Photovoltaic power prediction model training method and prediction method based on hierarchical clustering
By extracting and clustering the historical cloud map data of the photovoltaic power generation area, and training the photovoltaic power prediction model based on the hierarchical clustering algorithm, the problem of insufficient prediction accuracy under the influence of dynamic changes in the cloud layer in traditional methods is solved, and more efficient photovoltaic power prediction is achieved.
Patent Information
- Application Number
- CN202411981733.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-31
- Publication Date
- 2025-05-02
AI Technical Summary
Traditional photovoltaic power prediction methods are difficult to fully consider the dynamic changes of clouds, resulting in insufficient prediction accuracy.
The photovoltaic power prediction model training method based on hierarchical clustering is used to extract and cluster the historical cloud map data of each photovoltaic power generation area to obtain cloud feature clusters, and the photovoltaic power is predicted based on these clusters.
The prediction accuracy of photovoltaic power can be improved and the dynamic changes of clouds can be considered more effectively, thereby improving the accuracy of prediction.
Smart Images
Figure CN119918730A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of communication technology, and in particular to a photovoltaic power prediction model training method and a prediction method based on hierarchical clustering. Background Art
[0002] With the rapid development of renewable energy, photovoltaic power generation, as an important clean energy, has gradually been widely used in various scenarios. Photovoltaic power prediction is a complex process, which is highly affected by meteorological conditions and needs to consider multiple factors, such as solar irradiance, temperature, humidity, etc. Ground-based cloud map data plays an important role in this regard. Ground-based cloud maps can record sky conditions in real time, and photovoltaic output power is mainly affected by solar irradiance. The generation, movement and ablation of clouds in the sky will make solar irradiance random and volatile. In particular, cloud changes have a significant impact on the accuracy of ultra-short-term power prediction. Traditional photovoltaic power prediction methods mostly rely on historical power generation data and weather forecasts, and often fail to fully consider the dynamic changes of clouds, resulting in insufficient prediction accuracy. Summary of the invention
[0003] The present invention provides a photovoltaic power prediction model training method based on hierarchical clustering, a photovoltaic power prediction method, an electronic device and a storage medium, which can solve at least one of the above technical problems.
[0004] According to one aspect of the present invention, a photovoltaic power prediction model training method based on hierarchical clustering is provided, comprising:
[0005] Extracting features from the cloud map data of each photovoltaic power generation area at each historical time to obtain cloud layer features of each of the cloud map data;
[0006] Based on the hierarchical clustering algorithm, the cloud features corresponding to the cloud image data of each photovoltaic power generation area at each time are clustered to obtain multiple cloud feature clusters;
[0007] For each of the cloud layer feature clusters, determining the corresponding photovoltaic power generation power based on the historical time and photovoltaic power generation area corresponding to each cloud layer feature in the cloud layer feature cluster, so as to obtain a photovoltaic power generation power set corresponding to the cloud layer feature cluster;
[0008] Based on each of the cloud layer feature clusters and the photovoltaic power generation power sets corresponding to each of the cloud layer feature clusters, a photovoltaic power prediction model is trained to obtain a target photovoltaic power prediction model.
[0009] According to another aspect of the present invention, a photovoltaic power prediction method based on hierarchical clustering is provided, comprising:
[0010] Based on the target photovoltaic power prediction model, first cloud map data of the first photovoltaic power generation area at the first time is processed to obtain the predicted photovoltaic power generation power of the first photovoltaic power generation area at the first time;
[0011] The target photovoltaic power prediction model is generated by training according to the photovoltaic power prediction model training method based on hierarchical clustering described in any one of the embodiments of the present disclosure.
