Hydraulic engineering health state intelligent perception method fusing physical model constraint

Abstract

The application relates to the field of engineering state management, and discloses a water conservancy project health state intelligent sensing method fusing physical model constraints, which is used for solving the problem that when engineering health sensing is performed, the previous health threshold is no longer representative due to the fact that the structure has undergone maintenance or reinforcement operation, comprising the following steps: obtaining a water conservancy project structure health state index according to preprocessed monitoring data and theoretical response parameters; obtaining a structure performance recovery characteristic parameter; obtaining a structure influence index according to the structure performance recovery characteristic parameter; judging whether the current structure is influenced by maintenance and reinforcement according to the structure influence index; if it is judged that the current structure is influenced by maintenance and reinforcement, correcting an initial health threshold to obtain a corrected health threshold; and judging whether the current structure is in a healthy state according to the health state index and the corrected health threshold, so that the accuracy of engineering health state sensing is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of engineering status management, and more specifically to an intelligent sensing method for the health status of water conservancy projects that integrates physical model constraints. Background Technology

[0002] As a crucial infrastructure for the safe operation of the national economy and society, the structural health of water conservancy projects directly impacts the basin's scheduling capacity, storage and discharge safety, and downstream flood control and disaster prevention capabilities. With advancements in sensing technology and information technology, Internet of Things (IoT)-based online monitoring systems for water conservancy projects have been widely applied to typical structures such as dams, gates, dikes, and canals. These systems enable real-time acquisition and analysis of multi-dimensional data, including structural strain, seepage pressure, water level, vibration, and temperature. By deploying strain gauges, piezometers, water level gauges, and accelerometers at key stress points, the system continuously acquires structural operational status information and transmits it via wireless communication networks to the cloud or edge computing nodes, forming a visualized monitoring system for the health status of water conservancy projects. This provides data support for operation management and maintenance decisions.

[0003] Existing water conservancy project health monitoring systems generally adopt a judgment method based on comparing monitoring indicators with set thresholds. By calculating characteristic indicators such as strain amplitude, seepage pressure change rate or vibration energy, and comparing them with preset health thresholds, it is possible to determine whether there are any abnormalities in the structure.

[0004] However, the above-mentioned technologies have at least the following technical problems: In actual engineering operations, especially after a structure has undergone maintenance or reinforcement, the core physical parameters of the hydraulic engineering structure will change. This will cause the health thresholds set before maintenance based on historical data to gradually lose their representativeness. If the system still uses the old thresholds for status determination, it is easy to have problems such as "false anomalies" or "missed potential anomalies", which will affect the reliability of health assessment and the accuracy of engineering decisions. Summary of the Invention

[0005] To overcome the aforementioned deficiencies in the prior art, this invention provides an intelligent sensing method for the health status of water conservancy projects that integrates physical model constraints, thereby addressing the problems existing in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A method for intelligent sensing of the health status of hydraulic engineering projects constrained by a physical model includes the following steps: Step 1: Strain gauges, piezometers, water level gauges, and accelerometers are deployed at key parts of the hydraulic engineering project to collect real-time monitoring data on the structural operation of the hydraulic engineering project. The monitoring data include strain, seepage pressure, water level, and acceleration, forming a monitoring dataset; Step 2: The monitoring dataset is preprocessed to obtain preprocessed monitoring data; Step 3: Key characteristic indicators are calculated based on the preprocessed monitoring data. The key characteristic indicators include root mean square value, peak factor, energy entropy, dominant frequency, and frequency band energy ratio; Step 4: A finite element physical model is established based on the hydraulic structure design drawings and material parameters. Theoretical response parameters are obtained through the finite element physical model, including theoretical strain, theoretical seepage pressure, theoretical water level, and theoretical acceleration. Step 5: Based on the preprocessed monitoring data and theoretical response parameters, the structural health status index of the water conservancy project is obtained, and structural performance recovery characteristic parameters are acquired. These parameters include equivalent stiffness, strain response amplitude, and seepage pressure response. The structural impact index is then assessed based on these parameters. The impact index is used to determine whether the current structure is affected by maintenance and reinforcement. Step 6: If the current structure is affected by maintenance and reinforcement, an initial health threshold is obtained. This initial health threshold is then corrected to obtain a corrected health threshold. Step 7: Based on the health status index and the corrected health threshold, the current structure is determined to be in a healthy state. If it is determined to be in a healthy state, the process returns to Step 1 to continue data monitoring. If it is determined to be in a non-healthy state, an early warning is issued.

[0007] Preferably, the steps for obtaining the key feature indicators are as follows: For each monitoring quantity, firstly, acquire all sampling point data of the signal within the monitoring period to obtain a monitoring data sequence, and calculate the root mean square value of the monitoring data sequence; in the monitoring data sequence of each monitoring quantity, extract the maximum absolute value of the signal as the peak value, divide the peak value by the root mean square value of the signal to obtain the peak factor; perform energy decomposition on each monitoring quantity to obtain the energy distribution of the signal under each frequency component, and sum the negative values ​​of the product of the energy proportion of each frequency band and its natural logarithm to obtain the energy entropy; perform a fast Fourier transform on each monitoring quantity to obtain the spectral amplitude distribution, and identify the frequency component with the largest amplitude as the dominant frequency; according to the spectral amplitude distribution, obtain the target frequency band energy and the reference frequency band energy, and calculate the ratio of the target frequency band energy to the reference frequency band energy to obtain the frequency band energy ratio.

[0008] Preferably, the steps for obtaining the structural health status index are as follows: For each monitored quantity, the absolute difference between the monitored value and the corresponding theoretical value is subtracted, and then divided by the corresponding theoretical value to obtain the relative deviation value; within a monitoring period, the relative deviation values ​​are arranged in chronological order to form a deviation value sequence, and the average value and standard deviation of the deviation value sequence are calculated to obtain the deviation average value and deviation standard deviation; the deviation fluctuation coefficient is obtained by calculating the ratio of the deviation standard deviation to the deviation average value; between two adjacent monitoring times within the monitoring period, the time interval between the two adjacent monitoring times is obtained, and the absolute difference between the two relative deviation values ​​is divided by the product of the deviation average value and the time interval to obtain the deviation change rate; the sum of 1 and the product of the deviation fluctuation coefficient and the deviation change rate coefficient is logarithmically calculated to obtain the physical consistency constraint degree; for each monitored quantity, the relative deviation value is divided by the physical consistency constraint degree to obtain the health level, the health levels of all monitored quantities are obtained, and the mean is calculated to obtain the structural health status index.

