Power distribution network state detection method and device and nonvolatile storage medium

Abstract

The invention discloses a power distribution network state detection method and device and a nonvolatile storage medium. The method comprises the steps that real-time operation information of a target power distribution network is acquired, and the real-time operation information comprises fan real-time generated power, photovoltaic unit real-time generated power and electric vehicle real-time charging data; the real-time operation information is input into a preset target state detection model, a state detection result of the target power distribution network is obtained, the target state detection model is obtained based on training of an initial state detection model, and the initial state detection model is a convolutional neural network model; the state detection result comprises a safety state, a warning state, a general emergency state or a serious emergency state. According to the method, the technical problems that the real-time operation state of the power distribution network is inconvenient to identify and the accuracy is reduced due to the fact that transient stability analysis of the security domain and the stability domain is separated in a traditional method, and the source load fluctuation and the uncertainty thereof cannot be effectively handled are solved.

Description

Technical Field

[0001] This invention relates to the field of power system technology, and more specifically, to a method, device, and non-volatile storage medium for detecting the status of a power distribution network. Background Technology

[0002] As a critical link in the power system, the accurate identification of the distribution network's operating status is crucial for ensuring stable power system operation and improving power supply quality and efficiency. Traditional distribution network status identification methods are typically based on security domain analysis and stability domain analysis, treating these two analytical methods as independent processes. Security domain analysis focuses on assessing the distribution network's operating status before a fault, while stability domain analysis focuses on the system response after a fault occurs. However, distribution network operation is a continuous process; there are no significant time intervals between before, during, and after a fault. This fragmented analysis approach leads to inconsistencies in the judgment of state transitions before and after a fault, thus affecting the accuracy and efficiency of real-time distribution network operating status identification.

[0003] Furthermore, with the widespread integration of distributed energy resources and electric vehicles, the operating environment of power distribution networks has become more complex and volatile. The volatility of source and load increases the difficulty of state identification, especially when the output of wind turbines and photovoltaic units is highly uncertain, and the randomness of electric vehicle charging load is significant. Current technologies struggle to effectively handle and predict these extremely unbalanced operating states, leading to a decline in the accuracy and reliability of state identification.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This invention provides a distribution network status detection method, device, and non-volatile storage medium to at least solve the technical problems of traditional methods that separate transient stability analysis of the security domain and the stability domain, resulting in inconvenience and reduced accuracy in identifying the real-time operating status of the distribution network, and the inability to effectively cope with source load fluctuations and their uncertainties.

[0006] According to one aspect of the present invention, a distribution network status detection method is provided, comprising: acquiring real-time operating information of a target distribution network, wherein the real-time operating information includes real-time power generation of wind turbines, real-time power generation of photovoltaic units, and real-time charging data of electric vehicles, wherein the electric vehicles are located in the target distribution network; inputting the real-time operating information into a preset target status detection model to obtain a status detection result of the target distribution network, wherein the target status detection model is trained based on an initial status detection model, the initial status detection model being a convolutional neural network model, and the status detection result including a safe status, a warning status, a general emergency status, or a severe emergency status.

[0007] Optionally, the target state detection model is trained based on the initial state detection model, including: constructing a first extreme imbalance sample set based on the historical power generation of wind turbines and photovoltaic units through Latin hypercube sampling; constructing a second extreme imbalance sample set based on the historical charging data of electric vehicles through Monte Carlo random sampling; constructing an initial state detection model, wherein the initial state detection model includes an input layer, a hidden layer, and an output layer; determining the training set of the initial state detection model based on the first extreme imbalance sample set and the second extreme imbalance sample set; and training the initial state detection model based on the training set to obtain the target state detection model.

[0008] Optionally, by using Latin hypercube sampling, a first extreme imbalance sample set is constructed based on the historical power generation of wind turbines and photovoltaic units. This includes: determining the power generation parameter space; defining the value range of each power generation parameter based on the historical power generation of wind turbines and photovoltaic units; dividing the power generation parameter space, uniformly dividing the value range of each parameter into multiple equally probable intervals; and randomly selecting a sample point within each of the multiple equally probable intervals to obtain multiple samples, which serve as the first extreme imbalance sample set.

[0009] Optionally, a second extreme imbalance sample set is constructed based on historical charging data of electric vehicles using Monte Carlo random sampling. This includes: determining a probability distribution model of electric vehicle charging behavior based on historical charging data of electric vehicles; randomly selecting a sample point from the probability distribution model and determining whether to accept the sample point; repeating the above sampling and judgment process until the number of accepted sample points exceeds a preset acceptance threshold, resulting in multiple sample points as the second extreme imbalance sample set.

[0010] Optionally, based on the first extreme imbalance sample set and the second extreme imbalance sample set, the training set of the initial state detection model is determined, including: acquiring static information and environmental data of the target distribution network, wherein the static information includes the power grid structure, line parameters, and equipment capacity, and the environmental data includes weather information, production data, and living data; and based on the static information and environmental data, determining samples in the first extreme imbalance sample set and the second extreme imbalance sample set whose correlation with the operation of the target distribution network is higher than a preset threshold, as the training set.

