A method for building structural health monitoring that couples BIM models with real-time monitoring data

Abstract

This invention discloses a method for building structural health monitoring that couples BIM models with real-time monitoring data, belonging to the field of building monitoring technology. The method includes the following steps: acquiring real-time monitoring data of the building structure; fusing strain growth rate, stiffness degradation amplitude, and displacement evolution trajectory from the monitoring data; extracting key characteristic parameters reflecting the nonlinear evolution trend of the structure; and generating buckling evolution sensitive factors for dynamically characterizing the structure's approach to buckling state. This invention achieves full-process control of disturbance signals in time series, modalities, and energy transfer by constructing a disturbance identification and control mechanism based on buckling evolution sensitive factors, avoiding misjudgments under critical states and improving system robustness and stability. Simultaneously, it introduces a virtual residual domain and fractal disturbance decomposition method to complete energy buffering and multi-scale deconstruction, enhancing low-frequency trend feedback and suppressing high-frequency interference, making health assessment results smoother and more reliable, and providing precise support for structural early warning and operation and maintenance decisions.

Description

Technical Field

[0001] The present invention relates to the technical field of building monitoring, and specifically to a building structure health monitoring method that couples a BIM model with real-time monitoring data. Background Art

[0002] Coupling a BIM model with real-time monitoring data for building structure health monitoring means dynamically integrating and jointly calculating the three-dimensional information model (BIM model) formed during the design and construction stages of a building with the real-time monitoring data (such as sensor data of temperature, strain, vibration, displacement, load, etc.) collected during the operation of the building. In this way, the digital model of the building structure is no longer static but continuously updated according to the actual operating environment and stress state, thereby achieving dynamic perception and evaluation of the structural safety and health status. Specifically, the BIM model can provide the geometric shape, material properties, and spatial relationships of building components, while the real-time monitoring data reflects the actual working state of the building during service. After coupling the two, the stress distribution, deformation trend, and potential risk points of the structure can be intuitively displayed in the virtual space, assisting engineers to detect abnormalities in a timely manner, give early warnings, and perform maintenance, thereby constructing a digital twin-based building health monitoring system.

[0003] The prior art has the following deficiencies: In the prior art, when the building structure gradually approaches the buckling critical state, the originally extremely small disturbance signals in the real-time monitoring data tend to be amplified by the nonlinear effect during the coupling calculation with the BIM model, resulting in a sharp fluctuation in the numerical output. This amplification effect not only masks the true stress evolution trend but may also cause the model to generate multiple false instability indicators in a short period, ultimately leading to a significant deviation in the judgment of the overall safety status of the structure.

[0004] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0005] The object of the present invention is to provide a building structure health monitoring method that couples a BIM model with real-time monitoring data to solve the problems in the above background art.

[0006] To achieve the above object, the present invention provides the following technical solution: A building structure health monitoring method that couples a BIM model with real-time monitoring data, characterized by including the following steps: Construct a buckling critical characteristic response identification mechanism: acquire real-time monitoring data of building structures, integrate strain growth rate, stiffness degradation amplitude and displacement evolution trajectory in the monitoring data, extract key characteristic parameters reflecting the nonlinear evolution trend of the structure, and generate buckling evolution sensitive factors for dynamically characterizing the structure approaching buckling state. Construct a disturbance response mapping mechanism: map buckling evolution sensitive factors to the disturbance gain domain, identify the high excitation potential region corresponding to the buckling evolution sensitive factors through a preset response amplification function, and locate the local component region that is easy to trigger disturbance amplification in the BIM model of the building structure; Construct a phase modulation mechanism for disturbance signals: Based on the temporal evolution characteristics of the buckling evolution sensitive factors in the high excitation potential region, the phase of the disturbance signal in the real-time monitoring data is fine-tuned to regulate its coupling rhythm with the nonlinear response of the structure, so as to weaken the temporal overlap effect of the peak of multi-source disturbances. Construct a modal response transfer mechanism: Based on the phase-modulated perturbation signal, guide it to a highly flexible local modal path in the high excitation potential region, and set a response threshold attenuation condition to suppress modal crosstalk and cascade amplification effects triggered by perturbation in non-critical regions; Constructing a residual energy buffer mechanism: After completing the modal response transfer, extract the amplitude anomaly segments that are not fully decayed in the disturbance signal, inject them into the virtual residual domain, and perform a step-by-step dissipation operation in the residual domain to reduce the risk of violent abrupt changes in the main structure response channel; Constructing a disturbance frequency band deconstruction mechanism: Based on the output results of the residual energy buffer mechanism, the fractal disturbance decomposition method is used to divide the real-time monitoring data into two types of structures: high-frequency micro-disturbance and low-frequency trend. In the coupled calculation with the BIM model, the feedback strength of the low-frequency trend is enhanced, while a nonlinear scaling suppression strategy is applied to the high-frequency micro-disturbance signal, thereby improving the stability and credibility of the structural health assessment results.

