Stress wave waveform period normalization method and high strain pile detection method

Abstract

The present application belongs to the technical field of pile detection technology, and relates to a stress wave waveform period normalization method and a high-strain pile detection method. The method comprises the following steps: obtaining stress wave data of continuous multiple hammering, performing Fourier transform to obtain amplitude spectrum and phase spectrum, calculating actual period according to the amplitude spectrum; calculating theoretical period according to the pile length and wave velocity, constructing a correction factor according to the actual period and the theoretical period; scaling the frequency axis according to the correction factor and performing frequency domain interpolation, combining the phase spectrum to perform inverse Fourier transform to obtain a period normalized waveform; calculating the matching degree and the signal-to-noise ratio of the period normalized waveform, determining a weight coefficient according to the matching degree and the signal-to-noise ratio, performing weighted superposition according to the weight coefficient to obtain an enhanced waveform. The present application eliminates the waveform period drift caused by the cumulative deformation of the hammer pad through frequency domain phase spectrum correction, effectively enhances the coherence of the multiple hammering waveforms, and improves the signal-to-noise ratio of the superimposed waveform and the pile bottom reflection recognition rate.

Description

Technical Field

[0001] This invention relates to the field of pile foundation testing technology, and in particular to a stress wave waveform period normalization method and a high-strain pile foundation testing method. Background Technology

[0002] High-strain pile testing involves using a heavy hammer to impact the pile top and generate stress waves. Based on one-dimensional stress wave theory, the integrity of the pile and its vertical compressive bearing capacity are analyzed. This is one of the core methods for evaluating pile quality in building construction. In this testing method, to ensure the pile head is not damaged by impact and to regulate the impact pulse pattern, the standard practice is to place a natural wooden board as a hammer pad between the hammer head and the pile top. However, during continuous hammering, the natural wooden board undergoes irreversible cumulative compressive deformation, and its equivalent stiffness increases with each hammer blow, causing a systematic drift in the waveform period of the stress wave generated by each blow.

[0003] Specifically, during the test of 10 consecutive hammer blows, the wooden board was in its initial state during the first hammer blow, with a waveform period of approximately 3.2 ms; by the fifth hammer blow, the wooden board had been partially compacted, and the period was shortened to approximately 2.8 ms; and by the tenth hammer blow, the wooden board was significantly compacted, and the period was further shortened to approximately 2.4 ms.

[0004] This periodic drift causes the stress wave waveforms generated by each hammer blow to become misaligned in the time domain. Even if multiple waveforms are superimposed and averaged according to the specifications, the effective signals will cancel each other out due to phase inconsistency. This makes it impossible to achieve the goal of improving the signal-to-noise ratio by superimposing and averaging multiple waveforms, thus affecting the identification of the pile bottom reflection signal and the accurate determination of the defect location. Summary of the Invention

[0005] Therefore, the purpose of this invention is to overcome the problem that waveform period drift caused by cumulative deformation of the hammer pad, resulting in the inability to effectively coherently enhance waveforms after multiple hammer blows. This invention provides a stress wave waveform period normalization method and a high-strain pile detection method. By achieving stress wave period normalization through frequency domain phase spectrum correction and combining it with adaptive superposition with matching degree weighting, the waveform period drift caused by cumulative deformation of the hammer pad in high-strain pile detection is eliminated, enabling effective coherent enhancement of waveforms after multiple hammer blows. This effectively improves the signal-to-noise ratio of the superimposed waveform and the pile bottom reflection recognition rate.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a method for normalizing the period of a stress wave waveform, comprising: Acquire stress wave data generated by repeated hammer blows to the pile foundation; perform Fourier transform on the stress wave data to obtain the amplitude spectrum A. i (f) and phase spectrum φ i (f); Calculate the actual period T based on the frequency corresponding to the main peak value in the amplitude spectrum. i; i is the hammer strike number; The theoretical period T is calculated based on the pile length and stress wave velocity, combined with the actual period T. i Constructing a correction factor α with the theoretical period T i =T i / T;ɑ i is the correction factor for the stress wave of the i-th hammer impact; The frequency axis is scaled according to the correction factor, and the scaled amplitude spectrum A' is obtained by frequency domain interpolation. i (f); combined with the amplitude spectrum A' i (f) and phase spectrum φ i (f) Perform an inverse Fourier transform to obtain the period-normalized stress wave data W' i (t); For the stress wave data after period normalization, the matching degree between it and the actual period is calculated, and the waveform signal-to-noise ratio is calculated. The weighting coefficient is determined according to the matching degree and the waveform signal-to-noise ratio. All the waveforms after period normalization are weighted and superimposed according to their respective weighting coefficients to obtain the enhanced waveform for high strain pile detection.

[0007] Preferably, calculating the matching degree includes: ;ρ i T represents the matching degree of the i-th hammer strike; i Let be the actual period of the i-th hammer strike.

[0008] Preferably, the weighting coefficient ω is determined. i include: ;ω i SNR is the stress wave weighting coefficient for the i-th hammer blow; i denoted as , where is the stress wave signal-to-noise ratio of the i-th hammer strike; N is the number of hammer strikes; and j is the summation index variable.

[0009] Preferably, scaling the frequency axis according to the correction factor includes: f' = f / ɑ i f is the frequency value before scaling; f' is the frequency value after scaling.

