Distribution protection method and system based on distributed power supply access

Abstract

The invention discloses a power distribution protection method and system based on distributed power supply access, and relates to the technical field of intelligent power grids, and the method comprises the steps: collecting meteorological, historical and real-time sensor data, carrying out preprocessing preparation analysis, extracting features from the preprocessed data, building a power supply and demand prediction model, and predicting the future power demand. And automatically adjusting a power grid protection threshold based on a prediction result. And using a DBSCAN algorithm to identify an abnormal power grid area, establishing and applying a fault detection model, and processing the abnormal area. And computing resources are dynamically adjusted through a Kubernetes platform, so that efficient execution of key tasks is ensured. The accuracy of power supply and demand prediction is improved, and the stability of power grid operation is enhanced. The overall efficiency and economical efficiency of a power system are improved, and the continuity and reliability of power supply are enhanced. The fault detection and response time is greatly shortened, the system adaptability is enhanced, the maintenance cost is reduced, and the user satisfaction and the overall service quality of a power grid are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart grid, and in particular to a power distribution protection method and system based on distributed power supply access. Background Art

[0002] In the context of global energy transformation, the popularization of distributed generation (DGs) marks the entry of power systems into a new era. Its clean and flexible characteristics optimize the energy structure and improve efficiency. However, the randomness and intermittency of DGs challenge the stability of traditional power grids, especially in the balance and regulation of power supply and demand, and fault detection and response, where centralized management shows limitations. The development of advanced distribution protection technologies to enhance the flexibility and reliability of power grids has become a hot topic and industry frontier in power system research. Although existing technologies have partially alleviated the difficulty of DGs access, the prediction model ignores the impact of weather and has limited accuracy; anomaly detection relies on fixed thresholds and is prone to false alarms and missed alarms; the fault response mechanism lags, affecting power supply restoration and user experience; resource management is static and difficult to adapt to DGs fluctuations, which restricts the efficiency of monitoring decisions. Innovative methods are urgently needed to achieve high-precision prediction, dynamic anomaly detection, immediate fault response and intelligent resource allocation, comprehensively improve the efficiency of power grid management, empower power systems with intelligence, and meet the challenges of the new energy era. Summary of the invention

[0003] In view of the above existing problems, the present invention is proposed.

[0004] Therefore, the present invention provides a distribution protection method and system based on distributed power source access to solve the problems of unstable power supply and demand balance, slow fault detection response, unreasonable resource allocation and increased complexity of power grid management.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] In a first aspect, an embodiment of the present invention provides a power distribution protection method based on distributed power access, which includes integrating collected meteorological data, historical data, and real-time sensor data as initial data and performing preprocessing;

[0007] The preprocessed initial data is subjected to feature extraction through a feature function, and a power supply and demand prediction model is established based on the feature-extracted data to obtain a power supply and demand prediction result;

[0008] Automatically adjust protection thresholds based on power supply and demand forecasts;

[0009] When the power supply value exceeds the set protection threshold, the DBSCAN algorithm is used to divide the abnormal power grid area;

[0010] Establish a fault detection model based on the abnormal power grid area to detect faults in the abnormal power grid area, and process the abnormal power grid area according to the detection results;

[0011] Manage containers through Kubernetes and dynamically adjust resource allocation.

[0012] As a preferred solution of the power distribution protection method based on distributed power access described in the present invention, wherein: meteorological data including temperature, humidity, wind speed, wind direction, and light intensity are obtained from an official meteorological station; the distributed power generation and grid load data of the past five years are extracted from the power company database; the distributed power generation, grid load, voltage, and current are collected in real time from sensors;

[0013] Integrate the collected meteorological data, historical data, and real-time sensor data as initial data;

[0014] Define the initial data as X = (X1, X2, X3, X4, X5, X6, X7, X8, X9);

[0015] Among them, X1 represents temperature data, X2 represents humidity data, X3 represents wind speed data, X4 represents wind direction data, X5 represents light intensity data, X6 represents distributed power generation data, X7 represents grid load data, X8 represents voltage data, and X9 represents current data;

[0016] The initial data is preprocessed and the expression is:

[0017]

[0018] Among them, D i represents the initial data after preprocessing, X represents the set of initial data, i represents the initial data group, j represents the observation point, n represents the number of missing values, median(X i ) represents the median of the data, IQR(X i ) represents the discreteness of the data, and k represents the detection threshold.

