Online monitoring method and system for power transmission tower

Abstract

The invention relates to the technical field of power transmission tower safety monitoring, in particular to an on-line monitoring method and system for a power transmission tower, and the method comprises the steps: carrying out the fitting of the data of the power transmission tower, obtaining a stress data-time curve and a stress data-height curve, and calculating the noise performance degree and credibility of a stress data point, according to the method, a target data sequence is obtained, weighted average is carried out on the noise expression degree and credibility of stress data points of the target data sequence to obtain the noise level of the target data sequence, a denoising result of a CEEMDAN algorithm is accurately evaluated by using the noise level, and the optimal average frequency of the CEEMDAN algorithm is obtained according to feedback of an evaluation result, so that the limitation of manually setting parameters is avoided, and the accuracy of the CEEMDAN algorithm is improved. The data denoising effect is ensured, the stress data truly reflecting the stress state of the power transmission tower is obtained, and the accuracy of power transmission tower monitoring analysis is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of safety monitoring of power transmission towers, and in particular to an online monitoring method and system for power transmission towers. Background Art

[0002] In traditional power transmission systems, transmission towers are key infrastructure of the power transmission network, ensuring that electricity is safely and efficiently transmitted from power stations to users. These towers must withstand stress from multiple external factors such as load, wind, temperature changes, ice and snow loads, geological activities, and internal stress caused by the aging of the equipment itself. Long-term stress accumulation may lead to metal fatigue and structural damage. In severe cases, it may cause the tower to collapse, causing large-scale power outages, and having a significant impact on the social economy and people's lives. Therefore, real-time monitoring of the stress state of transmission towers and timely detection of potential problems are crucial to preventing failures and ensuring the safety of the power grid.

[0003] The existing technology mainly collects and analyzes the stress data of the transmission tower to monitor the transmission tower. For example, the patent document with the authorization announcement number CN104092308B provides a stress measurement transmission tower online safety monitoring system and method, which mainly installs stress sensors at key parts of the tower. The measured values ​​of the stress sensors are transmitted to the background server through the network. The background server calculates the overall reliability of the tower, which solves the problem of real-time stress monitoring of the tower and can evaluate the safety status of the tower in real time. However, the challenge faced in the actual monitoring process is that the outdoor environment where the transmission tower is located is complex and changeable, and the collected stress data is susceptible to various noise pollution, such as electromagnetic interference, environmental changes, etc. These noises will seriously affect the accuracy of the data and lead to misjudgment of the stress state of the tower. Therefore, it is necessary to denoise the stress data to improve the quality of the monitoring data and ensure the accuracy of the analysis.

[0004] For the denoising of stress data, the CEEMDAN (Adaptive Noise Complete Ensemble Empirical Mode Decomposition) algorithm can be used. This algorithm is a common denoising algorithm. For example, the patent document with the authorization publication number CN114992033B proposes a hydropower unit signal denoising method. First, the original signal is preprocessed with NLM (non-local mean) to reduce noise, and then the CEEMDAN method is used to decompose the preprocessed signal to obtain a series of IMF components. At the same time, the sample entropy of each component is calculated, and the components are divided into high-frequency noisy components, signal-noise mixed components and low-frequency effective components according to the sample entropy values. Finally, the noise components in the signal-noise mixed components are further filtered out by the wavelet threshold denoising method, and are filtered out from the original ferry signal together with the high-frequency noisy components, thereby completing the denoising of the hydropower unit ferry signal.