[0012] According to another aspect of the present invention, a photovoltaic power prediction model training device based on hierarchical clustering is provided, comprising:
[0013] A cloud feature extraction module is used to extract features from the cloud map data of each photovoltaic power generation area at each historical time to obtain cloud features of each cloud map data;
[0014] The cloud feature clustering module is used to cluster the cloud features corresponding to the cloud image data of each photovoltaic power generation area at each time based on a hierarchical clustering algorithm to obtain multiple cloud feature clusters;
[0015] A power generation acquisition module is used to determine the corresponding photovoltaic power generation power for each of the cloud layer feature clusters based on the historical time and photovoltaic power generation area corresponding to each cloud layer feature in the cloud layer feature cluster, so as to obtain the photovoltaic power generation power set corresponding to the cloud layer feature cluster;
[0016] The prediction model training module is used to train the photovoltaic power prediction model based on each of the cloud layer feature clusters and the photovoltaic power generation power sets corresponding to each of the cloud layer feature clusters to obtain a target photovoltaic power prediction model.
[0017] According to another aspect of the present invention, there is provided an electronic device, comprising:
[0018] at least one processor; and
[0019] a memory communicatively connected to the at least one processor; wherein,
[0020] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any photovoltaic power prediction model training method and photovoltaic power prediction method based on hierarchical clustering in the embodiments of the present invention.
[0021] According to another aspect of the present invention, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute any photovoltaic power prediction model training method and photovoltaic power prediction method based on hierarchical clustering according to the embodiments of the present invention.
[0022] By adopting the technical solution of the present invention, the cloud map data of each photovoltaic power generation area at each historical time is extracted to obtain the cloud layer features of each cloud map data; based on the hierarchical clustering algorithm, the cloud layer features corresponding to the cloud map data of each photovoltaic power generation area at each time are clustered to obtain multiple cloud layer feature clusters; for each of the cloud layer feature clusters, the corresponding photovoltaic power generation power is determined based on the historical time and photovoltaic power generation area corresponding to each cloud layer feature in the cloud layer feature cluster, so as to obtain the photovoltaic power generation power set corresponding to the cloud layer feature cluster; based on each cloud layer feature cluster and the photovoltaic power generation power set corresponding to each cloud layer feature cluster, the photovoltaic power prediction model is trained to obtain the target photovoltaic power prediction model. In this way, the prediction accuracy of photovoltaic power can be improved.
[0023] 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
[0024] The accompanying drawings are used to better understand the present invention and do not constitute a limitation of the present invention.
[0025] Figure 1 is a flow chart of a photovoltaic power prediction model training method based on hierarchical clustering according to an embodiment of the present invention;
[0026] Figure 2 is a flow chart of a photovoltaic power prediction method based on hierarchical clustering according to an embodiment of the present disclosure;
[0027] Figure 3 It is a structural block diagram of a photovoltaic power prediction model training device based on hierarchical clustering according to an embodiment of the present disclosure;
[0028] Figure 4 is a structural block diagram of a photovoltaic power prediction device based on hierarchical clustering according to an embodiment of the present disclosure;
[0029] Figure 5 The block diagram is a block diagram of an electronic device for implementing the method according to the embodiment of the present invention. DETAILED DESCRIPTION
[0030] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be appreciated by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope of the present invention. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.
[0031] Figure 1The present invention is a flowchart of a photovoltaic power prediction model training method based on hierarchical clustering according to an embodiment of the present invention.
[0032] like Figure 1 As shown, the photovoltaic power prediction model training method based on hierarchical clustering may include:
[0033] S110, extracting features from the cloud map data of each photovoltaic power generation area at each historical time to obtain cloud layer features of each cloud map data;
[0034] S120, clustering cloud features corresponding to the cloud image data of each photovoltaic power generation area at each time based on a hierarchical clustering algorithm to obtain a plurality of cloud feature clusters;
[0035] S130, for each cloud feature cluster, determining the corresponding photovoltaic power generation power based on the historical time and photovoltaic power generation area corresponding to each cloud feature in the cloud feature cluster, so as to obtain a photovoltaic power generation power set corresponding to the cloud feature cluster;
[0036] S140, training a photovoltaic power prediction model based on each cloud layer feature cluster and a photovoltaic power generation power set corresponding to each cloud layer feature cluster to obtain a target photovoltaic power prediction model.
