Tower crane structural member inspection early warning method and medium

Abstract

The invention discloses a tower crane structural member inspection early warning method and a medium, and relates to the technical field of inspection early warning, and the tower crane structural member inspection early warning method comprises the following steps: obtaining sound wave reflection signals of a bonding layer of a tower crane structural member in a curing process, and carrying out feature mapping in combination with preset strain gauge space coordinates to obtain a bonding layer delay matrix; analyzing the thickness change of the bonding layer based on the bonding layer delay matrix to obtain a bonding layer delay curve; extracting a strain signal set of a preset strain gauge, and performing delay feature extraction on the strain signal set based on the bonding layer delay curve to obtain a delay signal sequence; performing time sequence correction on the strain signal set based on the delay signal sequence to obtain a synchronous strain sequence set; and carrying out strain distribution characteristic calculation on the tower crane structural member based on the synchronous strain sequence set, and carrying out safety early warning based on a calculation result, so that the problems of strain conduction phase difference and asynchronous response of the edge and the center of a strain gauge caused by non-uniform thickness of an adhesive layer are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of inspection and early warning, and more particularly, to a tower crane structural member inspection and early warning method and medium. BACKGROUND

[0002] As an important hoisting equipment in the construction site, the tower crane bears the key operations such as lifting, moving and precise positioning of building materials, and its structural safety is directly related to the safety of construction personnel and the progress of the project. Among the many structural members of the tower crane, especially the jib, the connecting node and the intersection of the jib and the foundation support, the fixation and stress transmission of the members are usually achieved through adhesives, epoxy composites or other adhesive layers.

[0003] However, especially in the case of uneven thickness of the adhesive layer, the conduction of strain in the adhesive layer presents obvious non-uniformity. Due to the difference in local stiffness, density and acoustic impedance caused by the change in the thickness of the adhesive layer, the stress generated during the external load or the construction process cannot be uniformly distributed in the entire adhesive layer, so that the transmission speed and amplitude of the strain at different positions are different. This non-uniform conduction can cause the strain gauges arranged on the surface of the adhesive layer or at the node to have obvious phase difference, that is, under the action of the same load, the response of the edge of the strain gauge and the response of the center position do not arrive at their peak values or specific waveform points at the same time. This phenomenon is manifested as signal lag in actual measurement. This signal lag not only affects the accuracy of the strain gauge measurement data, but also may cause inaccurate strain distribution calculation and structure health assessment, making it difficult to truly reflect the local stress state and thickness change of the adhesive layer. Especially at the long jib or node of the tower crane, a small unevenness in the thickness of the adhesive layer can amplify the stress concentration effect and form a local fatigue or early delamination area.

[0004] In view of the above problems, the present application provides a solution. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a tower crane structural member inspection and early warning method and medium, which combines the thickness change of the adhesive layer of the tower crane structural member with the strain distribution for analysis, so as to solve the problem that uneven thickness of the adhesive layer can cause phase difference of strain conduction and the response of the edge of the strain gauge and the center is not synchronized.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical scheme: A tower structure member inspection and early warning method, comprising the following steps: obtaining an acoustic wave reflection signal of a bonding layer of a tower structure member during curing, and combining a preset strain gauge spatial coordinate to perform feature mapping to obtain a bonding layer delay matrix; analyzing thickness variation of the bonding layer based on the bonding layer delay matrix to obtain a bonding layer delay curve; extracting a strain signal set of the preset strain gauge, and performing delay feature extraction on the strain signal set based on the bonding layer delay curve to obtain a delay signal sequence; performing timing correction on the strain signal set based on the delay signal sequence to obtain 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 warning based on the calculation result.

[0007] In a preferred embodiment, the acoustic wave reflection signal of the bonding layer of the tower structure member during curing is obtained, and the bonding layer delay matrix is obtained by combining the preset strain gauge spatial coordinate for feature mapping, specifically: a wideband acoustic wave signal is emitted to the bonding layer of the tower structure member, and a preset acoustic wave sensor is used to receive a reflected echo signal from the bonding layer interface to obtain the acoustic wave reflection signal; time domain analysis is performed on the acoustic wave reflection signal to obtain the time difference of arrival and amplitude attenuation value of the acoustic wave reflection signal; the acoustic wave impedance variation value of the bonding layer is calculated based on the time difference of arrival and amplitude attenuation value of the acoustic wave reflection signal; the spatial coordinates of the preset strain gauge are aligned with the spatial coordinates of the preset acoustic wave sensor, and the acoustic wave impedance variation value is mapped to the strain gauge position by a spatial interpolation method to obtain the delay feature distribution value of the strain gauge; the delay feature distribution value of the strain gauge is mapped to the grid by a spatial mapping method to obtain the bonding layer delay matrix.

[0008] In a preferred embodiment, the thickness variation of the bonding layer is analyzed based on the bonding layer delay matrix to obtain the bonding layer delay curve, specifically: singular value decomposition is performed on the bonding layer delay matrix to extract a principal component feature vector; the propagation speed of the acoustic wave in the bonding layer is calculated based on the principal component feature vector, and the bonding layer thickness variation sequence is obtained according to the propagation speed; the delay time sequence of the strain gauge is extracted from the bonding layer delay matrix, and an associated function is constructed based on the bonding layer thickness variation sequence and the delay time sequence of the strain gauge; the bonding layer delay matrix is reconstructed based on the associated function to obtain the bonding layer delay curve.

[0009] In a preferred embodiment, the strain signal set of the preset strain gauge is extracted, and the strain signal set is extracted based on the bonding layer delay curve to obtain a delay signal sequence, specifically: collecting strain signals from each preset strain gauge to construct a strain signal set; analyzing the data of the bonding layer delay curve, and screening the delay reference points according to the analysis results; segmenting the strain signal set to obtain a plurality of segments of strain signals and adjusting the starting time of each segment of signals according to the delay reference points to obtain adjusted strain signals; 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 shift.

