Irregular breaking wave load signal processing and peak value extraction method

Abstract

This invention discloses a method for processing and extracting peak values ​​of irregular breaking wave load signals. The method acquires the total load time series from deep-water wave surfaces and shallow-water broken zones; uses LOESS smoothing to separate the total load into quasi-static and dynamic loads; sets candidate thresholds based on the quasi-static load, and identifies points crossing the thresholds in the same direction to divide the period; iteratively adjusts the thresholds to ensure that the relative difference between the average period of the quasi-static load and the input wave period is less than a preset critical value, determining the final period division point; accordingly, it simultaneously divides the three load sequences and extracts the peak values ​​for statistical analysis. The extraction method disclosed in this invention overcomes the period division distortion caused by high-frequency oscillations of breaking wave loads, achieves effective load separation and accurate peak value extraction, and improves the accuracy of load assessment for marine structures.

Description

Technical Field

[0001] This invention relates to the interdisciplinary field of marine engineering and signal processing, specifically to a method for processing irregular breaking wave load signals and extracting peak values, applicable to load signal analysis and feature identification of marine structures (such as piles, foundations, etc.) under the action of breaking waves in nearshore shallow waters. Background Technology

[0002] With the deepening development of marine resources, an increasing number of marine engineering structures (such as offshore wind power foundations, cross-sea bridge piers, nearshore breakwaters, and offshore platforms) are being constructed in shallow nearshore areas. In these areas, irregular waves undergo intense nonlinear deformation and even breakage due to the influence of seabed topography during propagation, forming breaking waves. The loads exerted by breaking waves on marine structures contain slowly varying quasi-static components, superimposed with transient high-frequency dynamic impact components with extremely strong destructive power. Therefore, accurately processing irregular breaking wave load signals and extracting their peak characteristics is of significant engineering importance for the safety design, fatigue assessment, and health monitoring of marine structures.

[0003] In existing wave load signal processing techniques, it is typically necessary to first divide the load time series into periods, and then extract the load peak value within each period for statistical analysis. Currently, conventional processing methods are prone to failure under breaking wave conditions. Existing technologies generally employ the "zero-crossing method" (such as crossing the upper or lower zero point) to divide the load time series into periods. However, under irregular breaking wave action, the load signal is superimposed with strong high-frequency dynamic impact components, causing the signal to generate violent high-frequency oscillations near zero (i.e., the "pseudo-zero-crossing" phenomenon). If the zero-crossing method is used directly, a true wave period will be incorrectly divided into multiple false short periods, or the true wave period will be missed due to oscillation masking, resulting in severe distortion of period identification and rendering subsequent peak value statistics meaningless.

[0004] In summary, how to effectively overcome the interference of high-frequency oscillations of broken waves on period identification and adaptively determine the optimal period division threshold to ensure the physical authenticity of load peak extraction is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] The technical problem to be solved

[0006] The purpose of this invention is to provide a method for processing irregular breaking wave load signals and extracting peak values, thereby addressing the technical deficiencies of existing zero-point crossing methods in processing breaking wave impact load signals. This method effectively overcomes the technical problems of pseudo-period division, inaccurate period identification, and inaccurate load peak value extraction caused by high-frequency oscillations and repeated local zero-point crossings in the signal near the zero point. This invention decomposes the total load into quasi-static and dynamic load components by performing local weighted regression smoothing on the total load time series. Using the quasi-static load time series as the period division object, a cross-threshold method is employed to identify the load period. Combined with input wave period constraints and an iterative optimization strategy for candidate thresholds, the optimal load threshold matching the incident wave period and the final period division point are determined, thus improving the accuracy of total load, quasi-static load, and dynamic load peak value extraction.

[0007] Technical solution

[0008] To address the aforementioned technical problems, this invention provides the following technical solution: a method for processing and extracting peak values ​​of irregular breaking wave load signals, comprising the following steps:

[0009] S1. Collect the wavefront time series of unbroken incident waves in deep water and the total load time series of marine structures in shallow water broken areas.