[0005] In the above scheme, when the CEEMDAN algorithm is used for denoising, the CEEMDAN algorithm performs denoising by continuously adding noise and decomposing the original data, wherein the average number is a parameter in the algorithm, which refers to the number of times noise is added and EMD decomposition is performed, and the average number is often a preset fixed value. Since the noise level in the transmission tower monitoring environment fluctuates with time and external conditions, the noise characteristics of data in different time periods are different, and a unified average number cannot flexibly respond to such dynamic changes, which can easily lead to insufficient or excessive denoising, damage the authenticity of the data, and thus reduce the accuracy of monitoring the transmission tower. Summary of the invention

[0006] In order to solve the problem that when using the CEEMDAN algorithm to denoise the stress data of the transmission tower, it is easy to cause insufficient or excessive denoising, damage the authenticity of the data, and reduce the accuracy of monitoring the transmission tower, the present invention proposes an online monitoring method and system for the transmission tower. The specific technical solution is as follows: On the one hand, the present invention provides an online monitoring method for a transmission tower, comprising: Collecting stress data sequences of the same height of a transmission tower at different collection times as target data sequences of the height; and evaluating the noise level of the target data sequences, including: Using data fitting to obtain a stress data-time curve corresponding to the target data sequence, and a stress data-height curve corresponding to the acquisition moment of each stress data in the target data sequence; According to the fitting loss of the stress data point in the stress data-time curve, the noise performance degree of the stress data point in the target data sequence is calculated, and the noise performance degree is positively correlated with the fitting loss; the difference between the derivative of the stress data point in the stress data-height curve and the mean of the derivative is used as the credibility of the noise performance degree of the stress data point; Performing weighted averaging on the noise performance degree and credibility of the stress data points of the target data sequence to obtain the noise level of the target data sequence; A plurality of average times are pre-generated, and the CEEMDAN algorithm denoises the target data sequence based on the average times, and re-evaluates the noise level of the target data sequence, and the average times and denoising results corresponding to the minimum noise level are taken as the optimal average times and optimal denoising results, and the transmission tower is monitored according to the optimal denoising results.

[0007] The above scheme performs time curve fitting on the stress data series at different time points at the same height, and performs spatial curve fitting on the stress data at different heights at the same time point. This method can accurately identify and quantify the degree of noise in the data. By associating the fitting loss with the noise intensity, the accuracy of the noise assessment is ensured, providing a scientific basis for subsequent denoising. The analysis of the derivative change of the stress data point in the height curve is introduced as a credibility indicator of the degree of noise. This measure not only takes into account the fluctuation of the data itself, but also combines the change trend in the physical sense, improving the robustness and pertinence of the denoising process. By pre-generating multiple average times and applying the CEEMDAN algorithm to conduct denoising experiments, different denoising methods are systematically explored. The improvement of data quality under intensity is automatically determined according to the principle of minimizing the noise level. This method can adaptively select the most effective denoising scheme for a specific data set, avoiding the overfitting or underfitting problems that may be caused by the artificially set denoising parameters (averaging times). After determining the optimal averaging times, the final target data sequence obtained has the lowest noise level, which means that the data sequence can more realistically and reliably reflect the stress state of the transmission tower. In summary, the above scheme greatly improves the quality and analysis efficiency of the transmission tower stress monitoring data through a sophisticated data processing process and an intelligent parameter optimization strategy, providing an effective data basis for the safe operation and maintenance management of the power system.

[0008] Preferably, the noise level of the target data sequence satisfies the following relationship:

[0009] In the formula, is the noise level of the target data sequence, j is the sequence number of the stress data point in the target data sequence, is the total number of stress data points contained in the target data series, is the noise performance degree of the jth stress data point in the target data sequence, is the credibility of the noise performance of the j-th stress data point, It is the sum of the credibility of the noise performance of all stress data points in the target data sequence.

[0010] The above scheme comprehensively considers the noise performance degree and credibility of each stress data point, realizes the accurate quantification of the overall noise level of the target data sequence, and improves the denoising effect; by introducing credibility and the sum of credibility, the evaluation can more comprehensively reflect the differences between stress data points, optimizes the trust metric level in the denoising process, and enhances data reliability.

[0011] Preferably, the noise performance degree of the stress data point satisfies the following relationship:

[0012] In the formula, is the noise performance degree of the jth stress data point in the target data sequence, is the normalization function, is the total number of stress data points contained in the local range of the jth stress data point, is the sequence number of the stress data point in the local range, For the local range stress data points and The fitting loss between the stress data points.