[0010] Preferably, after obtaining the period-normalized stress wave data W' i (t) then includes: calculating the phase consistency index of each waveform and the reference waveform within the pile bottom reflection time window; if the phase consistency index of any waveform does not meet the requirements, the phase spectrum of the waveform is adjusted: within the frequency range corresponding to the pile bottom reflection, a linear correction term φ=2π·Δt·f' is applied to the phase spectrum to align the pile bottom reflection of the waveform with the reference waveform; Δt is the compensation time shift amount.

[0011] Preferably, the phase consistency index between each waveform and the reference waveform within the pile bottom reflection time window is calculated, including: for the stress wave data W' i (t) and the reference waveform are used to extract waveform segments within the pile bottom reflection time window, and Fourier transforms are performed on the waveform segments to obtain their respective window phase spectra; based on the stress wave data W' i The phase difference between the window phase spectra of the reference waveform and the reference waveform is calculated to obtain a phase difference sequence. A first-order linear fit is performed on the phase difference sequence to obtain the fitting slope and fitting residual. The inverse of the fitting slope is used as the compensation time shift. The reciprocal of the sum of squares of the fitting residuals is used as the phase consistency index.

[0012] Secondly, to solve the above-mentioned technical problems, the present invention also provides a method for detecting high-strain foundation piles, comprising: A vertical pad is placed on the top of the pile, and force sensors and acceleration sensors are symmetrically installed on the pile body at a distance of twice the pile diameter from the top of the pile. The pile top is repeatedly hammered with a heavy hammer, and the stress wave data generated by each hammering is collected by the force sensor and acceleration sensor to obtain the original waveform sequence. The original waveform sequence is processed based on the stress wave waveform period normalization method described above to obtain an enhanced waveform; Defect detection of the pile foundation is performed based on the enhanced waveform.

[0013] Preferably, defect detection of the pile foundation based on the enhanced waveform includes: performing time-frequency analysis on the enhanced waveform to identify the pile bottom reflection signal and the pile body defect reflection signal; calculating the vertical compressive bearing capacity of a single pile based on the arrival time and amplitude of the pile bottom reflection signal; and calculating the defect location based on the arrival time of the pile body defect reflection signal.

[0014] Preferably, the drop pad is configured as a wooden board.

[0015] Preferably, after obtaining the original waveform sequence, the method further includes: calculating the signal-to-noise ratio (SNR) of the original waveform and calculating the median of the SNR of all waveforms; if the SNR of any waveform is less than one-third of the median, then the weighting coefficient of the corresponding waveform is assigned to zero.

[0016] Compared with the prior art, the above-described technical solution of the present invention has the following advantages: First, the theoretical period calculated using pile length and stress wave velocity is used as the normalization benchmark. This benchmark is independent of the state of the hammer pad and the number of hammer blows, and does not need to rely on any measured waveform as a reference, thus improving the consistency of the normalized waveform.

[0017] Secondly, frequency domain phase spectrum correction avoids the destruction of waveform transient characteristics by time domain interpolation. The preservation of the phase spectrum ensures that the internal temporal relationship of the waveform is completely preserved. This is crucial for pile foundation inspection that relies on the time delay of reflected waves for defect location. The accuracy of defect location will not be reduced by the period normalization operation.

[0018] Furthermore, the adaptive weighted superposition based on matching degree and signal-to-noise ratio automatically reduces the contribution of waveforms with large period deviations or low signal-to-noise ratios to the superposition result, effectively suppresses the contamination of abnormal waveforms, and significantly improves the signal-to-noise ratio after superposition.

[0019] In summary, the stress wave waveform period normalization method and high-strain pile detection method described in this invention achieve stress wave period normalization through frequency domain phase spectrum correction, and combine it with matching degree weighted adaptive superposition to eliminate waveform period drift caused by cumulative deformation of the hammer pad in high-strain pile detection, so that the waveform of multiple hammer blows can be effectively coherently enhanced, thereby effectively improving the signal-to-noise ratio of the superimposed waveform and the pile bottom reflection recognition rate. Attached Figure Description

[0020] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart of the stress wave waveform period normalization method in a preferred embodiment of the present invention; Figure 2 This is a flowchart illustrating the determination of phase consistency index in a preferred embodiment of the present invention; Figure 3 This is a flowchart of a high-strain foundation pile testing method in a preferred embodiment of the present invention. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0022] In high-strain pile testing, the standard practice is to place a natural wooden board as a hammer pad between the hammer and the pile top, using its plastic deformation to adjust the impact pulse pattern. With each consecutive hammer blow, the internal fibrous structure of the wooden hammer pad undergoes irreversible cumulative compressive deformation, and its equivalent stiffness increases progressively, causing a systematic drift in the period of the stress wave generated by each blow. This periodic drift makes the waveforms of each blow misaligned in the time domain. Even if multiple waveforms are superimposed and averaged according to current testing specifications, the effective signals will cancel each other out due to phase inconsistencies. The supposedly enhanced pile bottom reflection signal is instead weakened, affecting the accuracy of pile integrity assessment and the reliability of bearing capacity calculation.

[0023] For processing stress wave data, existing techniques propose methods to adjust waveform periods through time-domain resampling. Specifically, this method first calculates the actual period of each waveform, selects one waveform as a reference, then changes the sampling interval of the other waveforms and performs resampling using linear interpolation or spline interpolation to make the periods of all waveforms consistent with the reference waveform. While this method can theoretically unify the periods of different waveforms, its implementation has inherent limitations.

