Power transmission line galloping risk early warning and system based on multi-physics field fusion

Abstract

The invention discloses a power transmission line galloping risk early warning and system based on multi-physics field fusion, and the method comprises the steps: firstly obtaining a multi-physics field data set of a power transmission line galloping process, and generating a dynamic three-dimensional point cloud sequence through space-time alignment and three-dimensional attitude reconstruction; galloping energy distribution features, frequency spectrum features and space trajectory features are extracted and fused to obtain a fused feature vector; inputting the vector into a physical constraint enhanced bidirectional long-short-term memory network, predicting a galloping evolution trend and outputting a joint prediction matrix of galloping amplitude and frequency; inputting the matrix into a pre-trained risk propagation reasoning model, and calculating a multi-level risk quantitative index; and finally, generating a risk early warning prompt according to the index and sending the risk early warning prompt to corresponding risk management equipment so as to realize power transmission line galloping risk early warning. The power transmission line galloping evolution trend can be accurately predicted, so that the power transmission line galloping risk can be accurately early warned.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a transmission line galloping risk warning and system based on multi-physical field fusion. Background Art

[0002] During transmission line operation, conductor galloping, a self-excited vibration phenomenon caused by wind excitation, poses a significant threat to the safe and stable operation of the power grid due to its low-frequency and high-amplitude characteristics. Existing technologies typically use single-physics field monitoring methods or simple multi-parameter combined analysis methods for risk assessment. Specifically, traditional technologies rely on limited sensors installed on transmission lines (such as accelerometers, anemometers, and thermometers) to collect local data. Combined with meteorological parameters provided by meteorological stations, these methods analyze the line galloping status using threshold judgments or simple statistical models. Some methods use video monitoring to capture line vibration images and extract galloping features through image processing techniques. Alternatively, traditional machine learning algorithms (such as support vector machines and decision trees) are used to model and analyze historical monitoring data to predict galloping risks. These methods often only consider information in a single temporal or spatial dimension in data processing, lacking in-depth integration and analysis of multi-physics field data. Furthermore, the prediction models are often data-driven and fail to fully consider the physical constraints of transmission line galloping.

[0003] Therefore, the existing technical solutions have the following problems: the lack of in-depth analysis of multi-physical field data leads to inaccurate predictions of the evolution trend of transmission line galloping, and thus makes it impossible to accurately warn of the risk of transmission line galloping. Summary of the Invention

[0004] The embodiment of the present invention provides a transmission line galloping risk warning and system based on multi-physics field fusion, which can accurately predict the evolution trend of transmission line galloping, thereby accurately warning of transmission line galloping risks.

[0005] An embodiment of the present invention provides a transmission line galloping risk warning method based on multi-physics field fusion, comprising: Obtain a multiphysics dataset of the galloping process of a transmission line; Performing spatiotemporal alignment and three-dimensional posture reconstruction on the multi-physics field data set to generate a dynamic three-dimensional point cloud sequence of the transmission line dancing; Extracting dance energy distribution features, frequency spectrum features, and spatial trajectory features from the dynamic three-dimensional point cloud sequence, and fusing the three features to obtain a fused feature vector; Inputting the fused feature vector into a physical constraint enhanced bidirectional long short-term memory network to predict the dancing evolution trend, and outputting a joint prediction matrix of the dancing amplitude and frequency of the transmission line in the future; Inputting the fused feature vector and the joint prediction matrix of the galloping amplitude and frequency into a pre-trained risk propagation inference model to calculate multi-level risk quantification indicators of the transmission line, wherein the multi-level risk quantification indicators include the probability of conductor dynamic tension exceeding the limit and the probability of insulation gap breakdown risk; Generate corresponding risk warning prompts based on the multi-level risk quantification indicators and send them to corresponding risk management devices.

[0006] As an improvement to the above solution, performing spatiotemporal alignment and three-dimensional posture reconstruction on the multi-physics field dataset to generate a dynamic three-dimensional point cloud sequence of the power transmission line movement includes the following sub-steps: Performing phase synchronization calibration based on the dancing resonance frequency on the motion acceleration data and video stream data in the multi-physics field dataset to obtain a time-aligned dataset; the multi-physics field dataset includes motion acceleration data, ambient wind speed data, and video stream data; Based on the time-aligned data set, a wire rigid body dynamics constraint model is applied to perform spatial coordinate compensation on the ambient wind speed data to obtain spatially corrected data; The spatial correction data is input into a three-dimensional reconstruction algorithm that integrates the bending stiffness parameters of the conductor to generate a dynamic three-dimensional point cloud sequence of the transmission line's galloping.

[0007] As an improvement to the above solution, extracting dancing energy distribution features, frequency spectrum features, and spatial trajectory features from the dynamic three-dimensional point cloud sequence, and fusing the three features to obtain a fused feature vector, includes the following sub-steps: Based on the conductor aerodynamic structure coupling equation, the dynamic three-dimensional point cloud sequence is subjected to vortex-induced vibration mode separation to obtain the dancing energy distribution characteristics; Applying dancing dominant frequency adaptive filtering to the dynamic three-dimensional point cloud sequence to obtain frequency spectrum features; Performing trajectory curvature correction on the dynamic three-dimensional point cloud sequence based on the catenary equation of the wire to obtain spatial trajectory features; The dancing energy distribution feature, the frequency spectrum feature and the spatial trajectory feature are fused to obtain the fused feature vector.

[0008] As an improvement to the above solution, the method of inputting the fused feature vector into a physical constraint enhanced bidirectional long short-term memory network to predict the galloping evolution trend and outputting a joint prediction matrix of the galloping amplitude and frequency of the transmission line includes the following sub-steps: Reorganizing the fused feature vector in time series dimension to generate a feature tensor sequence with a fixed time window length; Synchronously inputting the feature tensor sequence into the forward computing layer and the reverse computing layer of the bidirectional long short-term memory network, and outputting a forward dance evolution hidden state sequence and a reverse dance evolution hidden state sequence respectively; A physical rule library is constructed based on the displacement constraint equation of the wire suspension point and the aerodynamic torque balance condition, and the forward dancing evolution hidden state sequence and the reverse dancing evolution hidden state sequence are dynamically weighted and fused to generate a physical constraint-corrected dancing evolution feature matrix; The physical constraint-corrected dance evolution feature matrix is ​​mapped into a joint prediction matrix of the dance amplitude and frequency.

[0009] As an improvement to the above solution, the joint prediction matrix of the galloping amplitude and frequency is input into a pre-trained risk propagation inference model to calculate the multi-level risk quantification index of the transmission line. The multi-level risk quantification index includes the probability of conductor dynamic tension exceeding the limit and the insulation gap breakdown risk probability, including the following sub-steps: performing amplitude-frequency parameter decoupling on the joint prediction matrix of the dancing amplitude and frequency to separate and obtain a dancing amplitude component sequence and a dancing frequency component sequence; Inputting the dancing amplitude component sequence into a pre-trained risk propagation inference model to calculate the dynamic tension of the conductor, and combining it with the static initial tension value of the transmission line to calculate the dynamic tension time history data of the transmission line; Based on the calculated dynamic tension time history data and a preset conductor tension safety threshold, a proportion of durations of tension exceeding a limit is calculated to obtain a probability of the conductor dynamic tension exceeding a limit; The dancing amplitude component sequence and the dancing frequency component sequence are input into a pre-trained risk propagation inference model to perform insulation gap electric field analysis. Combined with the inherent insulation configuration parameters of the transmission line, the minimum air gap distance change and the local electric field strength distribution are calculated. Based on the calculation results, the insulation gap breakdown risk probability is obtained.

