A data fusion processing method, system and device for power system
By preprocessing and determining the characteristic correlation of multi-source heterogeneous data of the power system and fusion using a modal network, the problem of poor data fusion effect in the prior art is solved, and high-quality multi-modal data fusion is achieved.
Patent Information
- Application Number
- CN202111340803.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-11-12
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2041-11-12
AI Technical Summary
The prior art fails to effectively consider the correlation and dependence between different mode data when processing multi-source, heterogeneous, and multi-mode power data, resulting in poor data fusion effect.
By collecting multi-source heterogeneous data from the power system, pre-processing and feature extraction, the correlation between modal features is determined, and relevant features are input into the modal network and the feature-level fusion network for modal fusion to generate the final fusion results.
It effectively improves the quality of data fusion, realizes the integration of internal and modal modes, and improves the integration consistency and accuracy of data.
Smart Images

Figure CN114239683B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of power data processing, and in particular to a data fusion processing method, system and device for a power system. Background Art
[0002] With the development of the Internet, people's demand for power grid data dispatch is also increasing. In order to meet the changing needs of the company and other users, the power grid system needs to develop new businesses and provide new services, such as line loss in the same period, online state grid, new energy consumption, spot market, etc.; these services require the support of background data such as EMS, water regulation, TMR, OMS, flue gas, AVC, etc. However, the background power data has the characteristics of different sources, inconsistent data formats, multi-modal data, difficult to connect business data, and complex technical environment. Then it is necessary to design and develop a method that can quickly break through the barriers of business systems and realize data fusion in a complex multi-source data environment.
[0003] At present, the processing of multi-source, heterogeneous, and multi-modal data in the existing technology is generally a simple splicing of sampled single-mode features or a simple combination and reclassification of single-mode decision results, without considering the correlation and dependency issues between different modal data, resulting in poor data fusion effect.
[0004] Therefore, the present application provides a data fusion processing method, system and device for a power system. Summary of the invention
[0005] The purpose of the embodiment of the present application is to provide a data fusion processing method for a power system to solve the current technical problem of poor data fusion quality. The specific technical solution is as follows:
[0006] In a first aspect, a data fusion processing method for a power system is provided, the method comprising:
[0007] Collect multi-source heterogeneous data from relevant business departments of the power system based on dispatch query requirements;
[0008] Performing a first preprocessing and a second preprocessing on the multi-source heterogeneous data in sequence to obtain preprocessed multi-source heterogeneous data;
[0009] Extracting n modal features of the preprocessed multi-source heterogeneous data, wherein the n modal features include image features, video features, voice features, and point cloud features;
[0010] Determine the correlation between n modal features;
[0011] The n related modal features are respectively input into the corresponding modal network and feature-level fusion network for modal fusion to obtain the corresponding modal decision and fusion decision;
[0012] The modal decisions output by each modal network and the fusion decision are fused to obtain a final fusion result.
[0013] Optionally, the first preprocessing comprises the following steps:
[0014] Cleaning the multi-source heterogeneous data to obtain cleaned multi-source heterogeneous data;
[0015] The cleaned multi-source heterogeneous data is subjected to denoising processing to obtain denoised multi-source heterogeneous data.
[0016] Optionally, the second preprocessing comprises the following steps:
[0017] Classifying the multi-source heterogeneous data that has undergone the first preprocessing to obtain structured data and unstructured data;
[0018] Converting the unstructured data into structured data and storing the data in a structured database;
[0019] Optionally, determining the correlation between the n modal features includes:
[0020] Performing feature enhancement processing and denoising processing on the n modes in sequence;
[0021] The semantic recognition processing is performed on the enhanced and denoised modal features to obtain the correlation between the modal features.
[0022] Optionally, performing semantic recognition processing on the enhanced and denoised modal features to obtain the correlation between the modal features includes:
[0023] Each enhanced and denoised modal feature is input into multiple regular graph convolutions and convolved in sequence; each graph convolution layer contains a graph construction module, a graph convolution module and a feature filtering module;
[0024] The convolutional features are input into the NLP (Natural Language Processing) layer for semantic segmentation to obtain the segmentation score and the global maximum pooling layer for classification to obtain the classification score;
[0025] The correlation between the features of each modality is determined according to the segmentation score and the classification score.