[0007] Preferably, generating a buckling evolution sensitivity factor includes the following steps: Real-time monitoring data of building structures are acquired, and the monitoring data is denoised, smoothed, and interpolated to obtain a multi-source response dataset that is time-aligned and has good continuity. Based on the multi-source response dataset, the strain growth rate, displacement evolution trajectory, and stiffness degradation amplitude are calculated separately, and the strain growth rate, displacement evolution trajectory, and stiffness degradation amplitude are normalized and combined into a multi-dimensional feature vector. The dynamic time warping method is used to compare the multidimensional feature vector with the historical feature trajectory, and the Kullback-Leibler divergence and time gradient are combined to calculate and identify key feature parameters to characterize the nonlinear evolution trend of the structure approaching the buckling state. By weighting and superimposing key feature parameters and combining them with geometric information, component boundary conditions, and connection relationships in the building information model, buckling evolution sensitivity factors are generated to achieve a visual representation of temporal evolution and spatial distribution in the building information model.

[0008] Preferably, in the process of generating the buckling evolution sensitive factor, a hyperbolic tangent function or an exponential function is used as a nonlinear amplification function to make the output value of the key characteristic parameter rise rapidly when it approaches the buckling critical state, so as to enhance the buckling evolution sensitive factor's ability to characterize the critical approach trend.

[0009] Preferably, the step of mapping the buckling evolution sensitivity factor to the perturbation gain domain includes: The buckling evolution sensitive factor sequence is input into the dynamic normalization stage, and the difference between the extreme values ​​and normal values ​​of the buckling evolution sensitive factor is amplified by combining the statistical distribution under historical operating conditions. The normalized buckling evolution sensitive factor is then mapped into the perturbation gain domain. The perturbation gain domain is calculated using a preset response amplification function. When the buckling evolution sensitivity factor continues to rise within a certain time window and the amplified response value exceeds the threshold, the region is determined to be a high excitation potential region. The coordinate information of the high excitation potential region is registered with the geometric topology of the components in the building information model to establish the correspondence between sensor coordinates and model nodes. The local component regions that are prone to triggering disturbance amplification are determined by combining the stiffness distribution and boundary conditions of the components and displayed in the building information model in an intuitive way.

[0010] Preferably, the step of performing phase fine-tuning on the disturbance signal in the real-time monitoring data includes: Based on the time series characteristics of the buckling evolution sensitive factor in the high excitation potential region, the phase information of the main energy concentration frequency band is extracted by fast Fourier transform, and the phase difference is calculated by combining the actual phase characteristics of the disturbance signal, and this is used as the phase reference value. The disturbance signal is segmented within a time window, and the Hilbert transform is used to obtain the continuous change function of the instantaneous phase in order to achieve time sequence reconstruction. The instantaneous phase of the disturbance signal is compared point by point with the phase reference value, and phase correction is achieved by introducing time delay, advance, or using a nonlinear phase offset function; Multiple phase-corrected disturbance signals are superimposed and analyzed to calculate phase correlation. When the peaks tend to overlap, a small phase delay is applied to make them staggered. The disturbance signal after comprehensive regulation is input into the nonlinear calculation of the building information model to verify whether the peak amplitude has decreased and the change has become stable, and the result is compared with that without regulation to quantify the improvement effect.

[0011] Preferably, the step of guiding the phase-modulated disturbance signal to a highly flexible local modal path and setting a response threshold attenuation condition includes: Finite element vibration analysis is performed on the overall modal distribution of the building structure to identify the flexibility modes with large displacement amplitude in the low frequency range, and the displacement participation coefficient and strain energy distribution of component elements in different modes in the high excitation potential region are calculated to determine the local modal path of large flexibility. The perturbation signal after phase modulation is subjected to frequency domain spectrum analysis to extract its main energy components and compare them with the frequency of the flexibility mode. Through modal orthogonality projection, only the components of the corresponding flexibility mode path are retained to achieve directional transmission of perturbation energy. In modal projection calculation, an energy threshold is set, modal components below the threshold are attenuated to zero, and nonlinear attenuation functions that gradually saturate with increasing amplitude are applied to components exceeding the threshold. Combined with the component boundary conditions in the building information model, local components that can accommodate energy are selected to avoid modal crosstalk and cascade amplification in non-critical areas. The disturbance signal, after projection and attenuation processing, is input into the overall structural dynamic response calculation. The response amplitude changes in the high excitation potential region and the non-critical region are compared to verify the effectiveness of the compliance mode for energy absorption. Dynamic feedback correction is achieved by adjusting the energy threshold or optimizing the attenuation function parameters.

