A method and system for detecting the hardening of shotcrete

Abstract

The present invention proposes a method and system for detecting the hardening of shotcrete, which relates to the field of material testing technology. The method comprises: determining detection requirements based on the physical properties of shotcrete and establishing a progressive multi-frequency excitation sequence; after each multi-frequency excitation, real-time monitoring of the echo signal within the concrete and dynamic adjustment of the excitation parameters based on the echo signal quality; preprocessing the collected echo signals; extracting multi-dimensional feature parameters from the preprocessed echo signals, including time domain features, frequency domain features, and energy domain features, to form a representative feature vector; eliminating redundant features based on the feature vector and calculating a hardening score using the remaining features; and outputting detection results and warning information based on the hardening score. The present invention can accurately assess the hardening state of shotcrete, thereby providing reliable technical support for engineering construction quality control.

Description

Technical Field

[0001] The present invention relates to the technical field of material detection, and in particular to a hardening detection method and system for shotcrete. Background Art

[0002] Shotcrete, a construction process in which concrete is sprayed onto a base surface using compressed air, is widely used in tunneling, slope support, underground engineering, and other fields. In practice, the hardening process of shotcrete is directly related to the quality and safety of the project. Inadequate or uneven hardening can lead to structural weakness, cracking, and spalling, seriously threatening the project's safety and service life.

[0003] Traditional methods for testing shotcrete hardening include rebound testing, penetration testing, and ultrasonic testing. Ultrasonic testing, due to its non-destructive nature, has gradually become the mainstream method. This method assesses the hardening state of concrete by emitting ultrasonic signals and receiving echoes based on the propagation characteristics of sound waves in concrete.

[0004] However, existing detection technologies still have some shortcomings, mainly manifested in low detection accuracy, poor anti-interference ability, and inability to achieve comprehensive assessments. Especially in complex construction environments, where there are numerous external interference factors, the reliability and accuracy of detection results are difficult to guarantee, and they cannot meet the actual needs of the project. Summary of the Invention

[0005] In view of this, the present invention proposes a method and system for detecting the hardening of shotcrete. By establishing a progressive multi-frequency excitation sequence, real-time monitoring of echo signal quality, multi-dimensional feature extraction and analysis, and environmental factor compensation, etc., accurate assessment of the hardening state of shotcrete can be achieved, thereby providing reliable technical support for engineering construction quality control.

[0006] The technical solution of the present invention is achieved as follows:

[0007] In one aspect, the present invention provides a method for detecting hardening of shotcrete, comprising:

[0008] S1. Determine the detection requirements based on the physical properties of shotcrete, establish a progressive multi-frequency excitation sequence, and transmit low-frequency reference signals, medium-frequency detection signals, and high-frequency adjustment signals in chronological order;

[0009] S2. After completing each multi-frequency excitation, the echo signal inside the concrete is monitored in real time, and the excitation parameters are dynamically adjusted according to the echo signal quality;

[0010] S3. Preprocess the collected echo signals, including bandpass filtering and denoising operations, to remove distortion caused by the construction environment and external high-frequency interference;

[0011] S4. Extracting multi-dimensional feature parameters from the pre-processed echo signal, including time domain features, frequency domain features, and energy domain features, to form a representative feature vector;

[0012] S5. Eliminate redundant features according to the feature vector, and calculate a hardening score using the remaining features;

[0013] S6. Output the test results and warning information according to the hardening score.

[0014] Based on the above scheme, preferably, in the progressive multi-frequency excitation sequence: the frequency range of the low-frequency reference signal is 20-40kHz; the frequency range of the intermediate-frequency detection signal is 40-80kHz; and the frequency range of the high-frequency adjustment signal is 80-120kHz.

[0015] Based on the above solution, preferably, the process of establishing the progressive multi-frequency excitation sequence includes:

[0016] Analyze the actual needs of the shotcrete construction site and determine the initial frequencies of the low-frequency reference signal, medium-frequency detection signal, and high-frequency adjustment signal based on the concrete thickness and material acoustic impedance characteristics;

[0017] By setting the timing control module, the low-frequency reference signal is first transmitted to obtain the deep layer information of the concrete, then the medium-frequency detection signal is used to collect the middle layer information, and finally the high-frequency adjustment signal is used to correct the surface layer detection;

[0018] The three frequency band signals are executed in sequence with an interval of no less than 50μs, so that the signals can achieve progressive coverage on the time axis;

[0019] Multi-frequency welding points are established and the signal waveforms are preliminarily debugged to form a set of scalable multi-frequency excitation sequences.

[0020] Based on the above solution, preferably, the dynamic adjustment of the excitation parameters adopts the following formula:

[0021]

[0022] Where, and are the frequencies before and after correction, is the excitation sensitivity coefficient, is the incremental value of the current excitation echo signal-to-noise ratio, is the average signal-to-noise ratio of the recent excitation echo, is the delay penalty coefficient, is the standard deviation of the echo arrival time.

[0023] Based on the above solution, preferably, in step S3, when preprocessing the collected echo signal, the bandpass filtering adopts a Chebyshev bandpass filter with adaptive bandwidth, and the bandwidth range is automatically adjusted with the frequency band; the denoising operation adopts a wavelet threshold algorithm, and the threshold function is:

[0024]

[0025] Where, is the j-th layer threshold; is the noise standard deviation; N is the signal length; is the j-th layer wavelet coefficient, is the regulating factor.