[0010] S2. The wave surface time series is divided into periods using the zero-point crossing method to obtain the average period of the incident wave in the deep water area, which is denoted as the input wave period.

[0011] S3. Perform local weighted regression smoothing on the total load time series to separate the total load time series into a quasi-static load time series and a dynamic load time series; wherein, the quasi-static load time series is the low-frequency smoothed load component obtained by local weighted regression smoothing, and the dynamic load time series is the difference between the total load time series and the quasi-static load time series.

[0012] S4. Using the quasi-static load time series as the periodic division object, set a quasi-static load candidate threshold, and identify adjacent cross-threshold points in the same direction in the quasi-static load time series.

[0013] S5. Determine the load cycle interval based on the adjacent cross threshold points in the same direction, extract the quasi-static load peak value in each load cycle interval, and calculate the cycle corresponding to each load cycle interval to obtain the quasi-static load average cycle.

[0014] S6. Compare the average period of the quasi-static load with the input wave period. If the relative difference between the two is less than or equal to a preset critical value, then determine the current quasi-static load candidate threshold as the final threshold and determine the current same-direction threshold crossing point as the final period division point. If the relative difference between the two is greater than the preset critical value, then adjust the quasi-static load candidate threshold according to the preset adjustment strategy and repeat steps S4 to S5 until a final threshold and a final period division point that meet the conditions are determined.

[0015] S7. Based on the final period division point, the total load time series, the quasi-static load time series, and the dynamic load time series are divided into corresponding load period intervals, and the total load peak value, quasi-static load peak value, and dynamic load peak value are extracted in each load period interval respectively.

[0016] S8. Statistically analyze the total load peak value, the quasi-static load peak value and the dynamic load peak value to obtain the load peak value statistical results of the marine structure under the action of irregular breaking waves.

[0017] Further, in step S2, the zero-point crossing method is the upper zero-point crossing method, the lower zero-point crossing method, or a combination of the two; when the upper zero-point crossing method is used, the time interval between adjacent upper zero-points in the wave surface time series is identified, and the input wave period is calculated based on the multiple time intervals.

[0018] Furthermore, in step S3, the local weighted regression smoothing process is LOESS smoothing, which obtains the quasi-static load time series by performing weighted polynomial regression fitting on the data points in the local neighborhood of the total load time series.

[0019] Further, in step S3, the method for selecting the span parameter Span of the LOESS smoothing process is to calculate the number of sampling points within a single average wave cycle based on the sampling frequency of the input wave cycle and the load signal, and set the size of the local neighborhood window of LOESS smoothing to include the number of data points containing 0.3 to 0.6 times the number of sampling points, thereby calculating the span parameter Span.

[0020] Further, in step S3, the dynamic load time series is determined as follows: F d (t) = F t (t)-F q (t), where F t (t) represents the total load time series, F q (t) is the quasi-static load time series, F d (t) represents the dynamic load time series, where t is the time variable;

[0021] Further, in step S4, the initial value of the quasi-static load candidate threshold is set to 0; the same-direction crossing threshold point is two adjacent upward crossing threshold points or two adjacent downward crossing threshold points; the upward crossing threshold point is the moment when the load value in the quasi-static load time series jumps from below the current quasi-static load candidate threshold to above the current quasi-static load candidate threshold; the downward crossing threshold point is the moment when the load value in the quasi-static load time series decreases from above the current quasi-static load candidate threshold to below the current quasi-static load candidate threshold.

[0022] Further, in step S5, the time interval between two adjacent same-direction threshold crossing points is defined as a load period interval; within the load period interval, the maximum value of the quasi-static load time series is selected as the quasi-static load peak value of the load period interval, and the difference between the corresponding times of two adjacent same-direction threshold crossing points is used as the period of the load period interval.

[0023] Further, in step S6, the preset threshold value is 2%~15%, preferably 5%; the relative difference is calculated as follows: δ = |T q -T w | / T w ×100%, where δ is the relative difference, T q T is the average period of the quasi-static load. w The input wave period;

[0024] Further, in step S6, the preset adjustment strategy is adjusted to: increase the quasi-static load candidate threshold by a preset step size, and re-identify the same-direction threshold crossing point, re-divide the load period interval, and recalculate the quasi-static load average period.

