Power distribution network self-healing control method and device, computer device and readable storage medium
By collecting and fusing multimodal data, an optimal self-healing control strategy generation model was constructed, realizing fault self-healing control of the distribution network. This solved the problem of decreased fault identification accuracy caused by distributed new energy access, and improved the self-healing capability and power supply reliability of the distribution network.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU POWER SUPPLY BUREAU GUANGDONG POWER GRID CO LTD
- Filing Date
- 2026-02-28
- Publication Date
- 2026-07-14
AI Technical Summary
The integration of distributed renewable energy sources into the power distribution network leads to power flow distortion and voltage exceeding limits, increasing the probability of fault occurrence. Traditional fault diagnosis methods are difficult to adapt, and the accuracy of fault identification decreases. How can we achieve fault self-healing control?
Multimodal data is collected, features are extracted and aligned and fused, and an optimal self-healing control strategy generation model is constructed. This generates a highly adaptable and reliable self-healing control strategy, which is then used to control the distribution network to perform fault self-healing operations by executing commands.
It significantly improves the ability to accurately identify fault conditions, quickly adapts to topology changes, shortens fault handling time, reduces reliance on manual intervention, and enhances self-healing capabilities and power supply reliability.
Smart Images
Figure CN122394071A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of power distribution automation technology, and in particular to a power distribution network self-healing control method, device, computer equipment, computer-readable storage medium, and computer program product. Background Technology
[0002] In recent years, distributed renewable energy sources such as photovoltaics and wind power have achieved large-scale grid connection due to their clean and low-carbon advantages, driving the transformation of the power supply structure of distribution networks from a traditional single power source to a multi-power source structure. Accompanying this transformation, the operating characteristics of distribution networks exhibit complex features such as strong randomness, high volatility, and dynamic topological changes.
[0003] As the final link connecting power sources and users in the power system, the power supply reliability of the distribution network is directly related to the normal operation of social production and life. However, the random fluctuations in the output of distributed renewable energy sources can easily lead to power flow distortion and voltage exceeding limits, thereby increasing the probability of line short circuits and equipment overload faults. At the same time, the integration of renewable energy sources changes the fault current distribution characteristics of the distribution network, making traditional fault diagnosis methods based on single electrical quantities difficult to adapt to this change, resulting in a decrease in fault identification accuracy.
[0004] Therefore, how to control the power distribution network to achieve fault self-healing is an urgent problem to be solved. Summary of the Invention
[0005] Therefore, it is necessary to provide a distribution network self-healing control method, device, computer equipment, computer-readable storage medium, and computer program product that can control the distribution network to achieve fault self-healing in order to address the above-mentioned technical problems.
[0006] In a first aspect, this application provides a self-healing control method for a power distribution network, comprising:
[0007] Collect multimodal data of the target power distribution network;
[0008] Feature extraction is performed on the multimodal data to obtain the multimodal features of the target distribution network; the multimodal features are aligned and fused to obtain the comprehensive feature vector of the target distribution network.
[0009] Construct an optimal self-healing control strategy generation model; input the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy for the target distribution network;
[0010] The optimal self-healing control strategy of the distribution network is converted into an execution command; the execution command is sent to the target distribution network; the execution command is used to control the target distribution network to perform fault self-healing operation.
[0011] In one embodiment, multimodal data of the target distribution network is collected, including:
[0012] Multimodal data of the target distribution network are collected using sensors, smart terminals, image acquisition equipment, and new energy interfaces deployed on the target distribution network.
[0013] In one embodiment, the multimodal data includes electrical quantity data, time-series operation data, equipment status image data, and distributed renewable energy output data;
[0014] Multimodal features include spatial features, time-dependent features, and statistical features;
[0015] Feature extraction is performed on the multimodal data to obtain the multimodal features of the target distribution network, including:
[0016] Feature extraction is performed on electrical quantity data to obtain the statistical characteristics of the target distribution network;
[0017] Feature extraction is performed on time-series operational data and distributed renewable energy output data to obtain the time-dependent features of the target distribution network.
[0018] Feature extraction is performed on the equipment status image data to obtain the spatial characteristics of the target power distribution network.
[0019] In one embodiment, before performing feature extraction on the electrical quantity data to obtain the statistical characteristics of the target distribution network, the method further includes:
[0020] The electrical quantity data, time-series operation data, equipment status image data, and distributed new energy output data are respectively subjected to noise reduction processing, dimension unification processing, background removal processing, and missing value filling processing.