[0026] Based on the above scheme, preferably, the time domain features include signal waveform attenuation rate, initial peak amplitude and first wave arrival time; the frequency domain features include main frequency peak, bandwidth and spectrum energy center; the energy domain features include total energy value, energy attenuation rate, frequency band energy ratio and energy concentration;

[0027] After multi-dimensional feature parameter extraction of the preprocessed echo signal, the three types of features are merged into a representative feature vector. The feature vector contains at least 9 dimensional parameters and is used to comprehensively characterize the hardening state of shotcrete within different depth ranges.

[0028] Based on the above solution, preferably, step S5 includes:

[0029] S51, constructing an initial feature set based on the feature vector, performing correlation analysis on parameters of each dimension, and calculating a correlation coefficient matrix between dimensions of the feature vector;

[0030] S52, setting a correlation threshold, eliminating redundant features with correlations higher than the correlation threshold, and obtaining optimized features;

[0031] S53, collecting environmental parameters, including temperature, humidity, and vibration intensity, and using an environmental compensation algorithm to correct the optimized features based on the environmental parameters to obtain effective features;

[0032] S54. Calculate the final hardening score based on the valid features.

[0033] Based on the above solution, preferably, in step S53, the obtained optimization feature is recorded as , i represents the characteristic dimension, and in the collected environmental parameters, 、 and Represents the deviation of temperature, humidity and vibration intensity from the reference value respectively, then the effective feature obtained after the correction of each dimension optimization feature is Calculated by the following formula:

[0034]

[0035] Where, is the adjustment coefficient, represents the hyperbolic tangent function, 、 、 are the compensation coefficients for temperature, humidity and vibration intensity, is the control parameter, is the baseline environmental factor;

[0036] In step S54, the stage segmentation threshold is set , the detection time t of shotcrete is divided into the early stage and later stages , using piecewise function to calculate the final hardening score :

[0037] when hour, The calculation formula is:

[0038]

[0039] when hour, The calculation formula is:

[0040]

[0041] in, For the initial score; 、 is the size and speed of the hardening increment over time in the early stage; is the weighted comprehensive value of all features, n is the number of effective features, is the weight of the i-th effective feature; for Hardening score at the moment; 、 is the size and speed of the hardening increment over time in the later stage, and 、 The value is less than the corresponding 、 .

[0042] Based on the above solution, preferably, the warning information output in step S6 is divided into three warning levels according to the hardening score:

[0043] When the hardening score is greater than or equal to 85, it displays a green safety status and no additional treatment is required;

[0044] When the hardening score is between 60-85, a yellow warning is triggered and repeated testing or appropriate extension of the curing time is required;

[0045] When the hardening score is below 60, a red alert is triggered and a plan for construction reinforcement or accelerated hardening is generated.

[0046] On the other hand, the present invention further provides a shotcrete hardening detection system, the system being configured to execute any of the above methods, the system comprising:

[0047] Controller, used to control the operation of the overall system;

[0048] a signal generating module, electrically connected to the controller, for generating a progressive multi-frequency excitation sequence, wherein the progressive multi-frequency excitation sequence includes a low-frequency reference signal, an intermediate-frequency detection signal, and a high-frequency adjustment signal;

[0049] an ultrasonic transducer, electrically connected to the signal generating module, for transmitting the progressive multi-frequency excitation sequence and receiving echo signals;

[0050] A signal acquisition module, electrically connected to the ultrasonic transducer, for collecting echo signals in real time and monitoring signal quality;

[0051] A signal processing module, electrically connected to the signal acquisition module, for performing bandpass filtering and wavelet threshold denoising preprocessing on the collected echo signal;

[0052] a feature extraction module, electrically connected to the signal processing module, for extracting time domain features, frequency domain features, and energy domain features of the echo signal and forming a feature vector;

[0053] Environmental sensor group, used to collect environmental parameters such as temperature, humidity and vibration intensity;

[0054] a score calculation module, electrically connected to the feature extraction module and the environmental sensor group, for eliminating redundant features according to the feature vector and calculating a hardening score in combination with environmental parameters;

[0055] The display output module is electrically connected to the score calculation module and is used to display the detection results and output warning information according to the hardening score.

[0056] The present invention has the following beneficial effects compared to the prior art:

[0057] The present invention achieves comprehensive and high-precision detection of the hardening state of shotcrete by organically combining technical means such as establishing a progressive multi-frequency excitation sequence, real-time monitoring of echo signal quality and dynamic adjustment of excitation parameters, multi-dimensional feature extraction, and environmental compensation, significantly improving the reliability and accuracy of the detection results.

[0058] A progressive multi-frequency excitation sequence is used, which is excited in sequence according to the low frequency, medium frequency and high frequency, and a reasonable time interval is set, so that the detection signal can effectively cover different depth layers of concrete, thereby obtaining more complete hardening status information;

[0059] By monitoring the echo signal quality in real time and using an adaptive algorithm to dynamically adjust the excitation parameters, the detection parameters can be automatically optimized according to the actual detection environment, effectively improving the adaptability of the detection system and the signal acquisition quality;

[0060] Adopting environmental compensation algorithm, taking into account the influence of environmental factors such as temperature, humidity and vibration intensity, it effectively corrects the characteristics, reduces the interference of environmental changes on the detection results, and improves the accuracy of detection;

[0061] The hardening score is calculated based on piecewise functions and a three-level early warning mechanism is set up to achieve quantitative assessment of the hardening status of concrete and timely early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0063] Figure 1 is a flow chart of the method of the present invention;

[0064] Figure 2 It is a detection flow chart of the present invention;

[0065] Figure 3 This is a system framework diagram of the present invention. DETAILED DESCRIPTION