[0010] As an improvement to the above solution, generating corresponding risk warning prompts based on the multi-level risk quantification indicators and sending them to corresponding risk management devices includes the following sub-steps: Determining the risk type, risk location, and risk level based on the conductor dynamic tension exceeding probability and the insulation gap breakdown risk probability in the multi-level risk quantification indicators; Generate a structured risk warning prompt based on the risk level, the risk warning prompt including risk type, risk location and risk level information; The risk warning prompt is dynamically sent to a corresponding risk management device based on the risk location.

[0011] Another embodiment of the present invention provides a transmission line galloping risk warning system based on multi-physics field fusion, including: An acquisition module is used to obtain a multi-physics field dataset of a transmission line galloping process; A generation module, configured to perform spatiotemporal alignment and three-dimensional posture reconstruction on the multi-physics field data set to generate a dynamic three-dimensional point cloud sequence of the transmission line dancing; a fusion module, configured to extract dancing energy distribution features, frequency spectrum features, and spatial trajectory features from the dynamic three-dimensional point cloud sequence, and fuse the three features to obtain a fused feature vector; A prediction module, configured to input the fused feature vector into a physical constraint enhanced bidirectional long short-term memory network to predict the galloping evolution trend, and output a joint prediction matrix of the galloping amplitude and frequency of the transmission line; a calculation module, configured to input the joint prediction matrix of the galloping amplitude and frequency into a pre-trained risk propagation inference model to calculate the multi-level risk quantification index of the transmission line; The early warning module is used to generate corresponding risk early warning prompts according to the multi-level risk quantification indicators and send them to the corresponding risk management equipment.

[0012] As an improvement to the above solution, the generation module is specifically used to: Performing phase synchronization calibration based on the dancing resonance frequency on the motion acceleration data and video stream data in the multi-physics field dataset to obtain a time-aligned dataset; the multi-physics field dataset includes motion acceleration data, ambient wind speed data, and video stream data; Based on the time-aligned data set, a wire rigid body dynamics constraint model is applied to perform spatial coordinate compensation on the ambient wind speed data to obtain spatially corrected data; The spatial correction data is input into a three-dimensional reconstruction algorithm that integrates the bending stiffness parameters of the conductor to generate a dynamic three-dimensional point cloud sequence of the transmission line's galloping.

[0013] As an improvement to the above solution, the fusion module is specifically used to: Based on the conductor aerodynamic structure coupling equation, the dynamic three-dimensional point cloud sequence is subjected to vortex-induced vibration mode separation to obtain the dancing energy distribution characteristics; Applying dancing dominant frequency adaptive filtering to the dynamic three-dimensional point cloud sequence to obtain frequency spectrum features; Performing trajectory curvature correction on the dynamic three-dimensional point cloud sequence based on the catenary equation of the wire to obtain spatial trajectory features; The dancing energy distribution feature, the frequency spectrum feature and the spatial trajectory feature are fused to obtain the fused feature vector.

[0014] As an improvement to the above solution, the prediction module is specifically used to: Reorganizing the fused feature vector in time series dimension to generate a feature tensor sequence with a fixed time window length; Synchronously inputting the feature tensor sequence into the forward computing layer and the reverse computing layer of the bidirectional long short-term memory network, and outputting a forward dance evolution hidden state sequence and a reverse dance evolution hidden state sequence respectively; A physical rule library is constructed based on the displacement constraint equation of the wire suspension point and the aerodynamic torque balance condition, and the forward dancing evolution hidden state sequence and the reverse dancing evolution hidden state sequence are dynamically weighted and fused to generate a physical constraint-corrected dancing evolution feature matrix; The physical constraint-corrected dance evolution feature matrix is ​​mapped into a joint prediction matrix of the dance amplitude and frequency.

[0015] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: By acquiring a multi-physics field dataset of the transmission line dancing process, multi-dimensional physical information related to the dancing is collected; the multi-physics field dataset is subjected to spatiotemporal alignment and three-dimensional posture reconstruction to generate a dynamic three-dimensional point cloud sequence, eliminating the spatiotemporal deviation of the data; the dancing energy distribution characteristics, frequency spectrum characteristics and spatial trajectory characteristics are extracted from the dynamic three-dimensional point cloud sequence and fused to obtain a fused feature vector, and the energy, frequency and spatial motion characteristics of the dancing are comprehensively characterized by multi-dimensional features to avoid the limitations of a single feature; the fused feature vector is input into a bidirectional long short-term memory network enhanced by physical constraints, and the bidirectional long short-term memory network is used to capture the forward and backward correlation information of the time series data and to enhance physical constraints to ensure that the prediction conforms to the actual physical laws, thereby better predicting the dancing evolution trend and outputting a joint prediction matrix of the dancing amplitude and frequency; the joint prediction matrix of the dancing amplitude and frequency is input into a pre-trained risk propagation inference model, and the predicted dancing characteristics are converted into hierarchical quantitative risk indicators to achieve multi-level risk quantification; finally, a risk warning prompt is generated based on the multi-level risk quantification indicator and sent to the corresponding risk management equipment to complete the warning. In summary, the embodiments of the present invention acquire multi-physical field data, integrate the data and extract and fuse multi-dimensional features, use a physical constraint-enhanced bidirectional long short-term memory network to accurately predict the galloping evolution trend, and quantify multi-level risks through a risk propagation reasoning model. It is possible to accurately predict the galloping evolution trend of transmission lines, thereby accurately warning of transmission line galloping risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of a transmission line galloping risk warning method based on multi-physics field fusion provided by one embodiment of the present invention; Figure 2It is a structural diagram of a transmission line galloping risk warning system based on multi-physical field fusion provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0018] See also Figure 1 , is a flow chart of a transmission line galloping risk warning method based on multi-physics field fusion provided by one embodiment of the present invention. The transmission line galloping risk warning method based on multi-physics field fusion includes: S10, obtaining a multi-physics field dataset of the galloping process of the transmission line; S11, performing spatiotemporal alignment and three-dimensional posture reconstruction on the multi-physics field dataset to generate a dynamic three-dimensional point cloud sequence of the transmission line dancing; S12, extracting dancing energy distribution features, frequency spectrum features, and spatial trajectory features from the dynamic three-dimensional point cloud sequence, and fusing the three features to obtain a fused feature vector; S13, inputting the fused feature vector into a physical constraint enhanced bidirectional long short-term memory network to predict the dancing evolution trend, and outputting a joint prediction matrix of the dancing amplitude and frequency of the transmission line in the future; S14, inputting the fused feature vector and the joint prediction matrix of the galloping amplitude and frequency into a pre-trained risk propagation inference model to calculate multi-level risk quantification indicators of the transmission line, wherein the multi-level risk quantification indicators include the probability of conductor dynamic tension exceeding the limit and the probability of insulation gap breakdown risk; S15: Generate corresponding risk warning prompts according to the multi-level risk quantification indicators and send them to corresponding risk management devices.