[0013] The above scheme effectively evaluates the noise performance of a single stress data point by accumulating the fitting losses of adjacent stress data points in a local range, thereby improving the targetedness of denoising. The accumulation of fitting losses in a local range takes into account the fluctuation of local data, making the monitoring reflect the noise level of the tower stress data more realistic, which is conducive to timely discovery of potential problems and enhances the reliability of system safety monitoring.

[0014] Preferably, the credibility of the noise performance level = , is the normalization function, is the first The derivative of a stress data point in the stress data-height curve, is the derivative mean of the stress data-height curve, To obtain the absolute value sign.

[0015] The above scheme measures the credibility of noise performance by calculating the absolute value of the deviation between the derivative of a single stress data point on the height curve and its average derivative, and normalizes it, thereby improving the accuracy of the assessment. The noise credibility quantified by the derivative deviation is directly related to the physical meaning of the data change, which improves the model's sensitivity to identifying the actual stress state and helps to quickly discover and respond to potential risks of transmission towers.

[0016] Preferably, monitoring the transmission tower according to the optimal denoising result includes: displaying the optimal denoising result in a visual manner to achieve monitoring of the transmission tower.

[0017] Preferably, the method for obtaining the fitting loss is: In the stress data-time curve, calculate the stress data points and stress data points and The integral area enclosed by the axis, and the stress data points and stress data points and The area of ​​the trapezoid enclosed by the axis, the absolute value of the difference between the integral area and the trapezoid area is taken as the first stress data points and The fitting loss between the stress data points.

[0018] The above scheme quantifies the local fitting loss by accurately calculating the difference between the integral area and the trapezoidal area of ​​the stress data between two points in the time series, improves the sensitivity measurement to data fluctuations, and enhances the analysis accuracy.

[0019] Preferably, the local range of the stress data point is composed of the stress data point and M stress data points on its left and right sides, where M is a preset value.

[0020] Preferably, the method for pre-generating a plurality of average times is as follows: presetting an initial average number as C, an iteration step as 1, and gradually increasing the iteration step to obtain a plurality of average times.

[0021] On the other hand, the present invention also provides an online monitoring system for a transmission tower, the monitoring system comprising a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the monitoring method described above.

[0022] The present invention has the following effects: The present invention accurately locates and quantifies data noise through the time curve and space height curve fitting technology of stress data points, innovatively incorporates derivative change analysis to obtain a more reasonable noise level, and uses the noise level to accurately evaluate the denoising result of the CEEMDAN algorithm. The optimal average number (average number) of the CEEMDAN algorithm is obtained according to the feedback of the evaluation result, avoiding the limitations of manual parameter setting, ensuring the data denoising effect, and preventing excessive or insufficient denoising. Stress data that truly reflects the stress state of the transmission tower can be obtained, providing reliable data support for the safe operation and maintenance of the power system, and improving the accuracy of monitoring and analysis of the transmission tower. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 It is a schematic flow chart of the method of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0025] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0026] Reference Figure 1 , an online monitoring method for a transmission tower, comprising steps S1-S3, specifically as follows: S1: Collect stress data of the same height of the transmission tower at different collection times as a target data sequence of the height.

[0027] In order to comprehensively and accurately evaluate the mechanical state of the transmission tower, a high-performance strain gauge sensor system is used to monitor the stress of the tower. These strain gauges are carefully installed on the key structural components of the tower, including but not limited to the base of the tower legs, the cross arm area that bears heavy loads, and critical interface locations such as connectors. Through such a layout, multi-point coverage of the overall stress distribution of the tower is ensured, thereby obtaining the most detailed data basis.

