Temperature correction method of flexible inclinometer for monitoring deflection of concrete panel

Abstract

The invention discloses a flexible inclinometer temperature correction method for monitoring the deflection of a concrete panel, and relates to the field of structural health monitoring, and the method comprises the following steps: S1, collecting the environmental data of the concrete panel, and constructing a structural state feature vector in combination with preset material parameters; s2, establishing a thermal drift correction model, and calculating a thermal drift response value; s3, stripping the original deflection data based on the response value to obtain a real deflection value; s4, calculating a deflection simulation value through finite element simulation, and comparing the deflection simulation value with a true value to obtain a drift prediction consistency residual error; s5, based on the spatial distance between the structural state feature vector and the historical feature vector, reconstructing a deflection correction value by adopting a semantic weighted interpolation algorithm; and S6, dynamically monitoring residual change according to the time window, and updating the thermal drift correction mode. Through multi-source sensing fusion, physical simulation feedback and tensor interpolation reconstruction, the accuracy, stability and robustness of thermal drift correction are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of structural health monitoring, specifically to a temperature correction method for a flexible inclinometer used for monitoring the deflection of concrete panels. Background Technology

[0002] Concrete structures are widely used in large-scale infrastructure projects, and their long-term service performance is affected by multiple factors such as environmental factors, material aging, and structural loads. Deflection, as an important parameter reflecting changes in the state of concrete structures, is of great significance for structural health monitoring and early warning of defects. To achieve high-precision deflection monitoring, flexible inclinometers are widely used for deformation monitoring in concrete panels, bridges, tunnels, and other structures due to their good adaptability and deployment flexibility.

[0003] However, in actual deployment, because the inclinometer itself and its sensitive elements are highly sensitive to temperature changes, its output signal usually contains significant temperature drift errors, especially under conditions of frequent temperature fluctuations on the surface of concrete structures. This type of temperature drift can seriously interfere with the accuracy of the original measurements, reduce the reliability of the monitoring data, and even lead to distortion of the structural condition assessment results.

[0004] Existing temperature correction methods are mostly based on univariate linear models or empirical temperature drift curves for compensation, lacking the ability to model complex environmental coupling factors and struggling to adapt to multi-source disturbances and nonlinear response characteristics. Furthermore, during the long-term service of concrete structures, factors such as structural thermal inertia, material parameter variability, and boundary condition uncertainties further exacerbate the coupling between temperature drift errors and actual deflection changes, increasing the difficulty of correction.

[0005] Some studies have introduced data-driven methods to model and fit drift trends, but the stability and generalization ability of these models remain limited in the face of sensor response non-stationarity and a lack of sufficient calibration data. Furthermore, most existing methods rely on static modeling or one-time corrections, lacking a mechanism for continuous modeling and iterative updates of the dynamic evolution of thermal drift, making it difficult to meet the practical engineering requirements for highly reliable, low-noise, and robust structural deflection monitoring.

[0006] Therefore, how to construct a temperature correction method that integrates multi-source environmental information, material thermal properties, and structural response behavior to improve the adaptability and measurement accuracy of flexible inclinometers in complex concrete structures remains an important technical challenge in the field of structural health monitoring. Summary of the Invention

[0007] Based on the shortcomings of the prior art described above, the purpose of this invention is to provide a temperature correction method for a flexible inclinometer used for monitoring the deflection of concrete panels, so as to solve the above-mentioned technical problems.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a temperature correction method for a flexible inclinometer used for monitoring the deflection of concrete panels, comprising:

[0009] S1: Use a sensor network to collect environmental datasets related to the concrete panel, and combine them with preset material parameters to construct a structural state feature vector;

[0010] S2: Establish a thermal drift correction model based on the structural state feature vector, and calculate the thermal drift response value at the target time;

[0011] S3: Based on the thermal drift response value, the original deflection data is stripped of temperature drift using a signal stripping method to obtain the true deflection data;

[0012] S4: Based on the environmental dataset and preset material parameters, the deflection simulation value is calculated using the finite element thermal response simulation function, and compared with the actual deflection data to calculate the drift prediction consistency residual;

[0013] S5: Based on the spatial distance between the structural state feature vector and the collected historical feature vector, the real deflection data is reconstructed using a semantic weighted interpolation algorithm to obtain the deflection correction value.

