Lead galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion

Abstract

The invention relates to the field of power transmission line operation monitoring, in particular to a lead galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion, which comprises a data acquisition and preprocessing module, a data fusion and three-dimensional reconstruction module, an edge intelligent identification module and a cloud decision and collaborative management module, beidou RTK positioning data, micro-meteorological data and lead image data are synchronously obtained through an integrated monitoring device installed on the unmanned aerial vehicle in a live-line mode, and the positioning data are filtered and corrected; dynamically reconstructing a three-dimensional space form curve of the wire based on multi-source data fusion and physical parameters of the wire; performing time-frequency analysis on the lead motion characteristics, and identifying specific working condition types and severity levels; and an early warning decision instruction and a structured diagnosis report are generated at the cloud in combination with historical cases, operation rules, weather and power grid operation data. The method improves the monitoring precision and real-time performance of the operation state of the conductor and the perspectiveness of operation and maintenance decision, and is suitable for safe operation management of the power transmission line.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of transmission line operation monitoring, and in particular to a conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion. BACKGROUND

[0002] As an important infrastructure of the power system, the transmission line is long-term operated in a complex and changeable natural environment, and is affected by many factors such as wind field change, temperature fluctuation, rainfall and icing. During the operation process, the conductor is prone to galloping, wind deviation, sag change and icing accumulation and other dynamic behaviors. Such conductor motion state not only causes the reduction of the distance between phases, the aggravation of the fittings fatigue, but also may induce serious power grid accidents such as tripping and conductor breakage. Therefore, continuous, accurate and real-time monitoring and evaluation of the conductor motion state is an important technical means to ensure the safe operation of the transmission line. With the development of Beidou high-precision positioning, multi-source sensing, edge computing and cloud computing technologies, the conductor state perception and intelligent analysis based on multi-source data fusion have gradually become an important development direction in the field of transmission line monitoring, and provide a technical basis for realizing the fine perception, intelligent identification and active early warning of the conductor motion state.

[0003] The existing conductor monitoring technology mainly relies on a single sensing method or local measurement information, and it is difficult to simultaneously consider the positioning accuracy, time continuity and environmental adaptability. Especially under strong wind and complex weather conditions, the positioning data is easily disturbed by noise, and it is difficult to stably reflect the real spatial form of the conductor. At the same time, some schemes only perform simple frequency domain analysis on the conductor vibration signal, and lack the three-dimensional space reconstruction capability combined with the physical characteristics of the conductor and the environmental load, which leads to limited identification accuracy of the galloping, wind deviation and icing conditions. In addition, the existing system mainly uploads data to the cloud for processing, which lacks real-time performance, and does not fully consider the evolution trend of the working condition and the spatial correlation between multiple towers, which makes it difficult to form a hierarchical early warning and collaborative disposal mechanism for operation and decision-making, and cannot meet the demand of modern power grid for intelligent and fine management of the safe operation of the transmission line. SUMMARY

[0004] The present application provides a conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion, which is used to solve the problems of insufficient positioning accuracy, discontinuous description of the three-dimensional spatial form of the conductor under complex weather conditions, lagging working condition type identification and lack of collaborative analysis in the existing conductor operation monitoring.

[0005] The conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion comprises a data acquisition and preprocessing module, a data fusion and three-dimensional reconstruction module, an edge intelligent identification module and a cloud decision and collaborative management module, wherein: The data acquisition and preprocessing module is configured to synchronously acquire Beidou RTK preliminary positioning data stream and multi-source auxiliary data set containing micro-weather data and conductor image through at least one integrated monitoring device mounted on the conductor of the target power transmission line segment by the unmanned aerial vehicle, and to perform real-time filtering and outlier rejection on the preliminary positioning data stream to generate clean high-precision Beidou positioning data sequence. The data fusion and three-dimensional reconstruction module is configured to receive the high-precision Beidou positioning data sequence and the multi-source auxiliary data set from the data acquisition and preprocessing module, perform time-space alignment and fusion thereon, and dynamically reconstruct a three-dimensional spatial form curve of the conductor segment on a continuous time scale based on the fused data and conductor physical parameters. The edge intelligent identification module is deployed on an edge computing node close to the integrated monitoring device, configured to receive the three-dimensional spatial form curve from the data fusion and three-dimensional reconstruction module, perform time-frequency domain analysis thereon to extract a feature vector, and match the feature vector with a pre-set typical working condition feature model to intelligently identify a specific working condition type of the conductor and a severity level thereof. The cloud-based decision-making and collaborative management module is configured to receive the specific working condition type, the severity level thereof and related feature data from the edge intelligent identification module, perform deep analysis and trend prediction based on cloud-based big data and a rule base to generate a final early warning decision-making instruction and a diagnosis report, and feed back the early warning decision-making instruction, the diagnosis report or a control instruction to a downstream system and the integrated monitoring device, while realizing visual display of the monitoring results.

