Real-time monitoring system for line loss of photovoltaic grid-connected transformer area based on data driving

Abstract

The invention discloses a photovoltaic grid-connected zone area line loss real-time monitoring system based on data driving, which relates to the field of power systems, and comprises a data collection and preprocessing module, a line loss anomaly detection module, a line loss reason identification module, a line loss prediction module and an optimization control module. A cloud big data platform is adopted to construct a comprehensive power grid model and a photovoltaic power generation model, a time sequence analysis method is used to mine the line loss trend of a distributed photovoltaic grid-connected zone area, line loss prediction and early warning are realized, and an optimization control strategy is made according to the prediction trend, the current power grid operation state and historical optimization records. Therefore, the stability and reliability of power grid operation are improved; the method solves the problems that the line loss of the photovoltaic grid-connected transformer area is difficult to identify and position quickly and accurately and a corresponding line loss optimization strategy is lacked, and can realize real-time monitoring of the line loss of the distributed photovoltaic grid-connected transformer area, effectively reduce the line loss rate and guarantee stable operation of a power grid.

Description

Technical Field

[0001] The present invention relates to the field of power systems, and in particular to a data-driven real-time monitoring system for line losses in photovoltaic grid-connected areas. Background Art

[0002] With the widespread application of photovoltaic power stations and the large-scale development of photovoltaic power, distributed photovoltaic power generation has been widely used as a clean and renewable energy technology around the world. However, in the process of rapid development of distributed photovoltaic power generation, the operation of the power grid also faces many challenges, especially in the management of line losses in the substation area. Substation line loss refers to the power loss generated by the power grid in the process of transmitting and distributing electric energy, which is an important indicator for measuring the efficiency of power grid operation. Traditional management methods can no longer meet the new demands brought about by the increasingly complex grid structure and photovoltaic power generation access. Therefore, the development of a system that can monitor the line loss of photovoltaic grid-connected substations in real time has become a core requirement for improving photovoltaic power generation efficiency, ensuring power quality, and reducing operating costs.

[0003] The line loss problem in the distributed photovoltaic grid-connected area is affected by many factors, including abnormal cable resistance, electromagnetic field vortex effect, alternating magnetic field effect, grid management loopholes, and metering device failures. These factors are intertwined, making the line loss problem complex and changeable. Traditional methods often rely on manual inspections and empirical judgments, which are difficult to quickly and accurately identify and locate line loss problems, resulting in problems not being solved in a timely manner; and because the voltage and frequency of the distributed photovoltaic grid-connected local power grid fluctuate, there is a lack of corresponding line loss optimization strategies in the current distribution network management, which affects the safe and stable operation of the power grid.

[0004] Therefore, a data-driven real-time monitoring system for line losses in photovoltaic grid-connected areas is needed to solve the above problems. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention discloses a data-driven real-time monitoring system for line losses in photovoltaic grid-connected stations. Through the organic combination of functions such as real-time data collection, precise anomaly detection, cause identification, trend prediction and optimization control, it realizes all-round and intelligent monitoring and management of line losses in distributed photovoltaic grid-connected stations, with a high degree of automation and intelligence.

[0006] The present invention adopts the following technical solutions:

[0007] A data-driven real-time monitoring system for line loss in a photovoltaic grid-connected area, the system comprising:

[0008] A data collection and preprocessing module, which uses an edge sensing monitoring interconnection device to collect the operating status data of the distributed photovoltaic grid-connected area in real time, and transmits the real-time collected data to the cloud big data platform through a wireless communication network for data preprocessing, storage and analysis operations. The operating status data at least includes photovoltaic power generation, voltage, current and load characteristics;

[0009] A line loss anomaly detection module, which uses a K-Medoids clustering algorithm to perform cluster analysis on the historically detected area line loss data to identify the cluster centers of the normal working mode and abnormal line loss mode of the distributed photovoltaic grid-connected area, and calculates the Euclidean distance between the real-time collected data and the cluster center of the abnormal line loss mode. When the Euclidean distance between the real-time collected data and the cluster center of the abnormal line loss mode exceeds a preset low threshold, the real-time collected data is marked as abnormal line loss data;

[0010] A line loss cause identification module, wherein the line loss cause identification module constructs a comprehensive power grid model and a photovoltaic power generation model of a distributed photovoltaic grid-connected station area through the cloud big data platform, and the cloud big data platform compares the simulated physical state data of the comprehensive power grid model and the photovoltaic power generation model with the abnormal line loss data to calculate the time-sharing line loss, and identifies the cause of the abnormal line loss based on the calculation result. The comprehensive power grid model is used to simulate the operating characteristics of the power grid in the distributed photovoltaic grid-connected station area, and the photovoltaic power generation model is used to simulate the output characteristics of the distributed photovoltaic grid-connected station area under different weather conditions;

[0011] A line loss prediction module, which uses a time series analysis method to mine the time series characteristics of photovoltaic power generation, voltage changes and load fluctuations to predict the line loss trend of distributed photovoltaic grid-connected stations;

[0012] The optimization control module formulates an optimization control strategy according to the line loss prediction trend, the current grid operation status and the historical optimization records.