[0014] S6: Based on a preset time window, dynamically monitor the changes in drift consistency residuals and thermal drift response values, and dynamically update the thermal drift correction model.

[0015] The present invention is further configured such that S1 includes:

[0016] Use sensor networks to collect environmental datasets related to concrete panels;

[0017] Five-dimensional vector fusion is performed based on environmental datasets to construct a five-dimensional physical perception vector;

[0018] The environmental dataset includes: temperature, raw deflection data, relative humidity, gravitational disturbance, and material acoustic emission intensity.

[0019] The present invention is further configured to perform data fusion based on preset material parameters to obtain a structural material mapping tensor;

[0020] Based on the five-dimensional physical perception vector processed by the nonlinear mapping algorithm, and combined with the structural material mapping tensor, a multi-dimensional structural state tensor is constructed.

[0021] The material parameters include: local elastic modulus, local specific heat capacity, and thermal conductivity of concrete.

[0022] The present invention is further configured to use a dimensionality reduction algorithm to process the structural state tensor to obtain the final structural state feature vector.

[0023] The present invention is further configured such that S2 includes:

[0024] The structural state feature vector is used as input and passed to the thermal drift prediction function to calculate the initial thermal drift response value.

[0025] By combining the preset residual mean and the preset drift change rate, the initial drift value is dynamically corrected to obtain the thermal drift response value at the target time.

[0026] The present invention is further configured such that S3 includes:

[0027] Based on the time series of raw deflection data in the environmental dataset, a filtering algorithm is used to estimate the residual error.

[0028] By combining the thermal drift response value and residual error, drift stripping is performed on the original deflection data to obtain temperature-corrected true deflection data.

[0029] The present invention is further configured such that S4 includes:

[0030] Based on the environmental dataset and preset material parameters, the thermal-structural coupling behavior of concrete structures is modeled and calculated using the finite element thermal response simulation function to obtain the simulated values ​​of structural response.

[0031] The present invention is further configured to compare the simulated structural response values ​​with the actual deflection data after temperature correction, and calculate the drift prediction consistency residual used for thermal drift correction model optimization.

[0032] The present invention is further configured such that S5 includes:

[0033] Based on the spatial distance relationship between the structural state feature vector and the historical structural state feature vector in the system execution record, the set of historical feature points with the smallest spatial distance is selected.

[0034] A weighted interpolation algorithm based on spatial distance is used to reconstruct the real deflection data corresponding to the historical feature point set, thereby generating the deflection correction value of the current structure position.

[0035] The present invention is further configured such that S6 includes:

[0036] By monitoring the residual between the corrected deflection value and the measured value, and by setting a fixed time window, the prediction consistency residual within the time window is statistically analyzed, and the mean residual is calculated.

[0037] When the mean residual exceeds the set update threshold, the retraining mechanism of the thermal drift correction model is triggered, and the thermal drift change rate is obtained by calculating the rate of change of the thermal drift response value within the time window.

[0038] The residual mean and drift rate of change are passed back to S2 to achieve dynamic updating of the thermal drift response value.