[0019] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: By acquiring a multi-physics field dataset of the transmission line dancing process, multi-dimensional physical information related to the dancing is collected; the multi-physics field dataset is subjected to spatiotemporal alignment and three-dimensional posture reconstruction to generate a dynamic three-dimensional point cloud sequence, eliminating the spatiotemporal deviation of the data; the dancing energy distribution characteristics, frequency spectrum characteristics and spatial trajectory characteristics are extracted from the dynamic three-dimensional point cloud sequence and fused to obtain a fused feature vector, and the energy, frequency and spatial motion characteristics of the dancing are comprehensively characterized by multi-dimensional features to avoid the limitations of a single feature; the fused feature vector is input into a bidirectional long short-term memory network enhanced by physical constraints, and the bidirectional long short-term memory network is used to capture the forward and backward correlation information of the time series data and to enhance physical constraints to ensure that the prediction conforms to the actual physical laws, thereby better predicting the dancing evolution trend and outputting a joint prediction matrix of the dancing amplitude and frequency; the joint prediction matrix of the dancing amplitude and frequency is input into a pre-trained risk propagation inference model, and the predicted dancing characteristics are converted into hierarchical quantitative risk indicators to achieve multi-level risk quantification; finally, a risk warning prompt is generated based on the multi-level risk quantification indicator and sent to the corresponding risk management equipment to complete the warning. In summary, the embodiments of the present invention acquire multi-physical field data, integrate the data and extract and fuse multi-dimensional features, use a physical constraint-enhanced bidirectional long short-term memory network to accurately predict the galloping evolution trend, and quantify multi-level risks through a risk propagation reasoning model. It is possible to accurately predict the galloping evolution trend of transmission lines, thereby accurately warning of transmission line galloping risks.

[0020] As one example of the above solution, performing spatiotemporal alignment and three-dimensional posture reconstruction on the multi-physics field dataset to generate a dynamic three-dimensional point cloud sequence of the transmission line movement includes the following sub-steps: Performing phase synchronization calibration based on the dancing resonance frequency on the motion acceleration data and video stream data in the multi-physics field dataset to obtain a time-aligned dataset; the multi-physics field dataset includes motion acceleration data, ambient wind speed data, and video stream data; Based on the time-aligned data set, a wire rigid body dynamics constraint model is applied to perform spatial coordinate compensation on the ambient wind speed data to obtain spatially corrected data; The spatial correction data is input into a three-dimensional reconstruction algorithm that integrates the bending stiffness parameters of the conductor to generate a dynamic three-dimensional point cloud sequence of the transmission line's galloping.

[0021] In this embodiment, phase synchronization calibration is first performed on the motion acceleration data and video stream data in the multi-physics field dataset based on the inherent resonant frequency characteristics of the transmission line galloping, achieving precise alignment of multi-source heterogeneous data in the time dimension. Then, the ambient wind speed data is spatially compensated using the conductor rigid body dynamics constraint model, correcting the conductor's spatial position deviation under wind load and eliminating environmental interference. Finally, the time-space aligned and spatially corrected data is input into a 3D reconstruction algorithm that incorporates the conductor's bending stiffness parameters. By introducing the mechanical property constraints of the conductor material, a dynamic 3D point cloud sequence is generated that truly reflects the deformation characteristics of the transmission line galloping process. Therefore, this embodiment solves the problem of inconsistent time bases in multi-source data through resonant frequency phase synchronization, eliminates the spatial positioning error of the ambient wind load through rigid body dynamics constraint compensation, and ensures that the point cloud sequence conforms to the physical deformation laws of the conductor through bending stiffness fusion reconstruction, thereby constructing a high-fidelity dynamic 3D model of the galloping.

[0022] Specifically, the working process of this embodiment is as follows: First, the multi-physics field data set is collected as follows: the motion acceleration data is collected by three-axis acceleration sensors deployed at the suspension point and midpoint of the wire, with a sampling frequency of 200Hz, recording the instantaneous acceleration in three directions: x (along the line), y (perpendicular to the line), and z (vertical); the video stream data is collected by a camera near the line; the ambient wind speed data is collected by a three-dimensional ultrasonic anemometer next to the camera, recording the instantaneous wind speed, wind direction, and turbulence intensity. Due to the hardware response delay and sampling trigger mechanism differences between the acceleration sensor and the camera, there is a time phase deviation between the two types of data, and phase synchronization calibration is required based on the dancing resonant frequency. Specifically, the motion acceleration data is first subjected to a fast Fourier transform (FFT) to extract the main dancing resonant frequency. (usually 0.1-5Hz, determined by the conductor type, spacing and wind speed); then As a benchmark, calculate the initial phase difference between acceleration data and video stream data (Assume that the acceleration data phase , video stream data phase ,but );according to Correct the video stream timestamp using the following formula: , The original timestamp of the video stream. The corrected video stream data, original motion acceleration data, and ambient wind speed data are integrated according to the unified timestamp to obtain a time-aligned dataset.

[0023] Based on the time alignment dataset, the YOLOv8 algorithm is used to extract key feature points such as the wire hanging point and the midpoint of the span from the video stream data, and the camera internal parameters (focal length 、Principal point coordinates( )) and external parameters to calculate the real-time three-dimensional coordinates of each feature point in the world coordinate system Then, the instantaneous velocity and displacement of the midpoint of the wire are obtained by quadratic integration of the motion acceleration data. Combining the rigid body motion characteristics of the wire (no obvious deformation during dancing, only rigid body motion) to build a constraint model, the real-time spatial posture parameters of the wire (midpoint displacement , deflection angle , pitch angle According to the attitude parameters, the environmental wind speed data (three components in the world coordinate system) ) is converted into the component in the local coordinate system of the conductor (along the conductor axis , vertical wire radial , vertical direction ), compensates for the wind speed deviation caused by the spatial distance between the anemometer and the conductor and the conductor posture, and forms spatial correction data.

[0024] Extract continuous frame images of video stream from spatial correction data and use SIFT feature matching algorithm to preliminarily generate original 3D point cloud of wire ( is the original coordinate ); At the same time, obtain the physical parameters of the wire (such as the elastic modulus of the wire , Sectional moment of inertia , calculate the bending stiffness ), combined with motion acceleration data Directional acceleration , the static theoretical shape of the wire is obtained through quadratic integration and catenary model, which is used as the deformation correction benchmark. A point cloud correction model integrating bending stiffness is constructed based on the bending theory of beams and combined with the aerodynamic torque of the wire. , correct the original point cloud, the correction formula is: , , is the corrected point cloud coordinate, for Directional deformation correction, for Direction unit vector, The coordinates of the point cloud along the wire axis are corrected one by one and arranged in chronological order to form a dynamic 3D point cloud sequence with a time interval consistent with the video stream frame rate, fully recording the spatial form and bending deformation of the wire.

