0012] The historical net load of the non-photovoltaic user under different irradiances during the sampling period is obtained.

[0013] Furthermore, the constructing of the graph data set based on the historical net load includes:

[0014] Draw a net load curve graph under different irradiances according to the sequence of sampling times within the sampling period and the historical net load;

[0015] Bundling the net load curve graphs of the same sampled user under different irradiances to obtain graph data corresponding to each sampled user;

[0016] The graph data corresponding to all sampled users are aggregated to obtain the graph data set.

[0017] Furthermore, the classification model is any one of a neural network model and a variant model of a neural network.

[0018] Furthermore, when the classification model is a convolutional neural network model, the classification model is trained using the graph data set to obtain a trained user identification model, including:

[0019] Using the convolutional layer in the convolutional neural network model to perform feature extraction processing on the graph data corresponding to each of the sampled users, so as to obtain image features corresponding to each of the sampled users;

[0020] Using the pooling layer in the convolutional neural network model to perform dimensionality reduction processing on the image features to obtain a graph feature pooling result;

[0021] The fully connected layer in the convolutional neural network model is used to perform classification and regression operations on the graph feature pooling results to obtain the trained user identification model.

[0022] Furthermore, the fully connected layer in the convolutional neural network model is used to perform classification and regression operations on the graph feature pooling result to obtain the trained user identification model, including:

[0023] Using the softmax activation function in the fully linked layer to obtain probability values ​​of the graph data belonging to each category; and determining the predicted user category corresponding to each of the graph data based on the probability values;

[0024] Calculate the loss value based on the actual user category of each sampled user and the predicted user category;

[0025] The parameters in the convolutional neural network model are adjusted based on the loss value to obtain the user identification model.

[0026] Furthermore, the converting the to-be-analyzed payload into to-be-analyzed graph data corresponding to the to-be-identified user includes:

[0027] Draw a net load curve graph to be analyzed under different irradiances according to the sequence of sampling times within the sampling period and the net load to be analyzed;

[0028] The net load curve graphs to be analyzed under different irradiances of the user to be identified are bundled to obtain the graph data to be analyzed corresponding to the user to be identified.

[0029] According to another aspect of the present invention, a photovoltaic user identification device is provided, comprising:

[0030] A construction module, used to respectively obtain the historical net load of each sampled user under different irradiances within a sampling period; and construct a graph data set based on the historical net load;

[0031] A training module, used to train a classification model using the graph data set to obtain a trained user identification model;

[0032] An acquisition and conversion module, used for acquiring the net load to be analyzed of the user to be identified under different irradiances, and converting the net load to be analyzed into the graph data to be analyzed corresponding to the user to be identified;

[0033] The classification and identification module is used to classify the graph data to be analyzed based on the user identification model to obtain a user type corresponding to the user to be identified.

[0034] Furthermore, the sampled users are divided into photovoltaic users and non-photovoltaic users; the construction module includes an acquisition unit and a construction unit; wherein the acquisition unit is used to:

[0035] Obtain the historical net load of the photovoltaic user under different irradiances during the sampling period;

[0036] The historical net load of the non-photovoltaic user under different irradiances during the sampling period is obtained.

[0037] Among them, the construction unit is used for:

[0038] Draw a net load curve graph under different irradiances according to the sequence of sampling times within the sampling period and the historical net load;

[0039] Bundling the net load curve graphs of the same sampled user under different irradiances to obtain graph data corresponding to each sampled user;

[0040] The graph data corresponding to all sampled users are aggregated to obtain the graph data set.

[0041] Furthermore, the classification model in the device is any one of a neural network model and a variant model of a neural network. When the classification model is a convolutional neural network model, the training module is used to:

[0042] Using the convolutional layer in the convolutional neural network model to perform feature extraction processing on the graph data corresponding to each of the sampled users, so as to obtain image features corresponding to each of the sampled users;

[0043] Using the pooling layer in the convolutional neural network model to perform dimensionality reduction processing on the image features to obtain a graph feature pooling result;

[0044] The fully connected layer in the convolutional neural network model is used to perform classification and regression operations on the graph feature pooling results to obtain the trained user identification model.