[0024] First, this method requires manually selecting a waveform as a reference, and the selection of the reference waveform directly affects the correction results of all waveforms. In actual testing, the deformation of the hammer pad exhibits a monotonic evolution trend, and the waveform period of the first and last hammer blows differs significantly. There is no statistically representative reference waveform, and the arbitrariness in the selection of the reference will introduce systematic bias.

[0025] Secondly, the time-domain resampling combined with interpolation method requires two interpolation operations: first, resampling at the new sampling interval, and then interpolating back to the original sampling interval. Stress wave waveforms have sharp rising edges and rapidly decaying oscillating characteristics. These transient components are key features of stress waveforms. Two interpolations will smooth out the steepness of the rising edge, the amplitude of the decay peak, and change the phase relationship between various frequency components, resulting in waveform distortion.

[0026] The purpose of this invention is to overcome the problem that waveform period drift caused by cumulative deformation of the hammer pad, resulting in ineffective coherent enhancement of waveforms after multiple hammer blows. It provides a stress wave waveform period normalization method and a high-strain pile detection method. The method achieves stress wave period normalization through frequency domain phase spectrum correction and combines it with adaptive superposition with matching degree weighting to eliminate waveform period drift caused by cumulative deformation of the hammer pad in high-strain pile detection. This enables effective coherent enhancement of waveforms after multiple hammer blows, thereby effectively improving the signal-to-noise ratio of the superimposed waveform and the pile bottom reflection recognition rate.

[0027] Example 1: This embodiment of the invention provides a stress wave waveform period normalization method for high-strain foundation pile testing, referring to... Figure 1 As shown, the method specifically includes the following steps: S100, Obtain stress wave data W generated by continuous multiple hammer blows on the pile foundation. i (t); Perform a Fourier transform on the stress wave data to obtain the amplitude spectrum A. i (f) and phase spectrum φ i (f); Calculate the actual period T based on the frequency corresponding to the main peak value in the amplitude spectrum. i ; i is the hammer strike number; S200. Calculate the theoretical period T based on the pile length and stress wave velocity, and combine it with the actual period T. i Constructing a correction factor α with the theoretical period T i =Ti / T;ɑ i is the correction factor for the stress wave of the i-th hammer impact; S300. Scale the frequency axis according to the correction factor, and obtain the scaled amplitude spectrum A' through frequency domain interpolation. i (f); combined with the amplitude spectrum A' i (f) and phase spectrum φ i (f) Perform an inverse Fourier transform to obtain the period-normalized stress wave data W' i (t); S400. For the stress wave data after period normalization, calculate the matching degree between it and the actual period, and calculate the waveform signal-to-noise ratio. Determine the weighting coefficient based on the matching degree and the waveform signal-to-noise ratio. Weight all the waveforms after period normalization are superimposed according to their respective weighting coefficients to obtain the enhanced waveform for high strain pile detection.

[0028] In specific implementation step S100, force sensors and acceleration sensors installed near the pile top at a distance of twice the pile diameter are used to collect stress wave data generated by multiple consecutive hammer blows; a set of stress wave data W is collected for each hammer blow. i (t), where i is the hammer strike number; t is the time; W i (t) represents the stress wave data of the i-th hammer blow.

[0029] For each stress wave data W i Perform a Fourier transform on (t) to obtain its expression in the frequency domain: F i (f)=A i (f)e jφ i (f) A i (f) represents the amplitude spectrum corresponding to the stress wave data of the i-th hammer blow, indicating the magnitude of each frequency component; φ i (f) represents the phase spectrum corresponding to the stress wave data of the i-th hammer blow, indicating the initial phase of each frequency component; j is the imaginary unit; e is the logarithm.

[0030] In the amplitude spectrum A i The frequency with the largest search amplitude in (f) is denoted as the dominant frequency f. i The actual period of the hammer impact stress wave was calculated based on the dominant frequency: T i =1 / f i Actual period T i The instantaneous stiffness of the hammer pad reflects the impact pulse state during the hammering: the harder the hammer pad, the narrower the impact pulse, the higher the main frequency, and the shorter the actual cycle.

[0031] In specific implementation step S200, the theoretical period T is calculated based on the pile length L and the propagation speed C of the stress wave in the pile: T = 2L / C. This theoretical period T represents the round-trip time required for the stress wave to propagate from the top of the pile to the bottom of the pile and back to the top of the pile. It is a constant value determined by the inherent physical parameters of the pile and does not depend on any measured waveform.

[0032] Then, for each hammer blow, combined with the actual cycle T i Based on the theoretical period T, construct the period correction factor: α i =T i / T; Periodic correction factor α i Reflecting the degree of expansion and contraction of the hammer impact waveform relative to the theoretical period, when α i >1 indicates that the actual period is greater than the theoretical period, and the waveform is stretched; when α i <1 indicates that the actual period is less than the theoretical period, and the waveform is compressed.