[0006] Optionally, the data acquisition and preprocessing module comprises: The integrated monitoring device mounted on the conductor of the target power transmission line segment by the unmanned aerial vehicle synchronously triggers the Beidou RTK module, the environmental sensor and the image acquisition module therein to perform data acquisition, so as to synchronously obtain original Beidou RTK preliminary positioning data stream and original monitoring information containing micro-weather data and conductor image data; The obtained Beidou RTK preliminary positioning data stream and the original monitoring information containing micro-weather data and conductor image data are strictly aligned and bound in time stamp and spatial coordinate system to form a time-space unified multi-source auxiliary data set; The Beidou RTK preliminary positioning data stream contained in the multi-source auxiliary data set after alignment and binding is subjected to real-time filtering and gross error rejection processing by using an adaptive filtering algorithm to eliminate signal interference and jump, so as to finally generate clean and continuous high-precision Beidou positioning data sequence.

[0007] Optionally, the strict alignment and binding of the time stamp and the spatial coordinate system is specifically to uniformly align the Beidou RTK preliminary positioning data stream and the original monitoring information containing microclimate data and conductor image data to the same reference time axis based on the time stamp generated thereby, and to associate to the unique spatial identifier of the integrated monitoring device.

[0008] Optionally, the real-time filtering and gross error elimination processing using an adaptive filtering algorithm is specifically to use an adaptive Kalman filtering algorithm and fuse the differential correction data from the reference station to process the Beidou RTK preliminary positioning data stream.

[0009] Optionally, the data fusion and three-dimensional reconstruction module comprises: Receiving the high-precision Beidou positioning data sequence and the multi-source auxiliary data set, associating each positioning point in the high-precision Beidou positioning data sequence with the microclimate data in the multi-source auxiliary data set with the corresponding time stamp, and extracting key visual features in the conductor image data to form a fusion data set that is unified in time and space and enhanced in features; Based on the fusion data set, and calling the pre-stored conductor physical parameters, a conductor spatial form dynamics model combining static catenary constraint and dynamic wind load response is constructed; Taking the fusion data set as input, the conductor spatial form dynamics model is driven to iteratively solve and update the state, and the spatial coordinate set of the monitoring points and the interpolation points on the conductor segment at different times is continuously output, so as to dynamically reconstruct the three-dimensional spatial form curve of the conductor segment on the continuous time scale.

[0010] Optionally, the key visual features in the conductor image data are extracted, specifically edge detection algorithm is used to extract conductor edge contour features, and image classification algorithm is used to identify whether there are ice adhesion features in the image.

[0011] Optionally, the edge intelligent recognition module comprises: Receiving the three-dimensional spatial form curve, performing joint analysis of the three-dimensional spatial form curve in time domain and frequency domain on the edge computing node, and extracting a feature vector containing amplitude, dominant frequency, vibration mode order and motion trajectory ellipse parameter; Performing multi-dimensional similarity matching calculation on the extracted feature vector and the typical working condition feature model pre-stored in the local knowledge base of the edge computing node; According to the multi-dimensional similarity matching calculation result, the specific working condition type and the severity level that can best represent the current conductor motion state are determined and output.

[0012] Optionally, the multi-dimensional similarity matching calculation is specifically calculating the Euclidean distance and dynamic time warping distance between the feature vector and each model feature vector in the typical working condition feature model, and fusing the two distance measurement results by weighting.

[0013] Optionally, the cloud decision and collaborative management module comprises: receiving the specific working condition type and its severity level, and synchronously converging the feature vector, three-dimensional space form curve abstract data and external associated data from meteorology and power grid operation related to the specific working condition type and its severity level to form a cloud data set to be analyzed; Based on the cloud data set, the historical case library and operation rule library in the cloud big data platform are called to perform time series-based working condition evolution trend prediction, spatial correlation analysis of multi-base tower monitoring data and comprehensive risk assessment, thereby generating a preliminary decision scheme including cause analysis, development trend and disposal suggestion; The generated preliminary decision scheme is formatted, packaged and audited to generate a final early warning decision instruction and a structured diagnostic report for execution, and the early warning decision instruction and the diagnostic report are pushed to the relevant operation and maintenance personnel terminal and downstream control system, while the early warning information and conductor state display on the visualization platform are updated.

[0014] Optionally, the time series-based working condition evolution trend prediction is specifically modeling and predicting the history and current sequence of the specific working condition type and its severity level in the cloud data set by using a long short-term memory neural network model.

[0015] The beneficial effects of the present application are: 1. The present application synchronously acquires Beidou RTK preliminary positioning data stream, microclimate data and conductor image data through the integrated monitoring device installed by the unmanned aerial vehicle, and introduces an adaptive Kalman filtering algorithm based on real-time wind speed dynamic adjustment process noise parameter in the data acquisition and preprocessing module to suppress environmental interference and measurement jump at the source level, thereby generating clean and continuous high-precision Beidou positioning data sequence;On this basis, the data fusion and three-dimensional reconstruction module strictly binds the high-precision Beidou positioning data sequence and the multi-source auxiliary data set with time stamp and space identifier, and combines the conductor physical parameters, static catenary constraint and dynamic wind load response to construct a conductor space form dynamics model, and realizes continuous state update through an extended Kalman filtering algorithm, thereby dynamically reconstructing the three-dimensional space form curve of the conductor on a continuous time scale, so that the complex motion state of the conductor galloping, wind deviation and icing can be stably and finely described, and the credibility and engineering applicability of the monitoring result are significantly improved.

