Multi-device cooperative line loss anomaly cooperative detection method and system

Abstract

The invention discloses a multi-device cooperative line loss anomaly cooperative detection method and system, and belongs to the technical field of electric power, and the method comprises the steps: collecting original time series data of multiple dimensions, carrying out the preprocessing, generating a dynamic space-time diagram based on the time series data to reflect the cooperative work characteristics of devices, extracting the cooperative feature vectors of the devices, and carrying out the fault detection. And meanwhile, self-optimization is carried out, a system strategy is adjusted, and a result is finally output to the equipment. According to the scheme, the problem that an existing method cannot adapt to dynamic changes of power equipment in real time is solved, and the adaptive capacity of the system to the complexity of a power network is enhanced through self-adaptive optimization of an intelligent algorithm.

Description

Technical Field

[0001] The present invention belongs to the field of electric power technology, and in particular to a method and system for collaborative detection of line loss anomalies by multiple devices. Background Art

[0002] In the power system, line loss (i.e., energy loss during transmission and distribution) has always been an important issue of concern to power companies. With the continuous expansion and intelligent development of power networks, the detection of line loss anomalies has become increasingly complex. Traditional line loss detection methods mostly rely on manual inspection, equipment status monitoring, and experience-based rule detection. These methods have many limitations, such as: they can only monitor the abnormal status of a single device and cannot obtain the collaborative working conditions between multiple devices in real time; they only rely on static rules or threshold judgments, resulting in slow response to abnormal signals in complex systems; and it is difficult to cope with the increasingly complex relationships and real-time status changes between devices in the power system.

[0003] At present, some existing anomaly detection technologies based on data analysis, such as machine learning models (such as support vector machines, decision trees, etc.), have improved this problem to a certain extent and can identify abnormal patterns through historical data. However, these methods are often unable to achieve accurate and effective detection when faced with the complexity and real-time requirements of the coordinated work of multiple devices in the power system. There are many types of equipment in the power system, which are widely distributed and have strong correlations between the equipment. Simply relying on traditional equipment monitoring methods and static machine learning models cannot fully capture the synergy between the equipment, resulting in low accuracy of anomaly detection. In addition, the time-varying nature of equipment status and complex environmental factors make it difficult for existing methods to adapt to the needs of dynamic changes in the power system. Therefore, the adaptability and robustness of existing technologies are greatly insufficient.

[0004] To this end, we propose a multi-device collaborative line loss anomaly collaborative detection method and system to solve the above problems. Summary of the invention

[0005] The purpose of the present invention is to solve the limitation problem in the prior art of line loss anomaly detection by multi-device collaboration, and to propose a method and system for collaborative line loss anomaly detection by multi-device collaboration.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: A multi-device collaborative line loss anomaly detection method, comprising: S1: Collect original time series data of multiple dimensions, perform time synchronization processing on all original time series data, remove outliers, and then perform feature-by-feature normalization processing to obtain time series feature data; S2: Based on the time series feature data obtained in step S1, a dynamic coordination degree matrix is ​​established to obtain the dynamic coordination degree between devices. Based on the adjacency relationship and dynamic coordination degree of the devices, a dynamic adjacency matrix is ​​established to generate a dynamic spatiotemporal graph to reflect the coordination perception characteristics of the devices. S3: extract the collaborative perception features of the device to obtain the collaborative feature vector of the device, and send the collaborative feature vector of the device to the device anomaly detection branch and the line loss collaborative anomaly detection branch respectively, and output the device anomaly detection result A and the line loss collaborative anomaly detection result S respectively; S4: Generate a strategy adjustment factor F based on the device's abnormal detection result A, line loss collaborative abnormal detection result S, power grid load status, and environmental risk index, and optimize the dynamic space-time graph in step 2 based on the strategy adjustment factor F to obtain an optimized dynamic space-time graph. The system simultaneously generates a "strategy optimization instruction" to dynamically adjust the abnormal detection threshold of each device in the next round of detection to achieve sensitivity adaptation; S5: Based on the optimized dynamic space-time diagram and the device anomaly detection result A and line loss collaborative anomaly detection result S, a multi-device collaborative response mechanism is executed. For the detected collaborative anomalies, combined with the collaborative relationship, partitioned, hierarchical, and dynamic response actions are implemented.

[0007] Preferably, the multiple dimensions in step S1 include current, voltage, active power, temperature and vibration amplitude of the equipment.

