A Disaster Resistance Assessment Method for Transmission Towers in Near-Fault Regions Considering Wind-Induced Fatigue Effect

Abstract

This invention relates to the field of disaster resistance assessment technology for transmission towers, and discloses a method for assessing the disaster resistance performance of transmission towers in near-fault areas considering wind-induced fatigue effects. The method includes acquiring wind-induced fatigue data of transmission towers in near-fault areas, including wind speed data, structural vibration data, and fatigue damage data; then identifying wind-induced fatigue anomalies through wind-induced fatigue baselines, single-mode anomalies, and secondary verification; subsequently, extracting features from the wind-induced fatigue data, assigning timestamped labels to the extracted features, and inputting them into an assessment unit; the assessment unit analyzes and judges the data using multi-level assessment progression, and outputs disaster resistance performance assessment results; a dynamic optimization unit receives the assessment results, adjusts the assessment progression step size in real time, and determines whether the transmission tower structure needs reinforcement. This method can accurately capture the dynamic process of wind-induced fatigue, improve the accuracy and timeliness of disaster resistance performance assessment, and provide support for transmission tower maintenance.

Description

Technical Field

[0001] This invention relates to the field of disaster resistance assessment technology for transmission towers, specifically a method for assessing the disaster resistance performance of transmission towers in near-fault areas considering wind-induced fatigue effects. Background Technology

[0002] Due to their unique geological structure, near-fault regions often experience complex wind environments. Transmission towers subjected to long-term wind loads in such environments are prone to wind-induced fatigue, leading to structural damage and even collapse, posing a serious threat to the safe and stable operation of the power system. Currently, while a certain technical system has been established for assessing the disaster resistance performance of transmission towers, existing methods still have significant shortcomings in the specific scenario of near-fault regions.

[0003] Traditional assessment methods largely rely on static load calculations, neglecting the dynamic characteristics of wind loads and the cumulative fatigue effects over long periods. In near-fault regions, wind speeds fluctuate dramatically and are accompanied by turbulence, causing continuous vibrations in transmission tower structures. These vibrations lead to repeated stress alternations within the components, gradually resulting in fatigue damage. Existing methods often fail to accurately capture this dynamic fatigue process, relying solely on structural parameters at a single time point, which is insufficient to reflect changes in the structure's disaster resistance throughout its service life.

[0004] Existing assessment systems lack effective anomaly detection mechanisms. In the early stages of wind-induced fatigue damage, changes in structural parameters are often subtle and difficult to identify using traditional monitoring methods, leading to the neglect of anomalies until damage accumulates to a certain extent and causes significant faults. Only then are repair measures taken, at which point not only are maintenance costs greatly increased, but there is also the potential for severe consequences such as power outages. Furthermore, existing assessment processes lack effective linkage between data processing and assessment models. The extracted wind-induced fatigue characteristics are difficult to accurately match real-time assessment requirements, and the fixed analysis step size of the assessment unit cannot be dynamically adjusted according to actual wind conditions and structural responses. This limits the timeliness and accuracy of assessment results, making it difficult to meet the needs of real-time monitoring and early warning of the disaster resistance performance of transmission towers in near-fault areas. Summary of the Invention

[0005] The purpose of this invention is to provide a disaster resistance performance evaluation method for transmission towers in near-fault areas considering wind-induced fatigue effects, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a method for evaluating the disaster resistance performance of transmission towers in near-fault areas considering wind-induced fatigue effects, the method comprising:

[0007] Acquire wind-induced fatigue data of transmission towers in the near-fault area, the wind-induced fatigue data including wind speed data, structural vibration data and fatigue damage data;

[0008] The determination of wind-induced fatigue anomalies includes wind-induced fatigue baseline, single-modal anomaly, and secondary verification.

[0009] Feature extraction is performed on the wind-induced fatigue data, and the extracted wind-induced fatigue features are assigned with timestamp labels and input into the evaluation unit;

[0010] The evaluation unit sets up a multi-level evaluation process to analyze and judge the wind-induced fatigue data and output the disaster resistance performance evaluation results.

[0011] A dynamic optimization unit that adjusts the evaluation step size of the evaluation unit in real time receives the disaster resistance performance evaluation results and determines whether the transmission tower structure needs to be reinforced.

[0012] Preferably, the judgment of wind-induced fatigue anomaly includes a wind-induced fatigue baseline for comparing whether the acquired wind speed data, structural vibration data and fatigue damage data are within the normal threshold range of wind-induced fatigue. If the wind speed data, structural vibration data and fatigue damage data are not within the normal threshold range of wind-induced fatigue, the judgment of wind-induced fatigue anomaly is triggered.

[0013] The single-mode anomaly is used to map whether there are abrupt changes, imbalances, and decouplings in wind speed data, structural vibration data, and fatigue damage data. If any of the wind speed data, structural vibration data, and fatigue damage data has a single data anomaly, a single wind-induced fatigue anomaly data report is output, and the data acquisition unit is checked for anomalies.

[0014] Preferably, the secondary verification includes temporal correlation verification and spatial propagation verification. The temporal correlation verification is used to establish the mutation, imbalance and decoupling of continuous timestamp wind speed data, and the spatial propagation verification is used to obtain the mutation, imbalance and decoupling of structural vibration data and fatigue damage data.

[0015] If the data mapped in the temporal correlation verification and spatial propagation verification is abnormal, the grid density is dynamically optimized by dividing the grid into high-resolution grids, extracting local neighborhood features.

[0016] Preferably, the feature extraction of wind-induced fatigue data includes performing dynamic fluctuation feature extraction processing on wind speed data to obtain wind speed fluctuation feature vector, performing structural response feature analysis processing on structural vibration data to obtain structural response feature vector, and performing cross-modal feature fusion processing on the wind speed fluctuation feature vector and the structural response feature vector to generate a fused wind-induced feature set.

[0017] The cross-modal feature fusion processing includes time dimension alignment processing, which maps the wind speed fluctuation feature vector to the same time granularity as the structural response feature vector, and constructs a wind speed-structure cross-attention mechanism to calculate the cross-modal correlation matrix between the fluctuation feature of each time step in the wind speed fluctuation feature vector and the response feature of the corresponding time step in the structural response feature vector.