[0039] This invention provides a temperature correction method for a flexible inclinometer used for monitoring the deflection of concrete panels. The method comprises the following steps: S1: Collecting environmental datasets related to the concrete panel using a sensor network, and constructing a structural state feature vector based on preset material parameters; S2: Establishing a thermal drift correction model based on the structural state feature vector, and calculating the thermal drift response value at the target time; S3: Based on the thermal drift response value, performing temperature drift stripping on the collected raw deflection data using a signal stripping method to obtain the true deflection data; S4: Calculating the simulated deflection value using a finite element thermal response simulation function based on the environmental dataset and preset material parameters, comparing it with the true deflection data, and calculating the drift prediction consistency residual; S5: Reconstructing the true deflection data using a semantic weighted interpolation algorithm based on the spatial distance between the structural state feature vector and the collected historical feature vectors to obtain the deflection correction value; S6: Dynamically monitoring changes in the drift consistency residual and thermal drift response value within a preset time window, and dynamically updating the thermal drift correction model. The resulting benefits include:

[0040] Achieving deep integration of multi-source environmental perception and structural response improves the accuracy and adaptability of thermal drift correction: By constructing a structural state feature vector that integrates multi-dimensional perception information such as temperature, humidity, gravity disturbance, and acoustic emission intensity, and combining it with local thermo-mechanical parameters of concrete, high-dimensional modeling and physical correlation of thermal drift effect are achieved, effectively improving the accuracy and generalization ability of temperature drift correction, and breaking through the limitations of traditional linear or univariate temperature compensation methods.

[0041] A finite element thermal response simulation feedback mechanism is introduced to achieve physical consistency calibration of model training: by comparing finite element thermal-structure coupling simulation with measured deflection data, the drift prediction consistency residual is calculated, and the residual is used as the optimization basis of the regressor. A closed-loop feedback correction framework based on physical mechanism is constructed, which effectively avoids the prediction bias and instability problems that are prone to occur in pure data-driven models under structural variations or complex working conditions.

[0042] A deflection interpolation reconstruction mechanism based on tensor space semantic distance is proposed to enhance the robustness and spatial generalization ability of data reconstruction: by calculating the semantic distance between the structural state feature vector and the historical state, the closest historical data point in the feature space is selected for weighted interpolation, which effectively improves the reconstruction accuracy of the deflection signal after stripping and ensures the monitoring continuity and data integrity in the context of data loss or drift interference.

[0043] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0045] Figure 1 A flowchart illustrating a temperature correction method for a flexible inclinometer used for monitoring the deflection of a concrete panel, as shown in an exemplary embodiment of the present invention.

[0046] Figure 2 A horizontal settlement diagram before temperature correction for a flexible inclinometer temperature correction method for monitoring the deflection of a concrete panel, as shown in an exemplary embodiment of the present invention.

[0047] Figure 3 The image shows a temperature-corrected horizontal settlement diagram of a flexible inclinometer temperature correction method for monitoring the deflection of a concrete panel, as illustrated in an exemplary embodiment of the present invention. Detailed Implementation

[0048] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0049] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0050] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0051] Example:

[0052] Temperature correction method for flexible inclinometers used for concrete panel deflection monitoring, such as Figure 1 As shown, it includes:

[0053] S1: Use a sensor network to collect environmental datasets related to the concrete panel, and combine them with preset material parameters to construct a structural state feature vector;

[0054] S2: Establish a thermal drift correction model based on the structural state feature vector, and calculate the thermal drift response value at the target time;

[0055] S3: Based on the thermal drift response value, the original deflection data is stripped of temperature drift using a signal stripping method to obtain the true deflection data;

[0056] S4: Based on the environmental dataset and preset material parameters, the deflection simulation value is calculated using the finite element thermal response simulation function, and compared with the actual deflection data to calculate the drift prediction consistency residual;

[0057] S5: Based on the spatial distance between the structural state feature vector and the collected historical feature vector, the real deflection data is reconstructed using a semantic weighted interpolation algorithm to obtain the deflection correction value.

[0058] S6: Based on a preset time window, dynamically monitor the changes in drift consistency residuals and thermal drift response values, and dynamically update the thermal drift correction model.

