An abnormality detection method, device, equipment, medium and program of a power transmission line

Abstract

The application discloses a kind of power transmission line anomaly detection method, device, equipment, medium and program.The method comprises: obtaining the multi-source operating state data of each key node in power transmission line;The multi-source operating state data is carried out multi-source fusion processing, and comprehensive feature vector is generated;By at least two preset anomaly detection models, respectively, the comprehensive feature vector is carried out anomaly analysis, and the corresponding anomaly detection result is generated;Based on the historical performance of anomaly detection model and the anomaly detection result, the model score of each anomaly detection model is determined, and the target anomaly detection result is determined from each anomaly detection result according to the model score.The technical scheme of the application can avoid the limitation that single detection model is poor in adaptability, improve the adaptability of anomaly detection, and also can realize comprehensive, accurately identify the abnormal situation of power transmission line.

Description

Technical Field

[0001] This invention relates to the field of power system safety technology, and in particular to a method, apparatus, equipment, medium, and procedure for detecting anomalies in transmission lines. Background Technology

[0002] As one of the important facilities of the power system, the safe and stable operation of transmission lines is crucial to the safe operation of the power system.

[0003] To ensure the safe and stable operation of power transmission lines, existing technologies typically involve using inspection robots equipped with sensors to move along the transmission lines, collect certain status data of the transmission lines, and transmit this data to a control and processing terminal to identify anomalies in the transmission lines using a pre-set detection model.

[0004] However, the existing technical solutions mentioned above not only rely on a single and fixed detection model, making it difficult to cope with the complex and ever-changing operating environment of transmission lines (such as wind deflection, galloping, and fluctuations in ambient temperature and humidity), but also the collected data is relatively limited, making it difficult to comprehensively identify the various types of anomalies in transmission lines. As a result, the accuracy and adaptability of the current transmission line anomaly detection methods are low, and they cannot meet the current needs for anomaly detection in transmission lines. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, medium, and program for detecting anomalies in power transmission lines. It avoids the limitations of poor adaptability of a single detection model, improves the adaptability of anomaly detection, and enables comprehensive and accurate identification of anomalies in power transmission lines.

[0006] According to a first aspect of the present invention, a method for detecting anomalies in a transmission line is provided, comprising:

[0007] Acquire multi-source operational status data of key nodes in transmission lines;

[0008] The multi-source operating status data is subjected to multi-source fusion processing to generate a comprehensive feature vector;

[0009] Anomaly analysis is performed on the comprehensive feature vector using at least two preset anomaly detection models to generate corresponding anomaly detection results.

[0010] Based on the historical performance of the anomaly detection models and the anomaly detection results, the model score of each anomaly detection model is determined, and the target anomaly detection result is determined from each anomaly detection result according to the model score.

[0011] According to a second aspect of the present invention, an anomaly detection device for a transmission line is provided, comprising:

[0012] The acquisition module is used to acquire multi-source operating status data of key nodes in the transmission line;

[0013] The fusion module is used to perform multi-source fusion processing on the multi-source operating status data to generate a comprehensive feature vector;

[0014] The detection module is used to perform anomaly analysis on the comprehensive feature vector using at least two preset anomaly detection models, and generate corresponding anomaly detection results.

[0015] The determination module is used to determine the model score of each anomaly detection model based on the historical performance of the anomaly detection model and the anomaly detection results, and to determine the target anomaly detection result from each anomaly detection result according to the model score.

[0016] According to a third aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0017] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform an anomaly detection method for a power transmission line according to any embodiment of the present invention.

[0018] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement an anomaly detection method for a power transmission line as described in any embodiment of the present invention.

[0019] According to a fifth aspect of the present invention, a computer program product is provided, the computer program product comprising a computer program that, when executed by a processor, implements an anomaly detection method for a power transmission line as described in any embodiment of the present invention.

