Ground wire pay-off tension self-adaptive control method and system

Abstract

The invention discloses an adaptive control method and system for paying-off tension of a ground wire, particularly relates to the technical field of automatic control, and is used for solving the problem of control misalignment caused by incapability of distinguishing and stripping multi-span coupling vibration interference in an existing tension control method. Firstly, high-frequency fluctuation components are identified by monitoring frequency band energy distribution of tension signals, then state space trajectory similarity and causal network topology of multi-point vibration signals are cooperatively analyzed, and local or global coupling of vibration sources is accurately judged; then, extracting a dominant vibration frequency and a phase difference, and separating a vibration interference component from the original tension signal through a phase compensation adaptive filter according to the dominant vibration frequency and the phase difference; and finally, subtracting the interference component from the original tension to obtain a real tension value for closed-loop adjustment of the output of the tension equipment. Accurate identification and dynamic compensation of vibration interference are achieved, and the control precision of the pay-off tension and the system stability are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology, and more specifically, to an adaptive control method and system for conductor and ground wire tension. Background Technology

[0002] In the construction of high-voltage transmission lines, the tension control of conductors and ground wires is a core aspect of ensuring construction quality and safety. Existing technologies typically use tension sensors installed at tensioning equipment to form a closed-loop control system, aiming to stabilize the tension of the conductors and ground wires at a set value. This control strategy based on single-point tension feedback plays a fundamental role in maintaining the static sag of the conductors and ground wires. The entire laying system involves traction equipment, tensioning equipment, and long-distance conductors and ground wires connected by multiple laying pulleys.

[0003] However, in actual field laying environments, conductors and ground wires are prone to vibration due to factors such as wind loads. This vibration can be transmitted and coupled between adjacent spans. The dynamic load formed by the coupled vibration of multiple spans will act on the tension sensing device in the tension control system, causing the detection signal to be mixed with non-real tension fluctuation components. Existing control methods are unable to distinguish and remove this vibration interference, causing the control system to adjust based on the distorted tension signal. This not only fails to effectively suppress real tension fluctuations but may also cause continuous oscillation of the system, ultimately affecting the accurate control and safety of conductor and ground wire sag during the laying process. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides an adaptive control method and system for conductor and ground wire tension to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: An adaptive control method for conductor and ground wire tension includes: S1. Acquire the raw tension signal output by the tension sensor and the vibration signal output by the vibration sensor set at different positions on the conductor. S2. Monitor the energy distribution of the original tension signal within the preset frequency band. When the energy appears simultaneously at multiple adjacent frequency points and continues to increase, it is determined that there is a high-frequency fluctuation component caused by multi-slot coupled vibration. S3. When it is determined that there are high-frequency fluctuation components, the similarity of the state space trajectories reconstructed from vibration signals at different locations and the network topology constructed by the causal flow between vibration signals are analyzed to determine whether the high-frequency fluctuation components originate from local vibration or global coupled vibration. S4. Based on the determination result that the high-frequency wave component originates from local vibration or global coupled vibration, extract the dominant vibration frequency and phase difference information corresponding to the high-frequency wave component from the vibration signal. S5. Based on the dominant vibration frequency and phase difference information, the vibration interference component is separated from the original tension signal using an adaptive filter; S6. Subtract the vibration interference component from the original tension signal to obtain the compensated true tension value, and adjust the output of the tension device according to the true tension value.

[0006] Furthermore, acquiring the raw tension signal output by the tension sensor and the vibration signal output by the vibration sensor located at different positions on the conductor includes: Determine the locations of multiple equally spaced measuring points on the conductor; Install a vibration sensor at each measuring point; The original tension signal output by the tension sensor and the vibration signal output by each vibration sensor are synchronously acquired at a preset sampling frequency. The vibration signals output from each vibration sensor are normalized to obtain vibration signals with a consistent amplitude range.

[0007] Furthermore, the energy distribution of the original tension signal within a preset frequency band is monitored. When the energy appears simultaneously at multiple adjacent frequency points and continues to increase, it is determined that there is a high-frequency fluctuation component caused by multi-span coupled vibration, including: The monitoring frequency band is determined based on the conductor span and material properties; The time spectrum is obtained by performing a windowed Fourier transform on the original tension signal; All frequency points whose energy exceeds the background noise threshold within the monitoring frequency band are identified as candidate frequency points; If there are at least three adjacent candidate frequency points and their energy maintains an increasing trend within a continuous time window, then it is determined that there is a high-frequency fluctuation component caused by multi-slot coupled vibration.

[0008] Furthermore, when high-frequency fluctuation components are determined to exist, the similarity of the state-space trajectories reconstructed from vibration signals at different locations and the network topology constructed from the causal flow between vibration signals are collaboratively analyzed to determine whether the high-frequency fluctuation components originate from local vibration or globally coupled vibration, including: The phase space of the normalized vibration signal is reconstructed to obtain the state space trajectory corresponding to each measuring point. Calculate the dynamic time warped distance between each pair of all state space trajectories, and obtain the average dynamic time warped distance; The causal flow direction intensity between different vibration signals is calculated based on the transfer entropy algorithm, and a causal flow direction matrix is ​​constructed. Generate a directed network graph based on the causal flow matrix, and calculate the standard deviation of the node degree distribution of each node in the directed network graph; If the average dynamic time warp distance is less than the similarity threshold and the standard deviation of the node degree distribution is lower than the network heterogeneity threshold, then the high-frequency fluctuation component is determined to originate from global coupling vibration; otherwise, the high-frequency fluctuation component is determined to originate from local vibration.

[0009] Furthermore, the causal flow intensity between different vibration signals is calculated based on the transfer entropy algorithm, and a causal flow matrix is ​​constructed. This includes: for each pair of normalized vibration signals, calculating the transfer entropy value from the historical sequence of one vibration signal to the future sequence of another vibration signal, and using the transfer entropy value as the causal flow intensity between the two vibration signals; traversing all vibration signal pairs, filling the causal flow intensity of each pair into the corresponding positions of the matrix, and constructing the causal flow matrix.

[0010] Furthermore, a directed network graph is generated based on the causal flow matrix, and the standard deviation of the node degree distribution of each node in the directed network graph is calculated, including: using the causal flow matrix as a directed weighted adjacency matrix, where each measurement point corresponds to a node, to construct a directed network graph; calculating the node degree of each node in the directed network graph; and calculating the standard deviation of the node degree distribution based on the node degree of all nodes.