[0018] Preferably, the evaluation unit includes a wind speed evaluation shift unit, a structural vibration evaluation shift unit, and a fatigue damage evaluation shift unit;

[0019] The wind speed assessment and shift unit takes wind speed features with timestamp labels as input, and obtains the difference between the wind speed features of the previous moment and the wind speed features of the current moment and the difference between the wind speed features of the next moment and the wind speed features of the current moment by real-time monitoring of wind speed features within a continuous time stamp, triggering anomaly marking within a continuous time period.

[0020] The structural vibration assessment and shift unit sets up a structural vibration prediction model, predicts the structural vibration data at the next moment through the structural vibration prediction model, and marks the structural risk through mutation detection.

[0021] Preferably, the dynamic optimization unit receives the disaster resistance performance assessment results, drives the wind speed assessment shift unit to receive the wind speed change rate within the continuous time stamp, and adjusts the wind speed parameter assessment shift step size by using the wind speed change rate within the continuous time stamp. If the wind speed data change rate per unit time is greater than the highest weight, the wind speed parameter assessment shift step size is shortened.

[0022] The step size is adjusted by adjusting the acceleration of the vibration change of the continuous timestamp structure. If the acceleration of the vibration change of the timestamp structure from time t-1 to time t is greater than the acceleration of the vibration change of the continuous timestamp structure from time t-2 to time t-1, and the acceleration increase is greater than the minimum threshold, then the step size ratio is compressed.

[0023] The fatigue damage signal is obtained by the fatigue damage assessment shift unit, and the step size is adjusted by comparing it with the damage feature library. If the damage signal captured by the fatigue damage spectrum is crack propagation and material degradation, the damage shift step size of the fatigue damage assessment shift unit is adjusted.

[0024] Preferably, a pre-trained evaluation network model is invoked to perform fatigue root cause weight allocation processing on the fused wind-induced feature set to generate an abnormal correlation score set corresponding to wind-induced fatigue, wherein the abnormal correlation score set contains the abnormal correlation distribution between fatigue damage data and various structural response features.

[0025] Based on the abnormal correlation distribution, joint root cause tracing processing of fatigue damage data and structural response feature vectors is performed to generate abnormal root cause tracing results of wind-induced fatigue.

[0026] Preferably, the step size boundary is set to receive the adjusted damage push step size, the structural vibration compression step size ratio, and the adjusted wind speed assessment push step size, and an intensity level is set, which includes low disturbance and medium-high disturbance.

[0027] By adjusting the step size boundary, the redundant boundary buffer layer is triggered to temporarily tolerate the step size adjustment. If the step size briefly exceeds the rated boundary, the actual total number of adjusted steps will not be greater than the maximum number of steps in the redundant boundary buffer layer. If the buffer layer stays for more than the threshold, it will immediately switch to the minimum safe step size.

[0028] Preferably, if the wind speed change rate, structural vibration gradient change rate, and fatigue damage spectrum characteristics of the evaluation unit are in the stable domain of the redundant boundary buffer layer after adjustment, the step size of the evaluation unit evaluation shift will be adjusted according to real-time data.

[0029] If the total compensation amount for adjusting the step size of the evaluation unit is not greater than the maximum step size of the redundant boundary buffer layer, and the buffer layer dwell time is greater than the threshold, then step size freezing is triggered.

[0030] Preferably, a disaster resilience optimization strategy is generated based on the root cause tracing results, and the disaster resilience optimization strategy is fed back to the evaluation unit to trigger parameter calibration operation;

[0031] If it is determined that the transmission tower structure needs to be reinforced, the reinforcement of the transmission tower structure connectors and foundation reinforcement components is triggered, the damper is activated, and the balancer is triggered. The balancer moves the load upward in the longitudinal and lateral directions to maintain a near-stable state, and finally completes the reinforcement of the transmission tower structure.

[0032] If the transmission tower structure does not require reinforcement, the damper is activated, and the balancer is triggered. The balancer moves the load upward in the longitudinal and lateral directions to maintain a near-stable state, triggering electrical isolation between the transmission tower structure and the foundation.

[0033] Compared with the prior art, the beneficial effects of the present invention are:

[0034] This disaster resistance performance assessment method for transmission towers in near-fault areas, considering wind-induced fatigue effects, firstly achieves multi-dimensional coverage of wind-induced fatigue-related information by comprehensively acquiring wind speed data, structural vibration data, and fatigue damage data. This overcomes the limitations of traditional assessment methods that rely on single and incomplete data collection, providing a sufficient data foundation for subsequent accurate assessments. Compared to traditional assessment methods that rely solely on static parameters, this method fully incorporates dynamic wind load data and structural vibration data, enabling a more realistic reflection of the correlation between the wind field environment in near-fault areas and the structural response of the transmission tower, avoiding assessment biases caused by neglecting dynamic factors.

[0035] In the wind-induced fatigue anomaly assessment stage, a multi-level assessment mechanism—comprising a wind-induced fatigue baseline, single-modal anomaly assessment, and secondary verification—effectively improves the accuracy and reliability of anomaly identification. The establishment of the wind-induced fatigue baseline provides a reasonable reference for anomaly assessment; single-modal anomaly assessment can quickly capture abnormal changes in a single parameter dimension; and secondary verification further reviews the initially identified anomalies, reducing misjudgments caused by accidental factors. This multi-level anomaly assessment system solves the problems of untimely and inaccurate anomaly identification in traditional assessment methods, enabling the detection of structural anomalies in the early stages of fatigue damage. This creates conditions for timely intervention and prevents further damage accumulation leading to more serious structural failures.

[0036] In terms of data processing and evaluation unit operation, the extracted wind-induced fatigue features are assigned timestamped labels, closely linking the feature data with the time dimension. This facilitates the evaluation unit in tracing changes in fatigue status at different time points and clearly presenting the cumulative process of fatigue damage. The multi-level evaluation shift mechanism set up by the evaluation unit can analyze and judge wind-induced fatigue data in stages and levels, rather than using the traditional fixed-mode one-time evaluation. This approach can more meticulously analyze the trend of structural disaster resistance performance changes over different time periods, improving the granularity of the evaluation results. At the same time, the introduction of the dynamic optimization unit can adjust the evaluation shift step size in real time based on the disaster resistance performance evaluation results output by the evaluation unit. When the evaluation results show that the structural state is stable, the step size can be appropriately increased to improve evaluation efficiency; when potential risks or abnormal trends are found in the structure, the step size can be decreased to enhance the sensitivity of the evaluation, achieving a dynamic balance between evaluation efficiency and accuracy, avoiding the problem of difficulty in balancing efficiency and accuracy in traditional fixed-step evaluation.