[0016] 2.The application, edge intelligent recognition module is arranged on the edge side, the three-dimensional space form curve is directly processed on the edge computing node close to the integrated monitoring device, key features such as amplitude, dominant frequency, vibration modal order and motion trajectory ellipse parameters are extracted through time domain and frequency domain joint analysis, and multi-dimensional similarity matching of Euclidean distance and dynamic time warping distance is carried out with the preset steady-state sag working condition characteristic model, the dancing working condition characteristic model, the wind deflection working condition characteristic model and the icing working condition characteristic model in the local knowledge base, so that the specific working condition type is quickly determined without relying on cloud computing; at the same time, the identification result is classified into light, moderate and severe severity levels based on the preset threshold system, so that the conductor anomaly is upgraded from 'whether abnormal' to 'abnormal type and risk level' fine identification, further meeting the actual needs of the operation state classification control of the transmission line.

[0017] 3.The application, through the cloud decision and collaborative management module, the specific working condition type and its severity level output on the edge side are unified with the feature vector, three-dimensional space form curve abstract data, and future 72-hour weather warning and forecasting data and real-time power grid load data, forming a structured cloud data set, and using the historical case library, operation rule library, long short-term memory neural network model and spatial correlation analysis method based on graph theory, the working condition evolution trend prediction, multi-pole tower disturbance propagation analysis and comprehensive risk assessment are realized, and finally the warning decision instruction and structured diagnosis report containing cause analysis, development trend and disposal suggestion are generated; by synchronously pushing to the operation and maintenance personnel terminal, downstream control system and visualization platform, the early warning and collaborative disposal of conductor dancing and icing and other risks are realized, so as to reduce the risk of sudden failure and improve the safety, controllability and intelligent level of the transmission line operation. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only illustrate the application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0019] Fig. 1 It is a system flowchart of the embodiment of the application. Fig. 2 It is a flowchart of the edge intelligent recognition module of the embodiment of the application. DETAILED DESCRIPTION

[0020] The application will be described in detail below with reference to the drawings and specific embodiments. Meanwhile, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best, preferred embodiments, and other alternative ways can also be implemented by those skilled in the art for some known technologies; and the drawings are only used to describe the embodiments more specifically, and are not intended to specifically limit the application.

[0021] It should be noted that in the specification, "one embodiment", "embodiment", "exemplary embodiment", "some embodiments" and the like indicate that the described embodiment can include a specific feature, structure or property, but not necessarily every embodiment includes the specific feature, structure or property. In addition, when a specific feature, structure or property is described in combination with an embodiment, it should be within the knowledge of those skilled in the related art to realize such a feature, structure or property in combination with other embodiments (whether or not explicitly described).

[0022] Generally, the terms can be understood at least in part from the use in context. For example, depending at least in part on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular or can be used to describe combinations of features, structures, or characteristics that are combinable into one or more instances. In addition, the term "based on" can be understood as not necessarily intended to convey a set of exclusive factors, but can instead, depending at least in part on the context, allow for the existence of other factors that are not necessarily explicitly described.

[0023] As shown in Figs. 1-2 The conductor dancing real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion includes a data acquisition and preprocessing module, a data fusion and three-dimensional reconstruction module, an edge intelligent identification module, and a cloud decision and collaborative management module, wherein The data acquisition and preprocessing module is used to synchronously obtain Beidou RTK preliminary positioning data stream and multi-source auxiliary data set containing micro-meteorological data and conductor image through at least one integrated monitoring device installed on the conductor of the target transmission line section by the unmanned aerial vehicle, and to perform real-time filtering and outlier rejection on the preliminary positioning data stream to generate clean high-precision Beidou positioning data sequence. The specific steps are as follows: The integrated monitoring device installed on the conductor of the target transmission line section by the unmanned aerial vehicle synchronously triggers the internal Beidou RTK module and the environmental sensor and image acquisition module to perform data acquisition, thereby synchronously obtaining the original Beidou RTK preliminary positioning data stream and the original monitoring information containing micro-meteorological data and conductor image data. Specifically, The integrated monitoring device is fixed and installed on the conductor of the target power transmission line segment by using the unmanned aerial vehicle live working platform. After completing mechanical fixation and electrical insulation verification, the integrated monitoring device is issued with a preset periodic wake-up instruction or a remote wake-up instruction from the reference station, so that the integrated monitoring device is switched from a low-power standby state to a working state. At the same time, the control unit in the integrated monitoring device sends a synchronous acquisition instruction to the Beidou RTK module, the environmental sensor and the image acquisition module. The Beidou RTK module continuously outputs a Beidou RTK preliminary positioning data stream containing carrier phase information and pseudo-range information at a preset sampling frequency. The environmental sensor synchronously acquires and outputs microclimate data reflecting the environmental state around the conductor. The image acquisition module images the appearance of the conductor within the same acquisition period and outputs corresponding conductor image data, so as to form original monitoring information under the same time reference.