[0011] Furthermore, based on the determination that the high-frequency wave component originates from local vibration or globally coupled vibration, the dominant vibration frequency and phase difference information corresponding to the high-frequency wave component are extracted from the vibration signal, including: When the determination result is global coupled vibration, the normalized vibration signal is subjected to spectrum analysis, and at least two frequencies with the highest energy are extracted as the dominant vibration frequencies. When the determination result is local vibration, the vibration signal output by the vibration sensor closest to the vibration source is subjected to spectrum analysis, and the frequency with the highest energy is extracted as the dominant vibration frequency. Using the location of the tension sensor as a reference point, the phase difference between the vibration signal output by each vibration sensor and the original tension signal at the dominant vibration frequency is calculated as the phase difference information.

[0012] Furthermore, based on the dominant vibration frequency and phase difference information, an adaptive filter is used to separate the vibration interference component from the original tension signal, including: A reference signal containing sine and cosine components is generated based on the dominant vibration frequency; Phase compensation is performed on the reference signal based on the phase difference information; The phase-compensated reference signal is used as the input to the adaptive filter, and the original tension signal is used as the desired signal. The coefficients of the adaptive filter are adjusted by the least mean square algorithm so that the output of the adaptive filter approximates the vibration disturbance component. The output of the adaptive filter is used as the vibration disturbance component separated from the original tension signal.

[0013] Furthermore, the vibration interference component is subtracted from the original tension signal to obtain the compensated true tension value, and the output of the tension device is adjusted according to the true tension value, including: The original tension signal is subtracted from the vibration disturbance component to obtain the compensated true tension value. Calculate the tension error between the actual tension value and the preset tension setting value; Based on the tension error, a proportional-integral-derivative control algorithm is used to generate control signals for the tension equipment. The tension control signal is output to the tension device to adjust the wire tension.

[0014] On the other hand, the present invention provides a conductor-to-ground wire tension adaptive control system, comprising: The signal acquisition module is used to acquire the raw tension signal output by the tension sensor and the vibration signal output by the vibration sensor set at different positions on the conductor; the signal acquisition module outputs the acquired vibration signal to the vibration determination module and the information extraction module respectively. The high-frequency identification module is used to monitor the energy distribution of the original tension signal within a preset frequency band. When the energy appears simultaneously at multiple adjacent frequency points and continues to increase, it is determined that there is a high-frequency fluctuation component caused by multi-slot coupled vibration. The vibration determination module is used to collaboratively analyze the similarity of the state space trajectories reconstructed from vibration signals at different locations and the network topology constructed from the causal flow between vibration signals when high-frequency fluctuation components are determined to exist, and to determine whether the high-frequency fluctuation components originate from local vibration or global coupled vibration. The information extraction module is used to extract the dominant vibration frequency and phase difference information corresponding to the high-frequency vibration component from the vibration signal based on the determination result that the high-frequency wave component originates from local vibration or global coupled vibration. The component separation module is used to separate the vibration interference component from the original tension signal using an adaptive filter based on the dominant vibration frequency and phase difference information. The tension output module is used to subtract the vibration interference component from the original tension signal to obtain the compensated true tension value, and to adjust the output of the tension device according to the true tension value.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. By collaboratively analyzing the spatiotemporal characteristics and causal relationships of multi-point vibration signals, the source and characteristics of vibration interference mixed in tension signals can be accurately identified. This not only effectively distinguishes between two different disturbance modes, local vibration and global coupled vibration, but also provides a key basis for subsequent interference separation by extracting accurate dominant vibration frequency and phase difference information. This interference analysis strategy based on vibration source characteristic identification significantly improves the perception accuracy of the real tension state under complex vibration environments.

[0016] 2. By employing phase-compensated adaptive filtering technology, the interference component synchronized with the vibration source can be dynamically separated from the original tension signal. This process enables real-time online compensation of the tension measurement value, allowing the control system to ultimately rely on the "true tension" that better reflects the actual stress state of the conductor and ground wire. This overcomes the signal distortion problem caused by vibration interference in the original control system, thereby significantly improving the accuracy of tension control and the stability of the system, and ensuring precise control of the conductor and ground wire sag during the laying process. Attached Figure Description

[0017] Figure 1 This is a flowchart of an adaptive control method for conductor and ground wire tension according to the present invention; Figure 2 This is a schematic diagram of the structure of an adaptive control system for conductor and ground wire tension according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: Figure 1 This invention provides an adaptive control method for conductor and ground wire tension, comprising: S1. Acquire the raw tension signal output by the tension sensor and the vibration signal output by the vibration sensor set at different positions on the conductor. S2. Monitor the energy distribution of the original tension signal within the preset frequency band. When the energy appears simultaneously at multiple adjacent frequency points and continues to increase, it is determined that there is a high-frequency fluctuation component caused by multi-slot coupled vibration. S3. When it is determined that there are high-frequency fluctuation components, the similarity of the state space trajectories reconstructed from vibration signals at different locations and the network topology constructed by the causal flow between vibration signals are analyzed to determine whether the high-frequency fluctuation components originate from local vibration or global coupled vibration. S4. Based on the determination result that the high-frequency wave component originates from local vibration or global coupled vibration, extract the dominant vibration frequency and phase difference information corresponding to the high-frequency wave component from the vibration signal. S5. Based on the dominant vibration frequency and phase difference information, the vibration interference component is separated from the original tension signal using an adaptive filter; S6. Subtract the vibration interference component from the original tension signal to obtain the compensated true tension value, and adjust the output of the tension device according to the true tension value.

[0020] To achieve precise adaptive control of the conductor tension, high-quality and consistent raw sensor data is first required. The specific implementation method of step S1 is described in detail below.