[0037] This method uses a dynamic optimization unit to determine whether the transmission tower structure needs reinforcement based on the evaluation results, forming a complete closed loop of "data acquisition—anomaly detection—feature extraction—evaluation and analysis—reinforcement decision." This closed-loop evaluation process ensures that disaster resistance performance evaluation is no longer an isolated data analysis process, but is directly linked to structural maintenance decisions. It can guide actual structural reinforcement work in a timely manner based on the evaluation results, avoiding the disconnect between traditional evaluation and actual maintenance, ensuring the pertinence and timeliness of maintenance measures, effectively extending the service life of transmission towers, and guaranteeing the safe and stable operation of the power system in near-fault areas. Attached Figure Description

[0038] Figure 1 This is a schematic diagram illustrating the working principle of the disaster resistance performance evaluation method for transmission towers in near-fault areas considering wind-induced fatigue effects as described in this invention.

[0039] Figure 2 A flowchart detailing the near-anomaly detection method;

[0040] Figure 3 A flowchart of the data feature extraction method;

[0041] Figure 4 This is a flowchart of the method for tracing the root causes of anomalies. Detailed Implementation

[0042] 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.

[0043] Please see Figure 1 This invention provides a method for evaluating the disaster resistance performance of transmission towers in near-fault areas considering wind-induced fatigue effects, the method comprising:

[0044] The system acquires wind-induced fatigue data of transmission towers in the near-fault area, including wind speed, structural vibration, and fatigue damage data. It then performs wind-induced fatigue anomaly assessment, which includes comparison with wind-induced fatigue baselines, single-mode anomaly identification, and secondary verification. Features are extracted from the wind-induced fatigue data, timestamped, and input into the evaluation unit. The evaluation unit analyzes and assesses the data through multi-level evaluation progression, outputting disaster resistance performance assessment results. The dynamic optimization unit receives these assessment results, adjusts the evaluation progression step size of the evaluation unit in real time, and determines whether the transmission tower structure requires reinforcement.

[0045] Example 1: See Figure 2 In the disaster resistance performance assessment of transmission towers near fault zones, the initial judgment of wind-induced fatigue anomalies is the fundamental step in the entire process. This judgment process relies on a multi-level, multi-modal verification system, the core of which lies in the comprehensive review and cross-verification of three types of data: wind speed, structural vibration, and fatigue damage.

[0046] Establishing a wind-induced fatigue baseline serves as the initial threshold for determining whether data is abnormal. This baseline is not a single value, but rather a dynamically changing set of threshold ranges, determined comprehensively based on historical meteorological data of the transmission tower's location, the tower's structural design parameters, and the fatigue characteristics of the materials. For example, for wind speed data, the normal threshold range considers the region's average wind speed, the probability distribution of maximum wind speed, and seasonal variations in wind direction; for structural vibration data, the threshold range is determined based on the tower's natural frequency, damping ratio, and the expected response spectrum under design wind load; and the threshold for fatigue damage data is closely related to the material's SN curve, cumulative damage model, and stress concentration at monitoring points. The system compares the raw data acquired in real time with these preset threshold ranges. Once a data item, such as the instantaneous wind speed at a certain height, the vibration acceleration amplitude of a specific mode of the tower, or the stress cycle count at a critical node, is found to continuously or momentarily exceed its corresponding normal threshold upper or lower limit, the system triggers the wind-induced fatigue anomaly judgment process, indicating that the data has entered a state requiring further verification.

[0047] The task at this stage is to deeply analyze whether there are abnormal patterns within a single data stream, mainly including three types: abrupt changes, imbalances, and decoupling. Abrupt changes refer to a sudden, discontinuous jump in the data over time. For example, a wind speed value recorded by an anemometer may experience a steep pulse within a very short period, an acceleration signal acquired by a structural vibration sensor may show a sharp peak, or the stress amplitude recorded by a fatigue damage monitoring unit may undergo a step change. The system captures such events by calculating the first-order difference (rate of change) and second-order difference (acceleration) of the data in real time and setting a reasonable abrupt change threshold. Imbalances refer to inconsistencies between different measurement points of the same type of data, or between the data and the expected theoretical relationship. For example, the values ​​measured by multiple anemometers installed at different heights on a transmission tower should normally exhibit a certain gradient distribution. If the wind speed value at a certain layer deviates significantly from this gradient pattern and shows a significant divergence from the data at adjacent layers, it is judged as an imbalance. For example, under a specific wind speed, the vibration response of the tower should theoretically be within a certain order of magnitude. If the measured vibration amplitude is much higher or lower than the theoretical expectation, it is also a manifestation of imbalance. Decoupling refers to the significant weakening or disappearance of the correlation between data items that should originally be highly correlated. For example, under wind load, the time series of wind speed and the time series of stress in the main members of the tower usually have a strong correlation. If a significant change in wind speed is detected, but the stress response does not change accordingly, or vice versa, it means that there may be a decoupling phenomenon between wind speed and structural response. Once any of the above-mentioned single-mode anomalies is identified in any data stream, the system will immediately generate a single wind-induced fatigue anomaly data report, recording in detail the anomaly type, occurrence time, location, and severity. At the same time, the system will initiate a self-check program to detect whether there are hardware problems such as faults, drift, or signal interruptions in the data acquisition units (such as sensors, transmission lines, and data acquisition cards) to eliminate false anomalies caused by faults in the measurement system itself.

