Load prediction method and device for orderly charging
By combining the timing dependency capture model and attention mechanism network in the load prediction model, the power load sample data and influence factors are used for prediction, the safety hazards caused by load monitoring and regulation delay are solved, and the power load is realized Short-term accurate prediction and orderly charging and regulation.
Patent Information
- Application Number
- CN202411460897.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-18
- Publication Date
- 2025-02-14
AI Technical Summary
In the prior art, load monitoring devices and cloud platforms have delays when transmitting and regulating charging power strategies, resulting in lag in regulation when the power load increases, posing a safety hazard.
A load prediction method for orderly charging is adopted. By collecting the power load sample data of the distribution station area of the residential community, the pre-trained load prediction model is input after standardization. The model includes a time-dependent capture model and attention mechanism network prediction model, using time dynamic characteristics and impact factors (such as regional occupancy, climate, holiday conditions, and major regional events) to predict.
It realizes short-term and accurate prediction of electricity load, ensures timely response to orderly charging and regulation, and ensures safe electricity use.
Smart Images
Figure CN119443854A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of power big data processing, and in particular to a method for load prediction of orderly charging. The present application also relates to a load prediction device for orderly charging, a computing device, and a computer-readable storage medium. Background Art
[0002] As the popularity of electric vehicles increases, there will be more and more charging scenarios in residential areas, so the popularity of intelligent and orderly charging technology will be greatly improved. Intelligent and orderly charging technology is to control the output power of charging piles by monitoring the power load of low-voltage areas, thereby achieving the safety of the distribution network.
[0003] However, since the load monitoring device monitors the load data of the substation and transmits it to the cloud platform, there is a certain delay. After the cloud platform generates the charging power control strategy and sends it to the charging pile for execution, there is also a certain delay, which may lead to a rapid increase in power load and safety hazards caused by the lag in control. Therefore, it is necessary to adopt certain technologies to predict the short-term power load of the substation in order to execute the orderly charging strategy more quickly. Summary of the invention
[0004] In view of this, the embodiment of the present application provides a load prediction method for orderly charging to solve the technical defects existing in the prior art. The embodiment of the present application also provides a load prediction device for orderly charging, a computing device, and a computer-readable storage medium.
[0005] According to a first aspect of an embodiment of the present application, a load forecasting method for orderly charging is provided, comprising:
[0006] Collecting power load sample data of the distribution station area in the residential area, and standardizing the power load sample data to obtain standard power load sample data;
[0007] Inputting the standard power load sample data into a pre-trained load forecasting model, wherein the load forecasting model includes a temporal dependency capture model and an attention mechanism network forecasting model;
[0008] The standard power load sample data is processed by the timing dependency capture model to obtain time dynamic characteristics;
[0009] The time dynamic features are processed through the attention mechanism network prediction model to obtain power load prediction data, wherein the influencing factors associated with the attention mechanism network prediction model include the regional occupancy rate, regional climate, holiday conditions and major regional events corresponding to the residential community distribution station area.
[0010] Optionally, also include:
[0011] Abnormal data in the power load sample data is identified and corrected.
[0012] Optionally, the step of standardizing the power load sample data to obtain standard power load sample data includes:
[0013] The power load sample data is standardized by using a normalization method to obtain the standard power load sample data.
[0014] Optionally, the processing of the standard power load sample data by the time series dependency capture model to obtain the time dynamic characteristics includes:
[0015] The time dependency capture model of the two-layer GRU neural network architecture is used to capture the time dependency relationship in the standard power load sample data to obtain the time dynamic characteristics.
[0016] Optionally, the attention mechanism network prediction model includes a discard layer, a 1-dimensional convolution layer, a 1-dimensional maximum pooling layer, an attention mechanism layer and a fully connected layer.
[0017] Optionally, the processing of the temporal dynamic features by the attention mechanism network prediction model to obtain power load prediction data includes:
[0018] Extracting the temporal dynamic features based on the 1-dimensional convolutional layer to obtain target features;
[0019] Based on the 1-dimensional maximum pooling layer, the target feature is reduced in dimension to obtain a reduced-dimensional target feature;
[0020] Based on the attention mechanism layer, processing the output results of the 1D convolution layer and the 1D maximum pooling layer;
[0021] The processed dimensionality reduction target features are processed based on the fully connected layer to obtain the power load prediction data, wherein during the data processing of the attention mechanism network prediction model, the discard layer randomly discards the neurons included in the attention mechanism network prediction model.
