Power grid switching method, system and device based on zhi-pan platform and storage medium

Abstract

The application discloses a power grid switching method, system and device based on a wisdom platform and a storage medium, and comprises the following steps: acquiring first physical information collected by a data collection device; performing time alignment and standardization processing on the first physical information to obtain second physical information; constructing an One-Class SVM model based on training data generated by a pre-configured digital twin model, training the One-Class SVM model to obtain a trained model; inputting the second physical information into the trained model for anomaly detection to obtain abnormal data and a power grid anomaly position; generating a power grid switching instruction according to the abnormal data and sending the power grid switching instruction to a power grid switching execution device according to the power grid anomaly position. The method can simultaneously solve the problems of noise reduction and anomaly identification, and quickly determine the power grid anomaly position from massive data.

Description

Technical Field

[0001] This invention relates to the field of power grid switching technology, and in particular to a power grid switching method, system, device and storage medium based on a smart viewing platform. Background Technology

[0002] With the development of society and the economy, electricity has become a crucial infrastructure supporting social operations. my country's power grid is continuously expanding, with a large influx of power electronic equipment and new energy sources, leading to increasing risks to the safe and stable operation of the power grid. In the process of ensuring the safe operation of the power grid, grid switching typically refers to the switching between different power sources or grid sections within the power system due to various reasons (such as equipment maintenance, fault handling, etc.) to ensure the continuity and stability of power supply.

[0003] The Southern Power Grid Smart View Platform is based on a unified power grid data model. Using a map as the entry point, it supports the management of graphics, topology, and ledgers for all types of power grid equipment, achieving the integration of geographical, physical, and management data. On this basis, it expands to include new energy and integrated energy models, deepening the platform's ability to support business operations and providing support for the design and operation monitoring of distributed power sources and intelligent planning of distribution networks under the new power system.

[0004] A common grid switching method involves noise reduction and fault detection of the collected grid data before generating grid switching commands. However, in "high-renewable energy" power systems (those with a high proportion of renewable energy and power electronic equipment), the number of dynamic components to be considered during grid switching can reach 10 due to the replacement of small power sources and the integration of massive distributed power sources, electric vehicles, and distributed energy storage. 5 ~10 6 The sheer scale of this has led to a "dimensionality explosion," posing serious challenges to the stable operation of the power grid, such as high computational and communication demands.

[0005] Faced with the ever-expanding scale of the power grid, the aforementioned power grid switching methods require filtering massive amounts of data, making it difficult to quickly identify the location of power grid anomalies, which is detrimental to the safe and stable operation of the power grid. Summary of the Invention

[0006] This invention provides a power grid switching method, system, device, and storage medium based on an intelligent observation platform, which simultaneously solves the problems of noise reduction and anomaly identification, and quickly determines the location of power grid anomalies from massive amounts of data.

[0007] In a first aspect, to address the aforementioned technical problems, the present invention provides a power grid switching method based on an intelligent observation platform, comprising:

[0008] Acquire first physical information collected by the data acquisition device, wherein the first physical information includes environmental data, equipment status data, and power grid operation data;

[0009] The first physical information is time-aligned and standardized to obtain the second physical information;

[0010] A One-Class SVM model is constructed based on training data generated from a pre-configured digital twin model. The One-Class SVM model is then trained to obtain the trained model. The training data includes noise and normal data that are fused from the entire power grid data.

[0011] The second physical information is input into the trained model to perform anomaly detection, thereby obtaining abnormal data and the location of power grid anomalies.

[0012] A power grid switching command is generated based on the abnormal data, and the power grid switching command is sent to the power grid switching execution device according to the location of the power grid abnormality.

[0013] As an optional implementation, the environmental data includes the temperature and humidity of the environment in which the data acquisition device is located, the device status data includes the location coordinates and mechanical wear of the data acquisition device, and the power grid operation data includes the power grid load, voltage, and current.