[0008] Preferably, in the time synchronization process in step S1, for low-sampling devices, a collaborative trend difference compensation algorithm is used to jointly compensate for missing data through local trends of neighboring devices in a collaborative relationship.

[0009] Preferably, the dynamic coordination degree calculation formula in step S2 is as follows: ; in, Indicates and Equipment at time The feature cosine similarity at each moment, is the time window, For equipment With equipment in the past The coordinated anomaly index of abnormal events within a cycle reflects the frequency of coordinated abnormal behaviors of two devices. This item significantly enhances the system's modeling capabilities for "collaborative abnormal risk areas" and solves the problem that pure "similarity" modeling cannot capture implicit collaborative abnormal risks.

[0010] Preferably, the formula for generating the dynamic space-time graph in step S2 is as follows: ; in, is a collection of device nodes. is the dynamic adjacency matrix; The dynamic adjacency matrix is ​​constructed by integrating the adjacency relationship and dynamic coordination degree of the devices, and the calculation formula is: ; in, For equipment and equipment Edge weights in physical topology, is the balance factor, For equipment and equipment The feature matrix is ​​the most recent The time series feature distribution difference term within time steps, The calculation formula introduces a difference penalty term. When the characteristics between devices change dramatically but the abnormal coordination is high, the system still retains the basic dependence on the physical topology, reducing the risk of false correlation. It is particularly suitable for graph relationship control when power equipment is affected by the environment and produces abnormalities.

[0011] Preferably, after entering the device anomaly detection branch in step S3, the anomaly probability of a single device is output through a single device detection head for each collaborative perception vector, and the device is determined to be abnormal when it exceeds 50%.

[0012] Preferably, after entering the line loss collaborative anomaly detection branch in step S3, the collaborative feature vectors of all devices are aggregated, and the global collaborative feature is obtained by collaborative weighting. The global collaborative feature is input into the global detection head, and the line loss collaborative anomaly probability is output. When it exceeds 50%, the whole network collaborative anomaly is determined.

[0013] Preferably, in step S4, the dynamic space-time graph in step 2 is structurally optimized based on the strategy adjustment factor F to obtain an optimized dynamic space-time graph, which is specifically: ; in, is the policy adjustment factor of device i, is the policy adjustment factor of device j, is the original edge weight of the dynamic space-time graph, is the adaptive adjustment coefficient. Comprehensive reflection equipment The current "abnormal risk perception intensity". Achieve "synergistic relationship enhancement" in high-risk areas, increase the focus of subsequent detection on abnormality-prone areas, and dynamically adjust the detection sensitivity through a risk grading mechanism to achieve adaptive sensitivity management and improve detection accuracy and robustness in dynamic power grid scenarios.

[0014] And a multi-device collaborative line loss anomaly collaborative detection system, comprising: The data acquisition module is configured to collect multi-device multi-source data and pre-process the collected raw data, including time synchronization processing of the raw data, removal of outliers, and feature-by-feature normalization processing to obtain time series feature data, and output the time series feature data to the dynamic graph module and the detection module; The dynamic graph module is configured to receive the time series feature data from the data acquisition module, and construct a dynamic coordination degree matrix based on the data, calculate the dynamic coordination degree, and integrate the adjacency relationship of the devices and the dynamic coordination degree to establish a dynamic adjacency matrix, generate a dynamic spatiotemporal graph, and output it to the detection module; The detection module is configured to receive the dynamic spatiotemporal graph and the time series feature data of the device, perform feature propagation through the dynamic spatiotemporal graph, extract the collaborative feature vector of each device, and send the collaborative feature vector to the device anomaly detection branch and the line loss collaborative anomaly detection branch respectively, output the device anomaly detection result A and the line loss collaborative anomaly detection result S respectively, and output them to the adaptive module and the collaborative module; The adaptive module is set to generate a strategy adjustment factor based on the current anomaly detection result A and the line loss collaborative anomaly detection result S, combined with the real-time grid load status and the environmental risk index , and adjust the factor Optimize the edge weights of the dynamic spatiotemporal graph and simultaneously detect the strategy adjustment factors The abnormal detection threshold of each device in the next round of detection is dynamically adjusted based on the range of the device, and output to the collaborative module; Collaboration module: It is configured to execute a multi-device collaborative response mechanism based on the optimized dynamic spatiotemporal graph and the device's anomaly detection result A and line loss collaborative anomaly detection result S. It implements partitioned, hierarchical, and dynamic response actions for the detected collaborative anomalies in combination with the collaborative relationship.