[0011] Optionally, the system further includes: when the status detection result is the safe state, no control operation is performed on the target distribution network; when the status detection result is the alert state, a first control operation is performed on the target distribution network, wherein the first control operation includes load transfer, increasing reserve capacity, and system hardening; when the status detection result is a general emergency state, a second control operation is performed on the target distribution network, wherein the second control operation includes fault isolation, load reduction, and activation of emergency power supply; and when the status detection result is a severe emergency state, a third control operation is performed on the target distribution network, wherein the third control operation includes system disconnection.

[0012] According to another aspect of the present invention, a battery data display device is also provided, comprising: an acquisition module, configured to acquire real-time operating information of a target power distribution network, wherein the real-time operating information includes real-time power generation of wind turbines, real-time power generation of photovoltaic units, and real-time charging data of electric vehicles, wherein the electric vehicles are located in the target power distribution network; and a detection module, configured to input the real-time operating information into a preset target state detection model to obtain a state detection result of the target power distribution network, wherein the target state detection model is trained based on an initial state detection model, the initial state detection model being a convolutional neural network model, and the state detection result including a safe state, a warning state, a general emergency state, or a severe emergency state.

[0013] According to another aspect of the present invention, a non-volatile storage medium is also provided, the non-volatile storage medium including a stored program, wherein, when the program is running, it controls the device where the non-volatile storage medium is located to execute any of the above-described battery data display methods.

[0014] According to another aspect of the present invention, a computer device is also provided, the computer device including a processor for running a program, wherein the program executes any of the above-described battery data display methods during runtime.

[0015] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements an emergency control method for a novel power distribution system under any of the above-described disaster conditions.

[0016] In this embodiment of the invention, a distribution network status detection method is adopted. This method acquires real-time operating information of the target distribution network, including real-time power generation of wind turbines, real-time power generation of photovoltaic units, and real-time charging data of electric vehicles located within the target distribution network. The real-time operating information is input into a preset target status detection model to obtain the status detection results of the target distribution network. The target status detection model is trained based on an initial status detection model, which is a convolutional neural network model. The status detection results include safe status, alert status, general emergency status, or severe emergency status. This achieves the goal of constructing a high-precision distribution network operation status identification system based on a convolutional neural network. This enables real-time and accurate identification of the distribution network's operation under different states, improving the network's flexibility and risk response capabilities. Furthermore, it solves the technical problems of traditional methods that separate transient stability analysis of the safety domain and stability domain, leading to inconvenience and reduced accuracy in real-time distribution network operation status identification, and the inability to effectively address source load fluctuations and their uncertainties. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0018] Figure 1 A hardware structure block diagram of a computer terminal for implementing a power distribution network status detection method is shown.

[0019] Figure 2 This is a flowchart illustrating the power distribution network status detection method provided according to an embodiment of the present invention;

[0020] Figure 3 This is a diagram showing the results of identifying the operating status of a power distribution network using a target detection model provided by an optional embodiment of the present invention.

[0021] Figure 4 This is a diagram illustrating the training process of a power distribution network state detection model according to an optional embodiment of the present invention.

[0022] Figure 5 This is a flowchart of the power distribution status detection model processing according to an optional embodiment of the present invention;

[0023] Figure 6 This is a structural block diagram of a power distribution network status detection device provided according to an embodiment of the present invention. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0025] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0026] According to an embodiment of the present invention, a method for detecting the state of a power distribution network is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0027] The method embodiment provided in Embodiment 1 of this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 A hardware block diagram of a computer terminal for implementing a method for power distribution network status detection is shown. Figure 1 As shown, the computer terminal 10 may include one or more processors (shown as 102a, 102b, ..., 102n in the figure) (the processor may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1The different configurations shown.

[0028] It should be noted that the aforementioned one or more processors and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10. As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0029] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the distribution network status detection method in this embodiment of the invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the above-mentioned distribution network status detection method of the application program. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0030] The display can be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10.

[0031] Figure 2 This is a flowchart illustrating a method for detecting the state of a power distribution network according to an embodiment of the present invention, as shown below. Figure 2 As shown, the method includes the following steps:

[0032] Step S201: Obtain real-time operation information of the target power distribution network, wherein the real-time operation information includes real-time power generation of wind turbines, real-time power generation of photovoltaic units, and real-time charging data of electric vehicles, wherein the electric vehicles are located in the target power distribution network.

[0033] In this step, real-time operational information of the target distribution network is fundamental for state identification. Acquiring this real-time operational information encompasses the dynamics of key energy sources within the distribution network, including real-time power generation from wind turbines, photovoltaic (PV) units, and electric vehicles (EVs) within the network. Specifically, wind turbines and PV units, as crucial components of distributed energy, experience fluctuations in power generation due to weather conditions such as wind speed and solar intensity. Therefore, collecting their real-time power generation data provides timely information on the current output status of renewable energy, offering critical source data for state identification. Real-time charging data for EVs includes charging power, battery capacity, initial charging time, and daily mileage, reflecting dynamic changes on the load side. Given the highly random and unpredictable nature of EV charging behavior, real-time detection of their charging demand is essential for accurately assessing the load status of the distribution network. By integrating this real-time information, a dataset reflecting the current operating status of the distribution network can be constructed. This data provides input for subsequent target state detection models, enabling immediate assessment of the distribution network's operational status and providing real-time, accurate data support for the development of preventative and emergency control strategies.