[0026] Optionally, the modal network and the feature-level fusion network both adopt a generative adversarial network structure, including n generators, n modal discriminators and a fusion discriminator, and the n associated modal features are respectively input into the respective corresponding modal networks and feature-level fusion networks for modal fusion to obtain the respective corresponding modal decisions and fusion decisions, including:
[0027] Extracting the noise of the n modal features;
[0028] Input the noise of n modal features into their respective generators to obtain the corresponding fake sample features;
[0029] Each modal feature and the corresponding fake sample feature are respectively input into the corresponding modal discriminator and fusion discriminator for feature fusion to obtain the corresponding modal decision and fusion decision.
[0030] Optionally, the method further comprises:
[0031] Obtain the parameter weight values of each modality network and feature-level fusion network;
[0032] Compare the weight value of each parameter with the preset threshold;
[0033] Parameters below the preset threshold are deleted.
[0034] In a second aspect, the present application provides a data fusion processing system for a power system, the system comprising:
[0035] A collection unit, used to collect multi-source heterogeneous data from relevant business departments of the power system based on dispatch query requirements;
[0036] A preprocessing unit, configured to sequentially perform a first preprocessing and a second preprocessing on the multi-source heterogeneous data to obtain preprocessed multi-source heterogeneous data;
[0037] An extraction unit, used to extract n modal features of the preprocessed multi-source heterogeneous data, wherein the n modal features include image features, video features, voice features, and point cloud features;
[0038] A determination unit, used to determine the correlation between n modal features;
[0039] The fusion unit is used to input the n related modal features into the corresponding modal networks and feature-level fusion networks respectively to perform modal fusion to obtain the corresponding modal decisions and fusion decisions; the modal decisions and fusion decisions output by each modal network are fused to obtain the final fusion result.
[0040] In a third aspect, the present application provides an electronic device, including a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;
[0041] Memory, used to store computer programs;
[0042] The processor is used to implement any method step described in the first aspect when executing the program stored in the memory.
[0043] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, any method step described in the first aspect is implemented.
[0044] In a fifth aspect, a computer program product comprising instructions is provided, which, when executed on a computer, enables the computer to execute any of the above-mentioned data fusion processing methods for power systems.
[0045] Beneficial effects of the embodiments of the present application:
[0046] The embodiment of the present application provides a data fusion processing method, system and device for the power system. The present application collects multi-source heterogeneous data from relevant business departments of the power system based on the dispatch query requirements; performs the first preprocessing and the second preprocessing on the multi-source heterogeneous data in sequence to obtain preprocessed multi-source heterogeneous data; extracts n modal features of the preprocessed multi-source heterogeneous data; determines the correlation between the n modal features; inputs the associated n modal features into the respective corresponding modal networks and feature-level fusion networks for modal fusion to obtain the respective corresponding modal decisions and fusion decisions; performs decision fusion on the modal decisions and fusion decisions output by each modal network to obtain the final fusion result. The present application performs semantic recognition on multi-source heterogeneous multi-modal power data to find the correlation between the features of each modal data, and inputs the associated modal features into the modal network and the feature-level fusion network for fusion, which not only realizes the fusion within the modality, but also realizes the data fusion between the modalities, effectively improving the data fusion quality.
[0047] Of course, implementing any product or method of the present application does not necessarily require achieving all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0049] Figure 1 A flow chart of a data fusion processing method for a power system provided in an embodiment of the present application;
[0050] Figure 2 A schematic diagram of a multimodal fusion process according to an embodiment of the present application;
[0051] Figure 3 is a schematic diagram of the structure of a modal network and a feature-level fusion network in an embodiment of the present application;
[0052] Figure 4 is a schematic diagram of the semantic recognition process of an embodiment of the present application;
[0053] Figure 5 A schematic diagram of the structure of a data fusion processing system for a power system provided in an embodiment of the present application;
[0054] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0056] The present application embodiment provides a data fusion processing method for a power system. The following will describe in detail a data fusion processing method for a power system provided by the present application embodiment in combination with a specific implementation method. Figure 1 As shown, the specific steps are as follows:
[0057] Step S101: Collect multi-source heterogeneous data from power system-related business departments based on scheduling query requirements.
