Information detection method and system for civil construction monitoring intelligent sensor under signal mismatch

Abstract

The invention provides a civil construction monitoring intelligent sensor information detection method and system under signal mismatch, and the method comprises the steps: carrying out the time-frequency analysis of temperature and strain fundamental frequency components, selecting a wavelet basis function adaptive to the concrete setting and strength increasing process, carrying out the self-adaptive threshold processing of a wavelet coefficient, removing high-frequency noise, and carrying out the detection of the temperature and strain fundamental frequency components. Low-frequency signal characteristics reflecting concrete strength increase are reserved; performing feature extraction on the denoised temperature and strain signals by adopting a singular value decomposition algorithm, detecting internal defects of the concrete through abrupt change of a singular value spectrum, and if crack and hole defects exist, causing strain signal abrupt change by stress concentration; if it is detected that defects exist in the concrete, that is, spatial heterogeneity of temperature and stress distribution exists, temperature and strain monitoring data of multi-point arrangement and a defect detection result serve as boundary conditions and constraints, and a temperature field and a stress field in the whole concrete structure are obtained through numerical simulation.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method and system for detecting information of an intelligent sensor for civil construction monitoring under signal mismatch. Background Art

[0002] During the pouring of large concrete structures, smart sensors face complex signal detection challenges. As the depth of concrete pouring increases, the internal temperature gradient and pressure changes become more significant. The cement hydration reaction releases a large amount of heat, causing the temperature around the sensor to rise sharply, affecting its sensitivity and accuracy. At the same time, the chemical reaction inside the concrete will produce electromagnetic interference, disrupting the sensor's signal acquisition. The high-frequency vibration of the vibrating rod and the impact force of the pumped concrete further aggravate the signal distortion. As the pouring time increases, the concrete gradually solidifies and hardens, and its physical properties continue to change. The coupling state between the sensor and the concrete also changes accordingly, causing signal mismatch. In this dynamically changing environment, how to accurately capture the temperature field distribution, stress-strain state, and microstructural evolution process inside the concrete has become a technical problem that needs to be solved urgently. Especially in deep concrete, the sensor signal attenuation is serious. How to achieve effective information extraction and data fusion under signal mismatch conditions to accurately evaluate the pouring quality and structural safety of concrete is a major challenge currently faced. Summary of the invention

[0003] The present invention provides a method for detecting information of an intelligent sensor for civil construction monitoring under signal mismatch, which mainly includes:

[0004] According to the temperature change characteristics caused by the heat of concrete hydration, the compensation coefficient of the temperature sensor embedded in the concrete is calculated, and the original temperature data collected is drift corrected to obtain the compensated concrete temperature data. At the same time, the embedded vibrating wire strain gauge is used to monitor the deformation of the concrete.

[0005] During the solidification and hardening process of concrete, concrete gradually acquires a certain stiffness and strength, and transmits the internal stress to the strain sensor. The strain sensor is used to collect strain signals, and the harmonic interference introduced by construction vibration is filtered out by setting the threshold of the strain harmonic component to obtain the fundamental frequency component reflecting the real strain state inside the concrete.

[0006] Conduct time-frequency analysis on the fundamental frequency components of temperature and strain, select the wavelet basis function that is suitable for the concrete setting and strength growth process, and remove high-frequency noise through adaptive threshold processing of wavelet coefficients to retain the low-frequency signal characteristics that reflect the strength growth of concrete.

[0007] The singular value decomposition algorithm is used to extract the features of the denoised temperature and strain signals. The internal defects of the concrete are detected through the mutation of the singular value spectrum. If there are cracks or holes, stress concentration will cause mutation of the strain signal.

[0008] If defects are detected inside the concrete, that is, there is spatial non-uniformity in the temperature and stress distribution, the temperature and strain monitoring data arranged at multiple points and the defect detection results are used as boundary conditions and constraints to obtain the temperature field and stress field inside the entire concrete structure through numerical simulation;

[0009] According to the known constitutive relationship between the strength of concrete materials and temperature and stress state, the strength of concrete is closely related to its temperature history and stress state. Appropriate high temperature promotes strength growth, while excessive temperature stress and shrinkage stress lead to cracking and strength reduction. By comparing the temperature field and stress field inside the entire concrete structure obtained by simulation with the constitutive relationship between strength, temperature and stress state, the spatial strength distribution of concrete is evaluated.

[0010] Based on the spatial strength distribution of concrete, the overall strength of concrete is evaluated, and the strength evaluation results are fuzzy processed to obtain the strength membership function. The fuzzy rules of the strength grade are set, and the membership of the concrete strength grade is obtained through fuzzy reasoning. The membership weighted average method is used to defuzzify and obtain the quantitative evaluation result of the concrete strength. If the strength meets the design requirements, the concrete reaches the design strength and enters the structural health monitoring stage. If the strength is insufficient or degradation or internal defects occur, timely warnings will be issued.

[0011] The present invention provides a civil construction monitoring intelligent sensor information detection system under signal mismatch, which mainly includes:

[0012] Data acquisition and correction module, used for temperature sensor data correction and vibrating wire strain gauge monitoring;

[0013] Signal processing module for interference filtering, time-frequency analysis and wavelet denoising;

[0014] Defect detection module, used for singular value decomposition feature extraction and mutation detection;

[0015] Temperature stress field simulation module, used to numerically simulate the temperature stress distribution inside the overall structure;

[0016] The intensity assessment module is used for spatial intensity distribution assessment and fuzzy reasoning intensity level judgment.