[0021] In one embodiment, constructing an optimal self-healing control strategy generation model includes:
[0022] Define the state space and action space; the state space includes the distribution network topology, fault information, real-time load distribution, and predicted output of new energy sources; the action space includes the opening and closing actions of smart switches, the adjustment of output of distributed new energy sources, and the load transfer path.
[0023] A multi-objective reward function is constructed with the objectives of minimizing fault isolation time, maximizing power restoration coverage, and minimizing fluctuations in renewable energy output.
[0024] Based on the state space, action space, and multi-objective reward function, the model to be trained is iteratively trained to generate the optimal self-healing control strategy generation model.
[0025] In one embodiment, the method further includes:
[0026] By inputting the comprehensive feature vector into the pre-trained distribution network fault diagnosis model, the fault type, fault location, and fault severity level of the target distribution network can be obtained.
[0027] Secondly, this application also provides a power distribution network self-healing control device, comprising:
[0028] The acquisition module is used to acquire multimodal data of the target power distribution network;
[0029] The extraction module is used to extract features from multimodal data to obtain the multimodal features of the target distribution network; and to align and fuse the multimodal features to obtain the comprehensive feature vector of the target distribution network.
[0030] The module is used to build an optimal self-healing control strategy generation model; the comprehensive feature vector is input into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy of the target distribution network.
[0031] The execution module is used to convert the optimal self-healing control strategy of the distribution network into execution instructions; to send the execution instructions to the target distribution network; and to control the target distribution network to perform fault self-healing operations.
[0032] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:
[0033] Collect multimodal data of the target power distribution network;
[0034] Feature extraction is performed on the multimodal data to obtain the multimodal features of the target distribution network; the multimodal features are aligned and fused to obtain the comprehensive feature vector of the target distribution network.
[0035] Construct an optimal self-healing control strategy generation model; input the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy for the target distribution network;
[0036] The optimal self-healing control strategy of the distribution network is converted into an execution command; the execution command is sent to the target distribution network; the execution command is used to control the target distribution network to perform fault self-healing operation.
[0037] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:
[0038] Collect multimodal data of the target power distribution network;
[0039] Feature extraction is performed on the multimodal data to obtain the multimodal features of the target distribution network; the multimodal features are aligned and fused to obtain the comprehensive feature vector of the target distribution network.
[0040] Construct an optimal self-healing control strategy generation model; input the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy for the target distribution network;
[0041] The optimal self-healing control strategy of the distribution network is converted into an execution command; the execution command is sent to the target distribution network; the execution command is used to control the target distribution network to perform fault self-healing operation.
[0042] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:
[0043] Collect multimodal data of the target power distribution network;
[0044] Feature extraction is performed on the multimodal data to obtain the multimodal features of the target distribution network; the multimodal features are aligned and fused to obtain the comprehensive feature vector of the target distribution network.
[0045] Construct an optimal self-healing control strategy generation model; input the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy for the target distribution network;
[0046] The optimal self-healing control strategy of the distribution network is converted into an execution command; the execution command is sent to the target distribution network; the execution command is used to control the target distribution network to perform fault self-healing operation.
[0047] The aforementioned distribution network self-healing control method, device, computer equipment, computer-readable storage medium, and computer program product collect multimodal data of the target distribution network; extract features from the multimodal data to obtain multimodal features of the target distribution network; and align and fuse the multimodal features to obtain a comprehensive feature vector of the target distribution network. In the process of generating the comprehensive feature vector, the fusion perception and feature alignment of multimodal data significantly improve the accurate identification capability of distribution network fault states under complex operating environments. By constructing an optimal self-healing control strategy generation model and inputting the comprehensive feature vector into it, the system can quickly adapt to dynamic topology changes caused by distributed new energy access, outputting an optimal self-healing control strategy for the target distribution network with strong adaptability and high reliability. The optimal self-healing control strategy of the distribution network is converted into execution instructions; these instructions are then sent to the target distribution network; and the execution instructions are used to control the target distribution network to perform fault self-healing operations. The execution instructions generated by the above process can control the distribution network to achieve fault self-healing, realizing an end-to-end closed loop from fault perception to control response, effectively shortening the fault handling time of the distribution network, reducing reliance on manual intervention, and thus comprehensively improving the self-healing capability and power supply reliability of the distribution network. Attached Figure Description
[0048] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0049] Figure 1 This is an application environment diagram of the power distribution network self-healing control method in one embodiment;
[0050] Figure 2 This is a flowchart illustrating a self-healing control method for a power distribution network in one embodiment.