[0024] 2. The obtained Beidou RTK preliminary positioning data stream and the original monitoring information containing microclimate data and conductor image data are strictly aligned and bound in time stamp and spatial coordinate system to form a unified time-space multi-source auxiliary data set, specifically: The time management unit of the integrated monitoring device adds high-precision time stamps corresponding to the acquisition time to the Beidou RTK preliminary positioning data stream, the microclimate data and the conductor image data respectively, and maps all the time stamps to the same reference time axis. At the same time, the unique space identifier pre-configured by the integrated monitoring device is called to establish a one-to-one correspondence between the space identifier and the spatial coordinate information of the target power transmission line segment. In the data arrangement process, the Beidou RTK preliminary positioning data stream, the microclimate data and the conductor image data are matched in the order of time stamps, and the matched data records are bound to the unique space identifier, so as to generate a multi-source auxiliary data set consistent in time dimension and spatial coordinate system.

[0025] 3. The Beidou RTK preliminary positioning data stream contained in the multi-source auxiliary data set after alignment and binding is processed by using an adaptive filtering algorithm for real-time filtering and gross error elimination to eliminate signal interference and jump, and finally a clean and continuous high-precision Beidou positioning data sequence is generated, specifically: An adaptive Kalman filtering algorithm is called in the edge computing unit, a state space model taking the spatial position change of the conductor as the state quantity is established for the Beidou RTK preliminary positioning data stream, and the differential correction data from the reference station is introduced into the filtering process as the observation correction quantity; in each filtering iteration period, first, the state prediction is performed according to the state estimation result at the last moment, and then the state update is performed in combination with the Beidou RTK observation value and the differential correction data at the current moment; at the same time, the real-time wind speed information in the micro-meteorological data corresponding to the time stamp in the multi-source auxiliary data set is read, and the process noise parameter in the adaptive Kalman filtering algorithm is dynamically adjusted according to the real-time wind speed information, so that the filtering model can match the motion characteristics of the conductor under different wind load conditions; after the state update is completed, the threshold value of the filtering residual is determined, the data exceeding the preset residual threshold value is determined as a gross error and is removed, and the continuous and physically reasonable positioning result is reserved, and finally the clean and continuous high-precision Beidou positioning data sequence is output.

[0026] The data fusion and three-dimensional reconstruction module is used for receiving the high-precision Beidou positioning data sequence and the multi-source auxiliary data set from the data acquisition and preprocessing module, performing spatio-temporal alignment and fusion, and dynamically reconstructing the three-dimensional spatial form curve of the conductor segment on the continuous time scale based on the fusion data and the physical parameters of the conductor, and the specific steps are as follows: 1. receiving the high-precision Beidou positioning data sequence and the multi-source auxiliary data set, associating each positioning point in the high-precision Beidou positioning data sequence with the micro-meteorological data in the multi-source auxiliary data set corresponding to the time stamp, and extracting the key visual features in the conductor image data to form a fusion data set that is unified in time and space and enhanced in features, specifically: The data fusion processing unit receives the high-precision Beidou positioning data sequence and reads each positioning point in it in chronological order; for each positioning point, the micro-meteorological data record with the same time stamp is searched and matched in the multi-source auxiliary data set according to the time stamp information carried by the positioning point, and the micro-meteorological data record is associated with the corresponding positioning point one by one; At the same time, the image processing submodule is called to perform preprocessing operation on the conductor image data corresponding to the time stamp in the multi-source auxiliary data set, and the preprocessing operation includes image graying and noise suppression to ensure the stability of subsequent feature extraction; After the preprocessing is completed, the edge detection algorithm is used to operate on the conductor image data, extract the conductor edge contour feature representing the spatial profile of the conductor, and convert the conductor edge contour feature into quantifiable visual feature parameters; Subsequently, the image classification algorithm is used to calculate the identification result of whether there is ice adhesion feature in the image; Finally, the positioning point information in the high-precision Beidou positioning data sequence, the micro-meteorological data associated therewith, the extracted conductor edge contour feature, and the icing attachment feature recognition result are uniformly packaged to form a unified spatiotemporal and feature-enhanced fusion data set consistent in time stamp and spatial identifier, denoted as: . wherein, denotes a piece of fusion record in the fusion data set at time on the reference time axis, denotes the positioning point coordinate vector of the high-precision Beidou positioning data sequence at time , expressed as in the spatial coordinate system adopted by the system, denotes the micro-meteorological data vector in the multi-source auxiliary data set aligned with the time stamp , defined as , wherein is the real-time wind speed, is the wind direction, denotes the conductor edge contour feature vector extracted from the conductor image data, obtained by using an edge detection algorithm, defined as , wherein is the conductor edge curvature feature scalar, is the conductor edge connected length feature scalar, denotes the icing attachment feature recognition result obtained by using an image classification algorithm, with the value defined as , wherein 1 indicates the presence of the icing attachment feature, and 0 indicates the absence of the icing attachment feature, and SID denotes the unique spatial identifier of the integrated monitoring device, used to bind the multi-source information at the same time stamp to the same spatial entity.

