Power distribution method based on AI artificial intelligence and intelligent power distribution cabinet

Abstract

The invention relates to the technical field of data processing, and discloses a power distribution method based on AI artificial intelligence and an intelligent power distribution cabinet. The method comprises the following steps: carrying out real-time data acquisition on key nodes of the power distribution system through an intelligent sensor network to obtain original operation data of the power distribution system, and carrying out cleaning and standardization processing on the original operation data to obtain a standardized data set; performing analysis modeling on the standardized data set through multi-level hybrid model training to obtain an AI analysis model of the power distribution system; processing and analyzing the power distribution data collected in real time to obtain a load prediction result and a dynamic balance strategy; performing abnormal mode identification and multi-path reasoning on the system operation parameters to obtain equipment health state evaluation and maintenance suggestions; and based on the equipment health state assessment and maintenance suggestions, performing multi-objective optimization calculation on renewable energy source supply and power consumption demands to obtain an intelligent energy scheduling scheme. According to the invention, the efficiency and accuracy of the power distribution method based on AI artificial intelligence are improved.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a power distribution method and an intelligent power distribution cabinet based on AI artificial intelligence. Background Art

[0002] With the rapid development and intelligent transformation of power systems, traditional distribution methods have been unable to meet the complex needs of modern power grids. At present, many distribution systems use analysis methods based on historical data and static models to maintain system stability through regular inspections and manual intervention. These methods have improved the reliability and efficiency of distribution systems to a certain extent, but they have obvious limitations when processing large-scale, real-time, and multi-variable data. At the same time, some advanced distribution systems have begun to introduce artificial intelligence technologies, such as machine learning algorithms and neural networks, for load forecasting and fault diagnosis, and have achieved certain results.

[0003] However, existing distribution methods still face many challenges. First, the lack of real-time performance of data collection and processing leads to a slow response of the system to emergencies. Second, traditional single models or simple AI models are difficult to fully capture the complex dynamic characteristics of distribution systems, especially in terms of integrating renewable energy and handling uncertainties. In addition, existing anomaly detection and fault diagnosis methods are often passive and lagging, making it difficult to achieve active prevention and timely intervention. Finally, energy scheduling optimization usually only considers a single objective, such as economy or reliability, and lacks the ability to comprehensively balance multiple objectives. Summary of the invention

[0004] The present application provides an AI-based power distribution method and an intelligent power distribution cabinet, which are used to improve the efficiency and accuracy of the AI-based power distribution method.

[0005] In a first aspect, the present application provides an AI-based power distribution method, which includes: performing real-time data collection on key nodes of a power distribution system through an intelligent sensor network to obtain original power distribution system operation data; cleaning and standardizing the original power distribution system operation data through a data preprocessing algorithm to obtain a standardized data set; analyzing and modeling the standardized data set through multi-level hybrid model training to obtain an AI analysis model of the power distribution system; processing and analyzing the real-time collected power distribution data through the AI ​​analysis model to obtain load prediction results and dynamic balancing strategies; based on the load prediction results and dynamic balancing strategies, performing abnormal pattern recognition and multi-path reasoning on system operating parameters to obtain equipment health status assessment and maintenance recommendations; based on the equipment health status assessment and maintenance recommendations, performing multi-objective optimization calculations on renewable energy supply and electricity demand to obtain an intelligent energy scheduling solution.

[0006] In a second aspect, the present application provides an intelligent power distribution cabinet, the intelligent power distribution cabinet comprising:

[0007] The acquisition module is used to collect real-time data from key nodes of the power distribution system through an intelligent sensor network to obtain the original power distribution system operation data;

[0008] A processing module, used for cleaning and standardizing the original power distribution system operation data through a data preprocessing algorithm to obtain a standardized data set;

[0009] A modeling module, used to analyze and model the standardized data set through multi-level hybrid model training to obtain an AI analysis model of the power distribution system;

[0010] An analysis module, used to process and analyze the real-time collected power distribution data through the AI ​​analysis model to obtain load prediction results and dynamic balancing strategies;

[0011] An identification module is used to perform abnormal pattern recognition and multi-path reasoning on system operation parameters based on the load prediction results and dynamic balancing strategy to obtain equipment health status assessment and maintenance suggestions;

[0012] The calculation module is used to perform multi-objective optimization calculations on renewable energy supply and electricity demand based on the equipment health status assessment and maintenance recommendations to obtain an intelligent energy scheduling solution.

[0013] In the technical solution provided by this application, real-time data collection of key nodes of the distribution system is carried out through an intelligent sensor network, which can comprehensively and timely obtain the system operation status, provide a high-quality and time-effective original data basis for subsequent analysis, and significantly improve the comprehensiveness and real-time nature of data collection. The data preprocessing algorithm is used to clean and standardize the original distribution system operation data, which not only effectively removes the noise and outliers in the data, but also realizes the unified representation of the data, laying a solid data foundation for subsequent model training and analysis, and greatly improving the data quality and availability. The introduction of multi-level hybrid model training enables the system to simultaneously capture the temporal characteristics, spatial characteristics and decision characteristics of the distribution data, significantly enhances the expression ability and generalization performance of the model, and can more accurately characterize the dynamic characteristics of complex distribution systems. Through the AI ​​analysis model to process and analyze the real-time collected distribution data, the system can quickly and accurately generate load prediction results and dynamic balance strategies, greatly improving the prediction accuracy and response speed of the distribution system, and providing reliable support for real-time scheduling and decision-making. Based on the load forecast results and dynamic balancing strategies, the abnormal pattern recognition and multi-path reasoning of the system operating parameters can not only timely detect potential equipment failures and system anomalies, but also provide accurate equipment health status assessment and maintenance suggestions, significantly improving the reliability and maintenance efficiency of the system. Finally, through multi-objective optimization calculation of renewable energy supply and electricity demand, the intelligent energy scheduling scheme can maximize the utilization of renewable energy while ensuring power supply reliability, and achieve a balance between economy and environmental protection, greatly improving the overall operating efficiency and sustainability of the distribution system. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] 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 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.

[0015] Figure 1 This is a schematic diagram of an embodiment of a power distribution method based on AI artificial intelligence in an embodiment of the present application;

[0016] Figure 2 This is a schematic diagram of an embodiment of an intelligent power distribution cabinet in an embodiment of the present application. DETAILED DESCRIPTION

[0017] The embodiment of the present application provides a power distribution method and intelligent distribution cabinet based on AI artificial intelligence. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0018] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of the power distribution method based on AI artificial intelligence includes:

[0019] Step S101: collect real-time data from key nodes of the power distribution system through an intelligent sensor network to obtain original power distribution system operation data;

[0020] Step S102: cleaning and standardizing the original power distribution system operation data by using a data preprocessing algorithm to obtain a standardized data set;

[0021] Step S103: Analyze and model the standardized data set through multi-level hybrid model training to obtain an AI analysis model of the power distribution system;

[0022] Step S104: Process and analyze the real-time collected power distribution data through the AI ​​analysis model to obtain load prediction results and dynamic balancing strategies;

[0023] Step S105: Based on the load prediction results and the dynamic balancing strategy, abnormal pattern recognition and multi-path reasoning are performed on the system operation parameters to obtain equipment health status assessment and maintenance recommendations;

[0024] Step S106: Based on the equipment health status assessment and maintenance recommendations, multi-objective optimization calculations are performed on renewable energy supply and electricity demand to obtain an intelligent energy scheduling solution.