[0028] In the specific operation, the strain gauges are arranged in order from bottom to top according to the height of the transmission tower, and the corresponding heights from the lowest point to the highest point are defined as , , ...、 ,in Representing the Tower The vertical height of the first measuring point, i.e. height, and N represents the total number of strain gauges deployed in the entire monitoring network. Indicates The monitoring is intensive and comprehensive. Each height is equipped with independent strain gauges to continuously record stress changes, thus forming a series of stress data sequences, such as The stress data sequence corresponding to the height is , Detailed record of the The dynamic changes of stress data at this height over time are recorded by high-frequency acquisition frequency (1Hz), ensuring that even the smallest stress fluctuations can be effectively captured, providing a detailed data foundation for subsequent in-depth analysis and safety assessment. In short, by scientifically deploying strain gauges and implementing high-frequency data acquisition, a comprehensive and high-precision transmission tower stress monitoring network has been built.

[0029] Furthermore, any height is selected, and the stress data of this height at different acquisition times are used to form a stress data sequence. The stress data sequence is used as the target data sequence of this height, and the noise level is evaluated and the noise is removed in subsequent steps. The stress data sequence of each height can be operated according to the same method.

[0030] S2: Evaluate the noise level of the target data sequence.

[0031] In order to analyze the change of stress data, taking the target data sequence as an example, data fitting, such as the least square method, is used to fit the target data sequence to obtain a stress data-time curve. The horizontal axis of the curve is the different acquisition times of the corresponding height of the target data sequence, and the vertical axis is the value of the stress data of the height at different acquisition times. The curve reflects the change of stress data corresponding to different acquisition times at the same height; Further, each stress data in the target data sequence is analyzed: stress data at different heights at the time of collection of each stress data is obtained, and data fitting is also performed on the stress data at different heights at the time of collection of the stress data to obtain a stress data-height curve corresponding to the time of collection of the stress data, wherein the horizontal axis is the different heights at the time of collection, and the vertical axis is the value of the stress data at different heights, and the curve reflects the change trend of stress data at different heights at the same collection time, wherein the data fitting can adopt the least squares method.

[0032] In order to more intuitively display the above collected stress data, as shown in Table 1: Table 1

[0033] In Table 1, 100m, 200m, and 300m are respectively different heights of the transmission tower, T1, T2, and T3 are different acquisition times, and D11 to D33 represent the stress data corresponding to each height at each acquisition time. Taking 200m as an example, the stress data sequence of this height is D21, D22, and D23. These three data are fitted to obtain the stress data-time curve of this height. The stress data of different heights corresponding to the acquisition time T1 of D21 are D11, D21, and D31. The data of D11, D21, and D31 are fitted to obtain the stress data-height curve at time T1. The stress data of different heights corresponding to the acquisition time T2 of D22 are D12, D22, and D32. The stress data-height curve at time T2 is obtained by fitting these three. The stress data of different heights corresponding to the acquisition time T3 of D23 are D13, D23, and D33. The stress data-height curve corresponding to time T3 is obtained by fitting these three.

[0034] The noise points caused by environmental factors often appear as local mutation points, and the mutation points will affect the fitting values ​​of themselves and the surrounding stress data points to deviate from the original values ​​during the curve fitting process. Therefore, the fitting loss of the stress data points in the curve fitting can be used to evaluate the degree of noise performance. The greater the fitting loss, the greater the noise performance.

[0035] To evaluate the noise level of the target data sequence, the noise performance degree of the stress data points and the credibility of the noise performance degree must first be calculated, and the noise level of the target data sequence is evaluated based on the noise level and its credibility.

[0036] The noise performance degree of the stress data point is calculated, specifically, according to the fitting loss of the stress data point in the stress data-time curve corresponding to the stress data sequence at a certain height, the noise performance degree of each stress data point of the stress data sequence at the height is calculated, and the noise performance degree is positively correlated with the fitting loss, that is, the greater the fitting loss, the greater the noise performance degree. For example, the stress data sequence at 200m is used as the target data sequence; The noise performance degree of each stress data point in the target data sequence is calculated using the following formula:

[0037] In the formula, is the first The noise level of each stress data point is is the normalization function, is the total number of stress data points contained in the local range of the jth stress data point, is the sequence number of the stress data point in the local range, For the local range stress data points and The formula is the average of the fitting losses between adjacent stress data points in the local range, which is used as the noise performance degree of the stress data point. For example, the noise performance degree of the stress data point D22 in the target data sequence is calculated according to the formula.

