Tower structure member inspection and early warning method and medium

Abstract

The application discloses a tower structure member inspection early warning method and medium, and relates to the technical field of inspection early warning, and comprises the following steps: acquiring the acoustic wave reflection signal of the bonding layer of a tower structure member in the curing process, combining a preset strain gauge space coordinate to perform feature mapping, and obtaining a bonding layer delay matrix; analyzing the thickness change of the bonding layer based on the bonding layer delay matrix, and obtaining a bonding layer delay curve; extracting a strain signal set of the preset strain gauge, performing delay feature extraction on the strain signal set based on the bonding layer delay curve, and obtaining a delay signal sequence; performing time sequence correction on the strain signal set based on the delay signal sequence, and obtaining a synchronous strain sequence set; performing strain distribution feature calculation on the tower structure member based on the synchronous strain sequence set, and performing safety early warning based on the calculation result, so that the problems that the phase difference of strain conduction is caused by uneven thickness of the glue layer and the edges and the center of the strain gauge are out of synchronization are solved.

Description

Technical Field

[0001] This invention relates to the field of inspection and early warning technology, and more specifically, to a method and medium for inspection and early warning of tower crane structural components. Background Technology

[0002] Tower cranes, as crucial lifting equipment on construction sites, undertake critical operations such as lifting, moving, and precisely positioning building materials. Their structural safety directly impacts the lives of construction workers and the progress of the project. Among the many structural components of a tower crane, especially the boom, connection nodes, and the junction between the boom and the foundation support, the components are typically fixed and stress is transferred through adhesives, epoxy composite materials, or other bonding layers.

[0003] However, especially when the adhesive layer thickness is uneven, strain transmission within the adhesive layer exhibits significant non-uniformity. Due to variations in adhesive layer thickness leading to differences in local stiffness, density, and acoustic impedance, stresses generated by external loads or construction processes cannot be uniformly distributed throughout the adhesive layer, resulting in different strain transmission rates and amplitudes at different locations. This non-uniform transmission causes significant phase differences in strain gauges placed on the adhesive layer surface or at nodes; that is, under the same load, the response at the edge of the strain gauge does not reach its peak value or a specific waveform point simultaneously with the response at the center. This phenomenon manifests as signal hysteresis in actual measurements. This signal hysteresis not only affects the accuracy of strain gauge measurement data but may also lead to inaccuracies in strain distribution calculations and structural health assessments, making it difficult to accurately reflect the local stress state and thickness variations of the adhesive layer. Especially at long booms or nodes of tower cranes, even minor variations in adhesive layer thickness can amplify stress concentration effects, forming areas of localized fatigue or early debonding.

[0004] To address the above problems, this invention proposes a solution. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method and medium for early warning of tower crane structural components during inspection. By combining the analysis of the thickness variation of the adhesive layer of the tower crane structural components with the strain distribution, the method aims to solve the problem of phase difference in strain transmission caused by uneven adhesive layer thickness and asynchronous response between the edge and center of the strain gauge.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for early warning inspection of tower crane structural components includes the following steps: acquiring acoustic wave reflection signals of the adhesive layer of the tower crane structural component during the curing process, and performing feature mapping based on the spatial coordinates of preset strain gauges to obtain an adhesive layer delay matrix; analyzing the thickness variation of the adhesive layer based on the adhesive layer delay matrix to obtain an adhesive layer delay curve; extracting the strain signal set of preset strain gauges, and extracting delay features from the strain signal set based on the adhesive layer delay curve to obtain a delay signal sequence; performing time-series correction on the strain signal set based on the delay signal sequence to obtain a synchronous strain sequence set; calculating the strain distribution characteristics of the tower crane structural component based on the synchronous strain sequence set, and issuing a safety warning based on the calculation results.

[0008] In a preferred embodiment, the step of acquiring the acoustic wave reflection signal of the adhesive layer of the tower crane structural component during the curing process and performing feature mapping in conjunction with the spatial coordinates of a preset strain gauge to obtain the adhesive layer delay matrix specifically involves: transmitting a broadband acoustic wave signal to the adhesive layer of the tower crane structural component and receiving the echo signal reflected from the adhesive layer interface using a preset acoustic wave sensor to obtain the acoustic wave reflection signal; performing time-domain analysis on the acoustic wave reflection signal to obtain the arrival time difference and amplitude attenuation value of the acoustic wave reflection signal; calculating the acoustic impedance change value of the adhesive layer based on the arrival time difference and amplitude attenuation value of the acoustic wave reflection signal; aligning the spatial coordinates of the preset strain gauge with the spatial coordinates of the preset acoustic wave sensor, and mapping the acoustic impedance change value to the strain gauge position using a spatial interpolation method to obtain the delay characteristic distribution value of the strain gauge; and mapping the delay characteristic distribution value of the strain gauge to a grid using a spatial mapping method to obtain the adhesive layer delay matrix.

[0009] In a preferred embodiment, the step of analyzing the thickness variation of the adhesive layer based on the adhesive layer delay matrix to obtain the adhesive layer delay curve specifically involves: performing singular value decomposition on the adhesive layer delay matrix to extract principal component eigenvectors; calculating the propagation velocity of the sound wave in the adhesive layer based on the principal component eigenvectors, and obtaining the adhesive layer thickness variation sequence based on the propagation velocity; extracting the strain gauge delay time sequence from the adhesive layer delay matrix, and constructing a correlation function based on the adhesive layer thickness variation sequence and the strain gauge delay time sequence; and reconstructing the adhesive layer delay moment based on the correlation function to obtain the adhesive layer delay curve.