[0066] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0067] like Figure 1 As shown, first, the present invention provides a method for detecting hardening of shotcrete, comprising:

[0068] S1. Determine the detection requirements based on the physical properties of shotcrete, establish a progressive multi-frequency excitation sequence, and transmit low-frequency reference signals, medium-frequency detection signals, and high-frequency adjustment signals in chronological order;

[0069] S2. After completing each multi-frequency excitation, the echo signal inside the concrete is monitored in real time, and the excitation parameters are dynamically adjusted according to the echo signal quality;

[0070] S3. Preprocess the collected echo signals, including bandpass filtering and denoising operations, to remove distortion caused by the construction environment and external high-frequency interference;

[0071] S4. Extracting multi-dimensional feature parameters from the pre-processed echo signal, including time domain features, frequency domain features, and energy domain features, to form a representative feature vector;

[0072] S5. Eliminate redundant features according to the feature vector, and calculate a hardening score using the remaining features;

[0073] S6. Output the test results and warning information according to the hardening score.

[0074] like Figure 2 As shown, the present invention first determines detection requirements based on the physical properties of shotcrete and actual construction site needs. A progressive multi-frequency excitation sequence is then established. This sequence utilizes an ordered combination of a low-frequency reference signal (20-40 kHz), a medium-frequency detection signal (40-80 kHz), and a high-frequency modulation signal (80-120 kHz), transmitted sequentially in a strict time sequence, to comprehensively detect concrete at different depths. During signal acquisition, the system monitors echo signal quality in real time and dynamically adjusts excitation parameters based on signal-to-noise ratio and delay characteristics to ensure high-quality detection data. The collected echo signals are then filtered using an adaptive-bandwidth Chebyshev bandpass filter and pre-processed with a wavelet threshold algorithm for denoising, effectively eliminating signal distortion caused by the construction environment and external high-frequency interference. Furthermore, the system extracts feature parameters from the time, frequency, and energy domains, including key indicators such as signal waveform attenuation rate, bandwidth, and energy concentration. This generates a feature vector with at least nine dimensions that comprehensively characterizes the hardened state of the concrete. Finally, the characteristic vector is optimized and compensated based on environmental parameters such as temperature, humidity and vibration intensity. The hardening score is calculated using a piecewise function. The score is divided into three levels: green safety, yellow warning and red alert, and the corresponding detection results and warning information are output.

[0075] Specifically, in one embodiment of the present invention, a shotcrete hardening detection system is first established, which mainly includes a controller, a signal generating module, an ultrasonic transducer, a signal acquisition module, a signal processing module, a feature extraction module, an environmental sensor group, a scoring calculation module and a display output module.

[0076] During testing, the ultrasonic transducer is first fixed to the surface of the shotcrete to be tested, ensuring close contact between the transducer and the concrete surface. If necessary, a coupling agent can be used to improve the transmission of sound waves. An environmental sensor group is arranged around the detection area to monitor environmental parameters in real time. After the system is powered on, the controller first performs a self-test and parameter initialization. Then, the signal generation module generates a progressive multi-frequency excitation sequence: first, a low-frequency reference signal of 20-40kHz is emitted to obtain information about the deep layer of concrete; after an interval of 50μs, a medium-frequency detection signal of 40-80kHz is emitted to collect information about the middle layer; and after an interval of another 50μs, a high-frequency adjustment signal of 80-120kHz is emitted to correct the surface detection results.

[0077] During signal acquisition, the system monitors the echo signal's signal-to-noise ratio and delay characteristics in real time, dynamically adjusting excitation parameters based on actual conditions. The acquired echo signal is processed using an adaptive bandwidth Chebyshev bandpass filter and wavelet threshold denoising to effectively eliminate environmental interference. Subsequently, the feature extraction module extracts characteristic parameters from the time, frequency, and energy domains, including key indicators such as waveform attenuation rate, bandwidth, and energy concentration, to form a feature vector.

[0078] The scoring module first performs correlation analysis on the feature vectors to eliminate redundant features. It then applies environmental compensation to these features, combining them with temperature, humidity, and vibration data collected by environmental sensors. The system uses a piecewise function to calculate the final hardening score based on the test duration and displays the results in real time on the display output module. A hardening score of 85 or greater indicates a green safety status, while a score between 60 and 85 triggers a yellow warning. A score below 60 triggers a red alert, with appropriate action taken.

[0079] The detection system in this embodiment has a reasonable structure and is easy to operate. It can achieve accurate evaluation of the hardening state of shotcrete and provide reliable technical support for engineering construction quality control.

[0080] Specifically, in one embodiment of the present invention, step S1 includes:

[0081] First, a detailed analysis of the actual conditions at the construction site is required, including the thickness of the shotcrete, construction conditions, and the acoustic impedance characteristics of its materials, in order to determine the overall testing requirements.

[0082] Based on this requirement, a three-stage excitation signal was designed: a low-frequency reference signal of 20-40kHz, used to penetrate deep concrete to obtain reference data; a medium-frequency detection signal of 40-80kHz, used to collect middle-layer information; and a high-frequency adjustment signal of 80-120kHz, used to correct and refine surface detection results.

[0083] The three frequency bands are programmed and controlled according to a preset time control module. Following strict timing requirements, the low-frequency reference signal is transmitted first, followed by the medium-frequency detection signal, and finally the high-frequency adjustment signal. The interval between each signal is guaranteed to be no less than 50μs, achieving progressive signal coverage along the time axis. This process also requires setting up and establishing multi-frequency welding points, and optimizing the signal waveforms obtained from preliminary debugging to form a multi-frequency excitation sequence that is highly scalable and adaptable to the complex conditions of the construction site, ensuring the accuracy and stability of subsequent signal acquisition and processing.