[0017] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0018] The present invention discloses a method for detecting information of intelligent sensors for monitoring civil construction under signal mismatch. The method collects original temperature and strain data by embedding a temperature sensor and a vibrating-wire strain gauge inside concrete. Drift correction is performed on the temperature data, and harmonic interference is filtered out on the strain signal to obtain a signal reflecting the real state inside the concrete and detect internal defects of the concrete. Combining multi-point monitoring data and defect detection results, the temperature field and stress field inside the concrete structure are obtained through numerical simulation. Based on the constitutive relationship between the strength of concrete materials and the temperature and stress state, the spatial strength distribution of concrete is estimated. A comprehensive evaluation is performed on the overall strength of concrete to obtain a quantitative evaluation result. The present invention realizes real-time monitoring and evaluation of concrete strength, can timely detect insufficient strength or abnormal conditions and issue warnings, and provides an effective means for quality control and safety management of concrete structures. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 The present invention is a flow chart of a method for detecting information of an intelligent sensor for monitoring civil construction under signal mismatch. DETAILED DESCRIPTION

[0020] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] like Figure 1 In this embodiment, a civil construction monitoring intelligent sensor information detection system under signal mismatch may specifically include:

[0022] Step S101, according to the temperature change characteristics caused by the hydration heat of concrete, the compensation coefficient of the temperature sensor embedded in the concrete is calculated, the drift correction is performed on the collected original temperature data, and the compensated concrete temperature data is obtained; at the same time, the embedded vibrating wire strain gauge is used to monitor the deformation of the concrete.

[0023] The internal temperature data of concrete collected by the multi-point temperature acquisition system is obtained, and the temperature-time curve is drawn according to the temperature data; the temperature change function caused by the heat of concrete hydration is determined by the polynomial regression method, and the compensation coefficient of the temperature sensor is calculated according to the temperature change function; the compensation coefficient is applied to the original temperature data for drift correction, and if the sliding window method is used to smooth the corrected temperature data, the influence of random noise is eliminated; the processed discrete temperature data is interpolated and fitted by the cubic spline interpolation algorithm to obtain a continuous concrete temperature field distribution; the concrete strain data is obtained from the vibrating wire strain gauge, and a temperature-strain relationship model is established in combination with the temperature field distribution data, and the functional relationship between temperature and strain is fitted by linear regression; the thermal expansion coefficient of concrete is calculated according to the compensated concrete temperature data and the true deformation data, and the thermal expansion coefficient is smoothed to obtain a thermal expansion coefficient curve that changes with temperature; the relationship function between the thermal expansion coefficient and temperature is fitted by the least squares method.

[0024] Exemplarily, a multi-point temperature acquisition system is used to continuously monitor the temperature at different depths and positions inside the concrete, and a temperature-time curve is drawn based on the collected temperature data. The temperature change function caused by the heat of concrete hydration is obtained by a polynomial regression method, and the compensation coefficient of the temperature sensor is calculated based on the function. The compensation coefficient is applied to the original temperature data for drift correction, and the corrected temperature data is smoothed by a sliding window method of 5 data points to eliminate the influence of random noise. Then, the processed discrete temperature data is interpolated and fitted by a cubic spline interpolation algorithm to obtain a continuous concrete temperature field distribution. Concrete strain data is obtained from a vibrating wire strain gauge, and a temperature-strain relationship model is established in combination with the temperature field distribution data. The functional relationship between temperature and strain is fitted by linear regression, the correlation coefficient is calculated, the regression equation is determined, the model fit is evaluated, the strain component caused by temperature is calculated, and the temperature strain is eliminated from the total strain to obtain the real deformation data of the concrete. The thermal expansion coefficient of the concrete is calculated based on the compensated concrete temperature data and the real deformation data, and the thermal expansion coefficient is smoothed to obtain a thermal expansion coefficient curve that changes with temperature, and the relationship function between the thermal expansion coefficient and temperature is fitted by the least squares method. The multi-point temperature acquisition system arranges 10 temperature sensors inside the concrete, with depths of 0.1m, 0.3m, 0.5m, 1m and 2m, respectively, and 2 sensors at each depth. The acquisition frequency is once an hour, and the monitoring is continuous for 7 days. After drawing the temperature-time curve, the temperature change function T(t)=a0+a1 t+a2t^2+a3t^3+a4t^4+a5t^5 is fitted using a 5th-order polynomial regression, where T is temperature and t is time. The calculated compensation coefficient is 1.02. The original temperature data is corrected by applying the compensation coefficient, and then smoothed using the 5-point sliding window method. The weighted average is used in the window, and the weights are 0.1, 0.2, 0.4, 0.2, and 0.1, respectively. The processed discrete temperature data is interpolated using the cubic spline interpolation algorithm, and the interpolation interval is 0.1 hours to obtain a continuous temperature field distribution. The vibrating wire strain gauge is arranged at the same position as the temperature sensor to synchronously collect strain data. When establishing the temperature-strain relationship model, linear regression ε=βT+α is used, where ε is strain, T is temperature, β is temperature coefficient, and α is constant term. Calculate the correlation coefficient r, determine the regression equation, and evaluate the model fit R^2. If R^2 is greater than 0.95, the fit is considered good. Calculate the strain component εT=βT caused by temperature, remove εT from the total strain εtota l, and obtain the true deformation εrea l=εtota l-εT. According to the compensated temperature data and true deformation data, calculate the thermal expansion coefficient of concrete α=Δεrea l / ΔT, perform 5-point moving average smoothing on α, and obtain the thermal expansion coefficient curve that changes with temperature.The relationship function between the thermal expansion coefficient and temperature is fitted using a quadratic polynomial α(T)=k0+k1T+k2T^2, where k0, k1, and k2 are unknown coefficients. The coefficients are solved using the least squares method to obtain a specific function of the thermal expansion coefficient changing with temperature.

[0025] Step S102: During the solidification and hardening process of concrete, the concrete gradually acquires a certain stiffness and strength, and transmits the internal stress to the strain sensor. The strain sensor is used to collect the strain signal, and the threshold of the strain harmonic component is set to filter out the harmonic interference introduced by the construction vibration, so as to obtain the fundamental frequency component reflecting the real strain state inside the concrete.

[0026] The original signal of the internal strain of the concrete collected by the strain sensor is obtained, and the original signal is converted into a digital signal by the data acquisition device; the digital signal is smoothed by the five-point moving average method, and denoised by wavelet transform to obtain the preprocessed strain signal; the preprocessed strain signal is subjected to fast Fourier transform to determine the spectrum distribution and main frequency components of the signal, and the fundamental frequency component and each harmonic component are identified from the spectrum; it is determined whether the amplitude of the harmonic component in the spectrum distribution exceeds the preset threshold, and if it exceeds the preset threshold, the harmonic component is filtered out by an adaptive filter; the filtered frequency domain signal is subjected to inverse fast Fourier transform to obtain a time domain signal reflecting the real strain state inside the concrete; the current strain data of the concrete is determined according to the time domain signal and the pre-established concrete age-strength relationship curve.