[0010] In a preferred embodiment, the step of extracting the strain signal set of preset strain gauges and extracting delay features from the strain signal set based on the delay curve of the adhesive layer to obtain a delay signal sequence specifically involves: acquiring strain signals from each preset strain gauge to construct a strain signal set; performing data analysis on the delay curve of the adhesive layer and selecting delay reference points based on the analysis results; segmenting the strain signal set to obtain several strain signal segments and adjusting the start time of each signal segment according to the delay reference points to obtain adjusted strain signals; and extracting delay features from the adjusted strain signals to obtain a delay signal sequence, wherein the delay signal sequence includes a delay time difference and a signal phase offset.

[0011] In a preferred embodiment, the step of performing data analysis on the delay curve of the adhesive layer and selecting delay reference points based on the analysis results specifically involves: performing feature analysis on the delay curve of the adhesive layer to obtain the curve slope change points; dividing the curve into several delay segments using the curve slope change points, and identifying the delay peak point in each delay segment as an alignment reference point; setting a search window near the alignment reference point, and retrieving the feature response point corresponding to the time position of the alignment reference point from the strain signal corresponding to the search window; using the time of the feature response point as the start time of the delay segment, and performing time-shift correction on the strain signal of the delay segment based on the start time to obtain the delay reference point.

[0012] In a preferred embodiment, the step of performing time-series correction on the strain signal set based on the delayed signal sequence to obtain a synchronous strain sequence set specifically involves: performing time-series matching between the delayed signal sequence and the strain signal set, calculating the time-series offset of each strain signal; constructing a time-series correction function based on the time-series offset, and resampling and interpolating the strain signal set based on the time-series correction function to obtain the synchronous strain sequence set.

[0013] In a preferred embodiment, the step of calculating the strain distribution characteristics of the tower crane structural components based on the synchronous strain sequence set and issuing a safety warning based on the calculation results specifically involves: performing spatial differentiation calculations from the synchronous strain sequence set to obtain the strain gradient distribution value at each time point; identifying abnormal sequences in the synchronous strain sequence set based on the strain gradient distribution value and predicting the damage level of the structural components based on the abnormal sequences; and triggering a safety warning signal and outputting the warning location and severity of the tower crane structural components when the damage level exceeds a preset threshold.

[0014] The technical effects and advantages of the tower crane structural component inspection and early warning method and medium of the present invention are as follows:

[0015] This invention acquires acoustic wave reflection signals from the adhesive layer of tower crane structural components during curing or operation, and performs feature mapping using preset strain gauge spatial coordinates. This enables precise identification of adhesive layer thickness variations and strain propagation delays. Based on the adhesive layer delay matrix and delay curve, delay feature extraction and timing correction of the strain signals are performed to obtain multi-point synchronous strain sequences, thereby accurately reflecting the strain distribution characteristics of tower crane structural components at various spatial locations. Furthermore, by identifying abnormal sequences in the strain gradient distribution and predicting the degree of damage, a quantitative assessment of potential damage to the structural components is achieved. When the degree of damage exceeds a preset threshold, a safety warning is automatically triggered, outputting the warning location and severity. This method has several significant advantages: First, it effectively overcomes the limitations of traditional inspection methods that rely on manual visual inspection or single-point detection, providing high-precision, real-time, and continuous structural health monitoring. Second, through delayed signal extraction and timing correction, it eliminates the strain signal hysteresis problem caused by uneven adhesive layer thickness or material aging, enabling the synchronization of multi-point strain measurement data and ensuring the accuracy of strain distribution calculation. Third, it enables early detection and quantitative assessment of local damage to tower crane structural components, providing a scientific and quantifiable basis for construction safety decisions. Finally, this method is applicable to comprehensive inspection of all key structural components of tower cranes, is easy to operate, highly automated, and significantly improves the operational safety of tower cranes and the efficiency of structural maintenance. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a tower crane structural component inspection and early warning method according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1, Figure 1 This invention provides a method for early warning and inspection of tower crane structural components, comprising the following steps:

[0019] S1, obtain the acoustic wave reflection signal of the adhesive layer of the tower crane structural component during the curing process, and perform feature mapping in combination with the preset strain gauge spatial coordinates to obtain the adhesive layer delay matrix;

[0020] In this example, the acoustic wave reflection signal of the adhesive layer of the tower crane structural component during the curing process is obtained, and feature mapping is performed in combination with the preset strain gauge spatial coordinates to obtain the adhesive layer delay matrix, specifically:

[0021] Broadband acoustic signals are emitted to the adhesive layer of the tower crane structural components, and the echo signals reflected from the adhesive layer interface are received by a preset acoustic sensor to obtain the acoustic reflection signals.

[0022] Time-domain analysis of the reflected sound wave signal yields the arrival time difference and amplitude attenuation value.

[0023] The acoustic impedance change of the adhesive layer is calculated based on the arrival time difference and amplitude attenuation value of the reflected acoustic signal.

[0024] Align the spatial coordinates of the preset strain gauge with the spatial coordinates of the preset acoustic sensor, and map the acoustic impedance change value to the strain gauge position using a spatial interpolation method to obtain the delay characteristic distribution value of the strain gauge.

[0025] The spatial mapping method is used to map the delay characteristic distribution values ​​of the strain gauges onto the mesh to obtain the delay matrix of the adhesive layer.

[0026] In this example, the formula for calculating the change in acoustic impedance of the adhesive layer is as follows:

[0027]

[0028] in, This represents the change in acoustic impedance of the adhesive layer. The reference acoustic impedance of the adhesive layer, The amplitude of the reflected sound wave. The amplitude of the incident sound wave. The time difference of arrival of the reflected sound wave signal. This is the ideal sound wave propagation time.