[0009] Preferably, the steps for obtaining the structural impact index are as follows: During the monitoring period, obtain the acceleration and equivalent vibration mass before and after maintenance and reinforcement; calculate the structural stiffness recovery coefficient based on the acceleration and equivalent vibration mass before and after maintenance and reinforcement; during the monitoring period, obtain the strain sequence before and after maintenance and reinforcement; calculate the strain response offset coefficient based on the strain sequence before and after maintenance and reinforcement; during the monitoring period, obtain the seepage pressure sequence and theoretical seepage pressure sequence before and after maintenance and reinforcement; calculate the seepage response offset coefficient based on the seepage pressure sequence before and after maintenance and reinforcement and the theoretical seepage pressure sequence; normalize the structural stiffness recovery coefficient, strain response offset coefficient, and seepage response offset coefficient; and calculate the structural impact index based on the normalized structural stiffness recovery coefficient, strain response offset coefficient, and seepage response offset coefficient. The specific steps are as follows: In the formula, This is expressed as an index indicating the impact on the structure. This is expressed as the normalized structural stiffness restitution coefficient. This is expressed as the strain response offset coefficient after normalization. This is expressed as the normalized seepage response offset coefficient. , , It represents the weighting coefficients of the normalized structural stiffness recovery coefficient, the normalized strain response offset coefficient, and the normalized seepage response offset coefficient.

[0010] Preferably, the steps for obtaining the structural stiffness restoration coefficient are as follows: During the monitoring period, the acceleration before and after maintenance and reinforcement at the same measuring point are obtained respectively. Spectral analysis is performed on the acceleration before and after maintenance and reinforcement to obtain the spectral amplitude distribution. The first-order dominant frequency is read at the low-order peak with the largest spectral amplitude to obtain the first-order dominant frequency before and after maintenance. The equivalent vibration mass corresponding to the measuring point is obtained from the finite element physical model to obtain the equivalent vibration mass before and after maintenance. The equivalent stiffness before maintenance is calculated based on the first-order dominant frequency before maintenance and the equivalent vibration mass before maintenance. The equivalent stiffness after maintenance is calculated based on the first-order dominant frequency after maintenance and the equivalent vibration mass after maintenance. The absolute difference between the equivalent stiffness after maintenance and the equivalent stiffness before maintenance is calculated, and then the ratio is calculated with the equivalent stiffness before maintenance to obtain the structural stiffness restoration coefficient.

[0011] Preferably, the steps for obtaining the strain response offset coefficient are as follows: within the monitoring period, select the strain sequence of the same strain gauge after pre-treatment before maintenance and reinforcement and the strain sequence after pre-treatment after maintenance and reinforcement; calculate the average value of the strain sequence after pre-treatment before maintenance and reinforcement by taking the absolute value; calculate the average value of the strain sequence after pre-treatment after maintenance and reinforcement by taking the absolute value; calculate the ratio of the absolute difference between the average strain amplitude after maintenance and the average strain amplitude before maintenance to the average strain amplitude before maintenance to obtain the strain response offset coefficient.

[0012] Preferably, the steps for obtaining the seepage response offset coefficient are as follows: Within the monitoring period, for the same monitoring point, the seepage pressure sequence before maintenance and reinforcement and the seepage pressure sequence after maintenance and reinforcement are obtained respectively, and the theoretical seepage pressure data sequence output by the finite element physical model under the same working conditions is also obtained; the seepage pressure value in the seepage pressure sequence before maintenance and reinforcement is divided by the theoretical seepage pressure value at the same time to obtain the seepage pressure ratio sequence before maintenance; similarly, the seepage pressure value in the seepage pressure sequence after maintenance and reinforcement is divided by the theoretical seepage pressure value to obtain the seepage pressure ratio sequence after maintenance; the median of the seepage pressure ratio sequence before maintenance and the seepage pressure ratio sequence after maintenance are calculated respectively to obtain the seepage pressure calibration factor before maintenance and the seepage pressure calibration factor after maintenance; the absolute difference between the calibration factor after maintenance and the calibration factor before maintenance is calculated and divided by the calibration factor before maintenance to obtain the seepage response offset coefficient.

[0013] Preferably, the step of determining whether the current structure is affected by maintenance and reinforcement based on the structural impact index is as follows: compare the structural impact index with the impact threshold; if the structural impact index is greater than or equal to the impact threshold, it is determined that the current structure is affected by structural maintenance and reinforcement; if the structural impact index is less than the impact threshold, it is determined that the current structure is not affected by structural maintenance and reinforcement.

[0014] Preferably, the step of obtaining the corrected health threshold is as follows: Obtain the initial health threshold; within the monitoring period, obtain the equivalent stiffness, average strain amplitude, and average seepage pressure before and after maintenance and reinforcement, respectively, to obtain the pre-equivalent stiffness, post-equivalent stiffness, pre-average strain amplitude, post-average strain amplitude, pre-average seepage pressure, and post-average seepage pressure; subtract the pre-equivalent stiffness from the post-equivalent stiffness and divide by the pre-equivalent stiffness to obtain the equivalent stiffness direction ratio; subtract the pre-average strain amplitude from the post-average strain amplitude and divide by the pre-average strain amplitude. The strain direction ratio is obtained. The pressure direction ratio is obtained by subtracting the previous average seepage pressure from the subsequent average seepage pressure and dividing by the previous average seepage pressure. The absolute values ​​of the equivalent stiffness direction ratio, strain direction ratio, and pressure direction ratio are taken to obtain the equivalent stiffness amplitude coefficient, strain amplitude coefficient, and pressure amplitude coefficient. The affected index amplitude is calculated based on these coefficients. The sign function of the equivalent stiffness amplitude coefficient is calculated, and the negative value is taken to obtain the stiffness direction sign. The sign function of the strain amplitude coefficient is then calculated. The process involves calculating the strain direction sign, performing a sign function on the permeability amplitude coefficient to obtain the permeability direction sign, and determining that if at least two of the stiffness, strain, and permeability direction signs are positive, the health threshold needs to be increased; if at least two of the stiffness, strain, and permeability direction signs are negative, the health threshold needs to be decreased; otherwise, the three signs are deemed not majority-represented, and the health threshold remains unchanged. If the health threshold needs to be increased, the initial health threshold is multiplied by 1 plus the magnitude of the affected exponent to obtain the corrected health threshold. If the health threshold needs to be decreased, the initial health threshold is compared with the value of 1 plus the magnitude of the affected exponent. If the three signs are deemed not majority-represented, the initial health threshold is not corrected, and it is directly used as the corrected health threshold. The historical health period quantile limit is obtained as the boundary for increasing or decreasing the health threshold. When the corrected health threshold exceeds the boundary, the nearest boundary value is used as the corrected health threshold.