[0034] Step S202: Input the real-time operation information into the preset target state detection model to obtain the state detection result of the target distribution network. The target state detection model is trained based on the initial state detection model, which is a convolutional neural network model. The state detection result includes safe state, warning state, general emergency state, or severe emergency state.

[0035] In this step, an initial state detection model is trained and optimized using Convolutional Neural Networks (CNNs) in deep learning to form a target state detection model. CNNs can automatically extract features from multi-dimensional input data and are suitable for processing image or time-series data. In this embodiment, it is used to capture and understand complex patterns in distribution network operation information, such as fluctuations in power output and changes in load demand, which directly affect the stability of the power system. Through the CNN training process, the target state detection model can accurately extract key features from the real-time operation data of the target distribution network and identify its real-time state. Finally, the target state detection model can output four possible state detection results: safe state, alert state, general emergency state, or severe emergency state.

[0036] Specifically, a safe state refers to a state in which the system can withstand specified faults without exceeding any constraints; a warning state refers to a state in which the system cannot remain stable when a specified fault event occurs. If the distribution network is in normal operation or a warning state, and if it is subjected to a sufficiently large disturbance, and the system can operate stably after the automatic protection device clears the fault, then the current fault state is a general emergency state; if the system enters an emergency state, and the system becomes unstable after the automatic protection device clears the fault, then the current fault state is a severe emergency state.

[0037] Figure 3 This is a diagram showing the results of the target detection model for identifying the operating status of the power distribution network according to an optional embodiment of the present invention, such as... Figure 3 As shown, the target state detection model achieves identification accuracy rates of 95.74%, 95.34%, and 95.78% in normal, alert, and fault states, respectively, with an average accuracy rate of 95.62%. This result verifies the efficiency and accuracy of the model in identifying the operating state of the target distribution network, indicating that the model can quickly and accurately identify the operating state of the distribution network.

[0038] Through the above steps, the goal of constructing a high-precision distribution network operation status identification system based on convolutional neural networks is achieved. This enables the real-time and accurate identification of the operation status of the distribution network under different conditions, improving the flexibility and risk response capability of the distribution network. Furthermore, it solves the technical problems of traditional methods that separate the transient stability analysis of the security domain and the stability domain, resulting in inconvenience and reduced accuracy in identifying the real-time operation status of the distribution network, and the inability to effectively cope with source load fluctuations and their uncertainties.

[0039] As an optional embodiment, the target state detection model is trained based on the initial state detection model, including: constructing a first extreme imbalance sample set based on the historical power generation of wind turbines and photovoltaic units through Latin hypercube sampling; constructing a second extreme imbalance sample set based on the historical charging data of electric vehicles through Monte Carlo random sampling; constructing an initial state detection model, wherein the initial state detection model includes an input layer, a hidden layer, and an output layer; determining the training set of the initial state detection model based on the first extreme imbalance sample set and the second extreme imbalance sample set; and training the initial state detection model based on the training set to obtain the target state detection model.

[0040] Optionally, firstly, power distribution system operation information is retrieved from the power distribution system simulation database. Using Latin hypercube sampling, a first extreme imbalance sample set is constructed based on the historical power generation of wind turbines and photovoltaic units. This sampling strategy ensures that sample points are evenly distributed across different dimensions in the parameter space, effectively covering various extreme operating conditions even when facing the high dimensionality and complex probability distribution of wind and solar power output, providing comprehensive and balanced data support for model training. Secondly, a second extreme imbalance sample set is generated using Monte Carlo random sampling based on historical electric vehicle charging data. This process simulates the uncertainty of electric vehicle charging behavior, enhancing the model's adaptability to electric vehicle load fluctuations by randomly sampling data from different charging scenarios. Subsequently, an initial state detection model is constructed using a convolutional neural network (CNN) architecture. This model includes an input layer, hidden layers (including convolutional layers, pooling layers, and fully connected layers), and an output layer, aiming to capture and analyze the complex characteristics of the power distribution network's operating state through deep learning technology.

[0041] Specifically, the input layer accepts data input, which is the operation information of the power distribution network. Hidden layers are typically constructed from convolutional layers, pooling layers, and fully connected layers. Specifically, the function of the convolutional layer is to extract features from the input data. The convolutional kernel performs element-wise matrix multiplication and summation of the input features within the receptive field, and then adds the bias.

[0042]

[0043] in, and Indicates the first l +1 layer inputs and outputs; Indicates the first l +1 layer weight coefficient; b This is the deviation amount; The convolution operation between the convolution kernel and the input features, and the bias. b Add them together, and then pass them through a nonlinear excitation function. f ( x The output features are obtained.