[0025] Furthermore, in step S6, the iterative adjustment process also includes a safe termination condition judgment: if the current quasi-static load candidate threshold exceeds the preset proportion of the maximum peak value of the quasi-static load time series, or the current number of iterations exceeds the preset maximum number of iterations, then the loop is forcibly terminated; when the loop is forcibly terminated, the quasi-static load candidate threshold corresponding to the iteration with the smallest relative difference calculated in each iteration is extracted and determined as the final threshold.

[0026] Further, the preset step size is 0.005~0.05, preferably 0.01; or, the preset step size is 0.005 to 0.05 times the load amplitude or normalized load amplitude of the quasi-static load time history, preferably 0.01 times.

[0027] Further, in step S7, the final period division point is determined by the quasi-static load time series and is simultaneously applied to the total load time series and the dynamic load time series, so that the total load peak, the quasi-static load peak and the dynamic load peak are extracted correspondingly within the same load period interval.

[0028] Further, in step S8, the statistical analysis includes: arranging the total load peak sequence, quasi-static load peak sequence, and dynamic load peak sequence in descending order of peak value; calculating the average value of each peak sequence as the average load of the corresponding load; calculating the average value of the largest third peak in each peak sequence as the effective load of the corresponding load; and calculating the average value of the largest tenth peak in each peak sequence as 1 / 10 of the load of the corresponding load.

[0029] Beneficial effects

[0030] Compared with the prior art, the present invention has at least the following beneficial effects:

[0031] 1. This invention uses the quasi-static load time series after LOESS smoothing as the basis for period division, which can effectively reduce the interference of high-frequency oscillations caused by break wave impact on zero point identification and reduce the problem of missegmentation caused by pseudo-zero point crossing.

[0032] 2. This invention uses the average period of unbroken incident waves in deep water as a physical constraint. The final threshold is determined by comparing the relative difference between the average period of the quasi-static load and the period of the input wave, so that the load period division result has a good correspondence with the characteristics of the incident wave.

[0033] 3. This invention, through candidate threshold iterative optimization and safe termination mechanism, can automatically determine a suitable threshold under different wave conditions, different structural forms and different load amplitudes, thereby improving the adaptability and stability of the load period division process.

[0034] 4. This invention uses a unified final period division point to synchronously segment and extract peak values ​​of total load, quasi-static load and dynamic load, ensuring that the peak values ​​of different load components have a corresponding relationship within the same wave period, which is beneficial for carrying out load composition analysis and statistical characteristic comparison. Attached Figure Description

[0035] Figure 1 A flowchart of a method for processing and extracting peak values ​​of irregular breaking wave load signals provided by the present invention;

[0036] Figure 2a This is a full-length curve of the total time series of the breaking wave load obtained from actual measurements;

[0037] Figure 2b for Figure 2a A magnified curve of the period from 110s to 140s.

[0038] Figure 3a This is a full-length curve of the quasi-static load time series obtained by LOESS smoothing.

[0039] Figure 3b for Figure 3a A magnified curve of the period from 110s to 140s.

[0040] Figure 4a The full-length curve of the dynamic load time series obtained by separation;

[0041] Figure 4b for Figure 4a A magnified curve of the period from 110s to 140s.

[0042] Figure 5a This is a schematic diagram of the quasi-static load peak value and threshold crossing point extracted using the method of the present invention;

[0043] Figure 5b for Figure 5a A magnified view of the localized results during the 110s to 140s time period;

[0044] Figure 6a This is a schematic diagram of the dynamic load peak value extracted using the method of the present invention;

[0045] Figure 6b for Figure 6a A magnified view of the localized results during the 110s to 140s time period;

[0046] Figure 7a This is a schematic diagram of the peak value of the total load of the breaking wave extracted using the method of this invention.