[0015] Preferably, the step of determining whether the current structure is in a healthy state based on the health status index and the corrected health threshold is as follows: compare the health status index with the corrected health threshold; if the health status index is greater than or equal to the corrected health threshold, then the current structure is determined to be in a healthy state; if the health status index is less than the corrected health threshold, then the current structure is determined to be in a unhealthy state.

[0016] The technical effects and advantages of this invention are as follows: Based on the preprocessed monitoring data and theoretical response parameters, the structural health status index of the water conservancy project is obtained, structural performance recovery characteristic parameters are acquired, and the structural impact index is obtained based on the structural performance recovery characteristic parameters. Based on the structural impact index, it is determined whether the current structure is affected by maintenance and reinforcement. If it is determined that the current structure is affected by maintenance and reinforcement, the initial health threshold is corrected to obtain the corrected health threshold. Based on the health status index and the corrected health threshold, it is determined whether the current structure is in a healthy state, which effectively improves the accuracy of the health status perception of the project. Attached Figure Description

[0017] Figure 1 A flowchart illustrating the intelligent sensing method for the health status of water conservancy projects that incorporates physical model constraints, as provided in this application embodiment. Detailed Implementation

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The intelligent sensing method for the health status of water conservancy projects that integrates physical model constraints involved in the present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] This invention provides an intelligent sensing method for the health status of hydraulic engineering projects that integrates physical model constraints, such as... Figure 1 As shown, it includes the following steps: Step 1: Install strain gauges, piezometers, water level gauges and acceleration sensors at key parts of the water conservancy project to collect real-time monitoring data on the structural operation of the water conservancy project. The monitoring data include strain, seepage pressure, water level and acceleration, forming a monitoring dataset. Step 2: Preprocess the monitoring dataset to obtain preprocessed monitoring data. The preprocessing involves using wavelet packet decomposition to extract low-frequency trend components and using sliding median filtering to smooth high-frequency noise, effectively eliminating abnormal fluctuations caused by environmental interference or sensor errors during the monitoring process, and providing a clean and continuous signal basis for subsequent feature extraction. It should be noted that the use of wavelet packet decomposition to extract low-frequency trend components and the use of sliding median filtering to smooth high-frequency noise are existing technologies, and this embodiment will not provide a detailed description of their specific steps.

[0020] Step 3: Based on the preprocessed monitoring data, key feature indicators are calculated, including root mean square value, peak factor, energy entropy, dominant frequency, and frequency band energy ratio. These features comprehensively reflect the energy distribution, abrupt change, and spectral characteristics of the monitored quantities, thereby transforming the complex raw signal into a quantifiable feature matrix, laying the data foundation for subsequent matching with the physical model; In this embodiment, it should be specifically explained that the steps for obtaining key feature indicators are as follows: For each monitored quantity (strain, seepage pressure, water level, acceleration), firstly, all sampling point data of the signal are acquired within the monitoring period to obtain the monitoring data sequence, and then the root mean square value is calculated on the monitoring data sequence to reflect the overall energy level of the signal. The larger the value, the higher the signal energy and the stronger the structural response amplitude. The monitoring period refers to the time interval within which a complete set of data is collected and analyzed during the continuous operation of a hydraulic engineering structure. This time interval is determined by the sampling frequency and data storage strategy set by the system, ensuring the comparability and synchronicity of various monitored quantities at the same time scale. The length of the monitoring period can be configured according to the structure type and operational characteristics, such as dividing it by hour, day, or week, to ensure temporal continuity and representativeness during statistical analysis and feature extraction.

[0021] In the monitoring data sequence of each monitoring quantity, the maximum absolute value of the signal is extracted as the peak value. The peak value is divided by the root mean square value of the signal to obtain the peak factor, which reflects the strength of sudden impact or abnormal vibration in the signal. The larger the value, the more obvious the abrupt change or pulse component in the signal. For each monitored quantity, energy decomposition is performed to obtain the energy distribution of the signal at each frequency component. The energy entropy is obtained by summing the negative products of the energy proportion of each frequency band and its natural logarithm. The energy entropy is used to characterize the uniformity of the signal energy distribution; the larger the value, the more dispersed the energy distribution and the higher the signal complexity. It should be noted that energy decomposition of monitored quantities to obtain the energy distribution at different frequency components is a prior art. This process can be implemented using publicly available signal analysis methods such as short-time Fourier transform, wavelet transform, or wavelet packet decomposition to calculate the energy proportion of the signal in each frequency band. Perform a fast Fourier transform on each monitoring quantity to obtain the spectral amplitude distribution, identify the frequency component with the largest amplitude as the dominant frequency, which is used to characterize the dominant frequency features of structural vibration or seepage response, and can reflect the structural characteristic stiffness and vibration mode changes. The Fast Fourier Transform (FFT) is a mathematical analysis method that converts discrete-time series signals into a frequency domain representation. Its basic principle is to optimize the computation process of the Discrete Fourier Transform using a divide-and-conquer approach, decomposing the time-domain signal into several frequency components, thereby significantly reducing computational complexity and improving computational efficiency. The FFT can be used to obtain the amplitude and phase spectra of the monitored quantity at different frequencies, which can be used to identify the dominant frequency component and energy distribution characteristics. Since the FFT algorithm is existing technology, and its mathematical principles and implementation methods have been widely disclosed and applied in the field of signal processing, this embodiment will not further explain its algorithm details.