[0025] It is understandable that the execution subject of the present application may be a smart power distribution cabinet, or a terminal or a server, which is not limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0026] Specifically, real-time data collection is performed on key nodes of the power distribution system through an intelligent sensor network. These intelligent sensors are arranged in locations such as substations, distribution cabinets and user terminals to collect parameters such as voltage, current, power factor, harmonic content, etc. The acquisition frequency can reach milliseconds to ensure the real-time nature of the data. The collected raw data is cleaned and standardized by the data preprocessing algorithm. During the cleaning process, the median filter algorithm is used to remove short-term fluctuations and outliers, and the missing data is interpolated. The standardization process includes Z-score standardization and unique hot encoding, so that different types of data can be processed uniformly. The preprocessed data forms a standardized data set, which provides a basis for subsequent modeling. The standardized data set is analyzed and modeled through multi-level hybrid model training. This process includes building multiple structures such as deep neural networks, long short-term memory networks, attention mechanisms, and convolutional neural networks. Deep neural networks are used to capture nonlinear relationships, long short-term memory networks process time series data, attention mechanisms focus on important features, and convolutional neural networks analyze spatial features. These network structures form an AI analysis model that can comprehensively analyze the characteristics of the power distribution system through iterative training and parameter optimization.

[0027] The AI ​​analysis model processes and analyzes the real-time collected distribution data to generate load forecast results and dynamic balancing strategies. Load forecasting includes short-term (15 minutes to 1 hour), medium-term (1 day to 1 week) and long-term (1 month to 1 year) forecasts. The forecasting process considers many factors such as historical data patterns, weather forecasts, and social events. The dynamic balancing strategy is based on the load forecast results and automatically adjusts the distribution parameters, such as the transformer tap position and the capacitor bank switching status, through the reinforcement learning algorithm to optimize the power quality and reduce line losses. Based on the load forecast results and dynamic balancing strategy, abnormal pattern recognition and multi-path reasoning are performed on the system operation parameters. Abnormal pattern recognition identifies potential abnormal situations by comparing real-time parameters with historical normal operation data. Multi-path reasoning uses pre-built knowledge graphs combined with abnormal patterns to infer possible causes of failures. Output equipment health status assessment and maintenance recommendations to provide a basis for preventive maintenance.

[0028] Finally, based on the equipment health status assessment and maintenance recommendations, a multi-objective optimization calculation is performed on the renewable energy supply and electricity demand. This step considers multiple factors such as renewable energy generation forecast, user electricity demand forecast, equipment health status, etc., and generates an intelligent energy scheduling plan through a multi-objective optimization algorithm. This plan maximizes the utilization of renewable energy while meeting electricity demand, and balances multiple goals such as economy, reliability, and environmental protection.

[0029] For example, in a regional power distribution system, on a weekday in spring, the smart sensor network collects data every minute between 8:00 and 9:00 in the morning. In the original data, the voltage of a distribution cabinet at 8:30 was 220.5V, the current was 100.2A, and the power factor was 0.92. After data preprocessing, the voltage value was standardized to 0.025 (assuming a mean of 220V and a standard deviation of 20V), and the current value was standardized to -0.01 (assuming a mean of 101A and a standard deviation of 80A). After processing these data, the AI ​​analysis model predicted that the peak load from 12:00 to 13:00 on that day would be 5.2MW, an increase of 8% over the same period of the previous week. Based on this prediction, the dynamic balancing strategy recommends that a group of nearby capacitors be put into use at 11:30 to cope with the upcoming load peak. At the same time, the system detected that the temperature of a transformer was 2°C higher than usual. Through multi-path reasoning, it was inferred that the efficiency of the cooling system might have decreased, and it was recommended to focus on inspection during routine maintenance next week. Finally, considering that the weather was fine and photovoltaic power generation was expected to be good, the intelligent dispatching plan recommended transferring part of the load to the photovoltaic power generation system between 13:00 and 15:00 to reduce the use of conventional power sources, which is expected to reduce carbon emissions by about 2 tons. This series of decisions and operations demonstrates how AI-based power distribution methods can achieve comprehensive, real-time, and intelligent grid management.

[0030] In the embodiment of the present application, by collecting real-time data of key nodes of the power distribution system through an intelligent sensor network, the system operation status can be obtained comprehensively and timely, providing a high-quality and time-effective original data basis for subsequent analysis, significantly improving the comprehensiveness and real-time nature of data collection. The data preprocessing algorithm is used to clean and standardize the original distribution system operation data, which not only effectively removes the noise and outliers in the data, but also realizes the unified representation of the data, laying a solid data foundation for subsequent model training and analysis, and greatly improving the data quality and availability. The introduction of multi-level hybrid model training enables the system to simultaneously capture the time series characteristics, spatial characteristics and decision characteristics of the distribution data, significantly enhancing the expression ability and generalization performance of the model, and more accurately characterizing the dynamic characteristics of complex distribution systems. Through the AI ​​analysis model to process and analyze the real-time collected distribution data, the system can quickly and accurately generate load prediction results and dynamic balancing strategies, greatly improving the prediction accuracy and response speed of the distribution system, and providing reliable support for real-time scheduling and decision-making. Based on the load forecast results and dynamic balancing strategies, the abnormal pattern recognition and multi-path reasoning of the system operating parameters can not only timely detect potential equipment failures and system anomalies, but also provide accurate equipment health status assessment and maintenance suggestions, significantly improving the reliability and maintenance efficiency of the system. Finally, through multi-objective optimization calculation of renewable energy supply and electricity demand, the intelligent energy scheduling scheme can maximize the utilization of renewable energy while ensuring power supply reliability, and achieve a balance between economy and environmental protection, greatly improving the overall operating efficiency and sustainability of the distribution system.

[0031] In a specific embodiment, the process of executing step S101 may specifically include the following steps:

[0032] (1) Intelligent sensors are deployed in the substations, distribution cabinets and user terminals of the power distribution system to obtain a multi-level data acquisition network, and communication protocols are configured for the multi-level data acquisition network to obtain a high-speed real-time data transmission channel;

[0033] (2) Continuously monitor voltage, current, power factor and harmonic content through a high-speed real-time data transmission channel to obtain multi-dimensional power parameter data, and timestamp the multi-dimensional power parameter data to obtain a time-series-related original data stream;

[0034] (3) Performing a data integrity check on the original data stream associated with the time series to obtain a valid data set, and performing outlier detection on the valid data set to obtain a preliminarily cleaned data set;

[0035] (4) Smoothing the preliminarily cleaned data set through a sliding window algorithm to obtain a denoised data sequence, and interpolating missing values ​​in the denoised data sequence to obtain a complete data sequence;

[0036] (5) Standardize and normalize the complete data sequence to obtain a standard data set with unified dimensions, and perform feature extraction on the standard data set with unified dimensions to obtain a feature vector after dimensionality reduction;

[0037] (6) Using the principal component analysis algorithm to identify key features of the feature vector after dimensionality reduction, obtain a core feature set, and perform correlation analysis on the core feature set to obtain a feature correlation matrix;

[0038] (7) A multidimensional data cube is constructed based on the feature association matrix to obtain a multi-angle data analysis view, and data clustering is performed on the multi-angle data analysis view to obtain the data distribution pattern.