[0021] Determining the locations of multiple equally spaced measuring points on the conductor is fundamental to data acquisition. The process of determining these locations must consider the length of the entire laying section and the spatial resolution requirements for vibration monitoring. Specifically, the total length L of the laying section can be obtained by measuring the actual distance between the traction equipment and the tensioning equipment at the construction site. The number of measuring points N must balance economic efficiency with technical necessity; typically, N should be no less than 3, for example, 4, 5, or 6, to ensure effective capture of the spatial distribution characteristics of vibration along the conductor. The interval D between each measuring point can be calculated using the formula D=L / (N-1). The first measuring point is located at the nearest installable point on the conductor to the tensioning equipment, the last measuring point is located at the nearest installable point on the conductor to the traction equipment, and the remaining N-2 measuring points are arranged at equal intervals between these two points according to the calculated interval D. For example, when L is 500 meters and N is set to 4, then D equals 500 / (4-1)≈166.7 meters. The four measuring points are located at approximately 0 meters, 166.7 meters, 333.3 meters, and 500 meters away from the tension device, respectively. This evenly spaced distribution method aims to provide a uniform spatial sampling basis for subsequent analysis of the propagation of vibration waves.

[0022] After completing the measurement point location planning, a vibration sensor needs to be installed at each predetermined measurement point location. The vibration sensor should be a high-sensitivity inertial measurement unit or accelerometer suitable for the field environment and capable of effectively detecting the lateral or longitudinal vibration of the conductor. During installation, ensure that the vibration sensor housing is reliably fixed to the surface of the conductor to avoid relative sliding introducing measurement noise. For example, a special clamp can be used to tightly hold the sensor to the conductor. All vibration sensors should have consistent models and performance parameters to ensure consistent data acquisition. Tension sensors are usually integrated into tension devices to detect the real-time tension of the conductor; their output signal is the original tension signal.

[0023] During the data acquisition phase, the raw tension signal output from the tension sensor and the vibration signals output from each vibration sensor are synchronously acquired at a preset sampling frequency. The preset sampling frequency fs must adhere to the Nyquist sampling theorem, meaning fs must be at least twice the highest vibration frequency component to be analyzed. Considering that the conductor vibration frequency may reach tens of hertz, for example, if the highest analysis frequency is set to 50 hertz, then fs must be at least greater than 100 hertz. To retain sufficient margin, fs is typically set to 1000 hertz to ensure distortion-free capture of high-frequency fluctuation components. Synchronous acquisition means that the start time of data acquisition for all sensor signals is completely consistent with the sampling time interval. This can be achieved by triggering multiple synchronous acquisition channels with a unified sampling clock signal, thereby ensuring a strict time correspondence between all signals and laying the foundation for subsequent analysis of phase differences and causal relationships between signals. The acquisition process continues, and each signal data point is stored sequentially.

[0024] Due to slight differences in the installation location and response characteristics of each vibration sensor, the amplitude range of their original output vibration signals may vary. To eliminate the impact of amplitude differences on subsequent analyses such as trajectory similarity comparison, the vibration signals output by each acquired vibration sensor need to be normalized to obtain vibration signals with a consistent amplitude range. The normalization process employs a linear scaling method. Specifically, for a set of discrete-time sequence signal data output by any vibration sensor, the maximum absolute value Vmax of the sequence signal is first calculated. Then, each data point in the sequence is divided by the calculated maximum absolute value Vmax. After this processing, the amplitude range of the data sequence of each vibration sensor output signal is scaled to the interval [-1, 1].

[0025] This step ensures that all vibration signals are on the same amplitude range, allowing subsequent analysis to focus on the signal's shape, frequency, and phase characteristics, rather than its absolute amplitude. After processing, the resulting vibration signals with consistent amplitude ranges will be used, along with the synchronously acquired raw tension signals, as input data for subsequent analysis steps.

[0026] After successfully acquiring the synchronously collected raw tension signal and the vibration signal with the same amplitude range obtained after normalization, the next step is to monitor whether there are high-frequency fluctuation components in the raw tension signal caused by multi-span coupled vibration. The specific implementation method of step S2 is described in detail below.

[0027] Determining the monitoring frequency band based on the conductor span and material properties is a prerequisite for targeted frequency domain analysis. The basis for determining the monitoring frequency band is the fundamental vibration characteristics of the conductor. The first-order vertical vibration frequency of a single conductor span, i.e., the fundamental frequency fbase, can be estimated using the following formula: Where Lspan represents the typical span, in meters; T represents the average operating tension of the conductor / ground wire, in Newtons; and μ represents the mass per unit length of the conductor / ground wire, in kilograms per meter. Multi-span coupled vibrations will excite higher-order modes, whose frequency components will be distributed around integer multiples of the fundamental frequency. Therefore, the lower limit frequency flow of the monitoring band can be set slightly higher than the fundamental frequency fbase, for example, flow = 1.2 × fbase, to avoid including too many low-frequency tension fluctuations. The upper limit frequency fhigh of the monitoring band needs to cover a sufficient number of higher-order modes; for example, it can be set to fhigh = 10 × fbase. For example, if a typical span (Lspan) of a project is 300 meters, the tension (T) is 20,000 Newtons, and the linear density (μ) is 1.5 kg / m, the fundamental frequency (fbase) is calculated to be approximately 0.2 Hz according to the formula. Therefore, the monitoring frequency band can be set from 0.24 Hz to 2.0 Hz; where 0.24 Hz represents the lower limit frequency (flow) and 2.0 Hz represents the upper limit frequency (fhigh). It should be noted that this frequency band is an example; in actual applications, it needs to be calculated and determined based on specific line parameters. This monitoring frequency band will serve as the frequency range for subsequent energy analysis.

[0028] The original tension signal is subjected to a windowed Fourier transform to obtain its time spectrum, allowing observation of the frequency components changing over time. The windowed Fourier transform uses a sliding time window. First, a window function is selected, such as the Hanning window, with a length of Nwindows (sampling points). The corresponding time length Twindow = Nwindows / fs, where fs is the preset sampling frequency set in step S1. The window length Twindow should balance frequency and time resolution, typically requiring it to be much larger than the period of the lowest frequency component of interest; for example, Twindow can be 10 seconds. Adjacent windows overlap, for example, with an overlap ratio of 50%, to smooth the time-varying analysis results. For each time window, the window function is applied to the original tension signal data points within that time period. Then, a fast Fourier transform is performed on the windowed signal segment to obtain the complex spectrum of the signal within that time window. The square of the spectrum amplitude is the energy spectral density estimate of the signal within that time window. By moving the window sequentially and repeating the above process, the time spectrum of the original tension signal changing over time can be obtained. This time spectrum is a two-dimensional array, with the dimension being the number of frequency points multiplied by the time window number.