[0047] Figure 7b for Figure 7a A magnified view of the localized results during the 110s to 140s time period. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0049] like Figure 1 As shown, the present invention provides a method for processing and extracting peak values ​​of irregular breaking wave load signals, which mainly includes the following steps:

[0050] Step S1: Collect the wavefront time series of unbroken incident waves in deep water and the total load time series corresponding to the marine structure in the shallow water fractured area. In a specific embodiment of the present invention, the wavefront time series η(t) is collected by a wave height meter deployed in the deep water area, and the total load time series F is collected by a force sensor deployed on the marine structure (such as a monopile foundation or pier) in the shallow water fractured area. t (t), In order to ensure that the high-frequency dynamic impulse characteristics of the signal are not lost due to aliasing, the sampling frequency fs is preferably set to 100 Hz or higher (e.g., 1000 Hz).

[0051] Step S2: Divide the wave surface time series into periods using the zero-point crossing method to obtain the average period of the incident wave in the deep water area, denoted as the input wave period T. w Specifically, the zero-point crossing method can employ the upper zero-point crossing method, the lower zero-point crossing method, or a combination of both. In this embodiment, the upper zero-point crossing method is preferred, which involves identifying the time interval between adjacent upper zero-point crossings (i.e., the moments when the wave surface crosses the still water surface zero point from bottom to top) in the wave surface time series η(t) to obtain the individual period of each incident wave. Then, the average of all individual periods is calculated to obtain the input wave period T. w .

[0052] Step S3: Perform local weighted regression smoothing on the total load time series to separate it into a quasi-static load time series and a dynamic load time series. Specifically, LOESS (Locally Estimated Scatterplot Smoothing) smoothing is used on F. t (t) is separated to obtain a low-frequency smooth quasi-static load time series F. q (t); Dynamic load time series F d (t) according to formula F d (t) = F t (t)-F q (t) is determined.

[0053] To ensure effective separation, the span parameter (Span) for LOESS smoothing is selected using the following strategy: based on the average period T of the input wave obtained in step S2... w And the sampling frequency fs, to calculate the number of sampling points N within a single average wave period. w =T w ×fs. Set the size of the local neighborhood window for LOESS smoothing to include 0.3 to 0.6 times (preferably 0.4 times) N. wThe number of data points is used to calculate the span parameter Span. This method ensures a reasonable match between the LOESS smoothing window and the input wave period corresponding to the breaking wave load, preventing the smoothing window from being too small, which would retain too much high-frequency oscillation, or too large, which would excessively weaken the quasi-static load variation characteristics.

[0054] Step S4: Using the quasi-static load time series as the periodic division object, set a quasi-static load candidate threshold and identify adjacent same-direction threshold crossing points in the quasi-static load time series. Specifically, the initial value of the quasi-static load candidate threshold is set to 0. The same-direction threshold crossing points include two adjacent upward threshold crossing points or two adjacent downward threshold crossing points. In this embodiment, it is preferable to use two adjacent upward threshold crossing points, where the upward threshold crossing point refers to F q The moment when the load value (t) jumps from below the current candidate threshold to above the current candidate threshold. By identifying two adjacent upper threshold points, the "pseudo-zero crossing" interference caused by high-frequency oscillations near zero in the total load signal is effectively avoided. It is understood that in other equivalent embodiments, two adjacent lower threshold points (i.e., F) can also be used. q The periodic division of (t) is based on the time point when the load value drops from above the current candidate threshold to below the current candidate threshold. The basic principle is the same as that of using the upper threshold point, and both are within the protection scope of this invention.

[0055] Step S5: Determine the load period interval based on adjacent cross-threshold points in the same direction. Extract the quasi-static load peak value within each load period interval and calculate the corresponding period for each load period interval to obtain the quasi-static load average period T. q Specifically, the time interval between two adjacent threshold points is defined as a load cycle interval. Within this interval, F is selected. q The maximum value of (t) is taken as the quasi-static load peak value, and the difference between the corresponding times of two adjacent upper threshold points is taken as the period of the interval. The average value of all the divided periods is used to obtain T. q .