[0044] The function of pooling layers is to perform feature selection and information filtering on the features extracted by convolutional layers. The general representation of a pooling model is as follows:

[0045]

[0046] in, For the first l Parameters of the convolutional layer; p These are pre-specified parameters, when p When = 1, it is mean pooling. When the maximum value is used, it is pooled.

[0047] The function of a fully connected layer is to perform non-linear combinations of extracted features to obtain the output, that is, to achieve the learning objective by utilizing existing high-order features.

[0048] Based on the constructed first and second extreme imbalance sample sets, the training set for the initial state detection model can be determined. This training set contains operational information reflecting the distribution network under extreme conditions, providing rich learning resources for model training. Finally, deep learning training is performed by inputting the training set data into the initial state detection model. Figure 4 This is a diagram illustrating the training process of a distribution network state detection model according to an optional embodiment of the present invention, such as... Figure 4 As shown, the performance of the state detection model varies with parameter settings. By continuously adjusting the model parameters until the state detection model achieves the expected classification effect, the optimized target state detection model is obtained. This model can accurately identify the operating state of the distribution network, thus providing strong technical support for real-time monitoring and emergency control of the distribution network.

[0049] As an optional implementation, a first extreme imbalance sample set is constructed using Latin hypercube sampling based on the historical power generation of wind turbines and photovoltaic units. This includes: determining the power generation parameter space; defining the value range of each power generation parameter based on the historical power generation of wind turbines and photovoltaic units; dividing the power generation parameter space, uniformly dividing the value range of each parameter into multiple equally probable intervals; and randomly selecting a sample point within each of the multiple equally probable intervals to obtain multiple samples, which serve as the first extreme imbalance sample set.

[0050] Optionally, Latin hypercube sampling is a method for approximate random sampling from a multivariate parameter distribution, belonging to stratified sampling techniques. The core of this method is ensuring that each dimension is uniformly divided, and that there is exactly one sample point within each division interval.

[0051] Specifically, in Latin hypercube sampling, assuming there are N variables, each variable can be divided into M intervals with equal probability. Then, M sample points satisfying the Latin hypercube condition can be selected. Latin hypercube sampling requires that the number of partitions M for each variable be the same, but it does not require that the sample size M increase in the same way as the number of variables increases. This ensures that the structure of the sample is similar to the structure of the population, improving the accuracy of the estimation. It is particularly suitable for simulations with large parameter spaces or those requiring computationally expensive simulations. The steps of Latin hypercube sampling are as follows:

[0052] First, given a n First, define the power generation parameter space. Second, based on the historical power generation of wind turbines and photovoltaic units, define each dimension. iThe range of values ​​for is [ a i ,b i Subsequently, the power generation parameter space is partitioned. For each parameter, its value range can be divided into M equally probable intervals, where M is the expected number of samples. By partitioning the power generation parameter space, an N×M grid can be obtained, and each interval of each parameter can be regarded as a "layer" of the parameter space. Finally, a point is randomly selected within each sub-interval of each dimension, but it is ensured that the value of each sample in each dimension lies within a different sub-interval. If Indicates the first j The sample at the th i For values ​​in a dimension, select... The method is as follows:

[0053]

[0054] in, rand j ( i ) is a random number uniformly distributed in the interval [0,1]. It is a guarantee Falling in i The first dimension j Integers within a sub-interval. Generated by randomly combining the selected values ​​from each parameter. M Each "sample point" ensures that no two sample points have the same value for any parameter, as if the sample points in each dimension were stretched into the vertices of a hypercube.

[0055] As an optional embodiment, a second extreme imbalance sample set is constructed based on historical charging data of electric vehicles using Monte Carlo random sampling. This includes: determining a probability distribution model of electric vehicle charging behavior based on historical charging data of electric vehicles; randomly selecting a sample point from the probability distribution model and determining whether to accept the sample point; repeating the above sampling and judgment process until the number of accepted sample points exceeds a preset acceptance threshold, resulting in multiple sample points as the second extreme imbalance sample set.

[0056] Optionally, Monte Carlo random sampling is a computational method based on probability and statistics theory. Its core lies in simulating and estimating solutions to complex problems through random sampling. This method differs fundamentally from general numerical computation methods, and it has unique advantages, especially in dealing with problems that are difficult to solve using traditional numerical methods. In Monte Carlo random sampling, if we want to find the probability P(A) of an event A, or the expected value E(X) of a random variable X, we can obtain the frequency of event A or the arithmetic mean of several specific observations of the random variable X through some kind of experiment, thus approximating the solution to the problem.

[0057] Specifically, firstly, based on historical charging data of electric vehicles, the probability distribution of electric vehicle charging behavior is determined as P(x). Secondly, a probability distribution with a probability density function of q(x) is selected as the proposed distribution, satisfying that for any x, c times q(x) is always greater than P(x), where c>0. Subsequently, sample points x' are randomly sampled according to the proposed distribution q(x), and then u is sampled uniformly within the range (0,1). Finally, if... Then x' If the sampling result is not obtained, the sampling and judgment process continues. The above sampling and judgment process is repeated until the number of accepted sample points exceeds the preset acceptance threshold, resulting in multiple sample points x', which serve as the second extreme imbalanced sample set.