[0022] Optionally, processing the output results of the 1-dimensional convolutional layer and the 1-dimensional maximum pooling layer based on the attention mechanism layer includes:
[0023] Different weights are assigned to the target sub-features in the target feature, and different weights are assigned to the reduced dimension target sub-features in the reduced dimension target feature according to the regional occupancy rate, the regional climate, the holiday conditions, and the major regional events.
[0024] According to a second aspect of an embodiment of the present application, a load prediction device for orderly charging is provided, comprising:
[0025] A collection module is configured to collect power load sample data of a residential community distribution station area, and standardize the power load sample data to obtain standard power load sample data;
[0026] The processing module is configured to input the standard power load sample data into a pre-trained load forecasting model, wherein the load forecasting model includes a timing dependency capture model and an attention mechanism network prediction model; the standard power load sample data is processed by the timing dependency capture model to obtain time dynamic features; the time dynamic features are processed by the attention mechanism network prediction model to obtain power load forecasting data, wherein the influencing factors associated with the attention mechanism network prediction model include the regional occupancy rate, regional climate, holiday conditions and major regional events corresponding to the residential community distribution station area.
[0027] According to a third aspect of an embodiment of the present application, a computing device is provided, including:
[0028] Memory and processor;
[0029] The memory is used to store computer-executable instructions, and the processor implements the steps of the load prediction method for orderly charging when executing the computer-executable instructions.
[0030] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the load prediction method for orderly charging are implemented.
[0031] According to a fifth aspect of an embodiment of the present application, a chip is provided, which stores a computer program, and when the computer program is executed by the chip, the steps of the load prediction method for orderly charging are implemented.
[0032] The load forecasting method for orderly charging provided in the present application collects the power load sample data of the distribution station area of the residential community, and standardizes the power load sample data to obtain standard power load sample data; inputs the standard power load sample data into the pre-trained load forecasting model, wherein the load forecasting model includes a time series dependency capture model and an attention mechanism network prediction model; processes the standard power load sample data through the time series dependency capture model to obtain time dynamic features; processes the time dynamic features through the attention mechanism network prediction model to obtain power load forecasting data, wherein the influence factors associated with the attention mechanism network prediction model include the regional occupancy rate, regional climate, holiday conditions and major regional events corresponding to the distribution station area of the residential community. Realizing short-term and accurate prediction of power load in intelligent orderly technology for residential communities provides conditions for timely response to orderly charging regulation and control, and ensures safe power use. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0034] Figure 1 This is a flow chart of a load prediction method for orderly charging provided by an embodiment of the present application;
[0035] Figure 2 It is a model schematic diagram of a load prediction method for orderly charging provided by an embodiment of the present application;
[0036] Figure 3 This is a schematic diagram of the attention mechanism structure of a load prediction method for orderly charging provided by an embodiment of the present application;
[0037] Figure 4 It is a training flow chart of a load prediction method for orderly charging provided by an embodiment of the present application;
[0038] Figure 5 It is a structural schematic diagram of a load prediction device for orderly charging provided by an embodiment of the present application;
[0039] Figure 6 It is a structural block diagram of a computing device provided in one embodiment of the present application. DETAILED DESCRIPTION
[0040] Many specific details are described in the following description to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present application, so the present application is not limited by the specific implementation disclosed below.
[0041] The terms used in one or more embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of the present application. The singular forms of "a", "said" and "the" used in one or more embodiments of the present application and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of the present application refers to and includes any or all possible combinations of one or more associated listed items.
[0042] It should be understood that, although the terms first, second, etc. may be used to describe various information in one or more embodiments of the present application, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of the present application, the first may also be referred to as the second, and similarly, the second may also be referred to as the first.
[0043] In the present application, a method for load prediction of orderly charging is provided. The present application also relates to a load prediction device for orderly charging, a computing device, and a computer-readable storage medium, which are described in detail one by one in the following embodiments.
[0044] Figure 1 A flow chart of a load prediction method for orderly charging provided according to an embodiment of the present application is shown, which specifically includes the following steps:
[0045] Step S102: collecting power load sample data of the distribution station area of the residential community, and standardizing the power load sample data to obtain standard power load sample data;
[0046] Step S104: inputting the standard power load sample data into a pre-trained load forecasting model, wherein the load forecasting model includes a temporal dependency capture model and an attention mechanism network forecasting model;
[0047] Step S1041: Processing the standard power load sample data through the timing dependency capture model to obtain time dynamic characteristics;
[0048] Step S1042: The time dynamic features are processed through the attention mechanism network prediction model to obtain power load prediction data, wherein the influencing factors associated with the attention mechanism network prediction model include the regional occupancy rate, regional climate, holiday conditions and major regional events corresponding to the residential community distribution station area.