[0010] In a preferred embodiment, the data of the bonding layer delay curve is analyzed, and the delay reference points are screened according to the analysis results, specifically: performing feature analysis on the glue layer delay curve to obtain a curve slope change point; dividing the curve into a plurality of delay sections using the curve slope change point, and identifying the point of the delay peak value in each delay section as an alignment reference point; setting a search window near the alignment reference point, searching for a feature response point corresponding to the time position of the alignment reference point in the strain signal corresponding to the search window; taking the time of the feature response point as the starting time of the delay section, and performing time shift correction on the strain signals of the delay section according to the starting time to obtain the delay reference point.

[0011] In a preferred embodiment, the strain signal set is time-corrected based on the delay signal sequence to obtain a synchronous strain sequence set, specifically: time-matching the delay signal sequence with the strain signal set to calculate the time sequence offset of each strain signal; constructing a time correction function according to the time sequence offset and resampling and interpolating the strain signal set based on the time correction function to obtain a synchronous strain sequence set.

[0012] In a preferred embodiment, the strain distribution characteristics of the tower structure are calculated based on the synchronous strain sequence set, and a safety warning is given based on the calculation results, specifically: performing spatial differential calculation on the synchronous strain sequence set to obtain strain gradient distribution values at each time point; identifying abnormal sequences in the synchronous strain sequence set based on the strain gradient distribution values, and predicting the damage degree value of the structure based on the abnormal sequences; when the damage degree value exceeds a preset threshold, triggering a safety warning signal and outputting the warning position and severity of the tower structure.

[0013] The technical effects and advantages of the tower structure inspection and early warning method and medium of the present application are as follows: The present application can realize accurate identification of the thickness change and strain conduction delay of the adhesive layer by acquiring the acoustic wave reflection signal of the adhesive layer of the tower crane structural member during curing or operation, combining with the characteristic mapping of the preset strain gauge spatial coordinates; the delay characteristics of the strain signal can be extracted and the time sequence can be corrected based on the delay matrix and delay curve of the adhesive layer, so that the multi-point synchronous strain sequence can be obtained, thereby accurately reflecting the strain distribution characteristics of the tower crane structural member at each position in space; further, through abnormal sequence identification and damage degree prediction of the strain gradient distribution, quantitative evaluation of the potential damage of the structural member can be realized, and when the damage degree exceeds the preset threshold, a safety warning is automatically triggered, and the warning position and severity are output. The method has the following advantages: first, it can effectively overcome the limitations of traditional inspection methods relying on manual visual inspection or single-point detection, and provide high-precision, real-time and continuous structural health monitoring; second, through delay signal extraction and time sequence correction, the strain signal delay problem caused by uneven adhesive layer thickness or material aging can be eliminated, so that the multi-point strain measurement data can be synchronized, and the accuracy of strain distribution calculation can be ensured; third, early detection and quantitative evaluation of local damage of the tower crane structural member can be realized, providing scientific and quantifiable decision basis for construction safety; finally, the method is suitable for all-around inspection of all key structural members of the tower crane, and is simple to operate and has high automation, which significantly improves the operation safety and structural maintenance efficiency of the tower crane. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 FIG. 1 is a flowchart of a tower crane structural member inspection and early warning method according to the present application. DETAILED DESCRIPTION

[0015] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0016] Embodiment 1, Figure 1 The present application provides a tower crane structural member inspection and early warning method, which comprises the following steps: S1, acquiring the acoustic wave reflection signal of the adhesive layer of the tower crane structural member during curing, and combining with the characteristic mapping of the preset strain gauge spatial coordinates to obtain an adhesive layer delay matrix; In this example, the acoustic wave reflection signal of the adhesive layer of the tower crane structural member during curing is acquired, and the characteristic mapping of the preset strain gauge spatial coordinates is combined to obtain an adhesive layer delay matrix, which is specifically: A wideband acoustic wave signal is emitted to the adhesive layer of the tower crane structural member, and a preset acoustic wave sensor is used to receive the echo signal reflected from the adhesive layer interface to obtain the acoustic wave reflection signal; The time-domain analysis is performed on the sound wave reflection signal to obtain the time difference of arrival and amplitude attenuation value of the sound wave reflection signal; The sound wave impedance change value of the bonding layer is calculated based on the time difference of arrival and amplitude attenuation value of the sound wave reflection signal; The spatial coordinates of the preset strain gauges are aligned with the spatial coordinates of the preset sound wave sensors, and the sound wave impedance change value is mapped into the strain gauge position by a spatial interpolation method to obtain the delay characteristic distribution value of the strain gauges; The delay characteristic distribution value of the strain gauges is mapped into the grid by a spatial mapping method to obtain the delay matrix of the bonding layer.

[0017] In this example, the calculation formula of the sound wave impedance change value of the bonding layer is as follows:

[0018] wherein, is the sound wave impedance change value of the bonding layer, is the reference sound wave impedance of the bonding layer, is the reflection sound wave amplitude, is the incident sound wave amplitude, is the time difference of arrival of the sound wave reflection signal, is the ideal sound wave propagation time.