[0084] Specifically, in one embodiment of the present invention, step S2 includes:

[0085] After completing a multi-frequency excitation, the system first collects the echo signal inside the concrete in real time through the ultrasonic transducer and quantitatively evaluates the signal quality, mainly examining the signal-to-noise ratio (SNR) and the change in the echo arrival time. Specifically, the system calculates the increment of the current excitation echo signal-to-noise ratio ( ) and the average signal-to-noise ratio of the recent stimulated echo ( ), and calculate the standard deviation of the echo arrival time ( ). Next, adjust the excitation parameter formula dynamically:

[0086]

[0087] The system sets the current excitation frequency According to the real-time monitored signal indicators, corrections are made to obtain a new excitation frequency .in, is the excitation sensitivity coefficient, which is used to amplify or weaken the effect of signal-to-noise ratio changes on frequency adjustment; The delay penalty coefficient is used to penalize fluctuations in echo arrival time, ensuring more conservative frequency adjustment when the delay varies significantly.

[0088] The specific adjustment process is as follows:

[0089] when When it is a positive value, it indicates that the current signal quality has improved and the system will appropriately increase the excitation frequency; when When it is a negative value, it indicates that the signal quality is degraded and the system will reduce the excitation frequency accordingly.

[0090] pass and The ratio of δ and the excitation sensitivity coefficient δ determine the basic amplitude of the frequency adjustment.

[0091] Taking into account the fluctuation of echo arrival time, a delay penalty term is introduced .when When it is large, it indicates that the signal propagation is unstable. At this time, the frequency adjustment amplitude will be reduced through the exponential decay term to ensure the reliability of the detection.

[0092] Finally, the new excitation frequency is calculated , and apply it to the next excitation. It is important to note that the system will ensure that the adjusted frequency is still within the allowed range of each frequency band (low frequency 20-40kHz, medium frequency 40-80kHz, high frequency 80-120kHz).

[0093] It should be noted that when the system is dynamically adjusted, the signal quality indicators include the signal-to-noise ratio (SNR) and the stability of the echo arrival time. That is, the conditions for meeting the signal quality requirements include: 1. The SNR of the currently collected echo signal reaches the preset minimum value; at the same time, the system monitors the and The ratio of the echo signal arrival time is kept in a relatively stable and reasonable range; 2. The standard deviation of the echo signal arrival time If the delay characteristic of the echo signal is below the preset threshold, it indicates that the delay characteristics of the echo signal are stable. When the echo signal has a high signal-to-noise ratio and changes smoothly (low delay fluctuation), the system determines that the signal quality meets the requirements and proceeds to the next step of signal preprocessing.

[0094] Through this dynamic adjustment mechanism, the system can adaptively optimize the excitation parameters according to the actual detection environment and changes in the concrete state to ensure high-quality detection data.

[0095] Specifically, in one embodiment of the present invention, step S3 preprocesses the collected echo signals to eliminate signal distortion caused by construction environment interference and external high-frequency noise, thereby ensuring the accuracy of the subsequent feature extraction data. The preprocessing process includes two main steps: bandpass filtering and denoising.

[0096] First, the collected echo signal is bandpass filtered. This embodiment uses an adaptive-bandwidth Chebyshev bandpass filter, which exhibits excellent passband flatness and stopband attenuation. The filter's bandwidth automatically adjusts based on the frequency band of the signal: for low-frequency reference signals (20-40kHz), the bandwidth is set to ±5kHz; for intermediate-frequency detection signals (40-80kHz), the bandwidth is set to ±8kHz; and for high-frequency modulation signals (80-120kHz), the bandwidth is set to ±10kHz. This adaptive bandwidth setting effectively preserves the effective components of signals in each frequency band while filtering out out-of-band interference.

[0097] Next, perform wavelet threshold denoising on the filtered signal. The specific steps are as follows:

[0098] Select an appropriate wavelet basis function (such as db4 wavelet) and perform three-layer wavelet decomposition on the signal to obtain the wavelet coefficients of each layer. ; Calculate the threshold of each layer according to the threshold function:

[0099]

[0100] Where, is the j-th layer threshold; is the noise standard deviation, estimated by dividing the median of the first layer wavelet coefficients by 0.6745; N is the signal length; is the j-th layer wavelet coefficient, is the regulating factor, which is determined through experimental optimization.

[0101] Perform soft threshold processing on the wavelet coefficients of each layer:

[0102] when When , set the coefficient to 0; when When , the coefficients are shrunk; the processed wavelet coefficients are used to reconstruct the signal to obtain the denoised signal.

[0103] Through the dual processing of bandpass filtering and wavelet threshold denoising, the signal distortion caused by external interference sources such as random noise in the construction environment, industrial electromagnetic interference, and mechanical vibration can be effectively filtered out, while the effective characteristics of the original signal are well maintained, providing a high-quality signal foundation for subsequent feature extraction.

[0104] Specifically, in one embodiment of the present invention, step S4 mainly uses a multi-dimensional feature parameter extraction algorithm to comprehensively analyze the echo signal after bandpass filtering and wavelet threshold denoising, thereby obtaining a feature vector that comprehensively represents the hardening state of the shotcrete. The specific implementation process is as follows:

[0105] 1. Time Domain Feature Extraction

[0106] The system first analyzes the denoised signal in the time domain and extracts parameters such as the signal waveform attenuation rate, initial peak amplitude, and first wave arrival time.