[0014] As an optional implementation, the first physical information is time-aligned and normalized to obtain the second physical information, including:

[0015] Time alignment is performed on the first physical information to generate a multi-dimensional feature vector. Among them are: This represents the temperature in the i-th time window. This represents the humidity in the i-th time window. This represents the position coordinates of the i-th time window. This represents the degree of mechanical wear in the i-th time window. This represents the grid load in the i-th time window. This represents the voltage in the i-th time window. This represents the current in the i-th time window;

[0016] The multidimensional feature vector is standardized over multiple time windows to generate the second physical information as shown in the following formula:

[0017]

[0018] in, T The number of time windows is indicated; each row in the second physical information represents the multidimensional feature vector within a time window; each column in the second physical information represents the change of the first physical information over time.

[0019] As an optional implementation, the step of constructing a One-Class SVM model based on training data generated from a pre-configured digital twin model, training the model to obtain the trained model, includes:

[0020] A One-Class SVM model is constructed using a radial basis function kernel as the kernel function and reasonable hyperparameters are set. The radial basis function kernel is used to map the input data to a high-dimensional space, and the hyperparameters are used to determine the hypersphere of the One-Class SVM model.

[0021] The training data generated by the digital twin model is passed to the Fit function configured in the One-Class SVM model for training, and the trained model is obtained.

[0022] As an optional implementation, the high-dimensional space is used to separate similar data points from dissimilar data points.

[0023] The hypersphere is a boundary generated by the One-Class SVM model based on the hyperparameters and the radial basis function kernel. It is used to determine abnormal data and normal data containing noise. The test sample data within the hypersphere boundary is the normal data containing noise, and the test sample data within the hypersphere boundary is the abnormal data.

[0024] As an optional implementation, the step of inputting the second physical information into the trained model for anomaly detection to obtain abnormal data and the location of power grid anomalies includes:

[0025] The second physical information is passed as a test sample to the Predict function configured in the trained model for prediction, and a prediction result is obtained; wherein, the Predict function is used to calculate the distance between the data of the test sample and the hypersphere configured in the trained model, the data with a first distance value is the abnormal data, the data with a second distance value is the normal data containing noise, and the abnormal location of the power grid is the location coordinate of the data acquisition device corresponding to the abnormal data.

[0026] As an optional implementation, the power grid switching command includes: connecting, disconnecting, or isolating the power supply equipment at the abnormal location of the power grid.

[0027] Secondly, the present invention provides a power grid switching system based on a smart observation platform, comprising: a data collection module for acquiring first physical information, wherein the first physical information includes: environmental data, equipment status data, and power grid operation data;

[0028] The preprocessing module is used to perform time alignment and normalization on the first physical information to obtain the second physical information;

[0029] The model training module is used to construct a One-Class SVM model based on the training data generated by the pre-configured digital twin model, train the One-Class SVM model to obtain the trained model, wherein the training data includes noise and normal data that fuses the entire power grid data;

[0030] An anomaly detection module is used to input the second physical information into the trained model to perform anomaly detection, and obtain abnormal data and the location of power grid anomalies;

[0031] The instruction sending module is used to generate a power grid switching instruction based on the abnormal data, and send the power grid switching instruction to the power grid switching execution device according to the location of the power grid abnormality.

[0032] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the program to implement the power grid switching method based on a big data platform as described above.

[0033] Fourthly, the present invention also provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the power grid switching method based on a big data platform as described above.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] This invention provides a power grid switching method, system, and storage medium based on an intelligent observation platform. The method is executed by a data terminal and includes: acquiring first physical information collected by a data acquisition device; performing time alignment and standardization on the first physical information to obtain second physical information; constructing a One-Class SVM model based on training data generated from a pre-configured digital twin model, training the One-Class SVM model to obtain a trained model; inputting the second physical information into the trained model for anomaly detection to obtain abnormal data and power grid anomaly locations; generating a power grid switching command based on the abnormal data, and sending the power grid switching command to a power grid switching execution device based on the power grid anomaly locations.

[0036] The proposed method does not require additional filtering operations. It solves both noise reduction and anomaly detection problems simultaneously through a One-Class SVM model. It can quickly identify the location of power grid anomalies from massive amounts of data, enabling fast and efficient power grid switching operations, which is beneficial to the stable operation of the power grid. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of a power grid switching method based on a smart viewing platform provided in an embodiment of the present invention.