[0019] It should be noted that a wideband piezoelectric transducer with a center frequency of 2 MHz and a frequency band width of 1.5 MHz can be used as the sound wave emitting device. The transducer generates a wideband pulse sound wave signal through a dedicated pulse exciter, with a pulse width of about 2 μs, a peak voltage of 100 V, and a repetition frequency of 200 Hz. The sound wave is uniformly conducted to the bonding layer at the tower crane arm node through a coupling agent (such as silicone or epoxy glue), and the bonding layer has a thickness of generally 2.5 mm and is formed by the combination of an epoxy resin-based glue layer and a steel substrate. When the wideband sound wave propagates to the bonding layer and the metal interface, part of the energy is reflected back. Eight preset sound wave sensors are arranged on the surface of the tower crane structure, with a model of Olympus V133-RM, a sensor spacing of about 50 mm, and a sampling frequency of 10 MHz. Each sensor receives the echo signal within a time window of 0-10 μs after the sound wave is emitted. The system synchronously records the reflection signal waveforms received by each sensor through a data acquisition module, and the peak voltage of the signal is generally between 15-60 mV, forming an original sound wave reflection signal data set.

[0020] In addition, the acquired reflection signals are first band-pass filtered (0.5-3 MHz) to suppress ambient noise, and then time-domain envelope extraction is performed to identify the main echo peaks. The signal waveform of each sensor is time-stamped within a sampling window of 0-10 μs, and the first main peak is selected as the reference wave, and the second main peak is selected as the adhesive layer interface reflection signal. By comparing the time positions of the two main peaks, the time difference of arrival of the signal can be obtained, for example, in actual testing, the time difference of arrival of sensor No. 1 is 1.32 μs, sensor No. 2 is 1.47 μs, and sensor No. 8 is 1.65 μs. At the same time, the amplitude attenuation value can be obtained by normalizing the peak voltage of the two main peaks, and the typical result is between 0.42 and 0.58. The time-domain analysis results show that the local wave velocity and energy loss of the adhesive layer have spatial differences, providing basic data for subsequent calculation of acoustic impedance.

[0021] Further, based on the time difference of arrival and amplitude attenuation value obtained by the above time-domain analysis, the acoustic characteristics change inside the adhesive layer can be inferred by the sound wave propagation delay and reflection energy change. Specifically, when the adhesive layer has micro-bubbles or insufficiently solidified areas, the sound wave reflection intensity will weaken, which is manifested as an increase in amplitude attenuation value; when the adhesive layer density increases or the thickness slightly decreases, the reflection time difference will decrease. In this embodiment, the acoustic impedance reference value of the sampling point is set to , the change value corresponding to sensor No. 1 is , sensor No. 2 is , and sensor No. 8 is . By statistically analyzing the impedance change trend of all sensors, it can be found that there is a slight density gradient distribution from left to right in the adhesive layer during the solidification process, indicating that the local solidification rate is different.

[0022] Secondly, in addition to the acoustic sensors, the surface of the tower crane arm node is also uniformly arranged with 5 strain gauges, numbered S1 to S5, with a strain gauge spacing of 60 mm, and a three-dimensional coordinate system is established with the node center as the origin. The coordinates of the acoustic sensors are obtained by a three-dimensional scanner with an accuracy of ±0.2 mm. First, the two sets of coordinate data are aligned in the coordinate system, and a rigid body registration algorithm based on least squares error is used to align the spatial distribution of the sensors and the strain gauges. After alignment, the distance between sensor No. 1 and strain gauge S1 is 15 mm, and the distance between sensor No. 8 and strain gauge S5 is 12 mm. In order to obtain the acoustic impedance change value at the position of the strain gauges, an inverse distance weighted (IDW) spatial interpolation method is used to interpolate the impedance change values of the 8 sensors to the positions of the 5 strain gauges. The final delay characteristic distribution values of strain gauges S1-S5 are +0.13, +0.10, +0.04, –0.05, and –0.09 (unit same as before), respectively.

[0023] Finally, to obtain the delay spatial distribution of the whole bonding layer, the delay characteristic distribution values of the strain gauges are further mapped into a two-dimensional grid model. The grid is divided into 10x10 cells in a 100mmx100mm analysis plane, and each cell represents a local area of the bonding layer. First, the delay characteristic values of the strain gauges S1-S5 are projected onto the grid according to their coordinates, and then the delay values of the remaining areas in the grid are filled by using a bilinear interpolation method. In actual calculation, the delay value of the center area of the grid (about S3 position) is about , the upper left corner area (close to S1) reaches , and the lower right corner (close to S5) is . The delay matrix of the bonding layer finally generated by the system is 10x10 dimension, and the matrix elements correspond to the delay characteristic values of each cell. After color mapping display, it can be directly observed that the left side of the tower node bonding layer is fully cured and the right side is slightly delayed, which provides intuitive basic data for subsequent thickness change analysis.

[0024] S2, analyzing the thickness change of the bonding layer based on the bonding layer delay matrix to obtain a bonding layer delay curve; In this example, the thickness change of the bonding layer is analyzed based on the bonding layer delay matrix to obtain a bonding layer delay curve, which is specifically: Perform singular value decomposition on the bonding layer delay matrix to extract principal component feature vectors; Calculate the propagation speed of sound waves in the bonding layer based on the principal component feature vectors, and obtain the bonding layer thickness change sequence according to the propagation speed; Extract the delay time sequence of the strain gauges from the bonding layer delay matrix, and construct a correlation function based on the bonding layer thickness change sequence and the delay time sequence of the strain gauges; Reconstruct the bonding layer delay matrix based on the correlation function to obtain the bonding layer delay curve.