[0033] In specific implementation step S300, a frequency domain phase spectrum correction operation is performed on each waveform. First, based on the correction factor α... i Scaling the frequency axis: f' = f / ɑ i f is the frequency value before scaling, and f' is the frequency value after scaling. This operation achieves scaling correction of the time-domain waveform in the frequency domain: when α i When α > 1, the frequency axis is compressed, and the time-domain waveform after the inverse transform is compressed and shortened; when α i When the value is less than 1, the frequency axis is expanded, and the time-domain waveform is stretched and lengthened after the inverse transformation.

[0034] After scaling the frequency axis, the original amplitude spectrum A i (f) Since the sampling point position on the frequency axis changes, interpolation needs to be performed on the new frequency axis to obtain the scaled amplitude spectrum A'. i (f). Frequency domain interpolation uses linear interpolation or cubic spline interpolation methods, and the frequency resolution of the amplitude spectrum remains unchanged after interpolation.

[0035] Specifically, after scaling the frequency axis, the discrete sequence on the new frequency axis f' is denoted as f'. k Where k = 1, 2, 3, ..., K, K is the total number of frequency sampling points; for each discrete point f' on the new frequency axis k Find the two adjacent frequency points f on the original frequency axis f. m and f m+1 Make it satisfy f m ≤f' k ≤f m+1 The scaled amplitude spectrum A' is calculated using a linear interpolation formula. i (f' k ): A i (f) represents the original amplitude spectrum; m is the frequency index on the original frequency axis; by traversing all k=1…K, the complete scaled amplitude spectrum A' can be obtained. i (f); f' k This is the kth frequency value on the new frequency axis.

[0036] During amplitude spectrum adjustment, phase spectrum φ i (f) Keeping it unchanged, the scaled amplitude spectrum A' i (f) and the original phase spectrum φ i (f) Combining these steps, we obtain the corrected frequency domain signal F'. i (f): F' i (f)=A' i (f)e jφ i (f) .

[0037] For F' i (f) Perform an inverse Fourier transform to obtain the period-normalized stress wave data W' i (t), at this time the stress wave data W' i The period of (t) has been corrected to near the theoretical period T, the residual period deviation has been significantly reduced, and the waveforms tend to be aligned on the overall scale.

[0038] In the specific implementation step S400, the matching degree of the normalized waveform for each cycle is first calculated, which is the stress wave data W'. i The matching degree of (t) reflects the degree of residual period deviation of the waveform after period normalization.

[0039] In the specific solution, the matching degree is calculated as follows: ;ρ i T represents the matching degree of the i-th hammer strike; i Let be the actual period of the i-th hammer strike.

[0040] When the actual period T i When the period is equal to the theoretical period T, the matching degree ρ i =1; when the actual period T i When the deviation from the theoretical period T, the matching degree ρ i The corresponding decrease. Matching degree ρ i The value range is [0,1], and a larger value indicates that the actual period is closer to the theoretical period. If Then let ρ i =0.

[0041] Simultaneously, the signal-to-noise ratio (SNR) for each waveform is calculated. iIn specific calculations, the root mean square of noise is calculated within the time window when there is no reflected signal at the end of the waveform, and the ratio of the peak value of the waveform to the root mean square of noise is used as the signal-to-noise ratio.

[0042] Based on the matching degree and signal-to-noise ratio, determine the weighting coefficients for each waveform: ;

[0043] The denominator of the weighting coefficient represents the summation over all waveforms, and j is the summation index variable. The weighting coefficient has the following properties: the sum of the weighting coefficients of all waveforms is 1; waveforms with high matching degree and high signal-to-noise ratio receive higher weights; waveforms with low matching degree or low signal-to-noise ratio automatically have lower weights.

[0044] Finally, the waveforms after all periods are normalized are weighted and superimposed according to their respective weight coefficients to obtain the enhanced waveform W. N (t): N represents the total number of hammer blows; this enhanced waveform has a high signal-to-noise ratio and clear pile bottom reflection characteristics, and can be directly used for subsequent pile integrity determination and bearing capacity calculation.

[0045] This invention abandons the traditional time-domain resampling method and instead achieves period normalization in the frequency domain. The core advantage of frequency domain processing is that the amplitude spectrum usually presents a smooth structure with a single or double peak in the frequency domain, and the frequency domain interpolation error is much smaller than that of interpolation for transient signals in the time domain. More importantly, this invention keeps the phase spectrum unchanged during frequency axis scaling, which allows the relative timing relationship of each frequency component in the waveform to be accurately preserved. The steepness of the rising edge, the accurate position of the peak point, and the relative time delay between each reflected wave are not affected by the period normalization operation.

[0046] Specifically, the frequency domain phase spectrum correction path avoids the destruction of waveform transient characteristics caused by time domain interpolation. During interpolation in the frequency domain, the smooth structure of the amplitude spectrum keeps interpolation errors at extremely low levels; the preservation of the phase spectrum ensures the complete retention of the waveform's internal timing relationships. This means that the waveform, which originally expanded due to the deformation of the hammer pad, recovers to a scale consistent with the theoretical period after frequency domain correction, while its key characteristics such as rising edge, peak point, and reflected wave delay remain undamaged. For defect location, the preservation of the phase spectrum directly guarantees the measurement accuracy of the reflected wave delay, keeping the defect location error within 0.1 meters.