[0020] The technical solution of this invention can acquire multi-source operational status data of key nodes in a transmission line and perform multi-source fusion processing on the multi-source operational status data to generate a comprehensive feature vector. This can fully leverage the complementary value of the transmission line's operational status data, facilitating a more comprehensive and accurate identification of various anomalies such as apparent defects, overheating, and partial discharge. Then, using at least two preset anomaly detection models, anomaly analysis is performed on the comprehensive feature vector to generate corresponding anomaly detection results. Finally, based on the historical performance of the anomaly detection models and the anomaly detection results, the model score of each anomaly detection model is determined, and the target anomaly detection result is determined from the anomaly detection results based on the model score. This avoids the limitation of poor adaptability of a single detection model and improves the adaptability of anomaly detection.

[0021] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of an anomaly detection method for power transmission lines provided in Embodiment 1 of the present invention;

[0024] Figure 2 This is a flowchart of an anomaly detection method for power transmission lines provided in Embodiment 2 of the present invention;

[0025] Figure 3 This is a schematic diagram of the structure of an anomaly detection device for a power transmission line provided in Embodiment 3 of the present invention;

[0026] Figure 4 This is a schematic diagram of the structure of an electronic device that implements an anomaly detection method for power transmission lines according to an embodiment of the present invention. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0029] Example 1

[0030] Figure 1 This is a flowchart of a method for detecting anomalies in transmission lines according to Embodiment 1 of the present invention. This embodiment is applicable to the detection of anomalies in transmission lines within a power system. The method can be executed by an anomaly detection device for the transmission line, which can be implemented in hardware and / or software and configured in a power inspection system or edge electronic equipment. Figure 1 As shown, the method includes:

[0031] S101. Obtain multi-source operating status data of each key node in the transmission line.

[0032] Key nodes can be structural locations in transmission lines that are prone to failure or require close monitoring. Multi-source operational status data can be status data of transmission lines collected from different types of sensors, such as visible light images, infrared temperature data, ultrasonic signals, vibration signals, and environmental meteorological data.

[0033] Understandably, existing anomaly detection solutions typically rely on a single type of sensor (such as visible light or infrared) to collect data, making it difficult to comprehensively reflect the multi-dimensional state of transmission lines. Therefore, this embodiment utilizes multiple sensors deployed at key nodes of the transmission line to simultaneously collect various types of operational status data, thereby obtaining richer and more complementary status information.

[0034] For example, this embodiment can collect various operational status data of the transmission line in real time by deploying visible light cameras, infrared thermal imagers, ultrasonic sensors, vibration sensors, and meteorological sensors at key nodes of the transmission line. Then, the various operational status data of the transmission line can be transmitted to the anomaly detection device of the transmission line for anomaly detection through wireless communication networks or fiber optic communication technology.

[0035] Among them, visible light cameras can be used to acquire visible light images of the power transmission line itself and its surrounding environment; infrared thermal imagers can be used to detect the temperature distribution of power transmission line equipment and conductors; ultrasonic sensors can be used to monitor the sound wave signals generated by partial discharge of power transmission lines; vibration sensors can be used to sense the vibration characteristics of power transmission lines caused by wind deflection, galloping, or external impact; and meteorological sensors can be used to record parameters such as temperature, humidity, wind speed, air pressure, and precipitation intensity of the environment in which the power transmission line is located.

[0036] S102. Perform multi-source fusion processing on the multi-source operating status data to generate a comprehensive feature vector.

[0037] Among them, the comprehensive feature vector can be a feature vector that can comprehensively reflect the multi-dimensional state of the transmission line.

[0038] It should be noted that multi-source operational status data contains various types of operational status data, which differ in format and physical meaning. Directly using these data for anomaly analysis can easily lead to information fragmentation or redundancy. Therefore, this embodiment uses multi-source fusion processing to convert multi-source operational status data into feature vectors that comprehensively reflect the multi-dimensional state of transmission lines, thereby supporting subsequent anomaly detection analysis.

[0039] For example, after acquiring multi-source operating status data, features can be extracted from various types of operating status data in the multi-source operating status data to obtain initial feature vectors corresponding to various types of operating status data. Then, the initial feature vectors can be concatenated to obtain a comprehensive feature vector that can comprehensively reflect the operating status of the transmission line.

