A monitoring method for intelligent correction of multi-source information data of a coastal city

Abstract

The application discloses a kind of coastal city intelligent correction multi-source information data monitoring method, by obtaining the structure monitoring data in the long-term operation process of structure and the tidal level data in corresponding time period;Pretreatment is carried out;Frequency domain analysis is carried out, and dominant frequency of tide is determined;Optimal delay time that tidal level change acts on structure response is determined using cross-correlation analysis method;Time-varying harmonic model is used to establish time-varying tidal response model, and tidal influence term is obtained, and sliding time window least square method is used to dynamically update time-varying tidal response model parameters;Tidal influence term is separated, and the structure monitoring data after tidal influence correction is obtained;Long-term trend is extracted to the structure monitoring data after tidal influence correction, and the long-term service response characteristic quantity of structure is obtained, which can effectively improve the characterization ability of monitoring data to the true service state of structure, to improve the accuracy and reliability of long-term health monitoring and safety evaluation of coastal infrastructure.

Description

Technical Field

[0001] This application relates to the field of structural health monitoring technology, and in particular to a method for monitoring multi-source information data with intelligent correction in coastal cities. Background Technology

[0002] With the continuous expansion of coastal urban infrastructure, long-term service structures such as bridges, tunnels, underground utility tunnels, port projects, seawalls, and marine engineering structures face complex marine environmental conditions, making their service safety and durability increasingly important. To achieve real-time perception of structural operating status and long-term safety assessment, structural health monitoring technology has been widely applied in the operation and maintenance of coastal infrastructure.

[0003] Existing structural health monitoring systems typically consist of monitoring sensors, data acquisition equipment, data transmission equipment, and a data processing platform. The monitoring sensors include strain gauges, displacement gauges, inclinometers, water pressure sensors, vibration sensors, GNSS (Global Navigation Satellite System) monitoring equipment, and fiber optic sensors, used to acquire real-time information on structural strain, displacement, settlement, tilt, water pressure, and vibration. The data acquisition equipment is responsible for synchronously acquiring the output signals from each monitoring unit. The data transmission equipment transmits the monitoring data to the data processing platform. The data processing platform then stores, analyzes, and performs safety assessments on the monitoring data.

[0004] Existing monitoring data processing methods mainly include temperature effect compensation, random noise filtering, outlier identification, and sensor drift correction. Commonly used methods include moving average filtering, wavelet analysis, empirical mode decomposition, frequency domain filtering, and regression analysis. By processing the monitoring data in the time or frequency domain, the stability and interpretability of the monitoring data can be improved.

[0005] However, for infrastructure in coastal environments, monitoring data is significantly affected not only by factors such as vehicle loads, wind loads, and temperature changes, but also by the effects of tidal environments. Existing research shows that tidal changes cause fluctuations in groundwater levels, changes in pore water pressure, and a redistribution of effective stress in the foundation, leading to periodic strain, settlement, displacement, or tilt responses in structures. This type of tidal-induced response is particularly pronounced in underground structures, port engineering, seawalls, and near-shore infrastructure.

[0006] Furthermore, tidal changes can alter the environmental field surrounding the sensor, including local temperature and humidity, water pressure, and soil constraint conditions, resulting in low-frequency oscillations in the monitoring system output corresponding to the tidal cycle. For long-term continuous monitoring systems, this type of low-frequency periodic response can easily couple with the actual service response of the structure, leading to long-term trend drift, misjudgment of abnormal conditions, and decreased accuracy in structural safety assessments.

[0007] Currently, existing technologies for handling tidal environmental effects still have shortcomings. Some methods employ fixed-period filtering or simple empirical models to address tidal influences, but due to the significant time-varying, regional, and non-stationary characteristics of tidal forces, traditional fixed-parameter methods struggle to accurately describe the dynamic characteristics of tidal responses as environmental conditions change. Furthermore, factors such as groundwater seepage and diffusion, pore water pressure propagation, and structural response inertia cause time delays between tidal level changes and structural responses, making it difficult for traditional synchronous filtering methods to effectively separate tidal influences from the actual structural response.

[0008] Furthermore, traditional frequency domain filtering methods typically weaken low-frequency response components directly, which can easily lead to the loss of true long-term structural response information while eliminating tidal effects, thus affecting the results of long-term structural trend analysis and damage identification. Therefore, existing technologies struggle to simultaneously achieve both the accuracy of tidal effect correction and the ability to preserve the true structural response.

[0009] Therefore, there is an urgent need to propose a monitoring method for intelligent correction of multi-source information data in coastal cities, so as to effectively separate the tidal environmental effects from the actual service response of the structure, improve the accuracy, stability and long-term trend identification ability of long-term monitoring data, and provide a reliable data foundation for the health monitoring and safety assessment of coastal infrastructure structures. Summary of the Invention

[0010] This invention aims to provide a monitoring method for intelligent correction of multi-source information data in coastal cities. It addresses the problem in existing technologies where changes in tidal levels cause fluctuations in groundwater levels, changes in pore water pressure, and redistribution of foundation stress, resulting in significant low-frequency periodic oscillations in structural monitoring data. This affects the stability and accuracy of long-term monitoring data, easily leading to long-term trend drift of structures, misjudgment of abnormal states, and decreased accuracy of safety assessments.