[0025] It should be noted that the bonding layer delay matrix is a 10x10 two-dimensional matrix, and each element of the matrix represents the acoustic wave delay characteristic value of the area, with a unit 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 difference. To analyze the main variation trend, the delay matrix is imported into the signal feature analysis module for singular value decomposition operation. Three significant eigenvector groups are obtained by decomposition, of which the singular value of the first eigenvector accounts for 78.4% of the total energy, the second accounts for 13.2%, and the third accounts for 4.9%, and the remaining eigenvalues can be ignored. The principal component eigenvector reflects the main spatial mode of the adhesive layer delay distribution. Taking the first eigenvector as an example, its components gradually transition from positive to negative in the left-up to right-down direction, indicating that there is a significant thickness or density gradient in the adhesive layer. This vector is extracted as the principal component eigenvector input for subsequent calculations.

[0026] Secondly, in order to verify whether the spatial features extracted by singular value decomposition are consistent with the acoustic signal changes, the acoustic reflection signals corresponding to the same region are analyzed again in the time domain. This analysis uses the original signals measured by the acoustic sensors, and each sensor records the waveform sample in the 0-10 μs time window. Taking sensors 1 and 8 as examples, after envelope extraction, the main echo peak of sensor 1 appears at 1.28 μs, with a peak amplitude of 55 mV; the main echo of sensor 8 appears at 1.62 μs, with an amplitude of 43 mV. The comparison results show that the time difference in the region is 0.34 μs, and the amplitude attenuation ratio is about 0.78. By processing all 8 sensors in the same way, the arrival time difference is 0.30-0.39 μs, and the amplitude attenuation value is distributed between 0.75 and 0.85. These data provide basic acoustic parameters for the thickness variation calculation of the adhesive layer, and serve as the verification basis for the principal component vectors in the singular value decomposition matrix.

[0027] Further, after obtaining the principal component eigenvector, it is associated with the time difference distribution of each sensor to estimate the propagation speed of sound waves in different regions. Under experimental conditions, the adhesion layer of the tower crane is an epoxy-based composite material, and the sound speed after curing is about 2400 m / s, while due to the internal temperature gradient and the influence of bubbles during the curing process, the local sound speed will fluctuate. Based on the spatial pattern of the principal component eigenvector, the adhesion layer is divided into three types of regions: high density area (left side), neutral area (middle), and low density area (right side). The calculation results show that the average propagation speed of the high density area is 2455 m / s, the neutral area is 2380 m / s, and the low density area is 2260 m / s. According to the difference of the propagation speed of each region, combined with the time difference of the arrival of sound waves, the thickness variation sequence of the adhesion layer is obtained: the left region is about 2.46 mm, the middle region is 2.53 mm, and the right region is 2.67 mm. It can be seen that the thickness of the adhesion layer increases slightly from left to right, indicating that the shrinkage of the right side of the glue layer is not sufficient during the curing process. The thickness variation sequence is recorded as the dynamic thickness characteristics of the adhesion layer.

[0028] Secondly, in order to further establish the coupling relationship between the thickness variation of the adhesion layer and the response of the strain gauge, the delay time sequence corresponding to the coordinates of the strain gauges S1-S5 is extracted from the aforementioned 10x10 delay matrix. The specific results are as follows: S1 is 1.29 μs, S2 is 1.34 μs, S3 is 1.41 μs, S4 is 1.53 μs, and S5 is 1.61 μs. After arranging these delay time points in chronological order to form a delay time sequence, it is associated with the thickness variation sequence (2.46, 2.53, 2.60, 2.64, 2.67 mm) obtained in the third step. Through the data analysis module, the correlation between the two is calculated, and the correlation coefficient is about 0.92, indicating that the strain gauge delay time and the adhesion layer thickness variation have a high degree of coupling. Based on this relationship, an associated function is automatically established to represent the delay response law caused by the thickness variation. The physical meaning of this function is that when the local thickness of the adhesion layer increases, the sound wave propagation path lengthens, the delay time increases, and the phase signal corresponding to the strain gauge lags.

[0029] Finally, after obtaining the correlation function between the thickness variation and the delay time, it is applied to the original adhesion layer delay matrix to modify and reconstruct the delay eigenvalue of each grid element. The modification process smooths the sudden values in the original matrix caused by noise or local abnormal points through interpolation. The reconstructed matrix presents 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 line data of the matrix along the principal strain direction is extracted and plotted into the adhesion layer delay curve. The curve shape shows that the delay value remains slowly decreasing in the range of 0-50 mm, and there is a significant steep drop in the range of 50-100 mm, indicating that the right side of the adhesion layer cures slowly. The peak-to-valley difference of the curve is about 0.2, which is consistent with the thickness variation sequence obtained in the third step. This result indicates that the adhesion layer thickness variation sequence can be used to represent the dynamic characteristics of the adhesion layer. The bonding layer delay curve is in good agreement with the corresponding relationship between the sound wave propagation speed and the thickness change. The bonding layer delay curve can directly reflect the dynamic changes of the sound wave propagation characteristics and the uniformity of the bonding layer during the curing process, providing key basis for subsequent strain signal delay correction and structure safety evaluation.

[0030] S3, extracting a strain signal set of the preset strain gauges, and performing delay feature extraction on the strain signal set based on the bonding layer delay curve to obtain a delay signal sequence; In this example, a strain signal set of the preset strain gauges is extracted, and delay feature extraction is performed on the strain signal set based on the bonding layer delay curve to obtain a delay signal sequence, specifically as follows: Collecting strain signals from each preset strain gauge to construct a strain signal set; Performing data analysis on the bonding layer delay curve, and selecting delay reference points according to the analysis results; Segmenting the strain signal set to obtain a plurality of segments of strain signals and adjusting the starting time of each segment of signals according to the delay reference points to obtain adjusted strain signals; Extracting delay features from the adjusted strain signals to obtain a delay signal sequence, the delay signal sequence including delay time differences and signal phase shifts.