[0039] Specifically, intelligent sensors are arranged in the substations, distribution cabinets and user terminals of the distribution system to form a multi-level data acquisition network. These intelligent sensors include voltage sensors, current sensors, power factor measurement devices and harmonic analyzers. By configuring high-speed communication protocols such as 5G or industrial Ethernet, a high-speed real-time data transmission channel is established to ensure the timeliness and integrity of the data. This multi-level network structure can fully cover all key nodes of the distribution system and provide comprehensive and detailed system operation status information. Using these high-speed transmission channels, the voltage, current, power factor and harmonic content are continuously monitored. The monitoring frequency can reach milliseconds to form multi-dimensional power parameter data. Each piece of data is attached with an accurate timestamp to generate a time-series-related raw data stream. This time-series association ensures the time consistency of the data and provides an important basis for subsequent analysis. Accurate recording of timestamps is crucial for analyzing system dynamic changes and identifying transient anomalies.

[0040] Perform data integrity checks on the original data streams associated with the time series, including checking the integrity of the data packets and the continuity of the timestamps, to obtain a valid data set. Then perform outlier detection on the valid data set, use statistical methods or machine learning algorithms to identify and mark abnormal data points, and obtain a preliminary cleaned data set. This step can effectively filter out invalid or erroneous data caused by factors such as sensor failure and communication interference. Smooth the preliminary cleaned data set using a sliding window algorithm. The core idea of ​​the sliding window algorithm is to perform average or median filtering on the data within a fixed-size window, and slide the window along the time axis to smooth the entire data sequence. This process effectively removes short-term fluctuations and noise to obtain a denoised data sequence. For missing values ​​in the denoised data sequence, interpolation methods such as linear interpolation or spline interpolation are used to complete the missing values ​​to obtain a complete data sequence. The size of the sliding window has an important impact on the results and needs to be adjusted according to data characteristics and application requirements.

[0041] The complete data sequence is standardized and normalized to unify data of different dimensions to the same scale. The Z-score method is usually used for standardization to convert the data into a distribution with a mean of 0 and a standard deviation of 1. The Min-Max method is often used for normalization to map the data to the interval [0,1]. After this processing, a standard data set with unified dimensions is obtained, which is convenient for subsequent feature extraction and model training. Then, feature extraction is performed on the standard data set, such as extracting time domain features and frequency domain features, to obtain the feature vector after dimensionality reduction. The principal component analysis (PCA) algorithm is used to identify key features of the feature vector after dimensionality reduction. The core of PCA is to calculate the eigenvalues ​​and eigenvectors of the covariance matrix, and select the eigenvectors corresponding to the largest eigenvalues ​​as the principal components. This step can effectively reduce the data dimension while retaining the main information of the data. The eigenvectors corresponding to the first k largest eigenvalues ​​are selected to form the core feature set. The core feature set is subjected to correlation analysis, and the correlation coefficient matrix between features is calculated to obtain the feature correlation matrix. This matrix reflects the relationship between different features and is of great significance for understanding system behavior and optimizing model structure. A multidimensional data cube is constructed based on the feature association matrix, with each dimension representing a core feature. This multidimensional structure allows data to be analyzed from different angles, forming a multi-angle data analysis view. The multidimensional data cube facilitates flexible query and in-depth analysis of data. Finally, data clustering is performed on the multi-angle data analysis view, such as using the K-means algorithm to classify similar data points into one category to obtain the data distribution pattern. This clustering result can help identify the typical operating status and abnormal patterns of the system.

[0042] For example, a power distribution system collects data every 5 minutes in a day, and a total of 288 sets of data are collected. Each set of data contains the voltage, current, power factor and harmonic content of 10 measurement points. In the original data, the voltage of a certain measurement point at 8:00 is 220.5V, the current is 100.2A, the power factor is 0.92, and the total harmonic distortion is 2.5%. The data integrity check found that 2 sets of data were missing, and the outlier detection identified 5 abnormal points. A 30-minute (6 data points) sliding window was used for smoothing, and the abnormal 220.5V was smoothed to 219.8V. After standardization, the voltage value becomes 0.025 (assuming a mean of 220V and a standard deviation of 20V). After PCA analysis, 5 principal components that explain 95% of the variance are retained. K-means clustering (k=3) divides the data into three categories: low load, medium load and high load, of which low load accounts for 25%, medium load accounts for 60%, and high load accounts for 15%.

[0043] In a specific embodiment, the process of executing step S102 may specifically include the following steps:

[0044] (1) Time series alignment is performed on the original distribution system operation data to obtain a synchronized data stream, and data integrity check is performed on the synchronized data stream to obtain a valid data set;

[0045] (2) Perform outlier detection on the valid data set to obtain a preliminary cleaned data set, and remove noise from the preliminary cleaned data set using a median filter algorithm to obtain a smoothed data sequence;

[0046] (3) Interpolate missing values ​​in the smoothed data series to obtain a complete data series, and remove duplicate values ​​from the complete data series to obtain a streamlined data set;

[0047] (4) converting the data type of the streamlined data set to obtain a data set in a unified format, and performing dimension normalization processing on the data set in the unified format to obtain a standardized data set;

[0048] (5) Extract features from the standardized data set to obtain a feature vector set, and perform dimensionality reduction on the feature vector set through principal component analysis to obtain a key feature set;

[0049] (6) Perform correlation analysis on the key feature set to obtain a feature correlation matrix, and perform data clustering based on the feature correlation matrix to obtain the data distribution pattern.

[0050] Specifically, the original distribution system operation data is time-series aligned to ensure that the data from different measurement points are consistent in time. Time series alignment checks the timestamp of each data point and rearranges the data from different sources at a uniform time interval to obtain a synchronized data stream. Subsequently, the synchronized data stream is checked for data integrity, including checking whether there are missing values, duplicate values, or obvious outliers, so as to obtain a valid data set. Outlier detection is performed on the valid data set. Outlier detection can be performed using statistical methods such as the triple standard deviation method or the quartile method, or machine learning algorithms such as isolation forests. The detected outliers are marked or temporarily removed to obtain a preliminary cleaned data set. Then, the preliminary cleaned data set is de-noised using the median filtering algorithm. The principle of median filtering is to replace the value at the center of the window with the median of the data in the sliding window. This method is particularly suitable for removing impulse noise while retaining the edge characteristics of the data well. After median filtering, a smoothed data series is obtained.

[0051] The next key step is to interpolate the missing values ​​of the smoothed data series. Missing value interpolation methods include simple linear interpolation, more complex spline interpolation, or machine learning-based methods such as KNN interpolation. The choice of an appropriate interpolation method depends on the characteristics of the data and the distribution pattern of the missing values. After interpolation, a complete data series is obtained. Subsequently, duplicate values ​​are removed from the complete data series. This step can reduce data redundancy and improve the efficiency of subsequent processing, and finally a streamlined data set is obtained. The streamlined data set is converted to ensure that all data is stored in a unified format, such as converting all numerical values ​​to floating point numbers. This can avoid calculation errors caused by inconsistent data types. Then, the unified format data set is dimensionally normalized. Common normalization methods include Min-Max normalization and Z-score normalization. Normalization converts data of different dimensions to the same scale, which facilitates subsequent feature extraction and model training, and finally obtains a standardized data set.