[0027] Exemplarily, according to the physical properties of concrete during the setting and hardening process and the pre-established concrete age-strength relationship curve, a strain sensor is used to monitor the internal strain of concrete in real time, and the original strain signal is obtained by a data acquisition device with a sampling frequency of 1000Hz and a resolution of 16 bits, and the collected strain signal is converted into a digital signal for storage. The original digital signal is smoothed using the 5-point moving average method, and then denoised using the wavelet transform to obtain the preprocessed strain signal. The preprocessed strain signal is subjected to a fast Fourier transform to obtain the spectrum distribution of the signal, and the main frequency components of the signal are determined by power spectral density analysis. The fundamental frequency component and each harmonic component are identified from the spectrum, and the amplitude and phase of each component are calculated. For the harmonic interference caused by construction vibration, the harmonic component amplitude threshold is set to 1.5 times the average amplitude through statistical analysis of experimental data, and the harmonic components exceeding the threshold are filtered out using an adaptive filter with a minimum mean square error algorithm, retaining the fundamental frequency component and harmonic components with smaller amplitudes. The filtered frequency domain signal is converted back to the time domain using the inverse fast Fourier transform to obtain the time domain signal reflecting the true strain state inside the concrete. Combined with the current strength level of the concrete, the final concrete strain data is obtained.

[0028] After the concrete is poured, the strength of the concrete at different time points is estimated based on the pre-established concrete age-strength relationship curve. For example, for ordinary Portland cement concrete, the strength is about 50% of the design strength at 3 days, about 70% at 7 days, and reaches the design strength at 28 days. 10 strain sensors are buried in the concrete, with depths of 0.1m, 0.3m, 0.5m, 1m and 2m, respectively, and 2 sensors at each depth. The strain data is continuously collected using a data acquisition device with a sampling frequency of 1000Hz and a resolution of 16 bits, and stored once every hour. The collected raw digital signal is first smoothed using the 5-point moving average method, with a window size of 5ms and a weight of [0.1, 0.2, 0.4, 0.2, 0.1]. Then, the db4 wavelet transform is used for 5-layer decomposition, and the high-frequency coefficients are soft-thresholded, with the threshold set to 3 times the standard deviation, and the signal is reconstructed to achieve denoising. The pre-processed signal is subjected to a 2048-point fast Fourier transform to obtain the spectrum distribution. By calculating the power spectrum density, the peak frequency is identified as the fundamental frequency, which is usually in the range of 0.1-10Hz. The harmonic component amplitude threshold is set to 1.5 times the average amplitude. If the amplitude of a frequency component exceeds this threshold, it is considered to be harmonic interference. An adaptive filter with a minimum mean square error algorithm is used, with a step size parameter set to 0.01 and a filter order of 32 to filter out harmonic components that exceed the threshold. Finally, a 2048-point inverse fast Fourier transform is used to convert the filtered frequency domain signal back to the time domain, and combined with the current concrete strength level (such as 70% of the design strength at 7 days of age), the final concrete strain data is calculated.

[0029] Step S103, performing time-frequency analysis on the fundamental frequency components of temperature and strain, selecting a wavelet basis function suitable for the concrete setting and strength growth process, removing high-frequency noise through adaptive threshold processing of wavelet coefficients, and retaining low-frequency signal characteristics reflecting the concrete strength growth.

[0030] The time series data of the fundamental frequency components of temperature and strain are input into the short-time Fourier transform processing unit, and the short-time Fourier transform processing unit calculates the time-frequency spectrum according to the preset window length and overlap rate; by analyzing the time-frequency spectrum, different frequency characteristics in the process of concrete solidification and strength growth are obtained; for the temperature and strain signals, a pre-selected Daubechies 4th-order wavelet basis function is used to perform 5-layer wavelet decomposition to obtain wavelet coefficients of different scales; for the wavelet coefficients, if an adaptive threshold processing method based on Bayesian estimation is used, the noise variance at each scale is calculated according to the preset noise distribution and signal distribution; the wavelet coefficients are subjected to soft threshold shrinkage processing using the adaptive threshold processing method to obtain the denoised wavelet coefficients; the signal is reconstructed according to the denoised wavelet coefficients to obtain the temperature and strain signals after removing the high-frequency noise; for the temperature and strain signals after removing the high-frequency noise, a Butterworth low-pass filter with a preset cutoff frequency and order is used to extract the low-frequency signal characteristics reflecting the strength growth of concrete to obtain the final temperature and strain data.

[0031] Exemplarily, according to the time series data of the fundamental frequency components of temperature and strain, the short-time Fourier transform is used to perform time-frequency analysis on the signal, the window length is set to 1 / 10 of the signal length, the overlap rate is 50%, and the time-frequency spectrum is calculated, from which the frequency characteristics of different time periods are extracted. By analyzing the time-frequency spectrum, the frequency characteristics related to the coagulation and strength growth of concrete are identified, such as the low-frequency components (0.1-1Hz) in the coagulation process and the medium-frequency components (1-10Hz) in the strength growth process. The performance of different wavelet basis functions such as Haar, Daubechies, and Symlets in characterizing the characteristics of concrete signals is compared, and the most suitable Daubechies 4th order wavelet is selected as the basis function. The temperature and strain signals are decomposed into 5 layers of wavelets to obtain wavelet coefficients of different scales. For the wavelet coefficients, an adaptive threshold processing method based on Bayesian estimation is used. It is assumed that the noise obeys Gaussian distribution and the signal obeys generalized Gaussian distribution. The noise variance at each scale is calculated, and the threshold size is automatically adjusted according to the signal-to-noise ratio, and the wavelet coefficients are soft-threshold shrunk. The processed wavelet coefficients are used for signal reconstruction to obtain the temperature and strain signals after removing high-frequency noise. The low-frequency signal characteristics reflecting the growth of concrete strength are extracted by designing a Butterworth low-pass filter with a cutoff frequency of 10Hz and an order of 4, and the final temperature and strain data are obtained. After the concrete is poured, the time series data of the fundamental frequency components of temperature and strain are collected with a sampling frequency of 100Hz for 72 hours. These data are subjected to short-time Fourier transform, and the window length is set to 7200 data points, i.e. 1 hour, with an overlap rate of 50%, to obtain a time-frequency spectrum. Analysis of the time-frequency spectrum shows that within 0-24 hours, the energy of the 0.1-1Hz frequency band gradually increases, corresponding to the concrete setting process; within 24-72 hours, the energy of the 1-10Hz frequency band slowly increases, reflecting the growth of strength. Comparing different wavelet basis functions, it is found that the Daubech i es 4th-order wavelet performs best in maintaining signal smoothness and capturing mutation characteristics, and it is selected for 5-layer wavelet decomposition. The wavelet coefficients were subjected to Bayesian adaptive threshold processing. Assuming that the noise variance was 0.01 and the signal prior distribution parameter was 1.5, the optimal threshold of each layer was calculated by maximum a posteriori probability estimation, such as 0.05 for the first layer and 0.002 for the fifth layer. After soft threshold processing, the signal was reconstructed to obtain the denoised temperature and strain data. A 4th-order Butterworth low-pass filter with a cutoff frequency of 10 Hz was designed to extract low-frequency features through filtering. The final temperature data showed that the internal temperature of the concrete began to rise rapidly after 12 hours and reached a peak of 78°C after 24 hours; the strain data showed that the shrinkage strain of the concrete reached 350 microstrains at the age of 28 days, which was consistent with the strength growth curve.