[0058] In this step, if the scheduling query demand is, for example, to query the power consumption, it is necessary to obtain the user's identity information and the user's power account information, collect the sensor data of the corresponding equipment based on the user's power account information, and integrate the power data, identity information data, and balance information, and then present it through the foreground.
[0059] In one example, multi-source heterogeneous data includes power grid operation data, power transaction data, power marketing data, and equipment asset data.
[0060] Step S102: performing a first preprocessing and a second preprocessing on the multi-source heterogeneous data in sequence to obtain preprocessed multi-source heterogeneous data.
[0061] Step S103: extracting n modal features of the preprocessed multi-source heterogeneous data, wherein the n modal features include image features, video features, voice features, and point cloud features.
[0062] Step S104: Determine the correlation between n modal features.
[0063] Step S105: inputting the n associated modal features into their respective corresponding modal networks and feature-level fusion networks for modal fusion to obtain their respective corresponding modal decisions and fusion decisions.
[0064] Step S106: Perform decision fusion on the modal decisions output by each modal network and the fusion decision to obtain a final fusion result. Figure 2 As shown, through the modal network, the feature-level fusion network, the modal decision and the fusion decision, not only the feature fusion within the modality is achieved, but also the feature fusion between the modalities is achieved, which effectively improves the fusion quality.
[0065] Optionally, the first preprocessing comprises the following steps:
[0066] Cleaning the multi-source heterogeneous data to obtain cleaned multi-source heterogeneous data;
[0067] The cleaned multi-source heterogeneous data is subjected to denoising processing to obtain denoised multi-source heterogeneous data.
[0068] Optionally, the second preprocessing comprises the following steps:
[0069] Classifying the multi-source heterogeneous data that has undergone the first preprocessing to obtain structured data and unstructured data;
[0070] Converting the unstructured data into structured data and storing the data in a structured database;
[0071] Optionally, determining the correlation between the n modal features includes:
[0072] Performing feature enhancement processing and denoising processing on the n modes in sequence;
[0073] In the embodiment of the present application, the denoising process may adopt an adaptive denoising method, which can effectively remove outlier noise points and maintain the original geometric characteristics of the data.
[0074] The semantic recognition processing is performed on the enhanced and denoised modal features to obtain the correlation between the modal features.
[0075] Optionally, performing semantic recognition processing on the enhanced and denoised modal features to obtain the correlation between the modal features includes:
[0076] Each enhanced and denoised modal feature is input into multiple regular graph convolutions and convolved in sequence; each graph convolution layer contains a graph construction module, a graph convolution module and a feature filtering module;
[0077] In one example, if Figure 3As shown in the figure, the feature of one of the modalities is the point cloud feature. The point cloud feature is input into multiple graph convolutional layers for graph construction, filtering and convolution. The graph Laplacian matrix is used to describe the feature relationship between layers, which solves the invariance of point cloud permutation while maintaining good robustness to noise and density.
[0078] The convolutional features are input into the NLP layer for semantic segmentation to obtain the segmentation score and the global maximum pooling layer for classification to obtain the classification score;
[0079] The correlation between the features of each modality is determined according to the segmentation score and the classification score.
[0080] In the embodiment of the present application, modal features having close segmentation scores and classification scores may be determined as interrelated features.
[0081] Alternatively, if Figure 4 As shown, the modal network and the feature-level fusion network both adopt a generative adversarial network structure, including n generators, n modal discriminators and a fusion discriminator. The n associated modal features are respectively input into the respective corresponding modal networks and feature-level fusion networks for modal fusion to obtain the respective corresponding modal decisions and fusion decisions, including:
[0082] Extracting the noise of the n modal features;
[0083] Input the noise of n modal features into their respective generators to obtain the corresponding fake sample features;
[0084] Each modal feature and the corresponding fake sample feature are respectively input into the corresponding modal discriminator and fusion discriminator for feature fusion to obtain the corresponding modal decision and fusion decision.