[0037] Exemplarily, historical cloud image data of various historical times are obtained from ground meteorological stations in photovoltaic power generation areas. These cloud image data are denoised, enhanced and normalized to facilitate the acquisition of effective cloud features. The cloud image data may include one or more cloud images from different perspectives.
[0038] Exemplarily, image processing technology is used to extract features from cloud image data to obtain cloud layer features of each cloud image data. Cloud layer features may include cloud image pixel features, cloud image texture features, cloud image color features, cloud size information, cloud distribution features, and the like.
[0039] Exemplarily, the image processing technology may include an encoder, a decoder, a gray-level co-occurrence matrix, a color histogram, and an image segmentation edge detection algorithm, etc.
[0040] Exemplarily, the hierarchical clustering algorithm may include an agglomerative hierarchical clustering algorithm, a top-down hierarchical clustering algorithm, a bottom-up hierarchical clustering algorithm, and the like.
[0041] It can be understood that each cloud layer feature cluster corresponds to a photovoltaic power generation set. The cloud layer feature cluster includes multiple cloud layer features, and the photovoltaic power generation set includes multiple photovoltaic power generation powers.
[0042] Exemplarily, a support vector machine algorithm may be used to learn each cloud layer feature cluster and the photovoltaic power generation set corresponding to each cloud layer feature cluster to generate a target photovoltaic power prediction model.
[0043] Exemplarily, an adaptive learning rate may be used to update the model parameters of the photovoltaic power prediction model to increase the update speed of the photovoltaic power prediction model.
[0044] According to the above implementation, feature extraction is performed on the cloud map data of each photovoltaic power generation area at each historical time to obtain the cloud features of each cloud map data; based on the hierarchical clustering algorithm, the cloud features corresponding to the cloud map data of each photovoltaic power generation area at each time are clustered to obtain multiple cloud feature clusters; for each of the cloud feature clusters, the corresponding photovoltaic power generation power is determined based on the historical time and photovoltaic power generation area corresponding to each cloud feature in the cloud feature cluster to obtain the photovoltaic power set corresponding to the cloud feature cluster; based on each cloud feature cluster and the photovoltaic power set corresponding to each cloud feature cluster, the photovoltaic power prediction model is trained to obtain the target photovoltaic power prediction model. In this way, the prediction accuracy of photovoltaic power can be improved.
[0045] In one embodiment, feature extraction is performed on cloud map data of each photovoltaic power generation area at each historical time to obtain cloud layer features of each cloud map data, including: using an encoder to extract features of the cloud map data to obtain cloud map pixel features; using a grayscale co-occurrence matrix to calculate the cloud map pixel features to obtain cloud map texture features; using a color histogram to process the cloud map pixel features to obtain cloud map color features; using an image segmentation edge detection algorithm to process the cloud map pixel features to obtain cloud size information; calculating the cloud map texture features, cloud map color features and cloud size information to obtain cloud distribution features; and determining the cloud layer features corresponding to the cloud map data based on the cloud map texture features, cloud map color features, cloud size information and cloud distribution features.
[0046] It can be understood that while updating the model parameters of the photovoltaic power prediction model, the model parameters of the encoder, grayscale co-occurrence matrix, color histogram and image segmentation edge detection algorithm are also updated to obtain the target encoder, target grayscale co-occurrence matrix, target color histogram and target image segmentation edge detection algorithm.
[0047] Exemplarily, the cloud image texture features may include texture indicators such as contrast and entropy of the cloud image.
[0048] Exemplarily, the cloud image color feature may include the color distribution of the cloud image.
[0049] Exemplarily, the cloud size information may include the area and perimeter of each cloud in the cloud image.
[0050] Exemplarily, the cloud cluster distribution characteristics may include the position and relative distribution position of each cloud cluster in the cloud map.