[0107] Extract the signal waveform attenuation rate: By fitting the signal envelope curve, solve the exponential attenuation coefficient and quantify the energy attenuation of the sound wave when propagating in concrete; Determine the initial peak amplitude: Locate the first obvious peak in the signal, reflecting the acoustic characteristics of the concrete surface and near-surface layer; Measure the first wave arrival time: Record the moment when the first waveform exceeding the noise level appears, thereby reflecting the sound speed and propagation characteristics in the concrete.

[0108] 2. Frequency Domain Feature Extraction

[0109] Frequency domain analysis methods such as Fast Fourier Transform (FFT) are used to perform spectrum analysis on the signal and extract parameters such as the main frequency peak, bandwidth, and spectrum energy center.

[0110] Calculate the peak of the main frequency: After transforming the signal using the Fast Fourier Transform (FFT), find the frequency component with the largest amplitude in the spectrum; Determine the bandwidth: Using -3dB as the judgment standard, calculate the half-value width of the spectrum near the peak of the main frequency to describe the distribution range of the signal's frequency domain energy; Find the center of the spectrum energy: Calculate the weighted average frequency of the spectrum to characterize the area where the main energy of the signal is concentrated.

[0111] 3. Energy Domain Feature Extraction

[0112] In the energy domain, the system calculates the overall energy of the signal and further extracts the total energy value, energy attenuation rate, frequency band energy ratio and energy concentration.

[0113] Calculate the total energy value: integrate the energy of the signal to obtain the overall energy of the echo signal; evaluate the energy decay rate: obtain the energy decay parameter over time by comparing the energy values of the signal in different time periods; analyze the frequency band energy ratio: count the energy in the low-frequency, medium-frequency and high-frequency bands respectively, and calculate the ratio between them to reflect the energy distribution of information at different depths; measure the energy concentration: analyze the distribution and concentration of the signal energy on the time axis to reveal changes in internal physical properties.

[0114] 4. Construction of feature vector

[0115] After extracting features from the time, frequency, and energy domains, these three categories of features are combined to form a representative feature vector. This feature vector contains at least nine dimensional parameters, corresponding to: time domain: signal waveform decay rate, initial peak amplitude, and first wave arrival time; frequency domain: dominant frequency peak, bandwidth, and spectral energy center; energy domain: total energy value, energy decay rate, frequency band energy ratio, or energy concentration. The resulting feature vector comprehensively reflects the hardening state of shotcrete at different depths, providing efficient and accurate basic data for subsequent feature optimization and scoring calculations.

[0116] Specifically, in one embodiment of the present invention, step S5 includes:

[0117] S51 . Construct an initial feature set based on the feature vector, perform correlation analysis on parameters of each dimension, and calculate a correlation coefficient matrix between dimensions of the feature vector.

[0118] The multi-dimensional features extracted from the echo signal after preprocessing in step S4 include three types of parameters: (1) time domain features: signal waveform attenuation rate, initial peak amplitude, and first wave arrival time; (2) frequency domain features: main frequency peak, bandwidth, and spectrum energy center; (3) energy domain features: total energy value, energy attenuation rate, frequency band energy ratio, and energy concentration. These features are arranged in a certain order to form the initial feature vector , where n is at least 9.

[0119] To analyze the correlation between features, the Pearson correlation coefficient between each two features was calculated. Defined as:

[0120]

[0121] Where E represents the expected value, and Characteristics After all features are calculated, an n×n correlation coefficient matrix R is formed, and its (i, j)th element is For example, if the initial feature vector contains 9 dimensions, then R is a 9×9 matrix where all diagonal elements are 1 and the off-diagonal elements reflect the strength of the linear correlation between features.

[0122] S52: Set a correlation threshold, remove redundant features with correlations higher than the correlation threshold, and obtain optimized features.

[0123] According to experimental data and domain expert experience, a predetermined correlation threshold (such as 0.85) is set. Exceeding this threshold (i.e. ), they are considered to have high redundancy.

[0124] For any pair of redundant features, the following principles apply:

[0125] Analyze the information content of each feature (for example, determine the variance of the feature, the signal-to-noise ratio, or the discriminative ability in historical data).

[0126] Retain the features with large information content and more representative in the hardening state assessment; remove the redundant parts to obtain the optimized feature vector, which is recorded as Where m<n.

[0127] S53. Collect environmental parameters, including temperature, humidity, and vibration intensity, and use an environmental compensation algorithm to correct the optimized features based on the environmental parameters to obtain effective features.

[0128] The environmental sensor group is used to collect the environmental parameters of the current construction site in real time. Specifically, it includes:

[0129] temperature : Use a digital temperature sensor to measure the current temperature, compare it with the reference temperature (such as 20°C), and calculate the deviation ;humidity : Use the humidity sensor to measure the current humidity and compare it with the reference humidity (such as 65%RH) to obtain ; Vibration intensity : The current vibration intensity is collected by the vibration sensor installed on site, and the deviation from the preset benchmark (such as the vibration value in the laboratory under stable conditions) is calculated. . Then the effective features obtained after the optimization features of each dimension are corrected are Calculated by the following formula:

[0130]

[0131] Where, is the adjustment coefficient, which is used to control the impact of environmental compensation; represents the hyperbolic tangent function, which is used to limit the compensation value to the range of 0-1; 、 、 are the compensation coefficients for temperature, humidity and vibration intensity, respectively, which are determined by fitting the test data; It is a control parameter used to amplify the nonlinear effect of the deviation; It is the benchmark environmental factor, which is used for normalization processing to ensure the consistency of the dimensions of each compensation item.