[0019] As a preferred solution of the distribution protection method based on distributed power access described in the present invention, the pre-processed initial data is subjected to feature extraction using the principles of wavelet transform and Fourier transform, and the expression is:

[0020]

[0021] Among them, F(x) represents the data after feature extraction, σ represents the standard deviation, and D irepresents the preprocessed data sequence, N represents the total number of preprocessed data sequences, f represents the frequency, φ represents the phase angle, s represents the time point, u represents the signal point, and i represents the index of the sequence;

[0022] Based on the feature-extracted data, the LSTM network is used to establish a power supply and demand forecasting model. The feature-extracted data is input into the power supply and demand forecasting model to obtain the power supply and demand forecasting result, which is expressed as:

[0023]

[0024] Among them, P t represents the final power supply and demand forecast output, F(x) represents the element in the fused feature vector, softmax represents the activation function, h represents the hidden state vector, and W a Represents the weight matrix, LSTM t Represents the output of the LSTM network at time point t.

[0025] As a preferred solution of the power distribution protection method based on distributed power access described in the present invention, a nonlinear transformation is performed on the power supply and demand forecast output result, and the protection threshold is automatically adjusted according to the power supply and demand forecast result after the nonlinear transformation, and its expression is:

[0026]

[0027] Among them, T represents the final calculated threshold, Q p represents the quantile function, ω represents the angular frequency, and φ represents the phase shift;

[0028] The protection threshold T will be based on the power supply and demand forecast output result P t Make dynamic adjustments.

[0029] As a preferred solution of the distribution protection method based on distributed power access described in the present invention, the DBSCAN algorithm is used to divide the abnormal points into regions to determine the abnormal power grid region, and its expression is:

[0030]

[0031] Among them, R represents the region division result, h i represents a single point in the abnormal set, C represents the core point set, and h j Indicates point h i Points in the neighborhood, B represents the boundary point set, C represents the core point set, h l represents the core point, st represents the additional conditions, N(h i ) means h i 's neighborhood.

[0032] As a preferred solution of the distribution protection method based on distributed power access described in the present invention, a fault detection model is established using SVM based on the abnormal power grid area, and its expression is:

[0033] SVM(R)={y:y=sign(w T Φ(x)+b)};

[0034] Where SVM(R) represents the fault detection model, w represents the weight vector, b is the bias, Φ(x) represents the mapping function, and y represents the output of the fault detection model;

[0035] Perform fault detection on abnormal areas. After confirming the fault, cut off power to the faulty area, isolate the faulty area, and restore power to non-faulty areas.

[0036] As a preferred solution of the power distribution protection method based on distributed power supply access described in the present invention, Kubernetes manages the deployment and resource allocation of containers in the cluster through a component called a scheduler. During the dynamic resource allocation process, Kubernetes's agent component kubelet continuously monitors the CPU usage and memory occupancy of each task. Kubelet regularly collects this information and reports it to the API server. When the resource demand of critical tasks increases, the Kubernetes scheduler will re-evaluate resource allocation.

[0037] In a second aspect, the present invention provides a power distribution protection system for accessing a distributed power source, comprising:

[0038] The data integration and preprocessing module integrates data from weather stations, historical databases and real-time sensors and performs preprocessing.

[0039] Feature extraction and modeling module, which extracts meaningful features from preprocessed data and builds a power supply and demand prediction model;

[0040] The protection threshold dynamic adjustment module automatically calculates and adjusts the protection threshold according to the power supply and demand forecast results;

[0041] The anomaly detection and area division module uses the DBSCAN clustering algorithm to identify abnormal power grid areas when the power supply exceeds the protection threshold;

[0042] The fault detection and processing module uses the fault detection model to identify faults based on abnormal areas, isolate the faulty areas, and restore power supply to non-faulty areas;

[0043] The resource dynamic allocation module uses the Kubernetes container management platform to dynamically adjust computing resource allocation according to the resource requirements of key tasks.

[0044] In a third aspect, an embodiment of the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the distribution protection method based on distributed power source access as described in the first aspect of the present invention is implemented.