[0029] After obtaining the time spectrum, it is necessary to identify all frequency points with energy exceeding the background noise threshold as candidate frequencies within the determined monitoring band, from the lower limit frequency (flow) to the upper limit frequency (fhigh). The background noise threshold (Enoisethreshold) is determined as follows: During a period of stable system operation without significant external excitation, such as a windless nighttime period, a segment of the original tension signal is recorded, and its average energy spectral density within the entire monitoring band is calculated. This average value is then multiplied by K and set as the background noise threshold (Enoisethreshold), where K is a safety factor, for example, 3 to 5, determined based on common statistical significance levels. For each time window in the time spectrum, all discrete frequency points within the monitoring band are scanned, and those frequency points whose energy spectral density values ​​consistently exceed the Enoisethreshold are marked. These marked frequency points constitute the candidate frequency set within that time window. Candidate frequency points represent potential vibration frequency components with energy significantly higher than the background noise at that moment.

[0030] Finally, a determination is made based on the distribution of candidate frequencies and the trend of energy changes. If, within a consecutive preset time window, such as three time windows representing a continuous observation period, there are at least three candidate frequencies with adjacent frequency indices within the monitoring frequency band (i.e., they are continuous or closely spaced on the frequency axis), and the energy spectral density values ​​corresponding to these candidate frequencies increase in each subsequent time window compared to the previous time window, i.e., maintain an increasing trend), then it is determined that there is a high-frequency fluctuation component caused by multi-slot coupled vibration. This pattern of coordinated energy growth of multiple adjacent frequency points is a typical characteristic of modal coupling excitation in multi-slot systems, distinguishing it from single frequency component fluctuations. Once this determination condition is met, further analysis in step S3 is triggered; otherwise, monitoring in step S2 continues.

[0031] After step S2 determines that there is a high-frequency fluctuation component caused by multi-segment coupled vibration, it is necessary to further analyze the vibration source characteristics of this component. The specific implementation of step S3 is to collaboratively analyze the similarity of the state-space trajectories reconstructed from vibration signals at different locations, as well as the network topology constructed from the causal flow between vibration signals, thereby determining whether the high-frequency fluctuation component originates from local vibration or global coupled vibration.

[0032] Phase space reconstruction is performed on the vibration signals with consistent amplitude ranges obtained in step S1 to obtain the state space trajectories corresponding to each measurement point. The phase space reconstruction uses the time delay method. For each measurement point's vibration signal time series X, which contains M data points and is represented as X={X1,X2,...,XM}, two key parameters need to be determined: the embedding dimension m and the time delay τ. The embedding dimension m can be determined using the spurious nearest neighbor algorithm, which determines the minimum sufficient embedding dimension by gradually increasing the dimension and observing the proportion of spurious nearest neighbors decreasing to a sufficiently low level (e.g., below 5%). The time delay τ can be determined using the autocorrelation function method, i.e., selecting the delay time of the first zero-crossing point of the time series autocorrelation function. For a vibration signal of length M, the reconstructed state space trajectory consists of a series of m-dimensional vectors Yu, where Yu=[Xu,Xu+τ,Xu+2τ,...,Xu+(m-1)τ], and the value of u ranges from 1 to M-(m-1)τ. Each measurement point will obtain such a state space trajectory characterizing its dynamic state evolution.

[0033] The dynamic time warping distance (DTW) between each pair of state-space trajectories is calculated, and the average DTW is obtained. The dynamic time warping algorithm is used to measure the similarity between two time series of different lengths or with nonlinear deformation. Specifically, for any two state-space trajectories of measurement points A and B, the Euclidean distance matrix between their corresponding m-dimensional vectors is first calculated. Then, dynamic programming is used to find the optimal path from the upper left corner to the lower right corner of the Euclidean distance matrix, which must satisfy boundary conditions, monotonicity, and continuity constraints. The minimum cumulative distance on the optimal path is the DTW between these two state-space trajectories. All measurement point pairs are traversed, and the DTW between each pair is calculated. The total number of pairs is calculated using the combination formula; if the number of measurement points is 10, the total number of pairs is 10(10-1) / 2. Finally, all pairwise distances are summed and divided by the total number of pairs to obtain the average DTWmean, which reflects the overall similarity of vibration modes across all measurement points.

[0034] The causal flow intensity between different vibration signals is calculated based on the transfer entropy algorithm, and a causal flow matrix is ​​constructed. Transfer entropy is an information-theoretic metric used to measure the flow of information from one time series to another. For each pair of vibration signals with consistent amplitude ranges, denoted as signal X and signal Y, the transfer entropy from X to Y is calculated. Transfer entropy is defined as the reduction in uncertainty of the prediction of Y's future based on the history of X, given the history of Y itself. The specific calculation requires estimating the joint probability distribution. When using the histogram method, the data needs to be discretized into B intervals. The selection of B is based on the sample size, for example, by using the square root rule. .

[0035] The formula for calculating the transfer entropy is: ;in, It represents the transfer entropy value from signal X to signal Y, characterizing the intensity of information flow; express , and The joint probability; This represents the value of signal Y at a future time t+1; The historical embedding vector of signal Y contains historical values ​​traced back k time steps from time t; The historical embedding vector of signal X contains historical values ​​traced back l time steps from time t; Indicates that in the known and Under the conditions, The conditional probability of occurrence; Indicates that only known Under the conditions, The conditional probability of occurrence; k represents the dimension (time window length) of the historical embedding vector of signal Y; l represents the dimension (time window length) of the historical embedding vector of signal X.

[0036] Construct a matrix representing the number of vibration sensors, with columns indicating the number of sensors. The element in the i-th row and j-th column of this matrix represents the causal flow intensity from the i-th vibration signal to the j-th vibration signal. This matrix is ​​the causal flow matrix, which quantitatively describes the direction and intensity of vibration energy transfer between all measuring points.