[0022] Based on the spectral amplitude distribution, the target frequency band energy and the reference frequency band energy are obtained. The ratio of the target frequency band energy to the reference frequency band energy is calculated to obtain the frequency band energy ratio, which reflects the degree of energy concentration of the structure in a specific response frequency band. The target frequency band energy refers to the total energy calculated in the frequency range where the energy is concentrated and the amplitude changes significantly in the spectrum of the monitored quantity. This frequency band reflects the main response components of the signal. The reference frequency band energy refers to the total energy calculated in the frequency range where the energy distribution is relatively stable and the amplitude changes are small in the signal spectrum. It is used as a reference for energy changes.

[0023] Step 4: Based on the hydraulic structure design drawings and material parameters, establish a finite element physical model. Obtain theoretical response parameters through the finite element physical model. The theoretical response parameters include theoretical strain, theoretical seepage pressure, theoretical water level, and theoretical acceleration. It should be noted that establishing a finite element physical model based on the hydraulic structure design drawings and material parameters is an existing technology, and this embodiment will not provide a detailed description of its specific steps. The finite element method (FEM) physical model refers to a numerical model based on structural mechanics and seepage theory. It discretizes a continuous hydraulic engineering structure into several finite elements, establishes equilibrium equations at each element node, and solves for stress, strain, displacement, and seepage distribution within the structure by considering material parameters, boundary conditions, and external loads. This model can simulate the theoretical response of the structure under different working conditions, provided that the design drawings and material performance parameters are known. It offers a physical reference for monitoring data and forms the basis for structural health status assessment and physical consistency analysis.

[0024] Step 5: Based on the pre-processed monitoring data and theoretical response parameters, the structural health status index of the water conservancy project is obtained, and the structural performance recovery characteristic parameters are acquired. The structural performance recovery characteristic parameters include equivalent stiffness, strain response amplitude and seepage pressure response. Based on the structural performance recovery characteristic parameters, the structural impact index is obtained, and based on the structural impact index, it is determined whether the current structure is affected by maintenance and reinforcement. In this embodiment, it should be specifically explained that the steps for obtaining the structural health status index are as follows: For each monitored quantity (such as strain, seepage pressure, water level, acceleration), the absolute difference between the monitored value and the corresponding theoretical value is subtracted, and then divided by the theoretical value to obtain the relative deviation value. Within a continuous monitoring period, the relative deviation values ​​are arranged in chronological order to form a deviation value sequence. The mean and standard deviation of the deviation value sequence are calculated to obtain the deviation mean and deviation standard deviation. The deviation fluctuation coefficient is obtained by calculating the ratio of the deviation standard deviation to the deviation mean. Within the monitoring period, the time interval between two adjacent monitoring moments is obtained. The absolute difference between the two relative deviation values ​​is divided by the product of the average deviation and the time interval to obtain the rate of change of deviation. The physical consistency constraint degree is obtained by summing the product of 1 and the deviation fluctuation coefficient and the deviation change rate coefficient, and then performing logarithmic calculation on the summation. For each monitoring quantity, the relative deviation value is divided by the physical consistency constraint to obtain the health level. The health levels of all monitoring quantities are obtained and the average value is calculated to obtain the structural health status index.

[0025] By comparing the differences between monitoring data and theoretical model responses, and characterizing the consistency of structural operation through stability and rate of change, a quantitative assessment of structural health can be achieved. This comprehensively reflects the actual stress and seepage state of the structure.

[0026] In this embodiment, it should be specifically explained that the steps for obtaining the structural impact index are as follows: During the monitoring period, the acceleration and equivalent vibration mass before and after maintenance and reinforcement are obtained, and the structural stiffness recovery coefficient is calculated based on the acceleration and equivalent vibration mass before and after maintenance and reinforcement. During the monitoring period, the strain sequence before and after maintenance and reinforcement is obtained, and the strain response offset coefficient is calculated based on the strain sequence before and after maintenance and reinforcement. During the monitoring period, the seepage pressure sequence before and after maintenance and reinforcement and the theoretical seepage pressure sequence are obtained, and the seepage response offset coefficient is calculated based on the seepage pressure sequence before and after maintenance and reinforcement and the theoretical seepage pressure sequence. The structural stiffness recovery coefficient, strain response offset coefficient, and seepage response offset coefficient are normalized. Specifically, in this embodiment, vector normalization can be used to normalize these coefficients. This involves constructing a three-dimensional vector from these three coefficients, calculating the square root of the sum of the squares of each component to obtain the norm of the vector, and then dividing each coefficient by this norm to complete the normalization process. The purpose of this normalization method is to ensure that the structural stiffness recovery coefficient, strain response offset coefficient, and seepage response offset coefficient have a uniform order of magnitude and scale standard when calculating the structural impact index, avoiding calculation deviations or local amplification effects caused by differences in monitoring types or response parameter value ranges. This process ensures that the three coefficients have consistent physical dimensions and sensitivity in the impact index calculation, thereby improving the calculation stability and reliability of the structural impact index. Since the vector normalization method is an existing technology and its mathematical calculation principle has been widely used in the fields of structural health monitoring and multi-source data fusion analysis, this embodiment will not further elaborate on its specific algorithm steps. The structural impact index is calculated based on the normalized structural stiffness recovery coefficient, strain response offset coefficient, and seepage response offset coefficient. The specific steps for obtaining this index are as follows: ; In the formula, This is expressed as an index indicating the impact on the structure. This represents the normalized structural stiffness restitution coefficient. When a structure undergoes maintenance or reinforcement, the more significant the change in overall stiffness and the higher the degree of stiffness recovery, the greater its contribution to the structural impact index. In other words, a larger stiffness restitution coefficient indicates a more significant adjustment in the structure's mechanical properties compared to before maintenance, reflecting a stronger impact of maintenance and reinforcement on the structure's dynamic characteristics, thus leading to an increase in the structural impact index. This represents the strain response offset coefficient after normalization. When a structure undergoes maintenance or reinforcement, the greater the shift in its strain response compared to before maintenance, the higher its contribution to the structural impact index. In other words, a larger strain response offset coefficient indicates a more significant change in the deformation characteristics or stress distribution of the structure under external loads compared to before maintenance. This suggests a stronger impact of maintenance and reinforcement on the local or overall mechanical behavior of the structure, thus leading to an increase in the structural impact index. This represents the normalized seepage response offset coefficient. When a structure undergoes maintenance or reinforcement, a greater offset in its seepage response compared to before maintenance indicates a higher contribution to the structural impact index. In other words, a larger seepage response offset coefficient indicates a more significant change in the seepage path, pore distribution, or pressure transfer characteristics within the structure compared to before maintenance. This suggests a more significant impact of maintenance and reinforcement on the seepage system, thus increasing the structural impact index. , , This represents the weighting coefficients of the normalized structural stiffness recovery coefficient, the normalized strain response offset coefficient, and the normalized seepage response offset coefficient, and... ,For example , , It can be 0.4, 0.3, or 0.3. , , The weights can be obtained through the Analytic Hierarchy Process (AHP), a weight determination method based on multi-index decision theory. Its basic principle is to decompose complex multi-factor problems into different levels such as the target level, criterion level, and indicator level. A judgment matrix is ​​established through pairwise comparisons, the relative importance of each factor to the upper-level target is calculated, and the weight coefficients of each indicator are obtained after consistency testing.