[0059] The present invention is further configured such that S1 includes:

[0060] Use sensor networks to collect environmental datasets related to concrete panels;

[0061] Five-dimensional vector fusion is performed based on environmental datasets to construct a five-dimensional physical perception vector;

[0062] The environmental dataset includes: temperature, raw deflection data, relative humidity, gravitational disturbance, and material acoustic emission intensity;

[0063] Data fusion is performed based on preset material parameters to obtain the structural material mapping tensor;

[0064] Based on the five-dimensional physical perception vector processed by the nonlinear mapping algorithm, and combined with the structural material mapping tensor, a multi-dimensional structural state tensor is constructed.

[0065] The material parameters include: local elastic modulus of concrete, local specific heat capacity, and thermal conductivity;

[0066] The structural state tensor is processed using a dimensionality reduction algorithm to obtain the final structural state feature vector.

[0067] Specifically, a sensor network is constructed using temperature and humidity sensors, a flexible inclinometer, a triaxial accelerometer, and an acoustic emission / AE sensor. This sensor network is deployed in key areas of the concrete panel for monitoring, collecting five types of raw data. Each type of data is sampled synchronously to form a multi-dimensional time-series environmental dataset with a unified time axis. At each sampling time, the five types of data are combined into a five-dimensional vector. This five-dimensional vector is then processed using various methods such as principal component analysis, standardized weight stacking, or deep fusion networks. The output is a five-dimensional physical perception vector with stable features, consistent dimensions, and temporal physical meaning. The local elastic modulus of concrete represents the concrete's resistance to stress in a localized area, reflecting the stiffness change of the panel under thermal loads or environmental stresses. It is a crucial factor affecting deflection response. The local elastic modulus of concrete is obtained from design data during the construction and deployment phases. Local specific heat capacity represents the amount of heat absorbed by a unit mass of concrete to increase its temperature by 1°C, determining the concrete's heat storage capacity under temperature changes. Its specific value is determined based on the concrete mix design and preset according to material composition. Thermal conductivity represents the ability of a unit area of ​​material to conduct heat along a temperature gradient per unit time, used to control the diffusion rate of the temperature field within the concrete. The specific value of thermal conductivity is also determined based on the concrete mix design and material parameters set during the structural design phase. For example, the elastic modulus of conventional C30 concrete is approximately 30 GPa, its specific heat capacity is approximately 900 J / (kg·K), and its thermal conductivity is approximately 1.7 W / (m·K). By fusing the local material parameters—local elastic modulus, local specific heat capacity, and thermal conductivity—into tensor form, the structural material mapping tensor is obtained. An initial structural state tensor is obtained by fusing a five-dimensional physical sensing vector with a structural material mapping tensor using a nonlinear mapping function. This nonlinear mapping function can be an RBF-based mapping network or a deep autoencoder. The initial structural state tensor is a high-dimensional tensor composed of multi-source sensor data and multi-dimensional material physical parameters. Collinearity or redundancy may exist between many dimensions. Directly using it in downstream modules without dimensionality reduction would introduce noise or redundant information, affecting the model's learning efficiency and prediction stability. Furthermore, using high-dimensional features as input significantly increases the number of model parameters, resulting in high training and inference overhead. Dimensionality reduction reduces unnecessary input complexity and improves the processing efficiency of subsequent regressors, interpolators, simulation comparison modules, etc. To improve system input consistency, reduce redundancy, enhance generalization ability, and improve modeling efficiency, the initial structural state tensor is reduced in dimensionality using a Diffusion Map / Auto Encoder algorithm, ultimately obtaining the structural state feature vector.

[0068] The present invention is further configured such that S2 includes:

[0069] The structural state feature vector is used as input and passed to the thermal drift prediction function to calculate the initial thermal drift response value.

[0070] By combining the preset residual mean and the preset drift change rate, the initial drift value is dynamically corrected to obtain the thermal drift response value at the target time.