[0015] In summary, the technical effects and advantages of the present invention are as follows: This solution can comprehensively monitor the operating status of each device in the power system by introducing a dynamic detection mechanism for multi-device collaboration, combined with real-time data analysis and adaptive learning algorithms, and can capture line loss anomalies caused by the collaborative work of devices in real time, and can autonomously optimize the anomaly detection strategy according to the actual status of the equipment and environmental changes, and improve real-time performance and accuracy. Through deep data fusion, the system can not only analyze the individual status of each device, but also comprehensively evaluate the synergy between devices, thereby identifying line loss anomalies caused by device interactions. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the method flow in the present invention; Figure 2 It is a schematic diagram of the system structure in the present invention. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0018] like Figure 1 As shown, the multi-device collaborative line loss anomaly collaborative detection method includes: S1: Collect original time series data of multiple dimensions, perform time synchronization processing on all original time series data, remove outliers, and then perform feature-by-feature normalization processing to obtain time series feature data; S2: Based on the time series feature data obtained in step S1, a dynamic coordination degree matrix is ​​established to obtain the dynamic coordination degree between devices. Based on the adjacency relationship and dynamic coordination degree of the devices, a dynamic adjacency matrix is ​​established to generate a dynamic time-space graph to reflect the collaborative working characteristics of the devices. S3: extract the collaborative perception features of the device to obtain the collaborative feature vector of the device, and send the collaborative feature vector of the device to the device anomaly detection branch and the line loss collaborative anomaly detection branch respectively, and output the device anomaly detection result A and the line loss collaborative anomaly detection result S respectively; S4: Generate a strategy adjustment factor F based on the device's abnormal detection result A, line loss collaborative abnormal detection result S, power grid load status, and environmental risk index, and optimize the dynamic space-time graph in step 2 based on the strategy adjustment factor F to obtain an optimized dynamic space-time graph. The system simultaneously generates a "strategy optimization instruction" to dynamically adjust the abnormal detection threshold of each device in the next round of detection to achieve sensitivity adaptation; S5: Based on the optimized dynamic space-time diagram and the device anomaly detection result A and line loss collaborative anomaly detection result S, a multi-device collaborative response mechanism is executed. For the detected collaborative anomalies, combined with the collaborative relationship, partitioned, hierarchical, and dynamic response actions are implemented.

[0019] Example 1

[0020] In step S1, the system first performs time synchronization processing, using the highest sampling frequency in the entire network (e.g. 1 second) as a benchmark, all device data are uniformly resampled to 1 second granularity. For low-sampling devices, a collaborative trend interpolation algorithm is used to consider neighboring devices that have a collaborative relationship physically or electrically (the set is denoted as ) and jointly interpolate missing data. For example, when device 3 is When the second data is missing, the system reference device 3 is The historical average And the average values ​​of neighbor devices 1 and 2 in the same time period , , which is completed by the following collaborative weighted formula: ; in, For device 3 The interpolated value of is the harmonic coefficient, and its value range is ,For example To strengthen your own data trends, is the collaborative neighbor set, Indicates the device exist The local mean within the window.

[0021] After interpolation, the system introduces a collaborative exception elimination mechanism based on the device With Neighbors The local mean of and standard deviation , define the rejection threshold ,For example If a data point , it is considered as an outlier and is removed. For example, device 1 Current value , but the current range of adjacent devices is , the system identifies the point as an outlier across the collaborative interval and directly removes it.

[0022] Finally, the system cleans the data Perform feature-by-feature normalization, and map all eigenvalues ​​to interval, eliminating the differences in characteristic scales of different devices and forming structured time-series alignment data , each for The matrix ( To unify the step size, is the feature dimension), as the input for the next step of “spatiotemporal dependency modeling”.

[0023] In step S2, the system receives the output from the previous step , that is, in the power system The multi-dimensional feature matrix of the time-series alignment and standardization of each device is used as the input for dynamic graph modeling. In order to achieve the goal of "line loss anomaly detection under multi-device collaboration", a dynamic spatiotemporal collaborative graph modeling mechanism (DSTGM mechanism) is proposed, which is specially customized for the collaborative characteristics of power system equipment and actual scenarios to solve the shortcomings of existing technologies in dealing with dynamic collaborative relationships. The collaborative relationship between equipment in the power system is not fixed. Affected by factors such as geographical location, equipment operating status, and environmental load fluctuations, the equipment collaboration strength has significant spatiotemporal dynamics, and traditional static topology diagrams or simple diagrams based on physical connections cannot accurately reflect this change.