[0031] It should be noted that the bonding layer region of the tower crane jib node is uniformly arranged with 5 strain gauges numbered S1 to S5, and the strain gauge model is BF350-3AA type resistance strain gauge with a resistance value of 350Ω and a sensitivity coefficient of 2.0. Each strain gauge is connected to a high-precision dynamic strain acquisition module through a full-bridge measurement circuit, with a sampling frequency of 5 kHz and a sampling accuracy of ±2με. During acquisition, the tower crane structural member is in the curing monitoring stage, the temperature is kept at 25℃, and the external load is a periodic simulated wind load with a frequency of 0.5 Hz and an amplitude of 0.2 kN. The system continuously acquires strain signals for 10 minutes, and each strain gauge obtains about 300000 data points. After noise filtering (using a 3rd order Butterworth low-pass filter) and baseline correction, the acquired data form an original strain signal curve. The actual data shows that the strain range of S1 is ±45με, that of S2 is ±52με, that of S3 is ±49με, that of S4 is ±60με, and that of S5 is ±57με. The time sequence signals of the five strain gauges are combined into a five-dimensional signal set according to the sensor number, and the constructed strain signal set is stored in a matrix form, in which the rows correspond to the sampling time points and the columns correspond to the strain gauge channels. The strain signal set serves as a basic data source for subsequent delay feature extraction.

[0032] Further, in order to further analyze the strain response characteristics of the adhesive layer at different stages, the continuously collected strain signals are segmented and processed according to time windows. In this embodiment, each 60 seconds is selected as an analysis interval, so that the 10-minute signal is divided into 10 segments, each containing 30000 sampling points. Subsequently, the adhesive layer delay curve is imported into the system, and key delay reference points are identified from the delay curve. For example, in actual measurement, a total of 3 main delay reference points are identified, which are located at the 120th second, the 320th second and the 480th second of the curing monitoring process, corresponding to the turning points of the adhesive layer curing rate change. The system takes these delay reference points as the time reference, and corrects each segment of the strain signal in time alignment: when it is detected that the peak value of a segment of the strain signal appears after the delay of about 0.12 seconds from the reference point, the system automatically moves the whole signal forward by 0.12 seconds; if the peak value is ahead of 0.08 seconds, it is moved backward by 0.08 seconds. After time shift correction, the signal peak values of all strain gauges are re-aligned at the reference point. The adjusted strain signal curve shows that the peak time difference between S1 and S5 signals originally reduced from 0.31 seconds to 0.04 seconds, greatly improving the time synchronization. The 10 segments of strain data after correction are recombined to form a synchronized strain signal set, which provides a stable time reference for delay feature extraction.

[0033] Finally, after the completion of the time correction, the adjusted strain signal of each segment is analyzed for delay feature extraction. The system first calculates the peak value correspondence and phase change trend between each strain gauge signal to identify the local delay characteristics. In this embodiment, S3 is taken as the reference channel, and the relative responses of S1 to S5 are compared. The analysis results show that in the first segment (0-60 s), the delay time difference of S1 relative to S3 is 0.06 s, S2 is 0.03 s, S4 is 0.09 s, and S5 is 0.11 s; by the fifth segment (240-300 s), the delay time difference gradually decreases to between 0.02-0.05 s, indicating that the curing of the adhesive layer gradually becomes uniform. In addition to the time difference, the signal phase shift value is also extracted, and the average phase shift of S1-S5 relative to S3 is identified using a phase tracking algorithm, which is +8.2°, +4.5°, -5.1°, -7.4°, and -10.3°, respectively. The delay signal sequence is obtained, including the delay time difference sequence and the phase shift sequence. The results show that the delay time difference and the phase shift both show a decaying trend with the curing time, and tend to be stable after 480 seconds, indicating that the adhesive layer structure has reached an acoustic stable state. The final output delay 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°], which serves as an important input for subsequent timing correction and synchronous strain calculation.

[0034] In this example, the adhesive layer delay curve is analyzed, and the delay reference points are selected based on the analysis results, specifically: Characteristic analysis is performed on the adhesive layer delay curve to obtain the curve slope change point; The curve is divided into several delay segments using the curve slope change point, and the points of the delay peaks in each delay segment are identified as alignment reference points; A search window is set near the alignment reference point, and a feature response point corresponding to the time position of the alignment reference point is searched in the strain signal corresponding to the search window; The time of the feature response point is taken as the start time of the delay segment, and the strain signal of the delay segment is time-shifted according to the start time to obtain the delay reference point.

[0035] It should be noted that the adhesive layer delay curve obtained in the previous step is first analyzed. The delay curve is a time delay distribution curve obtained by extracting the adhesive layer delay matrix along the principal strain direction, with the horizontal axis representing the adhesive layer position (0-100 mm) and the vertical axis representing the acoustic wave delay characteristic value (unit: ). In the experimental measurement, the left segment of the curve has a delay value of about +0.12, the middle segment slowly decreases to +0.05, and the right segment rapidly decreases to -0.09. The change rate of the curve is analyzed by using the piecewise difference method, and by calculating the delay change rate of adjacent sampling points, the area where the slope of the curve changes significantly is identified. Two significant slope change points are actually detected: the first is located at a position of about 30 mm, and the second is located at a position of about 68 mm. The slope change point of the curve here refers to the inflection point where the slope of the delay curve (i.e., the rate of change of the delay value with the spatial position) changes or changes from positive to negative, which usually corresponds to the boundary position of the internal physical property change or the curing speed mutation of the adhesive layer. In other words, these points mark the interfaces where the acoustic properties of the adhesive layer change significantly, and are an important basis for subsequent segmentation and feature matching.