[0038] In the stress data-time curve, calculate the stress data points and stress data points and The integral area enclosed by the axis, and the stress data points and stress data points and The area of ​​the trapezoid enclosed by the axis, the absolute value of the difference between the integral area and the trapezoid area is taken as the first stress data points and The larger the difference in area, the greater the fitting loss.

[0039] The fitting loss is expressed as:

[0040] In the formula, Indicates the local range stress data points and The area enclosed by the stress data point and the x-axis is expressed as an integral, which is the stress data point in the local range. Indicates the zth stress data point and the The area of ​​the trapezoid enclosed by the stress data points and the x-axis.

[0041] The local range of the stress data point is composed of the stress data point and M stress data points on its left and right sides, and M is a preset value. For example, the local range of the stress data point D22 in the target data sequence is composed of D21 and D23 on its left and right sides, and the value of M is 1. Since Table 1 only provides the simplest example, in the implementation process, this embodiment can set M to 5, that is, each stress data point and a total of 11 stress data points on its left and right sides are used as the local range, and the calculation method of the noise performance degree is the same.

[0042] The method for obtaining the fitting loss is: The credibility of the noise performance degree of the stress data point is specifically: taking the difference between the derivative of the stress data point in the stress data-height curve and the mean of the derivative as the credibility of the noise performance degree of the stress data point.

[0043] Although the noise performance degree of stress data points is clearly determined by calculation, the sudden increase or decrease of stress caused by sudden wind speed changes or abnormalities at specific locations of the tower is not fully considered. Therefore, when evaluating the noise level, the credibility of the noise performance degree of each stress data point must be included, and the credibility is used for weighted averaging to obtain the comprehensive noise level of the data series, where higher credibility is given greater weight. In the transmission tower scenario, the wind speed and stress at the same monitoring point are positively correlated, and the wind speed changes with the increase in height. Therefore, through the analysis of the height relationship between multiple points, the credibility of the noise performance degree of the quantitative stress data point can be inferred. Specifically, if the value of the stress data point is closely consistent with the height change trend under a certain instantaneous state, it indicates that it is less affected by noise, so the credibility of the noise performance degree is higher. Therefore, by introducing the credibility index and combining the height correlation, the noise level of the stress data can be quantified more accurately, ensuring the accuracy of the monitoring analysis and the reliability of the tower safety assessment.

[0044] The credibility of the noise performance of the stress data point is expressed by the formula: =

[0045] In the formula, is the normalization function, is the first The derivative of a stress data point in the stress data-height curve, is the derivative mean of the stress data-height curve, To obtain the absolute value sign, the stress data-height curve is The stress data-height curve corresponding to the collection time of each stress data point.

[0046] For example, the stress data-height curve corresponding to the acquisition time of the stress data point D22 in the target data sequence is the stress data-height curve at time T2.

[0047] This formula evaluates the impact of noise by comparing the derivative value of a stress data point at a specific height with the difference between its adjacent points. Specifically, if the stress data at a certain height deviates significantly from the surrounding data, it indicates that the point may encounter stronger noise interference. The greater the deviation of the derivative of a stress data point at a certain height from the derivative of stress data points at other heights, the greater the inconsistency between the trend of the stress data point at that height and other points, and the higher the credibility of the noise performance of this stress data point. In summary, the derivative difference not only highlights the inconsistency of the data, but also maps the strength of the credibility of the noise performance, providing a scientific basis for accurate evaluation.