[0012] Preferably, after completing the modal response transfer, the steps of extracting the incompletely decayed amplitude anomaly segments from the disturbance signal and performing a stepwise dissipation operation in the residual domain include: The perturbation signal after modal response transfer processing is scanned in its entire time series. An amplitude anomaly segment exceeding the threshold is extracted using a sliding window, and the threshold is dynamically adjusted according to the changes in the buckling evolution sensitivity factor. Amplitude anomaly segments are extracted from the main channel signal and injected into the virtual residual domain, maintaining temporal continuity and scaling to the standard energy scale of the residual domain during the injection process. In the residual domain, the amplitude anomalous segments are frequency-divided, an exponential decay function is applied to the high-frequency part, a logarithmic decay function is applied to the low-frequency part, and a time-weighted decay factor is superimposed to achieve stepwise dissipation. The residual domain dissipation results are compared with the evolution trend of the buckling evolution sensitive factor in the main channel. When the buckling evolution sensitive factor tends to be stable, the dissipation is deemed effective; otherwise, the dissipation function parameters are adjusted. By superimposing and analyzing the main channel response curve and the residual domain energy dissipation curve, we can verify whether the residual energy is completely eliminated and archived as parameters within a preset time, so as to ensure the long-term effectiveness of the residual energy buffering mechanism.

[0013] Preferably, the steps of dividing the output results based on the residual energy buffering mechanism into two types of structures—high-frequency perturbations and low-frequency trends—using the fractal perturbation decomposition method and processing them separately include: The signal after residual energy buffering is input into the fractal decomposition framework, and multi-scale decomposition is performed by combining wavelet packet decomposition and empirical mode decomposition methods to obtain high-frequency perturbation sequence and low-frequency trend curve. The low-frequency trend curve is matched with the mechanical response parameters of the components in the 3D building information model, and the low-frequency trend feedback is weighted by an amplification function to highlight the structural evolution trend. Amplitude distribution modeling is performed on high-frequency perturbation signals, peak segments are compressed using a nonlinear scaling function, and low-amplitude random noise is suppressed using an energy thresholding method while maintaining the continuity of the time series. The enhanced low-frequency trend signal and the suppressed high-frequency perturbation signal are combined into a new input sequence, which is then re-inputted into the building 3D information model for coupled calculation. The results are then verified to see if it eliminates false instability indicators and improves the stability and reliability of structural health assessment.

[0014] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention effectively solves the problem of nonlinear amplification of small-amplitude disturbance signals in building structures approaching the critical buckling state by constructing a disturbance response identification and control mechanism centered on buckling evolution sensitive factors. It achieves precise control of the disturbance signal throughout its temporal rhythm, modal path, and energy transfer processes. Before the disturbance enters the main structural response channel, disturbance source identification, amplification trend assessment, and energy guidance processing are completed, fundamentally avoiding misjudgments of structural safety due to false triggering under high sensitivity conditions. By identifying highly flexible modal paths within high excitation potential regions and setting response threshold attenuation conditions, disturbance energy is effectively confined to specific local areas, preventing diffusion to non-critical components, thereby significantly improving the robustness and response stability of the health monitoring system under critical conditions.

[0015] This invention introduces a virtual residual domain and fractal perturbation decomposition method to achieve energy buffering and multi-scale structural deconstruction of incompletely attenuated perturbation signals, enabling the building structural health monitoring system to possess stronger adaptive interference suppression and trend perception capabilities. In the residual domain, anomalous amplitude segments are dissipated step-by-step, effectively reducing the impact risk caused by the superposition of multi-source perturbation peaks. Simultaneously, by dividing real-time monitoring data into two structural categories—high-frequency micro-perturbations and low-frequency trends—the feedback strength of low-frequency trends is specifically enhanced, suppressing numerical interference from high-frequency components, significantly improving the smoothness and reliability of health assessment results. In the coupled calculation with Building Information Modeling (BIM), a deep integration of data-driven and physical modeling is further achieved, providing more stable and accurate support for subsequent structural safety early warning and operation and maintenance decisions. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0017] Figure 1 This is a flowchart of the method for monitoring the health of building structures by coupling BIM models with real-time monitoring data, as described in this invention. Detailed Implementation