[0048] To further improve the accuracy of anomaly detection and prevent false alarms, the system introduces a two-level verification mechanism after single-mode identification, including temporal correlation verification and spatial propagation verification. Temporal correlation verification focuses on analyzing the continuity and development of anomaly patterns from a temporal perspective. It doesn't view isolated abrupt changes, imbalances, or decouplings of a single data point, but rather places them within a continuous time-stamp sequence to examine whether the anomaly pattern is temporally correlated. For example, an instantaneous wind speed pulse might just be a measurement disturbance, but if continuous, high-frequency abrupt changes in wind speed occur over multiple consecutive timestamps, or if the imbalance persists and gradually worsens, or if decoupling spreads over time, the authenticity and severity of the anomaly are greatly enhanced. Spatial propagation verification examines the expansion of anomalies from a structural spatial perspective. For structural vibration and fatigue damage data, anomalies may not be limited to an isolated measurement point. For example, after a high-stress anomaly occurs at a node, will this anomaly propagate along the component to adjacent nodes? Will a local vibration anomaly mode excite other modes of the overall structure? Spatial propagation verification determines whether anomalies are localized or pose a global risk by analyzing the spatiotemporal correlation patterns of anomalies in sensor data from different locations. The system achieves this verification by constructing a topological network of all measurement points and analyzing the propagation paths and velocities of anomalies within this network.

[0049] When both time-series correlation verification and spatial propagation verification indicate that the anomaly does exist and has a spatiotemporal correlation, the system initiates high-resolution grid analysis. This grid covers the entire transmission tower structure and its surrounding wind field space. The system dynamically adjusts the grid density based on the spatiotemporal characteristics of the anomaly signal. In areas where the anomaly signal is strong or changes drastically, such as near nodes with sudden increases in vibration amplitude or wind field layers with extreme wind speed gradients, the grid is automatically densified for more refined local neighborhood feature extraction. This extraction is achieved using sliding window analysis technology within the densified grid. Data within the window is used for feature extraction to more accurately locate the anomaly source and quantify its intensity. The purpose of the entire process is to transform the initial anomaly alarm into a precise judgment of the anomaly's nature, location, severity, and potential impact through in-depth verification and analysis, providing high-quality, high-reliability anomaly input information for subsequent assessment units.

[0050] Example 2: See Figure 3The system employs dynamic fluctuation feature extraction for wind speed data processing. Because wind fields near fault zones often exhibit strong non-stationarity and random pulsation characteristics, simple average or maximum wind speed values ​​are insufficient to describe their dynamic load characteristics. Therefore, the system performs in-depth analysis of real-time acquired wind speed time series to capture their fluctuation details. For example, for a wind speed record under sustained strong winds, the processing not only extracts the average wind speed and standard deviation but also focuses on analyzing its turbulence intensity, gust factor, power spectral density characteristics of pulsating wind speeds, and the frequency and amplitude of wind direction changes. These parameters collectively constitute a multidimensional wind speed fluctuation feature vector, which can quantify the dynamic input energy of wind loads and its frequency distribution, thus surpassing static indicators and more accurately describing the excitation essence of wind on structures.

[0051] The processing of structural vibration data involves structural response feature analysis. The vibration response of a transmission tower under wind load is a complex, high-dimensional signal containing the superposition of multiple modes. The system acquires raw vibration signals through accelerometers deployed at key nodes of the tower (such as the tower head, crossarm connections, and tower slope change points). The analysis process first filters the signal to remove high-frequency noise and low-frequency drift. Then, modal parameter identification technology is used to separate the main vibration modes of the structure from the response, such as first-order bending, second-order bending, and torsional modes. For each identified mode, the analysis process extracts its current frequency (to monitor changes in structural stiffness), damping ratio (to assess energy dissipation characteristics), and mode shape amplitude (to quantify vibration intensity). In addition, the time-domain characteristics of the vibration signal, such as peak value and root mean square value, as well as frequency-domain characteristics, such as the energy proportion within a specific frequency band, are analyzed. All these analyzed parameters are integrated into a structural response feature vector, which comprehensively characterizes the dynamic state of the structure under the current wind load.

[0052] After acquiring the wind speed fluctuation feature vector and the structural response feature vector, the system performs cross-modal feature fusion processing to generate a fused wind-induced feature set. This fusion is not a simple data splicing, but a deep integration process emphasizing spatiotemporal correlation. First, time dimension alignment is performed. Due to potential subtle differences in the acquisition frequency and transmission delay of wind speed and vibration data, they must be mapped to a unified timestamp sequence before fusion to ensure that wind excitation and structural response are strictly corresponding in time within each analysis time step. Subsequently, a wind speed-structure cross-attention mechanism is constructed. The core function of this mechanism is to calculate the correlation between the fluctuation characteristics (such as turbulent energy at a specific frequency) of each time step in the wind speed fluctuation feature vector and the corresponding response characteristics (such as the vibration amplitude of a certain mode) of the structural response feature vector at the corresponding time step. Through extensive calculations, a cross-modal correlation matrix is ​​formed. This matrix quantifies the coupling strength between different wind characteristics and different structural response modes. For example, the matrix may reveal a high positive correlation between turbulent energy in a specific frequency range and the vibration amplitude of the first-order bending mode of the tower, while the correlation with the response of the torsional mode is weaker. This in-depth correlation analysis ensures that the fused feature set not only contains information about each mode, but also contains mechanistic information about the interaction between wind and structure.

[0053] These fused wind-induced features, each with precise timestamps, are input into the evaluation unit. This unit consists of three parallel sub-modules, each performing a shift assessment for wind, vibration, and damage, respectively. The wind speed assessment shift unit takes timestamped wind speed features (from the fused feature set) as input. It operates by dynamically comparing the differences in feature vectors across different time periods based on a continuous time series. It not only calculates the difference between the current and previous time features to determine the drasticness of the change, but also predicts the possible state of the features at the next time period based on recent trends and compares it with the subsequently input features. This continuous time-stamped difference analysis can trigger the labeling of persistent abnormal wind conditions, such as identifying wind speed fluctuations that are continuously deviating from normal patterns, even if their instantaneous values ​​do not exceed thresholds.