[0051] Figure 3 This is a flowchart illustrating the self-healing control method for a power distribution network in another embodiment;
[0052] Figure 4 This is a structural block diagram of a power distribution network self-healing control device in one embodiment;
[0053] Figure 5 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation
[0054] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0055] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.
[0056] The self-healing control method for power distribution networks provided in this application can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or placed on a cloud or other network server. Specifically, terminal 102 or server 104 executes a distribution network self-healing control method, which includes: collecting multimodal data of the target distribution network; extracting features from the multimodal data to obtain multimodal features of the target distribution network; aligning and fusing the multimodal features to obtain a comprehensive feature vector of the target distribution network; constructing an optimal self-healing control strategy generation model; inputting the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy of the target distribution network; converting the optimal self-healing control strategy of the distribution network into execution instructions; issuing the execution instructions to the target distribution network; and using the execution instructions to control the target distribution network to perform fault self-healing operations.
[0057] Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, and projection equipment. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted displays. Head-mounted displays can be virtual reality (VR) devices, augmented reality (AR) devices, and smart glasses. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.
[0058] In one exemplary embodiment, such as Figure 2 As shown, a self-healing control method for a distribution network is provided, which is applied to... Figure 1 Taking terminal 102 as an example, the explanation includes the following steps 202 to 208. Wherein:
[0059] Step 202: Collect multimodal data of the target distribution network.
[0060] The multimodal data includes electrical quantity data, time-series operation data, equipment status image data, and distributed renewable energy output data.
[0061] Optionally, data acquisition equipment deployed around the target distribution network can be used to collect multimodal data of the target distribution network.
[0062] Step 204: Extract features from the multimodal data to obtain the multimodal features of the target distribution network; align and fuse the multimodal features to obtain the comprehensive feature vector of the target distribution network.
[0063] Multimodal features include spatial features, time-dependent features, and statistical features.
[0064] For example, a cost matrix is constructed based on the cosine distance; based on the cost matrix, the multimodal features are aligned and fused to obtain the comprehensive feature vector of the target distribution network.
[0065] Alternatively, the formula for constructing the cost matrix is:
[0066]
[0067] in, For the first The single-modal feature and the first Alignment cost of a single modal feature and For feature vectors of different modalities, Let be the cosine similarity between the two feature vectors.
[0068] Optionally, the mathematical expression for the comprehensive feature vector is:
[0069]
[0070] in, For the comprehensive feature vector, For the optimal transfer matrix, , and The weighting coefficients for image features, time-series features, and electrical quantity features are respectively, satisfying... , , and These are three types of single-modal feature matrices, "" indicates the process of vertically concatenating the feature matrices.
[0071] Step 206: Construct the optimal self-healing control strategy generation model; input the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy of the target distribution network.
[0072] Among them, the optimal self-healing control strategy generation model is a reinforcement learning model.
[0073] For example, a state space and an action space are defined; a multi-objective reward function is constructed; and the model to be trained is iteratively trained based on the state space, action space, and multi-objective reward function to obtain the optimal self-healing control strategy generation model.
[0074] Step 208: Convert the optimal self-healing control strategy of the distribution network into an execution command; send the execution command to the target distribution network; the execution command is used to control the target distribution network to perform fault self-healing operation.
[0075] Among them, the fault self-healing operation includes fault isolation, load transfer and new energy output adjustment operations.
[0076] For example, the execution instructions are sent to the execution module of the target distribution network. The execution module includes an intelligent vacuum circuit breaker, a load transfer controller, and a new energy inverter controller.
[0077] In the aforementioned distribution network self-healing control method, multimodal data of the target distribution network is collected; features are extracted from the multimodal data to obtain multimodal features of the target distribution network; and the multimodal features are aligned and fused to obtain a comprehensive feature vector of the target distribution network. In the process of generating the comprehensive feature vector, the fusion perception and feature alignment of multimodal data significantly improve the accuracy of distribution network fault status identification under complex operating environments. By constructing an optimal self-healing control strategy generation model and inputting the comprehensive feature vector into it, the method can quickly adapt to dynamic topology changes caused by distributed renewable energy access, outputting an optimal self-healing control strategy for the target distribution network with strong adaptability and high reliability. The optimal self-healing control strategy of the distribution network is converted into execution instructions; these instructions are then sent to the target distribution network to control it to perform fault self-healing operations. The execution instructions generated in the above process can control the distribution network to achieve fault self-healing, realizing an end-to-end closed loop from fault perception to control response, effectively shortening the distribution network fault handling time, reducing reliance on manual intervention, and thus comprehensively improving the self-healing capability and power supply reliability of the distribution network.