[0038] Figure 2 This is a schematic diagram of a power grid switching system based on a smart viewing platform provided in an embodiment of the present invention. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0040] This invention provides a power grid switching method based on a big data platform, comprising the following steps:

[0041] S11, acquire the first physical information collected by the data acquisition device, wherein the first physical information includes environmental data, equipment status data and power grid operation data;

[0042] S12, perform time alignment and normalization processing on the first physical information to obtain the second physical information;

[0043] S13, Construct a One-Class SVM model based on the training data generated by the pre-configured digital twin model, train the One-Class SVM model to obtain the trained model, wherein the training data includes noise and normal data that fuses the entire power grid data;

[0044] S14, input the second physical information into the trained model to perform anomaly detection, and obtain abnormal data and the location of power grid anomalies;

[0045] S15, Generate a power grid switching command based on the abnormal data, and send the power grid switching command to the power grid switching execution device according to the abnormal location of the power grid.

[0046] In this invention, the big data platform is a platform or data processing terminal that uses digital means to re-engineer processes, drive service transformation, and enhance user experience. This platform, based on mapping technology and driven by geographic, physical, and management data, is one of the next-generation digital power grid operation infrastructure platforms serving the entire network and society. It integrates data from the power generation sector, explores multi-dimensional and comprehensive data fusion analysis and display, and achieves seamless integration of business standards and IT system construction. Furthermore, the big data platform also supports graphical, topological, and ledger management of all types of power grid equipment, achieving the fusion of geographic, physical, and management data, and expanding applications to include new energy and integrated energy services.

[0047] Optionally, the big data platform includes the intelligent observation platform, or other data processing terminals with comprehensive data processing functions. This invention does not limit the scope of the application.

[0048] Accordingly, refer to Figure 1 This invention provides a power grid switching method based on a smart observation platform, comprising the following steps:

[0049] S11, acquire the first physical information collected by the data acquisition device, wherein the first physical information includes environmental data, equipment status data and power grid operation data;

[0050] S12, perform time alignment and normalization processing on the first physical information to obtain the second physical information;

[0051] S13, Construct a One-Class SVM model based on the training data generated by the pre-configured digital twin model, train the One-Class SVM model to obtain the trained model, wherein the training data includes noise and normal data that fuses the entire power grid data;

[0052] S14, input the second physical information into the trained model to perform anomaly detection, and obtain abnormal data and the location of power grid anomalies;

[0053] S15, Generate a power grid switching command based on the abnormal data, and send the power grid switching command to the power grid switching execution device according to the abnormal location of the power grid.

[0054] In step S11, it should be noted that the environmental data includes the temperature and humidity of the environment in which the data acquisition equipment is located. The equipment status data includes the location coordinates and mechanical wear level of the data acquisition equipment, and the power grid operation data includes the power grid load, voltage, and current for multiple time windows within the current detection cycle.

[0055] Furthermore, the data acquisition equipment comprises various sensing devices configured for the power grid, including a power grid measurement unit, a temperature and humidity measurement unit, a coordinate positioning unit, and a network access unit. For example, the power grid measurement unit measures power grid load, voltage, and current, contributing to power grid safety. The temperature and humidity measurement unit measures ambient temperature and humidity; monitoring ambient temperature helps identify fires or other temperature anomalies, while monitoring ambient humidity helps prevent equipment problems caused by humidity. The coordinate positioning unit is used for coordinate positioning of the data collection module, acquiring geographical location data, such as three-dimensional spatial coordinates. The network access unit transmits the various physical information acquired by the data acquisition equipment to the data terminal.

[0056] In step S12, it should be noted that time alignment is performed on the first physical information. Time alignment ensures that different physical information has a consistent time base, generating a multi-dimensional feature vector. This includes: This represents the temperature in the i-th time window. This represents the humidity in the i-th time window. This represents the position coordinates of the i-th time window. This represents the degree of mechanical wear in the i-th time window. This represents the grid load in the i-th time window. This represents the voltage in the i-th time window. The current in the i-th time window is represented; then, the multidimensional data is standardized across multiple time windows, including superposition processing within multiple time windows, to generate the second physical information as shown in the following formula:

[0057]

[0058] in, T The number of time windows is indicated; each row in the second physical information represents the multidimensional feature vector within a time window; each column in the second physical information represents the change of the first physical information over time.