[0027] 2. Based on the fusion data set, and by calling the pre-stored conductor physical parameters, a conductor spatial form dynamics model combining static catenary constraint and dynamic wind load response is constructed, specifically: In the model construction unit, first, the conductor physical parameters corresponding to the target power transmission line segment are called from the system parameter library, including the conductor span, the height difference between the two suspension points, the conductor unit length mass, the conductor cross-sectional area, and the rated operating tension; Based on the conductor physical parameters, a static catenary constraint model of the conductor under the condition of no external disturbance is established according to the catenary theory, used to describe the reference spatial form of the conductor under the action of gravity; Subsequently, real-time wind speed and wind direction data in the microclimate data corresponding to the moment are read from the fusion data set, and real-time wind speed and wind direction data are converted into wind pressure load acting on the unit length of the conductor according to the aerodynamic formula, forming a dynamic wind load response term; The static catenary constraint model is uniformly modeled with the dynamic wind load response term, and a conductor spatial form dynamics model capable of reflecting the inherent geometric constraints of the conductor and the dynamic response characteristics under the action of wind load is constructed, which is represented as: ; Wherein, represents the air density parameter, which is preset as a constant by the system and remains unchanged during operation, represents the wind pressure load acting on the unit length of the conductor, which is used to represent the dynamic wind load response, represents the air density parameter, which is preset as a constant by the system and remains unchanged during operation, represents the real-time wind speed in the multi-source auxiliary data set, which is taken from the microclimate data, represents the air dynamic resistance coefficient parameter of the conductor, which is preset as a constant by the system and corresponds to the shape of the conductor, represents the equivalent outer diameter parameter of the conductor, which is preset by the system and corresponds to the physical parameters of the conductor.

[0028] Description: The above As a wind load input term in the conductor spatial form dynamics model, it jointly constrains the state evolution of the conductor spatial form with the static catenary constraint, and the wind direction is used to determine the direction of action and map the load component to the spatial coordinate system.

[0029] 3. The fusion data set is taken as input to drive the conductor spatial form dynamics model to iteratively solve and update the state, and the spatial coordinate set of the monitoring point and the interpolation point on the conductor segment at different moments is continuously output, so as to dynamically reconstruct the three-dimensional spatial form curve of the conductor segment on the continuous time scale, which is specifically: In the state estimation unit, the fusion data set is taken as the observation input, the conductor spatial form dynamics model is introduced as the state transition model, and the extended Kalman filter algorithm is used to iteratively solve and update the state of the conductor space; In each iteration period, first, the state prediction calculation is performed based on the conductor spatial state estimation result at the last moment, and then the predicted state is updated in combination with the positioning observation value in the high-precision Beidou positioning data sequence and the conductor edge contour feature in the fusion data set; ​In the iteration process of the extended Kalman filtering algorithm, the real-time wind speed and wind direction data in the microclimate data in the fusion data set are read in real time, and the process noise matrix in the extended Kalman filtering algorithm is dynamically adjusted according to the real-time wind speed and wind direction data, so that the state estimation process can reflect the change of the conductor motion uncertainty under different wind field conditions. After completing the state update, the spatial coordinates of each monitoring point on the conductor segment at the current time are output, and spatial interpolation calculation is performed based on the spatial coordinate results of adjacent monitoring points to obtain the spatial coordinate set of the interpolation point at the same time. The extended Kalman filtering algorithm iteration solution and state update and output three-dimensional space form curve, expressed as: , wherein, represents the process noise matrix of the extended Kalman filtering algorithm at time , for realizing "dynamically adjusting the process noise matrix by using real-time wind speed and wind direction data", represents the process noise matrix constant under the no-wind reference condition, which is obtained by offline calibration of the system and pre-stored, represents the wind speed sensitive coefficient constant, which is used to quantify the amplification effect of real-time wind speed on process uncertainty, and is obtained by offline calibration of the system and pre-stored, represents the real-time wind speed in the multi-source auxiliary data set, represents the wind direction sensitive coefficient constant, which is used to quantify the amplification effect of wind direction deflection on process uncertainty, and is obtained by offline calibration of the system and pre-stored, represents the wind direction data in the multi-source auxiliary data set, represents the reference direction angle constant of the conductor axial direction in the spatial coordinate system, which is determined by the line topology and installation attitude and pre-stored.

[0030] Description: is used for the prediction-update iteration of the extended Kalman filtering algorithm, so that the conductor space form dynamics model can maintain stable state update under strong wind and variable wind direction conditions, thereby continuously outputting the monitoring point and interpolation point spatial coordinate set, and further forming a three-dimensional space form curve.

[0031] By repeatedly executing the above iteration solution and output process at different times, a three-dimensional space form curve of the conductor segment on a continuous time scale is continuously generated.