[0013] Furthermore, the working method of the K-Medoids clustering algorithm comprises the following steps:

[0014] S1. Select K data points from the historically detected area line loss data set by random selection as the initial abnormal line loss pattern clustering center, set the maximum number of iterations and the threshold of the cluster center change, and then record the initial cluster center index or actual data point value;

[0015] S2, calculating the Euclidean distance between the real-time collected data points and the cluster center of the abnormal line loss pattern, and assigning each data point to the closest cluster center;

[0016] S3, recalculate the sum of the distances from all points in each cluster to the cluster center, select the cluster center with the smallest sum of distances as the new abnormal line loss pattern cluster center, and then update the index or actual data point value of the cluster center;

[0017] S4, repeating the S2 operation, and using the new abnormal line loss pattern cluster center to redistribute the data points until the change in the abnormal line loss pattern cluster center is less than the threshold or the maximum number of iterations is reached;

[0018] S5. Use the silhouette coefficient and CH index clustering quality evaluation indicators to evaluate whether the clustering results are consistent with the expected abnormal line loss pattern. If not, adjust the clustering parameters and repeat S1.

[0019] Furthermore, in S2, firstly, the time series features and space series features of the preprocessed real-time collected data points are extracted, and the extracted features are regularized and represented by feature vectors. The extracted time series features and space series feature vectors are represented as follows:

[0020]

[0021] In formula (1), X represents the feature vector set of data points collected in real time. is the time series feature vector of the i-th real-time data point, 1≤i≤n, i is the ordinal number of the real-time data point, n is the number of time series feature vectors, t represents the time series feature vector, is the spatial sequence feature vector of the ith real-time data point, y represents the spatial sequence feature vector, and the cluster center set of abnormal line loss pattern is V = {v1,v2,...,v g ,...,v k},v g represents the g-th abnormal line loss pattern cluster center, k represents the number of abnormal line loss pattern cluster centers, and the Euclidean distance measurement method is used to calculate the distance between the time series characteristics and the spatial series characteristics and the abnormal line loss pattern cluster center. The output function formula is:

[0022]

[0023] In formula (2), is the distance from the time series feature vector of the i-th real-time data point to the cluster center of the abnormal line loss pattern, ξ is the clustering hyperparameter of the time series feature vector, 0<ξ≤1, is the distance from the spatial sequence feature vector of the ith real-time data point to the cluster center of the abnormal line loss pattern, δ is the clustering hyperparameter of the spatial sequence feature vector, 0<δ≤1, and the Euclidean distance calculation formula between the ith real-time data point and the cluster center of the abnormal line loss pattern is:

[0024]

[0025] In formula (3), D i It represents the Euclidean distance between the i-th real-time collected data point and the cluster center of the abnormal line loss pattern.

[0026] Furthermore, the comprehensive power grid model and photovoltaic power generation model use a physical engine to simulate the operating status of distributed photovoltaic grid-connected stations in different time periods and under different operating conditions, and perform real-time calculation of time-sharing line losses based on the simulation results. When the time-sharing line loss calculation result exceeds the set time-sharing line loss threshold, an alarm is automatically triggered, and then the real-time monitored physical status data and the time-sharing line loss data are correlated and analyzed to locate the cause of the abnormal line loss.

[0027] Furthermore, the time series analysis method uses a long short-term memory network (LSTM) to construct a power prediction model, and trains the model by minimizing the error between the predicted value and the actual value, and then uses the trained model to predict short-term power. The long short-term memory network (LSTM) uses mean square error as a loss function, and uses root mean square error and mean absolute error indicators to evaluate the prediction performance of the training model.

[0028] Furthermore, the optimization control module adopts a game theory model to analyze the user's electricity consumption behavior to identify abnormal electricity consumption patterns, and formulates an optimization control strategy based on the game theory analysis results. The game theory model quantifies the benefits or costs of different electricity consumption strategies for power grid operators and users by establishing a utility function, and identifies the optimal strategy combination for users and operators by calculating Nash equilibrium.