[0018] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0019] This invention provides, for example Figure 1 The building structure health monitoring method shown, which couples BIM models with real-time monitoring data, includes the following steps: Construct a buckling critical characteristic response identification mechanism: acquire real-time monitoring data of building structures, integrate strain growth rate, stiffness degradation amplitude and displacement evolution trajectory in the monitoring data, extract key characteristic parameters reflecting the nonlinear evolution trend of the structure, and generate buckling evolution sensitive factors for dynamically characterizing the structure approaching buckling state. To construct a buckling critical characteristic response identification mechanism, it is first necessary to obtain complete and high-precision real-time monitoring data from the operation of the building structure. To achieve this, various types of sensors can be deployed at key stress-bearing components of the building structure. For example, high-precision resistance strain gauges can be attached to the surfaces of main load-bearing beams and columns to collect strain data; laser displacement gauges or fiber optic displacement sensors can be installed at structural nodes to collect relative displacement data; and three-dimensional force sensors can be deployed in some key areas to obtain external load input data. During monitoring, a unified clock signal should be used to strictly synchronize the data collected by each sensor to avoid phase misalignment caused by sampling delay. The collected data often contains high-frequency noise, electromagnetic interference, or abnormal deviations at individual sampling points. To ensure the continuity and authenticity of the data, high-frequency noise can be removed first using wavelet multi-scale decomposition, then local fluctuations can be smoothed using a moving average filter, and finally, missing or outlier values ​​can be supplemented using an interpolation-based repair method. Through this process, a multi-source response dataset that is perfectly aligned in time, has controllable noise levels, and good continuity can be obtained, serving as the basis for subsequent analysis.

[0020] After obtaining high-quality monitoring data, it is necessary to fuse response data from different dimensions such as strain, displacement, and stiffness to establish a multidimensional response feature space that can comprehensively characterize the buckling critical state. Specifically, the processing steps for strain data involve calculating the growth rate over a continuous time period. This involves obtaining the instantaneous strain change through first-order differencing in the collected strain time series, then averaging the values ​​using a sliding window to eliminate instantaneous fluctuations and obtain a stable strain growth rate curve. For displacement data, it is necessary to calculate the evolution trajectory of displacement at each node over time. The displacement increment can be obtained using a differencing method, and the curvature and deviation trend of the trajectory curve can be analyzed using polynomial fitting to determine the geometric deformation characteristics of the structure. For stiffness characteristics, the equivalent stiffness value can be calculated in real time under known load input and displacement response conditions. The stiffness degradation rate is obtained by comparing the changes in stiffness values ​​between adjacent time periods. After extracting the three indicators, to ensure comparability between parameters of different dimensions, it is necessary to normalize the strain growth rate, displacement evolution parameters, and stiffness degradation amplitude, unifying their numerical ranges to the same interval. Subsequently, these normalized feature parameters are combined into a multidimensional feature vector, and the principal component analysis method is used to compress redundant information, retaining only the main features that can explain most of the change trends, thus forming a multidimensional response feature space that can centrally reflect the structural approach buckling process.

[0021] After constructing the multidimensional response feature space, it is necessary to identify and extract the key feature parameters that truly characterize the nonlinear evolution trend of the structure. Specifically, this is achieved by first using dynamic time warping to align the currently acquired multidimensional feature trajectories with the historical feature trajectories of the structure during its normal service phase, identifying the feature components with the largest deviations, thereby determining which parameters have undergone anomalous evolution. To further verify whether the evolution trends of these parameters are significant, the Kullback-Leibler divergence calculation method can be introduced to compare the feature distributions at different time points, quantify the magnitude of distribution differences, and reflect the degree of abrupt changes in state. For the identified significantly changed parameters, it is also necessary to further calculate their gradients in the time dimension, i.e., to calculate the derivative of the parameter with time at each time point, to identify the trend of accelerated evolution near the critical point. To avoid oversensitivity to fluctuations far from the critical point, a time increment amplification factor needs to be introduced when calculating the gradient. This factor assigns higher weight to the change trend in the time interval near the buckling critical point, thereby highlighting the feature changes at key moments. Through this series of processes, a set of key parameters with high sensitivity and obvious trends can be extracted, which can accurately reflect the critical approximation characteristics of the structure in the nonlinear evolution stage.