[0032] The edge intelligent identification module is deployed on the edge computing node close to the integrated monitoring device, which is used to receive the three-dimensional space form curve from the data fusion and three-dimensional reconstruction module, perform time-frequency domain analysis to extract feature vectors, and match the feature vectors with the pre-set typical working condition feature model, intelligently identify the specific working condition type and severity level of the current conductor, and the specific steps are as follows:​ 1. Receive the 3D spatial morphological curve, perform joint time-domain and frequency-domain analysis on the 3D spatial morphological curve at the edge computing nodes, and extract feature vectors containing amplitude, dominant frequency, vibration mode order, and motion trajectory ellipse parameters, specifically: The signal analysis unit of the edge computing node receives the three-dimensional spatial morphology curve and performs discrete sampling and reconstruction of the three-dimensional spatial morphology curve in time sequence to obtain the three-dimensional coordinate time sequence data of each monitoring point at continuous sampling time. Based on three-dimensional coordinate time-series data, the displacement changes of the conductor in each spatial direction are calculated during the time-domain analysis, and the amplitude parameters of the conductor motion are determined by the difference between the maximum and minimum displacement changes. During the frequency-domain analysis, a fast Fourier transform is performed on the displacement time-series signal corresponding to the three-dimensional spatial shape curve to obtain the frequency-domain energy distribution spectrum, and the dominant frequency of the conductor motion is determined based on the frequency corresponding to the energy peak in the frequency-domain energy distribution spectrum. Based on this, by sorting and grouping multiple significant frequency components in the frequency domain energy distribution spectrum, the vibration mode order involved in the motion during the conductor vibration process is determined; Meanwhile, based on the displacement trajectory of the conductor in different spatial directions within the same sampling time window, a two-dimensional projection operation is performed on the three-dimensional spatial shape curve, and the least squares method is used to fit the ellipse equation of the motion trajectory, and the ellipse parameters of the motion trajectory are extracted from the ellipse equation. Finally, the amplitude, dominant frequency, vibration mode order, and motion trajectory ellipse parameters are combined in a predefined order to form a feature vector that characterizes the motion state of the conductor.

[0033] 2. Perform multi-dimensional similarity matching calculations between the extracted feature vectors and the typical working condition feature models pre-installed in the local knowledge base of the edge computing node. Specifically: The matching calculation unit of the edge computing node reads the preset typical working condition feature model from the local knowledge base. The typical working condition feature model includes the steady-state sag working condition feature model, the galloping working condition feature model, the wind deflection working condition feature model and the icing working condition feature model. For each typical working condition feature model, its corresponding model feature vector is read; the Euclidean distance between the feature vector and each model feature vector is calculated to obtain the first distance metric result that reflects the difference in feature amplitude. Simultaneously, dynamic time warping distance calculation is performed on the time series of feature vectors and feature vectors of each model to obtain a second distance metric that reflects the differences in time evolution. After calculating the two distance metrics, the Euclidean distance and the dynamic time warp distance are weighted and fused according to preset weighting coefficients to generate a comprehensive similarity score for each typical working condition feature model, expressed as: ; in, This represents the comprehensive similarity score calculated for a typical working condition feature model, used to characterize the degree of matching between the feature vector and the model. This represents the preset weighting coefficients, which satisfy... This is used to perform a weighted fusion of the contributions from Euclidean distance and dynamic time-warped distance. The Euclidean distance is calculated dimension-by-dimensionally from the eigenvectors and the corresponding model eigenvectors in the typical working condition characteristic model, and is used to characterize the difference in feature magnitude. The dynamic time-warped distance is calculated by expanding the time series of the feature vectors and model feature vectors using dynamic time warping, and is used to characterize the differences in temporal evolution. , This represents a normalized mapping term that converts the distance metric into a similarity component; the smaller the distance, the higher the similarity.

[0034] During implementation, a characteristic model for steady-state sag, galloping, wind deflection, and icing conditions was calculated for each condition. .

[0035] 3. Based on the multi-dimensional similarity matching calculation results, determine and output the specific working condition type and its severity level that best characterizes the current conductor motion state, specifically: The judgment unit compares and analyzes the comprehensive similarity scores, and selects the typical working condition feature model with the best comprehensive similarity score that exceeds the corresponding preset threshold as the specific working condition type corresponding to the current conductor motion state. After determining the specific working condition type, the comprehensive similarity score is compared with the set of grading thresholds corresponding to that working condition type. The severity level of the specific working condition type is determined according to the score range, and the severity level is divided into mild, moderate or severe. ; in, Indicates the severity level, with values ​​limited to mild, moderate, and severe. This represents the overall similarity score for the corresponding specific working condition type. For this specific working condition type, preset thresholds at various levels are required to meet the following conditions. It is pre-configured by the local knowledge base of the edge computing node and used for hierarchical determination.

[0036] When the specific working condition is determined to be icing, the edge computing node generates an image acquisition command and sends the image acquisition command to the integrated monitoring device to trigger the integrated monitoring device to perform high-definition image acquisition operation. Finally, the specific working condition type, severity level, and corresponding recognition results are stored and reported as edge intelligent recognition output results.