[0132] After environmental compensation, the features of each dimension are corrected to effective features ,These features can more accurately reflect the real hardening state of concrete and ,reduce the errors introduced by environmental fluctuations.

[0133] S54. Calculate the final hardening score based on the valid features.

[0134] The effective features after environmental compensation are combined into vectors . Assign each feature a weight (Weight determination can be based on expert experience, data statistics, or machine learning model training results) and calculate the overall comprehensive value S:

[0135]

[0136] Here, S represents the weighted comprehensive index of all effective features, reflecting the overall physical properties of hardened concrete.

[0137] According to the detection time t, the concrete hardening process is divided into the early stage ( ) and later stages ( ),in is the preset stage segmentation threshold (for example, 4 hours).

[0138] when hour, The calculation formula is:

[0139]

[0140] when hour, The calculation formula is:

[0141]

[0142] in, is the initial score (5-10 is acceptable); 、 is the size and speed of the hardening increment over time in the early stage; for Hardening score at the moment; 、 is the size and speed of the hardening increment over time in the later stage, and 、 The value is less than the corresponding 、 .

[0143] This embodiment first constructs an initial feature vector and calculates the Pearson correlation coefficient between each dimension to form a correlation coefficient matrix (e.g., a 9×9 matrix) to quantify redundant information between features. Highly correlated redundant features are then eliminated based on a preset threshold to obtain optimized features. Next, environmental sensors are used to collect parameters such as temperature, humidity, and vibration intensity in real time, and the optimized features are corrected using a preset compensation formula to eliminate environmental influences. Finally, a weighted summation of each valid feature is performed, and a piecewise logarithmic function is used to calculate the combined features based on the detection duration. This comprehensive feature is converted into a shotcrete hardening score, which provides a basis for outputting warning information. Through this series of processes, the hardening score not only accurately reflects the internal physical state of the concrete but also dynamically adapts to changes in the construction site environment, improving the reliability and applicability of the detection system.

[0144] Specifically, in one embodiment of the present invention, step S6 includes:

[0145] Warning information is divided into three warning levels according to the hardening score:

[0146] When the hardening score is greater than or equal to 85, the green safety status is displayed, indicating that the shotcrete has hardened sufficiently and does not require additional treatment;

[0147] When the hardening score is between 60-85, a yellow warning is triggered, indicating that there is a certain risk in the test result. The user should repeat the test or extend the curing time appropriately.

[0148] When the hardening score is lower than 60, a red alarm is triggered, indicating that the shotcrete is not hardened enough and construction reinforcement or initiation measures are taken to accelerate hardening.

[0149] Obtained in the score calculation module Afterwards, the corresponding warning information can be automatically selected according to the scoring range through the established rule library or preset program.

[0150] In the inspection system, the display output module receives the warning calculation results and displays the hardening score, grading results, and prompts in real time via a display screen or communication interface. This can be implemented using a graphical interface, directly presenting the score value, color code (e.g., green, yellow, red), and text description to the operator. The system can also upload the results to a central monitoring platform or send warning text messages or emails to ensure timely response to on-site issues.

[0151] Through the above steps, the system can monitor the hardening status of shotcrete in real time under actual construction scenarios, provide timely feedback on abnormal situations, and ensure the safety and quality of project construction.

[0152] In addition, if Figure 3As shown, the present invention also provides a shotcrete hardening detection system, the system is used to perform any of the above methods, the system comprising:

[0153] The controller controls the overall system operation. The system utilizes an industrial-grade embedded processor with a main frequency of 2GHz and 16GB of built-in storage. As the "brain" of the entire system, the controller is responsible for overall control and coordination. Following the pre-set concrete testing plan, it directs the orderly operation of subsequent modules, including signal generation, data acquisition, signal processing, feature extraction, and score calculation. It also manages communication with the display output module and the remote monitoring platform.

[0154] The signal generation module, electrically connected to the controller, generates a progressive multi-frequency excitation sequence consisting of a low-frequency reference signal, a medium-frequency detection signal, and a high-frequency adjustment signal. The signal generation module utilizes a programmable signal generator with a frequency range of 20-120 kHz, capable of switching between low-frequency, medium-frequency, and high-frequency signal outputs. The low-frequency signal (20-40 kHz) is used to detect deep concrete layers, the medium-frequency signal (40-80 kHz) collects information from mid-layers, and the high-frequency signal (80-120 kHz) is used to correct surface detection. The module features adjustable output power, allowing the excitation signal to be dynamically adjusted based on the actual echo signal-to-noise ratio and echo arrival time.

[0155] An ultrasonic transducer, electrically connected to the signal generation module, is used to transmit the progressive multi-frequency excitation sequence and receive echo signals. Made of piezoelectric ceramic, the ultrasonic transducer exhibits broadband response characteristics. This module not only converts the electrical signals generated by the signal generation module into ultrasonic signals for transmission, but also exhibits uniform transmission and reception sensitivity, ensuring accurate capture of echo signals returned after propagation through the concrete.

[0156] The signal acquisition module, electrically connected to the ultrasonic transducer, collects echo signals in real time and monitors signal quality. It utilizes a 16-bit high-speed data acquisition card with a sampling rate of at least 1 MHz, ensuring high-speed acquisition of complex echo signals from within the concrete. This module monitors echo signal quality in real time. Through high-speed data acquisition and monitoring, it provides high-fidelity signal data to the signal processing module, reducing errors caused by environmental interference on site.