[0029] Furthermore, the optimization control strategy includes the following aspects:

[0030] (1) To address the voltage fluctuation problem, intelligent substation devices, coordinated voltage control and reactive power compensation methods are used to control and regulate voltage fluctuations;

[0031] (2) To address the problem of unbalanced power load, load balancing is achieved by adjusting the distribution network structure and power load;

[0032] (3) In view of the problem of distributed energy access and distribution, the photovoltaic access location and access method are adjusted, and inverters and monitoring devices are used to monitor photovoltaic power generation and grid load;

[0033] (4) To solve the problem of power equipment failure, the power equipment failure can be eliminated through online fault monitoring, prediction and fault diagnosis.

[0034] Furthermore, the cloud big data platform adopts distributed storage technology to store massive line loss data, and adopts a real-time computing framework to stream process the real-time line loss data, thereby realizing real-time monitoring and analysis of line loss.

[0035] The beneficial effects of the present invention are:

[0036] 1. The present invention collects the operating status data of distributed photovoltaic grid-connected areas in real time through edge sensing monitoring interconnection devices to ensure the timeliness and accuracy of the data, and adopts the K-Medoids clustering algorithm to perform cluster analysis on line loss data, which can quickly identify abnormal line loss patterns, and realize rapid marking of abnormal data by calculating the Euclidean distance between real-time data and the cluster center of the abnormal pattern, thereby improving the accuracy and timeliness of abnormal detection.

[0037] 2. The present invention constructs a comprehensive power grid model and photovoltaic power generation model to simulate the power grid operation characteristics and photovoltaic power generation output characteristics, providing a scientific basis for the accurate identification of the causes of abnormal line losses. By comparing the simulated physical state data with the abnormal line loss data, the time-sharing line loss calculation is performed, which can more accurately identify the specific causes of abnormal line losses and provide strong support for subsequent optimization control.

[0038] 3. The present invention adopts the time series analysis method to deeply explore the time series characteristics of photovoltaic power generation, voltage changes and load fluctuations, improves the accuracy and reliability of line loss prediction, and the prediction model constructed based on time series characteristics can accurately predict the line loss trend of distributed photovoltaic grid-connected stations, providing an important reference for the planning and scheduling of power grids.

[0039] 4. The present invention formulates an optimization control strategy based on the line loss prediction trend, the current grid operation status and the historical optimization records, realizes the intelligent management of the grid operation, and through the implementation of the optimization control strategy, can effectively reduce the line loss rate of the distributed photovoltaic grid-connected area, improve energy utilization efficiency and reduce energy waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic diagram of the overall system architecture of the present invention;

[0041] Figure 2 Schematic diagram of the workflow of the K-Medoids clustering algorithm in the present invention;

[0042] Figure 3 Schematic diagram of the workflow of the time series analysis method of the present invention;

[0043] Figure 4 It is a schematic diagram of the workflow of the data-driven model predictive control MPC strategy in the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the attached embodiment of the present invention Figure 1 To Attachment Figure 4 , the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] The embodiment of the present invention discloses a data-driven real-time monitoring system for line loss in a photovoltaic grid-connected area, the system comprising:

[0046] Data collection and preprocessing module: This module uses existing edge sensing monitoring interconnection devices and installs sensor equipment such as power meters, voltmeters, ammeters, temperature sensors, etc. to collect photovoltaic power generation, voltage fluctuations, load characteristics and line loss data in real time, and performs data cleaning, missing value filling and outlier detection and other processing operations to ensure the accuracy and completeness of the data, and transmit these data to the cloud big data platform through the wireless communication network.

[0047] Line loss anomaly detection module: This module uses the K-Medoids clustering algorithm to perform cluster analysis on the historically detected substation line loss data to identify the cluster centers of the normal working mode and the abnormal line loss mode. Specifically, the algorithm performs multiple clustering operations on the historical data to obtain all cluster centers, and on this basis calculates the Euclidean distance between the real-time collected data within a specific time and the cluster center of the abnormal line loss mode. When it exceeds the preset low threshold, the data is marked as abnormal line loss data and needs to be processed.

[0048] Line loss cause identification module: This module first uses the cloud big data platform to build a comprehensive power grid model and photovoltaic power generation model for distributed photovoltaic grid-connected areas. Specifically, the comprehensive power grid model is used to simulate the operating characteristics of the distributed photovoltaic grid-connected area power grid, and the photovoltaic power generation model is used to simulate the output characteristics of the distributed photovoltaic grid-connected area under different weather conditions. Secondly, the simulated physical state data is compared with the abnormal line loss data to calculate the time-sharing line loss. Finally, the cause of the abnormal line loss is identified based on the calculation results.