[0029] It should be noted that a broadband piezoelectric transducer with a center frequency of 2MHz and a bandwidth of 1.5MHz can be used as the acoustic wave transmitting device. This transducer generates a broadband pulse acoustic wave signal through a dedicated pulse exciter. The pulse width is approximately 2μs, the peak voltage is 100V, and the repetition frequency is 200Hz. The acoustic wave is uniformly transmitted to the adhesive layer at the tower crane boom joint through a coupling agent (such as silicone grease or epoxy resin). The adhesive layer is typically 2.5mm thick and is formed by bonding an epoxy resin-based adhesive layer to a steel substrate. When the broadband acoustic wave propagates to the interface between the adhesive layer and the metal, some of the energy is reflected back. Eight preset acoustic wave sensors (Olympus V133-RM type) are arranged on the surface of the tower crane structural components, with a sensor spacing of approximately 50mm and a sampling frequency of 10MHz. Each sensor receives the echo signal within a 0-10μs time window after the acoustic wave emission. The system synchronously records the waveforms of the reflected signals received by each sensor through the data acquisition module. The peak voltage of the signal is generally between 15-60mV, forming a raw acoustic wave reflection signal dataset.

[0030] Furthermore, the acquired reflected signals were first bandpass filtered (0.5-3 MHz) to suppress environmental noise, and then time-domain envelope extraction was performed to identify the main echo peaks. The signal waveforms of each sensor were time-calibrated within a sampling window of 0-10 μs, with the first main peak selected as the reference wave and the second main peak as the reflection signal from the adhesive layer interface. The arrival time difference of the signals can be obtained by comparing the time positions of the two main peaks. For example, in actual testing, the arrival time difference for sensor 1 was 1.32 μs, for sensor 2 it was 1.47 μs, and for sensor 8 it was 1.65 μs. Simultaneously, the amplitude attenuation value was obtained by normalizing the peak voltages of the two main peaks, with typical results between 0.42 and 0.58. The time-domain analysis results indicate spatial differences in the local wave velocity and energy loss of the adhesive layer, providing fundamental data for subsequent calculations of acoustic impedance.

[0031] Furthermore, based on the arrival time difference and amplitude attenuation values ​​obtained from the above time-domain analysis, changes in the acoustic properties within the adhesive layer can be inferred through variations in sound wave propagation delay and reflected energy. Specifically, when microbubbles or insufficiently cured areas exist in the adhesive layer, the sound wave reflection intensity weakens, manifested as an increase in amplitude attenuation; when the density of the adhesive layer increases or its thickness decreases slightly, the reflection time difference decreases. In this embodiment, the acoustic impedance reference value at the sampling point is set to... The change value corresponding to sensor number 1 is Number 2 is Number 8 is By statistically analyzing the impedance change trends of all sensors, a slight density gradient distribution was observed in the adhesive layer from left to right during the curing process, indicating different local curing rates.

[0032] Secondly, in addition to the acoustic wave sensor, five strain gauges, numbered S1 to S5, are evenly distributed on the surface of the tower crane boom node, with a spacing of 60 mm between them. A three-dimensional coordinate system is established with the node center as the origin. The coordinates of the acoustic wave sensor are obtained through a 3D scanner with an accuracy of ±0.2 mm. First, the two sets of coordinate data are unified into a single coordinate system, and a rigid body registration algorithm based on least squares error is used to align the spatial distribution of the sensor and strain gauge. After alignment, the distance between acoustic wave sensor No. 1 and strain gauge S1 is 15 mm, and the distance between sensor No. 8 and strain gauge S5 is 12 mm. To obtain the acoustic impedance change value at the strain gauge location, the inverse distance weighted (IDW) spatial interpolation method is used to interpolate and map the impedance change values ​​of the eight sensors to the five strain gauge locations. Finally, the delay characteristic distribution values ​​of strain gauges S1-S5 are obtained as follows: +0.13, +0.10, +0.04, –0.05, –0.09 (units as before).

[0033] Finally, to obtain the overall spatial distribution of the retardation of the adhesive layer, the retardation characteristic distribution values ​​of the strain gauges were further mapped onto a two-dimensional mesh model. This mesh uses a 100mm × 100mm area as the analysis plane, divided into 10 × 10 cells, each cell representing a local region of the adhesive layer. First, the retardation characteristic values ​​of strain gauges S1-S5 were projected onto the mesh according to their coordinates. Then, bilinear interpolation was used to fill the remaining regions of the mesh with retardation values. In the actual calculation, the retardation value in the central region of the mesh (approximately corresponding to position S3) is approximately... The upper left corner area (close to S1) reached The bottom right corner (near S5) is The system ultimately generates a 10×10 dimension adhesive layer delay matrix, with each element corresponding to a delay feature value in each cell. Color mapping visually shows that the adhesive layer on the left side of the tower crane node is fully cured, while the right side shows a slight delay, providing intuitive basic data for subsequent thickness variation analysis.

[0034] S2, Based on the adhesive layer delay matrix, the thickness variation of the adhesive layer is analyzed to obtain the adhesive layer delay curve;

[0035] In this example, the thickness variation of the adhesive layer is analyzed based on the adhesive layer delay matrix to obtain the adhesive layer delay curve, as follows:

[0036] Singular value decomposition is performed on the delay matrix of the adhesive layer to extract the principal component eigenvectors;

[0037] The propagation speed of sound waves in the adhesive layer is calculated based on the principal component eigenvectors, and the thickness variation sequence of the adhesive layer is obtained based on the propagation speed.

[0038] The delay time series of strain gauges are extracted from the delay matrix of the adhesive layer, and a correlation function is constructed based on the thickness variation sequence of the adhesive layer and the delay time series of the strain gauges.

[0039] The delay moment of the adhesive layer is reconstructed based on the correlation function to obtain the delay curve of the adhesive layer.