[0157] The signal processing module, electrically connected to the signal acquisition module, performs bandpass filtering and wavelet threshold denoising preprocessing on the collected echo signals. The collected echo signals first pass through the signal processing module, which preprocesses the signals using an adaptive-bandwidth Chebyshev bandpass filter, automatically adjusting the bandwidth to match the signal characteristics of different frequency bands. Furthermore, the module uses a wavelet threshold algorithm to perform denoising operations to eliminate signal distortion caused by the construction environment and external high-frequency interference.

[0158] A feature extraction module, electrically connected to the signal processing module, extracts the time, frequency, and energy domain features of the echo signal and generates a feature vector. The processed data enters the feature extraction module, where it extracts multi-dimensional feature parameters from the echo signal. These parameters include time domain features (such as signal waveform decay rate, initial peak amplitude, and first wave arrival time), frequency domain features (such as dominant frequency peak, bandwidth, and spectral energy center), and energy domain features (such as total energy value, energy decay rate, frequency band energy ratio, and energy concentration). This processing generates a comprehensive feature vector containing at least nine dimensional parameters, reflecting the hardening state of concrete at different depths.

[0159] The environmental sensor suite collects environmental parameters such as temperature, humidity, and vibration intensity. The system is equipped with a temperature sensor, a humidity sensor, and a triaxial vibration sensor. The temperature sensor has a measurement range of -20°C to 60°C with an accuracy of ±0.5°C; the humidity sensor has a measurement range of 0-100%RH with an accuracy of ±2%RH; and the triaxial vibration sensor has a measurement range of ±16g and a bandwidth of DC to 2kHz. This sensor suite collects construction site environmental parameters in real time, and this data is used to compensate for environmental factors in the feature scoring calculation.

[0160] The scoring calculation module is electrically connected to the feature extraction module and the environmental sensor group, respectively, and is used to eliminate redundant features based on the feature vector and calculate the hardening score in combination with environmental parameters. The scoring calculation module receives the feature vector output by the feature extraction module and, in combination with the environmental data collected by the environmental sensor group, constructs a redundant feature elimination and environmental compensation process. First, the module performs a correlation analysis on each dimension of the feature vector using a correlation coefficient matrix, eliminating redundant features above a preset threshold to obtain optimized features. Subsequently, the optimized features are corrected according to a preset environmental compensation formula, and ultimately the effective features of each dimension are calculated. The module further calculates a weighted comprehensive value based on the weight of each effective feature and uses a piecewise function to calculate the final hardening score in combination with the detection time, thereby accurately evaluating the hardening state of the concrete.

[0161] The display and output module, electrically connected to the scoring calculation module, displays the test results and outputs warning information based on the hardening score. The display and output module displays the hardening score and warning information output by the scoring calculation module graphically, numerically, and visually. Based on the scoring results, the system categorizes the status as green safety (hardening score ≥ 85), yellow warning (between 60-85), and red alert (<60). The system then promptly transmits specific actions to on-site operators and the remote monitoring platform to assist in decision-making and the deployment of subsequent measures.

[0162] In one embodiment, the system workflow is as follows: First, the controller transmits a progressive multi-frequency excitation sequence through the signal generation module based on the sprayed concrete inspection requirements. The ultrasonic transducer converts these excitation signals into ultrasonic waves and transmits them into the concrete, where the echo signals are then received by the same transducer. The signal acquisition module uses a high-speed acquisition card to acquire the echo signals at high speed and monitor the signal quality. The acquired signals are sent to the signal processing module for bandpass filtering and wavelet threshold denoising to obtain pure signal data. The processed data enters the feature extraction module, which extracts key features in the time, frequency, and energy domains to form a multidimensional feature vector representing the hardening state of the concrete. Simultaneously, the environmental sensor group collects on-site temperature, humidity, and vibration data in real time. This data, along with the optimized features, is input into the score calculation module, which performs redundant feature elimination, environmental compensation, and final hardening score calculation. Finally, the display output module outputs the inspection results and warning information based on the signal score, which is displayed in real time on the on-site control screen or remote monitoring terminal. A warning of abnormal conditions is also issued through audible and visual alarms to ensure construction safety.

[0163] 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 in the scope of protection of the present invention.

Claims

1. A method for detecting hardening of shotcrete, characterized in that: include: S1. Determine the detection requirements based on the physical properties of shotcrete, establish a progressive multi-frequency excitation sequence, and transmit low-frequency reference signals, medium-frequency detection signals, and high-frequency adjustment signals in chronological order; S2. After completing each multi-frequency excitation, the echo signal inside the concrete is monitored in real time, and the excitation parameters are dynamically adjusted according to the echo signal quality; S3. Preprocess the collected echo signals, including bandpass filtering and denoising operations, to remove distortion caused by the construction environment and external high-frequency interference; S4. Extracting multi-dimensional feature parameters from the pre-processed echo signal, including time domain features, frequency domain features, and energy domain features, to form a representative feature vector; Time domain features include signal waveform attenuation rate, initial peak amplitude and first wave arrival time; frequency domain features include main frequency peak, bandwidth and spectrum energy center; energy domain features include total energy value, energy attenuation rate, frequency band energy ratio and energy concentration; After extracting multi-dimensional feature parameters from the pre-processed echo signal, the three types of features are combined into a representative feature vector. The feature vector contains at least nine dimensional parameters and is used to comprehensively characterize the hardening state of shotcrete at different depths. S5. Eliminate redundant features according to the feature vector, and calculate a hardening score using the remaining features; Step S5 includes: S51, constructing an initial feature set based on the feature vector, performing correlation analysis on parameters of each dimension, and calculating a correlation coefficient matrix between dimensions of the feature vector; S52, setting a correlation threshold, eliminating redundant features with correlations higher than the correlation threshold, and obtaining optimized features; S53, collecting environmental parameters, including temperature, humidity, and vibration intensity, and using an environmental compensation algorithm to correct the optimized features based on the environmental parameters to obtain effective features; S54, calculating the final hardening score based on the valid features; S6. Output the test results and warning information according to the hardening score.