[0051] Exemplarily, cloud layer features include cloud texture features, cloud color features, cloud size information, and cloud distribution features.
[0052] According to the above implementation, a variety of image processing technologies can be used to process the cloud image pixel features respectively to obtain a plurality of different cloud layer features.
[0053] In one embodiment, a photovoltaic power prediction model is trained based on each cloud feature cluster and the photovoltaic power generation power set corresponding to each cloud feature cluster to obtain a target photovoltaic power prediction model, including: determining the labeled photovoltaic power generation power corresponding to the cloud feature cluster based on the photovoltaic power generation power set corresponding to the cloud feature cluster; analyzing the cloud feature cluster based on the photovoltaic power prediction model to obtain the predicted photovoltaic power generation power; determining the predicted loss based on the labeled photovoltaic power generation power and the predicted photovoltaic power generation power; and adjusting the parameters of the photovoltaic power prediction model based on the predicted loss to obtain the target photovoltaic power prediction model.
[0054] Exemplarily, the photovoltaic power generation sets are clustered, and the cluster centers are used as labeled photovoltaic power generation.
[0055] Exemplarily, the photovoltaic power generation powers in the photovoltaic power generation power set are averaged to obtain the labeled photovoltaic power generation power.
[0056] Exemplarily, the difference between the marked photovoltaic power generation and the predicted photovoltaic power generation is calculated, and the difference is calculated using an error algorithm to obtain the predicted loss.
[0057] Exemplarily, based on the gradient information corresponding to the predicted loss, the parameters of the photovoltaic power prediction model are adjusted.
[0058] It can be understood that the above operations can be performed multiple times to obtain a target photovoltaic power prediction model.
[0059] According to the above implementation, the photovoltaic power prediction model is trained based on each cloud layer feature cluster and the photovoltaic power generation power set corresponding to each cloud layer feature cluster, so as to obtain a target photovoltaic power prediction model.
[0060] In one embodiment, based on the predicted loss, the parameters of the photovoltaic power prediction model are adjusted, including: using a decoder to reconstruct the cluster center of the cloud feature cluster to obtain a reconstructed cloud map; determining the reconstruction loss based on the cloud map data corresponding to the cluster center of the cloud feature cluster and the reconstructed cloud map; determining the comprehensive loss based on the predicted loss and the reconstruction loss; when the comprehensive loss does not meet the preset loss condition, adjusting the model parameters of the photovoltaic power prediction model based on the gradient information corresponding to the comprehensive loss; when the comprehensive loss meets the preset loss condition, determining the target photovoltaic power prediction model based on the photovoltaic power prediction model.
[0061] Exemplarily, while updating the model parameters of the photovoltaic power prediction model, the model parameters of the decoder are also updated.
[0062] Exemplarily, the reconstruction loss is determined based on the cloud image data and the distance between the reconstructed cloud images.
[0063] Exemplarily, a weighted sum of the prediction loss and the reconstruction loss is performed to obtain a comprehensive loss;
[0064] Exemplarily, the loss condition may be that the comprehensive loss is less than a preset loss threshold.
[0065] Exemplarily, after adjusting the model parameters of the photovoltaic power prediction model, based on the adjusted model parameters of the photovoltaic power prediction model, the above steps are returned to be executed to adjust the model parameters of the photovoltaic power prediction model.
[0066] According to the above implementation, the comprehensive loss consisting of the prediction loss and the reconstruction loss can be used to adjust the model parameters of the photovoltaic power prediction model, so as to improve the model accuracy of the photovoltaic power prediction model.
[0067] In one embodiment, it also includes: adjusting the model parameters in the encoder, gray-level co-occurrence matrix, color histogram, image segmentation edge detection algorithm and decoder based on the gradient information corresponding to the comprehensive loss.