[0059] In step S13, it should be noted that a One-Class SVM model is constructed by using a radial basis function kernel (RBF) as the kernel function and setting reasonable hyperparameters. The radial basis function kernel (RBF) is used to map the input data to a high-dimensional space. The nonlinear problem that is difficult to handle is transformed into a linear problem after being mapped to a high-dimensional space, so as to separate similar data points from dissimilar data points.

[0060] In this embodiment, the hyperparameter is pre-defined and defined as the upper limit of the proportion of samples violating constraints to all samples. It is the upper limit of the sum of all slack variables divided by the number of samples, used to determine the hypersphere of the One-Class SVM model. For example, the hyperparameter can be set to 0.00001, and this parameter can be adjusted as needed based on the proportion of anomalous data in historical data. The One-Class SVM model is an unsupervised machine learning model for anomaly detection. It belongs to the Support Vector Machine (SVM) family of models, but unlike traditional SVMs, the One-Class SVM model is specifically designed to handle cases with only one class of positive data (i.e., no explicit negative examples), identifying normal data in the dataset and treating any samples deviating from this pattern as anomalous data.

[0061] It's worth noting that the hypersphere is a sphere in a feature space of three or more dimensions. The boundary is generated by the One-Class SVM model based on hyperparameters and training data. It's used to distinguish between anomalous data and normal data containing noise. Test sample data outside the hypersphere boundary is considered anomalous and is marked with a first distance value "-1" in the One-Class SVM model; test sample data inside the hypersphere boundary is considered normal data containing noise and is marked with a second distance value "1" in the One-Class SVM model. The hypersphere can be described by the following formula:

[0062]

[0063] in, O It is the center of the hypersphere. r It is the radius of the hypersphere. ξ i It is a slack variable. l i The penalty coefficient is... i It is a positive integer. x i For training data. Further, slack variables. ξ i Indicates that training data is allowed. x i Penalty coefficient for violating the definition of a hypersphere l i The coefficients of manually defined slack variables are used to weight the importance of training data; the penalty coefficients are also mentioned. l i The larger the value, the more important the training data is, and the less likely the training data is to exceed the hypersphere boundary.

[0064] Furthermore, the training data generated by the big data platform is passed to the Fit function configured in the One-Class SVM model for training, resulting in a trained model. The big data platform includes the Intelligent View Platform, which integrates full-domain power grid data to establish a digital twin model. A digital twin model is a highly detailed virtual model that uses data from the physical world to create a digital replica of a real system or object. This technology is primarily used to simulate, analyze, and predict the performance of complex systems for testing and optimization without affecting the real system. Digital twin models typically contain a large amount of real-time data as well as historical event data collected under similar past conditions. This data is collected from its physical counterpart via sensors and used to update the model to reflect real-time conditions. The model integrates geographical, physical, managerial, and operational information to simulate the actual power grid, assisting power grid personnel in making power grid switching decisions. By integrating and simulating real-world power grid systems, the Intelligent View Platform's digital twin model can generate a large amount of data suitable for training machine learning models.

[0065] In one implementation, the digital twin model performs digital simulations of various devices in an actual power grid system, using simulation data from each data acquisition device in the digital twin model as training data. It should be noted that the simulation data generated in the digital twin model contains noise.

[0066] It should be noted that the training data generated by the digital twin model is time-aligned and standardized according to the first physical information, and then passed to the Fit function configured in the One-Class SVM model for machine learning model training to obtain the trained model.

[0067] The Fit function is a core component of the One-Class SVM model. It receives time-aligned and normalized training data and the One-Class SVM model, treating the training data as normal data of the same class. Internally, the Fit function constructs and solves an optimization problem to find a minimal hypersphere that contains as much of the training data as possible. During the optimization process, the Fit function determines support vectors located on or near the hypersphere boundary; these support vectors are crucial for defining the optimal hypersphere. The output of the Fit function is the trained model, which generates a minimal hypersphere based on predefined hyperparameters and the training data, aiming to contain the training data within the hypersphere boundary. This hypersphere is used for anomaly detection on test samples.

[0068] In step S14, the second physical information is input into the trained model for anomaly detection to obtain abnormal data and the location of power grid anomalies. It should be noted that the second physical information is used as a test sample and passed to the Predict function configured in the trained model for prediction to obtain the prediction result.