[0040] Optionally, the step of performing multi-source fusion processing on the multi-source operational status data to generate a comprehensive feature vector may include:

[0041] Feature extraction is performed on each running status data in the multi-source running status data to obtain the corresponding initial feature vector;

[0042] The initial feature vectors are weighted according to the preset feature weights to generate the comprehensive feature vector.

[0043] The initial feature vector can be a vector representing the key information of each type of operational status data (such as visible light images, infrared temperature data, and ultrasonic signals) after feature extraction processing. Feature weights can be used to measure the relative importance of each type of operational status data in constructing the comprehensive feature vector.

[0044] For example, this embodiment can take visible light images, infrared thermal imaging data, ultrasonic signals, vibration data and meteorological data as examples of multi-source operating status data to illustrate the implementation process of generating comprehensive feature vectors based on feature weights.

[0045] First, using feature extraction algorithms, features are extracted from visible light images, infrared thermal imaging data, ultrasonic signals, vibration data, and meteorological data in the multi-source operational status data, respectively, to obtain the initial feature vectors corresponding to the multi-source operational status data. ,in, In this context, 'i' can represent different sensors (e.g., 'v' can represent a visible light camera, 'ir' can represent an infrared thermal imager, 'us' can represent an ultrasonic sensor, 'va' can represent a vibration sensor, and 'ws' can represent a weather sensor). These can be represented as the initial feature vectors corresponding to the operating status data collected by the i-th type of sensor.

[0046] Then, based on the importance of the operational status data collected by the sensors or the importance of the sensor deployment location, corresponding weights can be assigned to the i-th type of sensors. , which serve as feature weights corresponding to various types of operational status data. The sum of the weights for each type of sensor must be non-negative and equal to 1.

[0047] Finally, the initial feature vectors corresponding to the operating status data collected by the i-th type of sensor can be sequentially generated. Weights corresponding to the i-th type of sensor We perform weighted summation to obtain the comprehensive feature vector.

[0048] It should be noted that feature weights can also be adjusted in real time using adaptive update rules. For example, the weights can be adjusted in real time based on the sensor's effectiveness score within the most recent time window.

[0049] Specifically, firstly, the weights of the i-th type of sensor can be initially adjusted:

[0050] ;

[0051] in, It can be the initial adjusted weight of the i-th type of sensor at the next time step after time t (the current time); It can be the weight at time t (the current time); It can represent the effectiveness score of the i-th type of sensor within the most recent time window (such as based on AUC, accuracy, information gain, etc.). It can represent the average of the effectiveness scores of all sensors; The preset learning rate can be used to control the magnitude of weight updates; exp() can be a natural exponential function.

[0052] Then, after the initial adjustment of the weights of various sensors, the weights of various sensors can be adjusted again as a whole through normalization to obtain the final weights of various sensors, so as to ensure that the sum of the final weights of various sensors is 1.

[0053] S103. Using at least two preset anomaly detection models, perform anomaly analysis on the comprehensive feature vector to generate corresponding anomaly detection results.

[0054] The anomaly detection model can be an algorithmic model used to determine whether there are abnormal states in transmission lines. For example, it can be a traditional machine learning model (such as random forest, SVM), a lightweight deep learning model (such as the MobileNet series, ShuffleNet), or a rule model based on domain knowledge.

[0055] It should be noted that, to avoid adaptability issues caused by the rigid structure or limited scenarios of a single model, this embodiment can pre-set multiple anomaly detection models to work in parallel. By leveraging the differences and complementarities among these models, the overall robustness of anomaly detection can be improved. Each anomaly detection model can be deployed in parallel as a Docker container, with each container independently loading its own anomaly detection inference logic to detect one or more anomalies in the transmission line.