[0024] Therefore, this step designs a dynamically updateable graph structure ,in is a collection of device nodes. It is a dynamic adjacency matrix. The specific generation process is as follows: The system first uses the output of step 1 , construct a dynamic coordination degree matrix based on equipment coordination relationship , which measures the recent spatiotemporal dependencies between devices. The calculation of the synergy not only considers the local feature similarity between devices, but also introduces a synergy compensation factor designed specifically for power systems to reflect the contribution of abnormal consistency in physical or electrical relationships.

[0025] The following synergy calculation formula is defined: ; in, Indicates and Equipment at time The feature cosine similarity at each moment, is the time window, For equipment With equipment in the past The coordinated abnormality index of abnormal events (such as abnormal line loss and frequent load fluctuations) within a cycle reflects the frequency of coordinated abnormal behaviors of the two devices. is the compensation factor (e.g. ), improve the system's sensitivity to "potential collaborative anomalies".

[0026] For example, when device 1 and device 2 are in the nearest If there are 6 coordinated anomalies in the abnormal events within the hour, then , if the cosine similarity is ,but Compensated to This item significantly enhances the system's modeling capabilities for "collaborative abnormality risk areas" and solves the problem that pure "similarity" modeling cannot capture implicit collaborative abnormality risks.

[0027] Next, the system integrates static physical adjacencies Dynamic synergy with the above , construct a dynamic adjacency matrix , the formula is: ; in, For equipment and equipment Edge weights in physical topology, is the balance factor (e.g. ), For equipment and equipment The feature matrix is ​​recently The difference term of the time series feature distribution within time steps (Frobenius norm), is the regularization factor (e.g. ), to prevent excessive expansion of edge weights caused by abnormal similarity. This formula introduces a difference penalty term. When the characteristics between devices change drastically but the abnormal coordination is high, the system still retains the basic dependence on the physical topology, reducing the risk of false associations. It is particularly suitable for graph relationship control when power equipment is affected by the environment (such as sudden weather, system-level fluctuations) and abnormalities occur.

[0028] For example, when device 1 (transformer) and device 3 (distribution line) are Under the physical relationship of Under the influence of However, since both The feature fluctuations are large within each time step ( ), the final dynamic edge weight , the edge weight is moderately suppressed to balance the collaborative risk perception and feature differences.

[0029] Through the above dynamic adjacency matrix The system generates a dynamic space-time graph , which can accurately reflect the collaborative working characteristics, abnormal propagation risks and potential weak or strong dependencies of power system equipment in the current period, and significantly improve the reliability of subsequent abnormality detection. Based on the cleaned high-quality data, this step establishes a dynamic graph structure of "collaborative abnormality perception + timing difference penalty" with the characteristics of the power industry, specifically solving the industry pain points of "weak collaborative anomalies are difficult to detect and the graph structure cannot be dynamically updated" in the collaborative anomaly detection of multiple power devices, and significantly improves the system's ability to detect hidden problems such as line loss anomalies.

[0030] final, As a complete dynamic graph structure, it is input into the subsequent detection model, providing graph structure support for efficient and intelligent anomaly detection under multi-device collaboration.

[0031] In step S3, this step receives the dynamic graph structure output from step 2 and The timing characteristics matrix of each device , designed an electric power collaborative multi-task detection mechanism (ECMTD mechanism) with "collaborative perception capability" to achieve synchronous detection of "single device abnormality" and "line loss abnormality under multi-device collaboration" in the power system.

[0032] The system first uses the graph structure Perform feature propagation to extract collaborative perception features of each device , specifically: ; in, For equipment With Neighbors The edge weight of are trainable graph convolution parameters, is the activation function. This mechanism combines the device and the characteristics of its neighboring devices, we get , i.e., equipment The collaborative feature vector of .

[0033] Then, the system enters the dual-branch detection phase, and each device Will be sent to two detection branches respectively: 1. Device anomaly detection branch: for each , through a single device detection head , output the abnormal prediction probability of a single device ,Right now: ; in, and is the parameter of the device anomaly detection branch, is Sigmoid activation, It is The abnormal probability of a device, value , Then determine the device abnormal.