[0071] Specifically, the thermal drift response value represents the predicted deflection drift caused by temperature at the current moment. The structural state feature vector is passed as input to the thermal drift prediction function, which is a prediction function based on a deep regression model, such as Transformer or attention network, used to model the mapping from the structural state feature vector to the thermal drift response value. The calculation logic for the thermal drift response value is as follows: ,in, This is the thermal drift response value; The thermal drift prediction function is a learned prediction function used to map the structural state feature vector to thermal drift-related outputs, using a Transformer structure, GNN-Tensor model, or attention network for prediction. The structural state feature vector; These are the model weight parameters, which are automatically obtained after the model is trained. This is the residual mean, initially set to 0 by default, and is a dynamically adjusted value based on subsequent steps. This is the drift change rate, initially set to 0 by default, and will be automatically adjusted based on the rate of change of the thermal drift response value. This is a thermal drift correction model. This correction model is not enabled during the first system run; only the thermal drift correction model is used. The initial thermal drift response value is calculated, and the initial drift value is dynamically corrected only after the system is executed a second time to obtain the thermal drift response value at the target time.

[0072] The present invention is further configured such that S3 includes:

[0073] Based on the time series of raw deflection data in the environmental dataset, a filtering algorithm is used to estimate the residual error.

[0074] By combining the thermal drift response value and residual error, drift stripping is performed on the original deflection data to obtain temperature-corrected true deflection data.

[0075] Specifically, residual error refers to the portion of the original deflection data that does not belong to the structural response or temperature drift. It typically manifests as low-amplitude, high-frequency random disturbances. Residual error is obtained by extracting high-frequency noise components from the time-series signal using a filter. While using filters to extract residual error is an existing technique, this invention does not limit the type of filter used; wavelet packet decomposition, SG filters, or Kalman filters can be used. The true deflection data is obtained by subtracting the thermal drift response value (containing temperature drift components) and the residual error (containing random disturbances) from the original deflection data. The horizontal settlement maps before and after temperature correction can be plotted on the subsequent system visualization page to highlight the correction results. The horizontal settlement maps include: 1. A horizontal settlement map before temperature correction plotted based on the original deflection data in the environmental dataset, such as... Figure 2 As shown; 2. A temperature-corrected horizontal settlement diagram drawn based on actual deflection data, as shown. Figure 3 As shown.

[0076] The present invention is further configured such that S4 includes:

[0077] Based on the environmental dataset and preset material parameters, the thermal-structural coupling behavior of concrete structures is modeled and calculated using the finite element thermal response simulation function to obtain the simulated values ​​of structural response.

[0078] The simulated structural response values ​​are compared with the actual temperature-corrected deflection data to calculate the drift prediction consistency residuals used for thermal drift correction model optimization.

[0079] Specifically, the structural response simulation value refers to the response result of a concrete structure under environmental temperature and humidity conditions, obtained through finite element simulation methods such as ANSYS and COMSOL, after the application of internal thermal stress. The calculation logic for the structural response simulation value is as follows: ,in, These are the simulated values ​​of the structural response; This is a finite element thermal response simulation function, used to numerically calculate the structural response based on a thermo-mechanical coupling model. For temperature; Relative humidity; Thermal conductivity; This refers to the local specific heat capacity. The geometric distance from the heat source to the observation point is used to estimate the delay and local influence of the thermal response. It is obtained by calculating the Euclidean distance between the heat source and the sensor / measuring point using BIM engineering drawings. The drift prediction consistency residual measures the deviation between the actual deflection data after drift stripping and the simulated structural response values ​​obtained from finite element thermo-structural simulation. A smaller residual indicates a higher consistency between the thermal drift stripping results and the structural physical behavior, and thus higher reliability. The calculation logic for the drift prediction consistency residual is as follows: ,in, For drift prediction consistency residuals; This is based on actual deflection data; These are the simulated values ​​for the structural response.

[0080] The present invention is further configured such that S5 includes:

[0081] Based on the spatial distance relationship between the structural state feature vector and the historical structural state feature vector in the system execution record, the set of historical feature points with the smallest spatial distance is selected.