[0069] Because the training data contains noise, the trained model classifies noise and normal data into one category. Therefore, from acquiring the first physical information from the data acquisition device at the data terminal to inputting the second physical information into the trained model, no noise reduction operation is required. Compared with existing power grid switching methods, this reduces the noise reduction step and helps save computing resources.

[0070] The `Predict` function calculates the distance between the test sample data and the hyperspherical boundary configured in the trained model, and then inputs it into the sign function (`sgn`). The sign function (`sgn`) classifies the input value: it returns "1" for positive values ​​and "-1" for negative values. Therefore, if the `Predict` function returns the first distance value "-1", the test sample data is considered outlier; if it returns the second distance value "1", the test sample data is considered normal data containing noise. The `Predict` function can be expressed as follows:

[0071]

[0072] in, x m For test sample data, r For the hypersphere radius, O The center of the hypersphere, This is to test the distance of the sample data to the center of the hypersphere in high-dimensional space.

[0073] Then, based on the second physical information where the abnormal data is located, the location coordinates of the data acquisition device corresponding to the abnormal data are found, which is the location of the power grid anomaly. It is worth noting that the second physical information includes location coordinates.

[0074] In step S15, a power grid switching command is generated based on the abnormal data, and the command is sent to the power grid switching execution device according to the location of the power grid anomaly. It should be noted that the data terminal analyzes the abnormal data, determines the anomaly type, and generates a corresponding power grid switching command based on the anomaly type. For example, the anomaly type may include fire, typhoon, or other situations requiring power grid switching; this invention does not limit this. The power grid switching command includes connecting, disconnecting, or isolating the power supply equipment at the location of the power grid anomaly.

[0075] This invention also provides a power grid switching system based on a big data platform, which can implement the entire process of the above-described power grid switching method based on a big data platform.

[0076] Reference Figure 2 This is a schematic diagram of an embodiment of the power grid switching system based on a big data platform provided by the present invention.

[0077] This invention provides a power grid switching system based on a big data platform, comprising:

[0078] The data collection module 201 is used to acquire first physical information, wherein the first physical information includes: environmental data, equipment status data, and power grid operation data;

[0079] Preprocessing module 202 is used to perform time alignment and standardization processing on the first physical information to obtain the second physical information;

[0080] The model training module 203 is used to construct a One-Class SVM model based on the training data generated by the pre-configured digital twin model, train the One-Class SVM model to obtain the trained model, wherein the training data includes noise and normal data that fuses the whole-domain power grid data;

[0081] Anomaly detection module 204 is used to input the second physical information into the trained model to perform anomaly detection, and obtain abnormal data and power grid anomaly locations;

[0082] The instruction sending module 205 is used to generate a power grid switching instruction based on the abnormal data, and send the power grid switching instruction to the power grid switching execution device according to the abnormal location of the power grid.

[0083] In a specific embodiment, the data collection module collects first physical information from the various sensing units configured in the data acquisition device. The environmental data includes the temperature and humidity of the environment in which the data acquisition device is located. The device status data includes the location coordinates and mechanical wear of the data acquisition device. The power grid operation data includes the power grid load, voltage, and current.

[0084] Optionally, the data collection module 201 includes:

[0085] The power grid measurement unit is used to measure power grid load, voltage, and current, which helps to ensure power grid safety.

[0086] Temperature and humidity measurement units are used to measure the temperature and humidity of the environment. Monitoring the ambient temperature helps to identify fires or other abnormal temperature conditions, while monitoring the ambient humidity helps to prevent equipment problems caused by humidity.

[0087] The coordinate positioning unit is used for coordinate positioning of the data collection module to obtain geographic location data, such as three-dimensional spatial coordinates.

[0088] The network access unit is used to transmit the various first physical information collected by the data acquisition device to the data terminal.

[0089] Optionally, the preprocessing module 202 includes:

[0090] The feature vector generation unit is used to perform time alignment on the first physical information to generate a multidimensional feature vector. Among them are: This represents the temperature in the i-th time window. This represents the humidity in the i-th time window. This represents the position coordinates of the i-th time window. This represents the degree of mechanical wear in the i-th time window. This represents the grid load in the i-th time window. This represents the voltage in the i-th time window. This represents the current in the i-th time window;

[0091] The standardization processing unit is used to standardize the multidimensional feature vector within multiple time windows to generate the second physical information as shown in the following formula:

[0092]

[0093] in, T The number of time windows is indicated; each row in the second physical information represents the multidimensional feature vector within a time window; each column in the second physical information represents the change of the first physical information over time.