[0011] To solve the above-mentioned technical problems, the present invention provides the following technical solution: A method for monitoring multi-source information data with intelligent correction in coastal cities includes: Step 1: Obtain structural monitoring data and tide level data for the corresponding time period during the long-term operation of the structure; Step 2: Preprocess the structural monitoring data; Step 3: Perform frequency domain analysis on the preprocessed structural monitoring data to determine the dominant tidal frequency; Step 4: Use cross-correlation analysis to study the time delay relationship between tidal level changes and structural response. By searching for the time delay corresponding to the maximum correlation coefficient, determine the optimal delay time for tidal level changes to affect the structural response. Step 5: Based on the dominant tidal frequency and the optimal delay time, a time-varying tidal response model is established using a time-varying harmonic model to obtain the tidal influence term. The parameters of the time-varying tidal response model are then dynamically updated using the sliding time window least squares method. Step 6: Separate the tidal influence term from the preprocessed structural monitoring data to obtain the tidal influence corrected structural monitoring data; Step 7: Extract long-term trends from the tidal-affected structural monitoring data to obtain the structural long-term service response characterization quantities.

[0012] Preferably, in the above-mentioned monitoring method for intelligent correction of multi-source information data in coastal cities, the structural monitoring data is obtained through a structural health monitoring platform. The structural health monitoring platform includes a data monitoring unit for acquiring structural monitoring data, a data acquisition unit, a tide level data acquisition unit for acquiring tide level data, a data transmission unit, and a data processing platform. The data monitoring unit includes any one or more of the following: a strain monitoring unit, a displacement monitoring unit, a tilt angle monitoring unit, a water pressure monitoring unit, and a temperature monitoring unit. The data acquisition unit enables synchronous acquisition by each data monitoring unit, and the data transmission unit enables real-time transmission of structural monitoring data and tide level data to the data processing platform.

[0013] Preferably, in the above-mentioned monitoring method for intelligent correction of multi-source information data in coastal cities, step 2, the preprocessing of structural monitoring data includes: establishing a unified time axis through time synchronization for structural monitoring data with different sampling frequencies; using linear interpolation to complete missing data points in the structural monitoring data; using statistical discrimination methods to identify and remove outliers in the structural monitoring data; and performing noise reduction processing on the structural monitoring data.

[0014] Preferably, in the above-mentioned monitoring method for intelligent correction of multi-source information data in coastal cities, in step 3, a fast Fourier transform is used to perform frequency domain analysis on the preprocessed structural monitoring data.

[0015] Preferably, in the above-mentioned intelligent correction monitoring method for multi-source information data in coastal cities, in step 3, the energy distribution of structural monitoring data in different frequency ranges is analyzed by power spectral density function, and the tidal dominant frequency is determined by searching for the peak value of the power spectrum.

[0016] Preferably, in the above-mentioned monitoring method for intelligent correction of multi-source information data in coastal cities, in step 4, when studying the time-delay relationship between tidal level change and structural response using cross-correlation analysis, the time-delay cross-correlation function between tidal level change and structural response is defined as: in, For structural monitoring data and tide data in time lag Normalized cross-correlation coefficients under the given conditions; The discrete time delay index is... Furthermore, a positive time delay is defined to represent the structural response lagging behind the tide level; Let A be the set of sampling points where both sets of data have valid values ​​under the corresponding time delay conditions. ; and These are the mean structural response and mean tide level for the corresponding valid data segments, respectively; For the first Preprocessed structural monitoring data at each sampling time point For the first Tide level data at each sampling time Optimal Discrete Time Delay for: in, The preset time delay search interval; According to the sampling interval The optimal discrete time delay Optimal delay time for structural response to tidal changes : .

[0017] Preferably, in the above-mentioned monitoring method for intelligent correction of multi-source information data in coastal cities, in step 5, based on the dominant tidal frequency and the optimal delay time, a time-varying tidal response model is established using a time-varying harmonic model to obtain the tidal influence term: in: This refers to the tidal influence term calculated using a time-varying tidal response model. For the first The time-varying tidal response amplitude at each sampling time; The dominant frequency is tidal. The sampling time interval; The optimal delay time for the structural response to tidal changes This is the residual time-varying phase correction amount used to characterize factors such as local environmental state and model errors.

[0018] Preferably, in the above-mentioned intelligent correction method for monitoring multi-source information data in coastal cities, step 5 considers the modulation effect of environmental factors on the intensity of tidal response and establishes an environmental coupling amplitude model: in, For the first The ambient temperature at each sampling time For the first Rainfall at each sampling time or within the corresponding statistical time scale For the first Groundwater level at each sampling time , , , These are the parameters for the environmental coupling amplitude model.