2. A method for detecting hardening of shotcrete according to claim 1, characterized in that: In the progressive multi-frequency excitation sequence: the frequency range of the low-frequency reference signal is 20-40 kHz; the frequency range of the intermediate-frequency detection signal is 40-80 kHz; and the frequency range of the high-frequency adjustment signal is 80-120 kHz.

3. A method for detecting hardening of shotcrete according to claim 1, characterized in that: The process of establishing the progressive multi-frequency excitation sequence includes: Analyze the actual needs of the shotcrete construction site and determine the initial frequencies of the low-frequency reference signal, medium-frequency detection signal, and high-frequency adjustment signal based on the concrete thickness and material acoustic impedance characteristics; By setting the timing control module, the low-frequency reference signal is first transmitted to obtain the deep layer information of the concrete, then the medium-frequency detection signal is used to collect the middle layer information, and finally the high-frequency adjustment signal is used to correct the surface layer detection; The three frequency band signals are executed in sequence with an interval of no less than 50μs, so that the signals can achieve progressive coverage on the time axis; Multi-frequency welding points are established and the signal waveforms are preliminarily debugged to form a set of scalable multi-frequency excitation sequences.

4. A method for detecting hardening of shotcrete according to claim 1, characterized in that: The dynamic adjustment of the excitation parameters adopts the following formula: Where, and are the frequencies before and after correction, is the excitation sensitivity coefficient, is the incremental value of the current excitation echo signal-to-noise ratio, is the average signal-to-noise ratio of the recent excitation echo, is the delay penalty coefficient, is the standard deviation of the echo arrival time.

5. A method for detecting hardening of shotcrete according to claim 1, characterized in that: In step S3, when pre-processing the collected echo signal, a band-pass filter adopts a Chebyshev band-pass filter with adaptive bandwidth, and the bandwidth range is automatically adjusted according to the frequency band; The denoising operation uses the wavelet threshold algorithm, and the threshold function is: Where, is the j-th layer threshold; is the noise standard deviation; N is the signal length; is the j-th layer wavelet coefficient, is the regulating factor.

6. A method for detecting hardening of shotcrete according to claim 1, characterized in that: In step S53, the obtained optimized features are recorded as , i represents the characteristic dimension, and in the collected environmental parameters, 、 and Represents the deviation of temperature, humidity and vibration intensity from the reference value respectively, then the effective feature obtained after the correction of each dimension optimization feature is Calculated by the following formula: Where, is the adjustment coefficient, represents the hyperbolic tangent function, 、 、 are the compensation coefficients for temperature, humidity and vibration intensity, For the control parameters, is the baseline environmental factor; In step S54, the stage segmentation threshold is set , the detection time t of shotcrete is divided into the early stage and later stages , using piecewise function to calculate the final hardening score : when hour, The calculation formula is: when hour, The calculation formula is: in, For the initial score; 、 is the size and speed of the hardening increment over time in the early stage; is the weighted comprehensive value of all features, n is the number of effective features, is the weight of the i-th effective feature; for Hardening score at the moment; 、 is the size and speed of the hardening increment over time in the later stage, and 、 The value is less than the corresponding 、 .

7. A method for detecting hardening of shotcrete according to claim 1, characterized in that: The warning information output in step S6 is divided into three warning levels according to the hardening score: When the hardening score is greater than or equal to 85, it displays a green safety status and no additional treatment is required; When the hardening score is between 60-85, a yellow warning is triggered and repeated testing or appropriate extension of the curing time is required; When the hardening score is below 60, a red alert is triggered and a plan for construction reinforcement or accelerated hardening is generated.

8. A shotcrete hardening detection system, characterized in that: The system is used to execute the method according to any one of claims 1 to 7, and the system includes: Controller, used to control the operation of the overall system; a signal generating module, electrically connected to the controller, for generating a progressive multi-frequency excitation sequence, wherein the progressive multi-frequency excitation sequence includes a low-frequency reference signal, an intermediate-frequency detection signal, and a high-frequency adjustment signal; an ultrasonic transducer, electrically connected to the signal generating module, for transmitting the progressive multi-frequency excitation sequence and receiving echo signals; A signal acquisition module, electrically connected to the ultrasonic transducer, for collecting echo signals in real time and monitoring signal quality; A signal processing module, electrically connected to the signal acquisition module, for performing bandpass filtering and wavelet threshold denoising preprocessing on the collected echo signal; a feature extraction module, electrically connected to the signal processing module, for extracting time domain features, frequency domain features, and energy domain features of the echo signal and forming a feature vector; Environmental sensor group, used to collect environmental parameters such as temperature, humidity and vibration intensity; a score calculation module, electrically connected to the feature extraction module and the environmental sensor group, for eliminating redundant features according to the feature vector and calculating a hardening score in combination with environmental parameters; The display output module is electrically connected to the score calculation module and is used to display the detection results and output warning information according to the hardening score.

Images