[0052] Feature extraction of standardized data sets is an important step to further reduce data dimensions and extract key information. Feature extraction can include time domain features (such as mean, variance, peak), frequency domain features (such as Fourier transform coefficients), etc. The extracted features constitute a feature vector set. Subsequently, the feature vector set is subjected to dimensionality reduction processing by principal component analysis (PCA). PCA calculates the eigenvalues ​​and eigenvectors of the covariance matrix, selects the principal components with the largest contribution, achieves data dimensionality reduction, and retains the main information of the data to obtain the key feature set. Finally, the key feature set is subjected to correlation analysis, and the correlation coefficients between the features are calculated to obtain the feature association matrix. This matrix reflects the strength of the linear relationship between different features. Based on the feature association matrix, data clustering can be performed, such as using K-means or hierarchical clustering algorithms to classify similar data points into one category to obtain the data distribution pattern. This clustering result helps to understand the intrinsic structure and distribution characteristics of the data.

[0053] For example, a power distribution system collects data every 5 minutes, including voltage, current, power factor and harmonic content at 10 measurement points. In one day's collection, there are 288 groups of raw data (24 hours × 12 times / hour). After time series alignment, it was found that there were data asynchronization problems at 3 time points, which were corrected by linear interpolation. Data integrity check found that 5 groups of data were missing and 2 groups of data were repeated. Outlier detection used the triple standard deviation method to identify 7 abnormal points, including 3 voltage abnormalities and 4 current abnormalities. Median filtering used a 5-point window to effectively smooth the data curve. Missing value interpolation used the spline interpolation method to successfully complete 5 groups of missing data. After removing duplicate values, the amount of data was reduced to 285 groups. The data type was uniformly converted to 64-bit floating point numbers. Normalization used the Min-Max method to map all data to the [0,1] interval. Feature extraction obtained 20 features for each measurement point, totaling 200 features. PCA analysis retained the top 15 principal components that explained 95% of the variance. Correlation analysis found that there is a strong correlation between voltage and power factor, and between current and harmonic content (correlation coefficient is greater than 0.8). Finally, K-means clustering (k=3) divides the data into three categories: low load, medium load and high load, accounting for 23%, 62% and 15% respectively.

[0054] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0055] (1) Perform data segmentation on the standardized data set to obtain a training set and a validation set, and perform feature engineering on the training set to obtain an enhanced feature set;

[0056] (2) A deep neural network with five fully connected layers is constructed for the enhanced feature set to obtain the basic network structure, and a batch normalization layer and a rectified linear unit activation function are added between every two layers in the basic network structure to obtain a multi-layer neural network architecture;

[0057] (3) Add three bidirectional long short-term memory layers after the input layer of the multi-layer neural network architecture, each layer contains 128 neurons, and obtain the time series analysis network. Then apply the multi-head self-attention mechanism on the output of the long short-term memory layer of the time series analysis network, with the number of heads set to 8, to obtain the weight distribution of key features;

[0058] (4) Input the key feature weight distribution into three layers of two-dimensional convolutional layers, using a 3x3 convolution kernel and 32 filters in each layer to obtain a spatial feature analysis network, and add a skip connection after each convolution layer in the spatial feature analysis network to obtain a deep feature representation;

[0059] (5) The deep feature representation is input into a dual-depth Q network structure, which includes two identical Q networks, each of which contains three fully connected layers with 256, 128, and 64 neurons, respectively, to obtain a decision optimization network, and the pre-trained knowledge graph embedding vector is connected to the second layer output of the decision optimization network to obtain a knowledge-enhanced decision representation;

[0060] (6) A transfer learning structure consisting of two fully connected layers is added after the knowledge-enhanced decision representation, with each layer containing 64 neurons to obtain an adaptive optimization network. The learning rate, batch size, and random inactivation rate of the adaptive optimization network are tuned through 5-fold cross validation to obtain an AI analysis model of the distribution system.

[0061] Specifically, the standardized data set is split, usually 80% as the training set and 20% as the validation set. The training set is subjected to feature engineering, including feature selection, feature combination and feature transformation, to obtain an enhanced feature set. Feature engineering aims to improve the expressiveness and generalization performance of the model. A deep neural network with 5 fully connected layers is constructed based on the enhanced feature set to form a basic network structure. Batch normalization layers and rectified linear unit activation functions are added between every two layers to obtain a multi-layer neural network architecture. Batch normalization helps alleviate the problem of internal covariate shift and accelerate network convergence, while the rectified linear unit activation function can effectively alleviate the problem of gradient vanishing and improve the nonlinear expressiveness of the model.

[0062] Three bidirectional long short-term memory layers are added after the input layer of the multi-layer neural network architecture, each layer contains 128 neurons to form a time series analysis network. The bidirectional long short-term memory network can capture the long-term dependencies of the data while considering the context information. The multi-head self-attention mechanism is applied to the output of the long short-term memory layer of the time series analysis network, and the number of heads is set to 8 to obtain the weight distribution of key features. The multi-head self-attention mechanism allows the model to focus on information in different subspaces, enhancing the diversity of feature extraction. The key feature weight distribution is input into three layers of two-dimensional convolutional layers, each layer uses a 3x3 convolution kernel and 32 filters to construct a spatial feature analysis network. The convolution operation can effectively capture local features and spatial correlation. A skip connection is added after each convolution layer in the spatial feature analysis network to obtain a deep feature representation. The skip connection helps to alleviate the gradient vanishing problem of the deep network while retaining the shallow feature information.

[0063] The deep feature representation is input into a dual-depth Q network structure, which contains two identical Q networks, each of which contains 3 fully connected layers with 256, 128, and 64 neurons, forming a decision optimization network. The dual-depth Q network structure can improve the stability and accuracy of decision making. The pre-trained knowledge graph embedding vector is connected to the second-layer output of the decision optimization network to obtain a knowledge-enhanced decision representation. The introduction of the knowledge graph provides domain knowledge support for the model and improves the interpretability and accuracy of the decision. Finally, a transfer learning structure containing 2 fully connected layers, each containing 64 neurons, is added after the knowledge-enhanced decision representation to obtain an adaptive optimization network. The transfer learning structure enables the model to adapt to new scenarios and tasks. The learning rate, batch size, and random dropout rate of the adaptive optimization network are tuned through 5-fold cross validation, and the AI ​​analysis model of the distribution system is finally obtained.

[0064] For example, the data set of a power distribution system contains one year of historical data, a total of 365 days, 288 time points per day (one sampling point every 5 minutes), and each time point contains the voltage, current, power factor and harmonic content data of 10 measurement points. After data segmentation, the training set contains 292 days of data, and the validation set contains 73 days of data. After feature engineering processing, the original features are expanded from 40 (10 measurement points × 4 parameters) to 100, including time features (such as hours, days of the week, whether it is a holiday), lag features (data from the previous 1 hour, the previous 24 hours) and statistical features (such as sliding average, standard deviation).