[0045] Furthermore, the training module is also used for:

[0046] Using the softmax activation function in the fully linked layer to obtain probability values ​​of the graph data belonging to each category; and determining the predicted user category corresponding to each of the graph data based on the probability values;

[0047] Calculate the loss value based on the actual user category of each sampled user and the predicted user category;

[0048] The parameters in the convolutional neural network model are adjusted based on the loss value to obtain the user identification model.

[0049] Furthermore, the acquisition and conversion module is also used for:

[0050] Draw a net load curve graph to be analyzed under different irradiances according to the sequence of sampling times within the sampling period and the net load to be analyzed;

[0051] The net load curve graphs to be analyzed under different irradiances of the user to be identified are bundled to obtain the graph data to be analyzed corresponding to the user to be identified.

[0052] According to another aspect of the present invention, a storage medium is provided, wherein the storage medium stores at least one executable instruction, and the executable instruction enables a processor to execute operations corresponding to the above-mentioned photovoltaic user identification method.

[0053] According to another aspect of the present invention, there is provided a computer device, comprising a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus;

[0054] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the above-mentioned photovoltaic user identification method.

[0055] By means of the above technical solution, the technical solution provided by the embodiment of the present invention has at least the following advantages:

[0056] The present invention provides a method and device for identifying photovoltaic users, a storage medium, and a computer device. Compared with the prior art, the present invention obtains the historical net load of each sampled user under different irradiances during a sampling period; constructs a graph data set based on the historical net load; uses the graph data set to train a classification model to obtain a trained user identification model; obtains the net load to be analyzed of the user to be identified under different irradiances, and converts the net load to be analyzed into graph data to be analyzed corresponding to the user to be identified; classifies the graph data to be analyzed based on the user identification model to obtain the user type corresponding to the user to be identified, thereby achieving accurate identification of photovoltaic users and non-photovoltaic users. The present invention plots the sampling data in the form of a curve graph, and then uses image classification technology based on the curve graph to classify and identify photovoltaic user types and non-photovoltaic user types, thereby solving the problems of difficult data collection and low identification accuracy in existing photovoltaic user identification methods.

[0057] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0059] Figure 1 A schematic diagram of a process of identifying a photovoltaic user provided by an embodiment of the present invention is shown;

[0060] Figure 2 A schematic diagram showing the difference in net load between photovoltaic users and non-photovoltaic users provided by an embodiment of the present invention is shown;

[0061] Figure 3 A schematic diagram showing differences in net loads of photovoltaic users under different irradiances provided by an embodiment of the present invention is shown;

[0062] Figure 4 A schematic flow chart of another photovoltaic user identification method provided by an embodiment of the present invention is shown;

[0063] Figure 5 A schematic diagram showing a flow chart of another photovoltaic user identification method provided by an embodiment of the present invention;

[0064] Figure 6 A schematic diagram of a flow chart of another photovoltaic user identification method provided by an embodiment of the present invention is shown;

[0065] Figure 7 A schematic diagram of the structure of a photovoltaic user identification device provided by an embodiment of the present invention is shown;

[0066] Figure 8 A schematic structural diagram of a computer device provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0067] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0068] The embodiment of the present invention provides a method for identifying photovoltaic users. Figure 1 As shown, the method includes:

[0069] 101. Obtain the historical net load of each sampled user under different irradiances during a sampling period respectively; and construct a graph data set based on the historical net load;

[0070] In the embodiment of the present invention, the current execution end obtains the historical net load of each sampled user under different irradiances during the sampling period. The sampling period is generally one day, or can be adjusted to a time period when the user's net load changes greatly as needed, such as from 5:00 to 23:00 in a day, etc., which is not specifically limited in the embodiment of the present invention. The sampled users are real electricity users, which can be divided into two categories, including photovoltaic users with photovoltaic devices installed and non-photovoltaic users without photovoltaic devices installed; the historical net load of each sampled user under different irradiances during the sampling period is obtained respectively, including: obtaining the historical net load of photovoltaic users under different irradiances during the sampling period; and obtaining the historical net load of non-photovoltaic users under different irradiances during the sampling period. The net load is obtained through the measurement meter data on the user side. For non-photovoltaic users, the net load is not affected by the photovoltaic output, and its net load is equal to the load consumption; for photovoltaic users, the net load is affected by the photovoltaic output, and the specific impact is shown in the following formula:

[0071] N t =L t -G t

[0072] Among them, N t is the net load of photovoltaic users at time t, L t is the load consumption of photovoltaic users at time t, G t is the photovoltaic output of the photovoltaic user at time t. Therefore, the net load of the photovoltaic user and the non-photovoltaic user differs due to the existence of photovoltaic output, such as Figure 2 As shown. Figure 2 It can be seen that during the daytime, the photovoltaic output of photovoltaic users increases, and the net load is affected by the photovoltaic output, and the net load curve of photovoltaic users presents an obvious concave shape. In addition, when users are allowed to reverse power to the grid side, if the photovoltaic output is greater than the load consumption, the net load of photovoltaic users will be negative.

[0073] It should be noted that, based on the above analysis, although there are differences in the shape of the net load curve between photovoltaic users and non-photovoltaic users, due to the different electricity consumption behaviors of different users, it is difficult to extract the characteristics that can fully represent the two types of users with a single curve. In order to more comprehensively identify whether the user is equipped with a photovoltaic device, it is necessary to further analyze the impact of photovoltaic output on the net load curve. Photovoltaic output is affected by many factors. Under different weather conditions, such as sunny and rainy days, there are obvious differences in photovoltaic output power. The size of solar irradiance can characterize different weather conditions to a certain extent, and has a significant impact on photovoltaic output, so solar irradiance is selected as a reference factor. Figure 3The load curves of a photovoltaic user who has installed a photovoltaic device under two different solar irradiances show that the photovoltaic output power is larger under higher solar irradiance, and the net load curve is more concave. For non-photovoltaic users who do not have photovoltaic devices installed, since the net load does not include photovoltaic output components, the difference between their net load curves under different solar irradiances is smaller than that of photovoltaic users. By comparing the difference in net load curves under different solar irradiances, it is possible to identify whether a user has installed a photovoltaic device.

[0074] In the embodiment of the present invention, the current execution end reflects the difference between the two types of users through multiple net load curves under different solar irradiance conditions for each user. Therefore, it is necessary to obtain the historical net load of each user under different irradiances and construct a graph data set based on the historical net load.

[0075] 102. Using the graph data set to train a classification model to obtain a trained user identification model;

[0076] In the embodiment of the present invention, the current execution end uses the graph data set constructed in step 101 to train the classification model to obtain a trained user identification model. The classification model is any one of a neural network model and a variant model of a neural network. The variant model of the neural network includes a convolutional neural network model, a recurrent neural network model, a residual neural network model, etc., which is not specifically limited in the embodiment of the present invention.

[0077] 103. Obtain the net load to be analyzed of the user to be identified under different irradiances, and convert the net load to be analyzed into graph data to be analyzed corresponding to the user to be identified;

[0078] In an embodiment of the present invention, the current execution end obtains the net load to be analyzed of the user to be identified under different irradiances, and uses the same method as constructing the graph data set in step 101 to convert the obtained multiple groups of net loads to be analyzed into graph data to be analyzed corresponding to the user to be identified.

[0079] 104. Classify the graph data to be analyzed based on the user identification model to obtain a user type corresponding to the user to be identified.

[0080] In the embodiment of the present invention, the current execution end classifies the image data to be analyzed obtained in step 103 based on the user identification model trained in step 102 to obtain the user type corresponding to the user to be identified, that is, determines whether the user to be identified is a photovoltaic user or a non-photovoltaic user.

[0081] Furthermore, as a refinement and extension of the specific implementation of the above embodiment, in order to increase the feature difference between the sampled users and to use images to represent the feature difference between the sampled users, another photovoltaic user identification method is provided, such as Figure 4 As shown, the steps are based on the historical net load construction diagram data set, including:

[0082] 201. Draw a net load curve graph under different irradiances according to the sequence of sampling times within the sampling period and the historical net load;

[0083] In the embodiment of the present invention, the current execution end draws a net load curve according to the sequence of sampling times within the adoption cycle and the historical net load, such as Figure 2 As shown, a net load curve is drawn according to the sequence of 24 hours a day. For a single sampling user, if the set irradiance conditions are n, then n net load curves are drawn accordingly; if the number of sampling users is k, then a total of k*n net load curves need to be drawn, which is not specifically limited in the embodiment of the present invention.