[0056] Anomaly detection results can be the judgment results output after analyzing the comprehensive feature vector input to the anomaly detection model. These results typically include information such as whether an anomaly exists, the type of anomaly, and the confidence level. Anomaly analysis can include target detection in visible light images to identify apparent defects in the line structure, such as insulator damage, broken conductor strands, or overheating connectors; temperature threshold analysis of infrared thermal imaging data, combined with meteorological parameters to correct for the influence of ambient temperature and humidity on equipment temperature, to identify overheating fault points; spectral analysis of ultrasonic signals to extract the frequency characteristics of partial discharge signals and determine the discharge type and severity; and time-frequency domain analysis of vibration data to identify abnormal vibration modes of the line caused by wind deflection, galloping, or external forces.

[0057] For example, in this embodiment, at least two anomaly detection models with different anomaly detection inference logics can be pre-built and deployed. Then, after generating a comprehensive feature vector, the comprehensive feature vector can be input into each anomaly detection model in parallel, so that each anomaly detection model can independently perform analysis and inference and output its corresponding anomaly detection results, thereby realizing parallel judgment by multiple models.

[0058] S104. Based on the historical performance of the anomaly detection models and the anomaly detection results, determine the model score of each anomaly detection model, and determine the target anomaly detection result from each anomaly detection result according to the model score.

[0059] The historical performance of the anomaly detection model can be considered as an evaluation of its performance in past detection tasks, such as historical detection accuracy and recall. The model score can be used to quantify the reliability of the anomaly detection model's current detections. The target anomaly detection result can be the anomaly detection result selected from the outputs of multiple anomaly detection models that can be used as the final judgment criterion.

[0060] For example, after obtaining the anomaly detection results output in parallel by each anomaly detection model, the long-term reliability of the anomaly detection model can be determined by analyzing the historical performance evaluation of each anomaly detection model. Then, combined with the current anomaly detection results, the model score of each anomaly detection model can be determined. Finally, based on the model scores, the anomaly detection result corresponding to the anomaly detection model with the highest model score can be selected as the final target anomaly detection result.

[0061] The technical solution of this embodiment can acquire multi-source operational status data of key nodes in a transmission line and perform multi-source fusion processing on the multi-source operational status data to generate a comprehensive feature vector. This can fully leverage the complementary value of the transmission line's operational status data, facilitating a more comprehensive and accurate identification of various anomalies such as apparent defects, overheating, and partial discharge. Then, using at least two preset anomaly detection models, anomaly analysis is performed on the comprehensive feature vector to generate corresponding anomaly detection results. Finally, based on the historical performance of the anomaly detection models and the anomaly detection results, the model score of each anomaly detection model is determined, and the target anomaly detection result is determined from the anomaly detection results according to the model score. This avoids the limitation of poor adaptability of a single detection model and improves the adaptability of anomaly detection.

[0062] Based on the above embodiments, the present invention also provides an optional embodiment, which can be further optimized based on the above embodiments, and may specifically include:

[0063] Based on the target anomaly detection results, and combined with the temporal and spatial information of the multi-source operating status data, the transmission line is subjected to fault location and propagation path analysis to generate fault source tracing results.

[0064] The time information of the multi-source operational status data can be the data timestamp when each operational status data is collected or recorded, or the time interval between different sensors triggering alarms or recording anomalies when an anomaly is detected. The spatial information of the multi-source operational status data can be the deployment location information of each sensor in the transmission line, and the network structure information of the connection relationship between various components of the transmission line (towers, insulators, conductor segments, etc.).

[0065] It should be noted that, based on the determination that an anomaly exists in the transmission line, this invention can further determine the precise physical location of the fault and infer the range and direction (propagation path) of the fault's possible impact along the transmission line structure.

[0066] For example, this optional embodiment can integrate the timestamps of different sensor alarm events and their spatial coordinates (latitude and longitude, tower number, relative position, etc.) based on the time and space information of multi-source operating status data, analyze the temporal proximity and spatial correlation of multi-source abnormal signals, and thus screen out the sections or nodes most likely to fail, and achieve preliminary positioning.