[0049] Among them, the residential community distribution station area is one or more pre-determined communities, and the power load sample data represents the charging status of the charging piles contained in the residential community distribution station area within a certain period of time. The charging status may include voltage data, power data, etc. The specific data type included is determined by the actual usage scenario and is not limited in this embodiment.
[0050] Based on this, the power load sample data is standardized and input into the load forecasting model, which then outputs the power load forecasting data. The load forecasting model consists of two sub-models: the temporal dependency capture model and the attention mechanism network prediction model.
[0051] Specifically, the temporal dependency capture model uses the GRU neural network architecture, while the attention mechanism network prediction model uses the CNN neural network architecture. Since the GRU neural network is good at processing time series data and can capture the temporal dependency in the power load, and the CNN neural network is good at extracting local features in the data, it performs multi-level feature extraction on the input data through convolutional layers and pooling layers. Combining GRU and CNN to obtain a load prediction model with a GRU-CNN architecture, it uses GRU to capture the temporal dynamic features in the power load data, and then extracts and analyzes these features through CNN to ensure the efficiency of understanding the changing laws of the power load.
[0052] Moreover, in actual usage scenarios, the changes in short-term power load in residential areas are sudden and highly disturbing, and show certain seasonal and synergistic characteristics. In order to further improve the prediction accuracy, the attention mechanism is introduced into the attention mechanism network prediction model and inserted before its fully connected layer. In the design of the attention mechanism, the accuracy of short-term power load prediction and the speed of prediction are considered at the same time. More specifically, through the performance evaluation and experimental data verification of different attention mechanisms, the scaled dot product attention mechanism is selected, and the influencing factors of regional occupancy rate, regional climate, holiday conditions and major regional events corresponding to the residential distribution station area are added to control the stability of attention distribution.
[0053] Among them, major regional events can be social activities such as concerts and night markets, or public transportation changes such as subway openings and bus route changes, or geological disasters. The event type corresponding to the major regional event and the frequency of personnel movement caused by the event are used as independent variables, and specific influencing factor data are generated through the preset practical reference equation.
[0054] Furthermore, before the power load sample data is standardized to obtain the standard power load sample data, the power load sample data needs to be preprocessed. In this embodiment, the specific implementation process is as follows:
[0055] Abnormal data in the power load sample data is identified and corrected.
[0056] Furthermore, the process of standardizing the power load sample data to obtain the standard power load sample data is specifically implemented as follows in this embodiment:
[0057] The power load sample data is standardized by using a normalization method to obtain the standard power load sample data.
[0058] Among them, for the power load sample data, the first thing to do is to identify whether there is abnormal data in it. If there is, the identified abnormal data is corrected, and the corrected power load sample data is standardized by the normalization method to obtain the standard power load sample data; if there is no abnormal data in the power load sample data, it is directly standardized. Standardization is achieved through the normalization method to ensure the consistency and applicability of the data.
[0059] Furthermore, the standard power load sample data is processed by the time series dependency capture model to obtain the process of time dynamic characteristics. In this embodiment, the specific implementation process is as follows:
[0060] The time dependency capture model of the two-layer GRU neural network architecture is used to capture the time dependency relationship in the standard power load sample data to obtain the time dynamic characteristics.
[0061] Furthermore, the attention mechanism network prediction model includes a discard layer, a 1-dimensional convolution layer, a 1-dimensional maximum pooling layer, an attention mechanism layer and a fully connected layer.
[0062] Furthermore, the attention mechanism network prediction model is used to process the temporal dynamic features to obtain the power load prediction data. In this embodiment, the specific implementation process is as follows:
[0063] Based on the 1-dimensional convolution layer, feature extraction is performed on the time dynamic feature to obtain the target feature; based on the 1-dimensional maximum pooling layer, dimension reduction is performed on the target feature to obtain the reduced dimension target feature; based on the attention mechanism layer, the output results of the 1-dimensional convolution layer and the 1-dimensional maximum pooling layer are processed; based on the fully connected layer, the processed reduced dimension target feature is processed to obtain the power load forecast data, wherein, during the data processing process of the attention mechanism network prediction model, the discard layer randomly discards the neurons included in the attention mechanism network prediction model.