[0048] For example, get the stress data points in the target data series The credibility of the noise performance is calculated first. At its collection time The derivative of the stress data-height curve at time t, and then calculate the mean of the derivative of the stress data-height curve at time t2, that is, , , The mean of the derivatives of these three stress data points in the curve is calculated based on the derivative and the mean of the derivative. The credibility of the noise representation level.

[0049] Wherein, evaluating the noise level of the target data sequence based on the noise level and its credibility includes:

[0050] In the formula, is the noise level of the target data sequence, is the sequence number of the stress data point in the target data sequence, is the total number of stress data points contained in the target data series, is the first The noise level of each stress data point is For the said The confidence level of the noise representation of each stress data point, is the sum of the credibility of the noise performance of all stress data points in the target data sequence, where is the normalization operation, which means The formula is a weighted average of the noise performance degree and credibility of each stress data point in the target data sequence, and the obtained value is taken as the noise performance level of the target data sequence.

[0051] The application of this formula is to comprehensively evaluate the overall noise level of the target data sequence. By combining the noise performance of each stress data point with its credibility in a weighted average manner, a comprehensive evaluation index is provided. Its specific significance is reflected in the following aspects: Ensure that the noise assessment not only focuses on the noise intensity of the stress data point itself, but also considers the reliability of the noisy stress data point in the overall data context; By using γ (the sum of the noise credibility of all stress data points) as the normalization factor, the normalization operation is performed to achieve a reasonable weight distribution of the noise level of each stress data point, ensuring that the influence of the stress data point on the overall noise level in the sequence is proportional to its credibility, thereby enhancing the rationality and effectiveness of the results; The degree and credibility of noise may vary in different data sets or environments. This formula adaptively reflects these changes by dynamically adjusting the weights, making the calculation of the noise level more accurately reflect the actual situation.

[0052] By carefully considering the noise performance and credibility of each stress data point, a comprehensive index is given in the form of weighted average to serve the decision support and optimization process, thereby improving the scientificity and practicality of data processing.

[0053] S3: Select the optimal averaging times of the CEEMDAN algorithm based on the noise level and monitor the transmission towers.

[0054] The noise level of the target data sequence directly indicates the quality of the denoising effect and provides a quantitative reference for subsequent analysis and decision-making. If the noise level is low, it indicates that the denoising effect is good and the data sequence is pure. Otherwise, it indicates that the denoising strategy may need to be optimized. Therefore, the noise level is an effective tool for evaluating the overall quality of the denoised data sequence.

[0055] Among them, the CEEMDAN algorithm decomposes and reconstructs the target data sequence according to the denoising parameters (average times), and finally obtains the denoised target data sequence. In order to obtain the optimal average times of the CEEMDAN algorithm so that it can achieve the optimal denoising effect, this step first pre-generates multiple average times. Specifically, the initial average times are preset to C. In this embodiment, C is 100 (empirical value), the iteration step is 1, and the iteration step is gradually increased to obtain several average times, such as 101, 102, 103, etc.

[0056] Furthermore, the target data sequence is denoised using the CEEMDAN algorithm based on the average number of times, and after each denoising, the noise level of the target data sequence is re-evaluated according to the method in S3. When the noise level is minimum after a certain denoising, the average number corresponding to this time is taken as the optimal average number of the CEEMDAN algorithm, and the corresponding denoising result (denoised target data sequence) at this time is taken as the optimal denoising result.

[0057] Increasing the number of averaging in the CEEMDAN algorithm may improve the statistical stability of the decomposition results and reduce the deviation introduced by random noise. However, excessively increasing the number of averaging may cause some IMFs to contain high-frequency components that should not be separated, but introduce new noise. Therefore, as the number of averaging increases, the noise level of the stress data series will first decrease and then increase. In this step, the number of averaging corresponding to the lowest noise level is selected as the optimal number of averaging. The algorithm denoises the target data series with the optimal number of averaging as a parameter, which improves the denoising effect and ensures the quality of the data.