[0034] 2. Line loss collaborative anomaly detection branch: The system aggregates the collaborative features of all devices , forming global collaborative features through collaborative weighting mechanism : ; Will Input to global detection head , output system-level line loss collaborative anomaly prediction value ,Right now: ; in, and are the parameters of the global detection branch, , as the predicted probability of whether there is "abnormal collaborative line loss" in the current period of the entire network, It is determined that there is an abnormality in the coordination of the entire network.

[0035] During the training process of the system, the loss function It consists of the following three parts: 1. Single device anomaly detection loss ; 2. System Collaborative Anomaly Detection Loss ; 3. Collaborative consistency regularity , which is used to avoid the “weak correlation false positives” in the figure.

[0036] The complete loss is expressed as: ; This step, as the patented core detection module, directly achieves the goal of "line loss anomaly detection under multi-device collaboration" and outputs and Provide input data for subsequent "dynamic optimization strategy" and "cooperative response mechanism" to realize intelligent collaborative abnormal management in the multi-device environment of the power system.

[0037] The technical solution in the above-mentioned embodiment of the present application has at least the following technical effects or advantages: This solution can comprehensively monitor the operating status of each device in the power system by introducing a dynamic detection mechanism of multi-device collaboration, combined with real-time data analysis and adaptive learning algorithms, and can capture line loss anomalies caused by the collaborative work of devices in real time, and can autonomously optimize the anomaly detection strategy according to the actual status of the equipment and environmental changes, and improve real-time performance and accuracy. Through deep data fusion, the system can not only analyze the individual status of each device, but also comprehensively evaluate the synergy between devices, thereby identifying line loss anomalies caused by device interactions.

[0038] Example 2

[0039] The present application also provides a multi-device collaborative line loss anomaly collaborative detection system, such as Figure 2 As shown, including: The data acquisition module is configured to collect multi-device multi-source data and pre-process the collected raw data, including time synchronization processing of the raw data, removal of outliers, and feature-by-feature normalization processing to obtain time series feature data, and output the time series feature data to the dynamic graph module and the detection module; The dynamic graph module is configured to receive the time series feature data from the data acquisition module, and construct a dynamic coordination degree matrix based on the data, calculate the dynamic coordination degree, and integrate the adjacency relationship of the devices and the dynamic coordination degree to establish a dynamic adjacency matrix, generate a dynamic spatiotemporal graph, and output it to the detection module; The detection module is configured to receive the dynamic spatiotemporal graph and the time series feature data of the device, perform feature propagation through the dynamic spatiotemporal graph, extract the collaborative feature vector of each device, and send the collaborative feature vector to the device anomaly detection branch and the line loss collaborative anomaly detection branch respectively, output the device anomaly detection result A and the line loss collaborative anomaly detection result S respectively, and output them to the adaptive module and the collaborative module; The adaptive module is set to generate a strategy adjustment factor based on the current anomaly detection result A and the line loss collaborative anomaly detection result S, combined with the real-time grid load status and the environmental risk index , and adjust the factor Optimize the edge weights of the dynamic spatiotemporal graph and simultaneously detect the strategy adjustment factors The abnormal detection threshold of each device in the next round of detection is dynamically adjusted based on the range of the device, and output to the collaborative module; Collaboration module: It is configured to execute a multi-device collaborative response mechanism based on the optimized dynamic spatiotemporal graph and the device's anomaly detection result A and line loss collaborative anomaly detection result S. It implements partitioned, hierarchical, and dynamic response actions for the detected collaborative anomalies in combination with the collaborative relationship.

[0040] The adaptive module receives the output from the detection module. (Device abnormality detection result) and (System collaborative line loss anomaly detection results), an electric power dynamic collaborative optimization mechanism (EDCOM mechanism) was designed to adaptively optimize subsequent detection sensitivity and dynamic graph structure according to the detection results and the current state of the power grid, and improve the adaptability and robustness of the system in a changeable power grid environment. The actual operating environment of the power system is highly dynamic. Affected by seasonal load fluctuations, equipment operating hours, sudden failures or weather, the collaborative relationship and abnormal risks between equipment vary greatly at different stages. To this end, the system has designed an adaptive strategy optimization mechanism of "joint drive of detection results and environmental characteristics", and at the same time, the "dynamic graph structure " and "Detection strategy parameters" are updated in real time.