[0058] As an optional embodiment, the training set of the initial state detection model is determined based on the first extreme imbalance sample set and the second extreme imbalance sample set, including: acquiring static information and environmental data of the target distribution network, wherein the static information includes the power grid structure, line parameters, and equipment capacity, and the environmental data includes weather information, production data, and living data; and based on the static information and environmental data, determining samples in the first extreme imbalance sample set and the second extreme imbalance sample set whose correlation with the operation of the target distribution network is higher than a preset threshold, as the training set.

[0059] Optionally, when determining the training set for the initial state detection model, static information and environmental data of the target distribution network are further introduced to enhance the model's understanding and adaptability to the actual operating environment of the distribution network. First, the acquisition of static information covers the basic architectural attributes of the power grid, such as its topology, specific line parameters like resistance and reactance, and capacity limitations of various devices. This information forms the basis of the distribution network's physical properties and is crucial for constructing the model's input features. Second, the collection of environmental data focuses on external influencing factors, including weather information (such as temperature, humidity, and wind speed), industrial production data (such as differences in production activities between weekdays and holidays), and residential data (such as the temporal distribution of peak and off-peak electricity consumption). This data not only enriches the model's learning perspective but also enables it to consider the impact of dynamic environmental factors on the distribution network's operating state.

[0060] Based on static information and environmental data, correlation analysis is used to select samples highly correlated with the target distribution network operation from the first and second extreme imbalance sample sets. This analysis process involves quantifying the degree of correlation between static information, environmental data, and the distribution network operation status. By setting a preset threshold, samples with low correlation are eliminated to ensure that every record in the training set can provide valuable training signals for the model, which helps the model learn and predict the distribution network operation status more accurately.

[0061] As an optional embodiment, the method further includes: when the status detection result is the safe state, no control operation is performed on the target distribution network; when the status detection result is the alert state, a first control operation is performed on the target distribution network, wherein the first control operation includes load transfer, increasing reserve capacity, and system hardening; when the status detection result is a general emergency state, a second control operation is performed on the target distribution network, wherein the second control operation includes fault isolation, load reduction, and activation of emergency power supply; and when the status detection result is a severe emergency state, a third control operation is performed on the target distribution network, wherein the third control operation includes system disconnection.

[0062] Optionally, when the target distribution network state detection result obtained through the model is a safe state, it indicates that the current operating conditions of the power grid are within a stable and normal range, requiring no additional control operations, and the system can maintain its current operation. This means that the power grid in a safe state has sufficient stability and margin, and can self-adapt and withstand expected disturbances without human intervention or automatic control measures.

[0063] However, when the status monitoring results indicate that the distribution network is in a state of alert, it means that the system is approaching the critical point of its stability boundary. At this time, it is necessary to implement the first control operation on the target distribution network to prevent potential instability. The first control operation specifically includes load transfer, increasing reserve capacity, and system hardening. Load transfer involves transferring some load from the current system to other more stable parts of the distribution network or the backup network to reduce the pressure on the current system; increasing reserve capacity means activating backup power generation resources or energy storage devices in advance to improve the grid's emergency response capability and redundancy; system hardening includes adjusting the grid's structure or parameters, such as optimizing the load distribution of transmission lines and strengthening the maintenance and inspection of key equipment, to enhance the grid's resilience to disturbances.

[0064] When the status monitoring results determine that the distribution network is in a general emergency state, it indicates that the grid has been significantly disturbed, but the system can still restore stable operation through the action of automatic protection devices. In this situation, the second control operation aims to quickly isolate the faulty area while maintaining continuous power supply to the grid as much as possible. The second control operation includes fault isolation, load shedding, and activation of emergency power sources. Fault isolation involves the activation of automatic protection devices to quickly disconnect faulty lines or equipment to prevent the fault from spreading; load shedding reduces the power supply to non-critical loads when necessary to alleviate grid pressure; activating emergency power sources includes starting diesel generators, energy storage systems, etc., to provide uninterrupted power support for critical loads.

[0065] Finally, when the status monitoring results indicate that the distribution network is in a severe emergency state, meaning that the stable operation of the grid cannot be restored through conventional means, the third control operation implemented is system disconnection, which involves dividing the distribution network into multiple independently operating parts. This operation is an emergency measure under extreme circumstances, designed to prevent large-scale system collapse, protect critical power facilities from damage, and create conditions for subsequent system recovery and fault repair. System disconnection involves measures such as disconnecting the main line connection and isolating the faulty subnetwork to ensure the continuous operation and power supply security of the remaining grid parts.