[0065] The number of neurons in the five fully connected layers of the deep neural network is 512, 256, 128, 64, and 32 respectively. The batch normalization layer normalizes the output of each layer, keeping the mean at 0 and the variance at 1. The bidirectional long short-term memory network processes the 24-hour historical data sequence to capture the intraday load change pattern. The multi-head self-attention mechanism calculates 8 different sets of attention weights to highlight the influence of important time points and features. The convolutional network treats the data of 10 measurement points as a 10×10 two-dimensional grid and extracts local correlations through a 3×3 convolution kernel. The jump connection directly adds the output of the first layer of convolution to the output of the third layer of convolution to fuse low-level and high-level features. The dual-depth Q network structure calculates the Q value of the current state and the target Q value respectively to reduce the over-estimation bias. The knowledge graph embedding contains information such as equipment topology and historical fault records, with a dimension of 128.

[0066] After the model training is completed, the transfer learning structure is fine-tuned with a small amount of new scenario data (such as newly added distribution lines) to adapt to the new working environment. During the 5-fold cross-validation process, the learning rate gradually decayed from 0.001 to 0.00001, the batch size tried three values ​​of 32, 64, and 128, and the random inactivation rate was adjusted between 0.1 and 0.5. The final model achieved 95% load prediction accuracy on the validation set (within ±5%), the F1 score of anomaly detection reached 0.92, and the dynamic balancing strategy improved the power quality by 12% and the overall system efficiency by 8%. This complex AI analysis model can not only accurately predict load changes, but also identify abnormal conditions in a timely manner.

[0067] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0068] (1) Extract features from the real-time collected power distribution data to obtain a real-time feature set, and input the real-time feature set into the AI ​​analysis model to obtain preliminary prediction results;

[0069] (2) Perform short-term load forecasting on the preliminary forecast results to obtain load forecast data for 15 minutes to 1 hour, and perform medium-term load forecasting on the load forecast data to obtain load forecast trends for 1 day to 1 week;

[0070] (3) Based on the load forecast trend, long-term load forecasting is performed to obtain a load forecast plan for 1 month to 1 year, and error analysis is performed on the load forecast plan to obtain a forecast credibility assessment;

[0071] (4) Analyze the load distribution of the distribution network according to the prediction credibility assessment, obtain the load prediction results, and generate a load balancing plan based on the load prediction results to obtain a preliminary dynamic balancing strategy;

[0072] (5) Conduct multi-scenario simulations on the preliminary dynamic balance strategy to obtain a strategy effectiveness evaluation, and optimize parameters based on the strategy effectiveness evaluation to obtain a dynamic balance strategy.

[0073] Specifically, the real-time collected distribution data is feature extracted to obtain a real-time feature set. The feature extraction process includes the calculation of time domain features (such as mean, standard deviation, peak factor), frequency domain features (such as power spectrum density, main frequency components) and statistical features (such as skewness, kurtosis). These features fully describe the current operating status of the distribution system. The extracted real-time feature set is input into a pre-trained AI analysis model, which may be a deep neural network, a long short-term memory network, or a hybrid structure of the two. The model processes these features and outputs preliminary prediction results, including short-term load trends and possible abnormalities. Based on the preliminary prediction results, short-term load forecasting is performed to obtain load forecast data for 15 minutes to 1 hour. Short-term forecasting uses time series analysis methods, such as autoregressive integrated moving average model (ARIMA) or exponential smoothing, taking into account recent load change trends and periodic patterns. Subsequently, medium-term load forecasting is performed on these short-term load forecast data to obtain load forecast trends for 1 day to 1 week. Medium-term forecasts incorporate more external factors such as weather forecasts, weekday / weekend patterns, special events, etc., using more complex models such as Support Vector Regression (SVR) or Random Forest algorithms.

[0074] Based on the medium-term load forecast trend, long-term load forecasting is performed to obtain a load forecast plan for 1 month to 1 year. Long-term forecasting needs to consider more macro factors, such as seasonal changes, economic growth trends, population changes, etc. Deep learning models such as long short-term memory networks (LSTM) or temporal convolutional networks (TCN) may be used here, which can capture long-term dependencies. Error analysis is performed on long-term load forecasting planning, and the deviation between historical forecasts and actual loads is calculated. Indicators such as root mean square error (RMSE) or mean absolute percentage error (MAPE) are used to obtain a forecast credibility assessment. This assessment reflects the reliability of the forecast results and provides a reference for subsequent decision-making. Based on the forecast credibility assessment, the load distribution of the distribution network is analyzed. This step takes into account factors such as network topology, substation capacity, and line load capacity. The load distribution in the network is analyzed using the flow calculation method to obtain detailed load forecast results. Based on these load forecast results, a load balancing solution is generated. The balancing solution may include measures such as adjusting transformer taps, switching capacitor banks, and dispatching controllable loads. The goal is to minimize network losses and balance the load levels of each node. This process forms a preliminary dynamic balancing strategy.

[0075] Finally, a multi-scenario simulation is performed on the preliminary dynamic balance strategy. The simulation process takes into account a variety of possible situations, such as sudden large loads, fluctuations in renewable energy generation, and equipment failures. Using the Monte Carlo simulation method, a large number of random scenarios are generated to evaluate the performance of the strategy under different conditions and obtain a strategy effect evaluation. Based on the evaluation results, the strategy parameters are optimized, and reinforcement learning algorithms such as deep Q networks (DQN) or policy gradient methods may be used to continuously adjust and improve the strategy, and finally a dynamic balance strategy with strong robustness and good adaptability is obtained.

[0076] For example, a smart power distribution system covers the commercial and residential areas of a medium-sized city and contains 100 distribution nodes. Real-time data collection shows that at 10 a.m. on weekdays, the total load is 50MW, the power factor is 0.92, and the voltage deviation is within the range of ±3%. After feature extraction, the real-time feature set obtained includes load growth rate (2% every 15 minutes), voltage fluctuation frequency (10 times per hour), etc. The AI ​​analysis model predicts that the load will continue to grow in the next hour and may reach 55MW. The short-term load forecast uses the ARIMA model to predict that the load will be 54.8MW after 1 hour, with an error range of ±1MW. The medium-term forecast takes into account the large-scale event in three days, and predicts that the peak load on that day may reach 65MW. The long-term forecast combines the city development plan and predicts that the average daily load will increase by 10% to 55MW one year later. Error analysis shows that the MAPE of short-term forecast is 2%, medium-term is 5%, and long-term is 8%.

[0077] Load distribution analysis found that the load rate of a substation in the commercial area has reached 85%, posing an overload risk. The initial dynamic balancing strategy recommends transferring part of the load to a nearby substation with a load rate of only 60%. Multi-scenario simulations took into account the situation where sudden high temperature weather causes a surge in air conditioning load. The simulation results showed that the initial strategy may cause local overload in extreme cases. After parameter optimization, the final dynamic balancing strategy reduced network losses by 7% and improved load balancing by 15% while ensuring power supply reliability.