[0032] Step S104, using a singular value decomposition algorithm to extract features from the denoised temperature and strain signals, and detecting internal defects of the concrete through the mutation of the singular value spectrum. If there are cracks or holes, stress concentration will cause a mutation in the strain signal.

[0033] The Hankel matrix of the denoised temperature signal and strain signal is constructed, and the singular value matrix of the Hankel matrix is ​​obtained by the singular value decomposition algorithm; the singular value sequence is extracted from the singular value matrix, and the local polynomial regression algorithm is used to smooth the singular value sequence to obtain the smoothed singular value spectrum curve; for the smoothed singular value spectrum curve, a sliding window is set to calculate the first-order difference of the curve, and the dynamic threshold algorithm is used to determine whether there is a mutation point in the singular value spectrum; if a mutation point is detected in the singular value spectrum, the signal is reconstructed using the left singular vector and the right singular vector, and the time-frequency features near the mutation point are extracted; the strain gradient of the strain signal is calculated, and the degree of mutation is quantified according to the strain gradient. Combined with the change of the temperature signal, the time-frequency features are compared through the pre-established feature library to determine whether there are cracks or hole defects.

[0034] Exemplarily, a Hankel matrix is ​​constructed for the denoised temperature and strain signals, and the number of rows and columns of the matrix is ​​determined according to the signal length, the number of rows is set to 1 / 3 of the signal length, and the number of columns is the remaining length. The singular value matrix, the left singular vector matrix, and the right singular vector matrix are calculated using the singular value decomposition algorithm. The singular value sequence is extracted from the singular value matrix, and the singular value spectrum curve is drawn. The singular value spectrum is smoothed using a third-order local polynomial regression algorithm, and the bandwidth is set to 10% of the number of singular values ​​to obtain a smoothed singular value spectrum curve, which reflects the main characteristics and energy distribution of the signal. For the smoothed singular value spectrum curve, a sliding window with a length of 5% of the number of singular values ​​and a step size of 1 / 4 of the window length is set to calculate the first-order difference of the curve. The mutation threshold of the differential value is determined by a dynamic threshold algorithm based on the signal variance, and the threshold is set to 3 times the signal standard deviation to determine whether there is a mutation point in the singular value spectrum. If a mutation point is detected in the singular value spectrum, the signal is reconstructed using the left singular vector and the right singular vector, and the reconstructed signal is subjected to time-frequency analysis to extract the time-frequency features near the mutation point. The strain gradient is calculated to quantify the degree of mutation, and the stress concentration is comprehensively judged by combining the temperature signal change. The mutation features are compared through the defect feature library accumulated by finite element analysis and experimental data to determine whether there are cracks or holes in the concrete. Taking the temperature and strain data collected within 72 hours after concrete pouring as an example, the signal length is 43,200 points and the sampling is once every 10 minutes. When constructing the Hankel matrix, the number of rows is set to 14,400 and the number of columns is 28,801. After performing singular value decomposition, 14,400 singular values ​​are obtained. A third-order local polynomial regression is applied to the singular value spectrum with a bandwidth set to 1440 to obtain a smooth curve. A sliding window of 720 points is set with a step size of 180, and the first-order difference is calculated. Assuming the signal standard deviation is 0.05, the mutation threshold is set to 0.15. Mutation points were detected near 24 hours and 48 hours, corresponding to the 720th and 1440th positions of the singular value sequence, respectively. The signals near these two time points were reconstructed, wavelet transform was performed, and the energy of the 0-10Hz frequency band was extracted. The strain gradient was calculated, and the maximum gradient at the 24-hour point was 5με / h, and at the 48-hour point was 8με / h. The change rates of the temperature signal at these two time points were 2℃ / h and 1.5℃ / h, respectively. Comparing the patterns in the defect feature library, it was found that the characteristics at the 48-hour point were similar to the microcrack pattern, with a strain gradient greater than 5με / h and a temperature change rate less than 2℃ / h. Comprehensively judging, microcracks may have appeared about 48 hours after the concrete was poured, which requires further monitoring and verification.

[0035] Step S105, if defects are detected inside the concrete, that is, there is spatial non-uniformity in temperature and stress distribution, the temperature and strain monitoring data arranged at multiple points and the defect detection results are used as boundary conditions and constraints to obtain the temperature field and stress field inside the entire concrete structure through numerical simulation.

[0036] A three-dimensional finite element mesh model is constructed based on the geometric characteristics of the concrete structure. The mesh model uses hexahedral units and the mesh size is set to a preset range. If a defective area is detected, the mesh size of the defective area is reduced to a preset value. Acquire temperature and strain monitoring data from multiple points, and use the Kriging interpolation algorithm to interpolate the monitoring data to the finite element mesh nodes. If it is a mesh node in the defective area, reduce the search radius to a preset value. Establish control equations for the temperature field and stress field, and the control equations are based on the thermomechanical coupling constitutive equations of concrete. Substitute the interpolated temperature and strain data into the control equations as boundary conditions and constraints. Use the Newton-Raphson iteration method to solve the control equations to obtain the temperature field and stress field distribution inside the concrete structure.