[0041] First, the system is based on the current detection results and , combined with real-time grid load status (such as the current load rate of the entire network) and environmental factors (such as temperature, humidity, wind speed, etc.) to generate strategy adjustment factors , used to dynamically adjust the device The sensitivity of is related to the graph. It is defined as: ; in, is the strategy coefficient (e.g. , wait), For equipment Environmental risk index (dynamically calculated based on the equipment's surrounding environment). Comprehensive reflection equipment The current “abnormal risk perception intensity”.

[0042] Then, the system uses Edge weights in dynamic graphs Carry out "graph structure optimization based on detection feedback" to realize the cross-step graph structure adaptation mechanism, specifically: ; in, The original edge weights of the dynamic graph built for the dynamic graph module, is the adaptive adjustment coefficient (such as ). This mechanism dynamically improves the edge weight of the "high-risk collaborative area" based on the current detection output of the equipment and the environmental situation, forming a "graph enhancement mechanism" for sensitive areas.

[0043] For example, when device 1 and device 2 , and the two devices , ,but , the system automatically strengthens the collaborative strength, and the subsequent detection model will pay more attention to the collaborative anomalies in this area to improve the detection accuracy.

[0044] The system simultaneously generates "strategy optimization instructions" to dynamically adjust the anomaly detection thresholds of each device in the next round of detection , to achieve sensitivity adaptation. The specific rules are: like (high risk threshold), reduce ; like (low risk threshold), appropriately increase ; The middle range maintains the current sensitivity.

[0045] For example, if the device of and ,but The next round of abnormal judgment threshold Will be from Down to , to achieve "improved detection sensitivity of key equipment" and respond to periodic sudden risks in the power system.

[0046] In the adaptive module, through "cross-step graph structure optimization + detection strategy tuning", the detection results can be used to reversely adjust the graph model. The innovation of this mechanism is: Combining multi-source environmental characteristics ( and ) and the test results to construct a "multi-factor driven risk index ”; Graph structure optimization: dynamic adjustment Achieve "synergistic relationship strengthening" in high-risk areas and increase the attention of subsequent testing to abnormal high-incidence areas; Strategy tuning: Through the risk grading mechanism, the detection sensitivity is dynamically adjusted to achieve adaptive sensitivity management, thereby improving the detection accuracy and robustness in dynamic power grid scenarios.

[0047] Finally, the system outputs: Optimized graph structure , as the input of the next round of collaborative detection model; The optimized detection strategy for each device (such as the adjusted ).

[0048] The collaborative module receives the optimized dynamic graph structure output by the adaptive module , detection strategy parameters , and the detection results of the detection module and , execute the multi-device coordinated response mechanism, and implement partitioned, hierarchical, and dynamic response actions for the detected coordinated anomalies in combination with the current coordinated relationship to achieve accurate abnormal handling of the power system.

[0049] First, the system is based on , the collaborative subgraph division is carried out in the power grid equipment network according to the following: ; in , indicating that devices with a synergy strength higher than 0.6 will be included in the same sub-graph , forming a "cooperative response unit" under a physical or electrical coupling relationship.

[0050] In this embodiment, the substation 4 downstream power distribution equipment , due to the close electrical connection and high load coordination, a sub-graph is formed ; Remote power terminal With neighboring terminals , coordination , divided into subgraphs .

[0051] Next, the system implements partition + hierarchical response actions based on the detection results.

[0052] Single device level response: If ,equipment Directly enter local response and execute actions such as power outage, load reduction, and interconnection line isolation; if ,equipment Maintain normal operation and only log warnings.

[0053] If the power distribution equipment Test results , the system immediately triggers The power reduction action reduces the node load by 30%.

[0054] Coordinated response within a subgraph: If ,or , then the subgraph Multiple devices are linked internally to execute "high-order joint response", such as branch isolation, sub-graph level load diversion, etc.

[0055] If in the subgraph ( )middle, and Abnormal, and , the system automatically executes The "regional group isolation" strategy cuts off Related branches, at the same time The load is diverted to the neighbor with the highest degree of coordination .

[0056] Linkage sequence and action allocation: based on Internal edge rights give priority to controlling high synergy strength equipment to avoid abnormal spread. For equipment pairs, perform "linked load adjustment", such as the faulty equipment The load is preferentially diverted to .

[0057] like and Edge Rights ,when After a power outage is triggered, the system will first transfer its remaining load to , ensuring the stability of local power supply in the system. Cross-subgraph response coordination: If the sub-figure and There is a high synergy between ),but If the abnormal threshold is not reached, The response strategy is to mitigate or delay the response to avoid false triggering.