[0040] It should be noted that the adhesive layer delay matrix is ​​a 10×10 two-dimensional matrix, where each element represents the acoustic delay characteristic value of that region, in units of... The values ​​in the left region of the matrix are concentrated between +0.10 and +0.14, the middle region between +0.04 and +0.08, and the right region between -0.05 and -0.09, reflecting the spatial delay differences. To analyze its main trends, the delay matrix was imported into the signal feature analysis module for singular value decomposition. The decomposition yielded three significant eigenvector groups, with the singular values ​​of the first eigenvector accounting for 78.4% of the total energy, the second for 13.2%, and the third for 4.9%, while the remaining eigenvalues ​​were negligible. The principal component eigenvectors reflect the main spatial pattern of the adhesive layer delay distribution. Taking the first eigenvector as an example, its components gradually transition from positive to negative in the upper left to lower right direction, indicating a significant thickness or density gradient in the adhesive layer. This vector was extracted as the input for subsequent principal component eigenvector calculations.

[0041] Secondly, to verify whether the spatial features extracted by singular value decomposition are consistent with the changes in acoustic signals, a second time-domain analysis was performed on the acoustic wave reflection signals corresponding to the same region. This analysis used raw signals measured by acoustic wave sensors, with each sensor recording waveform samples within a 0–10 μs time window. Taking sensors 1 and 8 as examples, after envelope extraction, the main echo peak of sensor 1 appeared at 1.28 μs with a peak amplitude of 55 mV; the main echo of sensor 8 appeared at 1.62 μs with an amplitude of 43 mV. The comparison results show that the arrival time difference within the region is 0.34 μs, and the amplitude attenuation ratio is approximately 0.78. By performing the same processing on all eight sensors, the arrival time difference range was found to be 0.30–0.39 μs, and the amplitude attenuation values ​​were distributed between 0.75 and 0.85. These data provide basic acoustic parameters for estimating the thickness variation of the adhesive layer and serve as a verification basis for the principal component vectors in the singular value decomposition matrix.

[0042] Furthermore, after obtaining the principal component eigenvectors, a correlation analysis was performed between these eigenvectors and the time difference distributions of each sensor to estimate the propagation speed of sound waves in different regions. Under experimental conditions, the tower crane adhesive layer was an epoxy-based composite material, and its cured sound velocity was approximately 2400 m / s. However, during the curing process, the local sound velocity fluctuated due to internal temperature gradients and air bubbles. Based on the spatial pattern of the principal component eigenvectors, the adhesive layer was divided into three regions: a high-density region (left side), a neutral region (middle), and a low-density region (right side). The calculation results showed that the average propagation speed in the high-density region was 2455 m / s, in the neutral region it was 2380 m / s, and in the low-density region it was 2260 m / s. Based on the differences in propagation speeds in each region and combined with the time difference of sound wave arrival, the thickness variation sequence of the adhesive layer was obtained: approximately 2.46 mm in the left region, 2.53 mm in the middle region, and 2.67 mm in the right region. It can be seen that the thickness of the adhesive layer increases slightly from left to right, indicating that the shrinkage of the adhesive layer on the right side was insufficient during the curing process. This thickness variation sequence was recorded as the dynamic thickness characteristics of the adhesive layer.

[0043] Secondly, to further establish the coupling relationship between the adhesive layer thickness variation and the strain gauge response, the delay time series corresponding to the coordinates of strain gauges S1-S5 were extracted from the aforementioned 10×10 delay matrix. Specifically, S1 was 1.29 μs, S2 was 1.34 μs, S3 was 1.41 μs, S4 was 1.53 μs, and S5 was 1.61 μs. These delay time points were arranged in chronological order to form a delay time series, which was then correlated with the thickness variation series obtained in the third step (2.46, 2.53, 2.60, 2.64, 2.67 mm). The correlation coefficient between the two was calculated using the data analysis module, yielding a correlation coefficient of approximately 0.92, indicating a high degree of coupling between the strain gauge delay time and the adhesive layer thickness variation. Based on this relationship, an automatic correlation function was established to characterize the delay response law caused by thickness variation. The physical meaning of this function is: when the local thickness of the adhesive layer increases, the sound wave propagation path lengthens, the delay time increases accordingly, and the phase signal corresponding to the strain gauge exhibits a hysteresis phenomenon.

[0044] Finally, after obtaining the correlation function between thickness variation and delay time, it was applied to the original adhesive layer delay matrix to correct and reconstruct the delay eigenvalues ​​of each grid cell. The correction process smoothed out abrupt changes in the original matrix caused by noise or local outliers through interpolation. The reconstructed matrix exhibited a continuous delay gradient distribution, smoothly transitioning from +0.12 in the upper left corner to -0.09 in the lower right corner. Subsequently, the diagonal data of the matrix was extracted along the principal strain direction and plotted as an adhesive layer delay curve. The curve shape shows that the delay value decreases slowly in the range of 0-50 mm, with a significant steep drop in the range of 50-100 mm, indicating that the right side of the adhesive layer cures more slowly. The peak-to-valley difference of the curve is approximately... The correlation between the acoustic wave propagation speed and thickness variation is well observed. This adhesive layer delay curve directly reflects the dynamic changes in acoustic wave propagation characteristics and adhesive layer uniformity during the curing process, providing crucial information for subsequent strain signal delay correction and structural safety assessment.

[0045] S3, extract the strain signal set of the preset strain gauge, and extract the delay feature of the strain signal set based on the delay curve of the adhesive layer to obtain the delay signal sequence;

[0046] In this example, the strain signal set of a preset strain gauge is extracted, and the delay feature of the strain signal set is extracted based on the delay curve of the adhesive layer to obtain the delay signal sequence, specifically:

[0047] Strain signals are collected from each preset strain gauge to construct a strain signal set;

[0048] Data analysis was performed on the delay curve of the adhesive layer, and delay reference points were selected based on the analysis results;

[0049] The strain signal set is segmented to obtain several strain signal segments. The start time of each signal segment is adjusted according to the delay reference point to obtain the adjusted strain signal.