[0037] Exemplarily, according to the geometric characteristics and actual size of the concrete structure, a three-dimensional finite element mesh model is constructed, the mesh unit type is selected as a hexahedral unit, the mesh size range is set to 0.1m-0.5m, and the mesh size is reduced to 0.05m for encryption in the area where defects are detected. The temperature and strain monitoring data arranged at multiple points are interpolated to the finite element mesh nodes, and the Kriging interpolation algorithm is used to realize the spatial distribution of the monitoring data. The spherical semivariogram model is selected, and the search radius is set to 2 times the spacing between the monitoring points. For the mesh nodes in the defective area, the search radius is reduced to 1 / 2 of the original. The quality of the interpolation results is evaluated by the cross-validation method, and the root mean square error is calculated. Based on the thermomechanical coupling constitutive equation of concrete, including the heat conduction equation and the thermal stress equation, the control equations of the temperature field and the stress field are established, and the interpolated temperature and strain data are substituted into the equation as boundary conditions and constraints. The Newton-Raphson iteration method is used to solve the nonlinear equations, and the initial value is set to the average value of the interpolation data, the error threshold is set to 10^-6, and the maximum number of iterations is 100. The temperature field and stress field distribution inside the entire concrete structure are obtained by numerical solution. The stress state inside the concrete is judged by the stress invariant criterion. When the first stress invariant is greater than 30% of the concrete compressive strength, it is judged as a high-pressure stress area. The maximum principal stress criterion is used to evaluate the risk of crack extension. When the maximum principal stress exceeds 80% of the concrete tensile strength, it is judged as a high-risk area. The spatial coefficient of variation of temperature and stress is calculated to quantify the spatial heterogeneity of the temperature field and stress field. Taking a 30m×10m×5m concrete pier as an example, a three-dimensional finite element mesh model is constructed. The mesh unit uses hexahedral units and the basic mesh size is set to 0.3m. In the area where microcracks are detected, which is located in the 2m×2m×1m range in the middle of the pier, the mesh size is reduced to 0.05m for encryption. 50 temperature sensors and 30 strain sensors are arranged, and the monitoring data are interpolated to the grid nodes using the Kriging interpolation algorithm. The spherical semivariogram model parameters are set as: base value 0.8, range 2m. For the microcrack area, the search radius is reduced from 1m to 0.5m. Cross-validation shows that the root mean square error of temperature interpolation is 0.5℃, and the root mean square error of strain interpolation is 5με. The thermomechanical coupling equation is established, the thermal conductivity is 2.5W / (m·K), and the linear expansion coefficient is 1×10^-5 / ℃. The Newton-Raphson method is used to solve the problem, the initial temperature field is set to 25℃, the initial stress field is set to 0MPa, and it converges after 100 iterations. The solution results show that the maximum temperature in the microcrack area reaches 65℃, and the average temperature in the surrounding area is 45℃; the maximum principal stress is 3.2MPa, which is close to 80% of the concrete tensile strength of 4MPa. The spatial coefficient of variation of the temperature field is 0.18, and the spatial coefficient of variation of the stress field is 0.25, indicating that the non-uniformity of the stress distribution is more significant. According to the stress invariant criterion, three high-pressure stress zones are identified with a total volume of approximately 0.5m 3 .

[0038] Step S106: According to the known constitutive relationship between the strength of concrete materials and the temperature and stress state, the strength of concrete is closely related to its temperature history and stress state. Appropriate high temperature promotes strength growth, while excessive temperature stress and shrinkage stress lead to cracking and strength reduction. The spatial strength distribution of concrete is evaluated by comparing the temperature field and stress field inside the entire concrete structure obtained by simulation with the constitutive relationship between strength and temperature and stress state.

[0039] The concrete mix ratio and curing conditions are obtained, and a constitutive relationship database between concrete strength and temperature and stress state is established according to the mix ratio and curing conditions; the temperature field and stress field data calculated by the finite element software are received, and the temperature and stress values ​​of each grid node are extracted from the temperature field and stress field data; for the temperature and stress values, the nearest neighbor interpolation method is used to retrieve the corresponding strength values ​​in the constitutive relationship database to obtain the preliminary concrete spatial strength distribution; if strength correction is required, the correction coefficient method is used to correct the preliminary concrete spatial strength distribution, and the correction coefficient is calculated based on the equivalent age theory; it is determined whether the stress exceeds the preset threshold value of the concrete tensile strength. If the stress exceeds the preset threshold value, the strength value of the area is multiplied by the preset strength reduction factor to obtain the final concrete spatial strength distribution.

[0040] Exemplarily, according to the concrete mix ratio and curing conditions, a constitutive relationship database between concrete strength and temperature and stress state is established, and the functional relationship between strength, temperature and stress is obtained through third-order polynomial regression analysis, and the functional relationship is discretized into a lookup table to store the corresponding strength values ​​under different temperature and stress combinations. The temperature field and stress field data calculated by finite element software are used to extract the temperature and stress values ​​of each grid node. These data are used as input parameters, and the corresponding strength values ​​are retrieved in the constitutive relationship lookup table using the nearest neighbor interpolation method to obtain a preliminary concrete spatial strength distribution. The preliminary strength distribution is statistically analyzed to calculate the average strength and coefficient of variation. Considering the growth characteristics of concrete strength over time, the preliminary strength distribution is corrected by the correction coefficient method. The correction coefficient is calculated based on the equivalent age theory, which contains a temperature acceleration factor to consider the effect of high temperature on strength growth. The temperature history curve is extracted through the time series of the temperature field data, the equivalent age and correction coefficient of each grid node are calculated, and the preliminary strength is multiplied by the correction coefficient to obtain the corrected strength distribution. In view of the cracking risk caused by temperature stress and shrinkage stress, a strength reduction factor is set. When the stress exceeds 70% of the tensile strength of concrete, the strength of the area is reduced to 80% of the original. The reduced strength value is compared with the corrected strength distribution, and the smaller value of the two is taken as the final concrete spatial strength distribution. Taking a 30m×10m×5m concrete bridge pier as an example, a strength-temperature-stress constitutive relationship database is established. Using third-order polynomial regression analysis, the strength function