[0037] A directed network graph is generated based on the causal flow matrix, and the standard deviation of the node degree distribution of each node in the directed network graph is calculated. The causal flow matrix is ​​regarded as a directed weighted adjacency matrix, where each measurement point corresponds to a node in the network graph, and the causal flow strength in the matrix represents the weight of the directed edge from the source node to the target node. To highlight significant causal connections, a threshold needs to be set to filter weak connections. The threshold can be set as the median or average of all non-zero causal flow strength values. Edges with weights greater than the threshold are retained, thus generating a simplified directed network graph. In this directed network graph, the node degree of each node is calculated. Node degree is divided into out-degree and in-degree. Here, the total node degree is calculated, which is the sum of the out-degree and in-degree of the node. The out-degree refers to the number of edges from the node to other nodes (based on the thresholded connections), and the in-degree refers to the number of edges from other nodes to the node. After obtaining the node degree values ​​of all nodes, the standard deviation of this set of node degree values ​​is calculated, denoted as σdegree. This standard deviation quantifies the dispersion of the node degree distribution in the network, that is, the heterogeneity of the network structure.

[0038] If the mean dynamic time warp distance (DTWmean) is less than the similarity threshold and the standard deviation of the node degree distribution (σdegree) is less than the network heterogeneity threshold, then the high-frequency fluctuation component is determined to originate from global coupled vibration; otherwise, the high-frequency fluctuation component is determined to originate from local vibration. The similarity threshold (θsimilarity) is set based on the following: on training or simulation data known to be globally coupled vibration, the mean dynamic time warp distance is calculated over multiple observations, and its upper statistical limit (e.g., the mean plus one standard deviation) is taken as θsimilarity. The network heterogeneity threshold (θheterogeneity) is set similarly, based on the upper statistical limit of the typical value of the standard deviation of the node degree distribution under a globally coupled vibration scenario. If both conditions are met simultaneously, i.e., DTWmean < θsimilarity and σdegree < θheterogeneity, it indicates that the vibration modes of all measurement points are highly similar and the causal network structure is uniform. This is a characteristic of globally coupled vibration, so the high-frequency fluctuation component is determined to originate from globally coupled vibration. If either condition is not met, such as low trajectory similarity or the existence of a significant central node in the network, i.e., a local vibration source, then the high-frequency fluctuation component is determined to originate from local vibration. This determination result will directly guide the extraction strategy of dominant vibration frequency and phase difference information in step S4.

[0039] After determining the source of the high-frequency wave component (global coupled vibration or local vibration) in step S3, the goal of step S4 is to accurately extract the dominant vibration frequency and phase difference information corresponding to the high-frequency wave component from the vibration signal based on the determination result, so as to provide key parameters for subsequent vibration interference separation.

[0040] When the determination result is global coupled vibration, spectral analysis is performed on the normalized vibration signal to extract at least two frequencies with the highest energy as the dominant vibration frequencies. Since global coupled vibration involves coordinated vibration across multiple spans, its energy is typically distributed across multiple modal frequencies. Specifically, the operation involves selecting data from the same time period where high-frequency fluctuations were determined to exist in step S2, and performing spectral analysis on the vibration signals output by each vibration sensor (corresponding to each measurement point location determined in step S1) with consistent amplitude ranges after normalization. The spectral analysis is implemented using Fast Fourier Transform.

[0041] For each vibration signal, its power spectral density is calculated. Then, the power spectral densities of all vibration sensors are averaged to obtain an average power spectral density curve. On this average power spectral density curve, all significant peaks within the monitoring frequency band determined in step S2 are identified. The criterion for a significant peak can be that its height exceeds a certain multiple of the average height of its two valleys, for example, 3 times. Then, from these significant peaks, the frequencies corresponding to the top P peaks with the largest power spectral density values ​​are selected, where P is an integer greater than or equal to 2, for example, P=2 or P=3. These selected frequencies are the dominant vibration frequencies under global coupled vibration, representing the most important vibration modes in the entire conductor-ground wire system.

[0042] When the vibration is determined to be localized, a spectral analysis is performed on the vibration signal output by the vibration sensor closest to the vibration source, and the frequency with the highest energy is extracted as the dominant vibration frequency. When the vibration is determined to be localized, the vibration source is usually located at a specific location, and the vibration sensor closest to the source receives the strongest signal, with minimal interference from vibrations from other parts. Determining the "vibration sensor closest to the vibration source" requires consideration of the analysis results from step S3.

[0043] In step S3, if the vibration is determined to be localized, its causal flow network topology will typically contain a central node with a significantly higher degree than other nodes. The location of the measuring point corresponding to this node can be identified as a potential local vibration source. The vibration sensor installed at this location is the closest vibration sensor to the vibration source. A spectral analysis is performed on the vibration signal output from this sensor, which has already undergone normalization and has a consistent amplitude range (power spectral density is calculated using Fast Fourier Transform). Within the monitoring frequency band determined in step S2, the frequency corresponding to the peak with the highest power spectral density is found. This single frequency is the dominant vibration frequency under localized vibration.

[0044] Using the location of the tension sensor as a reference point, the phase difference between the vibration signal output by each vibration sensor and the original tension signal at the dominant vibration frequency is calculated as phase difference information. Regardless of whether the vibration source is global or local, phase difference information is needed to characterize the relative delay of the vibration wave propagating to the tension sensor location. The location of the tension sensor is fixed, typically at the tensioning device. For each vibration sensor (i.e., each measuring point), the following steps are performed: First, extract the data of the vibration sensor signal and the original tension signal obtained in step S1 within the same time period. Then, bandpass filtering is applied to both the vibration signal and the original tension signal. The center frequency of the filter is sequentially set to each dominant vibration frequency extracted in the previous step, and the bandwidth must be sufficiently narrow to highlight the frequency component, for example, 5% of the center frequency. For each dominant vibration frequency fdom, the phase difference ΔΦ between the filtered vibration signal and the filtered original tension signal at that frequency is calculated.

[0045] The phase difference can be calculated using the Hilbert transform method: Perform Hilbert transforms on both the filtered vibration signal and the filtered original tension signal to obtain their respective analytic signals; the phase angle of the analytic signal is the instantaneous phase; calculate the difference between the two instantaneous phases at the same moment, and average (or take the mode of the main distribution area) the differences of this time series to obtain a stable phase difference estimate ΔΦ. This phase difference ΔΦ reflects the phase shift corresponding to the time delay experienced by the vibration propagating from the measuring point to the tension sensor position. This phase difference is calculated for each vibration sensor and each dominant vibration frequency, thus obtaining a complete set of phase difference information. This phase difference information will be used for phase compensation of the adaptive filter reference signal in step S5 to ensure accurate separation of interference components synchronized with the vibration source.