[0078] In one exemplary embodiment, collecting multimodal data of the target distribution network includes: collecting multimodal data of the target distribution network using sensors, smart terminals, image acquisition devices and new energy interfaces deployed on the target distribution network.
[0079] Optionally, the sensor can be a voltage or current sensor. The smart terminal can be a smart meter. The image acquisition device can be an infrared imager. The new energy interface can be a data interface of a new energy inverter.
[0080] In this embodiment, by deploying voltage and current sensors, smart meters, infrared imagers, and data interfaces of renewable energy inverters at various stages of the distribution network, comprehensive perception and multi-dimensional data acquisition of electrical and non-electrical quantities and equipment operating status of the distribution network are achieved. The multimodal data acquired through the above process covers not only the real-time operating parameters of the traditional power grid but also the output information of distributed renewable energy and the health status of key equipment. This provides a rich and accurate data foundation for subsequent fault diagnosis and self-healing decisions, significantly improving the comprehensiveness and precision of the distribution network's operational status perception.
[0081] In one embodiment, the multimodal data includes electrical quantity data, time-series operation data, equipment status image data, and distributed renewable energy output data; the multimodal features include spatial features, time-dependent features, and statistical features; feature extraction is performed on the multimodal data to obtain the multimodal features of the target distribution network, including: feature extraction of electrical quantity data to obtain the statistical features of the target distribution network; feature extraction of time-series operation data and distributed renewable energy output data to obtain the time-dependent features of the target distribution network; and feature extraction of equipment status image data to obtain the spatial features of the target distribution network.
[0082] For example, a convolutional neural network (CNN) is used to extract spatial features of equipment status image data; a long short-term memory network (LSTM) is used to extract time-dependent features of time-series operation data and new energy output data; and a support vector machine (SVM) is used to extract statistical features of electrical quantity data.
[0083] In this embodiment, by designing differentiated feature extraction strategies for different types of features, spatial features and time-dependent features are extracted simultaneously, improving the ability to identify complex faults. This effectively solves the problem of incomplete information and insufficient representation capabilities from a single data source, laying a high-quality data foundation for the subsequent deep fusion of multimodal features, and significantly improving the perception accuracy and fault representation capabilities of complex power distribution network operation status.
[0084] In one embodiment, before extracting features from electrical quantity data to obtain the statistical features of the target distribution network, the method further includes performing noise reduction processing, dimension unification processing, background removal processing, and missing value imputation processing on the electrical quantity data, time-series operation data, equipment status image data, and distributed renewable energy output data, respectively.
[0085] For example, wavelet transform is used to perform noise reduction on electrical quantity data. The mathematical expression for wavelet transform noise reduction is:
[0086]
[0087] in, The original electrical quantity signal, For wavelet basis functions, This is a scaling parameter used to control the scaling of the wavelet. These are translation parameters used to control the translation of the wavelet. These are the wavelet transform coefficients.
[0088] For example, Z-score normalization is used to unify the dimensions of time series data; the mathematical expression for Z-score normalization is:
[0089]
[0090] in, This is the original time series data. The mean of the time series data. The standard deviation of the time series data. This is standardized time-series data.
[0091] For example, adaptive threshold segmentation is used to remove background from device status image data, and the K-nearest neighbor algorithm is used to fill in missing values in various types of data.
[0092] In this embodiment, by performing noise reduction, dimension unification, background removal, and missing value filling operations on electrical quantity data, time-series operation data, equipment status image data, and distributed new energy output data respectively according to their data characteristics, high-quality cleaning and standardization alignment of multimodal data can be achieved, effectively eliminating noise interference, scale differences, and information loss in the original data.
[0093] In one embodiment, constructing an optimal self-healing control strategy generation model includes: defining a state space and an action space; the state space includes the distribution network topology, fault information, real-time load distribution, and predicted renewable energy output; the action space includes intelligent switch opening and closing actions, distributed renewable energy output adjustment amounts, and load transfer paths; constructing a multi-objective reward function with the objectives of minimizing fault isolation time, maximizing power restoration coverage, and minimizing renewable energy output fluctuations; and iteratively training the model to be trained based on the state space, action space, and multi-objective reward function to generate the optimal self-healing control strategy generation model.