[0022] After extracting the key feature parameters, these parameters need to be further integrated into a buckling evolution sensitivity factor that can dynamically quantify the critical evolution trend of the structure. Specifically, this is achieved by first weighting and superimposing all key feature parameters, with the weights determined based on the contribution of each parameter to the buckling trend in historical data. For example, the strain growth rate can be assigned a higher weight, while the displacement trajectory deviation and stiffness degradation magnitude are assigned medium weights based on their sensitivity to the overall buckling characteristics. Secondly, to ensure that the buckling evolution sensitivity factor reflects the characteristics of nonlinear acceleration as it approaches the critical state, a hyperbolic tangent function or an exponential function can be used as a nonlinear amplification function, causing the output value of the buckling evolution sensitivity factor to rise rapidly when the parameters reach a threshold. Then, the geometric information, component boundary conditions, and connectivity relationships from the BIM model are incorporated into the weighting process of the buckling evolution sensitivity factor, ensuring that the factor not only reflects the overall evolution trend but also accurately maps spatially to specific components. Finally, the temporal evolution results of the buckling evolution sensitivity factor are combined with the spatial distribution results to generate a heat map visualization in the BIM 3D environment. This makes the component areas with higher buckling evolution sensitivity factor values ​​appear as darker hot spots in the model, thus providing engineers with an intuitive view of the critical approach trend of the structure in different regions.

[0023] The continuous output curve of this buckling evolution sensitive factor can not only characterize the overall evolution state of the building structure in the critical buckling process in real time, but also provide data support and spatial positioning basis for subsequent disturbance response analysis and phase control, thus providing reliable, forward-looking and accurate technical support for building structure health monitoring.

[0024] Construct a disturbance response mapping mechanism: map buckling evolution sensitive factors to the disturbance gain domain, identify the high excitation potential region corresponding to the buckling evolution sensitive factors through a preset response amplification function, and locate the local component region that is easy to trigger disturbance amplification in the BIM model of the building structure; To construct a perturbation response mapping mechanism, the buckling evolution sensitivity factor generated in the previous step needs to be further incorporated into the characterization of the perturbation gain domain. Combined with the mechanical transfer characteristics of the building structure at different component locations, a complete mapping process from time-series parameters to spatial region location is established. This process mainly includes the following sub-steps: The buckling evolution sensitivity factor results are incorporated into the construction of the perturbation gain domain. Specifically, the buckling evolution sensitivity factor sequence obtained in the previous step is input into a dynamic normalization process in a time-continuous manner. During normalization, not only the numerical range of the buckling evolution sensitivity factor itself is considered, but also the statistical distribution under historical operating conditions is taken into account, so that the relative difference between the extreme values ​​and normal values ​​of the buckling evolution sensitivity factor can be amplified. Subsequently, the normalized buckling evolution sensitivity factor is mapped into the perturbation gain domain, which is defined as a spatiotemporal function field describing the structure's local ability to amplify perturbations. The value of this function field is determined by the combined amplitude and rate of change of the buckling evolution sensitivity factor. When the buckling evolution sensitivity factor shows a rapid upward trend, the corresponding perturbation gain domain will be assigned a higher response potential value within the corresponding time interval, thus laying the data foundation for the subsequent triggering of the amplification function.

[0025] After constructing the perturbation gain domain, it is necessary to identify the high excitation potential regions corresponding to buckling evolution sensitive factors using a preset response amplification function. Specifically, a response amplification function with nonlinear amplification characteristics is first defined. This function can be a hyperbolic tangent function or a piecewise exponential function, ensuring that the function output rises rapidly when the perturbation gain value approaches a threshold, thus highlighting the sensitivity to potential risks. The normalized buckling evolution sensitive factors are input into the response amplification function, and their amplified response values ​​over different time periods are calculated. These results are then compared point-by-point with the perturbation gain domain. When the amplified response value of a certain region exceeds the set threshold, that region is determined to be a high excitation potential region. In this process, not only the numerical value of the buckling evolution sensitive factor at a single point is considered, but its rate of change within a certain time window is also weighted, so that buckling evolution sensitive factors with a sustained upward trend receive a higher amplification weight. This process results in a clear high excitation potential distribution pattern within the perturbation gain domain, providing a basis for subsequent spatial positioning.

[0026] After identifying regions with high excitation potential, spatial localization is required within the 3D information model of the building structure to determine local component areas prone to perturbation amplification. Specifically, the coordinate information of the high excitation potential regions is registered with the geometric topological relationships of the components in the 3D information model. First, a correspondence is established between monitoring points and building model nodes, precisely aligning the sensor coordinates of the buckling evolution sensitivity factors with the component elements in the model. Second, considering the boundary conditions between different components, the numerical results of the high excitation potential are mapped to the surfaces or junctions of adjacent components using a finite element mesh topology. Finally, by calculating the stiffness distribution and constraint conditions of each component within this region, the local components most likely to experience perturbation amplification are selected. These selected local components are visually marked in the 3D information model using color gradients or increased brightness, enabling engineers to clearly identify areas requiring focused monitoring and reinforcement. This spatial localization process tightly integrates abstract buckling evolution sensitivity factors with their physical locations within the actual building structure, providing accurate input conditions for subsequent perturbation phase modulation and mode transfer.