[0054] The structural vibration assessment and regression unit is more complex, containing an internal structural vibration prediction model. This model, trained on long-term historical monitoring data, learns the normal evolution pattern of structural response under given wind-induced characteristic inputs. The unit inputs the fused features of the current moment into the model to predict the normal range of structural vibration data for the next moment. Subsequently, the actual monitored vibration data for the next moment is compared with the predicted value. If the actual value significantly exceeds the predicted range, especially with abrupt changes in amplitude or frequency, it indicates that the structure may have exhibited unexpected behavior, such as loosening of connections, accelerated material damage, or entry into the nonlinear response region. In this case, the unit immediately marks the location as a structural risk. This difference detection based on model prediction is more sensitive to capturing early structural anomalies than simple threshold judgment.

[0055] The fatigue damage assessment and regression unit receives direct or indirect damage signals from sensors and performs time-series regression analysis, focusing on changes in the damage accumulation rate. These three assessment and regression units operate in parallel, continuously and dynamically assessing the disaster resistance performance of the transmission tower from three dimensions: wind input, structural state, and damage results. Their output provides multi-dimensional judgment criteria for subsequent dynamic optimization and decision-making. The entire feature extraction and assessment regression process achieves a transformation from raw data to a deep understanding of the tower's state.

[0056] Example 3: See Figure 4 The dynamic optimization unit receives disaster resilience assessment results from the evaluation unit. Its core function is to adjust the analysis step size of each evaluation shift unit in real time based on the severity of data changes. For the wind speed evaluation shift unit, the optimization is based on the rate of change of wind speed within consecutive timestamps. This unit continuously calculates the magnitude of change in the wind speed characteristic vector per unit time, such as the rate of change of turbulence intensity or gust factor. If this rate of change exceeds a certain set high threshold, it indicates that the wind field is in an extremely unstable state, and the wind speed characteristics may evolve rapidly. At this time, the dynamic optimization unit will send an instruction to the wind speed evaluation shift unit to shorten its evaluation shift step size. This means that the system will perform comparisons of the differences between wind speed characteristics before and after time intervals more frequently, adjusting from possibly evaluating once per minute to evaluating once every ten seconds or less, capturing subtle changes in the wind field with higher temporal resolution, and preventing the omission of key dangerous wind condition information due to excessively long analysis intervals.

[0057] For the structural vibration assessment and progression unit, the optimization logic is more refined, based on the acceleration of structural vibration changes, i.e., the rate of change of vibration characteristics (such as the amplitude or frequency of a certain mode). The system not only calculates the vibration change rate from time t-2 to time t-1, but also from time t-1 to the current time t, and compares the difference between the two. If the rate of change in the later time interval shows a significant accelerating trend compared to the previous time interval, and this acceleration value exceeds a minimum threshold, it indicates that the structural response may be entering a rapid development stage of nonlinearity or instability. Faced with this scenario, the dynamic optimization unit issues an instruction to compress the step size ratio of the structural vibration assessment and progression unit. This allows the structural vibration prediction model to predict and compare with smaller step intervals, thereby enabling closer tracking of the deterioration trajectory of the structural state and providing earlier warnings of potential risks.

[0058] The optimization mechanism of the fatigue damage assessment shift unit is based on signal feature comparison. This unit continuously acquires and analyzes fatigue damage signals, such as waveform data from acoustic emission sensors or cumulative damage calculated based on strain history. These signals are decomposed into the frequency domain for analysis. The system internally constructs a damage feature library, which stores typical spectral feature patterns associated with different damage modes (such as microcrack initiation, stable macrocrack propagation, unstable propagation, loosening of rivets or bolts, and micro-degradation of materials). By comparing the real-time acquired damage spectrum with the patterns in the feature library, if the system detects a high degree of similarity between the current spectrum and the frequency band characteristics of specific damage modes such as "crack propagation" or "material degradation," it determines that the damage process is intensifying. This judgment immediately activates the step size adjustment mechanism, and the dynamic optimization unit correspondingly shortens the damage shift step size of the fatigue damage assessment shift unit. This allows the system to update the damage accumulation state assessment at a higher frequency and more sensitively monitor the rate of damage evolution.

[0059] While completing the dynamic step size adjustment, the system initiates a deep diagnostic process to trace the root cause of the anomaly. This process calls a pre-trained evaluation network model. This model receives the fused wind-induced feature set generated in Example 2 as input. Its internal mechanism performs in-depth analysis of the input features, quantifying and assigning values ​​to the correlation strength between each potential fatigue damage data point and each structural response feature in the fused feature set, i.e., performing fatigue root cause weight allocation processing. The output of this processing is an anomaly correlation score set, which is essentially a distribution matrix that clearly reveals the contribution of different structural response anomalies to the currently observed fatigue damage anomalies. For example, the score results may show that the current damage accumulation at the tower foot node is likely 80% related to the monitored first-order bending mode vibration amplitude anomaly of the tower body, while the direct correlation with other modes or wind speed characteristics is weak.

[0060] Based on this abnormal correlation distribution, the system performs joint root cause tracing. This process does not view damage or vibration in isolation, but rather deeply couples fatigue damage data with structural response feature vectors in a strongly correlated dimension, tracing the origin and propagation path of the anomaly in time and space. Its output is the root cause tracing result of wind-induced fatigue, which might be specifically stated as: "Currently, the fatigue damage in the main material on the southeast side of the third tower section is accumulating rapidly. The main root cause is the continuous vortex shedding excitation that occurred in the vicinity of this area in the past hour. This excitation resonated with the second-order vibration mode of the structure, leading to a significant increase in the local dynamic stress amplitude." This tracing result links abstract damage data with specific physical phenomena and structural components.

[0061] The core of the entire dynamic tuning and root cause analysis process is a closed loop of perception-analysis-adaptation. The dynamic tuning unit determines the granularity of analysis (analysis) based on the output (perception) of the evaluation unit. Its output step size adjustment instructions and root cause analysis results, in turn, guide the evaluation unit to work more focusedly and efficiently (adaptation). For example, when root cause analysis points to a specific mode of vibration, the structural vibration assessment and analysis unit can further refine the monitoring and analysis of that mode. This process enables the system to intelligently allocate computing resources, maintaining a regular monitoring rhythm during calm times and automatically entering a "wartime" state when storms approach or structural hazards arise, increasing the frequency and depth of analysis, thereby achieving efficient and accurate disaster resilience assessment.