[0027] In this embodiment, it should be specifically explained that the steps for obtaining the structural stiffness restoration coefficient are as follows: During the monitoring period, the acceleration before and after maintenance and reinforcement at the same measuring point is acquired. Spectral analysis (such as Fast Fourier Transform) is performed on the acceleration before and after maintenance and reinforcement to obtain the spectral amplitude distribution. The first-order dominant frequency is read from the low-order peak with the largest spectral amplitude to obtain the first-order dominant frequency before and after maintenance. It should be noted that performing spectral analysis on the acceleration before and after maintenance and reinforcement to obtain the spectral amplitude distribution is existing technology, and this embodiment will not describe its specific steps in detail. The equivalent vibrational mass corresponding to the measuring point is obtained from the finite element physical model. The equivalent vibrational mass before and after maintenance are obtained respectively. The equivalent vibrational mass is calculated using the mass matrix and modal relationship of the finite element physical model. It should be noted that calculating the equivalent vibrational mass of the structure under defined vibrational modes using the mass matrix and modal relationship of the finite element physical model falls within the scope of publicly available structural modal analysis techniques. By discretizing the structure into a finite element system, solving its modal vectors, and combining the mass matrix with projection calculations on the mass contribution corresponding to each mode, the equivalent vibrational mass distribution of a specific mode is obtained. Since this technique has been widely used in the fields of structural dynamics and modal analysis, its specific derivation steps will not be further explained in this embodiment. It should be specifically noted that the equivalent vibrational mass refers to the portion of the total mass of a structure that generates effective inertial force for a particular vibration mode. The equivalent vibrational mass of a corresponding measuring point or mode can be obtained by projecting the mass matrix of the structure onto the mode shape vector.

[0028] The equivalent stiffness before maintenance is calculated based on the first-order dominant frequency and the equivalent vibration mass before maintenance. The specific steps are as follows: ; In the formula, This represents the equivalent stiffness before maintenance. This represents the first-order main frequency before maintenance. This is expressed as the equivalent vibration mass before maintenance. The equivalent stiffness after maintenance is calculated based on the first-order dominant frequency and the equivalent vibration mass after maintenance. The specific steps for obtaining this stiffness are as follows: ; In the formula, This represents the equivalent stiffness after maintenance. This represents the first-order main frequency after maintenance. Represented as the equivalent vibration mass after maintenance; The structural stiffness recovery coefficient is obtained by calculating the absolute difference between the equivalent stiffness after maintenance and the equivalent stiffness before maintenance, and then comparing the ratio with the equivalent stiffness before maintenance. The larger the structural stiffness recovery coefficient, the more significant the stiffness change caused by maintenance and reinforcement.

[0029] In this embodiment, it should be specifically explained that the steps for obtaining the strain response offset coefficient are as follows: Within the monitoring period, the strain sequence of the same strain gauge before and after maintenance and reinforcement was selected; The average strain amplitude before maintenance is obtained by taking the absolute value of the strain sequence after pretreatment before maintenance and reinforcement and then calculating the mean value. The average strain amplitude after maintenance is obtained by taking the absolute value of the strain sequence after pretreatment and reinforcement. The strain response offset coefficient is obtained by calculating the absolute difference between the average strain amplitude after maintenance and the average strain amplitude before maintenance, and then comparing it with the average strain amplitude before maintenance. The larger the strain response offset coefficient, the more obvious the offset of the strain response relative to the period before maintenance.

[0030] In this embodiment, it should be specifically explained that the steps for obtaining the seepage response offset coefficient are as follows: During the monitoring period, for the same monitoring point, the seepage pressure sequence before and after maintenance and reinforcement were obtained respectively. At the same time, the theoretical seepage pressure data sequence output by the finite element physical model under the same working conditions was obtained. In order to ensure comparability, all monitoring values ​​and theoretical values ​​were synchronized at the same time interval. It should be noted that "under the same working conditions" means that the external environment and operating conditions of the hydraulic engineering structure remain consistent or comparable before and after maintenance or reinforcement, including key factors such as upstream and downstream water levels, seepage boundary conditions, external load intensity, gate opening and closing status, and meteorological conditions, without significant changes.

[0031] Divide the seepage pressure value in the seepage pressure sequence before maintenance and reinforcement by the theoretical seepage pressure value at the same time to obtain the seepage pressure ratio sequence before maintenance. Similarly, divide the seepage pressure value in the seepage pressure sequence after maintenance and reinforcement by the theoretical seepage pressure value to obtain the seepage pressure ratio sequence after maintenance. The medians of the pre-maintenance pressure ratio series and the post-maintenance pressure ratio series were calculated to obtain the pre-maintenance pressure calibration factor and the post-maintenance pressure calibration factor. The median is used to reflect the typical proportion level of the ratio series, which can effectively avoid the influence of individual outliers on the results. The absolute difference between the calibration factor after maintenance and the calibration factor before maintenance is calculated and then divided by the calibration factor before maintenance to obtain the seepage response offset coefficient. The larger the seepage response offset coefficient, the more significant the change in the seepage pressure system relative to the theoretical model.