[0050] Delay features are extracted from the adjusted strain signal to obtain a delayed signal sequence, which includes a delay time difference and a signal phase offset.

[0051] It should be noted that five strain gauges, numbered S1 to S5, were evenly arranged in the adhesive layer area of ​​the tower crane boom joint. These strain gauges were BF350-3AA type resistance strain gauges with a resistance of 350Ω and a sensitivity coefficient of 2.0. Each strain gauge was connected to a high-precision dynamic strain acquisition module via a full-bridge measurement circuit, with a sampling frequency of 5 kHz and a sampling accuracy of ±2με. During acquisition, the tower crane structural components were in the curing monitoring stage, the temperature was maintained at 25℃, and the external load was a periodic simulated wind load with a frequency of 0.5Hz and an amplitude of 0.2kN. The system continuously acquired strain signals for 10 minutes, obtaining approximately 300,000 data points for each strain gauge. The acquired data, after noise filtering (using a 3rd-order Butterworth low-pass filter) and baseline correction, formed the original strain signal curve. Actual data showed that the strain range for S1 was ±45με, S2 ±52με, S3 ±49με, S4 ±60με, and S5 ±57με. The time-series signals from the five strain gauges were combined into a five-dimensional signal set according to their sensor numbers. The constructed strain signal set was stored in matrix form, where rows corresponded to sampling time points and columns corresponded to strain gauge channels. This strain signal set served as the basic data source for subsequent delay feature extraction.

[0052] Furthermore, to further analyze the strain response characteristics of the adhesive layer at different stages, the continuously acquired strain signals were segmented according to time windows. In this embodiment, each 60-second interval was selected as an analysis interval, so the 10-minute signal was divided into 10 segments, each containing 30,000 sampling points. Subsequently, the adhesive layer delay curve was imported into the system, and key delay reference points were identified from the delay curve. For example, in the actual measurement, three main delay reference points were identified, located at the 120th, 320th, and 480th seconds of the curing monitoring process, corresponding to the inflection point of the adhesive layer curing rate change. The system used these delay reference points as time bases to perform time alignment correction on each strain signal segment: when the peak value of a certain strain signal was detected to appear approximately 0.12 seconds after the reference point, the system automatically shifted the entire signal segment forward by 0.12 seconds; if the peak value was advanced by 0.08 seconds, the entire signal segment was shifted backward by 0.08 seconds. After time shift correction, the signal peak values ​​of all strain gauges were realigned at the reference points. The adjusted strain signal curves show that the peak time difference between the original S1 and S5 signals has decreased from 0.31 seconds to 0.04 seconds, significantly improving time synchronization. The corrected 10 strain data segments were recombined to form a synchronized strain signal set, providing a stable time reference for delay feature extraction.

[0053] Finally, after time correction, delay feature extraction and analysis are performed on each adjusted strain signal segment. The system first calculates the peak correspondence and phase change trend between the signals of each strain gauge to identify local delay characteristics. In this embodiment, S3 is used as the reference channel to compare the relative responses of S1 to S5. The analysis results show that in the first segment (0-60 s), the delay time difference between S1 and S3 is 0.06 s, S2 is 0.03 s, S4 is 0.09 s, and S5 is 0.11 s; in the fifth segment (240-300 s), the delay time difference gradually decreases to between 0.02 and 0.05 s, indicating that the adhesive layer cures gradually and uniformly. In addition to the time difference, the signal phase offset value is also extracted. The average phase offset of S1 to S5 relative to S3 is identified using a phase tracking algorithm, which are +8.2°, +4.5°, -5.1°, -7.4°, and -10.3°, respectively. This yields a delayed signal sequence, comprising a time difference sequence and a phase shift sequence. The plotting results show that both the time difference and phase shift decrease with curing time and stabilize after 480 seconds, indicating that the adhesive layer structure has reached an acoustically stable state. The final output delayed signal sequence is: time difference sequence [0.06, 0.03, 0.09, 0.11] s to [0.02, 0.03, 0.04, 0.05] s, and phase shift sequence [+8.2°, +4.5°, –5.1°, –7.4°, –10.3°] to [+2.1°, +1.3°, –1.2°, –2.0°, –3.1°]. This result serves as an important input for subsequent timing correction and synchronous strain calculations.

[0054] In this example, data analysis is performed on the adhesive layer delay curve, and delay reference points are selected based on the analysis results, specifically as follows:

[0055] Characteristic analysis of the delay curve of the adhesive layer was performed to obtain the points where the curve slope changed;

[0056] The curve is divided into several delay segments by using the points where the slope of the curve changes, and the point of delay peak in each delay segment is identified as the alignment reference point.

[0057] A search window is set near the alignment reference point, and the characteristic response point corresponding to the time position of the alignment reference point is retrieved from the strain signal corresponding to the search window;

[0058] The time of the characteristic response point is used as the starting time of the delay section, and the strain signal of the delay section is time-shifted and corrected according to the starting time to obtain the delay reference point.

[0059] It should be noted that, firstly, data analysis is performed on the adhesive layer delay curve obtained in the previous step. This delay curve is a time delay distribution curve extracted from the adhesive layer delay matrix along the principal strain direction. The horizontal axis represents the adhesive layer position (0-100 mm), and the vertical axis represents the acoustic wave delay characteristic value (unit: In the experimental measurements, the delay value of the left segment of the curve was approximately +0.12, the middle segment slowly decreased to +0.05, and the right segment rapidly decreased to -0.09. The rate of change of the curve was analyzed using a piecewise difference method. By calculating the rate of change of delay at adjacent sampling points, regions where the curve slope changed abruptly were identified. Two significant slope change points were actually detected: the first was located at approximately 30 mm, and the second at approximately 68 mm. These slope change points refer to inflection points where the slope of the delay curve (i.e., the rate of change of the delay value with spatial location) changes abruptly or turns from positive to negative. These typically correspond to boundary locations where the internal physical properties of the adhesive layer change or the curing rate changes abruptly. In other words, these points mark the interfaces where the acoustic properties of the adhesive layer change significantly, and are important bases for subsequent segmentation and feature matching.