[0062] The core criterion for this adaptive adjustment process can be quantified by a dynamic weighting function:

[0063]

[0064] in: This represents the adjusted evaluation time interval, and its value changes dynamically. This represents the system's preset basic evaluation interval and is a reference constant. It is a scaling factor used to adjust the intensity of the overall adaptive adjustment. It is a sensitivity coefficient that determines the degree to which the system responds to accelerated changes. It represents the second derivative of the observed parameter (such as vibration amplitude or accumulated damage) with respect to time, that is, the acceleration of the change of the parameter, which directly reflects the drastic degree of the evolution of the system state.

[0065] The meaning of this relationship is: when the observed parameter changes with acceleration As the value increases, the value of the function's denominator also increases, leading to a change in the calculated adjusted interval. This means that the faster and more drastic the system state changes, the shorter the time interval between analyses performed by the evaluation unit, and the higher the sampling and analysis frequency, in order to achieve close tracking. Conversely, when the changes are stable, the interval regresses to the baseline value.

[0066] Example 4: The system sets a step size boundary constraint module. This module receives the adjusted damage shift step size, structural vibration compression step size ratio, and wind speed assessment shift step size from the dynamic optimization unit. These adjusted step size parameters are not unlimited but are constrained by preset boundaries. The boundary values ​​are not fixed but are set based on historical extreme operating condition data of the near-fault area where the transmission tower is located, taking into account factors such as maximum instantaneous wind speed, strongest ground motion record, and worst fatigue damage case. Simultaneously, the system defines intensity levels, which are divided according to the current level of environmental disturbance. Low disturbance levels correspond to normal wind conditions and structural stability; medium-high disturbance levels correspond to high-risk conditions such as continuous strong winds, increased abnormal structural vibration, or rapid accumulation of fatigue damage. Different intensity levels correspond to different step size boundary ranges. For example, under low disturbance conditions, a wider adjustment margin is allowed for the step size, while under medium-high disturbance conditions, the boundary is tightened, forcing the assessment to maintain a higher frequency.

[0067] By adjusting the step size boundary, the system triggers a redundancy boundary buffer layer mechanism. This buffer layer is not a physical structure, but a software-level fault-tolerant logic used to temporarily tolerate step size adjustments. Its core function is to allow the step size to briefly exceed the rated boundary under certain conditions without immediately triggering a system alarm or forced reset. The operation of the buffer layer relies on two key criteria: whether the actual adjusted total number of step sizes exceeds the maximum capacity limit of the buffer layer, and the residence time of the over-limit state within the buffer layer. For example, when a sudden surge in wind speed causes the dynamic tuning unit to significantly shorten the wind speed evaluation step size, the step size value may momentarily exceed the rated boundary set for the low-disturbance state. In this case, if the calculated total number of adjusted step sizes of all current evaluation units (which can be understood as the system processing load) does not exceed the preset maximum step size threshold of the redundancy boundary buffer layer, the system allows the over-limit state to exist temporarily, and the evaluation process continues to run at the adjusted step size, avoiding frequent switching of the evaluation rhythm due to short-term disturbances. However, this tolerance is limited. The system continuously monitors the residence time of the over-limit state within the buffer layer. If this time exceeds a preset threshold (e.g., 5 minutes), it is determined to be a continuous anomaly, the buffer layer mechanism immediately fails, and the system will force the step size of all evaluation units to switch to the preset minimum safe step size. The minimum safe step size is the most conservative setting for the system to ensure basic monitoring functions, ensuring that a minimum level of evaluation capability can still be maintained under the worst operating conditions.

[0068] The system continuously monitors the status of key parameters after the evaluation unit's adjustment, including the rate of change of wind speed, the rate of change of structural vibration gradient, and the characteristics of fatigue damage spectrum. If these parameters are determined to be within the stable region defined by the redundant boundary buffer layer, it indicates that the currently adjusted step size setting can effectively track changes in system state and that the system is operating smoothly. At this point, the system will continue to fine-tune the evaluation step size through the dynamic tuning unit based on the latest real-time data, achieving adaptive tracking. The range of the stable region is defined using stability theory methods to ensure that the system does not diverge or oscillate within the region.

[0069] The system also incorporates a step-size freeze mechanism. This mechanism monitors two conditions: the total compensation required for adjusting the step size of the evaluation units (which can be understood as the total number of step-size adjustments performed by the system to maintain stability), and the buffer layer dwell time. If the system calculates that the total compensation amount does not exceed the maximum step-size limit of the redundant boundary buffer layer, but the buffer layer dwell time has exceeded its set time threshold, step-size freeze is triggered. In the frozen state, the step size of all evaluation units will be locked at the current value, and they will no longer respond to the adjustment commands of the dynamic tuning units until the system detects that the key parameters (wind speed change rate, vibration gradient change rate, damage spectrum characteristics) have returned to the stable domain and remain there for a period of time, or receives an external reset command. The freeze mechanism aims to prevent the system from falling into invalid oscillations near the boundary due to continuous fine-tuning, wasting computational resources and potentially interfering with the evaluation results. Refer to Table 1, which shows the key parameters of the redundant boundary buffer layer and the system response logic under different states.

[0070] Table 1: Redundant boundary buffer layer state transition logic.

[0071]

[0072] Example 5: Based on the root cause tracing results output in Example 3, the system generates targeted disaster resistance performance optimization strategies. These strategies are not general solutions but are closely integrated with the specific root causes identified. For example, if the root cause tracing clearly points to excessive dynamic stress caused by local resonance in exacerbating fatigue damage at a specific connection node, the strategy may include measures to suppress vibrations of that mode. If the root cause is abnormal redistribution of internal forces due to uneven foundation settlement, the strategy may focus on foundation reinforcement. The generated strategies are fed back to the evaluation unit in real time, triggering parameter calibration. This operation adjusts the threshold parameters, weighting coefficients, or prediction model parameters within the evaluation unit according to the specific recommendations of the strategy. For example, if the strategy suggests focusing on vibrations of a specific mode, the structural vibration assessment and propagation unit will correspondingly increase the warning threshold sensitivity for that mode's amplitude or assign a higher weight to that mode in its prediction model. This feedback calibration enables the evaluation unit to more accurately monitor key indicators related to the strategy in subsequent work.