[0094] Optionally, the model training module 203 includes:

[0095] A model building unit is used to construct a One-Class SVM model using a radial basis function kernel as the kernel function and setting reasonable hyperparameters. The radial basis function kernel maps the input data to a high-dimensional space, and the hyperparameters determine the hypersphere of the One-Class SVM model. The high-dimensional space separates similar and dissimilar data points. The hypersphere is a boundary generated by the One-Class SVM model based on the hyperparameters and the radial basis function kernel, used to distinguish between anomalous data and normal data containing noise. Test sample data within the hypersphere boundary is considered normal data containing noise, and test sample data within the hypersphere boundary is considered anomalous data.

[0096] The training data generation unit is used to obtain training data by using simulation data from various data acquisition devices in the digital twin model as training data.

[0097] The training unit is used to pass the training data generated by the digital twin model to the Fit function configured in the One-Class SVM model for training, so as to obtain the trained model.

[0098] Optionally, the anomaly detection module 204 includes:

[0099] An abnormal data detection unit is used to pass the second physical information as a test sample to the Predict function configured in the trained model for prediction, and obtain a prediction result; wherein, the Predict function is used to calculate the distance between the data of the test sample and the hypersphere configured in the trained model, the data with a distance of a first distance value is the abnormal data, and the data with a distance of a second distance value is the normal data containing noise.

[0100] The power grid anomaly location detection unit is used to find the location coordinates of the data acquisition device corresponding to the abnormal data in the second physical information and determine it as the abnormal location of the power grid.

[0101] Optionally, the instruction sending module 205 includes:

[0102] The instruction generation unit is used to analyze abnormal data, determine the type of abnormality, and generate a corresponding power grid switching instruction based on the type of abnormality. For example, the abnormality type may include fire, typhoon, or other situations requiring power grid switching; this invention does not limit this. The power grid switching instruction includes: connecting, disconnecting, or isolating the power supply equipment at the location of the power grid abnormality.

[0103] The instruction transmission unit is used to transmit the power grid switching instruction to the power grid switching execution device according to the abnormal location of the power grid.

[0104] In summary, this invention provides a power grid switching method, system, device, and storage medium based on a big data platform. The method is executed by a data terminal and includes: acquiring first physical information collected by a data acquisition device; performing time alignment and standardization on the first physical information to obtain second physical information; constructing a One-Class SVM model based on training data generated from a pre-configured digital twin model; training the One-Class SVM model to obtain a trained model; inputting the second physical information into the trained model for anomaly detection to obtain abnormal data and power grid anomaly locations; generating a power grid switching command based on the abnormal data; and sending the power grid switching command to a power grid switching execution device based on the power grid anomaly locations. This method eliminates the need for additional filtering operations, simultaneously solving both noise reduction and anomaly detection problems through the One-Class SVM model. It can quickly determine the location of power grid anomalies from massive amounts of data, achieving fast and efficient power grid switching operations, which is beneficial for the stable operation of the power grid.

[0105] This invention also provides an electronic device. The terminal device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a projector licensing program. When the processor executes the computer program, it implements the steps described in the various projector licensing method embodiments above, for example... Figure 1 The step S11 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the above system embodiments.

[0106] For example, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0107] The terminal device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The terminal device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of terminal devices and do not constitute a limitation on the terminal device. It may include more or fewer components than described above, or a combination of certain components, or different components. For example, the terminal device may also include input / output devices, network access devices, buses, etc.