[0060] Next, based on the two slope change points obtained in the previous step, the entire delay curve is divided into three delay segments: the first segment (0-30 mm), the second segment (30-68 mm), and the third segment (68-100 mm). Within each segment, the main acoustic response point is identified by finding local maxima. Taking experimental data as an example, the delay curve in the first segment shows a local peak at approximately 18 mm, with a delay characteristic value of +0.11; the peak in the second segment is located at 52 mm, with a value of +0.06; and the peak in the third segment is located at 92 mm, with a value of -0.07. These peak points represent the moments when the energy of the reflected acoustic signal is strongest or the propagation path changes most significantly within the corresponding region, and are therefore defined as delay peak points. Delay peak points reflect the location of the maximum response delay of acoustic wave propagation within the adhesive layer, typically corresponding to local thickness abrupt changes, density anomalies, or curing front regions. The system determines the peak points of each delay segment as alignment reference points, denoted as a candidate set of alignment reference points, for subsequent strain signal matching.

[0061] Furthermore, after obtaining the delay peak points, a time search window is set around each alignment reference point to find significant response features in the strain signal corresponding to that location. The search window size is automatically set to ±0.2 seconds based on the acoustic wave propagation delay range. For example, for the second segment (alignment reference point time position is 3.25 seconds), the search window range is 3.05 to 3.45 seconds. Within this time interval, the system scans each channel of the synchronous strain signal set and analyzes its waveform change rate and amplitude characteristics. The results show that in the S3 strain gauge signal, a significant spike peak (amplitude +38με) appears at 3.27 seconds, while the S4 strain gauge shows a similar peak with an amplitude of +36με at 3.29 seconds. Such waveform abrupt changes usually correspond to the transient response of the stress transmission path in the adhesive layer, and the system identifies these abrupt peaks as characteristic response points. A characteristic response point refers to a significant change point in the strain signal in the time domain corresponding to a characteristic position of the acoustic wave delay curve, reflecting the synchronous relationship between acoustic wave delay and strain response. It typically manifests as signal peaks, troughs, or abrupt phase changes, serving as a crucial basis for time correction positioning.

[0062] Finally, once the characteristic response point is determined, its time position is used as the start time of the corresponding delay segment. Taking the second segment as an example, the characteristic response point time of channel S3 is 3.27 seconds, and the system sets the start time of the second segment to 3.27 seconds accordingly. Subsequently, the system performs time-shift correction on all strain signals within this segment, aligning different strain gauges at the same reference point. For example, if the S1 signal shows a similar peak at 3.30 seconds, delayed by 0.03 seconds from the reference point, the system shifts the S1 signal forward by 0.03 seconds; if the S5 signal is advanced by 0.05 seconds, it is shifted backward by 0.05 seconds. After correction, all channels show a synchronous peak at 3.27 seconds. This alignment time point is the delay reference point, representing the unified response time of each strain channel under the same physical event (such as changes in the bonding layer curing interface) after time-shift correction. In other words, the delay reference point is the time base after aligning the characteristic responses, representing both the synchronization starting point of the strain signals and the anchoring position for strain timing adjustments within the delay segment. After completing the time shift correction for the three delay segments, a continuous and aligned sequence of delay reference points can be obtained, providing an accurate time reference for the subsequent construction of synchronous strain sequences.

[0063] S4. Based on the delayed signal sequence, the strain signal set is time-corrected to obtain the synchronous strain sequence set;

[0064] In this example, the strain signal set is time-corrected based on the delayed signal sequence to obtain a synchronous strain sequence set, specifically:

[0065] The delayed signal sequence is time-matched with the strain signal set, and the time offset of each strain signal is calculated.

[0066] Based on the time offset, a time correction function is constructed, and the strain signal set is resampled and interpolated based on the time correction function to obtain a synchronous strain sequence set.

[0067] It should be noted that, firstly, a time-series matching process is performed between the delayed signal sequence (including delay time difference and signal phase offset) and the aforementioned strain signal set. The strain signal set consists of five strain gauges (S1-S5), with each channel sampling 10 minutes of dynamic strain data. The system calculates the alignment error between the strain signal and the standard time reference by analyzing the time offset characteristics of each channel in the delayed signal sequence. For example, the average delay time differences of S1, S2, S3, S4, and S5 in the delayed signal sequence are 0.06s, 0.03s, 0s, 0.04s, and 0.08s, respectively, with channel S3 set as the timing reference signal. The system uses a cross-correlation algorithm to detect the position of the maximum correlation peak of each strain channel relative to S3, thereby obtaining its timing offset. Actual calculation results show that the signal offset of S1 is -0.058s, S2 is -0.031s, S4 is +0.042s, and S5 is +0.083s. The timing offset reflects the lag or lead of each strain signal in the time domain, mainly caused by local acoustic delay in the adhesive layer, differences in sensor response, or uneven curing. This matching step accurately identifies the timing error of each strain channel, providing precise reference data for subsequent unified correction.