[0073] Based on the comprehensive output of the evaluation unit, especially the judgment of the dynamic optimization unit, the system makes a final decision on whether the transmission tower structure needs reinforcement. If reinforcement is deemed necessary, the system will trigger a series of structural enhancement operations. First, it triggers reinforcement of the transmission tower's structural connectors and foundation reinforcement components. Connector reinforcement targets identified weak or high-stress nodes and may include: replacing bolts or connecting plates with higher strength grades; adding additional angle steel or steel plate supports at critical nodes to distribute loads and suppress local deformation; and locally reinforcing or wrapping members showing early signs of damage. Foundation reinforcement targets foundation issues that may be involved in root cause analysis, such as: pressure grouting around the foundation to improve soil bearing capacity and uniformity; adding or tensioning ground anchors to enhance the foundation's pull-out and overturning resistance; and pouring reinforced concrete enclosures around the foundation cap to increase the foundation's base area and stability. These enhancement measures aim to improve the structure's load-bearing capacity and fatigue resistance from the source.

[0074] Regardless of whether major structural reinforcement is required, the system will trigger damper activation. Transmission tower structures are typically equipped with passive or semi-active tuned mass dampers or viscous dampers. The damper activation process is achieved through a control algorithm: the system calculates the optimal damper parameters (such as the tuning frequency and damping coefficient of the mass block) based on the currently monitored dominant frequency and amplitude of structural vibration. For passive dampers, it may be necessary to remotely adjust the internal variable damping valve or mass block position; for semi-active or active dampers, a real-time adjustment command is sent to their control system to make their output force opposite to the structural vibration, thereby effectively dissipating vibration energy, suppressing structural response, and reducing fatigue damage rate.

[0075] A balancer typically refers to a controllable counterweight system or hydraulic adjustment device installed at specific locations on the tower head or body. Its core function is to shift the load upwards in the longitudinal and lateral directions of the tilt to maintain a near-stable state. For example, when a transmission tower is detected to be tilting longitudinally (along the line direction) or laterally (perpendicular to the line direction) due to strong winds or uneven foundation settlement, the balancer control system is activated. In the case of longitudinal tilt, the balancer may use a hydraulic lifting device to finely adjust the load on the tower head (such as the equivalent mass at the conductor suspension point) upwards in the direction of tilt (i.e., the opposite direction of tilt). In the case of lateral tilt, the mass is moved upwards in the direction of tilt by adjusting the position of the counterweight sliders on both sides of the tower body. This active redistribution of load generates a restoring moment opposite to the tilt trend, effectively counteracting the tilting force, helping the structure restore or maintain a near-vertical stable state, reducing additional bending moments and stress concentrations caused by tilting, thereby reducing fatigue risk. After the structural connectors and foundation reinforcements are completed, the balancer operation, as the final step, works in conjunction with the damper to ultimately complete the overall reinforcement and stabilization of the transmission tower structure.

[0076] If the system determines that the transmission tower structure does not require reinforcement of major structural connectors and foundation reinforcements, a different operating procedure is executed. Similarly, the system triggers damper activation, adjusting damping parameters based on real-time vibration data to suppress vibration and reduce fatigue damage. Simultaneously, it also triggers balancer operation, using longitudinal and lateral load movement to maintain the structure in a near-stable state, counteracting any detected minor tilting tendencies and maintaining a good stress state for the structure.

[0077] Without requiring major reinforcement, the system triggers an electrical isolation operation between the transmission tower structure and its foundation. This operation aims to block stray current paths that may form between the tower and the foundation. When stray currents flow through the soil, if they form a circuit through the metal tower and foundation, they can cause electrochemical corrosion, accelerating the corrosion of metal connectors and foundation reinforcement, thereby reducing their fatigue life and load-bearing capacity. Electrical isolation is achieved physically: high-performance insulating materials are installed between the tower legs and the foundation abutment or at the foundation anchors. These insulating materials must possess extremely high mechanical strength to withstand structural loads, while also exhibiting excellent electrical insulation properties and long-term weather resistance. During installation, it is crucial to ensure that the insulating gaskets completely isolate the metal tower from the reinforcing mesh in the concrete foundation, thoroughly cutting off potential current paths. After electrical isolation is implemented, the system verifies the isolation effect using a dedicated insulation resistance monitoring device to ensure that the high resistance value required by the design is achieved.

[0078] Regardless of the path chosen (reinforcement required or not), the system re-enters the monitoring and evaluation cycle after all triggering operations are completed. The evaluation unit continues to receive real-time wind-induced fatigue data using calibrated parameters and performs multi-level evaluation progression. The dynamic optimization unit continues to adjust the evaluation step size based on the new data stream. The system continuously monitors the effectiveness of reinforcement or optimization measures, such as observing the degree of vibration amplitude attenuation after damper activation, the correction of tilting trends after balancer operation, the elimination effect of stray current after electrical isolation, and changes in the rate of stress or damage accumulation at critical nodes.