[0049] Line loss prediction module: In time series analysis, this module uses time series analysis methods to reveal the laws of photovoltaic power generation, voltage changes and load fluctuations, mines the time series characteristics of photovoltaic power generation, voltage changes and load fluctuations, and predicts the line loss trend of distributed photovoltaic grid-connected areas based on these characteristics.

[0050] Optimization control module: This module formulates optimization control strategies based on line loss prediction trends, current grid operation status, and historical optimization records. Specifically, based on the line loss prediction trend, the system predicts the line loss that may occur in the future and formulates corresponding optimization control plans; at the same time, based on the current grid operation status, analyzes the changes in parameters such as grid load in order to perform targeted optimization control; finally, using historical optimization records and model prediction results, continuous optimization and updating are performed to ensure the effectiveness and practicability of the optimization control strategy.

[0051] In short, the system is based on data-driven real-time monitoring and analysis, and uses a variety of technical means to solve the line loss problem in distributed photovoltaic grid-connected areas. It has the advantages of high efficiency, accuracy, and reliability.

[0052] The output end of the data collection and preprocessing module is connected to the input end of the line loss anomaly detection module, the line loss cause identification module, and the line loss prediction module; the output end of the line loss anomaly detection module is connected to the input end of the line loss cause identification module; the output end of the line loss anomaly detection module, the line loss cause identification module, and the line loss prediction module is connected to the input end of the optimization control module.

[0053] The working method of K-Medoids clustering algorithm includes the following steps:

[0054] S1: Initialize cluster center;

[0055] Data preparation: Obtain data from the historically detected substation line loss dataset.

[0056] Random selection: K different data points are selected from this set of data by random selection as the initial abnormal line loss pattern clustering centers.

[0057] Set parameters: Set the maximum number of iterations and the threshold for the change in cluster center.

[0058] Record the initial cluster centers: Record the indexes or actual data point values ​​of these K initial cluster centers for subsequent comparison and update.

[0059] S2: data point allocation;

[0060] Distance calculation: For each data point in the data set collected in real time, calculate the Euclidean distance between it and all K cluster centers.

[0061] Assign data points: Assign each data point to the cluster center with the smallest Euclidean distance to it, forming K clusters.

[0062] S3: Update cluster centers;

[0063] Calculate the internal distance sum: For each cluster, calculate the sum of the distances from all internal data points to the current cluster center.

[0064] Select new cluster centers: In each cluster, select the data point that minimizes the sum of internal distances as the new cluster center. This is usually done by traversing all data points in the cluster, calculating the sum of internal distances after replacing each point as the cluster center, and then selecting the point that minimizes the sum.

[0065] Update Cluster Centers: Update the cluster center indices or actual data point values ​​to the new selection.

[0066] S4: Iterative optimization;

[0067] Repeat assignment: Go back to S2, use the new cluster center to recalculate the distance from each data point to the cluster center, and redistribute the data points to the nearest cluster center.

[0068] Termination condition: Check whether the change in the cluster center is less than the preset threshold, or whether the maximum number of iterations has been reached. If any of the conditions are met, the iteration is stopped; otherwise, it returns to S3 for the next iteration.

[0069] S5: Clustering quality assessment;

[0070] Evaluation index calculation: Clustering quality evaluation indicators such as silhouette coefficient and CH index (Calinski-Harabasz Index) are used to evaluate the quality of clustering results.

[0071] The silhouette coefficient measures the compactness of data points within a cluster and the separation of data points between clusters. Its value range is between -1 and 1. The larger the value, the better the clustering effect.

[0072] The CH index evaluates the clustering effect by calculating the ratio of the trace of the intra-class dispersion matrix to the trace of the inter-class dispersion matrix. The larger the value, the better the clustering effect.

[0073] Result judgment: If the clustering result does not conform to the expected abnormal line loss pattern (for example, verified by expert judgment or business rules), the clustering parameters (such as K value, number of iterations, threshold, etc.) are adjusted, and the operations from S1 to S4 are repeated until the clustering result is satisfactory.