[0032] In this embodiment, it should be specifically explained that the step of determining whether the current structure is affected by maintenance and reinforcement based on the structural impact index is as follows: The structural impact index is compared with the impact threshold. If the impact index is greater than or equal to the impact threshold, the structure is considered to be affected by structural maintenance and reinforcement, and the initial health threshold needs to be corrected. If the impact index is less than the impact threshold, the structure is considered not to be affected by structural maintenance and reinforcement, and the initial health threshold does not need to be corrected. The impact threshold is obtained through an adaptive threshold method, which is a method for dynamically adjusting the threshold based on the statistical characteristics of monitoring data. Its basic principle is to correct the original fixed threshold in real time based on the mean, standard deviation, and trend of the monitoring data over different time periods, so that the threshold can be automatically adjusted according to changes in the structural operating status and environmental conditions. This method can avoid the problem of traditional fixed thresholds failing under environmental disturbances or changes in structural parameters, thus ensuring that the impact threshold always matches the current structural operating characteristics and improving the sensitivity and stability of threshold determination.

[0033] Step 6: If it is determined that the current structure is affected by maintenance and reinforcement, obtain the initial health threshold, and correct the initial health threshold according to the initial health threshold to obtain the corrected health threshold; In this embodiment, it should be specifically explained that the steps for obtaining the corrected health threshold are as follows: The initial health threshold is obtained, and the equivalent stiffness, average strain amplitude and average seepage pressure before and after maintenance and reinforcement are obtained during the monitoring period. The equivalent stiffness before maintenance and reinforcement, the average strain amplitude before maintenance and reinforcement, the average strain amplitude after maintenance and reinforcement, the average seepage pressure before maintenance and reinforcement and the average seepage pressure after maintenance and reinforcement are obtained. It should be specifically noted that the equivalent stiffness can be obtained through an accelerometer, and the dominant frequency is identified and calculated in combination with the known equivalent mass of the structure. This is an existing technology. The dominant frequency can be obtained from the measured acceleration spectrum by fast Fourier transform, and the mass comes from the structural design parameters or model identification. The average strain amplitude comes from strain gauges installed in key parts of the structure. The strain measurements are averaged over the monitoring period. The average seepage pressure comes from piezometers installed in key parts of the structure. The average sampled data is averaged over the same monitoring period.

[0034] The equivalent stiffness direction ratio is obtained by subtracting the previous equivalent stiffness from the subsequent equivalent stiffness and then dividing by the previous equivalent stiffness. The strain direction ratio is obtained by subtracting the previous average strain amplitude from the subsequent average strain amplitude and then dividing by the previous average strain amplitude. The seepage pressure direction ratio is obtained by subtracting the previous average seepage pressure from the subsequent average seepage pressure and then dividing by the previous average seepage pressure. The signs of the equivalent stiffness direction ratio, strain direction ratio, and seepage pressure direction ratio directly indicate the change in direction. A positive sign indicates that the quantity increases after maintenance, and a negative sign indicates that it decreases. The equivalent stiffness amplitude coefficient, strain amplitude coefficient, and seepage pressure amplitude coefficient are obtained by taking the absolute values ​​of the equivalent stiffness direction ratio, strain direction ratio, and seepage pressure direction ratio. The magnitude of the affected index is then calculated based on these coefficients. The specific steps for obtaining these coefficients are as follows: ; In the formula, This represents the magnitude of the affected index, and is greater than or equal to 0. Expressed as the equivalent stiffness amplitude coefficient, Represented as strain amplitude coefficient, This is expressed as the seepage pressure amplitude coefficient; The sign of the stiffness direction is obtained by calculating the equivalent stiffness amplitude coefficient using a sign function and taking a negative value. When the equivalent stiffness direction ratio is greater than zero, it means the structure is stiffer, the background deviation should be reduced, and the health threshold should be lowered; therefore, a negative sign is used. When the equivalent stiffness direction ratio is less than zero, it means the structure is more easily excited, and the health threshold should be raised; therefore, a positive sign is used. The sign of the strain amplitude coefficient and the sign of the seepage pressure amplitude coefficient are calculated using a sign function to obtain the sign of the seepage pressure direction. When calculating the direction sign, changes in structural stiffness and the direction of threshold adjustment are inversely related. When the equivalent stiffness increases after structural repair and reinforcement, it indicates improved structural load-bearing capacity and enhanced response stability; in this case, the health threshold should be appropriately lowered. Conversely, when stiffness decreases, structural resistance declines and abnormal responses are more likely to occur; in this case, the threshold should be raised. To maintain consistency in sign definition, with a positive sign representing an increase in the threshold and a negative sign representing a decrease, the stiffness direction sign must be negative to ensure a logical correspondence between stiffness changes and threshold correction directions.

[0035] If at least two of the stiffness direction sign, strain direction sign, and seepage pressure direction sign are positive, the health threshold needs to be increased; if at least two of the stiffness direction sign, strain direction sign, and seepage pressure direction sign are negative, the health threshold needs to be decreased; otherwise, the three signs are deemed not to have majority representation, and the health threshold remains unchanged. The absence of majority representation for the three symbols refers to a situation where at least two of the symbols for stiffness, strain, and seepage pressure are identical, meaning that the trends of change in each parameter cancel each other out or are inconsistent. In this case, it is impossible to determine whether the overall health threshold should be adjusted upwards or downwards. Therefore, this is defined as a state of no majority representation, indicating that no adjustment to the health threshold should be performed at this time to ensure the stability and rationality of the threshold correction.

[0036] If it is determined that the health threshold needs to be increased, the initial health threshold is multiplied by 1 plus the magnitude of the affected index to obtain the corrected health threshold. If it is determined that the health threshold needs to be decreased, the initial health threshold is multiplied by 1 plus the magnitude of the affected index to calculate the ratio. If it is determined that the three signs do not have majority representation, the initial health threshold is not corrected, and the initial health threshold is directly used as the corrected health threshold. The historical health period quantile boundary is obtained as the boundary for increasing or decreasing the health threshold. When the corrected health threshold exceeds the boundary, the nearest boundary value is used as the corrected health threshold.