[0066] As an optional embodiment, Figure 5 This is a flowchart of the power distribution status detection model processing provided by an optional embodiment of the present invention. Figure 5As shown, firstly, the operation information of the distribution network system is extracted from the simulation database, and then the data is preprocessed and samples are constructed. Specifically, based on the characteristics of wind turbines and photovoltaic power generation, Latin hypercube sampling is used to generate an extreme imbalance sample set to comprehensively cover all possible situations of wind and solar power output. Simultaneously, based on the characteristics of electric vehicle loads, another extreme imbalance sample set is constructed using Monte Carlo random sampling. Secondly, a feature subset is constructed based on the operation experience of the distribution network, and the optimal feature subset is selected. Subsequently, a distribution network operation status detection model is built using convolutional neural network (CNN) technology. Based on the above dataset, automatic learning is performed to extract features from the input data, and high-precision distribution network status classification is achieved. The model's recognition ability is improved through continuous iterative optimization of weight parameters. Finally, the optimized model is put into practical application; that is, with the input of real-time operating data, the model can quickly determine the real-time operation status of the distribution network, providing a decision-making basis for the prevention and emergency control of the distribution network.

[0067] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0068] Through the above description of the embodiments, those skilled in the art can clearly understand that the power distribution network status detection method according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0069] According to embodiments of the present invention, an apparatus for implementing the above-described power distribution network status detection method is also provided. Figure 6 This is a structural block diagram of a power distribution network status detection device provided according to an embodiment of the present invention, such as... Figure 6 As shown, the power distribution network status detection device includes an acquisition module 61 and a detection module 62. The power distribution network status detection device will be described below.

[0070] The acquisition module 61 is used to acquire real-time operating information of the target power distribution network, wherein the real-time operating information includes real-time power generation of wind turbines, real-time power generation of photovoltaic units, and real-time charging data of electric vehicles, wherein the electric vehicles are located in the target power distribution network.

[0071] The detection module 62, connected to the acquisition module 61, is used to input real-time operating information into a preset target state detection model to obtain the state detection result of the target power distribution network. The target state detection model is trained based on an initial state detection model, which is a convolutional neural network model. The state detection result includes a safe state, a warning state, a general emergency state, or a severe emergency state.

[0072] It should be noted that the acquisition module 61 and the detection module 62 mentioned above correspond to steps S201 to S202 in the embodiments. Multiple modules implement the same instances and application scenarios as their corresponding steps, but are not limited to the content disclosed in the above embodiments. It should also be noted that the above modules, as part of the device, can run on the computer terminal 10 provided in the embodiments.

[0073] Embodiments of the present invention may provide a computer device. Optionally, in this embodiment, the computer device may be located in at least one of a plurality of network devices in a computer network. The computer device includes a memory and a processor.

[0074] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the power distribution network status detection method and device in this embodiment of the invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby realizing the aforementioned power distribution network status detection method. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to a computer terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0075] The processor can access the information and application programs stored in the memory via a transmission device to perform the following steps: acquiring real-time operating information of the target power distribution network, wherein the real-time operating information includes real-time power generation of wind turbines, real-time power generation of photovoltaic units, and real-time charging data of electric vehicles located in the target power distribution network; inputting the real-time operating information into a preset target state detection model to obtain the state detection result of the target power distribution network, wherein the target state detection model is trained based on an initial state detection model, the initial state detection model is a convolutional neural network model, and the state detection result includes a safe state, a warning state, a general emergency state, or a severe emergency state.

[0076] Optionally, the processor may also execute program code for the following steps: the target state detection model is trained based on the initial state detection model, including: constructing a first extreme imbalance sample set based on the historical power generation of wind turbines and photovoltaic units through Latin hypercube sampling; constructing a second extreme imbalance sample set based on the historical charging data of electric vehicles through Monte Carlo random sampling; constructing an initial state detection model, wherein the initial state detection model includes an input layer, a hidden layer, and an output layer; determining the training set of the initial state detection model based on the first extreme imbalance sample set and the second extreme imbalance sample set; and training the initial state detection model based on the training set to obtain the target state detection model.

[0077] Optionally, the processor may also execute program code for the following steps: constructing a first extreme imbalance sample set based on the historical power generation of wind turbines and photovoltaic units through Latin hypercube sampling, including: determining the power generation parameter space; defining the value range of each power generation parameter based on the historical power generation of wind turbines and photovoltaic units; dividing the power generation parameter space, uniformly dividing the value range of each parameter into multiple equally probable intervals; randomly selecting a sample point in each of the multiple equally probable intervals to obtain multiple samples as the first extreme imbalance sample set.

[0078] Optionally, the processor may also execute program code for the following steps: constructing a second extreme imbalance sample set based on historical charging data of electric vehicles through Monte Carlo random sampling, including: determining a probability distribution model of electric vehicle charging behavior based on historical charging data of electric vehicles; randomly selecting a sample point from the probability distribution model and determining whether to accept a sample point; repeating the above sampling and judgment process until the number of accepted sample points exceeds a preset acceptance threshold, thereby obtaining multiple sample points as the second extreme imbalance sample set.