[0045] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the distribution protection method based on distributed power source access as described in the first aspect of the present invention is implemented.

[0046] The beneficial effects of the present invention are as follows: by collecting meteorological data, historical data and real-time sensor data and performing preprocessing, an initial data set that fully reflects the state of the power grid is formed, the accuracy of power supply and demand forecasts is improved, the forecast errors caused by data quality issues are reduced, and the stability of power grid operation is enhanced. Through accurate power supply and demand forecasts, the power grid can prepare in advance, avoiding overload or waste caused by supply and demand imbalances, and improving the overall efficiency and economy of the power system. The adaptive ability and safety of the power grid are improved, system failures caused by improper setting of fixed thresholds are reduced, and the continuity and reliability of power supply are enhanced. The fault detection and response time is greatly shortened, the maintenance cost is reduced, and the user satisfaction and the overall service quality of the power grid are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0048] Figure 1 This is a flow chart of the power distribution protection method based on distributed power source access in Example 1.

[0049] Figure 2 This is a decision diagram for automatically adjusting the protection threshold and fault detection in Example 1. DETAILED DESCRIPTION

[0050] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0051] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0052] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0053] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and provides a distribution protection method based on distributed power supply access, comprising the following steps:

[0054] S1. Obtain meteorological data including temperature, humidity, wind speed, wind direction, and light intensity from official weather stations; extract distributed power generation and grid load data for the past five years from the power company database; collect distributed power generation, grid load, voltage, and current from sensors in real time;

[0055] Integrate the collected meteorological data, historical data, and real-time sensor data as initial data;

[0056] Define the initial data as X = (X1, X2, X3, X4, X5, X6, X7, X8, X9);

[0057] Among them, X1 represents temperature data, X2 represents humidity data, X3 represents wind speed data, X4 represents wind direction data, X5 represents light intensity data, X6 represents distributed power generation data, X7 represents grid load data, X8 represents voltage data, and X9 represents current data;

[0058] The initial data is preprocessed and the expression is:

[0059]

[0060] Among them, Di represents the initial data after preprocessing, X represents the set of initial data, i represents the initial data group, j represents the observation point, n represents the number of missing values, median(Xi) represents the median of the data, IQR(Xi) represents the degree of discreteness of the data, and k represents the detection threshold.

[0061] S2. Use the principles of wavelet transform and Fourier transform to extract features from the preprocessed initial data. Wavelet transform is a time-frequency analysis tool that can analyze the details of the signal at different scales and is very suitable for analyzing non-stationary signals, that is, signals whose characteristics change over time. Wavelet transform can capture the local characteristics of the signal, which is very useful for analyzing power signals because the characteristics of power signals may change with time and frequency; Fourier transform is a mathematical tool that converts signals from the time domain to the frequency domain, which can reveal the frequency composition of the signal. For periodic signals in power systems, Fourier transform can reveal the fundamental frequency and harmonic components of the signal. Feature extraction is a key step in machine learning and deep learning, which involves extracting useful information from raw data to improve the predictive ability of the model. Its expression is:

[0062]

[0063] Among them, F(x) represents the feature function after feature extraction, σ represents the standard deviation, and D i represents the preprocessed data sequence, N represents the total number of preprocessed data sequences, f represents the frequency, φ represents the phase angle, s represents the time point, u represents the signal point, ds represents the integration at time s, du represents the integration at signal u, and i represents the index of the sequence;

[0064] Based on the characteristic function, the power supply and demand forecasting model is established using the LSTM network. LSTM is a special recurrent neural network that is particularly suitable for processing sequence data. It can remember long-term dependency information. The characteristic function is input into the power supply and demand forecasting model to obtain the power supply and demand forecasting result, which is expressed as:

[0065]

[0066] Among them, P t represents the final power supply and demand forecast output, F(x) represents the element in the fused feature vector, softmax represents the activation function, h represents the hidden state vector, and W a Represents the weight matrix, LSTM t Represents the output of the LSTM network at time point t.

[0067] The whole process starts with feature extraction, capturing the spatiotemporal features of power signals through wavelet transform and Fourier transform, and then inputting the extracted features into the LSTM network to use the memory and learning capabilities of LSTM to predict future power supply and demand. This prediction model is very important for power grid management because it can help power companies plan and adjust power production and distribution in advance to cope with future demand fluctuations.