[0067] Then, based on the network structure information of the connection relationship between the various components of the transmission line (towers, insulators, conductor segments, etc.), the distribution of abnormal intensity and the abnormal transmission relationship at different nodes can be analyzed. Furthermore, by combining the physical mechanism of the fault occurrence (such as electrical fault propagation along the line, mechanical fault localization, etc.), the possible origin of the fault and the diffusion path from the origin of the fault can be deduced.

[0068] The advantage of this setup is that it enables precise location of the fault source and clear analysis of the fault propagation path, significantly reducing troubleshooting time and maintenance costs.

[0069] Example 2

[0070] Figure 2 This is a flowchart of an anomaly detection method for transmission lines provided in Embodiment 2 of the present invention. This embodiment further optimizes step S104 of Embodiment 1, based on the historical performance of the anomaly detection model and the anomaly detection results, to determine the model score of each anomaly detection model, and then determines the target anomaly detection result from each anomaly detection result according to the model score. For example... Figure 2 As shown, the method includes:

[0071] S201. Obtain multi-source operating status data of each key node in the transmission line.

[0072] S202. Perform multi-source fusion processing on the multi-source operating status data to generate a comprehensive feature vector.

[0073] S203. Using at least two preset anomaly detection models, perform anomaly analysis on the comprehensive feature vector to generate corresponding anomaly detection results.

[0074] S204. Calculate the model score of the anomaly detection model based on the historical accuracy of the anomaly detection model, the confidence level corresponding to the anomaly detection results, the difference in anomaly detection results, and the model complexity.

[0075] Historical accuracy refers to the accuracy of anomaly detection models in identifying anomalies over a past period (such as the most recent N detection tasks), reflecting the long-term reliability of the anomaly detection model.

[0076] Confidence level represents the reliability of the anomaly detection model's output in this instance, typically expressed as a value between 0 and 1. 0 indicates low confidence, while 1 indicates high confidence. It's important to note that anomaly detection results can include, but are not limited to, the identified anomaly type and its probability. The probability value corresponding to the final anomaly category can be used as the confidence level of the anomaly detection result.

[0077] Anomaly detection result dissimilarity can be defined as the degree of difference between the current output of a particular anomaly detection model and the median or average of the outputs of all anomaly detection models.

[0078] Model complexity can be used as an indicator to quantify the computational or structural complexity of an anomaly detection model itself, such as the number of parameters, inference latency, or utilization of computing resources.

[0079] For example, the model score of the anomaly detection model can be calculated based on the model's historical accuracy, the confidence level of the anomaly detection results, the variance of the anomaly detection results, and the model complexity.

[0080] ;

[0081] Where i can represent the i-th anomaly detection model deployed in parallel; This can represent the model score corresponding to the i-th anomaly detection model; This can represent the historical accuracy of the i-th anomaly detection model; It can represent the confidence level of the anomaly detection result output by the i-th anomaly detection model; This can represent the degree of difference in anomaly detection results of the i-th anomaly detection model; Let represent the model complexity of the i-th anomaly detection model; are adjustable weight coefficients. These can be adjustable weighting coefficients used to balance the influence of various factors in the final model score.

[0082] S205. Determine the anomaly detection result corresponding to the anomaly detection model with the highest model score, and use it as the candidate anomaly detection result.

[0083] For example, determine the model score among all parallel-deployed anomaly detection models. The anomaly detection model with the highest value is denoted as the candidate anomaly detection model. Candidate anomaly detection model The output anomaly detection results are used as candidate anomaly detection results.

[0084] S206. Weight the confidence scores corresponding to each of the above-mentioned anomaly detection results to obtain the group confidence score.

[0085] Among them, the group confidence score can be used to translate the credibility of the anomaly detection results of the entire parallel-deployed anomaly detection model group.

[0086] For example, for i anomaly detection models deployed in parallel, the population confidence of the anomaly detection model group can be determined by the following formula:

[0087] ;

[0088] in, It can represent the population confidence of an anomaly detection model group; It can be the confidence fusion weight set based on the model score of the i-th anomaly detection model; It can represent the confidence level of the anomaly detection result output by the i-th anomaly detection model.