[0079] Optionally, the processor may also execute program code for the following steps: determining the training set of the initial state detection model based on the first extreme imbalance sample set and the second extreme imbalance sample set, including: acquiring static information and environmental data of the target distribution network, wherein the static information includes the power grid structure, line parameters, and equipment capacity, and the environmental data includes weather information, production data, and living data; and based on the static information and environmental data, determining samples in the first extreme imbalance sample set and the second extreme imbalance sample set whose correlation with the operation of the target distribution network is higher than a preset threshold, as the training set.

[0080] Optionally, the processor may also execute program code with the following steps: When the status detection result is a safe state, no control operation is performed on the target distribution network; when the status detection result is a warning state, a first control operation is performed on the target distribution network, wherein the first control operation includes load transfer, increasing reserve capacity, and system hardening; when the status detection result is a general emergency state, a second control operation is performed on the target distribution network, wherein the second control operation includes fault isolation, load reduction, and activation of emergency power supply; and when the status detection result is a severe emergency state, a third control operation is performed on the target distribution network, wherein the third control operation includes system disconnection.

[0081] This invention provides a scheme for power distribution network status detection. By acquiring real-time operating information of the target power distribution network, including real-time power generation of wind turbines, real-time power generation of photovoltaic units, and real-time charging data of electric vehicles located within the target power distribution network, the real-time operating information is input into a preset target status detection model to obtain the status detection results of the target power distribution network. The target status detection model is trained based on an initial status detection model, which is a convolutional neural network model. The status detection results include safe status, alert status, general emergency status, or severe emergency status. This achieves the goal of constructing a high-precision power distribution network operation status identification system based on a convolutional neural network, thereby solving the technical problems of traditional methods that separate transient stability analysis of the safety domain and the stability domain, leading to inconvenience and reduced accuracy in real-time power distribution network operation status identification, and the inability to effectively cope with source load fluctuations and their uncertainties.

[0082] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a non-volatile storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0083] Embodiments of the present invention also provide a non-volatile storage medium. Optionally, in this embodiment, the aforementioned non-volatile storage medium can be used to store the program code executed by the power distribution network status detection method provided in the above embodiments.

[0084] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0085] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: acquiring real-time operating information of the target distribution network, wherein the real-time operating information includes real-time power generation of wind turbines, real-time power generation of photovoltaic units, and real-time charging data of electric vehicles, wherein the electric vehicles are located in the target distribution network; inputting the real-time operating information into a preset target state detection model to obtain the state detection result of the target distribution network, wherein the target state detection model is trained based on an initial state detection model, the initial state detection model is a convolutional neural network model, and the state detection result includes a safe state, a warning state, a general emergency state, or a severe emergency state.

[0086] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: the target state detection model is trained based on the initial state detection model, including: constructing a first extreme imbalance sample set based on the historical power generation of wind turbines and photovoltaic units through Latin hypercube sampling; constructing a second extreme imbalance sample set based on the historical charging data of electric vehicles through Monte Carlo random sampling; constructing an initial state detection model, wherein the initial state detection model includes an input layer, a hidden layer, and an output layer; determining the training set of the initial state detection model based on the first extreme imbalance sample set and the second extreme imbalance sample set; and training the initial state detection model based on the training set to obtain the target state detection model.

[0087] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: constructing a first extreme imbalance sample set based on the historical power generation of wind turbines and photovoltaic units through Latin hypercube sampling, including: determining the power generation parameter space; defining the value range of each power generation parameter based on the historical power generation of wind turbines and photovoltaic units; dividing the power generation parameter space, uniformly dividing the value range of each parameter into multiple equally probable intervals; randomly selecting a sample point in each of the multiple equally probable intervals to obtain multiple samples as the first extreme imbalance sample set.

[0088] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: constructing a second extreme imbalance sample set based on historical charging data of electric vehicles through Monte Carlo random sampling, including: determining a probability distribution model of electric vehicle charging behavior based on historical charging data of electric vehicles; randomly selecting a sample point from the probability distribution model and determining whether to accept a sample point; repeating the above sampling and determination process until the number of accepted sample points exceeds a preset acceptance threshold, thereby obtaining multiple sample points as the second extreme imbalance sample set.

[0089] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: determining the training set of the initial state detection model based on the first extreme imbalance sample set and the second extreme imbalance sample set, including: acquiring static information and environmental data of the target distribution network, wherein the static information includes the power grid structure, line parameters, and equipment capacity, and the environmental data includes weather information, production data, and living data; and based on the static information and environmental data, determining samples in the first extreme imbalance sample set and the second extreme imbalance sample set whose correlation with the operation of the target distribution network is higher than a preset threshold, as the training set.

[0090] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: including: when the state detection result is the safe state, not performing any control operation on the target distribution network; when the state detection result is the alert state, performing a first control operation on the target distribution network, wherein the first control operation includes load transfer, increasing reserve capacity, and system hardening; when the state detection result is a general emergency state, performing a second control operation on the target distribution network, wherein the second control operation includes fault isolation, load reduction, and activation of emergency power supply; and when the state detection result is a severe emergency state, performing a third control operation on the target distribution network, wherein the third control operation includes system disconnection.