[0084] 202. Bundle the net load curve graphs of the same sampled user under different irradiances to obtain graph data corresponding to each sampled user;

[0085] In an embodiment of the present invention, the current execution end bundles the net load curve graphs of the same user under different irradiances, such as bundling n net load curve graphs under n irradiance conditions of sampled user A; bundling n net load curve graphs under n irradiance conditions of sampled user B; bundling n net load curve graphs under n irradiance conditions of sampled user C, etc., which are not specifically limited in the embodiment of the present invention. Among them, the bundling method of the net load curve graphs can be stored in a separate folder, the pictures can be spliced ​​in a certain order, and they can be bundled by naming the pictures, etc., which are not specifically limited in the embodiment of the present invention. After bundling, the graph data corresponding to each sampled user is obtained.

[0086] 203. Aggregate the graph data corresponding to all sampled users to obtain the graph data set.

[0087] In the embodiment of the present invention, the current execution end aggregates the graph data corresponding to all sampled users obtained in step 202 to obtain a graph data set for model training.

[0088] It should be noted that in order to make the feature difference between photovoltaic users and non-photovoltaic users in the graph data set greater, the span of irradiance can be increased. In addition, the number of irradiance conditions can be increased as needed, that is, the number of net load curve graphs corresponding to a single sampled user is increased, which is not specifically limited in the embodiment of the present invention.

[0089] Furthermore, as a refinement and extension of the specific implementation of the above embodiment, in order to make the obtained user identification model have a higher identification accuracy, another photovoltaic user identification method is provided, such as Figure 5 As shown, when the classification model is a convolutional neural network model, the step uses the graph data set to train the classification model to obtain a trained user identification model, including:

[0090] 301. Perform feature extraction processing on the graph data corresponding to each of the sampled users using the convolutional layer in the convolutional neural network model to obtain image features corresponding to each of the sampled users;

[0091] In an embodiment of the present invention, the convolutional neural network model (CNN) performs well for image processing problems. Based on convolution and pooling operations, the image features can be extracted and classified through the combination of multi-layer neural networks. Its structure is mainly composed of three layers: convolution layer, pooling layer and fully connected layer. The current execution end uses the convolution layer to perform feature extraction processing on the graph data corresponding to each of the sampled users to obtain image features corresponding to each sampled user. The convolution layer uses a convolution kernel to perform convolution operations with the image to extract different features of the image. The convolution kernel slides on the input image matrix according to the set step size. After each slide, the corresponding positions of the two matrices are multiplied and summed, and the final output image features are obtained after the convolution operation. The convolution operation with a step size of 1 is shown in the following formula:

[0092]

[0093] Where f(x, y) is the value of the output image feature (x, y) bit position; I is the input image matrix; K is the n-order convolution kernel. By using zero padding on the input image, each input point can be used as the center of each convolution, and the convolution kernel can be prevented from sliding exactly to the edge of the input image due to the set step size. After the convolution operation, an activation function is added to achieve nonlinear mapping to enhance the extraction of complex features.

[0094] 302. Perform dimensionality reduction processing on the image features using the pooling layer in the convolutional neural network model to obtain a graph feature pooling result;

[0095] In the embodiment of the present invention, the current execution end uses the pooling layer in the convolutional neural network model to perform dimensionality reduction processing on the image features to obtain the image feature pooling result. Dimensionality reduction processing can reduce the parameter scale while retaining important features. Common pooling operations include average pooling and maximum pooling, which divide the input feature map into rectangular areas of the same size as the pooling kernel and output the average or maximum value of each area.

[0096] 303. Use the fully connected layer in the convolutional neural network model to perform classification and regression operations on the graph feature pooling result to obtain the trained user recognition model.

[0097] In an embodiment of the present invention, the current execution end uses the full-link layer in the convolutional neural network model to perform classification and regression operations on the graph feature pooling results, specifically including: using the softmax activation function in the full-link layer to obtain the probability value that the graph data belongs to each category; and determining the predicted user category corresponding to each graph data based on the probability value; calculating the loss value based on the real user category of each sampled user and the predicted user category obtained by the full-link layer; for example, using the cross entropy loss function to calculate the loss value after classification, which is not specifically limited in the embodiment of the present invention. Finally, the parameters in the convolutional neural network model are adjusted based on the loss value to obtain the user identification model.