[0047] Furthermore, in this embodiment of the invention, the product of matching degree and signal-to-noise ratio is used as the weighting basis. The matching degree reflects the degree to which the waveform period deviates from the theoretical period. The larger the period deviation, the more severe the deformation of the hammer pad and the greater the waveform distortion, and its weight is automatically reduced. The signal-to-noise ratio reflects the degree to which the waveform is affected by noise. The lower the signal-to-noise ratio, the lower the weight of the waveform is also automatically reduced. This method automatically suppresses the influence of abnormal waveforms on the superposition result, while high-quality waveforms dominate.

[0048] Specifically, when an abnormal waveform is generated due to localized breakage of the hammer pad, hammer eccentricity, or environmental interference during a hammer impact, the actual period of this waveform will inevitably deviate from the theoretical period (reduced matching degree) or the signal-to-noise ratio will decrease, and its weighting coefficient will automatically decrease. In extreme cases, the weight of the abnormal waveform can drop below 0.05, and its influence on the superposition result is suppressed by more than 90%. This makes the enhanced waveform after superposition mainly contributed by the high-quality waveform, and the signal-to-noise ratio is improved by about 4 dB compared to equal-weight superposition.

[0049] The embodiment of this invention makes the pile bottom reflection signal clearer and more distinguishable in the enhanced waveform. The effective signals that were originally canceled out due to period drift are coherently enhanced after period normalization and weighted superposition, effectively improving the pile bottom reflection recognition rate. This provides high-quality input data for subsequent bearing capacity calculation and integrity determination, improving the overall reliability of high-strain pile detection without changing any hardware conditions.

[0050] In the above embodiment, the period-normalized waveform W' obtained after frequency domain phase spectrum correction (inverse Fourier transform while keeping the phase spectrum unchanged after frequency axis scaling) i (t), whose period has been corrected to near the theoretical period T, and the residual period deviation has been significantly reduced. However, residual time shifts still exist at the pile bottom reflection location for each waveform due to the following two reasons:

[0051] The first reason is the limited accuracy of actual period extraction. The frequency resolution of the Fourier transform is limited by the sampling frequency and the number of sampling points; the dominant frequency f... i The extraction accuracy has an inherent error, which is completely preserved throughout the entire propagation of the stress wave to the pile bottom and back. For a typical engineering pile with a pile length of 20 meters and a wave speed of 4000 meters per second, the arrival time of the reflection at the pile bottom is about 10 milliseconds. The small error in the periodic extraction, after propagating over a long distance, may cause a time shift of 0.1 to 0.2 milliseconds in the actual arrival time of the reflection at the pile bottom, corresponding to a positioning error of 0.2 to 0.4 meters.

[0052] The second reason is that the period normalization process only handles the overall scaling of the waveform, without performing local fine-tuning alignment for the specific peak of the pile bottom reflection. The pile bottom reflection is the characteristic wave with the longest propagation distance of the stress wave and is also the core basis for bearing capacity calculation. Its sensitivity to time shift errors is much higher than that of other parts of the waveform. The alignment accuracy of each waveform at the pile bottom reflection position directly determines the amplitude enhancement effect of the superimposed pile bottom reflection and the accuracy of the bearing capacity calculation.

[0053] To address the issue of residual time shifts in waveforms at the pile bottom reflection location after period normalization, which affect the superposition effect and the accuracy of bearing capacity calculation, this invention provides a preferred embodiment of a pile bottom reflection time shift compensation method based on a phase consistency index. This method involves truncating the pile bottom reflection time window, calculating the phase difference sequence, performing a first-order linear fit, using the fitting residual as a consistency index, and using the fitting slope as the compensation time shift. A linear correction term is applied to the phase spectrum in the frequency domain to achieve pure time shift compensation. The following is a combination of... Figure 2 The specific implementation plan will be described in detail, based on the aforementioned obtained period-normalized stress wave data W' i Following (t), the following steps are further included.

[0054] The first step is to select a reference waveform: The normalized waveform W'1(t) of the first hammer blow is used as the reference waveform, or the waveform with the highest signal-to-noise ratio among all periodic normalized waveforms is selected as the reference waveform; the reference waveform is used to measure the alignment of other waveforms at the pile bottom reflection position.

[0055] The second step is to determine the time window for pile bottom reflection: Centered on the aforementioned theoretical period T, a time window is set: [T-Δtw, T+Δtw]; Δtw is the half-width of the window, and its value is 1 / 2T~T; the pile bottom reflection time window is used to focus on the time domain range where the pile bottom reflection is located, and to eliminate interference from other waveform parts.

[0056] By employing a time window centered on the arrival time of the pile bottom reflection theory, the phase consistency check is focused on the time domain interval where the pile bottom reflection occurs, eliminating interference from the incident wave and other reflected waves. This ensures that the detected phase difference directly reflects the alignment of the pile bottom reflection position, rather than differences in other parts of the waveform.

[0057] The third step is to calculate the phase difference sequence: Normalize the waveform W' for each cycle i (t) and reference waveform W' ref (t), waveform segments within the pile bottom reflection time window are extracted respectively, and Fourier transforms are performed on the extracted waveform segments to obtain their respective window phase spectra φ*. i (f) and φ* ref(f); φ* i (f) is W' i (t) corresponds to the window phase spectrum; φ* ref (f) is W' ref The window phase spectrum corresponding to (t).

[0058] Within the frequency range corresponding to the pile bottom reflection, with the theoretical dominant frequency as the center and a bandwidth of ±20%, calculate the phase difference sequence Δφ*. i (f): Δφ* i (f)=φ* ref (f)-φ* ref (f). This phase difference sequence reflects the phase difference between the current waveform and the reference waveform within the pile bottom reflection frequency range.