[0074] In S2, firstly, the time series features and space series features of the preprocessed real-time collected data points are extracted, and the extracted features are regularized and represented by feature vectors. The extracted time series features and space series feature vectors are represented as:

[0075]

[0076] In formula (1), X represents the feature vector set of data points collected in real time. is the time series feature vector of the i-th real-time data point, 1≤i≤n, i is the ordinal number of the real-time data point, n is the number of time series feature vectors, t represents the time series feature vector, is the spatial sequence feature vector of the ith real-time data point, y represents the spatial sequence feature vector, and the cluster center set of abnormal line loss pattern is V = {v1,v2,...,v g ,...,v k},v g represents the g-th abnormal line loss pattern cluster center, k represents the number of abnormal line loss pattern cluster centers, and the Euclidean distance measurement method is used to calculate the distance between the time series characteristics and the spatial series characteristics and the abnormal line loss pattern cluster center. The output function formula is:

[0077]

[0078] In formula (2), is the distance from the time series feature vector of the i-th real-time data point to the cluster center of the abnormal line loss pattern, ξ is the clustering hyperparameter of the time series feature vector, 0<ξ≤1, is the distance from the spatial sequence feature vector of the ith real-time data point to the cluster center of the abnormal line loss pattern, δ is the clustering hyperparameter of the spatial sequence feature vector, 0<δ≤1, and the Euclidean distance calculation formula between the ith real-time data point and the cluster center of the abnormal line loss pattern is:

[0079]

[0080] In formula (3), D i It represents the Euclidean distance between the i-th real-time collected data point and the cluster center of the abnormal line loss pattern.

[0081] The hardware working environment of the K-Medoids clustering algorithm includes:

[0082] Processor (CPU): Strong processor performance can speed up data processing and calculation, especially for large-scale data sets, high-performance CPU is necessary. Multi-core CPU can process multiple tasks in parallel and improve the execution efficiency of the algorithm.

[0083] Memory (RAM): Sufficient memory is essential for processing large data sets. The K-Medoids algorithm needs to frequently access and update the data set during the iteration process, so sufficient RAM can reduce disk I / O operations and increase the speed of the algorithm.

[0084] Storage devices: Although the K-Medoids algorithm relies heavily on in-memory data processing, larger data sets may not be fully loaded into memory. In this case, fast storage devices (such as SSDs) can reduce data loading time and improve overall performance.

[0085] Graphics Processing Unit (GPU): For some K-Medoids algorithm implementations that support GPU acceleration, a powerful GPU can significantly increase the execution speed of the algorithm. GPUs are good at handling large-scale parallel computations and are particularly suitable for accelerating operations such as distance calculations and cluster updates.

[0086] Network conditions (for distributed or cloud environments): If the K-Medoids algorithm is run in a distributed system or cloud environment, network bandwidth and latency will affect the efficiency of data transmission and communication between nodes. Therefore, good network conditions are crucial to ensure the stable operation of the algorithm.

[0087] According to the annual line loss work requirements of the State Grid Sanmenxia Power Supply Company, a series of detailed on-site surveys were conducted at the Shanzhou Power Supply Company to understand the actual operation of the distributed photovoltaic grid-connected areas. The survey content included photovoltaic power generation, voltage fluctuations, load characteristics and line loss data, and these data were used to preliminarily identify the main sources and influencing factors of the line loss in the area.

[0088] A Core i9 64+1T memory computer was used for data analysis. The on-site experimental environment was set up, the simulation data accuracy was 95%, and the algorithm operation error did not exceed 2.5%. The K-Medoids algorithm (Group A) and the linear regression method (Group B) were used for comparative experiments. Four groups of historical line loss and wireless line loss mixed data were manually input, and the data capacity of each group was 500KB. Simulation work was performed in the experimental environment, and the detection accuracy and time consumption were recorded. The experiment was repeated 5 times to calculate the average value, and the records are shown in Table 1.

[0089] Table 1 Results Statistics

[0090]

[0091] The results show that on the test set, the K-Medoids algorithm has higher detection accuracy and lower error rate than the linear regression method. At the same time, the running time of the K-Medoids algorithm is also lower than that of the linear regression method, and it has higher computational efficiency. Therefore, under this specific experimental condition, the performance of the K-Medoids algorithm is better than that of the linear regression method.

[0092] The integrated power grid model and photovoltaic power generation model simulate the operating status of the distributed photovoltaic grid-connected area based on the physical engine, including parameters such as voltage, current, power, and electricity, and adjust the photovoltaic power generation model under different weather conditions to simulate the changes in photovoltaic power generation. During the simulation process, the integrated power grid model and photovoltaic power generation model collect and store simulation result data in real time through the cloud big data platform, including time-sharing electricity, time-sharing power, time-sharing voltage, time-sharing current and other parameters. The simulation result data is correlated with the real-time monitored physical state data in real time to locate the cause of abnormal line loss. If the time-sharing line loss calculation result exceeds the set time-sharing line loss threshold, an alarm is triggered. Based on the analysis results, the optimization control module will formulate an optimization control strategy, such as adjusting the inverter operation mode, adjusting the load, etc., to reduce the time-sharing line loss. At the same time, the line loss prediction module will mine the time series characteristics of photovoltaic power generation, voltage change and load fluctuation to predict the line loss trend of the distributed photovoltaic grid-connected area. The optimization control module will combine the line loss prediction trend and the current grid operation status to formulate a more targeted optimization control strategy.