[0098] It should be noted that the current execution end can use the Adaptive Moment Estimation (Adam) algorithm when adjusting parameters. This algorithm can dynamically adjust the learning rate of each parameter based on the first-order moment and second-order moment of the gradient. The update formula of the model parameters is as follows:

[0099]

[0100] Among them, θ t is the parameter of time step t, η is the initial learning rate, and is the unbiased corrected first-order and second-order moment estimate of the gradient, and ∈ is a small number that avoids the denominator being zero.

[0101] Furthermore, as a refinement and extension of the specific implementation of the above embodiment, in order to make the data obtained from the user to be identified suitable for the trained user identification model, another photovoltaic user identification method is provided, such as Figure 6 As shown, the step of converting the net load to be analyzed into graph data to be analyzed corresponding to the user to be identified includes:

[0102] 401. Draw a net load curve graph to be analyzed under different irradiances according to the sequence of sampling times in the sampling period and the net load to be analyzed;

[0103] In the embodiment of the present invention, the current execution end processes the net load to be analyzed based on the same sampling period in the graph data set. First, as shown in step 201, a net load curve graph to be analyzed under different irradiances is drawn according to the sequence of sampling times in the sampling period and the net load to be analyzed; if the set irradiance conditions are n, then n net load curve graphs to be analyzed are drawn accordingly, which is not specifically limited in the embodiment of the present invention.

[0104] 402. Bundle the net load curve graphs to be analyzed under different irradiances of the user to be identified to obtain the graph data to be analyzed corresponding to the user to be identified.

[0105] In the embodiment of the present invention, the current execution end bundles the net load curve graphs to be analyzed under different irradiances of the user to be identified obtained in step 401. The specific bundling method is the same as the bundling method selected in step 202. A separate folder can be used for storage, and the pictures can be spliced ​​in a certain order. The pictures can also be bundled by naming them, etc. The embodiment of the present invention does not make specific restrictions. After bundling, the to-be-analyzed graph data corresponding to the user to be identified is obtained.

[0106] In the embodiment of the present invention, the current execution end performs example calculations, and the example results show that the average values ​​of various indicators such as the model's precision and recall rate are all over 98%, the highest value of each indicator can reach 100%, and the minimum value exceeds 94%, which verifies the reliability of the photovoltaic user identification method of the present invention.

[0107] The embodiment of the present invention provides a method for identifying photovoltaic users. Compared with the prior art, the present invention obtains the historical net load of each sampled user under different irradiances during the sampling period; constructs a graph data set based on the historical net load; uses the graph data set to train a classification model to obtain a trained user identification model; obtains the net load to be analyzed of the user to be identified under different irradiances, and converts the net load to be analyzed into graph data to be analyzed corresponding to the user to be identified; classifies the graph data to be analyzed based on the user identification model to obtain the user type corresponding to the user to be identified, thereby achieving accurate identification of photovoltaic users and non-photovoltaic users. The present invention plots the sampled data in the form of a curve graph, and then uses image classification technology based on the curve graph to classify and identify photovoltaic user types and non-photovoltaic user types, thereby solving the problems of difficult data collection and low identification accuracy in existing photovoltaic user identification methods.

[0108] As the above Figure 1 The embodiment of the present invention provides a photovoltaic user identification device, such as Figure 7 As shown, the device comprises:

[0109] A construction module 51 is used to respectively obtain the historical net load of each sampled user under different irradiances within a sampling period; and construct a graph data set based on the historical net load;

[0110] A training module 52, used to train a classification model using the graph data set to obtain a trained user identification model;

[0111] An acquisition and conversion module 53 is used to acquire the net load to be analyzed of the user to be identified under different irradiances, and convert the net load to be analyzed into the graph data to be analyzed corresponding to the user to be identified;

[0112] The classification and identification module 54 is used to classify the graph data to be analyzed based on the user identification model to obtain the user type corresponding to the user to be identified.