[0059] The fourth step is to perform a first-order linear fit on the phase difference sequence: For the phase difference sequence Δφ* i (f) Perform a first-order linear fit, and the fitted model is: Δφ* i (f)=2π·Δt i ·f+φ0; Δt i φ is the fitting slope; φ0 is the fitting intercept; the least squares method is used for fitting to minimize the sum of squared fitting residuals, and the sum of squared fitting residuals is calculated simultaneously. The fitting intercept of this model is used to absorb constant phase deviations in the actual system that are independent of time shift, such as differences in sensor installation positions and filter phase delays.

[0060] The negative of the fitted slope is used as the compensation time shift. When the compensation time shift is positive, it means that the current waveform is delayed relative to the reference waveform and needs to be shifted backward. When the compensation time shift is negative, it means that the current waveform is ahead of the reference waveform and needs to be shifted forward.

[0061] The reciprocal of the sum of squared residuals is used as a phase consistency index. The smaller the sum of squared residuals, the closer the phase difference sequence is to a linear relationship, meaning the difference between the current waveform and the reference waveform is mainly caused by time shift and constant phase deviation, resulting in high waveform shape consistency. Conversely, the larger the sum of squared residuals, the more nonlinear distortion exists in the phase difference sequence, leading to differences in waveform shape. Therefore, a larger sum of squared residuals indicates higher phase consistency.

[0062] Step 5: Determine and compensate for residual time shift: When the phase consistency index fails to meet the requirements, the waveform is determined to have significant residual time shift or waveform distortion, requiring time shift compensation for the waveform W'. i (t) Perform frequency domain phase correction: maintain its amplitude spectrum A' i (f) Remain unchanged, update the phase spectrum as follows: φ i new (f')=φ i (f')-2π·Δt i ·f'.

[0063] Δt i φ represents the compensation time shift for the i-th hammer strike; i new (f) shows the updated phase spectrum; the corrected amplitude spectrum and the updated phase spectrum are combined and then subjected to an inverse Fourier transform to obtain the time-shift compensated waveform W''. i (t). When the phase consistency index meets the requirements, let W'' i (t)=W' i (t).

[0064] Finally, the reference waveform W' ref (t) and all time-shift compensated waveforms W'' i (t) (i ranges from 2 to N) together form the aligned waveform sequence, which is used for subsequent weighted superposition.

[0065] It should be noted that the phase consistency index of all waveforms can be determined based on the statistical distribution of the index. Waveforms that are lower than the mean minus two standard deviations are judged to have unsatisfactory phase consistency.

[0066] Compared to calculating the time shift through cross-correlation in the time domain, the solution in this embodiment of the invention determines the compensation time shift based on the fitting slope, avoiding the destruction of waveform shape by time domain differences, and achieving a clean time shift without introducing additional distortion.

[0067] The present invention decomposes the phase difference sequence into linear components (corresponding to time shift) and nonlinear components (corresponding to waveform distortion) through first-order linear fitting, and characterizes them respectively using the fitting slope and fitting residual. This decomposition allows the present invention to distinguish between time shift differences that can be compensated by translation and waveform distortions that cannot be compensated by translation. The former is eliminated through frequency domain phase correction, while the latter is reflected by the phase consistency index and its weight is automatically reduced in subsequent weighted superposition. This method of separating compensable and non-compensable components is more refined and effective than the method of directly calculating the time shift and then uniformly compensating.

[0068] Furthermore, frequency domain phase correction achieves time shift compensation by applying a linear correction term to the phase spectrum, avoiding the need for waveform resampling and interpolation required by time domain translation. Time domain translation requires shifting the waveform along the discrete time axis by a non-integer multiple of the sampling interval, inevitably introducing interpolation errors and destroying the transient characteristics of the waveform. In contrast, frequency domain linear phase correction utilizes the time-shifting property of the Fourier transform to achieve time shifts of arbitrary precision while maintaining the waveform shape, without affecting the steepness of the rising edge or the accurate location of the peak point.

[0069] Example 2: This embodiment of the invention provides a high-strain foundation pile testing method, applying the stress wave waveform period normalization method from Example 1 to the testing process, forming an end-to-end solution from sensor installation, data acquisition and waveform processing to defect detection. (Refer to...) Figure 3 As shown, it includes the following steps: Step T100: Place a vertical pad on the top of the pile, and symmetrically install force sensors and acceleration sensors on the pile body at a distance of twice the pile diameter from the top of the pile. Step T200: The pile top is repeatedly hammered with a heavy hammer, and the stress wave data generated by each hammering is collected by the force sensor and acceleration sensor to obtain the original waveform sequence. Step T300: Process the original waveform sequence based on the stress wave waveform period normalization method to obtain the enhanced waveform; Step T400: Perform defect detection on the pile foundation based on the enhanced waveform.

[0070] In the specific implementation step T100, the pile head of the pile to be tested is first treated by removing the laitance and loose concrete at the top of the pile to ensure that the pile top is flat. A hammer pad made of natural wood is placed on the top of the pile to buffer the impact of the heavy hammer and adjust the impact pulse pattern.