[0089] S207. If the model score corresponding to the candidate anomaly detection result and the group confidence score meet the preset conditions, then the candidate anomaly detection result is determined as the target anomaly detection result.

[0090] The preset conditions can be threshold conditions set to ensure the reliability of the final output results, such as score thresholds and / or group confidence thresholds. The thresholds can be set based on the experience of relevant personnel, or they can be appropriately determined automatically during the continuous anomaly detection process. This invention does not impose specific limitations on these thresholds. The target anomaly detection result can be the final determined anomaly judgment result.

[0091] For example, if the model score of the candidate anomaly detection model is not lower than a preset score threshold, and the corresponding group confidence score is not lower than a preset group confidence score threshold, then the current detection result can be considered reliable, and the candidate anomaly detection result can be determined as the final target anomaly detection result.

[0092] If any of the above conditions are not met, the candidate anomaly detection result can be marked as "requires review", prompting the operation and maintenance personnel to intervene manually or start a more detailed review process.

[0093] The technical solution of this embodiment can calculate the model score of the anomaly detection model based on the historical accuracy of the anomaly detection model, the confidence level corresponding to the anomaly detection result, the difference of the anomaly detection result, and the model complexity. The anomaly detection result corresponding to the anomaly detection model with the highest model score is determined as the candidate anomaly detection result. Finally, the confidence levels corresponding to each anomaly detection result are weighted to obtain the group confidence level. If the model score corresponding to the candidate anomaly detection result and the group confidence level meet the preset conditions, the candidate anomaly detection result is determined as the target anomaly detection result. This can significantly improve the accuracy of transmission line anomaly detection.

[0094] Based on the above embodiments, the present invention also provides an optional embodiment, which can be further optimized based on the above embodiments, specifically as follows:

[0095] After determining the target anomaly detection result based on the competition score, the method further includes updating the anomaly detection model based on the model score of the anomaly detection model.

[0096] It should be noted that after each anomaly detection task is executed and anomaly detection results are output, the internal parameters of the anomaly detection model (such as historical accuracy, model participation weights, etc.) or the allocation of external resources can be adjusted. This allows for the "survival of the fittest" of the anomaly detection model and continuous system evolution, thereby further avoiding the problem of a fixed anomaly detection model.

[0097] For example, after calculating the model score of the anomaly detection model, the historical accuracy of the anomaly detection model can be fine-tuned based on the difference between the model score and the average score (e.g., the historical accuracy of anomaly detection models with scores higher than the average can be improved, while those with lower scores can be reduced). Specifically, this can be done as follows:

[0098] ;

[0099] in, This can be the historical accuracy of the adjusted i-th anomaly detection model; It can be the current historical accuracy of the i-th anomaly detection model; This can represent the model score of the i-th anomaly detection model; It can represent the average model score of all anomaly detection models deployed in parallel; The maximum adjustment range for a single operation can be preset. It can be an adjustment factor.

[0100] Furthermore, the model participation weights of the anomaly detection model can be adjusted based on the difference between the model score and the average score. Specifically, this can be done as follows:

[0101] ;

[0102] in, This can represent the model participation weights of the i-th anomaly detection model at time t; γ can represent the model participation weights of the i-th anomaly detection model after adjustment; γ can represent the learning rate of the model participation weights.

[0103] It is understandable that the weights involved in the model can be directly or indirectly related to the allocation of computing resources (such as CPU / GPU time and memory) or affect the confidence fusion weights of the anomaly detection model.

[0104] In addition, long-term management strategies can be implemented based on model scores. For example, for anomaly detection models whose scores are consistently significantly higher than those of other models in the group, after verifying their stability, their computing resources can be increased (such as increasing CPU / GPU quotas or increasing task scheduling frequency) to enable them to play a greater role in subsequent anomaly detection. Alternatively, when the model score of an anomaly detection model or its updated historical accuracy falls below a preset elimination threshold multiple times consecutively, a retraining process (such as retraining it with the latest data) or a replacement process (such as using a backup model from the model library) can be triggered for that anomaly detection model.