[0068] According to the above implementation, while adjusting the model parameters of the photovoltaic power prediction model, the model parameters in the encoder, gray level co-occurrence matrix, color histogram, image segmentation edge detection algorithm and decoder are adjusted. In this way, the accuracy of the prediction model can be improved.
[0069] In one embodiment, based on a hierarchical clustering algorithm, cloud features corresponding to cloud map data of each photovoltaic power generation area at each time are clustered to obtain multiple cloud feature clusters, including: using Euclidean distance to calculate the distance between any two cloud features in each cloud feature; using an agglomerative hierarchical clustering algorithm, based on the distance between any two cloud features in each cloud feature, clustering each cloud feature to obtain multiple cloud feature clusters.
[0070] According to the above implementation, after extracting the cloud features, a hierarchical clustering algorithm is used to perform cluster analysis on the cloud features of different times and regions, so as to facilitate subsequent machine learning and improve the training accuracy of the model.
[0071] Figure 2 The present invention discloses a flow chart of a photovoltaic power prediction method based on hierarchical clustering according to an embodiment of the present invention.
[0072] like Figure 2 As shown, the prediction method includes:
[0073] S210, obtaining first cloud map data of a first photovoltaic power generation area at a first time;
[0074] S220, using a target encoder to perform feature extraction on the first cloud image data to obtain first cloud image pixel features;
[0075] S230, calculating the pixel features of the first cloud image using the target gray-level co-occurrence matrix to obtain the texture features of the first cloud image;
[0076] S240, using the target color histogram to process the first cloud image pixel feature to obtain a first cloud image color feature;
[0077] S250, processing the pixel features of the first cloud image using a target image segmentation edge detection algorithm to obtain first cloud size information;
[0078] S260, calculating the first cloud image texture feature, the first cloud image color feature, and the first cloud cluster size information to obtain the first cloud cluster distribution feature;
[0079] S270, determining a first cloud layer feature corresponding to the first cloud image data based on the first cloud image texture feature, the first cloud image color feature, the first cloud cluster size information, and the first cloud cluster distribution feature;
[0080] S280, based on the target photovoltaic power prediction model, processing the first cloud layer feature to obtain the predicted photovoltaic power generation power of the first photovoltaic power generation area at the first time.
[0081] Among them, for the training process of the above models and algorithms, reference can be made to the aforementioned embodiments and will not be described in detail here.
[0082] Figure 3 It is a structural block diagram of a photovoltaic power prediction model training device based on hierarchical clustering according to an embodiment of the present disclosure.
[0083] like Figure 3 As shown, the photovoltaic power prediction model training device based on hierarchical clustering includes:
[0084] The cloud layer feature extraction module 310 is used to extract features from the cloud image data of each photovoltaic power generation area at each historical time to obtain cloud layer features of each cloud image data;
[0085] The cloud feature clustering module 320 is used to cluster the cloud features corresponding to the cloud image data of each photovoltaic power generation area at each time based on a hierarchical clustering algorithm to obtain multiple cloud feature clusters;
[0086] The power generation acquisition module 330 is used to determine the corresponding photovoltaic power generation power for each of the cloud layer feature clusters based on the historical time and photovoltaic power generation area corresponding to each cloud layer feature in the cloud layer feature cluster, so as to obtain the photovoltaic power generation power set corresponding to the cloud layer feature cluster;
[0087] The prediction model training module 340 is used to train the photovoltaic power prediction model based on each of the cloud layer feature clusters and the photovoltaic power generation power sets corresponding to each of the cloud layer feature clusters to obtain a target photovoltaic power prediction model.