[0068] Furthermore, after obtaining the time offset of each strain gauge, a time correction function is automatically constructed based on the offset characteristics to achieve global time axis unification. This correction function is defined at the software level as a time mapping relationship, corresponding the sampling time of the original strain signal to the corrected time. To ensure a smooth correction process, the system performs piecewise fitting on the offset of each channel, enabling the correction function to simultaneously reflect the gradual characteristics of the curing stage. Taking this embodiment as an example, the initial offset of the S1 signal is –0.058s, remaining stable in the early curing stage (0-120 s), and then gradually decreasing to –0.035s; the initial offset of the S5 signal is +0.083s, decreasing to +0.046s in the later curing stage (480-600s). The system integrates these trends and establishes a function model that includes a nonlinear time correction relationship, making the time offset continuously adjustable throughout the measurement cycle. The output of this function is the corrected standard time coordinate, ensuring that signals acquired by different strain gauges have the same time signature when physical events (such as stress mutations or interface curing completion) occur. By constructing this function, the systematic drift and phase misalignment of the strain signal set on the time axis are eliminated, laying the foundation for subsequent synchronous resampling.

[0069] Finally, after constructing the timing correction function, time resampling and interpolation were performed on all strain signal channels to generate a synchronized strain sequence set. First, the correction function was applied to the original sampling time point of each channel, redefining its time coordinates. Then, linear interpolation was used to fill in the irregular sampling intervals caused by the time correction, ensuring that each channel was resynchronized at the same time step (0.0002s, i.e., 5kHz sampling rate). After resampling, the total number of points for signals S1-S5 remained at 300,000, with completely consistent time axes. To verify the synchronization effect, the system calculated the peak positions of the cross-correlation between each channel after correction. The results showed that the phase difference was controlled within ±0.002s, improving the synchronization accuracy by approximately 40 times compared to before correction (maximum 0.083s). The consistency of the amplitude change trend of the synchronized strain signals at the same time was significantly improved; for example, at 420s, the strain peaks of each channel appeared at the same time, with amplitude differences not exceeding 5%. The system defines the resampled strain dataset as a synchronous strain sequence set, which reflects the synchronous stress response characteristics of the tower crane bonding layer at the real time scale, providing high-precision input for subsequent strain distribution analysis and safety early warning calculation.

[0070] S5 calculates the strain distribution characteristics of tower crane structural components based on a synchronous strain sequence set, and provides safety warnings based on the calculation results.

[0071] In this example, the strain distribution characteristics of tower crane structural components are calculated based on a synchronous strain sequence set, and a safety warning is issued based on the calculation results. Specifically:

[0072] Spatial differential calculations are performed on the synchronous strain sequence set to obtain the strain gradient distribution value at each time point;

[0073] Based on the strain gradient distribution value, abnormal sequences in the synchronous strain sequence set are identified, and the damage degree of the structural component is predicted based on the abnormal sequences.

[0074] When the damage level exceeds the preset threshold, a safety warning signal is triggered and the warning location and severity of the tower crane structural component are output.

[0075] It should be noted that, in the specific implementation process, the system first uses a synchronous strain sequence set to calculate the strain gradient distribution on the surface of the tower crane structural components. Taking five measuring points (S1-S5) on the lower chord of the main beam of the tower crane boom as an example, the spacing between the measuring points is 0.5m, corresponding to the strain gauges arranged in the adhesive layer. The synchronous signal sampling frequency of each strain channel is 5kHz, and the total sampling time is 600s. After time synchronization, the system performs spatial difference processing on the strain values ​​of each channel at the same time, that is, calculates the strain change rate between adjacent measuring points to reflect the uneven stress distribution along the length of the structure. The actual measured strain values ​​of S1 to S5 at 420s are: 152με, 149με, 138με, 121με, and 118με, respectively. The system automatically calculates that the spatial strain gradient at this time is approximately -34με / m at a 0.5m spacing, indicating that the stress decreases from the root of the tower boom to the end. To facilitate visualization, the strain gradient data at each time point were constructed into a two-dimensional distribution map, with the horizontal axis representing time and the vertical axis representing spatial location. Analysis shows that the rate of change of the strain gradient increases significantly in the interval from 300s to 450s, with the maximum gradient occurring at 428s, reaching -62με / m. This indicates that the stress concentration in the local bonding layer of the tower crane structure intensifies at this time, and there may be a local damage initiation zone.

[0076] Furthermore, by dynamically analyzing the aforementioned spatial strain gradient distribution map, abnormal sequences within the synchronous strain sequence set were identified. An abnormal sequence refers to a signal segment whose strain gradient change rate or fluctuation amplitude significantly deviates from normal operating conditions within a certain time window. A sliding time window analysis method (window width of 10 s) was used to statistically analyze the strain gradient standard deviation and transient change rate of each channel. When the change rate of any channel exceeds 2.5 times the overall average, it is marked as an abnormal sequence. Taking 420-440 s as an example, the strain gradient change rates of channels S3 and S4 reached 5.8 με / m and 6.1 με / m per second, respectively, while other channels were only about 1.2 με / m. Based on this, the system determined that an anomaly occurred in the adhesive layer region between S3 and S4. Subsequently, the finite element model of the tower crane structure was called to perform an equivalent stiffness attenuation analysis on the adhesive layer of this local section, calculating that the effective modulus of the adhesive interface in this region decreased by approximately 12%. Based on historical experimental databases, a 10% to 15% decrease in modulus typically corresponds to early debonding or microcrack damage. Therefore, the system calibrates this damage level value as 0.13 (normalized to the 0-1 range). The system continues to track the strain signal in this area and finds that the damage level value further increases to 0.19 after 10 minutes, indicating that the damage is continuing to expand. Through this identification and prediction process, the system can achieve real-time damage monitoring and quantitative assessment of the tower crane's adhesive layer.