[0091] Embodiments of the present invention also provide a computer program product, including a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, it can: acquire real-time operating information of a target power distribution network, wherein the real-time operating information includes real-time power generation of wind turbines, real-time power generation of photovoltaic units, and real-time charging data of electric vehicles, wherein the electric vehicles are located in the target power distribution network; input the real-time operating information into a preset target state detection model to obtain the state detection result of the target power distribution network, wherein the target state detection model is trained based on an initial state detection model, the initial state detection model is a convolutional neural network model, and the state detection result includes a safe state, a warning state, a general emergency state, or a severe emergency state.

[0092] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0093] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0094] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0095] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0096] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0097] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0098] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for detecting the condition of a power distribution network, characterized in that, include: The real-time operation information of the target power distribution network is obtained, wherein the real-time operation information includes the real-time power generation of wind turbines, the real-time power generation of photovoltaic units, and the real-time charging data of electric vehicles, wherein the electric vehicles are located in the target power distribution network; The real-time operation information is input into a preset target state detection model to obtain the state detection result of the target distribution network. The target state detection model is trained based on an initial state detection model, which is a convolutional neural network model. The state detection result includes a safe state, a warning state, a general emergency state, or a severe emergency state.

2. The method according to claim 1, characterized in that, The target state detection model is trained based on the initial state detection model and includes: By using Latin hypercube sampling, a first extreme imbalance sample set is constructed based on the historical power generation of wind turbines and photovoltaic units. By using Monte Carlo random sampling, a second extreme imbalanced sample set was constructed based on historical charging data of electric vehicles; Construct the initial state detection model, wherein the initial state detection model includes an input layer, a hidden layer, and an output layer; Based on the first extremely imbalanced sample set and the second extremely imbalanced sample set, the training set of the initial state detection model is determined; Based on the training set, the initial state detection model is trained to obtain the target state detection model.

3. The method according to claim 1, characterized in that, The method employs Latin hypercube sampling to construct a first extreme imbalance sample set based on the historical power generation of wind turbines and photovoltaic units, including: Determine the power generation parameter space; Based on the historical power generation of the wind turbine and the historical power generation of the photovoltaic unit, the value range of each power generation parameter is defined; The power generation parameter space is divided, and the value range of each parameter is evenly divided into multiple equally probable intervals; Within each of the multiple equal probability intervals, a sample point is randomly selected to obtain multiple samples, which serve as the first extremely imbalanced sample set.

4. The method according to claim 1, characterized in that, The second extreme imbalanced sample set is constructed using Monte Carlo random sampling based on historical charging data of electric vehicles, including: Based on the historical charging data of electric vehicles, a probability distribution model of electric vehicle charging behavior is determined; Randomly select a sample point from the probability distribution model and determine whether to accept the sample point; Repeat the above extraction and judgment process until the number of accepted sample points exceeds the preset acceptance threshold, and obtain multiple sample points as the second extreme imbalanced sample set.

5. The method according to claim 1, characterized in that, The step of determining the training set for the initial state detection model based on the first extremely imbalanced sample set and the second extremely imbalanced sample set includes: Acquire static information and environmental data of the target power distribution network, wherein the static information includes power grid structure, line parameters, and equipment capacity, and the environmental data includes weather information, production data, and living data; Based on the static information and the environmental data, samples with a correlation higher than a preset threshold with the operation of the target distribution network are determined from the first extreme imbalance sample set and the second extreme imbalance sample set, and used as the training set.

6. The method according to claim 1, characterized in that, Also includes: If the status detection result indicates a safe state, no control operation will be performed on the target distribution network. Alternatively, if the status detection result indicates an alert status, a first control operation may be performed on the target distribution network, wherein the first control operation includes load transfer, increasing reserve capacity, and system hardening. Alternatively, if the status detection result indicates a general emergency state, a second control operation may be performed on the target distribution network, wherein the second control operation includes fault isolation, load reduction, and activation of emergency power supply. Alternatively, if the status detection result indicates a severe emergency state, a third control operation may be performed on the target distribution network, wherein the third control operation includes system disconnection.

7. A power distribution network condition detection device, characterized in that, include: The acquisition module is used to acquire real-time operating information of the target power distribution network, wherein the real-time operating information includes real-time power generation of wind turbines, real-time power generation of photovoltaic units, and real-time charging data of electric vehicles, wherein the electric vehicles are located in the target power distribution network; The detection module is used to input the real-time operation information into a preset target state detection model to obtain the state detection result of the target power distribution network. The target state detection model is trained based on an initial state detection model, which is a convolutional neural network model. The state detection result includes a safe state, a warning state, a general emergency state, or a severe emergency state.

8. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the non-volatile storage medium to execute the power distribution network status detection method according to any one of claims 1 to 6.

9. A computer device, characterized in that, include: Memory and processor The memory stores computer programs; The processor is configured to execute a computer program stored in the memory, wherein when the computer program is executed, the processor performs the power distribution network status detection 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 the processor, it implements the power distribution network status detection method according to any one of claims 1 to 6.

Images