[0025] The fusion feature vector is obtained by fusing the three features of the spatial trajectory feature, the frequency spectrum feature and the spatial trajectory feature, including the following sub-steps: Based on the conductor aerodynamic structure coupling equation, the dynamic three-dimensional point cloud sequence is subjected to vortex-induced vibration mode separation to obtain the dancing energy distribution characteristics; Applying dancing dominant frequency adaptive filtering to the dynamic three-dimensional point cloud sequence to obtain frequency spectrum features; Performing trajectory curvature correction on the dynamic three-dimensional point cloud sequence based on the catenary equation of the wire to obtain spatial trajectory features; The dancing energy distribution feature, the frequency spectrum feature and the spatial trajectory feature are fused to obtain the fused feature vector.

[0026] In this embodiment, the vortex-induced vibration mode separation of the dynamic three-dimensional point cloud sequence is first performed based on the conductor aerodynamic structure coupling equation, and the energy characteristics reflecting the dancing energy level distribution are extracted by analyzing the vibration modal characteristics of the conductor under the action of wind load; then the dancing dominant frequency adaptive filtering algorithm is applied to the point cloud sequence to dynamically identify and extract the key frequency components in the dancing process, forming a frequency spectrum feature that characterizes the periodic law of vibration; at the same time, the curvature of the point cloud trajectory is corrected in combination with the conductor catenary equation to eliminate the static deformation interference caused by gravity and restore the real dancing spatial trajectory characteristics; finally, the three types of physical constraint features are multi-dimensionally fused to construct a fusion feature vector containing energy, frequency and spatial elements. In summary, this embodiment improves the characterization accuracy and physical consistency of the dancing energy distribution, frequency characteristics and spatial trajectory characteristics through the physical constraint feature extraction method of vortex-induced vibration mode separation, dominant frequency adaptive filtering and catenary trajectory correction.

[0027] As an example, the working process of this embodiment is: From the dynamic 3D point cloud sequence, first extract the discrete points of the wire along the span every 0.5m in the main dancing direction ( An improved conductor aerodynamic-structural coupling equation is constructed by combining displacement time-history data (in both directions) with the conductor's mass per unit length (calculated from conductor density and cross-sectional area), damping coefficient, bending stiffness, and aerodynamic parameters. This equation optimizes the traditional linear model's depiction of the coupling relationship between aerodynamic loads and structural vibration by introducing a nonlinear lift coefficient term that varies dynamically with displacement and velocity. The equation is expressed as follows: , is the mass per unit length of the conductor, is the damping coefficient, is the bending stiffness, is the air density, is the wind speed, is the wire diameter, is the dynamic lift coefficient, For the The Galerkin method is used to perform modal decomposition on this equation, obtaining the natural frequencies and vibration modes of the first- and second-order vortex-induced vibration modes. The energy (including kinetic energy and elastic potential energy) of each mode at each discrete point is then calculated using the displacement time history data. The energy values ​​are then mapped to the entire length of the conductor at the discrete point locations, forming a dancing energy distribution feature that can clearly distinguish the energy contributions of different modes at different locations on the conductor.

[0028] The displacement time-history data of the conductor midpoint in the dynamic three-dimensional point cloud sequence (with the largest dancing amplitude and the most significant frequency characteristics) is selected. The time-frequency energy spectrum is first obtained through continuous wavelet transform, and the frequency interval with an energy share of more than 60% is identified as the candidate dominant frequency range. An adaptive filter is constructed based on this interval, and the filter passband frequency is dynamically adjusted with the energy spectrum: when the energy of a certain frequency accounts for ≥10% of the total energy, it is included in the passband; otherwise, it is considered as noise and filtered out. A power spectral density analysis is performed on the filtered displacement time-history data to extract the peak frequency (dominant frequency), 3dB bandwidth, and peak power. At the same time, the power spectrum data of the conductor suspension point and 1 / 4 span point are collected, and the dominant frequency deviation value of each position is calculated and integrated to form the frequency spectrum characteristics.

[0029] Extract the spatial point set of the conductor at a certain moment in the dynamic 3D point cloud sequence, fit the original trajectory curve using the least squares method, and then calculate the static theoretical trajectory based on the conductor catenary equation (no need to list the equation, directly use the existing catenary model parameters for calculation). Compare the original trajectory with the theoretical trajectory to obtain the deviation of each point. Correct the points where the deviation exceeds 1 / 5 of the conductor diameter. The correction formula is: , , 、 is the corrected coordinate, 、 is the original coordinate, is the deviation, 、 is the theoretical trajectory coordinate, and 0.9 is the correction factor. Based on the corrected trajectory, the curvature and torsion are calculated, and the coordinates of the highest and lowest points of the trajectory and the swing amplitude are extracted to form the spatial trajectory features.

[0030] Standardize the three types of features: reduce the dancing energy distribution feature (two-dimensional matrix) to a vector through PCA , organize the frequency spectrum features (including 6 parameters) into vectors , organize the spatial trajectory features (including 5 parameters) into vectors . Use random forest algorithm to determine feature weights 、 、 , through the weighted fusion formula: ,right 、 Zero pad to After calculating the dimension, we finally get the fusion feature vector ,This vector covers the core information of energy, frequency, and space, and meets the input requirements of the subsequent ,prediction model.

[0031] As one example of the above solution, inputting the fused feature vector into a physical constraint enhanced bidirectional long short-term memory network to predict the galloping evolution trend and outputting a joint prediction matrix of the galloping amplitude and frequency of the transmission line includes the following sub-steps: Reorganizing the fused feature vector in time series dimension to generate a feature tensor sequence with a fixed time window length; Synchronously inputting the feature tensor sequence into the forward computing layer and the reverse computing layer of the bidirectional long short-term memory network, and outputting a forward dance evolution hidden state sequence and a reverse dance evolution hidden state sequence respectively; A physical rule library is constructed based on the displacement constraint equation of the wire suspension point and the aerodynamic torque balance condition, and the forward dancing evolution hidden state sequence and the reverse dancing evolution hidden state sequence are dynamically weighted and fused to generate a physical constraint-corrected dancing evolution feature matrix; The physical constraint-corrected dance evolution feature matrix is ​​mapped into a joint prediction matrix of the dance amplitude and frequency.

[0032] In this embodiment, the fused feature vectors are reorganized in time series dimensions to construct a feature tensor sequence with a fixed time window length, using a standardized time series input format. The feature tensor sequence is then synchronously input into the forward and reverse computational layers of a bidirectional long short-term memory (BiLSTM) network. Bidirectional time series learning is used to capture the historical dependencies and future trends of dancing evolution, generating complementary forward and reverse dancing evolution hidden state sequences. Furthermore, a physical rule base is constructed based on the displacement constraint equations of the wire suspension points and the aerodynamic torque balance condition. The forward and reverse hidden state sequences are dynamically weighted and fused. The prediction bias of the data-driven model is corrected through physical mechanism constraints to generate a dancing evolution feature matrix with enhanced physical consistency. Finally, the corrected feature matrix is ​​mapped into a joint prediction matrix of dancing amplitude and frequency, achieving coupled output of multi-dimensional evolution trends. This embodiment of the present invention dynamically weights and fuses the hidden states of the bidirectional long short-term memory (BiLSTM) network using a physical rule base, improving the physical consistency and evolution trend capture accuracy of the joint prediction matrix of dancing amplitude and frequency.