[0094] The multi-objective reward function is as follows:
[0095]
[0096] in, For instant reward value, , and The weighting coefficients are and satisfy the following conditions: , For fault isolation time, To restore power coverage, Fluctuations in the output of new energy sources.
[0097] For example, a state space and an action space are defined; the state space includes the distribution network topology, fault information, real-time load distribution, and predicted renewable energy output; the action space includes intelligent switch opening and closing actions, distributed renewable energy output adjustment amounts, and load transfer paths; a multi-objective reward function is constructed with the objectives of minimizing fault isolation time, maximizing power restoration coverage, and minimizing renewable energy output fluctuations; the model to be trained executes actions in the action space to change the state of the state space; the multi-objective reward function is used to evaluate the actions to obtain evaluation results; the model to be trained is iteratively trained based on the evaluation results to generate an optimal self-healing control strategy generation model.
[0098] In this embodiment, by clearly defining the state space and action space, and constructing a multi-objective reward function with the objectives of minimizing fault isolation time, maximizing power restoration coverage, and minimizing renewable energy output fluctuations, refined modeling and multi-objective optimization of the distribution network self-healing control strategy are achieved. Specifically, during training, the model continuously explores control actions such as switch opening and closing, renewable energy output adjustment, and load transfer in the action space, observes changes in topology, fault information, load distribution, and renewable energy prediction values in the state space, and evaluates the quality of actions based on the feedback from the reward function. Through multiple rounds of iterative training and strategy optimization, the model can autonomously learn a coordinated control strategy that balances rapid fault isolation, maximizing power restoration, and mitigating renewable energy fluctuations in complex operating environments, ultimately generating an optimal self-healing control strategy generation model with strong adaptability and robustness. This model effectively addresses the shortcomings of traditional methods in multi-objective trade-offs and dynamic adaptation, significantly improving the self-healing capability and operational economy of the distribution network in the context of distributed renewable energy access.
[0099] In one embodiment, the method further includes: inputting the comprehensive feature vector into a pre-trained distribution network fault diagnosis model to obtain the fault type, fault location, and fault severity level of the target distribution network.
[0100] Among them, the pre-trained distribution network fault diagnosis model is an optimized CNN-LSTM hybrid model.
[0101] For example, real-time data collection of power distribution network operation after operation is used to feed back data preprocessing and fault diagnosis steps, and to dynamically optimize parameters and decision-making strategies.
[0102] In this embodiment, the integrated feature vector that integrates multi-source information is input into the optimized CNN-LSTM hybrid model. By making full use of the synergistic advantages of CNN in spatial feature extraction and LSTM in temporal dependency modeling, a multi-dimensional and accurate diagnosis of the fault status of the distribution network is achieved, and three key diagnostic results are output: fault type, fault location and fault severity level.
[0103] Next reference Figure 3The present application provides a method for self-healing control of a power distribution network, with a specific embodiment as an example.
[0104] Step 1: Collect multi-modal data from multiple sources in the target distribution network. Multi-modal data of the target distribution network is collected using sensors, smart terminals, image acquisition equipment, and renewable energy interfaces deployed on the network. This multi-modal data includes electrical quantity data, time-series operational data, equipment status image data, and distributed renewable energy output data.
[0105] Step 2: Perform data preprocessing on the multimodal data. Noise reduction, dimension unification, background removal, and missing value imputation operations are performed on the electrical quantity data, time-series operation data, equipment status image data, and distributed renewable energy output data, respectively.
[0106] Step 3: Extract multimodal features, fuse the multimodal features, and perform fault diagnosis. Feature extraction is performed on electrical quantity data to obtain the statistical features of the target distribution network; feature extraction is performed on time-series operational data and distributed renewable energy output data to obtain the time-dependent features of the target distribution network; feature extraction is performed on equipment status image data to obtain the spatial features of the target distribution network. The multimodal features are aligned and fused to obtain the comprehensive feature vector of the target distribution network. The comprehensive feature vector is input into a pre-trained distribution network fault diagnosis model to obtain the fault type, fault location, and fault severity level of the target distribution network.