[0056] Step S6: The average period T of the quasi-static load q With the input wave period T w The comparison is performed, and the threshold is iteratively optimized. The relative difference δ = |T q -T w | / T w If δ is less than or equal to the preset critical value, then the current candidate threshold is determined as the final threshold, and the current point crossing the threshold in the same direction is determined as the final period division point. If δ is greater than the preset critical value, then the candidate threshold is increased by a preset step size, and steps S4 to S5 are repeated.

[0057] In a preferred embodiment of the present invention, the above-mentioned iterative parameters were deeply optimized: (1) The preset critical value was set to 2%~15%, preferably 5%. The setting of this numerical range is based on the balance between the nonlinear deformation characteristics of irregular breaking waves and the convergence efficiency of the algorithm: If the preset critical value is set too small (for example, less than 2%): because the waves will produce strong nonlinear deformation and phase shift when breaking in shallow water, the load period and the incident wave period in deep water will have inherent physical differences. Too small a critical value will make it very difficult for the algorithm to meet the convergence condition in step S6, resulting in an increase in the number of iterations and a reduction in the real-time performance of signal processing. If the preset critical value is set too large (for example, greater than 15%): although the algorithm is very easy to converge, it will cause the divided load period to deviate too much from the real incident wave period, missing the real wave period (i.e., the 'leaking wave' phenomenon), which will lead to serious distortion of the quasi-static peak value and dynamic impact peak value extracted later, reducing the engineering reference value. (2) The preset step size is set to 0.005~0.05, preferably 0.01, or 0.005 times~0.05 times the load amplitude, preferably 0.01 times. If the step size is too small (e.g., less than 0.005), it will cause a surge in the number of iterations when processing 1000 Hz high-frequency data, which will not meet the requirements of quasi-real-time monitoring; if the step size is too large (e.g., greater than 0.05), it is easy to cross the optimal threshold point. (3) Introduce a safe termination condition: if the current candidate threshold exceeds 80% of the maximum peak value of the quasi-static load time series, or the current number of iterations exceeds the preset maximum number of iterations (e.g., 10,000 times), the loop will be forcibly terminated. When forcibly terminated, the candidate threshold corresponding to the iteration with the smallest relative difference δ calculated in each iteration is extracted and determined as the final threshold. This design prevents dead loops caused by extreme distorted signals and ensures the engineering robustness of the algorithm.

[0058] Step S7: Based on the final period division point, the total load time series, the quasi-static load time series, and the dynamic load time series are divided into corresponding load period intervals, and the peak values ​​of the total load, quasi-static load, and dynamic load are extracted from each interval. This method ensures that the peak values ​​of the total load, quasi-static load, and dynamic impact load are extracted within the exact same physical period interval, clarifying the correspondence between static and dynamic loads.

[0059] Step S8: Statistically analyze the extracted peak values ​​to obtain the load peak value statistics. Specifically, arrange the total load peak value sequence, quasi-static load peak value sequence, and dynamic load peak value sequence in descending order of peak value size. Calculate the average value of each sequence as the average load; calculate the average value of the largest third peak value in each sequence as the effective load of the corresponding load; calculate the average value of the largest tenth peak value in each sequence as 1 / 10 of the corresponding load, thus comprehensively reflecting the statistical characteristics of the breaking wave load.

[0060] To more intuitively demonstrate the processing effect of this invention, the following are a set of processing examples based on experimental data from physical model tests:

[0061] like Figure 2a and Figure 2b The figures shown are the full-length curve of the measured total load time series of the breaking wave and a magnified local curve in the 110s to 140s time period. From... Figure 2a The macroscopic fluctuation trend of the total load can be observed; and from Figure 2b The magnified details clearly show that, due to the impact of the breaking wave, the total load signal is superimposed with a strong high-frequency dynamic impact component, causing the signal to oscillate violently near the zero mean line. If the traditional zero-point crossing method is used directly... Figure 2b The dense "pseudo-zero crossings" will lead to a large number of false short cycles.