[0088] In one embodiment, the cloud feature extraction module 310 includes:
[0089] A feature encoding unit, used to extract features from the cloud image data using an encoder to obtain cloud image pixel features;
[0090] A texture feature extraction unit, used to calculate the cloud image pixel features using a gray level co-occurrence matrix to obtain cloud image texture features;
[0091] A color feature extraction unit, used to process the cloud image pixel features using a color histogram to obtain cloud image color features;
[0092] A size information determination unit, used to process the pixel features of the cloud image using an image segmentation edge detection algorithm to obtain cloud size information;
[0093] a distribution feature determination unit, configured to calculate the cloud image texture feature, the cloud image color feature and the cloud size information to obtain a cloud distribution feature; and
[0094] A cloud layer feature determination unit is used to determine the cloud layer features corresponding to the cloud image data based on the cloud image texture features, the cloud image color features, the cloud size information and the cloud distribution features.
[0095] In one embodiment, the prediction model training module 340 includes:
[0096] A marked power acquisition unit, used to determine the marked photovoltaic power generation power corresponding to the cloud layer feature cluster based on the photovoltaic power generation power set corresponding to the cloud layer feature cluster;
[0097] A power prediction unit, configured to analyze the cloud layer feature cluster based on the photovoltaic power prediction model to obtain predicted photovoltaic power generation power;
[0098] A predicted loss unit, configured to determine a predicted loss based on the marked photovoltaic power generation and the predicted photovoltaic power generation;
[0099] The first parameter adjustment unit is used to adjust the parameters of the photovoltaic power prediction model based on the predicted loss to obtain the target photovoltaic power prediction model.
[0100] In one implementation, the first parameter adjustment unit is specifically configured to:
[0101] Using a decoder, reconstructing the cluster center of the cloud feature cluster to obtain a reconstructed cloud image;
[0102] Determining a reconstruction loss based on the cloud image data corresponding to the cluster center of the cloud feature cluster and the reconstructed cloud image;
[0103] determining a comprehensive loss based on the predicted loss and the reconstructed loss;
[0104] When the comprehensive loss does not meet the preset loss condition, adjusting the model parameters of the photovoltaic power prediction model based on the gradient information corresponding to the comprehensive loss;
[0105] When the comprehensive loss meets the preset loss condition, the target photovoltaic power prediction model is determined based on the photovoltaic power prediction model.
[0106] In one embodiment, the prediction model training module 340 further includes:
[0107] The second parameter adjustment unit is used to adjust the model parameters in the encoder, the gray-level co-occurrence matrix, the color histogram, the image segmentation edge detection algorithm and the decoder based on the gradient information corresponding to the comprehensive loss.
[0108] In one embodiment, the cloud feature clustering module 320 includes:
[0109] A distance calculation unit, used for calculating the distance between any two cloud layer features among the cloud layer features by using Euclidean distance;
[0110] The feature clustering unit is used to cluster each of the cloud layer features based on the distance between any two of the cloud layer features by using an agglomerative hierarchical clustering algorithm to obtain a plurality of cloud layer feature clusters.
[0111] Figure 4 It is a structural block diagram of a photovoltaic power prediction device based on hierarchical clustering according to an embodiment of the present disclosure.
[0112] like Figure 4 As shown, the prediction device comprises:
[0113] The target cloud map acquisition module 410 is used to acquire first cloud map data of a first photovoltaic power generation area at a first time;
[0114] A first feature determination module 420 is used to extract features from the first cloud image data using a target encoder to obtain first cloud image pixel features;
[0115] The second feature determination module 430 is used to calculate the first cloud image pixel feature using the target gray level co-occurrence matrix to obtain the first cloud image texture feature;
[0116] The third feature determination module 440 is used to process the first cloud image pixel feature using the target color histogram to obtain the first cloud image color feature;
[0117] The fourth feature determination module 450 is used to process the pixel features of the first cloud image by using a target image segmentation edge detection algorithm to obtain the first cloud size information;
[0118] A fifth feature determination module 460 is used to calculate the first cloud image texture feature, the first cloud image color feature and the first cloud cluster size information to obtain the first cloud cluster distribution feature;
[0119] A sixth feature determination module 470, configured to determine a first cloud layer feature corresponding to the first cloud image data based on the first cloud image texture feature, the first cloud image color feature, the first cloud cluster size information, and the first cloud cluster distribution feature;
[0120] The power generation prediction module 480 processes the first cloud layer feature based on the target photovoltaic power prediction model to obtain the predicted photovoltaic power generation power of the first photovoltaic power generation area at the first time.