[0071] Force sensors and accelerometers are symmetrically installed at a distance of twice the pile diameter from the pile top. Specifically, the two force sensors are installed at corresponding positions on both sides of the pile, and the two accelerometers are similarly symmetrically installed. The purpose of symmetrical installation is to eliminate the influence of eccentric hammering on the measurement results; when the hammering is eccentric, the signal difference between the sensors on both sides can be used as a basis for judgment. Before installation, the mounting surface should be leveled and cleaned, and coupling agent should be used to ensure a tight fit between the sensor and the pile.

[0072] In the specific implementation step T200, a crane is used to lift the hammer to the predetermined height, and the hammer is released through the unhooking device, allowing the hammer to fall freely to impact the top of the pile; the mass of the hammer is selected according to the estimated bearing capacity, usually between 2 tons and 12 tons; the hammer drop distance is adjusted in stages according to the testing needs.

[0073] During each hammer blow, a force sensor acquires the force time history curve at the top of the pile, and an acceleration sensor acquires the acceleration time history curve. The acceleration signals are integrated to obtain the velocity time history curve. A set of stress wave data is collected for each hammer blow, arranged in the order of the blows, and denoted as the original waveform sequence W. i (t), i=1,2,...,N, where N is the total number of hammer blows and t is the time.

[0074] In specific implementation step T300, the original waveform sequence obtained in step T200 is processed based on the stress wave waveform period normalization method in Example 1. The core processing flow of this method includes: performing a Fourier transform on the stress wave data of each hammer blow to obtain the amplitude spectrum and phase spectrum; calculating the actual period based on the frequency corresponding to the main peak in the amplitude spectrum; calculating the theoretical period based on the pile length and stress wave velocity; constructing a period correction factor by combining the actual period and the theoretical period; scaling the frequency axis according to the correction factor; obtaining the scaled amplitude spectrum through frequency domain interpolation; performing an inverse Fourier transform based on the phase spectrum to obtain the period-normalized stress wave data; calculating the matching degree and waveform signal-to-noise ratio of each waveform with the actual period; determining the weighting coefficients based on the matching degree and signal-to-noise ratio; and weighting and superimposing all period-normalized waveforms according to the weighting coefficients to obtain the enhanced waveform.

[0075] The specific implementation method of this step has been described in detail in Embodiment 1, and will not be repeated here.

[0076] In specific implementation step T400, time-frequency analysis is performed on the enhanced waveform obtained in step T300 to identify the pile bottom reflection signal and the pile body defect reflection signal.

[0077] Specifically, it includes the following sub-steps: Sub-step T410: Identify the reflection signals from the pile bottom and the pile body defects. Based on the theoretical arrival time, the pile bottom reflection signal is searched at the tail of the waveform. This signal exhibits a distinct positive or negative peak, and its arrival time is close to the theoretical value. The pile defect reflection signal is then searched within the time range preceding the theoretical arrival time. The arrival time of the defect reflection signal is related to the defect depth d: defect depth d = (C·t) de ) / 2; t de This represents the arrival time of the defect reflection signal.

[0078] Sub-step T420: Calculate the vertical compressive bearing capacity of a single pile. Based on the arrival time and amplitude of the pile bottom reflection signal, the vertical compressive bearing capacity of a single pile is calculated using the Case method or the measured curve fitting method. The clarity of the pile bottom reflection signal directly affects the accuracy of the bearing capacity calculation. In the embodiment of this invention, the enhanced waveform used undergoes periodic normalization and weighted superposition processing, making the pile bottom reflection signal clearly distinguishable and significantly improving the repeatability of the bearing capacity calculation.

[0079] Furthermore, in the embodiment of the present invention, after obtaining the original waveform sequence, the method further includes: calculating the signal-to-noise ratio of the original waveform and calculating the median of the signal-to-noise ratio of all waveforms; if the signal-to-noise ratio of any waveform is less than one-third of the median, then the weighting coefficient of the corresponding waveform is assigned to zero.

[0080] Specifically, for each waveform in the original waveform sequence obtained in step T200, its signal-to-noise ratio is calculated. First, a noise window is determined. In the stress wave waveform, the effective signal is mainly concentrated in the interval from the arrival of the incident wave to the arrival of the pile bottom reflection. In the time period after the arrival of the pile bottom reflection, theoretically there is no effective signal, and the waveform in this time period mainly consists of environmental noise and system noise. Therefore, taking the theoretical arrival time of the pile bottom reflection as the starting point, a time interval with a length of one theoretical period is truncated as the noise window.

[0081] Then, within the noise window, the root mean square (RMS) value of the waveform amplitude is calculated.

[0082] Next, within the entire time range of the waveform, the maximum value of the waveform amplitude is searched as the signal peak value P.

[0083] Finally, calculate the signal-to-noise ratio (SNR): SNR = 20·log 10 (P / RMS); The higher the signal-to-noise ratio, the stronger the signal is relative to the noise, and the higher the waveform quality.

[0084] The signal-to-noise ratios (SNRs) of all calculated waveforms are sorted in ascending order. When N is odd, the median is the SNR of the (N+1) / 2th waveform after sorting; when N is even, the median is the average of the SNRs of the (N / 2)th and (N / 2)+1th waveforms after sorting. The median has better robustness than the average. When there are a few extreme abnormal waveforms (extremely low SNR), the average will be significantly shifted by these extreme values, while the median is not affected by extreme values ​​and can more stably reflect the typical SNR of most waveforms.