[0068] S3. Perform nonlinear transformation on the power supply and demand forecast output. The purpose of using nonlinear transformation is to map the power supply and demand forecast output Pt to a new space so that the transformed result can better reflect the dynamic characteristics of the power system. The protection threshold is automatically adjusted according to the power supply and demand forecast result after nonlinear transformation. The expression is:

[0069]

[0070] Among them, T represents the final calculated threshold, Q p represents the quantile function, ω represents the angular frequency, and φ represents the phase shift;

[0071] The protection threshold T will be based on the power supply and demand forecast output result P t Dynamic adjustment. The dynamic adjustment of the protection threshold T is based on the power supply and demand forecast results after nonlinear transformation. Its purpose is to adapt to the changing load demand in the power system and ensure that the power system can operate stably under various load conditions to avoid failures.

[0072] Furthermore, through nonlinear transformation, the power supply and demand forecast results are converted into a form that is more suitable for adjusting the protection threshold. Subsequently, based on the transformed results, the protection threshold T is dynamically adjusted by selecting appropriate quantiles, ensuring that the power system can automatically adapt to load changes, maintain the balance of power supply and demand, and avoid system overload or underload caused by improper setting of fixed thresholds, thereby improving the safety and stability of the power system. This dynamic adjustment mechanism is an important manifestation of the intelligence of the power system, and is particularly important for power grids connected to distributed power sources, because the randomness and intermittency of distributed power sources increase the complexity and uncertainty of power system operation.

[0073] S4. Use the DBSCAN algorithm to divide the outliers into regions. The DBSCAN algorithm is a density-based clustering algorithm that is particularly suitable for discovering clusters of arbitrary shapes and can identify and exclude noise points. The DBSCAN algorithm defines the neighborhood based on the distance between two points. Core points are those points that contain at least a certain number of points in their neighborhoods. The expression is:

[0074]

[0075] Among them, R represents the region division result, h i represents a single point in the abnormal set, C represents the core point set, and h j Indicates point h i Points in the neighborhood, B represents the boundary point set, C represents the core point set, h l represents the core point, st represents the additional conditions, N(h i ) means hi Neighborhood of;

[0076] Furthermore, the DBSCAN algorithm can identify abnormal areas in the power grid, which are composed of dense monitoring points and may indicate abnormal behavior in local areas of the power grid, such as overload and short circuit. Each cluster represents a potential abnormal power grid area. The application of the DBSCAN algorithm in power system anomaly detection provides strong support for timely identification and response to power grid anomalies, which helps to improve the stability and security of the power grid.

[0077] S5. Based on the abnormal power grid area, a fault detection model is established using SVM. SVM is a supervised learning algorithm that is particularly suitable for classification problems in high-dimensional space. It can find the optimal hyperplane to maximize the boundary distance between categories. Its expression is:

[0078] SVM(R)={y:y=sign(w T Φ(x)+b)};

[0079] Where SVM(R) represents the fault detection model, w represents the weight vector, b is the bias, Φ(x) represents the mapping function, and y represents the output of the fault detection model;

[0080] Furthermore, the SVM model is trained to distinguish between normal operation and fault conditions. The model takes data from abnormal grid areas as input and learns how to predict whether a fault has occurred based on the characteristics of this data. Once the model is trained, it can be applied to unknown abnormal areas to detect faults in real time.

[0081] After confirming the fault, the fault response mechanism is immediately activated to cut off the power supply to the fault area to prevent the fault from spreading further and ensure the stable operation of other parts of the power grid. After the fault area is isolated, the power grid is reconfigured to restore the power supply to the non-fault area to ensure that the normal operation of these areas is not affected. This process reflects the automation and intelligence of smart grid in fault detection and response, and is an important part of the management of distributed power access to the power grid.