[0113] Furthermore, the sampled users are divided into photovoltaic users and non-photovoltaic users; the construction module 51 includes an acquisition unit and a construction unit; wherein the acquisition unit is used to:

[0114] Obtain the historical net load of the photovoltaic user under different irradiances during the sampling period;

[0115] The historical net load of the non-photovoltaic user under different irradiances during the sampling period is obtained.

[0116] Among them, the construction unit is used for:

[0117] Draw a net load curve graph under different irradiances according to the sequence of sampling times within the sampling period and the historical net load;

[0118] Bundling the net load curve graphs of the same sampled user under different irradiances to obtain graph data corresponding to each sampled user;

[0119] The graph data corresponding to all sampled users are aggregated to obtain the graph data set.

[0120] Furthermore, the classification model in the device is any one of a neural network model and a variant model of a neural network. When the classification model is a convolutional neural network model, the training module 52 is used to:

[0121] Using the convolutional layer in the convolutional neural network model to perform feature extraction processing on the graph data corresponding to each of the sampled users, so as to obtain image features corresponding to each of the sampled users;

[0122] Using the pooling layer in the convolutional neural network model to perform dimensionality reduction processing on the image features to obtain a graph feature pooling result;

[0123] The fully connected layer in the convolutional neural network model is used to perform classification and regression operations on the graph feature pooling results to obtain the trained user identification model.

[0124] Furthermore, the training module 52 is also used for:

[0125] Using the softmax activation function in the fully linked layer to obtain probability values ​​of the graph data belonging to each category; and determining the predicted user category corresponding to each of the graph data based on the probability values;

[0126] Calculate the loss value based on the actual user category of each sampled user and the predicted user category;

[0127] The parameters in the convolutional neural network model are adjusted based on the loss value to obtain the user identification model.

[0128] Furthermore, the acquisition and conversion module 53 is also used for:

[0129] Draw a net load curve graph to be analyzed under different irradiances according to the sequence of sampling times within the sampling period and the net load to be analyzed;

[0130] The net load curve graphs to be analyzed under different irradiances of the user to be identified are bundled to obtain the graph data to be analyzed corresponding to the user to be identified.

[0131] The embodiment of the present invention provides a photovoltaic user identification device. Compared with the prior art, the present invention obtains the historical net load of each sampled user under different irradiances during the sampling period; constructs a graph data set based on the historical net load; uses the graph data set to train a classification model to obtain a trained user identification model; obtains the net load to be analyzed of the user to be identified under different irradiances, and converts the net load to be analyzed into graph data to be analyzed corresponding to the user to be identified; classifies the graph data to be analyzed based on the user identification model to obtain the user type corresponding to the user to be identified, thereby achieving accurate identification of photovoltaic users and non-photovoltaic users. The present invention plots the sampling data in the form of a curve graph, and then uses image classification technology based on the curve graph to classify and identify photovoltaic user types and non-photovoltaic user types, thereby solving the problems of difficult data collection and low identification accuracy in existing photovoltaic user identification methods.

[0132] According to an embodiment of the present invention, a storage medium is provided, wherein the storage medium stores at least one executable instruction, and the computer executable instruction can execute the photovoltaic user identification method in any of the above method embodiments.

[0133] Figure 8 A schematic diagram of the structure of a computer device provided according to an embodiment of the present invention is shown. The specific embodiment of the present invention does not limit the specific implementation of the computer device.

[0134] like Figure 8As shown, the computer device may include: a processor (processor) 602 , a communication interface (Communications Interface) 604 , a memory (memory) 606 , and a communication bus 608 .

[0135] The processor 602 , the communication interface 604 , and the memory 606 communicate with each other via a communication bus 608 .

[0136] The communication interface 604 is used to communicate with other devices such as clients or other servers.

[0137] The processor 602 is used to execute the program 610, and specifically can execute the relevant steps of the above-mentioned photovoltaic user identification method.

[0138] Specifically, the program 610 may include program codes, which include computer operation instructions.