[0046] After successfully extracting the dominant vibration frequency and phase difference information in step S4, the core task of step S5 is to use this information to accurately separate the vibration interference component from the original tension signal through adaptive filtering technology.

[0047] A reference signal containing sine and cosine wave components is generated based on the dominant vibration frequencies. For each dominant vibration frequency extracted in step S4, a pair of orthogonal sine and cosine wave components are generated. The sine wave component is represented as sin(2π fdom t), and the cosine wave component is represented as cos(2π fdom t), where t is a continuous time variable. In actual digital signal processing, t is discretized as the sampling point index multiplied by Ts, where Ts is the reciprocal of the preset sampling frequency fs set in step S1, i.e., the sampling period. The sine and cosine wave components corresponding to all dominant vibration frequencies are combined in sequence to form the reference signal vector X(n). For example, if two dominant vibration frequencies fdom1 and fdom2 are extracted, the reference signal vector at sampling time n is X(n) = [sin(2π fdom1 n Ts), cos(2π fdom1 n Ts), sin(2π fdom2 n Ts), cos(2π fdom2 n Ts)]. This reference signal vector contains all possible phase information of the vibration components, providing a basis for subsequent filtering.

[0048] Phase compensation is performed on the reference signal based on the phase difference information. In step S4, the phase difference ΔΦ between each vibration sensor signal and the original tension signal at each dominant vibration frequency fdom is calculated. The purpose of phase compensation is to adjust the phase of the reference signal to be in phase with the vibration interference component that actually propagates to the tension sensor. For each dominant vibration frequency fdom, a representative phase difference ΔΦcomp needs to be selected for compensation. In the case of globally coupled vibration, the average phase difference of all measuring points at this frequency can be used as ΔΦcomp. In the case of local vibration, the phase difference calculated by the vibration sensor closest to the vibration source can be directly used as ΔΦcomp. Phase compensation is achieved by rotating the vector: for the orthogonal pair [sin(2π fdom n Ts), cos(2π fdom n Ts)] corresponding to the frequency fdom in the reference signal vector, it is multiplied by a rotation matrix. The rotation matrix is ​​defined as [[cos(ΔΦcomp), -sin(ΔΦcomp)], [sin(ΔΦcomp), cos(ΔΦcomp)]]. After multiplication, a new orthogonal pair after phase compensation is obtained [sin(2π fdom n Ts + ΔΦcomp), cos(2π fdom n Ts + ΔΦcomp)]. This operation is performed sequentially on all components of the dominant vibration frequency to obtain the final phase-compensated reference signal vector Xcomp(n).

[0049] The phase-compensated reference signal is used as the input to the adaptive filter, and the original tension signal is used as the desired signal. The adaptive filter adopts a transverse filter structure, and its tap weight vector is denoted as W(n), with the same dimension as the phase-compensated reference signal vector Xcomp(n). At each sampling time n, the output y(n) of the adaptive filter is calculated as the inner product of the tap weight vector W(n) and the phase-compensated reference signal vector Xcomp(n), i.e., y(n) = the transpose of W(n) multiplied by Xcomp(n). The original tension signal d(n) (i.e., the value of the original tension signal obtained in step S1 at sampling time n) is set as the desired signal of the adaptive filter. The goal of the adaptive filter is to adjust the weight vector W(n) so that the filter output y(n) approximates as closely as possible to the components related to the reference signal contained in the original tension signal d(n), i.e., the vibration disturbance component.

[0050] The least mean square (LMS) algorithm is used to adjust the coefficients of the adaptive filter, making the output of the adaptive filter approximate the vibration interference component. The LMS algorithm updates the filter weight vector W(n) iteratively to minimize the square of the instantaneous squared error e(n), where e(n) = d(n) - y(n). The update formula for the weight vector is W(n+1) = W(n) + 2μe(n)Xcomp(n), where μ is the step size parameter, a key parameter controlling the convergence speed and stability of the algorithm. The step size parameter μ must satisfy the stability condition; typically, its value should be less than the reciprocal of the largest eigenvalue of the autocorrelation matrix of the input signal Xcomp(n). In practical applications, a small positive value, such as μ = 0.001, can be determined experimentally to ensure smooth convergence of the algorithm. The error signal e(n) = d(n) - y(n) is calculated after each iteration and used for the next weight update. Through continuous iteration, the weight vector W(n) of the adaptive filter will gradually converge to a set of optimal values, so that the filter output y(n) can best approximate the vibration interference component in the original tension signal.

[0051] The output of the adaptive filter is used as the vibration interference component separated from the original tension signal. After the adaptive filter reaches a stable convergence state through sufficiently long iterations, its output y(n) at each sampling time n is considered an estimate of the vibration interference component separated from the original tension signal at that current time. The convergence criterion can be that the norm of the weight vector W(n) changes less than a preset minimum threshold in multiple consecutive iterations. This separated vibration interference component y(n) accurately reflects the dominant vibration frequency component determined in step S4, and its phase is compensated and synchronized with the actual interference acting on the tension sensor. This vibration interference component will be used as the direct input for tension compensation in step S6.

[0052] After successfully separating the vibration interference component from the original tension signal in step S5, the core task of step S6 is to complete the final compensation of the tension and realize the closed-loop control of the wire tension, thereby suppressing the impact of vibration interference on the tension control accuracy.

[0053] The original tension signal is subtracted from the vibration interference component to obtain the compensated true tension value. This step is a direct data calculation. At each sampling time p, the original tension signal obtained from step S1 is denoted as d(p), and the vibration interference component obtained from step S5 is denoted as y(p). The compensated true tension value s(p) is calculated by a simple scalar subtraction: s(p) = d(p) - y(p). This operation is performed continuously for each sampling point, resulting in a time-varying sequence of true tension values ​​that have compensated for the identified vibration interference. This true tension value s(p) is considered to be a more accurate estimate of the static or quasi-static tension actually borne by the conductor, eliminating high-frequency fluctuations caused by multi-span coupled vibrations or local vibrations.