[0027] Through the above steps, the entire process is ultimately realized, from time-series data of buckling evolution sensitive factors to the construction of the perturbation gain domain, and then to the identification of the response amplification function and the localization of the building's 3D information model. This implementation method can not only dynamically capture potential high-excitation regions, but also accurately map them onto specific components of the building structure, providing reliable data support and spatial basis for subsequent technical measures to mitigate the perturbation amplification effect.

[0028] Construct a phase modulation mechanism for disturbance signals: Based on the temporal evolution characteristics of the buckling evolution sensitive factors in the high excitation potential region, the phase of the disturbance signal in the real-time monitoring data is fine-tuned to regulate its coupling rhythm with the nonlinear response of the structure, so as to weaken the temporal overlap effect of the peak of multi-source disturbances. To achieve a phase modulation mechanism for the perturbation signal, based on the identified high excitation potential region, the temporal evolution characteristics of the buckling evolution sensitive factors within this region are utilized to continuously adjust the phase and optimize the rhythm of the perturbation signal carried in the real-time monitoring data. This process effectively alters the coupling relationship between the perturbation signal and the structural nonlinear response, preventing peak superposition of multi-source perturbation signals in the time domain, thereby weakening the nonlinear amplification effect. This implementation method is achieved through the following steps: Based on the time-series characteristics of the buckling evolution sensitive factor in the high excitation potential region, an initial phase reference value for the perturbation signal is determined. Specifically, the evolution curve of the buckling evolution sensitive factor in this region is subjected to a Fast Fourier Transform (FFT), and the phase information of its main energy concentration frequency band is extracted as the reference value. Simultaneously, combined with the actual phase characteristics of the perturbation signal in the monitoring data, the phase difference between the two is calculated, and this phase difference is used as the benchmark for subsequent adjustments. In this way, a preliminary phase alignment relationship between the perturbation signal and the buckling evolution characteristics can be established, giving subsequent control a clear direction and target.

[0029] After obtaining the phase reference value, the perturbation signal needs to be reconstructed in time sequence to create conditions for phase fine-tuning. Specifically, the perturbation signal in the monitoring data is segmented within a time window to ensure that the signal characteristics remain relatively stable within each time period. A sliding window approach is used during segmentation, ensuring that the window size matches an integer multiple of the dominant frequency of the buckling evolution sensitivity factor to ensure consistency between phase modulation and the dominant rhythm of the nonlinear response. Subsequently, within each time window, the Hilbert transform is used to construct the analytical form of the signal, thereby obtaining a continuous function of the instantaneous phase change of the signal. This process not only ensures phase continuity but also provides a mathematical basis for subsequent precise adjustment.

[0030] After completing the time-series reconstruction, phase fine-tuning of the perturbation signal is required. Specifically, the instantaneous phase of the perturbation signal is compared point-by-point with the buckling evolution sensitivity factor phase reference value obtained in the previous step, and the phase of the perturbation signal is dynamically adjusted based on the difference. If the phase difference is too large, phase correction is achieved by introducing a small time delay or advance in the signal time series; if the phase difference is small, fine adjustment is performed using a nonlinear phase offset function. The offset function can employ sinusoidal modulation or a hyperbolic tangent function to ensure a smooth transition rather than abrupt changes in the phase adjustment process, thereby avoiding the generation of new high-frequency interference. During this process, it is also necessary to ensure that the amplitude of the phase-adjusted signal remains stable to prevent unnecessary energy fluctuations introduced by the phase correction.

[0031] After phase fine-tuning, the coupling rhythm between multiple perturbation signals needs to be comprehensively controlled to avoid peak superposition of multiple source signals in the time domain. Specifically, all perturbation signals originating from high excitation potential regions are superimposed and analyzed according to the phase fine-tuning results, and their phase correlation is calculated using a correlation function. When multiple signals tend to produce peak responses at the same time point, a small phase delay is applied to some signals to offset their peaks on the time axis, thereby weakening the superposition effect. This control process needs to be dynamically executed; that is, under the condition of continuous real-time monitoring data input, the phase change trend of each signal is continuously tracked, and corrections are implemented immediately when potential overlap risks are detected. Through this step, unrealistic high-amplitude responses caused by the concentrated superposition of multiple source perturbations in local components can be effectively avoided.