[0062] like Figure 3a and Figure 3b The figures shown are the full-length curve and a magnified portion of the quasi-static load time series obtained after LOESS smoothing. The LOESS smoothing process (setting the span to cover 0.4 average wave periods) in step S3 of this invention is used to obtain the quasi-static load time series. Figure 3b The magnified details show that the high-frequency oscillations of the load were successfully filtered out, resulting in a smooth quasi-static load time series curve that retains the true low-frequency fluctuation characteristics.

[0063] like Figure 4a and Figure 4b The figures show the full-length curve and a magnified portion of the separated dynamic load time series, respectively. It can be seen that the dynamic load curve exhibits typical transient high-frequency impact characteristics, showing a rapid rise to a peak within a short period followed by oscillating decay.

[0064] like Figures 5a to 7b Figure 1 shows a schematic diagram (Figure a) of the full-length results of the quasi-static load peak value, dynamic load peak value, and total breaking wave load peak value extracted using the method of this invention, and a schematic diagram (Figure b) of its partially enlarged results. Figure 5a and 5b In the diagram, the red dot represents the final period division point determined through iterative optimization in step S6 (where the final relative difference δ converges to 1.95%, satisfying the critical value of less than 5%). Figures 5a to 7b It can be seen that within each defined load cycle interval, a total load peak, a quasi-static peak, and a dynamic peak are accurately extracted, proving that the method of the present invention solves the problem of cycle identification distortion in traditional methods, achieves strict synchronous extraction of multi-dimensional load peaks, and can provide reliable data support for structural wave-resistant design.

[0065] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for processing and extracting peak values ​​of irregular breaking wave load signals, characterized in that, Includes the following steps: S1. Collect the wavefront time series of unbroken incident waves in deep water and the total load time series of marine structures in shallow water broken areas. S2. The wave surface time series is divided into periods using the zero-point crossing method to obtain the average period of the incident wave in the deep water area, which is denoted as the input wave period. S3. Perform local weighted regression smoothing on the total load time series to separate the total load time series into a quasi-static load time series and a dynamic load time series; wherein, the quasi-static load time series is the low-frequency smoothed load component obtained by local weighted regression smoothing, and the dynamic load time series is the difference between the total load time series and the quasi-static load time series. S4. Using the quasi-static load time series as the periodic division object, set a quasi-static load candidate threshold, and identify adjacent cross-threshold points in the same direction in the quasi-static load time series. S5. Determine the load cycle interval based on the adjacent cross threshold points in the same direction, extract the quasi-static load peak value in each load cycle interval, and calculate the cycle corresponding to each load cycle interval to obtain the quasi-static load average cycle. S6. Compare the average period of the quasi-static load with the input wave period. If the relative difference between the two is less than or equal to a preset critical value, then determine the current quasi-static load candidate threshold as the final threshold and determine the current same-direction threshold crossing point as the final period division point. If the relative difference between the two is greater than the preset critical value, then adjust the quasi-static load candidate threshold according to the preset adjustment strategy and repeat steps S4 to S5 until a final threshold and a final period division point that meet the conditions are determined. S7. Based on the final period division point, the total load time series, the quasi-static load time series, and the dynamic load time series are divided into corresponding load period intervals, and the total load peak value, quasi-static load peak value, and dynamic load peak value are extracted in each load period interval respectively. S8. Statistically analyze the total load peak value, the quasi-static load peak value, and the dynamic load peak value to obtain the statistical results of the load peak value of the marine structure under the action of irregular breaking waves.

2. The method for processing irregular breaking wave load signals and extracting peak values ​​according to claim 1, characterized in that, In step S2, the zero-point crossing method is the upper zero-point crossing method, the lower zero-point crossing method, or a combination of the two; when the upper zero-point crossing method is used, the time interval between adjacent upper zero-points in the wave surface time series is identified, and the input wave period is calculated based on the multiple time intervals.

3. The method for processing irregular breaking wave load signals and extracting peak values ​​according to claim 1, characterized in that, In step S3, the local weighted regression smoothing process is LOESS smoothing, which obtains the quasi-static load time series by performing weighted polynomial regression fitting on the data points in the local neighborhood of the total load time series.