[0093] The time series analysis method uses the long short-term memory network (LSTM) to build a power prediction model, and trains the model by minimizing the error between the predicted value and the actual value. The trained model is then used to predict short-term power. The long short-term memory network (LSTM) uses mean square error as the loss function, and the root mean square error and mean absolute error indicators are used to evaluate the prediction performance of the training model.

[0094] The time series analysis method is a short-term power prediction algorithm based on recurrent neural networks, which can capture the nonlinear and dynamic characteristics of photovoltaic power generation. Figure 3 As shown, the implementation details of the time series analysis method are as follows:

[0095] (S1). Data preprocessing: including data cleaning, standardization and feature engineering to improve the generalization ability of the model.

[0096] (S2). Model construction: Long short-term memory network (LSTM) is used as a variant of RNN to build a power prediction model. LSTM can solve the gradient vanishing problem in traditional RNN and improve learning efficiency.

[0097] (S3). Training process: The model is trained by minimizing the error between the predicted value and the actual value, usually using the mean squared error (MSE) as the loss function.

[0098] (S4). Prediction and evaluation: The short-term power is predicted using the trained model, and the prediction performance is evaluated using metrics such as root mean square error (RMSE) and mean absolute error (MAE).

[0099] The optimization control module uses game theory models to analyze users' electricity consumption behaviors, identify abnormal electricity consumption patterns, and optimize electricity consumption strategies. Based on the results of game theory analysis, optimization strategies are formulated to reduce line losses and improve grid operation efficiency. Game theory models quantify the benefits or costs of different electricity consumption strategies for grid operators and users by establishing utility functions, and identify the optimal strategy combination for users and operators by calculating Nash equilibrium.

[0100] In addition, data-driven methods such as reinforcement learning and policy optimization play an important role in optimizing control. The data-driven model predictive control (MPC) strategy can use historical data to eliminate the impact of parameter mismatch, thereby improving control performance. Figure 4 As shown, the implementation details of the MPC strategy include:

[0101] Step 1. Establish a photovoltaic system model: Establish an accurate system model based on the physical characteristics and operating parameters of the photovoltaic system.

[0102] Step 2. Define the objective function: This usually includes minimization of tracking error and smoothness of control input.

[0103] Step 3. Optimization process: Use the solver to perform online optimization to find the optimal control sequence that satisfies the system constraints.

[0104] Step 4. Real-time update: Continuously update the model and optimization process based on real-time data to adapt to changes in the status of the PV system.

[0105] The use of local energy storage technology can compensate for the difference between photovoltaic power and load through the existing power grid, better adapting to the flow of local grid-connected power. The control algorithm of the energy storage system can dynamically adjust the charging and discharging strategy of energy storage according to the needs of the power grid and the volatility of photovoltaic power generation to achieve optimal energy distribution and management.

[0106] The optimization control strategy includes the following aspects:

[0107] (1) To address the voltage fluctuation problem, intelligent substation devices, coordinated voltage control and reactive power compensation methods are used to control and regulate voltage fluctuations;

[0108] (2) To address the problem of unbalanced power load, load balancing is achieved by adjusting the distribution network structure and power load;

[0109] (3) In view of the problem of distributed energy access and distribution, the photovoltaic access location and access method are adjusted, and inverters and monitoring devices are used to monitor photovoltaic power generation and grid load;

[0110] (4) To solve the problem of power equipment failure, the power equipment failure can be eliminated through online fault monitoring, prediction and fault diagnosis.

[0111] In view of the voltage fluctuation problem, the automatic voltage control system was improved to reduce the voltage fluctuation. In view of the complexity of line loss, the line loss was reduced by optimizing the photovoltaic access location and access method. In order to solve the problem of massive IoT data, big data analysis technology was used to realize data sharing and comprehensive utilization among different systems and tap the potential value of data. In the specific implementation of the control strategy, a variety of technical means and algorithms were adopted. For example, the accurate calculation method of line loss in the grid-connected area of ​​distributed photovoltaic power generation with electric pile harmonics can accurately calculate the actual power loss value of the line by analyzing the influence of voltage, harmonics and islanding effects. The distributed photovoltaic line loss anomaly perception algorithm based on K-Medoids clustering combines the local anomaly factor (LOF) algorithm and granular calculation to optimize the clustering center, effectively avoiding the influence of isolated points on the anomaly perception effect, and accurately and effectively perceives the line loss anomaly in the distributed photovoltaic area. In addition, the distributed photovoltaic grid-connected capacity matching strategy based on network loss and voltage constraints was studied. From the perspective of safety and economy, the PV grid-connected capacity calculation formula suitable for actual engineering applications was obtained, and its correctness and effectiveness were verified through simulation.