[0019] Preferably, in the above-mentioned monitoring method for intelligent correction of multi-source information data in coastal cities, after obtaining the tidal-affected structural monitoring data in step 6 and before step 7, the following steps are also included: Discrete Fourier transform is performed on the tidal-corrected structural monitoring data to obtain the tidal-corrected frequency domain data. Based on the tidal dominant frequency identified in the aforementioned steps, a tidal dominant frequency band blocking filter function is constructed. The frequency domain data after tidal influence correction is processed using the tidal main band stop filter function to obtain the frequency domain monitoring data after tidal main band stop processing. The frequency domain monitoring data after tidal main frequency band blocking was subjected to inverse discrete Fourier transform and converted back to the time domain to obtain structural monitoring data with further weakened residual tidal periodic components, which served as the structural monitoring data after tidal influence correction.

[0020] Preferably, in the above-mentioned monitoring method for intelligent correction of multi-source information data in coastal cities, in step 7, the long-term trend extraction is performed on the structural monitoring data after further weakening the residual tidal periodic component to obtain the structural long-term service response characterization quantity.

[0021] As can be seen from the above-disclosed technical solutions, the beneficial effects of the present invention compared with the prior art are as follows: This invention establishes a dynamic coupling relationship between tidal level changes and structural response (i.e., a time-varying tidal response model based on the dominant tidal frequency and optimal delay time), and combines this with frequency domain dominant frequency identification, cross-correlation time delay analysis, and time-varying harmonic modeling to achieve effective separation between tidal environmental effects and the actual structural response. Compared to traditional fixed-period filtering methods, this invention considers the time-varying nature of tidal effects and environmental coupling characteristics, improving the stability and accuracy of tidal influence correction in complex marine environments.

[0022] After correction for tidal effects, the periodic low-frequency fluctuations in the structural monitoring data are significantly reduced, the long-term trend is more stable, and the true deformation characteristics of the structure are clearer, which is beneficial to improving the accuracy of structural damage identification, anomaly early warning, and long-term safety assessment. Meanwhile, this invention employs a combined time-domain dynamic modeling and frequency-domain filtering approach, which, while weakening the tidal effects, can better preserve the true structural response information, avoiding the problem of losing effective structural response caused by traditional filtering methods.

[0023] Furthermore, this invention does not rely on a specific type of sensor and can be applied to monitoring data of various types of structures, such as strain, displacement, settlement, tilt angle, water pressure, and vibration. It is compatible with various monitoring devices, including resistive sensors, vibrating wire sensors, microelectromechanical systems (MEMS) sensors, global navigation satellite system monitoring equipment, and fiber optic sensing systems, and has good versatility and engineering application value. Attached Figure Description

[0024] Figure 1 This is a flowchart illustrating a monitoring method for intelligent correction of multi-source information data in coastal cities. Detailed Implementation

[0025] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The technical content and features of the present invention will be described in detail below with reference to the listed embodiments and the accompanying drawings. It should be noted that the drawings are all in a very simplified form and use non-precise proportions, and are only used to facilitate and clarify the illustration of the embodiments of the present invention. For ease of description, the terms "upper" and "lower" used below are consistent with the upper and lower directions in the accompanying drawings, but this should not be construed as a limitation of the technical solution of the present invention.

[0026] Please see Figure 1 This embodiment discloses a monitoring method for intelligent correction of multi-source information data in coastal cities, including: Step 1: Obtain structural monitoring data and tide level data for the corresponding time period during the long-term operation of the structure; Preferably, in step 1, the structural monitoring data is obtained through a structural health monitoring platform. This platform includes a data monitoring unit for acquiring structural monitoring data, a data acquisition unit, a tide level data acquisition unit for acquiring tide level data, a data transmission unit, and a data processing platform. The data monitoring unit includes any one or more of the following: a strain monitoring unit, a displacement monitoring unit, a tilt angle monitoring unit, a water pressure monitoring unit, and a temperature monitoring unit. The data acquisition unit enables synchronous data acquisition by each monitoring unit, and the data transmission unit transmits the structural monitoring data and tide level data to the data processing platform in real time. The tide level data can originate from a tide level station, a hydrological monitoring system, or a marine environmental database.

[0027] The strain monitoring unit can use a resistance strain gauge, a vibrating wire strain gauge, or a fiber optic strain sensor; the displacement monitoring unit can use a GNSS (Global Navigation Satellite System) displacement monitoring device, a laser displacement gauge, or a displacement sensor; and the tilt monitoring unit can use a MEMS (Micro Electromechanical System) tilt meter.

[0028] The long-term structural monitoring data set is denoted as: The corresponding tidal data set is denoted as: in, For the first Each sampling time The acquired raw structural monitoring data, In order to be with the first Each sampling time The corresponding tide level data, This represents the total number of sampling points within the current data segment being analyzed. .

[0029] By establishing a unified time correspondence between structural response data and tidal level data, basic data is provided for subsequent tidal response identification and correction.