[0121] For the description of specific functions and examples of each module and submodule of the device in the embodiment of the present invention, reference can be made to the relevant description of the corresponding steps in the above method embodiment, which will not be repeated here.
[0122] In the technical solution of the present invention, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.
[0123] According to an embodiment of the present invention, the present invention also provides a system and a readable storage medium.
[0124] Figure 5 A schematic block diagram of an example electronic device 800 that can be used to implement an embodiment of the present invention is shown. 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 can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices, 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.
[0125] like Figure 5 As shown, the device 800 includes a computing unit 801, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 802 or a computer program loaded from a storage unit 808 into a random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the device 800 can also be stored. The computing unit 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.
[0126] A number of components in the device 800 are connected to the I / O interface 805, including: an input unit 806, such as a keyboard, a mouse, etc.; an output unit 807, such as various types of displays, speakers, etc.; a storage unit 808, such as a disk, an optical disk, etc.; and a communication unit 809, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 809 allows the device 800 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.
[0127] The computing unit 801 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 801 performs the various methods and processes described above, such as a photovoltaic power prediction model training method based on hierarchical clustering and a photovoltaic power prediction method. For example, in some embodiments, a photovoltaic power prediction model training method based on hierarchical clustering and a photovoltaic power prediction method may be implemented as a computer software program, which is tangibly included in a machine-readable medium, such as a storage unit 808. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 800 via the ROM 802 and / or the communication unit 809. When the computer program is loaded into the RAM 803 and executed by the computing unit 801, one or more steps of the photovoltaic power prediction model training method based on hierarchical clustering and the photovoltaic power prediction method described above may be performed. Alternatively, in other embodiments, the computing unit 801 may be configured to execute the photovoltaic power prediction model training method and the photovoltaic power prediction method based on hierarchical clustering in any other appropriate manner (for example, by means of firmware).
[0128] 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.
[0129] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.
[0130] In the context of the present invention, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable 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. 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.
[0131] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer 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 computer. 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 (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).
[0132] The systems and techniques described herein may be implemented in a computing system that includes back-end 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 front-end 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 back-end components, middleware components, or front-end 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), and the Internet.
[0133] A computer 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 relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.
[0134] 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 disclosed in the present invention can be achieved, and this document does not limit this.
[0135] 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 principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A photovoltaic power prediction model training method based on hierarchical clustering, characterized in that: The method comprises: Extracting features from the cloud map data of each photovoltaic power generation area at each historical time to obtain cloud layer features of each of the cloud map data; Based on the hierarchical clustering algorithm, the cloud features corresponding to the cloud image data of each photovoltaic power generation area at each time are clustered to obtain multiple cloud feature clusters; For each of the cloud layer feature clusters, determining the corresponding photovoltaic power generation power based on the historical time and photovoltaic power generation area corresponding to each cloud layer feature in the cloud layer feature cluster, so as to obtain a photovoltaic power generation power set corresponding to the cloud layer feature cluster; Based on each of the cloud layer feature clusters and the photovoltaic power generation power sets corresponding to each of the cloud layer feature clusters, a photovoltaic power prediction model is trained to obtain a target photovoltaic power prediction model.
2. The method according to claim 1, characterized in that The feature extraction of the cloud map data of each photovoltaic power generation area at each historical time to obtain the cloud layer features of each cloud map data includes: Using an encoder to extract features from the cloud image data to obtain cloud image pixel features; The gray level co-occurrence matrix is used to calculate the pixel features of the cloud image to obtain the texture features of the cloud image; Processing the cloud image pixel features using a color histogram to obtain cloud image color features; Using an image segmentation edge detection algorithm to process the pixel features of the cloud image to obtain cloud size information; Calculating the cloud image texture features, the cloud image color features and the cloud size information to obtain cloud distribution features; and Based on the cloud image texture features, the cloud image color features, the cloud size information and the cloud distribution features, cloud layer features corresponding to the cloud image data are determined.