[0139] The processor 602 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. The one or more processors included in the computer device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0140] The memory 606 is used to store the program 610. The memory 606 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0141] The program 610 may be specifically configured to enable the processor 602 to perform the following operations:

[0142] Obtain the historical net load of each sampled user under different irradiances during the sampling period respectively; and construct a graph data set based on the historical net load;

[0143] Using the graph data set to train a classification model to obtain a trained user identification model;

[0144] Obtaining the net load to be analyzed of the user to be identified under different irradiances, and converting the net load to be analyzed into graph data to be analyzed corresponding to the user to be identified;

[0145] The graph data to be analyzed is classified based on the user identification model to obtain a user type corresponding to the user to be identified.

[0146] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, and optionally, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0147] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for identifying photovoltaic users, characterized in that: include: Obtain the historical net load of each sampled user under different irradiances during the sampling period respectively; and construct a graph data set based on the historical net load; Using the graph data set to train a classification model to obtain a trained user identification model; Obtaining the net load to be analyzed of the user to be identified under different irradiances, and converting the net load to be analyzed into graph data to be analyzed corresponding to the user to be identified; The graph data to be analyzed is classified based on the user identification model to obtain a user type corresponding to the user to be identified.

2. The method according to claim 1, characterized in that The sampled users are divided into photovoltaic users and non-photovoltaic users; and the historical net load of each sampled user under different irradiances within the sampling period is obtained respectively, including: Obtain the historical net load of the photovoltaic user under different irradiances during the sampling period; The historical net load of the non-photovoltaic user under different irradiances during the sampling period is obtained.

3. The method according to claim 1, characterized in that: The constructing a graph data set based on the historical net load includes: Draw a net load curve graph under different irradiances according to the sequence of sampling times within the sampling period and the historical net load; Bundling the net load curve graphs of the same sampled user under different irradiances to obtain graph data corresponding to each sampled user; The graph data corresponding to all sampled users are aggregated to obtain the graph data set.

4. The method according to claim 1, characterized in that: The classification model is any one of a neural network model and a variant model of a neural network.

5. The method according to claim 4, characterized in that When the classification model is a convolutional neural network model, the adopting of the graph data set to train the classification model to obtain a trained user identification model includes: Using the convolutional layer in the convolutional neural network model to perform feature extraction processing on the graph data corresponding to each of the sampled users, so as to obtain image features corresponding to each of the sampled users; Using the pooling layer in the convolutional neural network model to perform dimensionality reduction processing on the image features to obtain a graph feature pooling result; The fully connected layer in the convolutional neural network model is used to perform classification and regression operations on the graph feature pooling results to obtain the trained user identification model.

6. The method according to claim 5, characterized in that The method of using the fully connected layer in the convolutional neural network model to perform classification and regression operations on the graph feature pooling result to obtain the trained user identification model includes: Using the softmax activation function in the fully linked layer to obtain probability values ​​of the graph data belonging to each category; and determining the predicted user category corresponding to each of the graph data based on the probability values; Calculate the loss value based on the actual user category of each sampled user and the predicted user category; The parameters in the convolutional neural network model are adjusted based on the loss value to obtain the user identification model.

7. The method according to any one of claims 1 to 6, characterized in that: The converting the to-be-analyzed payload into to-be-analyzed graph data corresponding to the to-be-identified user comprises: Draw a net load curve graph to be analyzed under different irradiances according to the sequence of sampling times within the sampling period and the net load to be analyzed; The net load curve graphs to be analyzed under different irradiances of the user to be identified are bundled to obtain the graph data to be analyzed corresponding to the user to be identified.

8. A photovoltaic user identification device, characterized in that: include: A construction module, used to respectively obtain the historical net load of each sampled user under different irradiances within a sampling period; and construct a graph data set based on the historical net load; A training module, used to train a classification model using the graph data set to obtain a trained user identification model; An acquisition and conversion module, used for acquiring the net load to be analyzed of the user to be identified under different irradiances, and converting the net load to be analyzed into the graph data to be analyzed corresponding to the user to be identified; The classification and identification module is used to classify the graph data to be analyzed based on the user identification model to obtain a user type corresponding to the user to be identified.

9. A storage medium, wherein at least one executable instruction is stored in the storage medium, and the executable instruction executes an operation corresponding to the photovoltaic user identification method according to any one of claims 1 to 7.

10. A computer device, comprising a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the photovoltaic user identification method according to any one of claims 1-7.

Images