[0112] The cloud big data platform uses distributed storage technologies such as Hadoop and HBase to store massive amounts of line loss data, ensuring high availability and reliability of the data. It also uses real-time computing frameworks such as Spark Streaming to stream real-time line loss data and achieve real-time monitoring and analysis of line losses.

[0113] The application effect of distributed photovoltaic grid-connected area line loss management technology in multiple scenarios is remarkable. In industry, commercial buildings, agricultural facilities and residential areas, distributed photovoltaic grid-connected systems can effectively reduce line loss rates and improve power supply efficiency and stability. For example, Liaoyang Power Supply Company significantly reduced line loss rates by adjusting the lines. In microgrid applications, distributed photovoltaic grid-connected systems are combined with energy storage, diesel generators, etc. to provide power, which improves the stability and reliability of the system.

[0114] Public facilities such as airports have achieved green and sustainable power supply by installing distributed photovoltaic grid-connected systems. In high-load locations, distributed photovoltaic grid-connected systems effectively absorb the impact of the volatility of renewable energy power generation and ensure the stable operation of the power grid. These application cases show that through reasonable planning and technological innovation, distributed photovoltaic grid-connected area line loss management technology can effectively improve the safety, stability and economy of power grid operation.

[0115] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A data-driven real-time monitoring system for line loss in photovoltaic grid-connected areas, characterized in that: The system comprises: A data collection and preprocessing module, which uses an edge sensing monitoring interconnection device to collect the operating status data of the distributed photovoltaic grid-connected area in real time, and transmits the real-time collected data to the cloud big data platform through a wireless communication network for data preprocessing, storage and analysis operations. The operating status data at least includes photovoltaic power generation, voltage, current and load characteristics; A line loss anomaly detection module, which uses a K-Medoids clustering algorithm to perform cluster analysis on the historically detected area line loss data to identify the cluster centers of the normal working mode and abnormal line loss mode of the distributed photovoltaic grid-connected area, and calculates the Euclidean distance between the real-time collected data and the cluster center of the abnormal line loss mode. When the Euclidean distance between the real-time collected data and the cluster center of the abnormal line loss mode exceeds a preset low threshold, the real-time collected data is marked as abnormal line loss data; A line loss cause identification module, wherein the line loss cause identification module constructs a comprehensive power grid model and a photovoltaic power generation model of a distributed photovoltaic grid-connected station area through the cloud big data platform, and the cloud big data platform compares the simulated physical state data of the comprehensive power grid model and the photovoltaic power generation model with the abnormal line loss data to calculate the time-sharing line loss, and identifies the cause of the abnormal line loss based on the calculation result. The comprehensive power grid model is used to simulate the operating characteristics of the power grid in the distributed photovoltaic grid-connected station area, and the photovoltaic power generation model is used to simulate the output characteristics of the distributed photovoltaic grid-connected station area under different weather conditions; A line loss prediction module, which uses a time series analysis method to mine the time series characteristics of photovoltaic power generation, voltage changes and load fluctuations to predict the line loss trend of distributed photovoltaic grid-connected stations; The optimization control module formulates an optimization control strategy according to the line loss prediction trend, the current grid operation status and the historical optimization records.

2. According to claim 1, a data-driven real-time monitoring system for line loss in photovoltaic grid-connected areas is characterized in that: The working method of the K-Medoids clustering algorithm includes the following steps: S1. Select K data points from the historically detected area line loss data set by random selection as the initial abnormal line loss pattern clustering center, set the maximum number of iterations and the threshold of the cluster center change, and then record the initial cluster center index or actual data point value; S2, calculating the Euclidean distance between the real-time collected data points and the cluster center of the abnormal line loss pattern, and assigning each data point to the closest cluster center; S3, recalculate the sum of the distances from all points in each cluster to the cluster center, select the cluster center with the smallest sum of distances as the new abnormal line loss pattern cluster center, and then update the index or actual data point value of the cluster center; S4, repeating the S2 operation, and using the new abnormal line loss pattern cluster center to redistribute the data points until the change in the abnormal line loss pattern cluster center is less than the threshold or the maximum number of iterations is reached; S5. Use the silhouette coefficient and CH index clustering quality evaluation indicators to evaluate whether the clustering results are consistent with the expected abnormal line loss pattern. If not, adjust the clustering parameters and repeat S1.