[0058] Furthermore, the stress data of the transmission tower is monitored according to the optimal denoising result. The optimal denoising result is a high-quality target data sequence with less noise after denoising, and it is displayed in a visual way, such as a curve chart, to facilitate engineers and decision makers to accurately monitor and analyze the status of the transmission tower. At the same time, monitoring reports are generated regularly to summarize and analyze the operating conditions and abnormal events of the transmission tower, providing a basis and reference for decision-making.

[0059] The present invention also provides an online monitoring system for a transmission tower, the monitoring system comprising a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the monitoring method.

[0060] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0061] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. An online monitoring method for a transmission tower, characterized in that: include: Collect stress data of the same height of the transmission tower at different collection times as a target data sequence of the height; Evaluating the noise level of the target data sequence, including: Using data fitting to obtain a stress data-time curve corresponding to the target data sequence, and a stress data-height curve corresponding to the acquisition moment of each stress data in the target data sequence; According to the fitting loss of the stress data point in the stress data-time curve, the noise performance degree of the stress data point in the target data sequence is calculated, and the noise performance degree is positively correlated with the fitting loss; the difference between the derivative of the stress data point in the stress data-height curve and the mean of the derivative is used as the credibility of the noise performance degree of the stress data point; Performing weighted averaging on the noise performance degree and credibility of the stress data points of the target data sequence to obtain the noise level of the target data sequence; A plurality of average times are pre-generated, and the CEEMDAN algorithm denoises the target data sequence based on the average times, and re-evaluates the noise level of the target data sequence, and the average times and denoising results corresponding to the minimum noise level are taken as the optimal average times and optimal denoising results, and the transmission tower is monitored according to the optimal denoising results.

2. The online monitoring method of a transmission tower according to claim 1, characterized in that: The noise level of the target data sequence satisfies the following relationship: ； In the formula, is the noise level of the target data sequence, is the sequence number of the stress data point in the target data sequence, is the total number of stress data points contained in the target data series, is the first The noise level of each stress data point is For the The confidence level of the noise representation of each stress data point, It is the sum of the credibility of the noise performance of all stress data points in the target data sequence.

3. The online monitoring method of a transmission tower according to claim 2, characterized in that: The noise performance degree of the stress data point satisfies the following relationship: ； In the formula, is the first The noise level of each stress data point is is the normalization function, For the The total number of stress data points contained in the local range of stress data points, is the sequence number of the stress data point in the local range, For the local range stress data points and The fitting loss between the stress data points.

4. The online monitoring method of a transmission tower according to claim 2, characterized in that: Confidence of the noise performance level , is the normalization function, is the first The derivative of a stress data point in the stress data-height curve, is the derivative mean of the stress data-height curve, To obtain the absolute value sign.

5. The online monitoring method of a transmission tower according to claim 3, characterized in that: The transmission tower is monitored according to the optimal denoising result, including: displaying the optimal denoising result in a visual manner to achieve monitoring of the transmission tower.

6. The online monitoring method of a transmission tower according to claim 3, characterized in that: The local range of the stress data point is composed of the stress data point and its left and right sides. stress data points together constitute is the default value.

7. The online monitoring method of a transmission tower according to claim 4, characterized in that: The method for obtaining the fitting loss is: In the stress data-time curve, calculate the stress data points and stress data points and The integral area enclosed by the axis, and the stress data points and stress data points and The area of ​​the trapezoid enclosed by the axis, the absolute value of the difference between the integral area and the trapezoid area is taken as the first stress data points and The fitting loss between the stress data points.

8. The online monitoring method of a transmission tower according to claim 1, characterized in that: The method for pre-generating a plurality of average times is as follows: presetting an initial average number as C, an iteration step as 1, and gradually increasing the iteration step to obtain a plurality of average times.

9. An online monitoring system for a transmission tower, characterized in that: The monitoring system comprises a memory and a processor, wherein a computer program is stored in the memory, and the processor executes the computer program to implement the steps of the monitoring method according to any one of claims 1 to 8.

Images