[0085] In the embodiment of the present application, a minority model data set can be increased by generating an adversarial network structure, thereby solving the problem of imbalanced model data.
[0086] Optionally, the method further comprises:
[0087] Obtain the parameter weight values of each modality network and feature-level fusion network;
[0088] Compare the weight value of each parameter with the preset threshold;
[0089] Parameters below the preset threshold are deleted.
[0090] In an example, the preset threshold can be set to 0.1, and the parameters corresponding to weight values lower than 0.1 are deleted. The closer the weight value is to 0, the less important the parameter is. Removing unimportant parameters can effectively compress the structure of the modal network, simplify the complexity of the network model, and ensure the accuracy of the modal network.
[0091] In the second aspect, based on the same inventive concept, the present application provides a data fusion processing system for a power system, such as Figure 5 As shown, the system comprises:
[0092] A collection unit 510 is used to collect multi-source heterogeneous data of power system-related business departments based on scheduling query requirements;
[0093] A preprocessing unit 520, configured to sequentially perform a first preprocessing and a second preprocessing on the multi-source heterogeneous data to obtain preprocessed multi-source heterogeneous data;
[0094] An extraction unit 530, configured to extract n modal features of the preprocessed multi-source heterogeneous data, wherein the n modal features include image features, video features, voice features, and point cloud features;
[0095] A determination unit 540, configured to determine the correlation between the n modal features;
[0096] The fusion unit 550 is used to input the associated n modal features into the respective corresponding modal networks and feature-level fusion networks for modal fusion to obtain the respective corresponding modal decisions and fusion decisions; and to fuse the modal decisions and fusion decisions output by each modal network to obtain the final fusion result.
[0097] Based on the same technical concept, an embodiment of the present invention further provides an electronic device, such as Figure 6 As shown, it includes a processor 601, a communication interface 602, a memory 603 and a communication bus 604, wherein the processor 601, the communication interface 602, and the memory 603 communicate with each other through the communication bus 604.
[0098] Memory 603, used for storing computer programs;
[0099] The processor 601 is used to implement the steps of the data fusion processing method for the power system when executing the program stored in the memory 603.
[0100] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.
[0101] The communication interface is used for communication between the above electronic device and other devices.
[0102] The memory may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.
[0103] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
[0104] In another embodiment of the present invention, a computer-readable storage medium is provided, in which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned data fusion processing methods for power systems are implemented.
[0105] In another embodiment provided by the present invention, a computer program product including instructions is also provided, which, when executed on a computer, enables the computer to execute any of the data fusion processing methods for power systems in the above embodiments.
[0106] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented by software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk Solid State Disk (SSD)), etc.
[0107] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0108] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest range consistent with the principles and novel features applied for herein.
Claims
1. A data fusion processing method for power system, characterized in that: The method comprises: Collect multi-source heterogeneous data from business departments related to the power system based on dispatch query requirements; the multi-source heterogeneous data includes power grid operation data, power transaction data, power marketing data and equipment asset data; Performing a first preprocessing and a second preprocessing on the multi-source heterogeneous data in sequence to obtain preprocessed multi-source heterogeneous data; Extracting n modal features of the preprocessed multi-source heterogeneous data, wherein the n modal features include image features, video features, voice features, and point cloud features; Determine the correlation between n modal features; The n related modal features are respectively input into the corresponding modal network and feature-level fusion network for modal fusion to obtain the corresponding modal decision and fusion decision; Performing decision fusion on the modal decisions output by each modal network and the fusion decision to obtain a final fusion result; The step of inputting the associated n modal features into the respective corresponding modal networks and feature-level fusion networks for modal fusion to obtain the respective corresponding modal decisions and fusion decisions includes: Extracting the noise of the n modal features; The noise of n modal features is input into their respective generators to obtain the corresponding fake sample features; each modal feature and the corresponding fake sample feature are input into the corresponding modal discriminator and fusion discriminator for feature fusion to obtain the corresponding modal decision and fusion decision.