[0030] Step 2: Preprocess the structural monitoring data. Since long-term structural monitoring data often suffers from inconsistent sampling frequencies, local missing data, abnormal fluctuations, and random noise, it is necessary to perform unified preprocessing on the structural monitoring data.

[0031] Preferably, in step 2, the preprocessing of the structural monitoring data includes: 1. For structural monitoring data with different sampling frequencies, a unified time axis is established through time synchronization; 2. For missing data points in the structural monitoring data, linear interpolation is used to complete them, and the completed data is denoted as follows: When the first When the original structural monitoring data corresponding to a sampling time is missing, the two most recent valid sampling times before and after it are determined to be the first two sampling times. The and the first Each sampling time satisfies The calculation is performed using linear interpolation. in, and The first The and the first Valid original structural monitoring data corresponding to each sampling time; and Missing sampling times The sequence number of the most recent valid sampling time. When the... If there are no missing original structural monitoring data corresponding to each sampling time, then: Third, after completing the missing value completion, outliers in the structural monitoring data are identified and removed using statistical discrimination methods, and the data after outlier processing is recorded as follows: The details are as follows: First, calculate the mean of the structural monitoring data within the current analysis data segment based on the structural monitoring data after missing values ​​have been filled in: and standard deviation: For the The structural monitoring data corresponding to each sampling time, when satisfying If the sampling time is not specified, the monitoring data corresponding to that sampling time will be identified as an outlier.

[0032] For the first value that was determined to be an outlier The sampling time is determined by identifying the two most recent valid and non-abnormal sampling times before and after it as the sampling time. The and the first Each sampling time satisfies And a linear interpolation method is used to repair outliers: in, and These are the abnormal sampling times. The most recent valid and non-abnormal structural monitoring data before and after, and These are the corresponding sampling time sequence numbers.

[0033] When the If the data corresponding to a sampling time point is not identified as an outlier, the original monitoring value remains unchanged: .

[0034] IV. To reduce the impact of high-frequency random noise on the identification of low-frequency tidal features, a further smoothing filtering method (Savitzky–Golaysmoothing filter) is used to denoise the structural monitoring data: in: For the first Preprocessed structural monitoring values ​​from each sampling point; The monitored value of the corresponding sampling point within the filtering window; These are the Savitzky-Golay filter coefficients; This is the position offset relative to the current center sampling point; The width of the filter window is half its width; the total length of the filter window is... .

[0035] Therefore, through the above processing, continuous, stable, and unified time-series monitoring data can be obtained: in, For the first The preprocessed structural monitoring values ​​from each sampling point This represents the total number of sampling points within the current data segment being analyzed. .

[0036] Step 3: Use Fast Fourier Transform to perform frequency domain analysis on the preprocessed structural monitoring data, analyze the energy distribution of the structural monitoring data in different frequency ranges through the power spectral density function, and determine the tidal dominant frequency by searching for the power spectral peak. Because the effects of tidal environments are distinctly periodic, they typically manifest as low-frequency periodic oscillations in structural monitoring data. To identify the dominant tidal frequency in structural monitoring data, this invention employs Fast Fourier Transform (FFT) for frequency domain analysis of the data. in, , For the first The Fourier transform results corresponding to each frequency unit For the first One pre-processed structural monitoring value, This represents the number of sampling points within the current frequency domain analysis window. The time sampling point number, This is the frequency unit number. Let be the imaginary unit, satisfying: ; No. The actual frequency corresponding to each frequency unit is: in, The sampling frequency.

[0037] Furthermore, the energy distribution of the structural monitoring data in different frequency ranges is analyzed using the power spectral density function, which is: in, For structural monitoring signals at frequency The power spectral density at that location.

[0038] Within the preset tidal candidate frequency range Internal search power spectrum peak: in, To identify the dominant tidal frequency, The candidate frequency range for tides is determined based on contemporaneous tide level data and local tidal characteristics.

[0039] According to the dominant tidal frequency Calculate the dominant tidal period : Tidal angular frequency for: Existing marine hydrological studies have shown that the semi-diurnal tidal main cycle is approximately 12.42 hours, and the diurnal tidal main cycle is approximately 24 hours. Therefore, when the identification frequency is close to the frequency corresponding to the above cycles, it can be determined that there is a significant tidal-induced response in the structural monitoring data.

[0040] Step 4: Considering that tidal changes typically have a propagation time lag in their effect on structural response—for example, groundwater seepage and diffusion, pore water pressure transmission, and foundation stress redistribution all have time delays—a cross-correlation analysis method is used to study the time-delay relationship between tidal changes and structural response. The time-delay cross-correlation function between tidal changes and structural response is defined as follows: in, For structural monitoring data and tide data in time lag Normalized cross-correlation coefficients under the given conditions; The discrete time delay index is... And define positive time delay (i.e. This indicates that the structural response lags behind the tide level; Let A be the set of sampling points where both sets of data have valid values ​​under the corresponding time delay conditions. ; and These are the mean structural response and mean tide level of the corresponding valid data segments, respectively.

[0041] Optimal Discrete Time Delay for: in, This is the preset time delay search interval.