[0041] f(T,σ)=a0+a1 T+a2σ+a3T^2+a4σ^2+a5Tσ+a6T^3+a7σ^3+a8T^2σ+a9Tσ^2, where T is temperature (℃) and σ is stress (MPa). The temperature range of 0-100℃ and the stress range of 0-40MPa were discretized into a 100×100 lookup table. The temperature and stress values ​​of 54,000 grid nodes were calculated using finite element software. The corresponding strength was retrieved from the lookup table by the nearest neighbor interpolation method to obtain the preliminary strength distribution. Statistical analysis showed that the average strength was 35MPa and the coefficient of variation was 0.15. The correction coefficient was calculated based on the equivalent age theory, the temperature acceleration factor used the Arrhenius equation, and the activation energy was 33.5kJ / mol. The temperature history curve was extracted, and the 28-day equivalent age of each node was calculated. The correction coefficient ranged from 0.9 to 1.2. The preliminary strength was multiplied by the correction coefficient to obtain the corrected strength distribution. The strength reduction factor is set. When the stress exceeds 70% of the tensile strength (3.5MPa), the strength is reduced to 80% of the original strength. The final spatial strength distribution of concrete is 28-42MPa, with an average value of 37MPa and a coefficient of variation of 0.12. The strength of the middle area of ​​the pier is higher, and the surface area is slightly lower due to the influence of temperature stress.

[0042] Step S107, based on the spatial strength distribution of concrete, evaluate the overall strength of concrete, fuzzify the strength evaluation result, obtain the membership function of strength, set the fuzzy rule of strength grade, obtain the membership of concrete strength grade through fuzzy reasoning, use the membership weighted average method to defuzzify, and obtain the quantitative evaluation result of concrete strength; if the strength meets the design requirements, the concrete reaches the design strength and enters the structural health monitoring stage; if the strength is insufficient or degradation or internal defects occur, timely warning will be given.

[0043] The spatial strength distribution data of concrete is obtained, and the kernel density estimation algorithm is used to fit the probability density distribution of the strength data. The bandwidth parameter of the kernel density estimation algorithm is automatically determined by the Silvman rule. According to the probability density distribution, the cumulative distribution function of the strength is obtained by integral operation, and the cumulative distribution function is used as the membership function of the strength. According to the membership function, the fuzzy rules of the concrete strength grade are set, and the strength is divided into three grades of low strength, medium strength and high strength. The trapezoidal membership function is used to describe the fuzzy set of each grade. Using the Mamdan i fuzzy reasoning mechanism, the strength membership function is matched with the pre-established fuzzy rule library to calculate the membership value of each strength grade. If the membership value is obtained, the centroid method is used to defuzzify the fuzzy reasoning result, and the quantitative evaluation result of the concrete strength is calculated by discretizing the membership function into a preset number of equally divided points. It is judged whether the quantitative evaluation result meets the design requirements. If it does not meet the design requirements, the corresponding level of early warning mechanism is triggered to generate early warning information containing the specific value of insufficient strength, location information and defect status.

[0044] Exemplarily, according to the spatial strength distribution data of concrete, the kernel density estimation algorithm is used to fit the probability density distribution of the strength data, and the bandwidth selection is automatically determined using the Silverman rule. The cumulative distribution function of the strength is obtained by integral operation, which is used as the membership function of the strength. The function characteristics reflect the central tendency and discrete degree of the strength distribution. The fuzzy rules of the concrete strength grade are set, and the strength is divided into three grades: low strength, medium strength and high strength. The trapezoidal membership function is used to describe the fuzzy set of each grade, and the function parameters are determined based on the percentage of the design strength: low strength is 0-80%, medium strength is 70-110%, and high strength is 100-130%. Combining expert knowledge and historical data, a fuzzy rule base is constructed, which includes strength grade judgment and defect impact assessment rules. Using the Mamdani fuzzy reasoning mechanism, the strength membership function is matched with the fuzzy rules, and the membership value of each strength grade is calculated. The centroid method is used to defuzzify the fuzzy reasoning results, and the membership function is discretized into 100 equally divided points. The quantitative evaluation results of concrete strength are obtained by weighted average calculation. The quantitative evaluation results are compared with the design requirements, and multi-level warning thresholds are set: the yellow warning is 90% of the design strength, the orange warning is 85%, and the red warning is 80%. If the strength meets the design requirements, a signal that the design strength is reached is output; if the strength is insufficient or degradation or internal defects are abnormal, the corresponding level of warning mechanism is triggered to generate warning information containing the specific value of insufficient strength, location information and defect status. Taking a 30m×10m×5m concrete pier as an example, the spatial strength distribution data contains the strength values ​​of 54,000 grid nodes. The kernel density estimation algorithm is used to fit the probability density distribution, and the optimal bandwidth calculated by the Si l verman rule is 1.2MPa. The probability density function after fitting is approximately normally distributed, with a mean of 38MPa and a standard deviation of 2.5MPa. The cumulative distribution function obtained by integration is used as the strength membership function, with a membership of 0.11 at 35MPa and 0.79 at 40MPa. Set the trapezoidal membership function parameters of the strength level, low strength

[0045] [0,0,28,32]MPa, medium strength [28,32,40,44]MPa, high strength [40,44,52,56]MPa. The fuzzy rule base contains 15 rules, such as "IF the strength is medium strength AND the degree of defect is slight THEN the structural condition is good". Mamdan i fuzzy reasoning calculation obtains the membership of low, medium and high strength levels as 0.05, 0.75 and 0.20 respectively. When the centroid method is used to defuzzify, the membership function is discretized into 100 points, and the calculated quantitative evaluation result is 37.8MPa. The design strength requirement is 35MPa, and the current strength meets the requirement. The warning threshold is set to 31.5MPa for yellow, 29.8MPa for orange, and 28MPa for red. Output the signal of "reaching the design strength" and record the positions of two microcracks (coordinates 15m, 5m, 2.5m and 22m, 3m, 1m) for subsequent structural health monitoring.