[0107] Step 4: Generate the optimal self-healing control strategy for the target distribution network. Define the state space and action space; the state space includes the distribution network topology, fault information, real-time load distribution, and predicted renewable energy output; the action space includes intelligent switch opening and closing actions, distributed renewable energy output adjustments, and load transfer paths; construct a multi-objective reward function with the objectives of minimizing fault isolation time, maximizing power restoration coverage, and minimizing renewable energy output fluctuations; based on the state space, action space, and multi-objective reward function, iteratively train the model to be trained to generate the optimal self-healing control strategy generation model. Input the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy for the target distribution network.
[0108] Step 5: Generate execution instructions to control the target distribution network to perform fault self-healing operations. The optimal self-healing control strategy of the distribution network is converted into execution instructions; these instructions are then sent to the target distribution network; and the execution instructions are used to control the target distribution network to perform fault self-healing operations.
[0109] Step 6: Collect the distribution network operation data after the operation in real time and feed it back to Step 2 or Step 3 for dynamic parameter optimization.
[0110] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.
[0111] Based on the same inventive concept, this application also provides a distribution network self-healing control device for implementing the aforementioned distribution network self-healing control method. The solution provided by this device is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more embodiments of the distribution network self-healing control device provided below can be found in the limitations of the distribution network self-healing control method described above, and will not be repeated here.
[0112] In one exemplary embodiment, such as Figure 4 As shown, a power distribution network self-healing control device 400 is provided, including: a data acquisition module 402, an extraction module 404, a construction module 406, and an execution module 408, wherein:
[0113] The acquisition module 402 is used to acquire multimodal data of the target power distribution network.
[0114] The extraction module 404 is used to extract features from multimodal data to obtain multimodal features of the target distribution network; and to align and fuse the multimodal features to obtain the comprehensive feature vector of the target distribution network.
[0115] Module 406 is used to construct the optimal self-healing control strategy generation model; the comprehensive feature vector is input into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy of the target distribution network.
[0116] The execution module 408 is used to convert the optimal self-healing control strategy of the distribution network into execution instructions; to send the execution instructions to the target distribution network; and to control the target distribution network to perform fault self-healing operations.
[0117] In one embodiment, the acquisition module is also used to acquire multimodal data of the target distribution network using sensors, smart terminals, image acquisition devices and new energy interfaces deployed on the target distribution network.
[0118] In one embodiment, the extraction module is further configured to perform feature extraction on electrical quantity data to obtain the statistical characteristics of the target distribution network; perform feature extraction on time-series operation data and distributed renewable energy output data to obtain the time-dependent characteristics of the target distribution network; and perform feature extraction on equipment status image data to obtain the spatial characteristics of the target distribution network.
[0119] In one embodiment, the extraction module is further configured to perform noise reduction processing, dimension unification processing, background removal processing, and missing value filling processing on electrical quantity data, time-series operation data, equipment status image data, and distributed new energy output data, respectively.
[0120] In one embodiment, the construction module is also used to define a state space and an action space; the state space includes the distribution network topology, fault information, real-time load distribution, and predicted output of renewable energy; the action space includes the opening and closing actions of intelligent switches, the adjustment amount of distributed renewable energy output, and the load transfer path; a multi-objective reward function is constructed with the objectives of minimizing fault isolation time, maximizing power restoration coverage, and minimizing renewable energy output fluctuation; based on the state space, action space, and multi-objective reward function, the model to be trained is iteratively trained to generate the optimal self-healing control strategy generation model.
[0121] In one embodiment, the distribution network self-healing control device further includes a fault diagnosis module, which is used to input the comprehensive feature vector into a pre-trained distribution network fault diagnosis model to obtain the fault type, fault location and fault severity level of the target distribution network.
[0122] Each module in the aforementioned power distribution network self-healing control device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.
[0123] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 5As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores multimodal data of the target power distribution network. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a power distribution network self-healing control method.
[0124] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0125] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:
[0126] Collect multimodal data of the target power distribution network;
[0127] Feature extraction is performed on the multimodal data to obtain the multimodal features of the target distribution network; the multimodal features are aligned and fused to obtain the comprehensive feature vector of the target distribution network.
[0128] Construct an optimal self-healing control strategy generation model; input the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy for the target distribution network;
[0129] The optimal self-healing control strategy of the distribution network is converted into an execution command; the execution command is sent to the target distribution network; the execution command is used to control the target distribution network to perform fault self-healing operation.
[0130] In one embodiment, when the processor executes the computer program, it also performs the following steps: collecting multimodal data of the target distribution network using sensors, smart terminals, image acquisition devices, and new energy interfaces deployed on the target distribution network.