[0077] Finally, in practical engineering applications, the system presets a damage threshold of 0.15. When the damage value of any monitoring unit exceeds this threshold, a safety warning mechanism is immediately triggered. Taking the data at time 428s as an example, the system detected a damage value of 0.19 in the S3-S4 section, exceeding the threshold of 0.15. The system automatically generates a first-level warning signal and provides prompts on the tower crane monitoring platform in three ways: 1. Audible and visual alarm—the red indicator light on the controller panel flashes continuously and is accompanied by a buzzer; 2. Data system alarm—the real-time monitoring interface highlights the S3-S4 section and outputs "Abnormal structural strain: suspected adhesive layer delamination, location: 2.0–2.5m below the main beam chord"; 3. Wireless push—the warning information is sent to the on-site tower crane operator's tablet terminal and the back-end safety monitoring center. At the same time, the system labels the strain gradient trend in this area as "rapidly decreasing type" and calculates the current severity index as 0.72 (range 0-1, 1 being severe). Subsequently, the system activated automatic data recording mode, temporarily amplifying the signal acquisition in this area at a sampling rate of 10 times (50kHz) for further analysis. After manual verification, it was found that a surface crack of approximately 0.3mm did indeed exist at the adhesive layer in the S3-S4 section of the tower crane, highly consistent with the system's prediction. This example demonstrates that, under actual working conditions, real-time identification, location, and severity-based early warning of adhesive layer damage to tower crane structural components can be achieved, significantly improving the reliability and response speed of structural safety monitoring.

[0078] The present invention also includes a computer-readable storage medium storing a computer program that, when executed by a processor, implements a method for early warning of inspection of tower crane structural components.

[0079] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0080] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0081] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0082] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0083] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0084] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for early warning and inspection of tower crane structural components, characterized in that, Includes the following steps: Acquire the acoustic wave reflection signal of the adhesive layer of the tower crane structural component during the curing process, and perform feature mapping based on the preset strain gauge spatial coordinates to obtain the adhesive layer delay matrix, specifically: Broadband acoustic signals are emitted into the adhesive layer of the tower crane structural components, and the echo signals reflected from the adhesive layer interface are received by a preset acoustic sensor to obtain the acoustic reflection signals; time-domain analysis is performed on the acoustic reflection signals to obtain the arrival time difference and amplitude attenuation value of the acoustic reflection signals. The acoustic impedance change of the adhesive layer is calculated based on the arrival time difference and amplitude attenuation value of the reflected acoustic signal; the spatial coordinates of the preset strain gauge are aligned with the spatial coordinates of the preset acoustic sensor, and the acoustic impedance change value is mapped to the strain gauge position through spatial interpolation to obtain the delay characteristic distribution value of the strain gauge. The spatial mapping method is used to map the delay characteristic distribution values ​​of the strain gauges onto the mesh to obtain the delay matrix of the adhesive layer; The thickness variation of the adhesive layer is analyzed based on the adhesive layer delay matrix, resulting in the adhesive layer delay curve, as follows: Singular value decomposition is performed on the delay matrix of the adhesive layer to extract principal component eigenvectors. The propagation velocity of the sound wave in the adhesive layer is calculated based on the principal component eigenvectors, and the thickness variation sequence of the adhesive layer is obtained based on the propagation velocity. The delay time sequence of the strain gauge is extracted from the delay matrix of the adhesive layer, and a correlation function is constructed based on the thickness variation sequence of the adhesive layer and the delay time sequence of the strain gauge. The delay moment of the adhesive layer is reconstructed based on the correlation function to obtain the delay curve of the adhesive layer. Extract the strain signal set of the preset strain gauge, and extract the delay feature of the strain signal set based on the delay curve of the adhesive layer to obtain the delay signal sequence, specifically: Strain signals are collected from each preset strain gauge to construct a strain signal set; data analysis is performed on the delay curve of the adhesive layer, and delay reference points are selected based on the analysis results; the strain signal set is segmented to obtain several strain signal segments, and the start time of each signal segment is adjusted according to the delay reference points to obtain the adjusted strain signal; delay features are extracted from the adjusted strain signal to obtain a delay signal sequence, which includes delay time difference and signal phase offset. A synchronous strain sequence set is obtained by performing time-series correction on the strain signal set based on the delayed signal sequence. The strain distribution characteristics of tower crane structural components are calculated based on the synchronous strain sequence set, and safety warnings are issued based on the calculation results.

2. The tower crane structural component inspection and early warning method according to claim 1, characterized in that, The process of analyzing the delay curve of the adhesive layer and selecting delay reference points based on the analysis results is as follows: Characteristic analysis of the delay curve of the adhesive layer was performed to obtain the points where the curve slope changed; The curve is divided into several delay segments by using the points where the slope of the curve changes, and the point of delay peak in each delay segment is identified as the alignment reference point. A search window is set near the alignment reference point, and the characteristic response point corresponding to the time position of the alignment reference point is retrieved from the strain signal corresponding to the search window; The time of the characteristic response point is used as the starting time of the delay section, and the strain signal of the delay section is time-shifted and corrected according to the starting time to obtain the delay reference point.

3. The tower crane structural component inspection and early warning method according to claim 2, characterized in that, The process of performing time-series correction on the strain signal set based on the delayed signal sequence to obtain a synchronous strain sequence set is as follows: The delayed signal sequence is time-matched with the strain signal set, and the time offset of each strain signal is calculated. Based on the time offset, a time correction function is constructed, and the strain signal set is resampled and interpolated based on the time correction function to obtain a synchronous strain sequence set.

4. The tower crane structural component inspection and early warning method according to claim 3, characterized in that, The calculation of strain distribution characteristics of tower crane structural components based on a synchronous strain sequence set, and the provision of safety warnings based on the calculation results, specifically includes: Spatial differential calculations are performed on the synchronous strain sequence set to obtain the strain gradient distribution value at each time point; Based on the strain gradient distribution value, abnormal sequences in the synchronous strain sequence set are identified, and the damage degree value of the structural component is predicted based on the abnormal sequences. When the damage level exceeds the preset threshold, a safety warning signal is triggered and the warning location and severity of the tower crane structural component are output.

5. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements a tower crane structural component inspection and early warning method as described in any one of claims 1 to 4.

Images