[0105] The advantages of this setup are: it enables the dynamic evolution of the anomaly detection model and the adaptive allocation of computing resources, achieves efficient competition and adaptive deployment of multiple anomaly detection models, and ensures the long-term adaptability of the anomaly detection model to complex operating environments.

[0106] Example 3

[0107] Figure 3 This is a schematic diagram of the structure of a method and apparatus for detecting anomalies in power transmission lines according to Embodiment 3 of the present invention. Figure 3 As shown, the device includes:

[0108] The acquisition module 31 can be used to acquire multi-source operating status data of each key node in the transmission line;

[0109] The fusion module 32 can be used to perform multi-source fusion processing on the multi-source operating status data to generate a comprehensive feature vector;

[0110] The detection module 33 can be used to perform anomaly analysis on the comprehensive feature vector using at least two preset anomaly detection models, and generate corresponding anomaly detection results.

[0111] The determination module 34 can be used to determine the model score of each anomaly detection model based on the historical performance of the anomaly detection model and the anomaly detection results, and to determine the target anomaly detection result from each anomaly detection result according to the model score.

[0112] The technical solution of this embodiment can acquire multi-source operational status data of key nodes in a transmission line and perform multi-source fusion processing on the multi-source operational status data to generate a comprehensive feature vector. This can fully leverage the complementary value of the transmission line's operational status data, facilitating a more comprehensive and accurate identification of various anomalies such as apparent defects, overheating, and partial discharge. Then, using at least two preset anomaly detection models, anomaly analysis is performed on the comprehensive feature vector to generate corresponding anomaly detection results. Finally, based on the historical performance of the anomaly detection models and the anomaly detection results, the model score of each anomaly detection model is determined, and the target anomaly detection result is determined from the anomaly detection results according to the model score. This avoids the limitation of poor adaptability of a single detection model and improves the adaptability of anomaly detection.

[0113] Optionally, the fusion module 32 can be specifically used to extract features from each running state data in the multi-source running state data to obtain the corresponding initial feature vector;

[0114] The initial feature vectors are weighted according to the preset feature weights to generate the comprehensive feature vector.

[0115] Optionally, the determining module 34 may include: a calculation unit;

[0116] The calculation unit can be used to calculate the model score of the anomaly detection model based on the historical accuracy of the anomaly detection model, the confidence level corresponding to the anomaly detection results, the difference in anomaly detection results, and the model complexity.

[0117] Optionally, the determining module 34 may further include: a detection unit;

[0118] The detection unit can be used to determine the anomaly detection result corresponding to the anomaly detection model with the highest model score, as a candidate anomaly detection result;

[0119] The confidence scores corresponding to each of the aforementioned anomaly detection results are weighted to obtain the group confidence score;

[0120] If the model score corresponding to the candidate anomaly detection result and the group confidence score meet the preset conditions, then the candidate anomaly detection result is determined as the target anomaly detection result.

[0121] Optionally, the determining module 34 may further include: an updating unit;

[0122] The updating unit can be used to update the anomaly detection model based on the model score of the anomaly detection model after the target anomaly detection result is determined based on the competition score.

[0123] Optionally, the apparatus further includes: a generation module;

[0124] The generation module can be used to perform fault location and propagation path analysis on the transmission line based on the target anomaly detection results and the time and space information of the multi-source operating status data, and generate fault source tracing results.

[0125] The abnormality detection device for transmission lines provided in this embodiment of the invention can execute the abnormality detection method for transmission lines provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0126] Example 4

[0127] Figure 4 A schematic diagram of an electronic device 40 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0128] like Figure 4As shown, the electronic device 40 includes at least one processor 41 and a memory, such as a read-only memory (ROM) 42 or a random access memory (RAM) 43, communicatively connected to the at least one processor 41. The memory stores computer programs executable by the at least one processor. The processor 41 can perform various appropriate actions and processes based on the computer program stored in the ROM 42 or loaded from storage unit 48 into the RAM 43. The RAM 43 may also store various programs and data required for the operation of the electronic device 40. The processor 41, ROM 42, and RAM 43 are interconnected via a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.