[0079] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0080] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for evaluating the disaster resistance performance of transmission towers in near-fault areas considering wind-induced fatigue effects, characterized in that, include: Acquire wind-induced fatigue data of transmission towers in the near-fault area, the wind-induced fatigue data including wind speed data, structural vibration data and fatigue damage data; The determination of wind-induced fatigue anomalies includes wind-induced fatigue baseline, single-modal anomaly, and secondary verification. Feature extraction is performed on the wind-induced fatigue data, and the extracted wind-induced fatigue features are assigned with timestamp labels and input into the evaluation unit; The evaluation unit sets up a multi-level evaluation process to analyze and judge the wind-induced fatigue data and output the disaster resistance performance evaluation results. A dynamic optimization unit that adjusts the evaluation step size of the evaluation unit in real time receives the disaster resistance performance evaluation results and determines whether the transmission tower structure needs to be reinforced. The feature extraction of wind-induced fatigue data includes performing dynamic fluctuation feature extraction processing on wind speed data to obtain wind speed fluctuation feature vector, performing structural response feature parsing processing on structural vibration data to obtain structural response feature vector, and performing cross-modal feature fusion processing on the wind speed fluctuation feature vector and structural response feature vector to generate a fused wind-induced feature set. The cross-modal feature fusion processing includes time dimension alignment processing, which maps the wind speed fluctuation feature vector to the same time granularity as the structural response feature vector, and constructs a wind speed-structure cross-attention mechanism to calculate the cross-modal correlation matrix between the fluctuation feature of each time step in the wind speed fluctuation feature vector and the response feature of the corresponding time step in the structural response feature vector. The assessment unit includes a wind speed assessment unit, a structural vibration assessment unit, and a fatigue damage assessment unit; The wind speed assessment and shift unit takes wind speed features with timestamp labels as input, and obtains the difference between the wind speed features of the previous moment and the wind speed features of the current moment and the difference between the wind speed features of the next moment and the wind speed features of the current moment by real-time monitoring of wind speed features within a continuous time stamp, triggering anomaly marking within a continuous time period. The structural vibration assessment and shifting unit sets up a structural vibration prediction model, predicts the structural vibration data at the next moment through the structural vibration prediction model, and marks the structural risk through abrupt change detection. The dynamic optimization unit receives the disaster resistance performance assessment results and drives the wind speed assessment shift unit to receive the wind speed change rate within the continuous time stamp through the wind speed change rate within the continuous time stamp. It adjusts the wind speed parameter assessment shift step size. If the wind speed data change rate per unit time is greater than the highest weight, the wind speed parameter assessment shift step size is shortened. The step size is adjusted by adjusting the acceleration of the vibration change of the continuous timestamp structure. If the acceleration of the vibration change of the timestamp structure from time t-1 to time t is greater than the acceleration of the vibration change of the continuous timestamp structure from time t-2 to time t-1, and the acceleration increase is greater than the minimum threshold, then the step size ratio is compressed. The fatigue damage signal is obtained by the fatigue damage assessment shift unit, and the step size is adjusted by comparing it with the damage feature library. If the damage signal captured by the fatigue damage spectrum is crack propagation and material degradation, the damage shift step size of the fatigue damage assessment shift unit is adjusted.

2. The disaster resistance performance evaluation method for transmission towers in near-fault areas considering wind-induced fatigue effects according to claim 1, characterized in that, The judgment of wind-induced fatigue anomaly includes using a wind-induced fatigue baseline to compare whether the acquired wind speed data, structural vibration data, and fatigue damage data are within the normal threshold range of wind-induced fatigue. If the wind speed data, structural vibration data, and fatigue damage data are not within the normal threshold range of wind-induced fatigue, the judgment of wind-induced fatigue anomaly is triggered. The single-mode anomaly is used to map whether there are abrupt changes, imbalances, and decouplings in wind speed data, structural vibration data, and fatigue damage data. If any of the wind speed data, structural vibration data, and fatigue damage data has a single data anomaly, a single wind-induced fatigue anomaly data report is output, and the data acquisition unit is checked for anomalies.

3. The disaster resistance performance evaluation method for transmission towers in near-fault areas considering wind-induced fatigue effects according to claim 2, characterized in that, The secondary verification includes temporal correlation verification and spatial propagation verification. The temporal correlation verification is used to establish the mutation, imbalance and decoupling of continuous timestamp wind speed data, and the spatial propagation verification is used to obtain the mutation, imbalance and decoupling of structural vibration data and fatigue damage data. If the data mapped in the temporal correlation verification and spatial propagation verification is abnormal, the grid density is dynamically optimized by dividing the grid into high-resolution grids, extracting local neighborhood features.

4. The disaster resistance performance evaluation method for transmission towers in near-fault areas considering wind-induced fatigue effects according to claim 1, characterized in that, The pre-trained evaluation network model is invoked to perform fatigue root cause weight allocation processing on the fused wind-induced feature set, generating an abnormal correlation score set corresponding to wind-induced fatigue, wherein the abnormal correlation score set contains the abnormal correlation distribution between fatigue damage data and response features of each structure. Based on the abnormal correlation distribution, joint root cause tracing processing of fatigue damage data and structural response feature vectors is performed to generate abnormal root cause tracing results of wind-induced fatigue.

5. The disaster resistance performance evaluation method for transmission towers in near-fault areas considering wind-induced fatigue effects according to claim 4, characterized in that, The step size is evaluated by setting step size boundary limits to receive adjusted damage push step size, structural vibration compression step size ratio and adjusted wind speed, and intensity levels are set, including low disturbance and medium-high disturbance. By adjusting the step size boundary, the redundant boundary buffer layer is triggered to temporarily tolerate the step size adjustment. If the step size briefly exceeds the rated boundary, the actual total number of adjusted steps will not be greater than the maximum number of steps in the redundant boundary buffer layer. If the buffer layer stays for more than the threshold, it will immediately switch to the minimum safe step size.

6. The disaster resistance performance evaluation method for transmission towers in near-fault areas considering wind-induced fatigue effects according to claim 5, characterized in that, If the wind speed change rate, structural vibration gradient change rate, and fatigue damage spectrum characteristics of the evaluation unit are within the stability domain of the redundant boundary buffer layer after adjustment, the step size of the evaluation unit's evaluation shift will be adjusted further based on real-time data. If the total compensation amount for adjusting the step size of the evaluation unit is not greater than the maximum step size of the redundant boundary buffer layer, and the buffer layer dwell time is greater than the threshold, then step size freezing is triggered.

7. The disaster resistance performance evaluation method for transmission towers in near-fault areas considering wind-induced fatigue effects according to claim 6, characterized in that, Based on the results of anomaly root cause tracing, a disaster resistance performance optimization strategy is generated, and the disaster resistance performance optimization strategy is fed back to the evaluation unit to trigger parameter calibration operation; If it is determined that the transmission tower structure needs to be reinforced, the reinforcement of the transmission tower structure connectors and foundation reinforcement components is triggered, the damper is activated, and the balancer is triggered. The balancer moves the load upward in the longitudinal and lateral directions to maintain a near-stable state, and finally completes the reinforcement of the transmission tower structure. If the transmission tower structure does not require reinforcement, the damper is activated, and the balancer is triggered. The balancer moves the load upward in the longitudinal and lateral directions to maintain a near-stable state, triggering electrical isolation between the transmission tower structure and the foundation.

Images