[0078] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0079] (1) Monitor the system operating parameters in real time to obtain a parameter change sequence, and compare the parameter change sequence with historical normal operating data to obtain a set of deviation values;

[0080] (2) Perform cluster analysis on the deviation value set to obtain an abnormal pattern candidate set, and perform feature extraction on the abnormal pattern candidate set to obtain an abnormal feature vector;

[0081] (3) Inputting the abnormal feature vector into the pre-trained anomaly detection model to obtain the abnormal probability distribution, and determining the abnormal level according to the abnormal probability distribution to obtain the abnormal state assessment of the equipment;

[0082] (4) Based on the abnormal status evaluation of the equipment, multi-path reasoning is performed in the knowledge graph to obtain a set of fault causes, and the set of fault causes is prioritized to obtain the fault diagnosis result;

[0083] (5) Generate an equipment health status report based on the fault diagnosis results to obtain an equipment health status assessment, and formulate a preventive maintenance plan based on the equipment health status assessment to obtain maintenance recommendations.

[0084] Specifically, the system operating parameters are monitored in real time to obtain parameter change sequences. These parameters include key indicators such as voltage, current, power factor, and harmonic content. Data is collected at a sampling frequency of milliseconds through high-precision sensors to form a continuous parameter change sequence. Subsequently, these parameter change sequences are compared with historical normal operation data. The comparison process uses sliding window technology to calculate the parameter statistical characteristics (such as mean, standard deviation, peak value, etc.) in the current time window, and compares them with the statistical characteristics of historical data in the same period to obtain a set of deviation values. Cluster analysis is performed on the deviation value set, and clustering algorithms such as K-means or DBSCAN are used to classify similar deviation patterns to obtain a candidate set of abnormal patterns. The clustering process considers multiple dimensions such as the amplitude, duration, and change trend of the deviation. Feature extraction is performed on the candidate set of abnormal patterns, including time domain features (such as mean, variance, peak factor), frequency domain features (such as power spectral density, main frequency components), and statistical features (such as skewness and kurtosis) to obtain abnormal feature vectors. These features comprehensively describe all aspects of the abnormal pattern.

[0085] The abnormal feature vector is input into the pre-trained anomaly detection model. The model can be an autoencoder based on deep learning or an isolation forest algorithm based on ensemble learning. The model outputs an abnormal probability distribution, which reflects the degree to which the current operating state deviates from the normal state. According to the abnormal probability distribution, different thresholds are set to divide the abnormal state into different levels such as slight, moderate, and severe, and obtain the abnormal state assessment of the equipment. This classification method helps to reasonably allocate maintenance resources. Based on the abnormal state assessment of the equipment, multi-path reasoning is performed in the pre-built knowledge graph. The knowledge graph contains information such as equipment topological relationships, historical fault records, and expert experience. Multi-path reasoning uses graph traversal algorithms, such as depth-first search or breadth-first search, starting from the abnormal equipment node, to explore possible fault propagation paths and obtain a potential set of fault causes. The set of fault causes is prioritized, and factors such as causal relationship strength, historical occurrence frequency, and potential impact are considered. A weighted scoring method is used to obtain the final fault diagnosis result.

[0086] Generate a device health status report based on the fault diagnosis results, including detailed information such as abnormal parameters, degree of abnormality, possible causes of failure and their probability. Develop a preventive maintenance plan based on the equipment health status assessment. The maintenance plan takes into account multiple factors such as equipment importance, fault urgency, maintenance cost and system operation impact, and obtains maintenance recommendations that balance various needs through multi-objective optimization algorithms such as genetic algorithms or particle swarm optimization.

[0087] For example, a transformer in a distribution station was abnormal during continuous monitoring. Real-time monitoring data showed that the oil temperature of the transformer showed a slow upward trend in the past 24 hours, from the normal 65°C to 75°C. Compared with the historical data of the same period, it was found that the temperature deviation exceeded 3 standard deviations. Cluster analysis classified this temperature anomaly pattern as a "slow temperature rise" type. The abnormal feature vectors obtained by feature extraction include temperature rise rate (0.42°C / hour), duration (24 hours), intraday fluctuation range (±1.5°C), etc. The anomaly detection model analyzes these features and outputs an abnormal probability of 0.87, classifying it as a moderate anomaly. During the knowledge graph reasoning process, considering the location of the transformer, the status of the connected equipment, and the historical fault records, three possible causes of the fault were derived: reduced cooling system efficiency (probability 0.6), excessive load (probability 0.3), and insulation aging (probability 0.1). The health status report records these findings in detail and gives maintenance recommendations: check the cooling system first, monitor the load at the same time, and arrange insulation testing at the next planned power outage if the problem persists.

[0088] The preventive maintenance plan recommends arranging a cooling system inspection within 72 hours. Considering that the transformer's power supply range includes an important industrial park, it is recommended to carry out maintenance on weekends when the load is lower to minimize the impact on users. This example shows how AI-driven distribution systems can achieve full-process intelligent management from anomaly detection to fault diagnosis to maintenance decision-making, greatly improving system reliability and operation and maintenance efficiency.

[0089] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0090] (1) Evaluate the health status of renewable energy power generation equipment to obtain a forecast of available power generation capacity, and use weather forecast data to forecast the power generation of renewable energy to obtain a forecast of renewable energy supply;

[0091] (2) Analyze electricity demand based on historical electricity consumption data and load forecast results to obtain electricity demand forecasts, and analyze the matching degree between the electricity demand forecasts and renewable energy supply forecasts to obtain a preliminary supply and demand balance plan;

[0092] (3) Perform multi-objective optimization calculations on the preliminary supply and demand balance plan to obtain a set of candidate scheduling plans, and perform simulation verification on the candidate scheduling plans to obtain a feasibility assessment of the plan;

[0093] (4) Select the optimal scheduling scheme based on the feasibility evaluation of the scheme and obtain the intelligent energy scheduling scheme.

[0094] Specifically, the health status of renewable energy power generation equipment is assessed, which includes a comprehensive analysis of the efficiency, wear and tear, and failure of equipment such as photovoltaic panels and wind turbines. By comparing real-time monitoring data with historical performance curves, the current operating efficiency of the equipment is obtained. Combining this information, the available power generation capacity is predicted using the following formula:

[0095]

[0096] Among them, C a is the total available generating capacity, C i is the rated capacity of the ith device, E i is its current operating efficiency, n is the total number of devices, and the power generation of renewable energy is predicted by combining weather forecast data. For photovoltaic power generation, the following model is used:

[0097] P pv =η×A×G×(1-0.005(T-25))

[0098] Here, P pv is the photovoltaic power generation, η is the photoelectric conversion efficiency, A is the photovoltaic panel area, G is the solar radiation intensity, and T is the ambient temperature. For wind power generation, a similar model is used to consider factors such as wind speed and air density. Power demand analysis is performed based on the user's historical power consumption data and load forecast results. Time series analysis methods, such as ARIMA models or long short-term memory networks (LSTM), are used to predict power demand for the next 24 hours. The forecasting model considers variables such as time factors (such as hours, weekdays / weekends), temperature, and special events. The power demand forecast is matched with the renewable energy supply forecast, and the supply and demand difference in each time period is calculated to obtain a preliminary supply and demand balance plan.