[0037] The cloud-based decision-making and collaborative management module receives specific operating condition types, severity levels, and related characteristic data from the edge intelligent identification module. Based on cloud-based big data and rule bases, it performs in-depth analysis and trend prediction to generate final early warning decision instructions and diagnostic reports. These instructions, reports, or control commands are then fed back to downstream systems and integrated monitoring devices. Simultaneously, the monitoring results are visualized. The specific steps are as follows: 1. Receive specific operating condition types and their severity levels, and simultaneously aggregate feature vectors, 3D spatial morphology curve summary data associated with specific operating condition types and their severity levels, as well as external correlation data from meteorology and power grid operation, to form a cloud dataset to be analyzed, specifically: The cloud data access unit receives the specific working condition type and its severity level uploaded by the edge intelligent recognition module, and uses the judgment result as the main index identifier. At the same time, in accordance with the unified data identification rules, the feature vector and three-dimensional spatial morphology curve summary data associated with the main index identifier are received synchronously. The three-dimensional spatial morphology curve summary data is the structured description data obtained by time window compression and key point extraction of the complete three-dimensional spatial morphology curve. After completing the internal data access, the cloud data access unit further obtains the next 72 hours of weather warning and forecast data from the meteorological system and the real-time power grid load data of the target transmission line segment from the power grid operation system from the external data interface; Subsequently, the above data was checked and rearranged according to timestamp, spatial identifier and transmission corridor number, and the data that passed the check was packaged into a cloud dataset for subsequent analysis.

[0038] 2. Based on cloud-based datasets, the system utilizes historical case libraries and operational rule libraries from the cloud-based big data platform to perform time-series-based prediction of operational condition evolution trends, spatial correlation analysis of multi-base tower monitoring data, and comprehensive risk assessment. This generates a preliminary decision-making plan that includes cause analysis, development trends, and remedial recommendations. Specifically: In the cloud analysis unit, the historical case library in the cloud big data platform is first called to retrieve historical work condition records that match the specific work condition type and its severity level, and the corresponding evolution path and handling results are extracted as reference samples. Subsequently, a long short-term memory neural network model was used to model the specific working condition types and their severity levels arranged in chronological order in the cloud dataset. By taking the joint input of the historical sequence and the current sequence, the evolution trend of the working condition within the future time window was predicted. After completing the time series prediction, the cloud analysis unit further constructs a spatial relationship graph of multiple towers in the same transmission corridor based on graph theory algorithm. The specific working condition type and severity level uploaded by each tower are used as node attributes. By analyzing the propagation direction and intensity between nodes, the spatial propagation relationship of the working condition is identified, thereby locating the disturbance source or identifying the weak section. Based on this, and combined with the pre-set safe operation rules in the operation rule base, the prediction results and spatial correlation analysis results are comprehensively evaluated to form a preliminary decision-making plan that includes analysis of the causes of operating conditions, judgment of future development trends, and targeted operation and maintenance suggestions.

[0039] 3. The generated preliminary decision-making scheme is formatted, packaged, and reviewed to generate final early warning decision instructions and structured diagnostic reports that can be issued and executed. These instructions and reports are then pushed to relevant maintenance personnel terminals and downstream control systems. Simultaneously, the early warning information and conductor status display on the visualization platform are updated. Specifically: The cloud-based decision output unit organizes the preliminary decision plan in a structured manner, and formats and encapsulates the working condition judgment results, evolution trend prediction conclusions, related risk analysis and operation and maintenance handling strategy suggestions according to a predefined data template; After encapsulation, an automatic review process based on a rule base is executed to verify the completeness and consistency of the preliminary decision-making scheme. Once the review is passed, a final early warning decision instruction and a structured diagnostic report are generated that can be issued and executed. Subsequently, through the cloud message distribution mechanism, the final early warning decision instructions and diagnostic reports are pushed to the terminals of relevant operation and maintenance personnel, and simultaneously sent to the downstream control system to support coordinated handling; At the same time, the visualization service interface is called to update the early warning information and conductor status display on the visualization platform, so that operation and maintenance personnel can grasp the current working status and its development trend in real time.

[0040] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0041] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A real-time monitoring system for conductor galloping based on Beidou high-precision positioning and multi-source data fusion, characterized in that, The system comprises a data acquisition and preprocessing module, a data fusion and three-dimensional reconstruction module, an edge intelligent identification module, and a cloud decision and collaborative management module. The data acquisition and preprocessing module is configured to acquire BeiDou RTK preliminary positioning data stream and multi-source auxiliary data set including microclimate data and conductor image through at least one integrated monitoring device installed on the conductor of the target transmission line segment by the unmanned aerial vehicle, and to filter and remove outliers of the preliminary positioning data stream in real time to generate a BeiDou positioning data sequence. The data fusion and three-dimensional reconstruction module is configured to receive the BeiDou positioning data sequence and the multi-source auxiliary data set from the data acquisition and preprocessing module, to perform time-space alignment and fusion on the data, and to dynamically reconstruct a three-dimensional spatial form curve of the conductor segment on a continuous time scale based on the fused data and physical parameters of the conductor. The edge intelligent identification module is deployed on an edge computing node close to the integrated monitoring device, configured to receive the three-dimensional spatial form curve from the data fusion and three-dimensional reconstruction module, to perform time-frequency domain analysis on the curve to extract a feature vector, and to match the feature vector with a preset typical working condition feature model to intelligently identify a specific working condition type and a severity level of the conductor. The cloud decision and collaborative management module is configured to receive the specific working condition type, the severity level, and related feature data from the edge intelligent identification module, to perform deep analysis and trend prediction based on cloud big data and a rule base, to generate a final early warning decision instruction and a diagnosis report, and to feed back the early warning decision instruction, the diagnosis report, or a control instruction to a downstream system and the integrated monitoring device, and to realize visual display of monitoring results.