[0108] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0109] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0110] Wherein, if the modules / units integrated in the terminal device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or system capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0111] It should be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and 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 network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0112] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A power grid switching method based on an intelligent observation platform, characterized in that, Executed by the data terminal, including: Acquire first physical information collected by the data acquisition device, wherein the first physical information includes environmental data, equipment status data, and power grid operation data; The first physical information is time-aligned and standardized to obtain the second physical information; A One-Class SVM model is constructed based on training data generated from a pre-configured digital twin model. The One-Class SVM model is then trained to obtain the trained model. The training data includes noise and normal data that are fused from the entire power grid data. The second physical information is input into the trained model to perform anomaly detection, thereby obtaining abnormal data and the location of power grid anomalies. A power grid switching command is generated based on the abnormal data, and the power grid switching command is sent to the power grid switching execution device according to the location of the power grid anomaly. The environmental data includes the temperature and humidity of the environment in which the data acquisition device is located; the device status data includes the location coordinates and mechanical wear of the data acquisition device; and the power grid operation data includes the power grid load, voltage, and current. The first physical information is time-aligned and normalized to obtain the second physical information, which includes: The first physical information is time-aligned to generate a multi-dimensional feature vector. Among them are: This represents the temperature in the i-th time window. This represents the humidity in the i-th time window. This represents the position coordinates of the i-th time window. This represents the degree of mechanical wear in the i-th time window. This represents the grid load in the i-th time window. This represents the voltage in the i-th time window. This represents the current in the i-th time window; The multidimensional feature vector is standardized over multiple time windows to generate the second physical information as shown in the following formula: in, T The second physical information represents the number of time windows; each row in the second physical information represents the multidimensional feature vector within a time window; each column in the second physical information represents the change of the first physical information over time. The step of constructing a One-Class SVM model based on training data generated from a pre-configured digital twin model, training the model to obtain the trained model, includes: A One-Class SVM model is constructed using a radial basis function kernel as the kernel function and reasonable hyperparameters are set; wherein, the radial basis function kernel is used to map the input data to a high-dimensional space, and the hyperparameters are used to determine the hypersphere of the One-Class SVM model; The hyperparameter is a preset ratio threshold used to control the proportion of samples in the training data that are allowed to deviate from the boundary. The ratio can be set in combination with the characteristics of historical abnormal data of the power grid. The training data generated by the digital twin model is passed to the Fit function configured in the One-Class SVM model for training, and the trained model is obtained. The digital twin model is built on the intelligent observation platform within the big data platform. The intelligent observation platform integrates full-domain power grid data and generates multi-scenario simulation data for different power grid devices. The simulation data is used to simulate the output of data acquisition devices under different operating conditions; the simulation data contains noise. The step of inputting the second physical information into the trained model for anomaly detection to obtain abnormal data and the location of power grid anomalies includes: The second physical information is passed as a test sample to the Predict function configured in the trained model for prediction, and a prediction result is obtained; wherein, the Predict function is used to calculate the distance between the data of the test sample and the hypersphere configured in the trained model, the data with a first distance value is the abnormal data, the data with a second distance value is normal data containing noise, and the abnormal location of the power grid is the location coordinate of the data acquisition device corresponding to the abnormal data; The power grid switching command includes: connecting, disconnecting, or isolating the power supply equipment at the abnormal location of the power grid.

2. The power grid switching method based on the intelligent observation platform according to claim 1, characterized in that, The high-dimensional space is used to separate similar data points from dissimilar data points; The hypersphere is a boundary generated by the One-Class SVM model based on the hyperparameters and the radial basis function kernel. It is used to determine abnormal data and normal data containing noise. Test sample data within the hypersphere boundary is the normal data containing noise, and test sample data outside the hypersphere boundary is the abnormal data.

3. A power grid switching system based on an intelligent observation platform, characterized in that, The method for implementing the power grid switching method based on the intelligent observation platform as described in any one of claims 1 to 2 includes: The data collection module is used to acquire first physical information, which includes: environmental data, equipment status data, and power grid operation data. The preprocessing module is used to perform time alignment and normalization on the first physical information to obtain the second physical information; The model training module is used to construct a One-Class SVM model based on the training data generated by the pre-configured digital twin model, train the One-Class SVM model to obtain the trained model, wherein the training data includes noise and normal data that fuses the entire power grid data; An anomaly detection module is used to input the second physical information into the trained model to perform anomaly detection, and obtain abnormal data and the location of power grid anomalies; The instruction sending module is used to generate a power grid switching instruction based on the abnormal data, and send the power grid switching instruction to the power grid switching execution device according to the location of the power grid abnormality.

4. An apparatus comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method according to any one of claims 1 to 2.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the power grid switching method based on the intelligent observation platform as described in any one of claims 1 to 2.

Images