[0099] The preliminary supply and demand balance scheme is subjected to multi-objective optimization calculation, taking into account objectives such as economy, reliability and environmental protection. The optimization problem can be expressed as:

[0100] minf(x)=[f1(x), f2(x), f3(x)]

[0101] subjectto: g i (x)≤0, i=1,…,m; hj(x)=0, j=1,…,n

[0102] Among them, f1(x) represents the operating cost, f2(x) represents the power supply reliability index, and f3(x) represents the carbon emissions. i (x) and h j (x) are inequality and equality constraints, respectively, representing various constraints on system operation. Use a multi-objective genetic algorithm or a particle swarm optimization algorithm to solve this problem, obtain a series of non-dominated solutions, and form a set of candidate scheduling solutions. The candidate scheduling solution set is simulated and verified, and the power system simulation software is used to simulate the system operation status under different solutions to evaluate the feasibility and stability of the solution. The simulation process takes into account abnormal conditions such as load fluctuations and equipment failures to comprehensively evaluate the robustness of the solution. According to the simulation results, the comprehensive score of each solution is calculated, and the solution with the highest score is selected as the final intelligent energy scheduling solution.

[0103] For example, a smart microgrid system includes 10MW photovoltaic power generation, 5MW wind power generation and 20MWh energy storage system. The health status assessment shows that the efficiency of the photovoltaic system is 92% and the efficiency of the wind system is 95%. The weather forecast shows that there will be plenty of sunshine the next day, and the photovoltaic power generation is expected to be 80MWh and the wind power generation is expected to be 40MWh. The power demand forecast results show that the demand during the peak power consumption period (10:00-14:00) is 15MW, and the demand during the off-peak period (02:00-06:00) is 5MW. The preliminary supply and demand balance plan recommends storing excess electricity in the energy storage system during the off-peak period and releasing the energy storage to supplement the power supply during the peak period. The multi-objective optimization calculation takes into account the three objectives of power cost, reliability and carbon emissions, and generates 10 candidate solutions. After simulation verification, the optimal solution releases 10MWh of electricity from the energy storage system during the peak period, while maintaining 2MW of spare capacity to cope with sudden demand. It is expected to reduce operating costs by 15%, improve power supply reliability by 8%, and reduce carbon emissions by 20 tons. This example demonstrates how AI-driven smart distribution systems can optimize the use of renewable energy, balance supply and demand, and achieve multiple goals of economy, reliability, and environmental protection.

[0104] The above describes the power distribution method based on AI artificial intelligence in the embodiment of the present application. The following describes the intelligent power distribution cabinet in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of the intelligent power distribution cabinet includes:

[0105] The acquisition module 201 is used to collect real-time data from key nodes of the power distribution system through an intelligent sensor network to obtain original power distribution system operation data;

[0106] A processing module 202 is used to clean and standardize the original power distribution system operation data through a data preprocessing algorithm to obtain a standardized data set;

[0107] A modeling module 203 is used to analyze and model the standardized data set through multi-level hybrid model training to obtain an AI analysis model of the power distribution system;

[0108] The analysis module 204 is used to process and analyze the power distribution data collected in real time through the AI ​​analysis model to obtain load prediction results and dynamic balancing strategies;

[0109] Identification module 205, used to perform abnormal pattern recognition and multi-path reasoning on system operation parameters based on the load prediction results and dynamic balancing strategy, and obtain equipment health status assessment and maintenance suggestions;

[0110] The calculation module 206 is used to perform multi-objective optimization calculations on renewable energy supply and electricity demand based on the equipment health status assessment and maintenance recommendations to obtain an intelligent energy scheduling solution.

[0111] Through the collaboration of the above components, the real-time data collection of key nodes of the distribution system can be carried out through the intelligent sensor network, so as to obtain the system operation status in a comprehensive and timely manner, provide high-quality and timely raw data foundation for subsequent analysis, and significantly improve the comprehensiveness and real-time nature of data collection. The data preprocessing algorithm is used to clean and standardize the original distribution system operation data, which not only effectively removes the noise and outliers in the data, but also realizes the unified representation of the data, laying a solid data foundation for subsequent model training and analysis, and greatly improving the data quality and availability. The introduction of multi-level hybrid model training enables the system to capture the time series characteristics, spatial characteristics and decision characteristics of distribution data at the same time, significantly enhances the expression ability and generalization performance of the model, and can more accurately characterize the dynamic characteristics of complex distribution systems. Through the AI ​​analysis model to process and analyze the real-time collected distribution data, the system can quickly and accurately generate load prediction results and dynamic balance strategies, greatly improving the prediction accuracy and response speed of the distribution system, and providing reliable support for real-time scheduling and decision-making. Based on the load forecast results and dynamic balancing strategies, the abnormal pattern recognition and multi-path reasoning of the system operating parameters can not only timely detect potential equipment failures and system anomalies, but also provide accurate equipment health status assessment and maintenance suggestions, significantly improving the reliability and maintenance efficiency of the system. Finally, through multi-objective optimization calculation of renewable energy supply and electricity demand, the intelligent energy scheduling scheme can maximize the utilization of renewable energy while ensuring power supply reliability, and achieve a balance between economy and environmental protection, greatly improving the overall operating efficiency and sustainability of the distribution system.

[0112] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A power distribution method based on AI artificial intelligence, characterized in that: The power distribution method based on AI artificial intelligence includes: Real-time data collection of key nodes of the power distribution system is carried out through the intelligent sensor network to obtain the original power distribution system operation data; Cleaning and standardizing the original power distribution system operation data by using a data preprocessing algorithm to obtain a standardized data set; The standardized data set is analyzed and modeled through multi-level hybrid model training to obtain an AI analysis model of the power distribution system; The AI ​​analysis model is used to process and analyze the power distribution data collected in real time to obtain load prediction results and dynamic balancing strategies; Based on the load prediction results and dynamic balancing strategy, abnormal pattern recognition and multi-path reasoning are performed on system operating parameters to obtain equipment health status assessment and maintenance recommendations; Based on the equipment health status assessment and maintenance recommendations, multi-objective optimization calculations are performed on renewable energy supply and electricity demand to obtain an intelligent energy scheduling solution.

2. The power distribution method based on AI artificial intelligence according to claim 1 is characterized in that: The real-time data collection of the key nodes of the power distribution system through the intelligent sensor network is used to obtain the original power distribution system operation data, including: Arrange intelligent sensors on the substation, distribution cabinet and user end of the power distribution system to obtain a multi-level data acquisition network, and configure the communication protocol of the multi-level data acquisition network to obtain a high-speed real-time data transmission channel; The voltage, current, power factor and harmonic content are continuously monitored through the high-speed real-time data transmission channel to obtain multi-dimensional power parameter data, and the multi-dimensional power parameter data is time-stamped to obtain a time-series associated original data stream; Performing a data integrity check on the original data stream associated with the time series to obtain a valid data set, and performing an outlier detection on the valid data set to obtain a preliminarily cleaned data set; Smoothing the preliminarily cleaned data set by a sliding window algorithm to obtain a denoised data sequence, and interpolating missing values ​​of the denoised data sequence to obtain a complete data sequence; Standardizing and normalizing the complete data sequence to obtain a standard data set with unified dimensions, and performing feature extraction on the standard data set with unified dimensions to obtain a feature vector after dimensionality reduction; Performing key feature identification on the feature vector after dimension reduction by using a principal component analysis algorithm to obtain a core feature set, and performing correlation analysis on the core feature set to obtain a feature correlation matrix; A multidimensional data cube is constructed based on the feature association matrix to obtain multi-angle data analysis views, and data clustering is performed on the multi-angle data analysis views to obtain a data distribution pattern.