[0046] Obviously, those skilled in the art can make various changes and modifications to the embodiments of the present application without departing from the spirit and scope of the embodiments of the present application. Thus, if these modifications and variations of the embodiments of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for detecting information of intelligent sensors for civil construction monitoring under signal mismatch, characterized in that: The method comprises: According to the temperature change characteristics caused by the heat of concrete hydration, the compensation coefficient of the temperature sensor embedded in the concrete is calculated, and the original temperature data collected is drift corrected to obtain the compensated concrete temperature data. At the same time, the embedded vibrating wire strain gauge is used to monitor the deformation of the concrete. During the solidification and hardening process of concrete, concrete gradually acquires a certain stiffness and strength, and transmits the internal stress to the strain sensor. The strain sensor is used to collect strain signals, and the harmonic interference introduced by construction vibration is filtered out by setting the threshold of the strain harmonic component to obtain the fundamental frequency component reflecting the real strain state inside the concrete. Conduct time-frequency analysis on the fundamental frequency components of temperature and strain, select the wavelet basis function that is suitable for the concrete setting and strength growth process, and remove high-frequency noise through adaptive threshold processing of wavelet coefficients to retain the low-frequency signal characteristics that reflect the strength growth of concrete. The singular value decomposition algorithm is used to extract the features of the denoised temperature and strain signals. The internal defects of the concrete are detected through the mutation of the singular value spectrum. If there are cracks or holes, stress concentration will cause mutation of the strain signal. If defects are detected inside the concrete, that is, there is spatial non-uniformity in the temperature and stress distribution, the temperature and strain monitoring data arranged at multiple points and the defect detection results are used as boundary conditions and constraints to obtain the temperature field and stress field inside the entire concrete structure through numerical simulation; According to the known constitutive relationship between the strength of concrete materials and temperature and stress state, the strength of concrete is closely related to its temperature history and stress state. Appropriate high temperature promotes strength growth, while excessive temperature stress and shrinkage stress lead to cracking and strength reduction. By comparing the temperature field and stress field inside the entire concrete structure obtained by simulation with the constitutive relationship between strength, temperature and stress state, the spatial strength distribution of concrete is evaluated. Based on the spatial strength distribution of concrete, the overall strength of concrete is evaluated, and the strength evaluation results are fuzzy processed to obtain the strength membership function. The fuzzy rules of the strength grade are set, and the membership of the concrete strength grade is obtained through fuzzy reasoning. The membership weighted average method is used to defuzzify and obtain the quantitative evaluation result of the concrete strength. If the strength meets the design requirements, the concrete reaches the design strength and enters the structural health monitoring stage. If the strength is insufficient or degradation or internal defects occur, timely warnings will be issued.

2. The method according to claim 1, characterized in that The method calculates the compensation coefficient of the temperature sensor embedded in the concrete according to the temperature change characteristics caused by the hydration heat of the concrete, performs drift correction on the collected original temperature data, obtains the compensated concrete temperature data, and simultaneously uses the embedded vibrating wire strain gauge to monitor the deformation of the concrete, including: Obtain the internal temperature data of the concrete collected by the multi-point temperature collection system, and draw a temperature-time curve based on the temperature data; The temperature variation function caused by the hydration heat of concrete is determined by using the polynomial regression method, and the compensation coefficient of the temperature sensor is calculated according to the temperature variation function; The compensation coefficient is applied to the original temperature data for drift correction. If the corrected temperature data is smoothed using the sliding window method, the influence of random noise is eliminated; The processed discrete temperature data are interpolated and fitted using the cubic spline interpolation algorithm to obtain the continuous concrete temperature field distribution. The concrete strain data is obtained from the vibrating wire strain gauge, and the temperature-strain relationship model is established by combining the temperature field distribution data. The functional relationship between temperature and strain is fitted by linear regression. The thermal expansion coefficient of concrete is calculated based on the compensated concrete temperature data and the real deformation data, and the thermal expansion coefficient data is smoothed to obtain a thermal expansion coefficient curve that changes with temperature; The relationship function between thermal expansion coefficient and temperature was fitted using the least squares method.

3. The method according to claim 1, characterized in that During the coagulation and hardening process of concrete, the concrete gradually acquires a certain rigidity and strength, transmits the internal stress to the strain sensor, collects the strain signal by using the strain sensor, and obtains the fundamental frequency component reflecting the real strain state inside the concrete by setting the threshold of the strain harmonic component and filtering out the harmonic interference introduced by the construction vibration, including: The original signal of the internal strain of the concrete collected by the strain sensor is obtained, and the original signal is converted into a digital signal by the data acquisition device; The digital signal is smoothed by using the five-point moving average method and denoised by wavelet transform to obtain the preprocessed strain signal. Perform fast Fourier transform on the preprocessed strain signal to determine the spectrum distribution and main frequency components of the signal, and identify the fundamental frequency component and each harmonic component from the spectrum; Determine whether the amplitude of the harmonic component in the spectrum distribution exceeds a preset threshold, and if so, use an adaptive filter to filter out the harmonic component; Perform inverse fast Fourier transform on the filtered frequency domain signal to obtain the time domain signal reflecting the real strain state inside the concrete; The current strain data of concrete is determined based on the time domain signal and the pre-established concrete age-strength relationship curve.

4. The method according to claim 1, characterized in that: The time-frequency analysis of the temperature and strain fundamental frequency components is performed, and a wavelet basis function suitable for the concrete coagulation and strength growth process is selected. The high-frequency noise is removed through adaptive threshold processing of the wavelet coefficients, and the low-frequency signal characteristics reflecting the concrete strength growth are retained, including: The time series data of the fundamental frequency components of temperature and strain are input into a short-time Fourier transform processing unit, and the short-time Fourier transform processing unit calculates a time-frequency spectrum according to a preset window length and overlap rate; By analyzing the time-frequency spectrum, different frequency characteristics of concrete solidification and strength growth process can be obtained; For temperature and strain signals, a pre-selected Daubechies 4th-order wavelet basis function is used to perform 5-layer wavelet decomposition to obtain wavelet coefficients of different scales. For the wavelet coefficients, if the adaptive threshold processing method based on Bayesian estimation is used, the noise variance at each scale is calculated according to the preset noise distribution and signal distribution; The wavelet coefficients are subjected to soft threshold shrinkage processing using the adaptive threshold processing method to obtain the denoised wavelet coefficients. Signal reconstruction is performed based on the denoised wavelet coefficients to obtain the temperature and strain signals after high-frequency noise is removed; For the temperature and strain signals after removing high-frequency noise, a Butterworth low-pass filter with a preset cutoff frequency and order is used to extract the low-frequency signal characteristics reflecting the growth of concrete strength to obtain the final temperature and strain data.