[0033] The working process of this embodiment is as follows: First, the acquired fusion feature vector is reorganized in the time series dimension. The fusion feature vector is initially a one-dimensional vector, and the dimension corresponds to the integration dimension of the three types of features: energy, frequency, and space. Considering that the dancing evolution has a time series correlation, it is necessary to set a fixed time window length based on the historical periodic characteristics of the transmission line dancing (such as the time series dependency of the past 10-30 sampling cycles). , the fusion feature vectors on the continuous time series are intercepted by sliding the time window, and each intercepted vector group is reorganized into a dimension of The feature tensor of is the dimension of the fusion feature vector), multiple tensors are arranged in chronological order to generate a feature tensor sequence, ensuring that the subsequent Bi-LSTM can capture the forward and backward correlation information in the time series.

[0034] The feature tensor sequence is synchronously input into the physical constraint enhanced Bi-LSTM, which consists of a forward computing layer and a backward computing layer. The forward computing layer processes the feature tensor sequence in the order from the past to the present, and learns the evolution law of the dance from the historical state to the current state through the hidden layer unit. The output dimension is The hidden state sequence of the forward dancing evolution ( is the number of Bi-LSTM hidden layer units); the reverse calculation layer processes the feature tensor sequence in reverse order from the present to the past, and learns the evolution law of the dance from the future (the subsequent state of the current moment) to the current state. The output dimension is also The reverse dancing evolution hidden state sequence During the two-layer calculation process, the activation function and state update logic of the hidden layer units follow the standard LSTM mechanism to ensure that long-term temporal dependencies can be effectively captured.

[0035] According to the displacement constraint equation of the conductor suspension point and the aerodynamic torque balance condition, a physical rule base is constructed to realize the dynamic weighted fusion of the forward and reverse hidden state sequences. In the physical rule base, the displacement constraint equation of the conductor suspension point stipulates that the suspension point has no displacement during the dancing process (because the suspension point is fixed to the tower, the displacement ), the aerodynamic moment balance condition stipulates that the sum of the aerodynamic moment and the structural resistance moment of any cross section of the conductor is 0 ( Based on this rule base, the conformity between the forward and reverse hidden state sequences and the physical constraints at each time step is calculated (such as the deviation rate between the suspension point displacement predicted by the hidden state and the constraint value, and the torque imbalance), and the conformity is converted into a dynamic weight coefficient and ( is the time step, and , the higher the degree of conformity, the greater the weight of the hidden state sequence). The forward and reverse hidden state sequences are fused using the improved dynamic weighted fusion formula, which is: , For the The hidden state vector after the time step physical constraint correction; all the time steps Arrange them in order, and the generated dimensions are The dancing evolution characteristic matrix corrected by physical constraints.

[0036] Finally, the dance evolution feature matrix corrected by physical constraints is input into the fully connected mapping layer, which transforms the dimension of the feature matrix from Convert to ( is the preset dancing evolution trend prediction time step, and 2 corresponds to the dancing amplitude and frequency respectively); the converted data is organized according to the dimension of "prediction time step - (amplitude + frequency)" to form a joint prediction matrix of the dancing amplitude and frequency of the transmission line. Each row in the matrix corresponds to the predicted dancing amplitude value and frequency predicted value of a prediction time step, and the core quantitative indicators of the dancing evolution trend are fully output.

[0037] As one example of the above solution, the joint prediction matrix of the galloping amplitude and frequency is input into a pre-trained risk propagation inference model to calculate the multi-level risk quantification index of the transmission line, wherein the multi-level risk quantification index includes the probability of conductor dynamic tension exceeding the limit and the insulation gap breakdown risk probability, including the following sub-steps: performing amplitude-frequency parameter decoupling on the joint prediction matrix of the dancing amplitude and frequency to separate and obtain a dancing amplitude component sequence and a dancing frequency component sequence; Inputting the dancing amplitude component sequence into a pre-trained risk propagation inference model to calculate the dynamic tension of the conductor, and combining it with the static initial tension value of the transmission line to calculate the dynamic tension time history data of the transmission line; Based on the calculated dynamic tension time history data and a preset conductor tension safety threshold, a proportion of durations of tension exceeding a limit is calculated to obtain a probability of the conductor dynamic tension exceeding a limit; The dancing amplitude component sequence and the dancing frequency component sequence are input into a pre-trained risk propagation inference model to perform insulation gap electric field analysis. Combined with the inherent insulation configuration parameters of the transmission line, the minimum air gap distance change and the local electric field strength distribution are calculated. Based on the calculation results, the insulation gap breakdown risk probability is obtained.

[0038] In this embodiment, the parameters of the joint prediction matrix of the dancing amplitude and frequency are decoupled to separate the independent dancing amplitude component sequence and the dancing frequency component sequence, thereby achieving accurate decomposition of the multi-dimensional prediction parameters; the amplitude component sequence is then input into the risk propagation inference model, and the dynamic tension is calculated in combination with the static initial tension value of the transmission line to generate time-course data reflecting the real-time change of tension, and the probability of the dynamic tension exceeding the limit of the conductor is quantified by statistically analyzing the proportion of the duration of tension exceeding the limit; at the same time, the amplitude component and the frequency component are synchronously input into the risk propagation inference model, and the insulation gap electric field analysis is performed in combination with the line insulation configuration parameters to calculate the minimum air gap distance change and the local electric field strength distribution, thereby deducing the insulation gap breakdown risk probability. It can be seen that this embodiment achieves accurate quantification of the conductor tension exceeding limit probability and the insulation gap breakdown risk probability through dual-channel physical modeling of dynamic tension time-course analysis and insulation gap electric field calculation.

[0039] For ease of understanding, the working process of this embodiment is described as follows: First, the amplitude-frequency parameter decoupling is performed on the joint prediction matrix of dance amplitude and frequency. The dimension of the joint prediction matrix is Parameter decoupling is accomplished by extracting by columns: extract all elements in the first column of the matrix and arrange them in the order of the prediction time step to obtain a length of The dancing amplitude component sequence ( For the The dancing amplitude of the time step); extract all the elements in the second column of the matrix and arrange them in the order of the same time step to get a length of The dancing frequency component sequence ( For the The dancing frequency of the time step is determined by the two parameters to ensure that the two types of parameters can be used for the calculation of different subsequent risk indicators.