[0082] A weighted interpolation algorithm based on spatial distance is used to reconstruct the real deflection data corresponding to the historical feature point set, thereby generating the deflection correction value of the current structure position.

[0083] Specifically, starting from the current input structural state feature vector, the Euclidean distance between it and each observation point in the historical structural state feature set in the system execution record is calculated. Historical feature points with fixed values ​​are sorted and filtered according to distance to form a similar sample set. The fixed values ​​are used to calculate the mean, and are set to 10 by default. For each sample in the similar sample set, an inverse distance weight is calculated. The inverse distance weight calculation logic is as follows: ,in, Inverse distance weights; The current structural feature vector to historical structural state feature vector Euclidean distance; For the currently calculated weights of the th A historical sample; For each index in the set of all selected similar historical samples, summation is performed. Indicates the first The original weights of each historical sample; This represents the sum of the weights of all selected samples, used to standardize all weights; it's the normalization operation in a weighted average. The deflection correction value for the current moment is reconstructed by weighting the true deflection values ​​of similar samples. The deflection correction value calculation logic is as follows: ,in, This is the deflection correction value; The similar sample set is constructed by calculating the Euclidean distance between the current structural feature state vector and each observation point in the historical structural feature set. It is built by summarizing the historical feature points with the first fixed values ​​and sorted according to the Euclidean distance. Inverse distance weights; The obtained real deflection data is used to reconstruct the real deflection data using a semantically weighted interpolation algorithm, resulting in a complete time-series-based deflection correction value sequence.

[0084] The present invention is further configured such that S6 includes:

[0085] By monitoring the residual between the corrected deflection value and the measured value, and by setting a fixed time window, the prediction consistency residual within the time window is statistically analyzed, and the mean residual is calculated.

[0086] When the mean residual exceeds the set update threshold, the retraining mechanism of the thermal drift correction model is triggered, and the thermal drift change rate is obtained by calculating the rate of change of the thermal drift response value within the time window.

[0087] The residual mean and drift rate of change are passed back to S2 to achieve dynamic updating of the thermal drift response value.

[0088] Specifically, the preset time window can be dynamically adjusted according to the accuracy requirements of the specific implementation project. For example, for scenarios with particularly high accuracy requirements, hourly windows can be used, retaining 60 one-minute sampling points per hour. The mean of the absolute values ​​of the prediction consistency residuals of all sampling points within the time window is used to obtain the mean residual, thus eliminating the influence of positive and negative deviations. The mean residual is monitored in real time to see if it exceeds the set update threshold. The default update threshold is set to 1.0. The typical accuracy of deflection sensors such as flexible inclinometers is ±0.5. Setting it to 1.0 can cover twice the measurement error and avoid false triggering. To prevent brief over-threshold exceedances due to noise, a number of consecutive over-threshold exceedances is set to improve robustness; the default number of consecutive exceedances is 3. When the mean residual is greater than the update threshold and the number of exceedances is greater than three, the time series of the thermal drift response values ​​within the time window is extracted. The rate of change is calculated using linear fitting or the difference method to obtain the thermal drift change rate. If the thermal drift change rate remains positive, it indicates accelerated thermal drift, requiring adjustment of the thermal drift prediction function in S2. Finally, the residual mean and thermal drift rate are passed back to the thermal drift prediction function in step S2 to update the preset residual mean and preset drift rate, and then the thermal drift response value is calculated to achieve adaptive dynamic update.