[0037] Step 7: Determine whether the current structure is in a healthy state based on the health status index and the corrected health threshold. If it is determined to be in a healthy state, return to Step 1 to continue data monitoring. If it is determined to be in a unhealthy state, issue an early warning to remind relevant personnel to carry out structural inspection and reinforcement in a timely manner. In this embodiment, it should be specifically explained that the step of determining whether the current structure is in a healthy state based on the health status index and the corrected health threshold is as follows: The health status index is compared with the corrected health threshold. If the health status index is greater than or equal to the corrected health threshold, the current structure is determined to be in a healthy state; if the health status index is less than the corrected health threshold, the current structure is determined to be in a unhealthy state.

[0038] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0039] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for intelligent sensing of the health status of hydraulic engineering projects that integrates physical model constraints, characterized in that: Includes the following steps: Step 1: Install strain gauges, piezometers, water level gauges and acceleration sensors at key parts of the water conservancy project to collect real-time monitoring data on the structural operation of the water conservancy project. The monitoring data include strain, seepage pressure, water level and acceleration, forming a monitoring dataset. Step 2: Preprocess the monitoring dataset to obtain preprocessed monitoring data; Step 3: Calculate key characteristic indicators based on the preprocessed monitoring data. Key characteristic indicators include root mean square value, peak factor, energy entropy, dominant frequency, and frequency band energy ratio. Step 4: Based on the hydraulic structure design drawings and material parameters, establish a finite element physical model and obtain the theoretical response parameters through the finite element physical model. The theoretical response parameters include theoretical strain, theoretical seepage pressure, theoretical water level, and theoretical acceleration. Step 5: Based on the pre-processed monitoring data and theoretical response parameters, the structural health status index of the water conservancy project is obtained, and the structural performance recovery characteristic parameters are acquired. The structural performance recovery characteristic parameters include equivalent stiffness, strain response amplitude and seepage pressure response. Based on the structural performance recovery characteristic parameters, the structural impact index is obtained, and based on the structural impact index, it is determined whether the current structure is affected by maintenance and reinforcement. Step 6: If it is determined that the current structure is affected by maintenance and reinforcement, obtain the initial health threshold, and correct the initial health threshold according to the initial health threshold to obtain the corrected health threshold; Step 7: Determine whether the current structure is in a healthy state based on the health status index and the corrected health threshold. If it is determined to be in a healthy state, return to Step 1 to continue data monitoring. If it is determined to be in a non-healthy state, issue an early warning.

2. The intelligent sensing method for the health status of hydraulic engineering projects that integrates physical model constraints according to claim 1, characterized in that: The steps for obtaining the key feature indicators are as follows: For each monitoring quantity, firstly, acquire all sampling point data of the signal within the monitoring period to obtain the monitoring data sequence, and then calculate the root mean square value of the monitoring data sequence; In the monitoring data sequence of each monitoring quantity, the maximum absolute value of the signal is extracted as the peak value, and the peak value is divided by the root mean square value of the signal to obtain the peak factor. For each monitored quantity, energy decomposition is performed to obtain the energy distribution of the signal under each frequency component. The energy entropy is obtained by summing the negative product of the energy proportion of each frequency band and its natural logarithm. Perform a fast Fourier transform on each monitored quantity to obtain the spectral amplitude distribution, and identify the frequency component with the largest amplitude as the dominant frequency; Based on the spectral amplitude distribution, the target frequency band energy and the reference frequency band energy are obtained. The ratio of the target frequency band energy to the reference frequency band energy is calculated to obtain the frequency band energy ratio.

3. The intelligent sensing method for the health status of hydraulic engineering projects integrating physical model constraints according to claim 1, characterized in that, The steps for obtaining the structural health status index are as follows: For each monitoring quantity, the absolute difference between the monitored value and the corresponding theoretical value is subtracted, and then divided by the corresponding theoretical value to obtain the relative deviation value. Within a monitoring period, the relative deviation values ​​are arranged in chronological order to form a deviation value sequence, and the mean and standard deviation of the deviation value sequence are calculated to obtain the deviation mean and deviation standard deviation. The deviation fluctuation coefficient is obtained by calculating the ratio of the deviation standard deviation to the deviation mean. Within the monitoring period, the time interval between two adjacent monitoring moments is obtained. The absolute difference between the two relative deviation values ​​is divided by the product of the average deviation and the time interval to obtain the rate of change of deviation. The physical consistency constraint degree is obtained by summing the product of 1 and the deviation fluctuation coefficient and the deviation change rate coefficient, and then performing logarithmic calculation on the summation. For each monitoring quantity, the relative deviation value is divided by the physical consistency constraint to obtain the health level. The health levels of all monitoring quantities are obtained and the average value is calculated to obtain the structural health status index.

4. The intelligent sensing method for the health status of hydraulic engineering projects integrating physical model constraints according to claim 1, characterized in that, The steps for obtaining the structural impact index are as follows: During the monitoring period, the acceleration and equivalent vibration mass before and after maintenance and reinforcement are obtained, and the structural stiffness recovery coefficient is calculated based on the acceleration and equivalent vibration mass before and after maintenance and reinforcement. During the monitoring period, the strain sequence before and after maintenance and reinforcement is obtained, and the strain response offset coefficient is calculated based on the strain sequence before and after maintenance and reinforcement. During the monitoring period, the seepage pressure sequence before and after maintenance and reinforcement and the theoretical seepage pressure sequence are obtained, and the seepage response offset coefficient is calculated based on the seepage pressure sequence before and after maintenance and reinforcement and the theoretical seepage pressure sequence. The structural stiffness recovery coefficient, strain response offset coefficient, and seepage response offset coefficient are normalized. Based on these normalized coefficients, the structural impact index is calculated. The specific steps are as follows: ； In the formula, This is expressed as an index indicating the impact on the structure. This is expressed as the normalized structural stiffness restitution coefficient. This is expressed as the strain response offset coefficient after normalization. This is expressed as the normalized seepage response offset coefficient. , , It represents the weighting coefficients of the normalized structural stiffness recovery coefficient, the normalized strain response offset coefficient, and the normalized seepage response offset coefficient.