3. The method according to claim 2, characterized in that The photovoltaic power prediction model is trained based on each of the cloud layer feature clusters and the photovoltaic power generation power sets corresponding to each of the cloud layer feature clusters to obtain a target photovoltaic power prediction model, including: Determining the annotated photovoltaic power generation power corresponding to the cloud layer feature cluster based on the photovoltaic power generation power set corresponding to the cloud layer feature cluster; Analyzing the cloud layer characteristic cluster based on the photovoltaic power prediction model to obtain predicted photovoltaic power generation; Determining predicted losses based on the marked photovoltaic power generation and the predicted photovoltaic power generation; Based on the predicted loss, parameters of the photovoltaic power prediction model are adjusted to obtain the target photovoltaic power prediction model.
4. The method according to claim 3, characterized in that The step of adjusting parameters of the photovoltaic power prediction model based on the predicted loss includes: Using a decoder, reconstructing the cluster center of the cloud feature cluster to obtain a reconstructed cloud image; Determining a reconstruction loss based on the cloud image data corresponding to the cluster center of the cloud feature cluster and the reconstructed cloud image; determining a comprehensive loss based on the predicted loss and the reconstructed loss; When the comprehensive loss does not meet the preset loss condition, adjusting the model parameters of the photovoltaic power prediction model based on the gradient information corresponding to the comprehensive loss; When the comprehensive loss meets the preset loss condition, the target photovoltaic power prediction model is determined based on the photovoltaic power prediction model.
5. The method according to claim 4, characterized in that Also includes: Based on the gradient information corresponding to the comprehensive loss, the encoder, the gray-level co-occurrence matrix, the color histogram, the image segmentation edge detection algorithm and the model parameters in the decoder are adjusted.
6. The method according to any one of claims 1 to 5, characterized in that Based on the hierarchical clustering algorithm, the cloud features corresponding to the cloud image data of each photovoltaic power generation area at each time are clustered to obtain multiple cloud feature clusters, including: Using Euclidean distance, calculate the distance between any two cloud features in each of the cloud features; An agglomerative hierarchical clustering algorithm is used to cluster the cloud features based on the distance between any two of the cloud features to obtain a plurality of cloud feature clusters.
7. A photovoltaic power prediction method based on hierarchical clustering, characterized in that: include: Based on the target photovoltaic power prediction model, first cloud map data of the first photovoltaic power generation area at the first time is processed to obtain the predicted photovoltaic power generation power of the first photovoltaic power generation area at the first time; The target photovoltaic power prediction model is generated by training according to the photovoltaic power prediction model training method based on hierarchical clustering according to any one of claims 1 to 6.
8. A photovoltaic power prediction model training device based on hierarchical clustering, characterized in that: include: A cloud feature extraction module is used to extract features from the cloud map data of each photovoltaic power generation area at each historical time to obtain cloud features of each cloud map data; The cloud feature clustering module is used to cluster the cloud features corresponding to the cloud image data of each photovoltaic power generation area at each time based on a hierarchical clustering algorithm to obtain multiple cloud feature clusters; A power generation acquisition module is used to determine the corresponding photovoltaic power generation power for each of the cloud layer feature clusters based on the historical time and photovoltaic power generation area corresponding to each cloud layer feature in the cloud layer feature cluster, so as to obtain the photovoltaic power generation power set corresponding to the cloud layer feature cluster; The prediction model training module is used to train the photovoltaic power prediction model based on each of the cloud layer feature clusters and the photovoltaic power generation power sets corresponding to each of the cloud layer feature clusters to obtain a target photovoltaic power prediction model.
9. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.
10. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-7.