[0089] 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 scope of the technology 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 of temperature correction for flexible inclinometers for concrete panel deflection monitoring, characterised in that, Comprise: S1: collect the environmental data set related to the concrete panel by using the sensor network, and construct the structural state feature vector by combining the preset material parameters; S2: establish a thermal drift correction model based on the structural state feature vector to calculate the thermal drift response value corresponding to the target time; S3: based on the thermal drift response value, the original deflection data collected is subjected to temperature drift stripping by a signal stripping method, and the true deflection data is corrected; S4: based on the environmental data set and the preset material parameters, the deflection simulation value is calculated by using the finite element thermal response simulation function, and compared with the true deflection data to calculate the drift prediction consistency residual; S5: based on the spatial distance between the structural state feature vector and the collected historical feature vector, the true deflection data is reconstructed by using a semantic weighted interpolation algorithm to obtain the deflection correction value; S6: based on the preset time window, the changes of the drift consistency residual and the thermal drift response value are dynamically monitored, and the thermal drift correction model is dynamically updated.

2. The flexible inclinometer temperature correction method for concrete panel deflection monitoring of claim 1, wherein, The S1 comprises: collecting the environmental data set related to the concrete panel by using the sensor network; fusing the five-dimensional vector based on the environmental data set to construct a five-dimensional physical perception vector; The environmental data set includes: temperature, original deflection data, relative humidity, gravity disturbance and material acoustic emission intensity.

3. The temperature correction method for the flexible inclinometer for monitoring the deflection of the concrete panel according to claim 2, wherein, data fusion is performed based on the preset material parameters to obtain a structural material mapping tensor; a multi-dimensional feature structural state tensor is constructed based on the five-dimensional physical perception vector processed by the nonlinear mapping algorithm and the structural material mapping tensor; The material parameters include: local elastic modulus, local specific heat capacity and thermal conductivity of concrete.

4. The temperature correction method for the flexible inclinometer for monitoring the deflection of the concrete panel according to claim 3, wherein, a dimension reduction algorithm is used to process the structural state tensor to obtain the final structural state feature vector.

5. The flexible inclinometer temperature correction method for concrete panel deflection monitoring of claim 1, wherein, The S2 comprises: the structural state feature vector is taken as an input and transmitted to a thermal drift prediction function to calculate an initial thermal drift response value; the initial drift value is dynamically corrected by combining the preset residual mean and the preset drift change rate to obtain the thermal drift response value at the target time.

6. The flexible tiltmeter temperature correction method for concrete panel deflection monitoring of claim 1, wherein, The S3 comprises: based on the time sequence of the original deflection data in the environmental data set, a filtering algorithm is used to estimate the residual error; the original deflection data is subjected to drift stripping processing by combining the thermal drift response value and the residual error to obtain the true deflection data subjected to temperature correction.

7. The flexible tiltmeter temperature correction method for concrete panel deflection monitoring of claim 1, wherein, The S4 comprises: based on the environmental data set and the preset material parameters, the thermal-structural coupling behavior of the concrete structure is modeled and calculated by using the finite element thermal response simulation function to obtain the structural response simulation value.

8. The temperature correction method for the flexible inclinometer for monitoring the deflection of the concrete panel according to claim 1, wherein, the structural response simulation value is compared with the true deflection data subjected to temperature correction to calculate the drift prediction consistency residual for optimizing the thermal drift correction model.

9. The flexible tiltmeter temperature correction method for concrete panel deflection monitoring of claim 1, wherein, The S5 comprises: Based on the spatial distance relationship between the structural state feature vector and the historical structural state feature vector in the system execution record, the historical feature point set with the minimum spatial distance is selected; A weighted interpolation algorithm based on spatial distance is used to weight the real deflection data corresponding to the historical feature point set, and the deflection correction value of the current structure position is generated.

10. The flexible inclinometer temperature correction method for concrete panel deflection monitoring of claim 1, wherein, The S6 includes: By monitoring the residual of the corrected deflection value and the measured value, setting a fixed time window, counting the prediction consistency residual in the time window, and calculating the residual mean value; When the residual mean value exceeds the set update threshold, the retraining mechanism of the thermal drift correction model is triggered, and the thermal drift change rate is obtained by calculating the change rate of the thermal drift response value in the time window; The residual mean value and the drift change rate are transmitted back to S2 to realize the dynamic update of the thermal drift response value.

Images