[0054] The tension error between the actual tension value and the preset tension setpoint is calculated. The preset tension setpoint Tset is a constant determined in advance based on factors such as transmission line construction specifications, conductor and ground wire types, and laying speed, and its unit is kilonewtons. This value is usually input into the control system by the operator before the laying operation begins. The tension error et(w) is calculated once in each control cycle w (the control cycle can be the same as the sampling cycle or an integer multiple of the sampling cycle). The calculation method is to subtract the filtered actual tension value sfiltered(w) of the current control cycle from the preset tension setpoint Tset. This error value et(w) reflects the magnitude and direction of the deviation between the current actual tension and the expected tension, and is the direct basis for subsequent control decisions.

[0055] The proportional-integral-derivative (PID) control algorithm is used to generate the control signal for the tension equipment based on the tension error. The PID algorithm controls the quantity by weighted summing of the proportional, integral, and derivative terms of the tension error. The control algorithm executes once in each control cycle w.

[0056] The discrete position calculation formula for the proportional-integral-derivative (PID) control algorithm is: u(w) = Kp × et(w) + Ki × Tc × Σet(j) + (Kd / Tc) × [et(w) - et(w-1)]; where u(w) is the control quantity calculated in the w-th control cycle; et(w) is the tension error in the w-th control cycle; et(w-1) is the tension error in the (w-1)-th control cycle; Σet(j) represents the sum of tension errors from the 0th control cycle to the w-th control cycle; Kp is the proportional gain coefficient; Ki is the integral gain coefficient; Kd is the derivative gain coefficient; and Tc is the control cycle.

[0057] The proportional gain coefficient Kp reflects the deviation signal et(w) of the control system proportionally. A larger coefficient results in a faster system response, but an excessively large coefficient can lead to system instability. The integral gain coefficient Ki is used to eliminate the steady-state error of the system and improve its accuracy. The differential gain coefficient Kd reflects the changing trend of the deviation signal, providing predictability and enabling anticipatory control. These three coefficients need to be tuned according to the dynamic characteristics of the controlled object (tension equipment and conductor system). For example, they can be determined through on-site debugging using the Ziegler-Nichols tuning method or trial and error. The control period Tc needs to match the system's response speed and is usually much larger than the sampling period, for example, 100 to 500 milliseconds. The calculated control quantity u(w), after being limited (e.g., restricted to between 0 and 10 volts), becomes the control signal for the tension equipment.

[0058] The tensioning equipment outputs a control signal to the tensioning device to adjust the wire tension. The tensioning equipment control signal u(w) is typically an analog voltage or current signal, but may also be a digital communication protocol command. This signal is transmitted to the actuator of the tensioning device, such as a proportional valve in a hydraulic system or a motor driver, via the analog output channel of a data acquisition card or a dedicated control bus. The actuator linearly adjusts its output force or torque according to the magnitude of the received control signal. For example, in a hydraulic tensioner, an increase in the control signal increases the opening of the proportional valve, raises the hydraulic cylinder pressure, and increases the brake clamping force, thus increasing the wire tension; conversely, a decrease in the control signal also increases the wire tension. Through this continuous feedback control, the actual tension value s(p) will be automatically adjusted and stabilized near the preset tension setpoint Tset, thereby achieving high-precision adaptive control of the wire tension and effectively overcoming the effects of vibration interference. The entire control process, from signal acquisition, processing, compensation to final output, forms a complete closed loop.

[0059] Example 2: Figure 2 A schematic diagram of a conductor-to-ground wire tension adaptive control system according to the present invention is provided. The conductor-to-ground wire tension adaptive control system includes: The signal acquisition module is used to acquire the raw tension signal output by the tension sensor and the vibration signal output by the vibration sensor set at different positions on the conductor; the signal acquisition module outputs the acquired vibration signal to the vibration determination module and the information extraction module respectively. The high-frequency identification module is used to monitor the energy distribution of the original tension signal within a preset frequency band. When the energy appears simultaneously at multiple adjacent frequency points and continues to increase, it is determined that there is a high-frequency fluctuation component caused by multi-slot coupled vibration. The vibration determination module is used to collaboratively analyze the similarity of the state space trajectories reconstructed from vibration signals at different locations and the network topology constructed from the causal flow between vibration signals when high-frequency fluctuation components are determined to exist, and to determine whether the high-frequency fluctuation components originate from local vibration or global coupled vibration. The information extraction module is used to extract the dominant vibration frequency and phase difference information corresponding to the high-frequency vibration component from the vibration signal based on the determination result that the high-frequency wave component originates from local vibration or global coupled vibration. The component separation module is used to separate the vibration interference component from the original tension signal using an adaptive filter based on the dominant vibration frequency and phase difference information. The tension output module is used to subtract the vibration interference component from the original tension signal to obtain the compensated true tension value, and to adjust the output of the tension device according to the true tension value.

[0060] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0061] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0062] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0063] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0064] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0065] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0066] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0067] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0068] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0069] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for adaptive control of conductor tension during ground wire laying, characterized in that, include: S1. Acquire the raw tension signal output by the tension sensor and the vibration signal output by the vibration sensor set at different positions on the conductor. S2. Monitor the energy distribution of the original tension signal within the preset frequency band. When the energy appears simultaneously at multiple adjacent frequency points and continues to increase, it is determined that there is a high-frequency fluctuation component caused by multi-slot coupled vibration. S3. When it is determined that there are high-frequency fluctuation components, the similarity of the state space trajectories reconstructed from vibration signals at different locations and the network topology constructed by the causal flow between vibration signals are analyzed to determine whether the high-frequency fluctuation components originate from local vibration or global coupled vibration. S4. Based on the determination result that the high-frequency wave component originates from local vibration or global coupled vibration, extract the dominant vibration frequency and phase difference information corresponding to the high-frequency wave component from the vibration signal. S5. Based on the dominant vibration frequency and phase difference information, the vibration interference component is separated from the original tension signal using an adaptive filter; S6. Subtract the vibration interference component from the original tension signal to obtain the compensated true tension value, and adjust the output of the tension device according to the true tension value.

2. The adaptive control method for conductor tension according to claim 1, characterized in that, Acquire the raw tension signal output by the tension sensor and the vibration signals output by vibration sensors located at different positions on the conductor, including: Determine the locations of multiple equally spaced measuring points on the conductor; Install a vibration sensor at each measuring point; The original tension signal output by the tension sensor and the vibration signal output by each vibration sensor are synchronously acquired at a preset sampling frequency. The vibration signals output from each vibration sensor are normalized to obtain vibration signals with a consistent amplitude range.