3. The power distribution method based on AI artificial intelligence according to claim 1 is characterized in that: The raw power distribution system operation data is cleaned and standardized by a data preprocessing algorithm to obtain a standardized data set, including: Performing time series alignment on the original power distribution system operation data to obtain a synchronized data stream, and performing data integrity check on the synchronized data stream to obtain a valid data set; Performing outlier detection on the valid data set to obtain a preliminary cleaned data set, and removing noise from the preliminary cleaned data set using a median filtering algorithm to obtain a smoothed data sequence; Interpolating missing values ​​on the smoothed data sequence to obtain a complete data sequence, and removing duplicate values ​​from the complete data sequence to obtain a streamlined data set; Performing data type conversion on the streamlined data set to obtain a data set in a unified format, and performing dimension normalization processing on the data set in the unified format to obtain a standardized data set; Performing feature extraction on the standardized data set to obtain a feature vector set, and performing dimensionality reduction processing on the feature vector set through principal component analysis to obtain a key feature set; A correlation analysis is performed on the key feature set to obtain a feature association matrix, and data clustering is performed based on the feature association matrix to obtain a data distribution pattern.

4. The power distribution method based on AI artificial intelligence according to claim 1 is characterized in that: The standardized data set is analyzed and modeled by multi-level hybrid model training to obtain an AI analysis model of the power distribution system, including: Performing data segmentation on the standardized data set to obtain a training set and a validation set, and performing feature engineering processing on the training set to obtain an enhanced feature set; Constructing a deep neural network including 5 fully connected layers for the enhanced feature set to obtain a basic network structure, and adding a batch normalization layer and a rectified linear unit activation function between every two layers in the basic network structure to obtain a multi-layer neural network architecture; Add three bidirectional long short-term memory layers after the input layer of the multi-layer neural network architecture, each layer including 128 neurons, to obtain a time series analysis network, and apply a multi-head self-attention mechanism on the output of the long short-term memory layer of the time series analysis network, with the number of heads set to 8, to obtain the weight distribution of key features; Input the key feature weight distribution into three layers of two-dimensional convolutional layers, using a 3x3 convolution kernel and 32 filters in each layer to obtain a spatial feature analysis network, and add a skip connection after each convolution layer in the spatial feature analysis network to obtain a deep feature representation; Inputting the deep feature representation into a dual-depth Q network structure, including two identical Q networks, each Q network including three fully connected layers with 256, 128, and 64 neurons, respectively, to obtain a decision optimization network, and connecting the pre-trained knowledge graph embedding vector with the second layer output of the decision optimization network to obtain a knowledge-enhanced decision representation; A transfer learning structure including two fully connected layers is added after the knowledge-enhanced decision representation, each layer including 64 neurons, to obtain an adaptive optimization network. The learning rate, batch size and random inactivation rate of the adaptive optimization network are tuned through 5-fold cross validation to obtain an AI analysis model of the distribution system.

5. The power distribution method based on AI artificial intelligence according to claim 1 is characterized in that: The AI ​​analysis model is used to process and analyze the real-time collected power distribution data to obtain load prediction results and dynamic balancing strategies, including: Extracting features from the real-time collected power distribution data to obtain a real-time feature set, and inputting the real-time feature set into the AI ​​analysis model to obtain a preliminary prediction result; Performing short-term load forecasting on the preliminary forecast results to obtain load forecast data for 15 minutes to 1 hour, and performing medium-term load forecasting on the load forecast data to obtain load forecast trends for 1 day to 1 week; Based on the load forecast trend, long-term load forecasting is performed to obtain a load forecast plan for 1 month to 1 year, and error analysis is performed on the load forecast plan to obtain a forecast credibility assessment; Performing load distribution analysis on the power distribution network according to the prediction credibility assessment to obtain a load prediction result, and generating a load balancing solution based on the load prediction result to obtain a preliminary dynamic balancing strategy; The preliminary dynamic balance strategy is simulated in multiple scenarios to obtain a strategy effect evaluation, and parameters are optimized based on the strategy effect evaluation to obtain a dynamic balance strategy.

6. The power distribution method based on AI artificial intelligence according to claim 1, characterized in that: Based on the load prediction results and dynamic balancing strategy, abnormal pattern recognition and multi-path reasoning are performed on system operation parameters to obtain equipment health status assessment and maintenance suggestions, including: Monitor the system operating parameters in real time to obtain a parameter change sequence, and compare the parameter change sequence with historical normal operating data to obtain a deviation value set; Performing cluster analysis on the deviation value set to obtain an abnormal pattern candidate set, and performing feature extraction on the abnormal pattern candidate set to obtain an abnormal feature vector; Inputting the abnormal feature vector into a pre-trained abnormality detection model to obtain an abnormality probability distribution, and determining an abnormality level according to the abnormality probability distribution to obtain an abnormal state assessment of the device; Based on the abnormal state evaluation of the equipment, multi-path reasoning is performed in the knowledge graph to obtain a set of fault causes, and the set of fault causes is prioritized to obtain a fault diagnosis result; An equipment health status report is generated according to the fault diagnosis result to obtain an equipment health status assessment, and a preventive maintenance plan is formulated based on the equipment health status assessment to obtain maintenance recommendations.

7. The power distribution method based on AI artificial intelligence according to claim 1, characterized in that: Based on the equipment health status assessment and maintenance recommendations, a multi-objective optimization calculation is performed on the renewable energy supply and electricity demand to obtain an intelligent energy scheduling solution, including: Evaluate the health status of renewable energy power generation equipment to obtain available power generation capacity forecasts, and use weather forecast data to predict renewable energy power generation to obtain renewable energy supply forecasts; Performing power demand analysis based on historical power consumption data of users and load forecast results to obtain a power demand forecast, and performing matching analysis between the power demand forecast and the renewable energy supply forecast to obtain a preliminary supply and demand balance plan; Performing multi-objective optimization calculation on the preliminary supply-demand balance plan to obtain a set of candidate scheduling plans, and performing simulation verification on the set of candidate scheduling plans to obtain a feasibility assessment of the plans; The optimal scheduling scheme is selected according to the feasibility evaluation of the scheme to obtain the intelligent energy scheduling scheme.

8. An intelligent power distribution cabinet, used to implement the power distribution method based on AI artificial intelligence as described in any one of claims 1 to 7, characterized in that: The intelligent power distribution cabinet comprises: The acquisition module is used to collect real-time data from key nodes of the power distribution system through an intelligent sensor network to obtain the original power distribution system operation data; A processing module, used for cleaning and standardizing the original power distribution system operation data through a data preprocessing algorithm to obtain a standardized data set; A modeling module, used to analyze and model the standardized data set through multi-level hybrid model training to obtain an AI analysis model of the power distribution system; An analysis module, used to process and analyze the real-time collected power distribution data through the AI ​​analysis model to obtain load prediction results and dynamic balancing strategies; An identification module is used to perform abnormal pattern recognition and multi-path reasoning on system operation parameters based on the load prediction results and dynamic balancing strategy to obtain equipment health status assessment and maintenance suggestions; The calculation module is used to perform multi-objective optimization calculations on renewable energy supply and electricity demand based on the equipment health status assessment and maintenance recommendations to obtain an intelligent energy scheduling solution.

Images