[0042] According to the sampling interval The optimal discrete time delay Optimal delay time for structural response to tidal changes : Step 4 can effectively reflect the propagation process of tidal action in the underground medium and the differences in sensitivity of different monitoring locations to tidal changes, thereby improving the accuracy of subsequent tidal response modeling.

[0043] Step 5, based on the dominant tidal frequency With optimal delay time A time-varying tidal response model was established using a time-varying harmonic model, and the tidal influence term was obtained: in: This refers to the tidal influence term calculated using a time-varying tidal response model. For the first The time-varying tidal response amplitude at each sampling time; The sampling time interval; This is the residual time-varying phase correction amount used to characterize factors such as local environmental state and model errors.

[0044] Considering the modulating effect of environmental factors on the intensity of tidal response, an environmental coupling amplitude model is established to determine the first element in the time-varying tidal response model. Time-varying tidal response amplitude at each sampling time : in, For the first The ambient temperature at each sampling time For the first Rainfall at each sampling time or within the corresponding statistical time scale For the first Groundwater level at each sampling time , , , These are the parameters for the environmental coupling amplitude model.

[0045] Subsequently, the parameters of the time-varying tidal response model are dynamically updated using the sliding time window least squares method. The specific method is as follows: Let the first The sliding analysis window is as follows: in: The sliding window number. ; This represents the total number of sliding time windows that can be constructed within the current complete data segment for analysis. For the first The starting sampling point of each window; The number of sampling points contained in each sliding window.

[0046] The actual time corresponding to the window length is: Preferred satisfaction: For the A sliding window defines the set of parameters for the tidal model: The parameters are dynamically updated using the sliding time window least squares method: in: For the first Optimal tidal model parameters within a sliding window; For the first The optimal local baseline term obtained by optimization within a sliding time window; For the first The optimal local trend slope obtained by optimization within a sliding time window; This is the local baseline item for the structural monitoring data within this window; This represents the slope of the local long-term trend within the window. This refers to the tidal influence term calculated under given model parameters. It is only used to prevent the long-term trend of the real structure from being absorbed by the tidal model, and will not be deducted in subsequent tidal corrections.

[0047] Step 6: After establishing the time-varying tidal response model and obtaining the tidal influence terms corresponding to each sampling time, the tidal influence terms are separated from the preprocessed structural monitoring data to obtain the tidal influence-corrected structural monitoring data. in, For the first Structural monitoring values ​​at each sampling time after tidal correction; For the first Preprocessed structural monitoring values ​​at each sampling time; The first time obtained from the time-varying tidal response model Tidal influence term corresponding to each sampling time; The sampling time number, and ; This represents the total number of sampling times contained within the current data segment being analyzed.

[0048] Step 7: Extract long-term trends from the tidal-affected structural monitoring data to obtain the structural long-term service response characterization quantities.

[0049] The step of extracting long-term trends from the corrected structural monitoring data includes: The Empirical Mode Decomposition (EMD) method was used to perform multi-scale decomposition of the structural monitoring data. in: For the first The eigenmode function of the first order is at the 1st order. The values ​​at each sampling point; ; This represents the total number of eigenmode functions obtained from empirical mode decomposition. To remove the residual trend terms after removing the intrinsic mode functions of each order; Long-term trend extraction is performed by retaining long-term trend items: in, The set of low-frequency intrinsic mode functions that meet the preset long-term characteristic period conditions and whose frequency does not belong to the tidal dominant frequency band; The extracted structural long-term service response characteristics are used for long-term deformation analysis, structural damage identification, safety status assessment, and remaining service prediction.

[0050] As a preferred approach, to further identify and mitigate any residual tidal frequency components that may still exist after time-domain tidal correction, firstly, the tidal-corrected structural monitoring data... Perform a discrete Fourier transform to obtain the first... Fourier transform results corresponding to each frequency unit : in, Structural monitoring data corrected for tidal effects on the first Fourier transform results at each frequency unit; For the first Structural monitoring values ​​at each sampling time after tidal correction; The sampling time number, and ; It is the frequency unit number, and ; This represents the total number of sampling times contained within the current data segment being analyzed. The imaginary unit satisfies ; No. The actual frequency corresponding to each frequency unit is: in, for The actual frequency corresponding to each frequency unit; This refers to the frequency unit number; The sampling frequency for structural monitoring data; This represents the total number of sampling times contained within the current data segment being analyzed.

[0051] Next, based on the dominant tidal frequency Construct the tidal main frequency band blocking filter function: in, For the first The tidal main band blocking filter function value corresponding to each frequency unit; For the first The actual frequency corresponding to each frequency unit; The dominant tidal frequency determined by the aforementioned frequency domain analysis; These are frequency scale parameters used to control the bandwidth attenuation of the tidal dominant frequency. The larger the value, the wider the attenuation range in the vicinity of the tidal dominant frequency; This is the frequency unit number.