[0036] Secondly, according to 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). In each segment, the main acoustic response point of the segment is identified by finding the local maximum value. Taking the experimental data as an example, the delay curve in the first segment has a local peak at about 18 mm, with a delay feature value of +0.11; the peak in the second segment is located at 52 mm, with a value of +0.06; 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 acoustic wave reflection signal is the strongest or the propagation path changes most significantly in the corresponding region, and therefore are defined as the delay peak points. The delay peak points reflect the position of the maximum response of the acoustic wave propagation delay in the adhesive layer, and usually correspond to the local thickness mutation, density anomaly, or curing front region. The system determines the peak point of each delay segment as the alignment reference point, denoted as the candidate set of alignment reference points, which is used for subsequent strain signal matching.

[0037] Further, after obtaining the delay peak point, a time search window is set around each alignment reference point to find a significant response feature in the strain signal corresponding to the position. The search window size is automatically set to ±0.2 seconds according to the sound wave propagation delay range. For example, for the second section (the alignment reference point time position is 3.25 seconds), the search window range is 3.05-3.45 seconds. Within this time interval, the system scans each channel of the synchronous strain signal set to analyze the waveform change rate and amplitude characteristics. As a result, it is found that in the S3 strain gauge signal, a significant peak value (amplitude +38με) appears at 3.27 seconds, and the S4 strain gauge appears a similar peak value of +36με at 3.29 seconds. Such waveform mutations usually correspond to the transient response of the stress transmission path in the adhesive layer, and the system identifies these mutation peaks as characteristic response points. The so-called characteristic response point refers to the significant change point in the time domain of the strain signal corresponding to the characteristic position of the sound wave delay curve, reflecting the synchronization relationship between the sound wave delay and the strain response. It usually appears as a signal peak, trough or phase mutation point, and is an important positioning basis for time correction.

[0038] Finally, when the characteristic response point is determined, the time position thereof is taken as the starting time of the corresponding delay section. Taking the second section as an example, the characteristic response point time of the S3 channel is 3.27 seconds, and the system sets the starting time of the second section as 3.27 seconds accordingly. Subsequently, the system performs time shift correction on all strain signals in the section to align different strain gauges at the same reference point. For example, the S1 signal appears a similar peak value at 3.30 seconds, which is delayed by 0.03 seconds from the reference point, so the system shifts the S1 signal forward by 0.03 seconds as a whole; the S5 signal is shifted forward by 0.05 seconds, so it is shifted backward by 0.05 seconds as a whole. After correction, the peak values of all channels appear synchronously at 3.27 seconds. At this time, the alignment time point is the delay reference point, which represents the unified response time of each strain channel to the same physical event (such as the curing interface change of the adhesive layer) after time shift correction. In other words, the delay reference point is the time reference after aligning the characteristic response, which represents the synchronous starting point of the strain signal and also serves as the anchor position for strain timing adjustment in the delay section. After completing the time shift correction of the three delay sections, a continuous and aligned delay reference point sequence is obtained, which provides an accurate time reference for subsequent synchronous strain sequence construction.

[0039] S4, performing timing correction on the strain signal set based on the delay signal sequence to obtain a synchronous strain sequence set; In this example, timing correction is performed on the strain signal set based on the delay signal sequence to obtain a synchronous strain sequence set, specifically as follows: The delay signal sequence is time-matched with the strain signal set to calculate the timing offset of each strain signal; According to 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.

[0040] It should be noted that, first, the delay signal sequence (including delay time difference and signal phase offset) is matched with the aforementioned strain signal set in time. The strain signal set is composed of five strain gauges (S1-S5), and each channel samples 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 delay signal sequence. For example, the average delay time difference of S1, S2, S3, S4, and S5 in the delay signal sequence is 0.06s, 0.03s, 0s, 0.04s, and 0.08s, respectively, and S3 channel is set as the time reference signal. The system uses the cross-correlation algorithm to detect the maximum correlation peak position of each strain channel relative to S3, thereby obtaining its time offset. The actual calculation result shows that the S1 signal offset is -0.058s, the S2 signal offset is -0.031s, the S4 signal offset is +0.042s, and the S5 signal offset is +0.083s. This time offset reflects the response lag or advance of each strain signal in the time domain, which is mainly caused by local acoustic delay of the bonding layer, sensor response difference, or uneven curing. Through this matching step, the time error of each strain channel can be accurately identified, providing accurate reference data for subsequent unified correction.

[0041] Further, after obtaining the time offset of each strain gauge, a time correction function is automatically constructed according to the offset characteristics for realizing the unification of the global time axis. The correction function is defined as a time mapping relationship at the software level, which corresponds the sampling time of the original strain signal to the corrected time. In order to ensure the smoothness of the correction process, the system performs piecewise fitting on the offset of each channel, so that the correction function can reflect the gradual change characteristics of the curing stage. Taking this embodiment as an example, the initial offset of S1 signal is -0.058s, which remains stable at the beginning of curing (0-120s) and then gradually decreases to -0.035s; the initial offset of S5 signal is +0.083s, which decreases to +0.046s at the later stage of curing (480-600s). The system establishes a function model containing nonlinear time correction relationship by comprehensively considering these change trends, so that the time offset is continuously adjustable in the entire measurement period. The output of this function is the corrected standard time coordinate, which ensures that the signals collected by different strain gauges have the same time identifier when a physical event (such as stress mutation or interface curing completion) occurs. Through the construction of this function, the systematic drift and phase misalignment of the strain signal set on the time axis are eliminated, laying a foundation for subsequent synchronous resampling.