[0082] S6. Kubernetes is an open source container orchestration platform for automating the deployment, expansion, and management of containerized applications. Kubernetes plays a particularly critical role in the distribution protection method for dealing with distributed power access, because it ensures that critical tasks such as fault detection and resource-intensive computing obtain the necessary computing resources, even when resource demand fluctuates greatly. Kubernetes continuously monitors resource usage on each node in the cluster through its agent component kubelet, including CPU usage and memory occupancy. Kubelet regularly collects this information and reports it to the API server, making the resource usage of the entire cluster transparent and visible. This data is not only used for monitoring, but also a key basis for dynamic resource allocation decisions. During the dynamic resource allocation process, Kubernetes continuously monitors the CPU usage and memory occupancy of each task, and when the resource demand of critical tasks increases, the Kubernetes scheduler re-evaluates resource allocation.

[0083] Furthermore, Kubernetes' dynamic resource allocation mechanism ensures that critical tasks can obtain the necessary resources even when resource demand fluctuates, thereby improving the system's responsiveness and efficiency. This is especially important for distribution protection systems that handle distributed power access, as such systems need to process large amounts of data in real time and make quick decisions to keep the power grid stable and secure.

[0084] This embodiment also provides a power distribution protection system based on distributed power supply access, including:

[0085] The data integration and preprocessing module integrates data from weather stations, historical databases and real-time sensors and performs preprocessing.

[0086] Feature extraction and modeling module, which extracts meaningful features from preprocessed data and builds a power supply and demand prediction model;

[0087] The protection threshold dynamic adjustment module automatically calculates and adjusts the protection threshold according to the power supply and demand forecast results;

[0088] The anomaly detection and area division module uses the DBSCAN clustering algorithm to identify abnormal power grid areas when the power supply exceeds the protection threshold;

[0089] The fault detection and processing module uses the fault detection model to identify faults based on abnormal areas, isolate the faulty areas, and restore power supply to non-faulty areas;

[0090] The resource dynamic allocation module uses the Kubernetes container management platform to dynamically adjust computing resource allocation according to the resource requirements of key tasks.

[0091] This embodiment also provides a computer device, which is suitable for the distribution protection method based on distributed power source access, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the distribution protection method based on distributed power source access proposed in the above embodiment.

[0092] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0093] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the distribution protection method based on distributed power supply access proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, disk or optical disk.

[0094] In summary, the present invention collects meteorological data, historical data and real-time sensor data, and performs preprocessing to form an initial data set that fully reflects the state of the power grid, thereby improving the accuracy of power supply and demand forecasts, reducing forecast errors caused by data quality issues, and enhancing the stability of power grid operation. Through accurate power supply and demand forecasts, the power grid can prepare in advance, avoiding overload or waste caused by imbalanced supply and demand, and improving the overall efficiency and economy of the power system. The adaptive ability and safety of the power grid are improved, system failures caused by improper setting of fixed thresholds are reduced, and the continuity and reliability of power supply are enhanced. The fault detection and response time is greatly shortened, the maintenance cost is reduced, and the user satisfaction and the overall service quality of the power grid are improved.

[0095] Example 2, referring to Table 1, is the second example of the present invention. To further verify the advancement of the present invention, experimental simulation data of a distribution protection method based on distributed power supply access is provided.

[0096] In order to verify the effectiveness and superiority of the distribution protection method based on distributed power access, this embodiment designs a simulation environment, which includes multiple types of distributed power sources (solar energy, wind energy) and complex power grid structures. In the experiment, meteorological data (temperature, humidity, wind speed, wind direction, light intensity) for one year, historical power grid data (distributed power generation and power grid load data for the past five years), and real-time sensor data (distributed power generation, power grid load, voltage, current) were collected. All data were preprocessed after collection, including data cleaning, missing value filling, outlier detection, etc., to ensure data quality.

[0097] Next, wavelet transform and Fourier transform were used to extract features from the preprocessed data and build a power supply and demand forecasting model. The model uses an LSTM network. Considering the dynamic characteristics of the power grid, the model can predict future power supply and demand based on the characteristic function. The accuracy and response speed of the forecasting model became the focus of this experiment.

[0098] Subsequently, the protection threshold is automatically adjusted based on the power supply and demand forecast results. In the experiment, the threshold adjustment mechanism is compared with the traditional fixed threshold method to evaluate its responsiveness and system stability when dealing with load fluctuations.

[0099] In the anomaly detection stage, the DBSCAN algorithm is used to identify abnormal power grid areas. In the experiment, the effectiveness of the proposed method in complex power grid environments is verified by comparing it with the traditional rule-based anomaly detection method.