[0032] After completing the coupling and modulation of multi-source signals, the phase modulation results need to be dynamically verified to ensure that the modulated disturbance signal can form a reasonable rhythmic match with the nonlinear response of the structure. Specifically, the modulated signal is re-inputted into the nonlinear calculation combined with the BIM model, and its performance on the structural response curve is observed. If, in the high excitation potential region, the modulated response curve shows a reduced peak amplitude, a more stable change, and no more abnormal transient spikes, then the phase modulation has achieved the expected effect. Simultaneously, this result needs to be compared with the response result before modulation to quantify the degree of improvement brought about by the modulation, including indicators such as the peak reduction ratio and the degree of energy distribution optimization. If local response anomalies are still found, the precision of the phase fine-tuning needs to be further refined to achieve a higher level of response stability.

[0033] By implementing the above steps, the perturbation signal in the high excitation potential region can be accurately matched with the temporal evolution characteristics of the buckling evolution sensitive factor. Furthermore, through continuous phase fine-tuning and rhythm optimization, the peak overlap effect of multi-source perturbation signals in the time domain is weakened. This not only improves the accuracy of building structural health monitoring but also provides more stable input conditions for subsequent modal response transfer and residual energy buffering.

[0034] Construct a modal response transfer mechanism: Based on the phase-modulated perturbation signal, guide it to a highly flexible local modal path in the high excitation potential region, and set a response threshold attenuation condition to suppress modal crosstalk and cascade amplification effects triggered by perturbation in non-critical regions; To achieve the modal response transfer mechanism, it is necessary to fully utilize the phase-tuned perturbation signal and, in conjunction with the spatial characteristics of the high excitation potential region, rationally guide the perturbation signal to a highly flexible local modal path. Simultaneously, to avoid meaningless modal crosstalk or energy cascade amplification in non-critical regions under perturbation, a clear response threshold attenuation condition must be established to ensure that energy attenuates within a reasonable range and is guided to an acceptable local modal channel. The entire process is implemented through the following steps: Identifying highly flexible local modal paths within the overall modal distribution of a building structure requires finite element vibration analysis of key components in the 3D building information model to extract global modal shapes and their corresponding natural frequencies. During the calculation, special attention should be paid to modes occurring in the low-frequency range and accompanied by large displacement amplitudes, as these modes typically correspond to highly flexible local components. Next, component elements within high excitation potential regions are projected one by one onto the global modal shape, and their displacement participation coefficients and strain energy distributions under different modes are calculated. If a mode exhibits high displacement participation and energy concentration within a region, it can be defined as a preferred flexible modal path. This approach establishes a precise mapping relationship from locally sensitive components to global modes, providing a physical channel for guiding disturbance signals.

[0035] It should be noted that: The displacement participation factor (FPD) is a quantitative indicator used in structural dynamics analysis to measure the contribution of a particular mode to a specific displacement response of a structure. It reflects the proportion of the displacement amplitude of a local node or component relative to the overall modal response under a given mode shape, thus indicating the dominance of that mode in the displacement response at that location. The method for obtaining FPD is typically: First, the modal shape vectors of the building structure are obtained through finite element modal analysis, with each component corresponding to the displacement amplitude of each node in that mode. Then, the displacement components of a target node or component are selected, and their ratios are calculated with the normalized results of the overall modal shape vectors to obtain the displacement participation coefficient at that location. If a regional index is required, a weighted average of the displacement components of multiple nodes within the region can be taken. The displacement participation coefficient obtained in this way can intuitively characterize the degree of participation of a certain locality in different modes, which helps to identify highly flexible local modal paths and provides a quantitative basis for disturbance signal guidance and modal response transfer.

[0036] After identifying the local modal path with high flexibility, the phase-tuned perturbation signal is matched with this modal path and guided into input. Specifically, the perturbation signal undergoes spectral analysis in the frequency domain to extract its main energy concentration frequency components, which are then compared with the flexibility mode frequency range obtained in the previous step. If the dominant frequency component of the perturbation signal matches or is close to the flexibility mode frequency, the signal can be projected onto this modal path using a frequency-weighted method. During the projection process, the modal orthogonality principle is employed to decompose the input quantity of the perturbation signal according to the direction of the mode shape, retaining only the component corresponding to the flexibility mode path. This ensures that the perturbation energy is preferentially introduced into the high flexibility region and does not diffuse to other irrelevant modes. This process not only guarantees the directional transmission of the perturbation signal but also lays the foundation for subsequent energy attenuation.