[0064] Furthermore, based on the attention mechanism layer, the output results of the 1D convolution layer and the 1D maximum pooling layer are processed. In this embodiment, the specific implementation process is as follows:
[0065] Different weights are assigned to the target sub-features in the target feature, and different weights are assigned to the reduced dimension target sub-features in the reduced dimension target feature according to the regional occupancy rate, the regional climate, the holiday conditions, and the major regional events.
[0066] Among them, Figure 2 As shown in the model schematic diagram of a load forecasting method for orderly charging, GRU1 and GRU2 represent the timing dependency capture model of the two-layer GRU neural network architecture.
[0067] It should be noted that for the 1D convolutional layer and the 1D maximum pooling layer, in addition to the single arrangement, they can also be arranged in multiple layers in a loop. Figure 2 Taking the model schematic diagram of a load prediction method for orderly charging as an example, the one-dimensional convolutional layers 1DCNN1, 1DCNN2 and the one-dimensional maximum pooling layers 1DMaxpooling1, 1DMaxpooling2 are arranged in a loop, and the processing results are output by the fully connected layer FC.
[0068] In addition, a dropout layer is arranged to randomly drop some neurons to improve the robustness of the neural network model. The attention mechanism layer is introduced to process the features after convolution and pooling, and different weights are given to different features, so that the model can focus on the most important information for prediction while ignoring irrelevant or minor features, thereby improving the accuracy of short-term load forecasting. Specifically, the attention structure of the attention mechanism layer is as follows: Figure 3 The attention mechanism structure diagram of a load prediction method for orderly charging is shown in FIG. t is the attention distribution vector, as shown in the following formula:
[0069]
[0070] Among them, e t =vtanh(Wht +b), W and v represent the attention weight, b represents the bias, Softmax is the normalized exponential function, and the final output y t , as shown in the following formula:
[0071]
[0072] As for the training process of the load forecasting model, Figure 4 As shown in the training flow chart of the load forecasting method for orderly charging, the input training material, that is, the training set load sample, is also subjected to abnormal data processing, data correction and data normalization. Then, the load forecasting model processes the training set load sample. Specifically, the data is processed by the time dependency capture model of the two-layer GRU neural network architecture in the load forecasting model, and the processing result is processed by the attention mechanism network prediction model including the dropout layer, 1DCNN layer, 1DMaxpooling layer, Attention mechanism layer and the full connection layer FC in the load forecasting model. The predicted result is compared with the actual result, and training is performed by constructing a loss function until the training stop condition is met.
[0073] After the training is completed, the test set load samples that need to be predicted are input into the trained model to output the prediction results. Specifically, the power load sample data in the residential community distribution station area is collected at a fixed time period, and the power load sample data of the current period is input into the trained load prediction model, and the output result is the power load prediction data, which represents the prediction of the power load sample data in the next time period.
[0074] Corresponding to the above method embodiment, the present application also provides an embodiment of a load prediction device for orderly charging, Figure 5 FIG. 1 is a schematic diagram showing the structure of a load prediction device for orderly charging provided by an embodiment of the present application. Figure 5 As shown, the device comprises:
[0075] The collection module 502 is configured to collect power load sample data of the distribution station area of the residential area, and standardize the power load sample data to obtain standard power load sample data;
[0076] The processing module 504 is configured to input the standard power load sample data into a pre-trained load forecasting model, wherein the load forecasting model includes a timing dependency capture model and an attention mechanism network prediction model; the standard power load sample data is processed by the timing dependency capture model to obtain time dynamic features; the time dynamic features are processed by the attention mechanism network prediction model to obtain power load forecasting data, wherein the influencing factors associated with the attention mechanism network prediction model include the regional occupancy rate, regional climate, holiday conditions and major regional events corresponding to the residential community distribution station area.
[0077] In an optional embodiment, the acquisition module 502 further includes:
[0078] The correction module is configured to identify abnormal data in the power load sample data and correct it.
[0079] In an optional embodiment, the acquisition module 502 is further configured to:
[0080] The power load sample data is standardized by using a normalization method to obtain the standard power load sample data.