[0040] In addition, the training method of the pre-trained risk propagation reasoning model is as follows: This model is an improved reasoning model based on graph neural network (GNN) combined with physical rules, which aims to characterize the propagation path and quantitative relationship of transmission line galloping risk in the "conductor-insulator-tower" system. The input layer of the model is defined as "risk source parameters" (such as dancing amplitude and frequency) and "system physical parameters" (such as conductor characteristics and insulation configuration), and the output layer is defined as "multi-level risk quantification indicators" (such as the probability of dynamic tension exceeding the limit and the probability of insulation gap breakdown); in the network structure, the nodes are divided into "risk unit nodes" (including conductor segment nodes, insulator nodes, and suspension point nodes) and "propagation relationship nodes" (representing the force / electric field transmission relationship between adjacent risk units), and the edge weights are determined through joint training of historical risk data and physical rules; the training process uses "actual dancing accident data and simulated dancing data" as the training set (including the dancing-risk correspondence under different conductor models, spans, and wind speeds), and adopts a mixed loss function of "physical loss and probabilistic loss" (physical loss ensures that the inference results conform to the laws of mechanics / electric fields, and probabilistic loss optimizes the accuracy of risk probability calculation). The training is iterated until the loss function converges, and finally a risk propagation inference model with generalization capabilities that meet engineering needs is obtained.

[0041] The dancing amplitude component sequence Input the "dynamic tension calculation module" of the risk propagation inference model. This module has a built-in conductor dynamic tension calculation model, combined with the static initial tension value of the transmission line (determined by the conductor model, span and installation conditions, as known input parameters of the model), by considering the coupled effects of the dancing amplitude and frequency on the tension fluctuation, the dynamic tension time history data is calculated. The dynamic tension is calculated using the improved formula: , For the The dynamic tension of the wire at each time step, is the amplitude-tension coupling coefficient (reflecting the influence of the dancing amplitude on the tension fluctuation, which is determined by the elastic modulus of the conductor and the pitch), is the frequency-tension correction coefficient (reflecting the effect of dancing frequency on the mean tension deviation, determined by the wire damping characteristics), is the dancing phase term (characterizing the temporal periodicity of tension fluctuations). The model calculates the dynamic tension step by step through this formula, and the length is Dynamic tension time history data .

[0042] Based on dynamic tension time history data and preset conductor tension safety threshold (determined by the rated tensile strength of the conductor and the safety factor, such as 80% of the rated tensile strength), calculate the probability of the conductor dynamic tension exceeding the limit. The specific process is: traverse each value in the dynamic tension time history data, determine Is it greater than , statistics satisfy The number of time steps ; The number of exceeded time steps is divided by the total prediction time steps The ratio of is taken as the probability of conductor dynamic tension exceeding the limit, that is: , The probability of conductor dynamic tension exceeding the limit. The larger the value, the higher the risk of conductor strand or wire breakage due to tension exceeding the limit.

[0043] The dancing amplitude component sequence With frequency component sequence The "Insulation Clearance Electric Field Analysis Module" is a common input to the risk propagation inference model. This module combines the inherent insulation configuration parameters of the transmission line (including the length of the insulator string) , Insulator string swing angle limit , rated creepage distance etc., as known input parameters of the model), first calculate the minimum air gap distance change through geometric analysis: based on the dancing amplitude The minimum air gap between the conductor and the tower / ground is derived from the swing characteristics of the insulator string. (The larger the dancing amplitude, The smaller the value); then calculate the local electric field intensity distribution through electric field simulation, and focus on extracting The maximum electric field strength at the corresponding position Based on the above calculation results, the improved insulation gap breakdown risk probability formula is adopted: , is the insulation gap breakdown risk probability, is the electric field-gap combined influence coefficient (corrected by air humidity and pressure), is the air breakdown field strength (about 30 kV / cm at standard atmospheric pressure, corrected according to the actual environment in the model), is the rated minimum air gap (determined by the line voltage level). The model calculates the , take all time steps The maximum value of is taken as the final insulation gap breakdown risk probability. The larger the value, the higher the risk of flashover of the insulator string due to gap breakdown.

[0044] Finally, the risk propagation reasoning model outputs the probability of conductor dynamic tension exceeding the limit Probability of insulation gap breakdown risk , and the two together constitute the multi-level risk quantification index of transmission lines.

[0045] As one example of the above solution, generating corresponding risk warning prompts according to the multi-level risk quantification indicators and sending them to corresponding risk management devices includes the following sub-steps: Determining the risk type, risk location, and risk level based on the conductor dynamic tension exceeding probability and the insulation gap breakdown risk probability in the multi-level risk quantification indicators; Generate a structured risk warning prompt based on the risk level, the risk warning prompt including risk type, risk location and risk level information; The risk warning prompt is dynamically sent to a corresponding risk management device based on the risk location.

[0046] In this embodiment, the risk type, location, and level are located by analyzing multi-level risk quantification indicators (the probability of exceeding the dynamic tension of the conductor and the probability of insulation gap breakdown risk), and a risk warning prompt containing structured information is generated. The warning prompt is dynamically sent to the corresponding risk management device based on the risk location, thereby achieving accurate and timely push of risk information.

[0047] Specifically, the working process of this embodiment is as follows: first, the probability of conductor dynamic tension exceeding the limit and the probability of insulation gap breakdown risk in the multi-level risk quantification indicators are used as the core judgment basis, and the risk tracing data output by the risk propagation reasoning model are combined to determine the key information: if the probability of conductor dynamic tension exceeding the limit is higher than the preset threshold (such as 60%), the risk type is determined to be "conductor dynamic tension exceeding the limit risk", and if the probability of insulation gap breakdown risk is higher than the threshold, it is determined to be "insulation gap breakdown risk"; at the same time, the transmission line topology data associated with the model (including span number, tower coordinates, and insulator installation position) is used to lock the specific location where the risk occurs (such as "110kV East Line 3 #28-#29 middle section of conductor" and "220kV South Line 5 #45 tower insulator string"), and the risk level is divided according to the probability interval (probability of 60%-75% is level 3 risk, 76%-90% is level 2 risk, and 91% and above is level 1 risk). A structured risk warning is then generated based on the determined risk level, integrating key content in a fixed information module format. For example, a Level 3 risk warning includes "Risk Type: Conductor Dynamic Tension Exceeding Limit; Risk Location: Middle Section of Conductor #28-#29 on the 110kV East Line 3; Risk Level: Level 3 Risk; Risk Probability: 68%." Level 2 and higher risk warnings additionally include "Risk Impact: May Cause Conductor Fatigue Damage / Insulator Flashover," ensuring complete information and consistent with O&M reading habits. Finally, based on the risk location, a pre-set device association mapping table (which pre-stores the O&M terminal numbers, monitoring center receiving addresses, and on-site warning device communication protocols for each line segment and tower) is queried. The structured risk warning is dynamically transmitted to the corresponding risk management device via a dedicated power communication network (such as the distribution automation system communication link). For Level 3 risks, this warning is sent to the O&M team's mobile inspection app, while Level 2 and higher risks are simultaneously sent to the transmission monitoring center's main system and the on-site tower's audible and visual alarm devices. This ensures differentiated and precise communication of different risk levels, supporting O&M personnel in rapidly addressing risks.