[0129] Multiple components in electronic device 40 are connected to I / O interface 45, including: input unit 46, such as keyboard, mouse, etc.; output unit 47, such as various types of monitors, speakers, etc.; storage unit 48, such as disk, optical disk, etc.; and communication unit 49, such as network card, modem, wireless transceiver, etc. Communication unit 49 allows electronic device 40 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0130] Processor 41 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 41 performs the various methods and processes described above, such as anomaly detection methods for power transmission lines.

[0131] In some embodiments, the transmission line anomaly detection method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 40 via ROM 42 and / or communication unit 49. When the computer program is loaded into RAM 43 and executed by processor 41, one or more steps of the transmission line anomaly detection method described above may be performed. Alternatively, in other embodiments, processor 41 may be configured to perform the transmission line anomaly detection method by any other suitable means (e.g., by means of firmware).

[0132] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0133] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0134] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0135] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0136] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0137] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0138] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0139] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for detecting anomalies in power transmission lines, characterized in that, include: Acquire multi-source operational status data of key nodes in transmission lines; The multi-source operating status data is subjected to multi-source fusion processing to generate a comprehensive feature vector; Anomaly analysis is performed on the comprehensive feature vector using at least two preset anomaly detection models to generate corresponding anomaly detection results. Based on the historical performance of the anomaly detection models and the anomaly detection results, the model score of each anomaly detection model is determined, and the target anomaly detection result is determined from each anomaly detection result according to the model score.

2. The method according to claim 1, characterized in that, The process of performing multi-source fusion processing on the multi-source operational status data to generate a comprehensive feature vector includes: Feature extraction is performed on each running status data in the multi-source running status data to obtain the corresponding initial feature vector; The initial feature vectors are weighted according to the preset feature weights to generate the comprehensive feature vector.

3. The method according to claim 1, characterized in that, Based on the historical performance of the anomaly detection models and the anomaly detection results, the model score of each anomaly detection model is determined, including: The model score of the anomaly detection model is calculated based on its historical accuracy, confidence level of the anomaly detection results, variance of the anomaly detection results, and model complexity.

4. The method according to claim 3, characterized in that, The target anomaly detection result is determined from each of the anomaly detection results based on the model score, including: The anomaly detection result corresponding to the anomaly detection model with the highest model score is identified as the candidate anomaly detection result; The confidence scores corresponding to each of the aforementioned anomaly detection results are weighted to obtain the group confidence score; If the model score corresponding to the candidate anomaly detection result and the group confidence score meet the preset conditions, then the candidate anomaly detection result is determined as the target anomaly detection result.

5. The method according to claim 4, characterized in that, After determining the target anomaly detection result based on the model score, the method further includes: The anomaly detection model is updated based on its model score.

6. The method according to claim 1, characterized in that, The method further includes: Based on the target anomaly detection results, and combined with the temporal and spatial information of the multi-source operating status data, the transmission line is subjected to fault location and propagation path analysis to generate fault source tracing results.

7. An anomaly detection device for a power transmission line, characterized in that, include: The acquisition module is used to acquire multi-source operating status data of key nodes in the transmission line; The fusion module is used to perform multi-source fusion processing on the multi-source operating status data to generate a comprehensive feature vector; The detection module is used to perform anomaly analysis on the comprehensive feature vector using at least two preset anomaly detection models, and generate corresponding anomaly detection results. The determination module is used to determine the model score of each anomaly detection model based on the historical performance of the anomaly detection model and the anomaly detection results, and to determine the target anomaly detection result from each anomaly detection result according to the model score.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform an anomaly detection method for a transmission line according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the abnormal detection method for a power transmission line according to any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements a method for detecting anomalies in a power transmission line according to any one of claims 1-6.

Images