[0081] In an optional embodiment, the processing module 504 is further configured to:
[0082] The time dependency capture model of the two-layer GRU neural network architecture is used to capture the time dependency relationship in the standard power load sample data to obtain the time dynamic characteristics.
[0083] In an optional embodiment, the attention mechanism network prediction model in the processing module 504 includes a discard layer, a 1-dimensional convolution layer, a 1-dimensional maximum pooling layer, an attention mechanism layer and a fully connected layer.
[0084] In an optional embodiment, the processing module 504 is further configured to:
[0085] Based on the 1-dimensional convolution layer, feature extraction is performed on the time dynamic feature to obtain the target feature; based on the 1-dimensional maximum pooling layer, dimension reduction is performed on the target feature to obtain the reduced dimension target feature; based on the attention mechanism layer, the output results of the 1-dimensional convolution layer and the 1-dimensional maximum pooling layer are processed; based on the fully connected layer, the processed reduced dimension target feature is processed to obtain the power load forecast data, wherein, during the data processing process of the attention mechanism network prediction model, the discard layer randomly discards the neurons included in the attention mechanism network prediction model.
[0086] In an optional embodiment, the processing module 504 is further configured to:
[0087] Different weights are assigned to the target sub-features in the target feature, and different weights are assigned to the reduced dimension target sub-features in the reduced dimension target feature according to the regional occupancy rate, the regional climate, the holiday conditions, and the major regional events.
[0088] The load prediction device for orderly charging provided in the present application obtains standard power load sample data by collecting power load sample data from the distribution station area of a residential community and standardizing the power load sample data; inputting the standard power load sample data into a pre-trained load prediction model, wherein the load prediction model includes a timing dependency capture model and an attention mechanism network prediction model; processing the standard power load sample data through the timing dependency capture model to obtain time dynamic features; processing the time dynamic features through the attention mechanism network prediction model to obtain power load prediction data, wherein the influencing factors associated with the attention mechanism network prediction model include the regional occupancy rate, regional climate, holiday conditions and major regional events corresponding to the distribution station area of the residential community. Realizing short-term and accurate prediction of power load in intelligent orderly technology for residential communities provides conditions for timely response to orderly charging regulation and control, and ensures safe use of electricity
[0089] The above is a schematic scheme of a load prediction device for orderly charging of this embodiment. It should be noted that the technical scheme of the load prediction device for orderly charging and the technical scheme of the load prediction method for orderly charging mentioned above belong to the same concept. For the details not described in detail in the technical scheme of the load prediction device for orderly charging, please refer to the description of the technical scheme of the load prediction method for orderly charging mentioned above. In addition, each component in the device embodiment should be understood as a functional module that must be established to implement each step of the program flow or each step of the method, and each functional module is not an actual functional division or separation definition. The device claim defined by such a group of functional modules should be understood as a functional module architecture that mainly implements the solution through the computer program recorded in the specification, and should not be understood as a physical device that mainly implements the solution through hardware.
[0090] Figure 6 The block diagram of a computing device 600 according to an embodiment of the present application is shown. The components of the computing device 600 include but are not limited to a memory 610 and a processor 620. The processor 620 is connected to the memory 610 via a bus 630, and the database 650 is used to store data.
[0091] The computing device 600 also includes an access device 640 that enables the computing device 600 to communicate via one or more networks 660. Examples of these networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 640 may include one or more of any type of network interface (e.g., a network interface card (NIC)) whether wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a World Wide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, a near field communication (NFC) interface, and the like.
[0092] In one embodiment of the present application, the above components of the computing device 600 and Figure 6 Other components not shown in the figure may also be connected to each other, for example, via a bus. It should be understood that Figure 6 The computing device structure block diagram shown is only for the purpose of illustration, and is not intended to limit the scope of the present application. Those skilled in the art may add or replace other components as needed.
[0093] The computing device 600 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or PC. The computing device 600 may also be a mobile or stationary server.
[0094] The processor 620 is used to execute computer executable instructions for each step of the load forecasting method for orderly charging.
[0095] The above is a schematic scheme of a computing device of this embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the above-mentioned orderly charging load prediction method belong to the same concept, and the details not described in detail in the technical scheme of the computing device can be referred to the description of the technical scheme of the above-mentioned orderly charging load prediction method.
[0096] An embodiment of the present application further provides a computer-readable storage medium storing computer instructions, which, when executed by a processor, are used to execute the steps of the load prediction method for orderly charging.