[0048] See also Figure 2 , is a schematic diagram of a transmission line galloping risk warning system based on multi-physics field fusion provided by one embodiment of the present invention. The transmission line galloping risk warning system based on multi-physics field fusion includes: An acquisition module 10 is used to acquire a multi-physics field dataset of a galloping process of a transmission line; A generating module 11 is configured to perform spatiotemporal alignment and three-dimensional posture reconstruction on the multi-physics field dataset to generate a dynamic three-dimensional point cloud sequence of the power transmission line; A fusion module 12 is configured to extract dancing energy distribution features, frequency spectrum features, and spatial trajectory features from the dynamic three-dimensional point cloud sequence, and fuse the three features to obtain a fused feature vector; A prediction module 13 is configured to input the fused feature vector into a physical constraint enhanced bidirectional long short-term memory network to predict the galloping evolution trend and output a joint prediction matrix of the galloping amplitude and frequency of the transmission line; A calculation module 14 is configured to input the joint prediction matrix of the dancing amplitude and frequency into a pre-trained risk propagation inference model to calculate the multi-level risk quantification index of the transmission line; The early warning module 15 is used to generate corresponding risk early warning prompts according to the multi-level risk quantification indicators and send them to the corresponding risk management equipment.

[0049] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: By acquiring a multi-physics field dataset of the transmission line dancing process, multi-dimensional physical information related to the dancing is collected; the multi-physics field dataset is subjected to spatiotemporal alignment and three-dimensional posture reconstruction to generate a dynamic three-dimensional point cloud sequence, eliminating the spatiotemporal deviation of the data; the dancing energy distribution characteristics, frequency spectrum characteristics and spatial trajectory characteristics are extracted from the dynamic three-dimensional point cloud sequence and fused to obtain a fused feature vector, and the energy, frequency and spatial motion characteristics of the dancing are comprehensively characterized by multi-dimensional features to avoid the limitations of a single feature; the fused feature vector is input into a bidirectional long short-term memory network enhanced by physical constraints, and the bidirectional long short-term memory network is used to capture the forward and backward correlation information of the time series data and to enhance physical constraints to ensure that the prediction conforms to the actual physical laws, thereby better predicting the dancing evolution trend and outputting a joint prediction matrix of the dancing amplitude and frequency; the joint prediction matrix of the dancing amplitude and frequency is input into a pre-trained risk propagation inference model, and the predicted dancing characteristics are converted into hierarchical quantitative risk indicators to achieve multi-level risk quantification; finally, a risk warning prompt is generated based on the multi-level risk quantification indicator and sent to the corresponding risk management equipment to complete the warning. In summary, the embodiments of the present invention acquire multi-physical field data, integrate the data and extract and fuse multi-dimensional features, use a physical constraint-enhanced bidirectional long short-term memory network to accurately predict the galloping evolution trend, and quantify multi-level risks through a risk propagation reasoning model. It is possible to accurately predict the galloping evolution trend of transmission lines, thereby accurately warning of transmission line galloping risks.

[0050] As one example of the above solution, the generating module is specifically configured to: Performing phase synchronization calibration based on the dancing resonance frequency on the motion acceleration data and video stream data in the multi-physics field dataset to obtain a time-aligned dataset; the multi-physics field dataset includes motion acceleration data, ambient wind speed data, and video stream data; Based on the time-aligned data set, a wire rigid body dynamics constraint model is applied to perform spatial coordinate compensation on the ambient wind speed data to obtain spatially corrected data; The spatial correction data is input into a three-dimensional reconstruction algorithm that integrates the bending stiffness parameters of the conductor to generate a dynamic three-dimensional point cloud sequence of the transmission line's galloping.

[0051] As one example of the above solution, the fusion module is specifically used to: Based on the conductor aerodynamic structure coupling equation, the dynamic three-dimensional point cloud sequence is subjected to vortex-induced vibration mode separation to obtain the dancing energy distribution characteristics; Applying dancing dominant frequency adaptive filtering to the dynamic three-dimensional point cloud sequence to obtain frequency spectrum features; Performing trajectory curvature correction on the dynamic three-dimensional point cloud sequence based on the catenary equation of the wire to obtain spatial trajectory features; The dancing energy distribution feature, the frequency spectrum feature and the spatial trajectory feature are fused to obtain the fused feature vector.

[0052] As one example of the above solution, the prediction module is specifically used to: Reorganizing the fused feature vector in time series dimension to generate a feature tensor sequence with a fixed time window length; Synchronously inputting the feature tensor sequence into the forward computing layer and the reverse computing layer of the bidirectional long short-term memory network, and outputting a forward dance evolution hidden state sequence and a reverse dance evolution hidden state sequence respectively; A physical rule library is constructed based on the displacement constraint equation of the wire suspension point and the aerodynamic torque balance condition, and the forward dancing evolution hidden state sequence and the reverse dancing evolution hidden state sequence are dynamically weighted and fused to generate a physical constraint-corrected dancing evolution feature matrix; The physical constraint-corrected dance evolution feature matrix is ​​mapped into a joint prediction matrix of the dance amplitude and frequency.

[0053] It should be noted that the system embodiment described above is merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the system embodiment provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art can understand and implement the present invention without inventive work.

[0054] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A transmission line galloping risk warning method based on multi-physics field fusion, characterized in that: include: Multidimensional Obtain a multiphysics dataset of the galloping process of a transmission line; Performing spatiotemporal alignment and three-dimensional posture reconstruction on the multi-physics field data set to generate a dynamic three-dimensional point cloud sequence of the transmission line dancing; Extracting dance energy distribution features, frequency spectrum features, and spatial trajectory features from the dynamic three-dimensional point cloud sequence, and fusing the three features to obtain a fused feature vector; Inputting the fused feature vector into a physical constraint enhanced bidirectional long short-term memory network to predict the galloping evolution trend, and outputting a joint prediction matrix of the galloping amplitude and frequency of the transmission line; Inputting the joint prediction matrix of the dancing amplitude and frequency into a pre-trained risk propagation inference model to calculate the multi-level risk quantification index of the transmission line; Generate corresponding risk warning prompts based on the multi-level risk quantification indicators and send them to corresponding risk management devices.

2. The transmission line galloping risk warning method based on multi-physics field fusion according to claim 1 is characterized in that: The step of performing spatiotemporal alignment and three-dimensional posture reconstruction on the multi-physics field dataset to generate a dynamic three-dimensional point cloud sequence of the power transmission line movement includes the following sub-steps: Performing phase synchronization calibration based on the dancing resonance frequency on the motion acceleration data and video stream data in the multi-physics field dataset to obtain a time-aligned dataset; the multi-physics field dataset includes motion acceleration data, ambient wind speed data, and video stream data; Based on the time-aligned data set, a wire rigid body dynamics constraint model is applied to perform spatial coordinate compensation on the ambient wind speed data to obtain spatially corrected data; The spatial correction data is input into a three-dimensional reconstruction algorithm that integrates the bending stiffness parameters of the conductor to generate a dynamic three-dimensional point cloud sequence of the transmission line's galloping.