[0100] The fault detection model is built based on the SVM algorithm. In the experiment, the accuracy and response time of the model, as well as the system's isolation and recovery capabilities when faults occur, are tested by simulating fault scenarios.

[0101] In the resource allocation phase, the Kubernetes platform was used to dynamically adjust resource allocation. In the experiment, the changes in system resource utilization and task response time when the demand for critical task resources increased were recorded and compared with the environment without a dynamic adjustment mechanism.

[0102] Table 1 Comparison between traditional power distribution protection technology and new power distribution protection technology

[0103]

[0104] By comparing the data of the examples, it can be seen that the method of the present invention has significant advantages over the existing technologies A and B in terms of power supply and demand prediction accuracy, protection threshold adjustment response time, anomaly detection accuracy, fault isolation time, and resource utilization improvement. The specific analysis is as follows:

[0105] Power supply and demand prediction accuracy: The prediction accuracy of the method of the present invention reaches 93.5%, which is much higher than 88.3% of the prior art A and 86.1% of the prior art B. This is due to the optimization of feature extraction and LSTM network model, which can more accurately predict future power supply and demand, providing a more reliable basis for power grid management.

[0106] Protection threshold adjustment response time: The response time of the method of the present invention is 150ms, which is much lower than 500ms of prior art A and 650ms of prior art B. The introduction of the dynamic adjustment mechanism enables the system to adjust the protection threshold more quickly when the load fluctuates, thereby enhancing the adaptive capability and stability of the power grid.

[0107] Anomaly detection accuracy: The anomaly detection accuracy of the method of the present invention reaches 98.2%, which is significantly better than 89.5% of the prior art A and 85.4% of the prior art B. The use of the DBSCAN algorithm improves the accuracy and robustness of anomaly detection, which helps to identify and respond to power grid anomalies earlier.

[0108] Fault isolation time: The fault isolation time of the method of the present invention is only 12 seconds, which is 34 seconds in prior art A and 48 seconds in prior art B, greatly shortening the fault response time, reducing the impact of the fault on the power grid, and improving the reliability of the power grid.

[0109] Improved resource utilization: The method of the present invention improves resource utilization by 23.4%, which is much higher than 15.2% of the prior art A and 12.5% ​​of the prior art B. The dynamic resource allocation mechanism of Kubernetes enables key tasks to obtain timely resource support, improving the operating efficiency and response speed of the overall system.

[0110] In summary, the method of the present invention has demonstrated significant superiority in multiple key indicators. It not only improves the accuracy of power supply and demand forecasting and enhances the self-regulation capability of the power grid, but also significantly improves the efficiency of fault detection and response, as well as the flexibility of resource allocation, providing strong technical support for building an intelligent, efficient and stable power grid system.

[0111] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A power distribution protection method based on distributed power supply access, characterized in that: include: Integrate the collected meteorological data, historical data, and real-time sensor data as initial data and perform preprocessing; The preprocessed initial data is subjected to feature extraction through a characteristic function, and a power supply and demand prediction model is established based on the feature-extracted data to obtain a power supply and demand prediction result; Automatically adjust protection thresholds based on power supply and demand forecasts; When the power supply value exceeds the set protection threshold, the DBSCAN algorithm is used to divide the abnormal power grid area; Establish a fault detection model based on the abnormal power grid area to detect faults in the abnormal power grid area, and process the abnormal power grid area according to the detection results; Manage containers through Kubernetes and dynamically adjust resource allocation.

2. The power distribution protection method based on distributed power supply access according to claim 1, characterized in that: Obtain meteorological data including temperature, humidity, wind speed, wind direction, and light intensity from official weather stations; extract distributed power generation and grid load data for the past five years from the power company database; collect distributed power generation, grid load, voltage, and current from sensors in real time; Integrate the collected meteorological data, historical data, and real-time sensor data as initial data; Define the initial data as X = (X1, X2, X3, X4, X5, X6, X7, X8, X9); Among them, X1 represents temperature data, X2 represents humidity data, X3 represents wind speed data, X4 represents wind direction data, X5 represents light intensity data, X6 represents distributed power generation data, X7 represents grid load data, X8 represents voltage data, and X9 represents current data; The initial data is preprocessed and the expression is: Among them, D i represents the initial data after preprocessing, X represents the set of initial data, i represents the initial data group, j represents the observation point, n represents the number of missing values, median(X i ) represents the median of the data, IQR(X i ) represents the discreteness of the data, and k represents the detection threshold.