[0052] Then, the frequency domain data after tidal influence correction is processed using the tidal main band stop filter function to obtain the frequency domain monitoring data after tidal main band stop processing: in, For the first Frequency domain monitoring data of each frequency unit after tidal main frequency band impedance processing; For the first The tidal main band blocking filter function value corresponding to each frequency unit; Structural monitoring data corrected for tidal effects on the first Fourier transform results at each frequency unit; This is the frequency unit number.

[0053] Finally, the frequency domain monitoring data after tidal main frequency band blocking was subjected to inverse discrete Fourier transform and transformed back to the time domain to obtain structural monitoring data with further reduction of residual tidal periodic components. : in, For the first Structural monitoring values ​​after time-varying tidal response correction and tidal main frequency band blocking at each sampling time; For the first Frequency domain monitoring data of each frequency unit after tidal main frequency band impedance processing; It is the frequency unit number, and ; The sampling time number, and ; This represents the total number of sampling times contained within the current data segment being analyzed. The imaginary unit satisfies .

[0054] Thus, by using the inverse discrete Fourier transform, the frequency-domain processed structural monitoring data is restored to the time domain, yielding structural monitoring data that can be used for subsequent long-term trend extraction. .

[0055] Accordingly, in step 7, the structural monitoring data after further weakening the residual tidal periodic component is analyzed. Long-term trend extraction is performed to obtain the structural long-term service response characterization.

[0056] The step of extracting long-term trends from the corrected structural monitoring data includes: The Empirical Mode Decomposition (EMD) method was used to perform multi-scale decomposition of the structural monitoring data. in: For the first The eigenmode function of the first order is at the 1st order. The values ​​at each sampling point; ; This represents the total number of eigenmode functions obtained from empirical mode decomposition. This refers to the residual trend term after removing the intrinsic mode functions of each order.

[0057] Long-term trend extraction is performed by retaining long-term trend items: in, The set of low-frequency intrinsic mode functions that meet the preset long-term characteristic period conditions and whose frequency does not belong to the tidal dominant frequency band; The extracted structural long-term service response characteristics are used for long-term deformation analysis, structural damage identification, safety status assessment, and remaining service prediction.

[0058] Specifically, taking a long-term health monitoring system for underground integrated pipe corridors in a coastal city as an example, this invention elaborates on the monitoring method for intelligent correction of multi-source information data in coastal cities: Two settlement sensors and two tilt sensors were installed on the top structure of the integrated utility tunnel, four strain sensors were installed at key locations on the side walls, and two groundwater pressure sensors were installed in the surrounding soil. Simultaneously, tidal level monitoring data from a tidal station approximately 1.8 km from the project site were acquired, forming a joint monitoring system for structural response and tidal environment. Settlement monitoring utilized GNSS (Global Navigation Satellite System) displacement monitoring equipment, tilt monitoring employed a MEMS (Micro Electro Mechanical System) inclinometer, strain monitoring used a vibrating wire strain gauge, and groundwater pressure monitoring employed a pore water pressure sensor. All monitoring equipment was connected to a unified data acquisition system, and the sampling frequency for each monitoring channel was set to 10 minutes per sampling.

[0059] The system operated continuously for 15 days, collecting a total of 2,160 sets of valid monitoring data. During the monitoring period, the local tide level varied from 0.8m to 3.3m, with an average tidal range of approximately 2.1m. The settlement monitoring data for the top of the integrated utility tunnel fluctuated from 4.9mm to 6.5mm, the tilt angle varied from 0.012° to 0.038°, and the groundwater pressure varied from 12.3kPa to 21.6kPa.

[0060] The system continuously collects the above data and synchronizes it using a unified clock. Missing data points are compensated using linear interpolation. For example, if settlement monitoring data is missing at a certain moment, interpolation is performed using data from two consecutive moments: if the monitoring value at 08:00 is 5.1 mm and the monitoring value at 08:20 is 5.5 mm, and the data at 08:10 is missing, then the compensated value is calculated to be 5.3 mm.

[0061] Significant abrupt changes are eliminated using the three-standard-deviation method. For example, if the average settlement over a certain period is 5.4 mm and the standard deviation is 0.25 mm, then the anomaly detection thresholds are: upper limit: 6.15 mm; lower limit: 4.65 mm. If the settlement value at a sampling point reaches 7.2 mm, it is identified as an outlier and eliminated.

[0062] For long-term monitoring data such as settlement and tilt angle, the Savitzky-Golay smoothing filter method was used for noise reduction. The filter window length was set to 11 sampling points, corresponding to a monitoring duration of 110 minutes. After filtering, the standard deviation of random fluctuations in the settlement data decreased from 0.21 mm to 0.08 mm.

[0063] The sliding window length was set to 72 sampling points, corresponding to an approximately 12-hour tidal cycle, with a window step size of 1 sampling point. Frequency domain analysis, correlation analysis, and periodic feature extraction were performed on each monitoring indicator within each window.

[0064] For example, some of the original settlement monitoring data for a certain tidal cycle are as follows: As can be seen from the above, the settlement data exhibits obvious periodic fluctuations with changes in tidal level, with a settlement peak-to-valley difference of 1.3 mm.