[0042] Finally, after the construction of the timing correction function, time resampling and interpolation are performed on all strain signal channels to generate a synchronized strain sequence set. First, the correction function is applied to the original sampling time points of each channel to redefine its time coordinates. Then, linear interpolation is used to fill the irregular sampling intervals caused by time correction, ensuring that the channels are resynchronized at the same time step (0.0002s, i.e. 5kHz sampling rate). After resampling, the total number of points for S1-S5 signals remains at 300000 points, and the time axis is completely consistent. To verify the synchronization effect, the system calculates the cross-correlation peak positions between the corrected channels, and the results show that the phase difference is controlled within ±0.002s, which is about 40 times more accurate than before correction (maximum 0.083s). The amplitude variation trend of the synchronized strain signals at the same time is significantly improved, for example, at 420s, the strain peak values of each channel appear at the same time, and the amplitude difference is not more than 5%. The system defines the resampled strain data set as a synchronized strain sequence set, which reflects the synchronous stress response characteristics of the tower crane bonding layer in the real time scale, and provides high-precision input for subsequent strain distribution analysis and safety warning calculation.

[0043] S5, based on the synchronized strain sequence set, the strain distribution characteristics of the tower crane structure are calculated, and safety warning is carried out based on the calculation results.

[0044] In this example, based on the synchronized strain sequence set, the strain distribution characteristics of the tower crane structure are calculated, and safety warning is carried out based on the calculation results, specifically: From the synchronized strain sequence set, spatial differentiation calculation is performed to obtain the strain gradient distribution value at each time point; Based on the strain gradient distribution value, the abnormal sequence in the synchronized strain sequence set is identified, and the damage degree value of the structure is predicted based on the abnormal sequence; When the damage degree value exceeds the preset threshold value, a safety warning signal is triggered and the warning position and severity of the tower crane structure are output.

[0045] It should be noted that in the specific implementation process, the system first calculates the strain gradient distribution on the surface of the tower crane structure member by using the synchronous strain sequence set obtained in claim 6. Taking five measuring points (S1-S5) on the lower chord of the tower crane boom main beam as an example, the distance between the measuring points is 0.5 m, and the strain gauges arranged on the adhesive layer correspond. The synchronous signal sampling frequency of each strain channel is 5 kHz, and the total sampling time is 600 s. 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 of the structure along the length direction. It is actually measured that at the 420th second, the strain values of S1-S5 are 152με, 149με, 138με, 121με, and 118με respectively. The system automatically calculates that the spatial strain gradient at this moment is about -34με / m on a 0.5 m interval, indicating that the stress decreases from the root to the end of the tower arm. In order to facilitate visualization, the strain gradient data at each moment is constructed into a two-dimensional distribution graph, with the horizontal axis representing time and the vertical axis representing spatial position. Analysis shows that in the interval of 300s to 450s, the change rate of strain gradient increases significantly, and the maximum gradient appears at the 428th second, reaching -62με / m, indicating that the local adhesive layer stress concentration of the tower crane structure intensifies at this time, and there may be a local damage initiation area.

[0046] Further, by dynamically analyzing the above spatial strain gradient distribution graph, the abnormal sequence existing in the synchronous strain sequence set is identified. The abnormal sequence refers to the signal segment whose strain gradient change rate or fluctuation amplitude significantly deviates from the normal working condition within a certain time window. The sliding time window analysis method (window width is 10s) is adopted to calculate the strain gradient standard deviation and transient change rate of each channel. When the change rate of any channel exceeds 2.5 times of the average value, it is marked as an abnormal sequence. Taking the 420th-440th second as an example, the strain gradient change rates of S3 and S4 channels are 5.8με / m and 6.1με / m per second respectively, while those of other channels are only about 1.2με / m. The system determines that the adhesive layer region between S3 and S4 is abnormal accordingly. Subsequently, the finite element model of the tower crane structure is called to perform equivalent stiffness attenuation analysis on the local adhesive layer, and it is calculated that the effective modulus of the adhesive interface in this region decreases by about 12%. According to the historical experimental database, a 10%=15% modulus drop usually corresponds to the early stage of delamination or micro-crack damage, so the system marks the damage degree value as 0.13 (normalized to the range of 0-1). The system continues to track the strain signal of this region and finds that the damage degree value further rises to 0.19 after 10 minutes, indicating that the damage is continuously expanding. Through this identification and prediction process, the system can realize real-time damage monitoring and quantitative evaluation of the adhesive layer of the tower crane.

[0047] Finally in the actual engineering application, the system presets the damage degree threshold value as 0.15, when the damage degree value of any monitoring unit exceeds the threshold value, the safety early warning mechanism is triggered immediately. Taking the data at the 428s time point as an example, the system detects that the damage degree value of the S3-S4 section is 0.19, which exceeds the threshold value 0.15. The system automatically generates a first level warning signal, and prompts on the tower crane monitoring platform through three ways: 1. Acoustic light alarm - the red indicator light on the controller panel flashes continuously and is accompanied by a buzzing prompt; 2. Data system alarm - the real-time monitoring interface highlights at the S3-S4 section, and outputs "structure strain anomaly: suspected adhesive layer delamination, location: 2.0-2.5m below the main beam chord"; 3. Wireless push - sends the warning information to the on-site tower crane operator tablet terminal and the background safety monitoring center. At the same time, the system marks the strain gradient trend of the region as "rapidly decreasing type", and calculates the current severity index as 0.72 (range 0-1, 1 is severe). Subsequently, the system starts the automatic data recording mode, and temporarily enhances the collection of signals in this region at a 10 times sampling rate (50kHz) in order to further analyze. After manual review, it is found that there is indeed a surface crack of about 0.3mm at the adhesive layer of the S3-S4 section of the tower crane, which is highly consistent with the system prediction result. This example shows that in actual working conditions, real-time identification, positioning and level warning of the adhesive layer damage of the tower crane structure can be realized, and the reliability and response speed of the structure safety monitoring are significantly improved.