2. The conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion according to claim 1, characterized in that, The data acquisition and preprocessing module comprises: The integrated monitoring device installed on the conductor of the target transmission line segment by the unmanned aerial vehicle is configured to trigger the BeiDou RTK module, the environmental sensor, and the image acquisition module inside the device to perform data acquisition simultaneously, so as to obtain original BeiDou RTK preliminary positioning data stream and original monitoring information including microclimate data and conductor image data. The obtained BeiDou RTK preliminary positioning data stream and the original monitoring information including microclimate data and conductor image data are strictly aligned and bound in time stamp and spatial coordinate system to form a time-space unified multi-source auxiliary data set. The BeiDou RTK preliminary positioning data stream included in the multi-source auxiliary data set after alignment and binding is processed by an adaptive filtering algorithm for real-time filtering and outlier removal to eliminate signal interference and jumps, and finally a BeiDou positioning data sequence is generated. 3.The conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion according to claim 2, characterized in that, The strict alignment and binding in time stamp and spatial coordinate system are specifically to align the BeiDou RTK preliminary positioning data stream and the original monitoring information including microclimate data and conductor image data to the same reference time axis based on the time stamps generated by the data, and to associate the data to the unique spatial identifier of the integrated monitoring device.

4. The conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion according to claim 2, characterized in that, The real-time filtering and gross error elimination processing is performed by using an adaptive filtering algorithm, specifically, an adaptive Kalman filtering algorithm, and differential correction data from a reference station is fused to process the Beidou RTK preliminary positioning data stream.

5. The conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion according to claim 2, characterized in that, The data fusion and three-dimensional reconstruction module comprises: The Beidou positioning data sequence and the multi-source auxiliary data set are received, each positioning point in the high-precision Beidou positioning data sequence is associated with the micro-meteorological data of the corresponding timestamp in the multi-source auxiliary data set, and visual features in the conductor image data are extracted to form a fusion data set that is unified in space and time and enhanced in features; Based on the fusion data set and by calling the pre-stored conductor physical parameters, a conductor space form dynamics model combining static catenary constraint and dynamic wind load response is constructed; The fusion data set is taken as input to drive the conductor space form dynamics model to iteratively solve and update the state, and a set of spatial coordinates of the monitoring points and interpolation points on the conductor segment at different times is continuously output, so that a three-dimensional space form curve of the conductor segment on a continuous time scale is dynamically reconstructed.

6. The conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion according to claim 5, characterized in that, The visual features in the conductor image data are extracted, specifically, edge detection algorithm is used to extract conductor edge contour features, and image classification algorithm is used to identify whether there are ice adhesion features in the image.

7. The conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion according to claim 5, characterized in that, The edge intelligent recognition module comprises: The three-dimensional space form curve is received, and time domain and frequency domain joint analysis is performed on the three-dimensional space form curve on the edge computing node to extract a feature vector containing amplitude, dominant frequency, vibration modal order and motion trajectory ellipse parameters; The extracted feature vector is subjected to multi-dimensional similarity matching calculation with a typical working condition feature model pre-stored in the local knowledge base of the edge computing node; According to the multi-dimensional similarity matching calculation result, the specific working condition type and the severity level of the current conductor motion state are determined and output.

8. The conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion according to claim 7, characterized in that, The multi-dimensional similarity matching calculation is specifically to calculate the Euclidean distance and dynamic time warping distance between the feature vector and each model feature vector in the typical working condition feature model, and to perform weighted fusion on the two distance measurement results.

9. The conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion according to claim 7, characterized in that, The cloud decision and collaborative management module comprises: The specific working condition type and the severity level are received, and the feature vector, three-dimensional space form curve abstract data and external associated data from meteorology and power grid operation related to the specific working condition type and the severity level are synchronously aggregated to form a cloud data set to be analyzed; Based on the cloud data set, the historical case library and operation rule library in the cloud big data platform are called to perform time series-based working condition evolution trend prediction, spatial correlation analysis of multi-pole tower monitoring data and comprehensive risk assessment, so as to generate a preliminary decision scheme containing cause analysis, development trend and disposal suggestion; The generated preliminary decision scheme is formatted, packaged, audited, and finally generates an executable early warning decision instruction and a structured diagnostic report, and the early warning decision instruction and the diagnostic report are pushed to the terminal of the related operation and maintenance personnel and the downstream control system, and the early warning information and the conductor state display on the visualization platform are updated.

10. The conductor galloping real-time monitoring system based on Beidou high-precision positioning and multi-source data fusion according to claim 9, characterized in that, The time series-based working condition evolution trend prediction is specifically modeling and predicting the history and current sequence of the specific working condition type and its severity level in the cloud data center by using a long short-term memory neural network model.

Images