2. A data fusion processing method for power system according to claim 1, characterized in that: The first preprocessing comprises the following steps: Cleaning the multi-source heterogeneous data to obtain cleaned multi-source heterogeneous data; The cleaned multi-source heterogeneous data is subjected to denoising processing to obtain denoised multi-source heterogeneous data.
3. The data fusion processing method for power system according to claim 1, characterized in that: The second preprocessing comprises the following steps: Classifying the multi-source heterogeneous data that has undergone the first preprocessing to obtain structured data and unstructured data; The unstructured data is converted into structured data and stored in a structured database.
4. The data fusion processing method for power system according to claim 1, characterized in that: Determining the correlation between n modal features includes: Performing feature enhancement processing and denoising processing on the n modes in sequence; The semantic recognition processing is performed on the enhanced and denoised modal features to obtain the correlation between the modal features.
5. The data fusion processing method for power system according to claim 4 is characterized in that: The semantic recognition processing of the enhanced and denoised modal features to obtain the correlation between the modal features includes: Each enhanced and denoised modal feature is input into multiple regular graph convolutions and convolved in sequence; each graph convolution layer contains a graph construction module, a graph convolution module and a feature filtering module; The convolutional features are input into the NLP layer for semantic segmentation to obtain the segmentation score and the global maximum pooling layer for classification to obtain the classification score; The correlation between the features of each modality is determined according to the segmentation score and the classification score.
6. The data fusion processing method for power system according to claim 1, characterized in that: The modality network and the feature-level fusion network both adopt a generative adversarial network structure, including n generators, n modality discriminators and a fusion discriminator. The n associated modality features are respectively input into the respective corresponding modality networks and feature-level fusion networks for modality fusion to obtain the respective corresponding modality decisions and fusion decisions, including: Extracting the noise of the n modal features; Input the noise of n modal features into their respective generators to obtain the corresponding fake sample features; Each modal feature and the corresponding fake sample feature are respectively input into the corresponding modal discriminator and fusion discriminator for feature fusion to obtain the corresponding modal decision and fusion decision.
7. The data fusion processing method for power system according to claim 1, characterized in that: The method further comprises: Obtain the parameter weight values of each modality network and feature-level fusion network; Compare the weight value of each parameter with the preset threshold; Parameters below the preset threshold are deleted.
8. A data fusion processing system for power systems, characterized in that: The system comprises: A collection unit, used to collect multi-source heterogeneous data from business departments related to the power system based on dispatch query requirements; A preprocessing unit, configured to sequentially perform a first preprocessing and a second preprocessing on the multi-source heterogeneous data to obtain preprocessed multi-source heterogeneous data; An extraction unit, used to extract n modal features of the preprocessed multi-source heterogeneous data, wherein the n modal features include image features, video features, voice features, and point cloud features; A determination unit, used to determine the correlation between n modal features; A fusion unit is used to input the associated n modal features into the respective corresponding modal networks and feature-level fusion networks for modal fusion to obtain the respective corresponding modal decisions and fusion decisions; and to perform decision fusion on the modal decisions and fusion decisions output by each modal network to obtain a final fusion result; The step of inputting the associated n modal features into the respective corresponding modal networks and feature-level fusion networks for modal fusion to obtain the respective corresponding modal decisions and fusion decisions includes: Extracting the noise of the n modal features; Input the noise of n modal features into their respective generators to obtain the corresponding fake sample features; Each modal feature and the corresponding fake sample feature are respectively input into the corresponding modal discriminator and fusion discriminator for feature fusion to obtain the corresponding modal decision and fusion decision.
9. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, for implementing the method steps described in any one of claims 1 to 7 when executing a program stored in a memory.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method steps described in any one of claims 1 to 7 are implemented.
Citation Information
Patent Citations
Multi-source heterogeneous data fusion method and apparatus
CN107545046A
Multi-modal fusion method and device based on normalized mutual information, medium and equipment
CN111461176A
Self-adaptive depth coupling convolution self-encoding multi-modal data fusion method
CN112528548A
Deep adversarial multi-modal data clustering method
CN113094566A
Multi-modal, multi-disciplinary feature discovery to detect cyber threats in electric power grid
US20180262525A1