[0085] For each waveform, its signal-to-noise ratio (SNR) is compared with the median. If the SNR of any waveform is lower than one-third of the median, the waveform is determined to be an interfered waveform. For waveforms that meet the above conditions, their weighting coefficients are directly assigned to zero. Waveforms with weighting coefficients of zero will not participate in subsequent period normalization and weighted superposition, directly eliminating the influence of pollution sources on the enhanced waveform and further increasing the SNR of the enhanced waveform in the presence of environmental interference.

[0086] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for normalizing the period of a stress wave waveform, characterized in that, include: Acquire stress wave data W generated by repeated hammer blows on the pile foundation i (t); Perform a Fourier transform on the stress wave data to obtain the amplitude spectrum A. i (f) and phase spectrum φ i (f); Calculate the actual period T based on the frequency corresponding to the main peak value in the amplitude spectrum. i ; i is the hammer strike number; The theoretical period T is calculated based on the pile length and stress wave velocity, combined with the actual period T. i Constructing a correction factor α with the theoretical period T i =T i / T;ɑ i is the correction factor for the stress wave of the i-th hammer impact; The frequency axis is scaled according to the correction factor, and the scaled amplitude spectrum A' is obtained by frequency domain interpolation. i (f); combined with the amplitude spectrum A' i (f) and phase spectrum φ i (f) Perform an inverse Fourier transform to obtain the period-normalized stress wave data W' i (t); For the stress wave data after period normalization, the matching degree between it and the actual period is calculated, and the waveform signal-to-noise ratio is calculated. The weighting coefficient is determined according to the matching degree and the waveform signal-to-noise ratio. All the waveforms after period normalization are weighted and superimposed according to their respective weighting coefficients to obtain the enhanced waveform for high strain pile detection.

2. The stress wave waveform period normalization method according to claim 1, characterized in that, Calculating the matching degree includes: ； ρ i T represents the matching degree of the i-th hammer strike; i Let be the actual period of the i-th hammer strike.

3. The stress wave waveform period normalization method according to claim 2, characterized in that, Determine the weighting coefficient ω i include: ； ω i SNR is the stress wave weighting coefficient for the i-th hammer blow; i denoted as , where is the stress wave signal-to-noise ratio of the i-th hammer strike; N is the number of hammer strikes; and j is the summation index variable.

4. The stress wave waveform period normalization method according to claim 1, characterized in that, The frequency axis is scaled according to the correction factor, including: f' = f / ɑ i f is the frequency value before scaling; f' is the frequency value after scaling.

5. The stress wave waveform period normalization method according to claim 4, characterized in that, After obtaining the period-normalized stress wave data W' i Following (t) is: Calculate the phase consistency index between each waveform and the reference waveform within the pile bottom reflection time window; If the phase consistency index of any waveform does not meet the requirements, the phase spectrum of the waveform is adjusted: within the frequency range corresponding to the pile bottom reflection, a linear correction term φ=2π·Δt·f' is applied to the phase spectrum to align the pile bottom reflection of the waveform with the reference waveform; Δt is the compensation time shift.

6. The stress wave waveform period normalization method according to claim 5, characterized in that, Calculate the phase consistency index between each waveform and the reference waveform within the pile bottom reflection time window, including: For stress wave data W' i (t) and the reference waveform, respectively extract waveform segments within the pile bottom reflection time window, and perform Fourier transform on the waveform segments to obtain their respective window phase spectra; Based on stress wave data W' i The phase difference between the window phase spectra of the reference waveform and the reference waveform is calculated to obtain the phase difference sequence. A first-order linear fit is performed on the phase difference sequence to obtain the fitting slope and fitting residual; The inverse of the fitted slope is used as the compensation time shift; the reciprocal of the sum of squares of the fitted residuals is used as the phase consistency index.

7. A method for testing high-strain foundation piles, characterized in that, include: A vertical pad is placed on the top of the pile, and force sensors and acceleration sensors are symmetrically installed on the pile body at a distance of twice the pile diameter from the top of the pile. The pile top is repeatedly hammered with a heavy hammer, and the stress wave data generated by each hammering is collected by the force sensor and acceleration sensor to obtain the original waveform sequence. The original waveform sequence is processed using the stress wave waveform period normalization method according to any one of claims 1-6 to obtain an enhanced waveform; Defect detection of the pile foundation is performed based on the enhanced waveform.

8. The high-strain pile testing method according to claim 7, characterized in that, Defect detection of the pile foundation based on the enhanced waveform includes: Time-frequency analysis was performed on the enhanced waveform to identify the pile bottom reflection signal and the pile body defect reflection signal; The vertical compressive bearing capacity of a single pile is calculated based on the arrival time and amplitude of the reflected signal at the pile bottom. The location of the defect is calculated based on the arrival time of the reflected signal from the pile defect.

9. The high-strain pile testing method according to claim 7, characterized in that, The drop pad is configured as a wooden board.

10. The high-strain pile testing method according to claim 7, characterized in that, After obtaining the original waveform sequence, the method further includes: calculating the signal-to-noise ratio (SNR) of the original waveform and calculating the median of the SNR of all waveforms. If the SNR of any waveform is less than one-third of the median, the weighting coefficient of the corresponding waveform is assigned to zero.

Images