3. The transmission line galloping risk warning method based on multi-physics field fusion according to claim 2 is characterized in that: The step of extracting the dancing energy distribution feature, the frequency spectrum feature, and the spatial trajectory feature from the dynamic three-dimensional point cloud sequence and fusing the three features to obtain a fused feature vector includes the following sub-steps: Based on the conductor aerodynamic structure coupling equation, the dynamic three-dimensional point cloud sequence is subjected to vortex-induced vibration mode separation to obtain the dancing energy distribution characteristics; Applying dancing dominant frequency adaptive filtering to the dynamic three-dimensional point cloud sequence to obtain frequency spectrum features; Performing trajectory curvature correction on the dynamic three-dimensional point cloud sequence based on the catenary equation of the wire to obtain spatial trajectory features; The dancing energy distribution feature, the frequency spectrum feature and the spatial trajectory feature are fused to obtain the fused feature vector.

4. The transmission line galloping risk warning method based on multi-physics field fusion according to claim 3 is characterized in that: The step of inputting the fused feature vector into a physical constraint enhanced bidirectional long short-term memory network to predict the galloping evolution trend and outputting a joint prediction matrix of the galloping amplitude and frequency of the transmission line comprises the following sub-steps: Reorganizing the fused feature vector in time series dimension to generate a feature tensor sequence with a fixed time window length; Synchronously inputting the feature tensor sequence into the forward computing layer and the reverse computing layer of the bidirectional long short-term memory network, and outputting a forward dance evolution hidden state sequence and a reverse dance evolution hidden state sequence respectively; A physical rule library is constructed based on the displacement constraint equation of the wire suspension point and the aerodynamic torque balance condition, and the forward dancing evolution hidden state sequence and the reverse dancing evolution hidden state sequence are dynamically weighted and fused to generate a physical constraint-corrected dancing evolution feature matrix; The physical constraint-corrected dance evolution feature matrix is ​​mapped into a joint prediction matrix of the dance amplitude and frequency.

5. The transmission line galloping risk warning method based on multi-physics field fusion according to claim 4 is characterized in that: The joint prediction matrix of the galloping amplitude and frequency is input into a pre-trained risk propagation inference model to calculate the multi-level risk quantification index of the transmission line, wherein the multi-level risk quantification index includes the probability of conductor dynamic tension exceeding the limit and the probability of insulation gap breakdown risk, including the following sub-steps: performing amplitude-frequency parameter decoupling on the joint prediction matrix of the dancing amplitude and frequency to separate and obtain a dancing amplitude component sequence and a dancing frequency component sequence; Inputting the dancing amplitude component sequence into a pre-trained risk propagation inference model to calculate the dynamic tension of the conductor, and combining it with the static initial tension value of the transmission line to calculate the dynamic tension time history data of the transmission line; Based on the calculated dynamic tension time history data and a preset conductor tension safety threshold, a proportion of durations of tension exceeding a limit is calculated to obtain a probability of the conductor dynamic tension exceeding a limit; The dancing amplitude component sequence and the dancing frequency component sequence are input into a pre-trained risk propagation inference model to perform insulation gap electric field analysis. Combined with the inherent insulation configuration parameters of the transmission line, the minimum air gap distance change and the local electric field strength distribution are calculated. Based on the calculation results, the insulation gap breakdown risk probability is obtained.

6. The transmission line galloping risk warning method based on multi-physics field fusion according to claim 5 is characterized in that: Generating corresponding risk warning prompts according to the multi-level risk quantification indicators and sending them to corresponding risk management devices includes the following sub-steps: Determining the risk type, risk location, and risk level based on the conductor dynamic tension exceeding probability and the insulation gap breakdown risk probability in the multi-level risk quantification indicators; Generate a structured risk warning prompt based on the risk level, the risk warning prompt including risk type, risk location and risk level information; The risk warning prompt is dynamically sent to a corresponding risk management device based on the risk location.

7. A transmission line galloping risk warning system based on multi-physics field fusion, characterized by: include: An acquisition module is used to obtain a multi-physics field dataset of a transmission line galloping process; A generation module, configured to perform spatiotemporal alignment and three-dimensional posture reconstruction on the multi-physics field data set to generate a dynamic three-dimensional point cloud sequence of the transmission line dancing; a fusion module, configured to extract dancing energy distribution features, frequency spectrum features, and spatial trajectory features from the dynamic three-dimensional point cloud sequence, and fuse the three features to obtain a fused feature vector; A prediction module, configured to input the fused feature vector into a physical constraint enhanced bidirectional long short-term memory network to predict the galloping evolution trend, and output a joint prediction matrix of the galloping amplitude and frequency of the transmission line; a calculation module, configured to input the joint prediction matrix of the galloping amplitude and frequency into a pre-trained risk propagation inference model to calculate the multi-level risk quantification index of the transmission line; The early warning module is used to generate corresponding risk early warning prompts according to the multi-level risk quantification indicators and send them to the corresponding risk management equipment.

8. The transmission line galloping risk warning system based on multi-physics field fusion according to claim 7 is characterized in that: The generation module is specifically used for: Performing phase synchronization calibration based on the dancing resonance frequency on the motion acceleration data and video stream data in the multi-physics field dataset to obtain a time-aligned dataset; the multi-physics field dataset includes motion acceleration data, ambient wind speed data, and video stream data; Based on the time-aligned data set, a wire rigid body dynamics constraint model is applied to perform spatial coordinate compensation on the ambient wind speed data to obtain spatially corrected data; The spatial correction data is input into a three-dimensional reconstruction algorithm that integrates the bending stiffness parameters of the conductor to generate a dynamic three-dimensional point cloud sequence of the transmission line's galloping.

9. The transmission line galloping risk warning system based on multi-physics field fusion according to claim 8, characterized in that: The fusion module is specifically used for: Based on the conductor aerodynamic structure coupling equation, the dynamic three-dimensional point cloud sequence is subjected to vortex-induced vibration mode separation to obtain the dancing energy distribution characteristics; Applying dancing dominant frequency adaptive filtering to the dynamic three-dimensional point cloud sequence to obtain frequency spectrum features; Performing trajectory curvature correction on the dynamic three-dimensional point cloud sequence based on the catenary equation of the wire to obtain spatial trajectory features; The dancing energy distribution feature, the frequency spectrum feature and the spatial trajectory feature are fused to obtain the fused feature vector.

10. The transmission line galloping risk warning system based on multi-physics field fusion according to claim 9, characterized in that: The prediction module is specifically used for: Reorganizing the fused feature vector in time series dimension to generate a feature tensor sequence with a fixed time window length; Synchronously inputting the feature tensor sequence into the forward computing layer and the reverse computing layer of the bidirectional long short-term memory network, and outputting a forward dance evolution hidden state sequence and a reverse dance evolution hidden state sequence respectively; A physical rule library is constructed based on the displacement constraint equation of the wire suspension point and the aerodynamic torque balance condition, and the forward dancing evolution hidden state sequence and the reverse dancing evolution hidden state sequence are dynamically weighted and fused to generate a physical constraint-corrected dancing evolution feature matrix; The physical constraint-corrected dance evolution feature matrix is ​​mapped into a joint prediction matrix of the dance amplitude and frequency.

Images