[0131] In one embodiment, when the processor executes the computer program, it further performs the following steps: extracting features from electrical quantity data to obtain statistical features of the target distribution network; extracting features from time-series operation data and distributed renewable energy output data to obtain time-dependent features of the target distribution network; and extracting features from equipment status image data to obtain spatial features of the target distribution network.
[0132] In one embodiment, when the processor executes the computer program, it further performs the following steps: performing noise reduction processing, dimension unification processing, background removal processing, and missing value filling processing on electrical quantity data, time-series operation data, equipment status image data, and distributed new energy output data, respectively.
[0133] In one embodiment, when the processor executes the computer program, it also performs the following steps: defining a state space and an action space; the state space includes the distribution network topology, fault information, real-time load distribution, and predicted output of new energy sources; the action space includes the opening and closing actions of intelligent switches, the adjustment amount of distributed new energy output, and the load transfer path; constructing a multi-objective reward function with the objectives of minimizing fault isolation time, maximizing power restoration coverage, and minimizing fluctuations in new energy output; and iteratively training the model to be trained based on the state space, action space, and multi-objective reward function to generate an optimal self-healing control strategy generation model.
[0134] In one embodiment, when the processor executes the computer program, it also performs the following steps: inputting the comprehensive feature vector into a pre-trained distribution network fault diagnosis model to obtain the fault type, fault location, and fault severity level of the target distribution network.
[0135] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:
[0136] Collect multimodal data of the target power distribution network;
[0137] Feature extraction is performed on the multimodal data to obtain the multimodal features of the target distribution network; the multimodal features are aligned and fused to obtain the comprehensive feature vector of the target distribution network.
[0138] Construct an optimal self-healing control strategy generation model; input the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy for the target distribution network;
[0139] The optimal self-healing control strategy of the distribution network is converted into an execution command; the execution command is sent to the target distribution network; the execution command is used to control the target distribution network to perform fault self-healing operation.
[0140] In one embodiment, when the computer program is executed by the processor, it also performs the following steps: collecting multimodal data of the target distribution network using sensors, smart terminals, image acquisition devices, and new energy interfaces deployed on the target distribution network.
[0141] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: extracting features from electrical quantity data to obtain statistical features of the target distribution network; extracting features from time-series operation data and distributed renewable energy output data to obtain time-dependent features of the target distribution network; and extracting features from equipment status image data to obtain spatial features of the target distribution network.
[0142] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: performing noise reduction processing, dimension unification processing, background removal processing, and missing value filling processing on electrical quantity data, time-series operation data, equipment status image data, and distributed new energy output data, respectively.
[0143] In one embodiment, when the computer program is executed by the processor, it also performs the following steps: defining a state space and an action space; the state space includes the distribution network topology, fault information, real-time load distribution, and predicted output of new energy sources; the action space includes the opening and closing actions of intelligent switches, the adjustment amount of distributed new energy output, and the load transfer path; constructing a multi-objective reward function with the objectives of minimizing fault isolation time, maximizing power restoration coverage, and minimizing fluctuations in new energy output; and iteratively training the model to be trained based on the state space, action space, and multi-objective reward function to generate an optimal self-healing control strategy generation model.
[0144] In one embodiment, when the computer program is executed by the processor, it also performs the following steps: inputting the comprehensive feature vector into a pre-trained distribution network fault diagnosis model to obtain the fault type, fault location, and fault severity level of the target distribution network.
[0145] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:
[0146] Collect multimodal data of the target power distribution network;
[0147] Feature extraction is performed on the multimodal data to obtain the multimodal features of the target distribution network; the multimodal features are aligned and fused to obtain the comprehensive feature vector of the target distribution network.
[0148] Construct an optimal self-healing control strategy generation model; input the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy for the target distribution network;
[0149] The optimal self-healing control strategy of the distribution network is converted into an execution command; the execution command is sent to the target distribution network; the execution command is used to control the target distribution network to perform fault self-healing operation.
[0150] In one embodiment, when the computer program is executed by the processor, it also performs the following steps: collecting multimodal data of the target distribution network using sensors, smart terminals, image acquisition devices, and new energy interfaces deployed on the target distribution network.
[0151] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: extracting features from electrical quantity data to obtain statistical features of the target distribution network; extracting features from time-series operation data and distributed renewable energy output data to obtain time-dependent features of the target distribution network; and extracting features from equipment status image data to obtain spatial features of the target distribution network.