3. A data-driven real-time monitoring system for line loss in photovoltaic grid-connected areas according to claim 2, characterized in that: In S2, firstly, the time series features and space series features of the preprocessed real-time collected data points are extracted, and the extracted features are regularized and represented by feature vectors. The extracted time series features and space series feature vectors are represented as follows: In formula (1), X represents the feature vector set of data points collected in real time. is the time series feature vector of the i-th real-time data point, 1≤i≤n, i is the ordinal number of the real-time data point, n is the number of time series feature vectors, t represents the time series feature vector, is the spatial sequence feature vector of the ith real-time data point, y represents the spatial sequence feature vector, and the cluster center set of abnormal line loss pattern is V = {v1,v2,...,v g ,...,v k },v g represents the g-th abnormal line loss pattern cluster center, k represents the number of abnormal line loss pattern cluster centers, and the Euclidean distance measurement method is used to calculate the distance between the time series characteristics and the spatial series characteristics and the abnormal line loss pattern cluster center. The output function formula is: In formula (2), is the distance from the time series feature vector of the i-th real-time data point to the cluster center of the abnormal line loss pattern, ξ is the clustering hyperparameter of the time series feature vector, 0<ξ≤1, is the distance from the spatial sequence feature vector of the ith real-time data point to the cluster center of the abnormal line loss pattern, δ is the clustering hyperparameter of the spatial sequence feature vector, 0<δ≤1, and the Euclidean distance calculation formula between the ith real-time data point and the cluster center of the abnormal line loss pattern is: In formula (3), D i It represents the Euclidean distance between the i-th real-time collected data point and the cluster center of the abnormal line loss pattern.

4. According to claim 1, a data-driven real-time monitoring system for line loss in photovoltaic grid-connected areas is characterized in that: The comprehensive power grid model and photovoltaic power generation model use a physical engine to simulate the operating status of distributed photovoltaic grid-connected stations in different time periods and under different working conditions, and perform real-time calculation of time-sharing line losses based on the simulation results. When the time-sharing line loss calculation result exceeds the set time-sharing line loss threshold, an alarm is automatically triggered, and then the real-time monitored physical status data and the time-sharing line loss data are correlated and analyzed to locate the cause of the abnormal line loss.

5. According to the data-driven real-time monitoring system for line loss in photovoltaic grid-connected areas according to claim 1, it is characterized in that: The time series analysis method adopts a long short-term memory network (LSTM) to construct a power prediction model, and trains the model by minimizing the error between the predicted value and the actual value, and then uses the trained model to predict short-term power. The long short-term memory network (LSTM) adopts mean square error as the loss function, and uses root mean square error and mean absolute error indicators to evaluate the prediction performance of the training model.

6. A data-driven real-time monitoring system for line loss in photovoltaic grid-connected areas according to claim 1, characterized in that: The optimization control module uses a game theory model to analyze the user's electricity consumption behavior to identify abnormal electricity consumption patterns, and formulates an optimization control strategy based on the game theory analysis results. The game theory model quantifies the benefits or costs of different electricity consumption strategies for power grid operators and users by establishing a utility function, and identifies the optimal strategy combination for users and operators by calculating Nash equilibrium.

7. A data-driven real-time monitoring system for line loss in photovoltaic grid-connected areas according to claim 6, characterized in that: The optimization control strategy includes the following aspects: (1) To address the voltage fluctuation problem, intelligent substation devices, coordinated voltage control and reactive power compensation methods are used to control and regulate voltage fluctuations; (2) To address the problem of unbalanced power load, load balancing is achieved by adjusting the distribution network structure and power load; (3) In view of the problem of distributed energy access and distribution, the photovoltaic access location and access method are adjusted, and inverters and monitoring devices are used to monitor photovoltaic power generation and grid load; (4) To solve the problem of power equipment failure, the power equipment failure can be eliminated through online fault monitoring, prediction and fault diagnosis.

8. The data-driven real-time monitoring system for line loss in photovoltaic grid-connected areas according to claim 1 is characterized in that: The cloud-based big data platform uses distributed storage technology to store massive amounts of line loss data, and uses a real-time computing framework to stream process real-time line loss data, thereby achieving real-time monitoring and analysis of line losses.

Images