5. The method according to claim 1, characterized in that The singular value decomposition algorithm is used to extract features from the denoised temperature and strain signals, and the internal defects of the concrete are detected by the mutation of the singular value spectrum. If there are cracks or holes, stress concentration will cause mutations in the strain signal, including: Construct the Hankel matrix of the denoised temperature signal and strain signal, and obtain the singular value matrix of the Hankel matrix through the singular value decomposition algorithm; Extracting a singular value sequence from a singular value matrix, smoothing the singular value sequence using a local polynomial regression algorithm, and obtaining a smoothed singular value spectrum curve; For the smoothed singular value spectrum curve, set a sliding window to calculate the first-order difference of the curve, and determine whether there is a mutation point in the singular value spectrum based on the dynamic threshold algorithm; If a mutation point is detected in the singular value spectrum, the signal is reconstructed using the left singular vector and the right singular vector to extract the time-frequency features near the mutation point; The strain gradient of the strain signal is calculated, and the degree of mutation is quantified according to the strain gradient. Combined with the change of the temperature signal, the time-frequency characteristics are compared through the pre-established feature library to determine whether there are cracks or hole defects.

6. The method according to claim 1, characterized in that If defects are detected inside the concrete, that is, there is spatial non-uniformity in temperature and stress distribution, the temperature and strain monitoring data arranged at multiple points and the defect detection results are used as boundary conditions and constraints to obtain the temperature field and stress field inside the entire concrete structure through numerical simulation, including: A three-dimensional finite element mesh model is constructed according to the geometric characteristics of the concrete structure. The mesh model uses hexahedral units and the mesh size is set to a preset range. If a defective area is detected, the grid size of the defective area is reduced to a preset value; Acquire temperature and strain monitoring data from multiple points, and use the Kriging interpolation algorithm to interpolate the monitoring data to the finite element grid nodes; If it is a grid node in the defect area, the search radius is reduced to a preset value; Establish the control equations of temperature field and stress field, which are based on the thermomechanical coupling constitutive equation of concrete; Substitute the interpolated temperature and strain data into the control equations as boundary conditions and constraints; The Newton-Raphson iteration method is used to solve the governing equations and obtain the temperature and stress field distributions inside the concrete structure.

7. The method according to claim 1, characterized in that According to the known constitutive relationship between the strength of concrete materials and temperature and stress state, the strength of concrete is closely related to its temperature history and stress state. Appropriate high temperature promotes strength growth, while excessive temperature stress and shrinkage stress lead to cracking and strength reduction. By comparing the temperature field and stress field inside the entire concrete structure obtained by simulation with the constitutive relationship between strength and temperature and stress state, the spatial strength distribution of concrete is evaluated, including: Obtain concrete mix ratio and curing conditions, and establish a constitutive relationship database between concrete strength and temperature and stress state based on the mix ratio and curing conditions; Receive temperature field and stress field data calculated by finite element software, and extract temperature and stress values ​​of each grid node from the temperature field and stress field data; According to the temperature and stress values, the nearest neighbor interpolation method is used to retrieve the corresponding strength values ​​in the constitutive relationship database to obtain the preliminary concrete spatial strength distribution; If strength correction is required, the preliminary concrete spatial strength distribution is corrected using the correction coefficient method, and the correction coefficient is calculated based on the equivalent age theory; It is determined whether the stress exceeds the preset threshold of the concrete tensile strength. If the stress exceeds the preset threshold, the strength value of the area is multiplied by the preset strength reduction factor to obtain the final concrete spatial strength distribution.

8. The method according to claim 1, characterized in that Based on the spatial strength distribution of concrete, the overall strength of concrete is evaluated, the strength evaluation result is fuzzified to obtain the strength membership function, the fuzzy rule of strength grade is set, the membership of concrete strength grade is obtained through fuzzy reasoning, and the membership weighted average method is used to defuzzify to obtain the quantitative evaluation result of concrete strength. If the strength meets the design requirements, the concrete reaches the design strength and enters the structural health monitoring stage. If the strength is insufficient or degradation or internal defects occur, timely warning is given, including: The spatial strength distribution data of concrete is obtained, and the probability density distribution of the strength data is fitted using the kernel density estimation algorithm. The bandwidth parameter of the kernel density estimation algorithm is automatically determined by the Silverman rule. According to the probability density distribution, the cumulative distribution function of the intensity is obtained through integral operation, and the cumulative distribution function is used as the membership function of the intensity; According to the membership function, the fuzzy rules of concrete strength grade are set, and the strength is divided into three grades: low strength, medium strength and high strength. The trapezoidal membership function is used to describe the fuzzy set of each grade. Using Mamdani fuzzy inference mechanism, the intensity membership function is matched with the pre-established fuzzy rule base to calculate the membership value of each intensity level; If the membership value is obtained, the centroid method is used to defuzzify the fuzzy reasoning result, and the quantitative evaluation result of the concrete strength is calculated by discretizing the membership function into a preset number of equally divided points; Determine whether the quantitative evaluation results meet the design requirements. If not, the corresponding level of early warning mechanism is triggered to generate early warning information including the specific value of insufficient strength, location information and defect status.

9. An intelligent sensor information detection system for civil construction monitoring under signal mismatch, characterized in that: The system comprises: Data acquisition and correction module, used for temperature sensor data correction and vibrating wire strain gauge monitoring; Signal processing module for interference filtering, time-frequency analysis and wavelet denoising; Defect detection module, used for singular value decomposition feature extraction and mutation detection; Temperature stress field simulation module, used to numerically simulate the temperature stress distribution inside the overall structure; The intensity assessment module is used for spatial intensity distribution assessment and fuzzy reasoning intensity level judgment.

Images