[0048] The application also includes a computer readable storage medium storing a computer program, which, when executed by a processor, implements a tower crane structure inspection and early warning method.

[0049] The above formulas are all dimensionless numerical calculations, and the formulas are obtained by software simulation of a large amount of data to obtain a formula of the nearest real situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0050] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized in the form of a computer program product wholly or partially.

[0051] Those skilled in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0052] In addition, each function module in each embodiment of the present application can be integrated in one processing module, or each module can be physically present alone, or two or more modules can be integrated in one module.

[0053] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0054] Finally: the above is only a preferred embodiment of the present application and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A tower structure member inspection and early warning method, characterized in that, The method comprises the following steps: acquiring an acoustic wave reflection signal of the adhesive layer of the tower structure during a curing process, and performing feature mapping in combination with preset spatial coordinates of strain gauges to obtain an adhesive layer delay matrix; analyzing thickness variation of the adhesive layer based on the adhesive layer delay matrix to obtain an adhesive layer delay curve; extracting a strain signal set of the preset strain gauges, and performing delay feature extraction on the strain signal set based on the adhesive layer delay curve to obtain a delay signal sequence; performing timing correction on the strain signal set based on the delay signal sequence to obtain a synchronous strain sequence set; performing strain distribution feature calculation on the tower structure based on the synchronous strain sequence set, and performing safety warning based on a calculation result.

2. The method according to claim 1, characterized in that, The method of acquiring the acoustic wave reflection signal of the adhesive layer of the tower structure during the curing process, and performing feature mapping in combination with the preset spatial coordinates of the strain gauges to obtain the adhesive layer delay matrix comprises the following steps: emitting a wideband acoustic wave signal to the adhesive layer of the tower structure, and receiving a reflected echo signal from an interface of the adhesive layer by 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 a time difference of arrival and an amplitude attenuation value of the acoustic wave reflection signal; calculating an acoustic wave impedance variation value of the adhesive layer based on the time difference of arrival and the amplitude attenuation value of the acoustic wave reflection signal; aligning the spatial coordinates of the preset strain gauges with the spatial coordinates of the preset acoustic wave sensor, and mapping the acoustic wave impedance variation value to the strain gauge positions by using a spatial interpolation method to obtain delay feature distribution values of the strain gauges; mapping the delay feature distribution values of the strain gauges to a grid by using a spatial mapping method to obtain the adhesive layer delay matrix.

3. The method according to claim 2, characterized in that, The method of analyzing the thickness variation of the adhesive layer based on the adhesive layer delay matrix to obtain the adhesive layer delay curve comprises the following steps: performing singular value decomposition on the adhesive layer delay matrix to extract principal component feature vectors; calculating a propagation speed of the acoustic wave in the adhesive layer based on the principal component feature vectors, and obtaining a thickness variation sequence of the adhesive layer according to the propagation speed; extracting a delay time sequence of the strain gauges from the adhesive layer delay matrix, and constructing a correlation function based on the thickness variation sequence of the adhesive layer and the delay time sequence of the strain gauges; reconstructing the adhesive layer delay matrix based on the correlation function to obtain the adhesive layer delay curve.

4. The method according to claim 3, characterized in that, The method of extracting the strain signal set of the preset strain gauges, and performing delay feature extraction on the strain signal set based on the adhesive layer delay curve to obtain the delay signal sequence comprises the following steps: collecting strain signals from each preset strain gauge to construct the strain signal set; performing data analysis on the adhesive layer delay curve, and selecting delay reference points according to an analysis result; segmenting the strain signal set to obtain a plurality of segments of strain signals, and adjusting a starting time of each segment of signal according to the delay reference points to obtain adjusted strain signals; extracting delay features from the adjusted strain signals to obtain the delay signal sequence, wherein the delay signal sequence comprises a delay time difference and a signal phase shift.

5. The method according to claim 4, wherein the method further comprises: The method of performing data analysis on the adhesive layer delay curve, and selecting delay reference points according to an analysis result comprises the following steps: performing feature analysis on the adhesive layer delay curve to obtain a curve slope change point; The curve is divided into a plurality of delay segments by using the change points of the curve slope, and a point of a delay peak value in each delay segment is identified as an alignment reference point; A search window is set near the alignment reference point, and a characteristic response point corresponding to the time position of the alignment reference point is searched in a strain signal corresponding to the search window; The time of the characteristic response point is taken as a start time of the delay segment, and the strain signal of the delay segment is time-shifted and corrected according to the start time to obtain a delay reference point.

6. The method according to claim 5, wherein the method further comprises: The strain signal set is time-corrected based on the delay signal sequence to obtain a synchronous strain sequence set, specifically as follows: The delay signal sequence is time-matched with the strain signal set, and a time offset of each strain signal is calculated; According to 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 the synchronous strain sequence set.

7. The method according to claim 6, wherein the method further comprises the steps of: determining the position of the tower structure member; and determining the distance between the tower structure member and the crane. The strain distribution characteristics of the tower structure are calculated based on the synchronous strain sequence set, and a safety warning is given based on the calculation result, specifically as follows: The strain gradient distribution value of each time point is obtained by spatial differentiation calculation from the synchronous strain sequence set; An abnormal sequence in the synchronous strain sequence set is identified based on the strain gradient distribution value, and a damage degree value of the structure is predicted based on the abnormal sequence; When the damage degree value exceeds a preset threshold, a safety warning signal is triggered, and a warning position and severity of the tower structure are output.

8. A computer readable storage medium storing a computer program, characterized in that, The computer program is executed by the processor to implement the tower structure inspection and warning method according to any one of claims 1 to 7.

Images