4. The method for processing irregular breaking wave load signals and extracting peak values ​​according to claim 3, characterized in that, The method for selecting the span parameter Span in the LOESS smoothing process involves calculating the number of sampling points within a single average wave cycle based on the input wave cycle and the sampling frequency of the load signal, setting the size of the local neighborhood window for LOESS smoothing to include 0.3 to 0.6 times the number of sampling points, and then calculating the span parameter Span.

5. The method for processing irregular breaking wave load signals and extracting peak values ​​according to claim 1, characterized in that, In step S3, the dynamic load time series is determined as follows: F d (t) = F t (t)-F q (t), where F t (t) represents the total load time series, F q (t) is the quasi-static load time series, F d (t) represents the dynamic load time series, where t is the time variable.

6. In the method for processing irregular broken wave load signals and extracting peak values ​​according to claim 1, in step S4, the initial value of the quasi-static load candidate threshold is set to 0; the same-direction crossing threshold point is two adjacent upper crossing threshold points or two adjacent lower crossing threshold points; the upper crossing threshold point is the moment when the load value in the quasi-static load time series jumps from below the current quasi-static load candidate threshold to above the current quasi-static load candidate threshold; the lower crossing threshold point is the moment when the load value in the quasi-static load time series decreases from above the current quasi-static load candidate threshold to below the current quasi-static load candidate threshold.

7. The method for processing and extracting peak values ​​of irregular breaking wave load signals according to claim 1, characterized in that, In step S5, the time interval between two adjacent same-direction threshold crossing points is defined as a load period interval; within the load period interval, the maximum value of the quasi-static load time series is selected as the quasi-static load peak value of the load period interval, and the difference between the corresponding times of two adjacent same-direction threshold crossing points is used as the period of the load period interval.

8. The method for processing irregular breaking wave load signals and extracting peak values ​​according to claim 1, characterized in that, In step S6, the preset threshold value is 2%~15%; the relative difference is calculated as follows: δ = |T q -T w | / T w ×100%, where δ is the relative difference, T q T is the average period of the quasi-static load. w The input wave period is the period of the wave.

9. The method for processing irregular breaking wave load signals and extracting peak values ​​according to claim 1, characterized in that, In step S6, the preset adjustment strategy is to increase the quasi-static load candidate threshold by a preset step size, and re-identify the same-direction threshold crossing point, re-divide the load period interval, and recalculate the quasi-static load average period.

10. The method for processing irregular breaking wave load signals and extracting peak values ​​according to claim 9, characterized in that, The preset step size is 0.005 to 0.05; or, the preset step size is 0.005 to 0.05 times the load amplitude or normalized load amplitude of the quasi-static load time history.

11. The method for processing irregular breaking wave load signals and extracting peak values ​​according to claim 1, characterized in that, In step S7, the final period division point is determined by the quasi-static load time series and is simultaneously applied to the total load time series and the dynamic load time series, so that the total load peak, the quasi-static load peak and the dynamic load peak are extracted correspondingly within the same load period interval.

12. The method for processing and extracting peak values ​​of irregular breaking wave load signals according to claim 1, characterized in that, In step S8, the statistical analysis includes: arranging the total load peak sequence, quasi-static load peak sequence, and dynamic load peak sequence in descending order of peak value; calculating the average value of each peak sequence as the average load of the corresponding load; calculating the average value of the largest third peak in each peak sequence as the effective load of the corresponding load; and calculating the average value of the largest tenth peak in each peak sequence as 1 / 10 of the load of the corresponding load.

13. The method for processing irregular breaking wave load signals and extracting peak values ​​according to claim 8, characterized in that, In the iterative adjustment process of step S6, a safe termination condition judgment is also included: if the current quasi-static load candidate threshold exceeds the preset proportion of the maximum peak value of the quasi-static load time series, or the current loop iteration number exceeds the preset maximum iteration number, the loop is forcibly terminated. When the loop is forcibly terminated, the quasi-static load candidate threshold corresponding to the iteration with the smallest relative difference calculated in each iteration is extracted and determined as the final threshold.

Images