3. The power distribution protection method based on distributed power supply access according to claim 2, characterized in that: The principle of wavelet transform and Fourier transform is used to extract features from the preprocessed initial data, and its expression is: Among them, F(x) represents the data after feature extraction, σ represents the standard deviation, and D i represents the preprocessed data sequence, N represents the total number of preprocessed data sequences, f represents the frequency, φ represents the phase angle, s represents the time point, u represents the signal point, and i represents the index of the sequence; Based on the feature-extracted data, the LSTM network is used to establish a power supply and demand forecasting model. The feature-extracted data is input into the power supply and demand forecasting model to obtain the power supply and demand forecasting result, which is expressed as: Among them, P t represents the final power supply and demand forecast output, F(x) represents the element in the fused feature vector, softmax represents the activation function, h represents the hidden state vector, and W a Represents the weight matrix, LSTM t Represents the output of the LSTM network at time point t.

4. The power distribution protection method based on distributed power supply access according to claim 3, characterized in that: The power supply and demand forecast output results are transformed nonlinearly, and the protection threshold is automatically adjusted according to the power supply and demand forecast results after nonlinear transformation. The expression is: Among them, T represents the final calculated threshold, Q p represents the quantile function, ω represents the angular frequency, and φ represents the phase shift; The protection threshold T will be based on the power supply and demand forecast output result P t Make dynamic adjustments.

5. The power distribution protection method based on distributed power supply access according to claim 4, characterized in that: The DBSCAN algorithm is used to divide the abnormal points into regions and determine the abnormal power grid area. The expression is: Among them, R represents the region division result, h i represents a single point in the abnormal set, C represents the core point set, and h j Indicates point h i Points in the neighborhood, B represents the boundary point set, C represents the core point set, h l represents the core point, st represents the additional conditions, N(h i ) means h i 's neighborhood.

6. The power distribution protection method based on distributed power supply access according to claim 5, characterized in that: Based on the abnormal power grid area, SVM is used to establish a fault detection model, which is expressed as: SVM(R)={y:y=sign(w T Φ(x)+b)}; Where SVM(R) represents the fault detection model, w represents the weight vector, b represents the deviation, Φ(x) represents the mapping function, and y represents the output of the fault detection model; Perform fault detection on abnormal areas. After confirming the fault, cut off power to the faulty area, isolate the faulty area, and restore power to non-faulty areas.

7. The power distribution protection method based on distributed power supply access according to claim 6, characterized in that: Kubernetes manages the deployment and resource allocation of containers in the cluster through a component called a scheduler. During the dynamic resource allocation process, Kubernetes's agent component kubelet continuously monitors the CPU usage and memory usage of each task. Kubelet regularly collects this information and reports it to the API server. When the resource demand of critical tasks increases, the Kubernetes scheduler will re-evaluate resource allocation.

8. A power distribution protection system based on distributed power access, based on the power distribution protection method based on distributed power access according to any one of claims 1 to 7, characterized in that: include: Data integration and preprocessing module, which integrates data from weather stations, historical databases and real-time sensors and performs preprocessing; Feature extraction and modeling module, which extracts meaningful features from preprocessed data and builds a power supply and demand prediction model; The protection threshold dynamic adjustment module automatically calculates and adjusts the protection threshold according to the power supply and demand forecast results; The anomaly detection and area division module uses the DBSCAN clustering algorithm to identify abnormal power grid areas when the power supply exceeds the protection threshold; The fault detection and processing module uses the fault detection model to identify faults based on abnormal areas, isolate the faulty areas, and restore power supply to non-faulty areas; The resource dynamic allocation module uses the Kubernetes container management platform to dynamically adjust computing resource allocation according to the resource requirements of key tasks.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the power distribution protection method based on distributed power supply access described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the power distribution protection method based on distributed power source access described in any one of claims 1 to 7 are implemented.

Images