3. The adaptive control method for conductor and ground wire tension according to claim 1, characterized in that, Monitor the energy distribution of the original tension signal within a preset frequency band. When the energy appears simultaneously at multiple adjacent frequency points and continues to increase, it is determined that there is a high-frequency fluctuation component caused by multi-span coupled vibration, including: The monitoring frequency band is determined based on the conductor span and material properties; The time spectrum is obtained by performing a windowed Fourier transform on the original tension signal; All frequency points whose energy exceeds the background noise threshold within the monitoring frequency band are identified as candidate frequency points; If there are at least three adjacent candidate frequency points and their energy maintains an increasing trend within a continuous time window, then it is determined that there is a high-frequency fluctuation component caused by multi-slot coupled vibration.

4. The adaptive control method for conductor tension according to claim 1, characterized in that, When high-frequency fluctuation components are determined to exist, the similarity of state-space trajectories reconstructed from vibration signals at different locations is analyzed collaboratively, along with the network topology constructed from the causal flow between vibration signals. This determines whether the high-frequency fluctuation components originate from local vibration or globally coupled vibration, including: The phase space of the normalized vibration signal is reconstructed to obtain the state space trajectory corresponding to each measuring point. Calculate the dynamic time warped distance between each pair of all state space trajectories, and obtain the average dynamic time warped distance; The causal flow direction intensity between different vibration signals is calculated based on the transfer entropy algorithm, and a causal flow direction matrix is ​​constructed. Generate a directed network graph based on the causal flow matrix, and calculate the standard deviation of the node degree distribution of each node in the directed network graph; If the average dynamic time warp distance is less than the similarity threshold and the standard deviation of the node degree distribution is lower than the network heterogeneity threshold, then the high-frequency fluctuation component is determined to originate from global coupling vibration; otherwise, the high-frequency fluctuation component is determined to originate from local vibration.

5. The adaptive control method for conductor and ground wire tension according to claim 4, characterized in that, The causal flow intensity between different vibration signals is calculated based on the transfer entropy algorithm, and a causal flow matrix is ​​constructed. This includes: for each pair of normalized vibration signals, calculating the transfer entropy value from the historical sequence of one vibration signal to the future sequence of another vibration signal, and using the transfer entropy value as the causal flow intensity between the two vibration signals; traversing all vibration signal pairs, filling the corresponding positions of the matrix with the causal flow intensity of each pair, and constructing the causal flow matrix.

6. The adaptive control method for conductor and ground wire tension according to claim 4, characterized in that, A directed network graph is generated based on the causal flow matrix. The standard deviation of the node degree distribution of each node in the directed network graph is calculated, including: using the causal flow matrix as a directed weighted adjacency matrix, where each measurement point corresponds to a node, to construct a directed network graph; calculating the node degree of each node in the directed network graph; and calculating the standard deviation of the node degree distribution based on the node degree of all nodes.

7. The adaptive control method for conductor tension according to claim 1, characterized in that, Based on the determination that the high-frequency wave component originates from local vibration or globally coupled vibration, the dominant vibration frequency and phase difference information corresponding to the high-frequency wave component are extracted from the vibration signal, including: When the determination result is global coupled vibration, the normalized vibration signal is subjected to spectrum analysis, and at least two frequencies with the highest energy are extracted as the dominant vibration frequencies. When the determination result is local vibration, the vibration signal output by the vibration sensor closest to the vibration source is subjected to spectrum analysis, and the frequency with the highest energy is extracted as the dominant vibration frequency. Using the location of the tension sensor as a reference point, the phase difference between the vibration signal output by each vibration sensor and the original tension signal at the dominant vibration frequency is calculated as the phase difference information.

8. The adaptive control method for conductor and ground wire tension according to claim 1, characterized in that, Based on the dominant vibration frequency and phase difference information, the vibration interference component is separated from the original tension signal using an adaptive filter, including: A reference signal containing sine and cosine components is generated based on the dominant vibration frequency; Phase compensation is performed on the reference signal based on the phase difference information; The phase-compensated reference signal is used as the input to the adaptive filter, and the original tension signal is used as the desired signal. The coefficients of the adaptive filter are adjusted by the least mean square algorithm so that the output of the adaptive filter approximates the vibration disturbance component. The output of the adaptive filter is used as the vibration disturbance component separated from the original tension signal.

9. The adaptive control method for conductor and ground wire tension according to claim 1, characterized in that, The compensated true tension value is obtained by subtracting the vibration interference component from the original tension signal, and the output of the tension device is adjusted according to the true tension value, including: The original tension signal is subtracted from the vibration disturbance component to obtain the compensated true tension value. Calculate the tension error between the actual tension value and the preset tension setting value; Based on the tension error, a proportional-integral-derivative control algorithm is used to generate control signals for the tension equipment. The tension control signal is output to the tension device to adjust the wire tension.

10. A conductor-to-ground wire tension adaptive control system, used to implement the conductor-to-ground wire tension adaptive control method according to any one of claims 1-9, characterized in that, include: The signal acquisition module is used to acquire the raw tension signal output by the tension sensor and the vibration signal output by the vibration sensor set at different positions on the conductor; the signal acquisition module outputs the acquired vibration signal to the vibration determination module and the information extraction module respectively. The high-frequency identification module is used to monitor the energy distribution of the original tension signal within a preset frequency band. When the energy appears simultaneously at multiple adjacent frequency points and continues to increase, it is determined that there is a high-frequency fluctuation component caused by multi-span coupled vibration. The vibration determination module is used to collaboratively analyze the similarity of the state space trajectories reconstructed from vibration signals at different locations and the network topology constructed from the causal flow between vibration signals when high-frequency fluctuation components are determined to exist, and to determine whether the high-frequency fluctuation components originate from local vibration or global coupled vibration. The information extraction module is used to extract the dominant vibration frequency and phase difference information corresponding to the high-frequency vibration component from the vibration signal based on the determination result that the high-frequency wave component originates from local vibration or global coupled vibration. The component separation module is used to separate the vibration interference component from the original tension signal using an adaptive filter based on the dominant vibration frequency and phase difference information. The tension output module is used to subtract the vibration interference component from the original tension signal to obtain the compensated true tension value, and to adjust the output of the tension device according to the true tension value.

Images