[0058] like (downstream of substation) and (Remote power terminal) collaborative relationship ; When the power failure response is triggered, If the abnormal conditions are not met, the system will only Send out a "warning signal" and plan to confirm again in the next cycle Whether actual actions need to be executed to prevent chained misoperations.

[0059] The final output of the system is: Sub-image level response action table , each It includes "linkage sequence + response type + diversion plan"; it is roughly as follows: Subgraph : - Linkage order: v1 ➡ v3 ➡ v2 - Response Type: v1: power outage + load transfer to v3 v3: Load takeover + tie line disconnection v2: Reduce load by 20% - Diversion plan: After v1 loses power, 80% of the load is transferred to v3 and 20% of the load is diverted to v2.

[0060] meaning: Clarify the execution order of response actions of multiple devices in a subgraph (which one comes first and which one comes later, to avoid chaotic responses); Specify the specific action type for each device (power off / load reduction / load takeover, etc.); Describe the load distribution strategy, which specifies how the load is transferred between high-coordination intensity devices.

[0061] The specific response instructions for each device, including the local action type (power off / downgrade / divergence) and action priority, are as follows: Device v1: - Action Type: Power Off - Priority: High (need to be executed first) Device v3: - Action type: Take over load + disconnect the tie line - Priority: Medium Device v2: - Action type: Load reduction 20% - Priority: Low illustrate: The local response actions of a single device have priorities to control the order of execution; It can be directly parsed by the scheduling or automation system to perform precise device-level actions.

[0062] Cross-subgraph slow-release control strategy to reduce the risk of "collaborative link misresponse"; Subgraph and : - Synergy Strength: 0.65 - Slow-release strategy: When performing hard isolation, The response action is delayed for 5 minutes; Postpone power outage and only enter the "load reduction + status observation" mode; like If the abnormal risk does not increase within 5 minutes, the response action will be canceled.

[0063] illustrate: In view of the high coordination relationship between different sub-graphs, the system automatically generates control strategies of "response delay" or "observation slow release" to avoid cross-zone false triggering.

[0064] The technical solutions in the above-mentioned embodiments of the present application have at least the following technical effects or advantages: it overcomes the problem that the existing methods cannot adapt to the dynamic changes of power equipment in real time, and enhances the system's adaptability to the complexity of the power network through adaptive optimization of intelligent algorithms.

[0065] The working principle is as follows: By introducing a dynamic detection mechanism for multi-device collaboration, combined with real-time data analysis and adaptive learning algorithms, it can comprehensively monitor the operating status of each device in the power system, and capture line loss anomalies caused by the collaborative work of devices in real time. According to the actual status of the equipment and environmental changes, it can autonomously optimize the anomaly detection strategy to improve real-time performance and accuracy. Through deep data fusion, the system can not only analyze the individual status of each device, but also comprehensively evaluate the synergy between devices, thereby identifying line loss anomalies caused by device interactions.

[0066] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A multi-device collaborative line loss anomaly detection method, characterized in that: include: S1: Collect original time series data of multiple dimensions, perform time synchronization processing on all original time series data, remove outliers, and then perform feature-by-feature normalization processing to obtain time series feature data; S2: Based on the time series feature data obtained in step S1, a dynamic coordination degree matrix is ​​established to obtain the dynamic coordination degree between devices. Based on the adjacency relationship and dynamic coordination degree of the devices, a dynamic adjacency matrix is ​​established to generate a dynamic spatiotemporal graph to reflect the collaborative perception characteristics of the devices. S3: extract the collaborative perception features of the device to obtain the collaborative feature vector of the device, and send the collaborative feature vector of the device to the device anomaly detection branch and the line loss collaborative anomaly detection branch respectively, and output the device anomaly detection result A and the line loss collaborative anomaly detection result S respectively; S4: Generate a strategy adjustment factor F based on the abnormal detection result A of the device, the line loss collaborative abnormal detection result S, the power grid load status and the environmental risk index, and perform structural optimization on the dynamic space-time graph in step 2 based on the strategy adjustment factor F to obtain the optimized dynamic space-time graph. The system synchronously generates a strategy optimization instruction to dynamically adjust the abnormal detection threshold of each device in the next round of detection to achieve sensitivity adaptation. S5: Based on the optimized dynamic space-time diagram and the device anomaly detection result A and line loss collaborative anomaly detection result S, a multi-device collaborative response mechanism is executed. For the detected collaborative anomalies, combined with the collaborative relationship, partitioned, hierarchical, and dynamic response actions are implemented.