5. The intelligent sensing method for the health status of hydraulic engineering projects integrating physical model constraints according to claim 4, characterized in that: The steps for obtaining the structural stiffness restoration coefficient are as follows: During the monitoring period, the acceleration before and after maintenance and reinforcement at the same measuring point were obtained respectively. The spectral analysis of the acceleration before and after maintenance and reinforcement was performed to obtain the spectral amplitude distribution. The first-order main frequency was read at the low-order peak with the largest spectral amplitude to obtain the first-order main frequency before and after maintenance. The equivalent vibration mass corresponding to the measuring point is obtained from the finite element physical model, and the equivalent vibration mass before and after maintenance are obtained respectively. The equivalent stiffness before maintenance is calculated based on the first-order dominant frequency and the equivalent vibration mass before maintenance. The equivalent stiffness after maintenance is calculated based on the first-order dominant frequency and the equivalent vibration mass after maintenance. The structural stiffness recovery coefficient is obtained by calculating the absolute difference between the equivalent stiffness after maintenance and the equivalent stiffness before maintenance, and then comparing the ratio with the equivalent stiffness before maintenance.

6. The intelligent sensing method for the health status of hydraulic engineering projects integrating physical model constraints according to claim 4, characterized in that: The steps for obtaining the strain response offset coefficient are as follows: Within the monitoring period, the strain sequence of the same strain gauge before and after maintenance and reinforcement was selected; The average strain amplitude before maintenance is obtained by taking the absolute value of the strain sequence after pretreatment before maintenance and reinforcement and then calculating the mean value. The average strain amplitude after maintenance is obtained by taking the absolute value of the strain sequence after pretreatment and reinforcement. The strain response offset coefficient is obtained by calculating the absolute difference between the average strain amplitude after maintenance and the average strain amplitude before maintenance, and then comparing the ratio with the average strain amplitude before maintenance.

7. The intelligent sensing method for the health status of hydraulic engineering projects integrating physical model constraints according to claim 4, characterized in that: The steps for obtaining the seepage response offset coefficient are as follows: During the monitoring period, for the same monitoring point, the seepage pressure sequence before maintenance and reinforcement and the seepage pressure sequence after maintenance and reinforcement were obtained respectively. At the same time, the theoretical seepage pressure data sequence output by the finite element physical model under the same working conditions was also obtained. Divide the seepage pressure value in the seepage pressure sequence before maintenance and reinforcement by the theoretical seepage pressure value at the same time to obtain the seepage pressure ratio sequence before maintenance. Similarly, divide the seepage pressure value in the seepage pressure sequence after maintenance and reinforcement by the theoretical seepage pressure value to obtain the seepage pressure ratio sequence after maintenance. The median of the pressure ratio series before and after maintenance was obtained to obtain the pressure calibration factor before and after maintenance. The seepage response offset coefficient is obtained by calculating the absolute difference between the calibration factor after maintenance and the calibration factor before maintenance, and then dividing it by the calibration factor before maintenance.

8. The intelligent sensing method for the health status of hydraulic engineering projects integrating physical model constraints according to claim 1, characterized in that: The steps for determining whether the current structure is affected by maintenance and reinforcement based on the structural impact index are as follows: The structural impact index is compared with the impact threshold. If the structural impact index is greater than or equal to the impact threshold, it is determined that the structure is currently affected by structural maintenance and reinforcement. If the structural impact index is less than the impact threshold, it is determined that the structure is not currently affected by structural maintenance and reinforcement.

9. The intelligent sensing method for the health status of hydraulic engineering projects integrating physical model constraints according to claim 5, characterized in that: The steps for obtaining the corrected health threshold are as follows: The initial health threshold is obtained, and the equivalent stiffness, average strain amplitude and average seepage pressure before and after maintenance and reinforcement are obtained during the monitoring period. The equivalent stiffness before maintenance and reinforcement, the average strain amplitude before maintenance and reinforcement, the average strain amplitude after maintenance and reinforcement, the average seepage pressure before maintenance and reinforcement and the average seepage pressure after maintenance and reinforcement are obtained. The equivalent stiffness direction ratio is obtained by subtracting the previous equivalent stiffness from the subsequent equivalent stiffness and dividing by the previous equivalent stiffness. The strain direction ratio is obtained by subtracting the previous average strain amplitude from the subsequent average strain amplitude and dividing by the previous average strain amplitude. The seepage pressure direction ratio is obtained by subtracting the previous average seepage pressure from the subsequent average seepage pressure and dividing by the previous average seepage pressure. The equivalent stiffness amplitude coefficient, strain amplitude coefficient, and seepage pressure amplitude coefficient are obtained by taking the absolute values ​​of the equivalent stiffness direction ratio, strain direction ratio, and seepage pressure direction ratio. The magnitude of the affected index is then calculated based on the equivalent stiffness amplitude coefficient, strain amplitude coefficient, and seepage pressure amplitude coefficient. The sign of the stiffness direction is obtained by calculating the equivalent stiffness amplitude coefficient using a sign function and taking its negative value; the sign of the strain direction is obtained by calculating the strain amplitude coefficient using a sign function; and the sign of the seepage pressure direction is obtained by calculating the seepage pressure amplitude coefficient using a sign function. If at least two of the stiffness direction sign, strain direction sign, and seepage pressure direction sign are positive, the health threshold needs to be increased; if at least two of the stiffness direction sign, strain direction sign, and seepage pressure direction sign are negative, the health threshold needs to be decreased; otherwise, the three signs are deemed not to have majority representation, and the health threshold remains unchanged. If it is determined that the health threshold needs to be increased, the initial health threshold is multiplied by 1 plus the magnitude of the affected index to obtain the corrected health threshold; if it is determined that the health threshold needs to be decreased, the initial health threshold is multiplied by 1 plus the magnitude of the affected index to calculate the ratio. If the three symbols are determined to be non-majority, the initial health threshold will not be corrected. Instead, the initial health threshold will be directly used as the corrected health threshold. The historical health period percentile will be obtained as the boundary for adjusting the health threshold up or down. When the corrected health threshold exceeds the boundary, the nearest boundary value will be used as the corrected health threshold.

10. The intelligent sensing method for the health status of hydraulic engineering projects integrating physical model constraints according to claim 1, characterized in that: The step of determining whether the current structure is in a healthy state based on the health status index and the corrected health threshold is as follows: The health status index is compared with the corrected health threshold. If the health status index is greater than or equal to the corrected health threshold, the current structure is determined to be in a healthy state; if the health status index is less than the corrected health threshold, the current structure is determined to be in a unhealthy state.

Images