[0065] Subsequently, a Fast Fourier Transform analysis was performed on the settlement data within the current window, and the power spectral density was calculated. The analysis results show that the frequency corresponding to the main peak of the settlement data spectrum is approximately: The corresponding period is approximately 12.4 hours, which is basically consistent with the local semi-diurnal tidal cycle, indicating that there is a significant tidal-induced response in the monitoring data.

[0066] Furthermore, cross-correlation analysis was used to study the time-delay relationship between tidal level changes and settlement response. According to the normalized cross-correlation analysis results, the correlation between the two reaches its maximum when the structural settlement response lags behind the tidal level change by 12 sampling points. Therefore, the optimal discrete time delay was determined to be... In this embodiment, the sampling time interval for each monitoring channel is 10 minutes, i.e. The corresponding optimal delay time is From this, the tidal angular frequency is obtained. The calculation results show that when the time delay is about 2 hours, the correlation between tide level and settlement response is the strongest, with a correlation coefficient of 0.83; while the maximum correlation coefficient between dip angle data and tide level change is about 0.71, corresponding to a time delay of about 1.5 hours; and the maximum correlation coefficient between groundwater pressure and tide level change reaches 0.91, corresponding to a time delay of about 2.4 hours.

[0067] Subsequently, a time-varying tidal response model is established, and in this embodiment, the tidal period is... The duration is 12.42 hours; tidal response amplitude The phase difference is 0.62 mm. It is 0.37 rad.

[0068] The parameters of the time-varying tidal response model are dynamically solved using the sliding time window least squares method, and the tidal influence term is separated from the original monitoring data. In other words, the parameters of the time-varying tidal response model are dynamically solved using the sliding time window least squares method. Within the current sliding window, the tidal response amplitude is obtained as follows: Residual phase parameters Therefore, according to the discrete time-varying tidal response model, within the current sliding window, the tidal influence term can be expressed as: .

[0069] After correction, the settlement data for the same time period are as follows: It can be seen that the peak-to-valley difference in the settlement data before correction was about 1.3 mm; after correction, the peak-to-valley difference was reduced to 0.2 mm, and the periodic fluctuations were significantly weakened.

[0070] Further statistical results show that: the standard deviation of the settlement data before correction was 0.48 mm; the standard deviation of the settlement data after correction decreased to 0.11 mm; the amplitude of periodic low-frequency fluctuations decreased by about 77%; and the stability of long-term trend fitting improved by about 65%.

[0071] In practical applications, when integrated utility tunnels are operating under conditions of high groundwater levels and strong tidal fluctuations, even if the instantaneous settlement value has not yet exceeded the traditional alarm threshold, if the long-term average settlement continues to rise, the periodic fluctuation of the tilt angle intensifies, the correlation between groundwater pressure and tidal level continues to increase, and the amplitude of the strain periodic response significantly increases, it indicates that the coupling effect of the tidal environment is strengthening. After correcting for tidal effects using the method of this invention, the periodic low-frequency drift in the monitoring data can be significantly reduced, making the true long-term deformation trend of the structure clearer, thereby identifying abnormal structural changes earlier.

[0072] In summary, this invention provides a monitoring method for intelligently correcting multi-source information data in coastal cities. This method addresses the problems of low-frequency periodic drift, long-term trend distortion, and difficulty in accurately identifying the true structural response caused by tidal changes during long-term structural health monitoring in coastal environments. By establishing a dynamic coupling relationship between tidal level changes and structural monitoring responses, this method identifies the tidal periodic response component in the monitoring data and constructs a time-varying tidal response model. This effectively separates the tidal environmental effects from the true structural service response, thereby improving the stability, accuracy, and interpretability of long-term monitoring data for coastal infrastructure.

[0073] This invention is applicable to long-term health monitoring scenarios for coastal infrastructure such as bridges, tunnels, underground utility tunnels, port projects, seawalls, underground structures, and marine engineering structures. It is suitable for correcting tidal environmental effects in various types of monitoring data, including strain, displacement, settlement, tilt angle, water pressure, vibration, and temperature.

[0074] The above description is merely a description of preferred embodiments of the present invention and is not intended to limit the scope of the present invention in any way. Any changes or modifications made by those skilled in the art based on the above disclosure shall fall within the protection scope of the claims.

Claims

1. A method for monitoring multi-source information data with intelligent correction in coastal cities, characterized in that, include: Step 1: Obtain structural monitoring data and tide level data for the corresponding time period during the long-term operation of the structure; Step 2: Preprocess the structural monitoring data; Step 3: Perform frequency domain analysis on the preprocessed structural monitoring data to determine the dominant tidal frequency; Step 4: Use cross-correlation analysis to study the time delay relationship between tidal level changes and structural response. By searching for the time delay corresponding to the maximum correlation coefficient, determine the optimal delay time for tidal level changes to affect the structural response. Step 5: Based on the dominant tidal frequency and the optimal delay time, a time-varying tidal response model is established using a time-varying harmonic model to obtain the tidal influence term. The parameters of the time-varying tidal response model are then dynamically updated using the sliding time window least squares method. Step 6: Separate the tidal influence term from the preprocessed structural monitoring data to obtain the tidal influence corrected structural monitoring data; Step 7: Extract long-term trends from the tidal-affected structural monitoring data to obtain the structural long-term service response characterization quantities.