2. The multi-device collaborative line loss anomaly detection method according to claim 1, characterized in that: The multiple dimensions in step S1 include current, voltage, active power, temperature and vibration amplitude of the equipment.

3. The multi-device collaborative line loss anomaly detection method according to claim 1, characterized in that: In the time synchronization process in step S1, for low-sampling devices, a collaborative trend difference compensation algorithm is used to jointly compensate for missing data through local trends of neighboring devices in a collaborative relationship.

4. The multi-device collaborative line loss anomaly detection method according to claim 1, characterized in that: The dynamic coordination degree calculation formula in step S2 is as follows: ； in, Indicates and Equipment at time The feature cosine similarity at each moment, is the time window, For equipment With equipment in the past The coordinated anomaly index of abnormal events within a cycle reflects the frequency of coordinated abnormal behaviors of two devices. is the compensation factor.

5. The multi-device collaborative line loss anomaly detection method according to claim 1, characterized in that: The formula for generating the dynamic space-time graph in step S2 is as follows: ； in, is a collection of device nodes. is a dynamic adjacency matrix; the dynamic adjacency matrix is ​​constructed by integrating the adjacency relationship and dynamic coordination degree of the devices, and the calculation formula is: ； in, For equipment and equipment Edge weights in physical topology, is the balance factor, For equipment and equipment The feature matrix is ​​recently The time series feature distribution difference term within time steps, is the regularization factor.

6. The multi-device collaborative line loss anomaly detection method according to claim 1, characterized in that: After entering the device anomaly detection branch in step S3, the anomaly probability of a single device is output through a single device detection head for each collaborative perception vector, and the device is determined to be abnormal when it exceeds 50%.

7. The multi-device collaborative line loss anomaly detection method according to claim 1, characterized in that: After entering the line loss collaborative anomaly detection branch in step S3, the collaborative feature vectors of all devices are aggregated, and the global collaborative feature is obtained by collaborative weighting. The global collaborative feature is input into the global detection head, and the line loss collaborative anomaly probability is output. When it exceeds 50%, the whole network collaborative anomaly is determined.

8. The multi-device collaborative line loss anomaly detection method according to claim 1, characterized in that: In step S4, the dynamic space-time graph in step 2 is structurally optimized based on the strategy adjustment factor F to obtain an optimized dynamic space-time graph, which is specifically: ； in, is the policy adjustment factor of device i, is the policy adjustment factor of device j, is the original edge weight of the dynamic space-time graph, is the adaptive adjustment coefficient, is the optimized edge weight of the dynamic spatiotemporal graph.

9. A multi-device collaborative line loss anomaly collaborative detection system, characterized in that: include: The data acquisition module is configured to collect multi-device multi-source data and pre-process the collected raw data, including time synchronization processing of the raw data, elimination of outliers, and feature-by-feature normalization processing to obtain time series feature data, and output the time series feature data to the dynamic graph module and the detection module; the dynamic graph module is configured to receive the time series feature data from the data acquisition module, and construct a dynamic coordination degree matrix based on the data, calculate the dynamic coordination degree, and integrate the adjacency relationship of the device and the dynamic coordination degree to establish a dynamic adjacency matrix, generate a dynamic space-time graph and output it to the detection module; the detection module is configured to accept the dynamic space-time graph and the time series feature data of the device, perform feature propagation through the dynamic space-time graph, extract the collaborative feature vector of each device, and send the collaborative feature vector to the device anomaly detection branch and the line loss collaborative anomaly detection branch respectively, output the device anomaly detection result A and the line loss collaborative anomaly detection result S respectively, and output them to the adaptive module and the collaborative module; The adaptive module is set to generate a strategy adjustment factor based on the current anomaly detection result A and the line loss collaborative anomaly detection result S, combined with the real-time grid load status and the environmental risk index , and adjust the factor Optimize the edge weights of the dynamic spatiotemporal graph and simultaneously detect the strategy adjustment factors The abnormal detection threshold of each device in the next round of detection is dynamically adjusted according to the range of the device, and the output is in the collaborative module; the collaborative module is set to execute the multi-device collaborative response mechanism based on the optimized dynamic spatiotemporal graph and the device's abnormal detection result A and the line loss collaborative anomaly detection result S. For the detected collaborative anomalies, combined with the collaborative relationship, partitioned, graded, and dynamic response actions are implemented.

Images