[0097] The above is a schematic scheme of a computer-readable storage medium of this embodiment. It should be noted that the technical scheme of the storage medium and the technical scheme of the load prediction method for orderly charging described above belong to the same concept, and the details not described in detail in the technical scheme of the storage medium can be referred to the description of the technical scheme of the load prediction method for orderly charging described above.
[0098] An embodiment of the present application further provides a chip storing a computer program, which implements the steps of the load prediction method for orderly charging when the computer program is executed by the chip.
[0099] The above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0100] The computer instructions include computer program codes, which may be in source code form, object code form, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.
[0101] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.
[0102] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0103] The preferred embodiments of the present application disclosed above are only used to help explain the present application. The optional embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of the present application. The present application selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present application, so that those skilled in the art can understand and use the present application well. The present application is only limited by the claims and their full scope and equivalents.
Claims
1. A load forecasting method for orderly charging, characterized in that: include: Collecting power load sample data of the distribution station area in the residential area, and standardizing the power load sample data to obtain standard power load sample data; Inputting the standard power load sample data into a pre-trained load forecasting model, wherein the load forecasting model includes a temporal dependency capture model and an attention mechanism network forecasting model; The standard power load sample data is processed by the timing dependency capture model to obtain time dynamic characteristics; The time dynamic features are processed through the attention mechanism network prediction model to obtain power load prediction data, wherein the influencing factors associated with the attention mechanism network prediction model include the regional occupancy rate, regional climate, holiday conditions and major regional events corresponding to the residential community distribution station area.
2. The method according to claim 1, characterized in that Before the power load sample data is standardized to obtain standard power load sample data, the method further includes: Abnormal data in the power load sample data is identified and corrected.
3. The method according to claim 1, characterized in that The step of standardizing the power load sample data to obtain standard power load sample data includes: The power load sample data is standardized by using a normalization method to obtain the standard power load sample data.
4. The method according to claim 1, characterized in that: The standard power load sample data is processed by the time series dependency capture model to obtain time dynamic characteristics, including: The time dependency capture model of the two-layer GRU neural network architecture is used to capture the time dependency relationship in the standard power load sample data to obtain the time dynamic characteristics.
5. The method according to claim 1, characterized in that The attention mechanism network prediction model includes a discard layer, a 1-dimensional convolution layer, a 1-dimensional maximum pooling layer, an attention mechanism layer and a fully connected layer.
6. The method according to claim 5, characterized in that The time dynamic features are processed by the attention mechanism network prediction model to obtain power load prediction data, including: Extracting the temporal dynamic features based on the 1-dimensional convolutional layer to obtain target features; Based on the 1-dimensional maximum pooling layer, the target feature is reduced in dimension to obtain a reduced-dimensional target feature; Based on the attention mechanism layer, processing the output results of the 1D convolution layer and the 1D maximum pooling layer; The processed dimensionality reduction target features are processed based on the fully connected layer to obtain the power load prediction data, wherein during the data processing of the attention mechanism network prediction model, the discard layer randomly discards the neurons included in the attention mechanism network prediction model.
7. The method according to claim 6, characterized in that The processing of the output results of the 1-dimensional convolutional layer and the 1-dimensional maximum pooling layer based on the attention mechanism layer includes: Different weights are assigned to the target sub-features in the target feature, and different weights are assigned to the reduced dimension target sub-features in the reduced dimension target feature according to the regional occupancy rate, the regional climate, the holiday conditions, and the major regional events.
8. A load prediction device for orderly charging, characterized in that: include: A collection module is configured to collect power load sample data of a residential community distribution station area, and standardize the power load sample data to obtain standard power load sample data; The processing module is configured to input the standard power load sample data into a pre-trained load forecasting model, wherein the load forecasting model includes a timing dependency capture model and an attention mechanism network prediction model; the standard power load sample data is processed by the timing dependency capture model to obtain time dynamic features; the time dynamic features are processed by the attention mechanism network prediction model to obtain power load forecasting data, wherein the influencing factors associated with the attention mechanism network prediction model include the regional occupancy rate, regional climate, holiday conditions and major regional events corresponding to the residential community distribution station area.
9. A computing device, characterized in that include: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the steps of the load forecasting method for orderly charging as described in any one of claims 1 to 9.
10. A computer-readable storage medium storing computer instructions, characterized in that: When the instruction is executed by the processor, the steps of the load prediction method for orderly charging described in any one of claims 1 to 9 are implemented.