[0152] In one embodiment, when the computer program is executed by the processor, it further performs the following steps: performing noise reduction processing, dimension unification processing, background removal processing, and missing value filling processing on electrical quantity data, time-series operation data, equipment status image data, and distributed new energy output data, respectively.
[0153] In one embodiment, when the computer program is executed by the processor, it also performs the following steps: defining a state space and an action space; the state space includes the distribution network topology, fault information, real-time load distribution, and predicted output of new energy sources; the action space includes the opening and closing actions of intelligent switches, the adjustment amount of distributed new energy output, and the load transfer path; constructing a multi-objective reward function with the objectives of minimizing fault isolation time, maximizing power restoration coverage, and minimizing fluctuations in new energy output; and iteratively training the model to be trained based on the state space, action space, and multi-objective reward function to generate an optimal self-healing control strategy generation model.
[0154] In one embodiment, when the computer program is executed by the processor, it also performs the following steps: inputting the comprehensive feature vector into a pre-trained distribution network fault diagnosis model to obtain the fault type, fault location, and fault severity level of the target distribution network.
[0155] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.
[0156] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.
[0157] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.
[0158] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A self-healing control method for a power distribution network, characterized in that, The method includes: Collect multimodal data of the target power distribution network; Feature extraction is performed on the multimodal data to obtain the multimodal features of the target distribution network; the multimodal features are aligned and fused to obtain the comprehensive feature vector of the target distribution network. Construct an optimal self-healing control strategy generation model; input the comprehensive feature vector into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy of the target distribution network; The optimal self-healing control strategy of the distribution network is converted into an execution instruction; the execution instruction is sent to the target distribution network; the execution instruction is used to control the target distribution network to perform fault self-healing operation.
2. The method according to claim 1, characterized in that, The multimodal data of the target distribution network to be collected includes: Multimodal data of the target distribution network are collected using sensors, smart terminals, image acquisition equipment, and new energy interfaces deployed on the target distribution network.
3. The method according to claim 1, characterized in that, The multimodal data includes electrical quantity data, time-series operation data, equipment status image data, and distributed new energy output data; The multimodal features include spatial features, temporal dependent features, and statistical features; The step of extracting features from the multimodal data to obtain the multimodal features of the target distribution network includes: Feature extraction is performed on the electrical quantity data to obtain the statistical characteristics of the target distribution network; Feature extraction is performed on the time-series operational data and the distributed renewable energy output data to obtain the time-dependent features of the target distribution network. Feature extraction is performed on the equipment status image data to obtain the spatial features of the target power distribution network.
4. The method according to claim 3, characterized in that, Before extracting features from the electrical quantity data to obtain the statistical characteristics of the target distribution network, the method further includes: The electrical quantity data, the time-series operation data, the equipment status image data, and the distributed new energy output data are respectively subjected to noise reduction processing, dimension unification processing, background removal processing, and missing value filling processing.
5. The method according to claim 1, characterized in that, The construction of the optimal self-healing control strategy generation model includes: Define a state space and an action space; the state space includes the distribution network topology, fault information, real-time load distribution, and predicted output of new energy sources; the action space includes the opening and closing actions of intelligent switches, the adjustment amount of distributed new energy output, and the load transfer path. A multi-objective reward function is constructed with the objectives of minimizing fault isolation time, maximizing power restoration coverage, and minimizing fluctuations in renewable energy output. Based on the state space, the action space, and the multi-objective reward function, the model to be trained is iteratively trained to generate the optimal self-healing control strategy generation model.
6. The method according to claim 1, characterized in that, The method further includes: The comprehensive feature vector is input into a pre-trained distribution network fault diagnosis model to obtain the fault type, fault location, and fault severity level of the target distribution network.
7. A power distribution network self-healing control device, characterized in that, The device includes: The acquisition module is used to acquire multimodal data of the target power distribution network; An extraction module is used to extract features from the multimodal data to obtain the multimodal features of the target distribution network; and to align and fuse the multimodal features to obtain the comprehensive feature vector of the target distribution network. A construction module is used to construct an optimal self-healing control strategy generation model; the comprehensive feature vector is input into the optimal self-healing control strategy generation model to obtain the optimal self-healing control strategy of the target distribution network; An execution module is used to convert the optimal self-healing control strategy of the distribution network into execution instructions; send the execution instructions to the target distribution network; and control the target distribution network to perform fault self-healing operations.
8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.
10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.