2. The monitoring method for intelligent correction of multi-source information data in coastal cities according to claim 1, characterized in that, The structural monitoring data is obtained through a structural health monitoring platform, which includes a data monitoring unit for acquiring structural monitoring data, a data acquisition unit for acquiring tidal data, a data transmission unit, and a data processing platform. The data monitoring unit includes any one or more of the following: a strain monitoring unit, a displacement monitoring unit, a tilt angle monitoring unit, a water pressure monitoring unit, and a temperature monitoring unit. The data acquisition unit enables synchronous acquisition by each data monitoring unit, and the data transmission unit enables real-time transmission of structural monitoring data and tidal data to the data processing platform.

3. The monitoring method for intelligent correction of multi-source information data in coastal cities according to claim 1, characterized in that, In step 2, the preprocessing of the structural monitoring data includes: establishing a unified time axis through time synchronization for structural monitoring data with different sampling frequencies; using linear interpolation to complete missing data points in the structural monitoring data; using statistical discrimination methods to identify and remove outliers in the structural monitoring data; and performing noise reduction processing on the structural monitoring data.

4. The monitoring method for intelligent correction of multi-source information data in coastal cities according to claim 1, characterized in that, In step 3, the Fast Fourier Transform is used to perform frequency domain analysis on the preprocessed structural monitoring data.

5. The monitoring method for intelligent correction of multi-source information data in coastal cities according to claim 1, characterized in that, In step 3, the energy distribution of structural monitoring data in different frequency ranges is analyzed by power spectral density function, and the tidal dominant frequency is determined by searching for power spectral peaks.

6. The monitoring method for intelligent correction of multi-source information data in coastal cities according to claim 1, characterized in that, In step 4, when studying the time-delay relationship between tidal level change and structural response using cross-correlation analysis, the time-delay cross-correlation function between tidal level change and structural response is defined as: in, For structural monitoring data and tide data in time lag Normalized cross-correlation coefficients under the given conditions; The discrete time delay index is... Furthermore, a positive time delay is defined to represent the structural response lagging behind the tide level; Let A be the set of sampling points where both sets of data have valid values ​​under the corresponding time delay conditions. ; and These are the mean structural response and mean tide level for the corresponding valid data segments, respectively; For the first Preprocessed structural monitoring data at each sampling time point For the first Tide level data at each sampling time Optimal Discrete Time Delay for: in, The preset time delay search interval; According to the sampling interval The optimal discrete time delay Optimal delay time converted to tidal level change acting on structural response : 。 7. The monitoring method for intelligent correction of multi-source information data in coastal cities according to claim 1, characterized in that, In step 5, based on the dominant tidal frequency and the optimal delay time, a time-varying tidal response model is established using a time-varying harmonic model to obtain the tidal influence term: in: This refers to the tidal influence term calculated using a time-varying tidal response model. For the first The time-varying tidal response amplitude at each sampling time; The dominant frequency is tidal. The sampling time interval; The optimal delay time for the structural response to tidal changes This is the residual time-varying phase correction amount used to characterize factors such as local environmental state and model errors.

8. The monitoring method for intelligent correction of multi-source information data in coastal cities according to claim 7, characterized in that, In step 5, considering the modulation effect of environmental factors on the tidal response intensity, an environmental coupling amplitude model is established: in, For the first The ambient temperature at each sampling time For the first Rainfall at each sampling time or within the corresponding statistical time scale For the first Groundwater level at each sampling time , , , These are the parameters for the environmental coupling amplitude model.

9. The monitoring method for intelligent correction of multi-source information data in coastal cities according to claim 1, characterized in that, In step 6, after obtaining the tidal-corrected structural monitoring data and before step 7, the following steps are also included: Discrete Fourier transform is performed on the tidal-corrected structural monitoring data to obtain the tidal-corrected frequency domain data. Based on the tidal dominant frequency identified in the aforementioned steps, a tidal dominant frequency band blocking filter function is constructed. The frequency domain data after tidal influence correction is processed using the tidal main band stop filter function to obtain the frequency domain monitoring data after tidal main band stop processing. The frequency domain monitoring data after tidal main frequency band blocking was subjected to inverse discrete Fourier transform and converted back to the time domain to obtain structural monitoring data with further weakened residual tidal periodic components, which served as the structural monitoring data after tidal influence correction.

10. The monitoring method for intelligent correction of multi-source information data in coastal cities according to claim 9, characterized in that, In step 7, the long-term trend of the structural monitoring data after further weakening the residual tidal periodic component is extracted to obtain the structural long-term service response characterization quantity.

Images