[0037] After the disturbance signal is guided to the flexibility modal path, a response threshold attenuation condition is set to effectively control modal crosstalk that may be triggered in non-critical areas. Specifically, an energy threshold can be set in the modal projection calculation. When the response energy of a modal component is lower than this threshold, it is automatically attenuated to zero to avoid meaninglessly transferring energy to non-critical modes. Simultaneously, for modal components exceeding the threshold, a nonlinear attenuation function can be applied, such as a hyperbolic tangent function that gradually saturates with increasing input amplitude, so that the attenuation rate gradually increases as the energy increases, thereby avoiding modal cascading amplification. At the spatial level, it is also necessary to combine the boundary conditions and stress states of the components in the BIM model to further screen out local components capable of accommodating the disturbance energy, and retain some modal responses on these components to achieve energy dispersion and dissipation. In this way, it can be ensured that the disturbance signal, while being guided, will not trigger additional modal resonance or coupling effects in non-critical areas of the structure.

[0038] After completing modal path guidance and threshold attenuation control, the effectiveness of the entire process is dynamically verified and adjusted based on feedback. Specifically, the perturbation signal, after projection and attenuation processing, is re-inputted into the overall structural dynamic response calculation, and the changes in response amplitude in the high excitation potential region and non-critical regions are compared. If the results show that the compliance mode in the high excitation potential region can effectively absorb most of the perturbation energy, and the response amplitude in the non-critical region is significantly weakened, it indicates that the modal response transfer mechanism has achieved the expected effect. Based on this, the time-series evolution curve of the buckling evolution sensitivity factor needs to be re-analyzed to verify whether the peak value of the perturbation signal is smoothed and whether the nonlinear amplification effect is weakened. If residual anomalous responses still exist in individual regions, the suppression effect can be further enhanced by dynamically adjusting the energy threshold or optimizing the parameters of the attenuation function. Through this feedback mechanism, it can be ensured that the modal response transfer not only holds true at the theoretical level but also possesses long-term stability and adaptability in actual operation.

[0039] By implementing the above steps, the phase-modulated disturbance signal can be accurately guided to a highly flexible local modal path. Simultaneously, the response threshold attenuation condition suppresses modal crosstalk and energy cascade amplification effects in non-critical areas. This implementation not only ensures a reasonable distribution of the disturbance signal in both spatial and modal dimensions but also enhances the robustness and disturbance resistance of the building structure under critical buckling conditions. It provides more stable and controllable input conditions for subsequent residual energy buffering mechanisms, thereby effectively improving the accuracy and reliability of building structural health monitoring.

[0040] Constructing a residual energy buffer mechanism: After completing the modal response transfer, extract the amplitude anomaly segments that are not fully decayed in the disturbance signal, inject them into the virtual residual domain, and perform a step-by-step dissipation operation in the residual domain to reduce the risk of violent abrupt changes in the main structure response channel; To implement the residual energy buffering mechanism, after completing the modal response transfer, additional processing is needed for the incompletely decayed amplitude anomaly segments remaining in the disturbance signal. Although the signal has been guided by the compliance mode path and threshold decay control, it may still retain some transient energy peaks in local regions. If this energy is directly fed back to the main structural response channel, it could cause drastic changes in the structure within a short period. Therefore, these undecayed signal segments need to be transferred to the virtual residual domain, where a stepwise dissipation operation is performed until the energy is gradually released and dissipated, thereby reducing the risk of abrupt changes in the main channel response. This process is achieved through the following steps: First, residual energy is extracted from the signal after modal response transfer. In this step, the perturbation signal, after transfer and threshold attenuation, is scanned across its entire time series, and its instantaneous amplitude is analyzed using a sliding window. Local segments exceeding a set threshold are marked as amplitude anomalous segments. The threshold can be dynamically adjusted based on changes in the buckling evolution sensitivity factor of the structure. For example, when the buckling evolution sensitivity factor shows a rapid upward trend, the threshold should be lowered to capture more potential risk signals, while when the buckling evolution sensitivity factor tends to stabilize, the threshold can be increased to reduce unnecessary energy segment extraction. In this way, selective capture of residual energy can be achieved, ensuring that subsequent processing only targets parts that may truly threaten structural stability.

[0041] The extracted amplitude anomaly fragments are injected into the virtual residual domain. Specifically, the anomaly fragments are separated from the original main channel signal, and their corresponding time-series trajectories are established in an independent residual domain. During the injection process, the time-series continuity of the residual signal must be maintained so that it can accurately reflect the dynamic characteristics of the original disturbance within the residual domain. Simultaneously, the signal amplitude needs to be recalibrated during injection, scaling it proportionally to the standard energy scale of the residual domain to avoid instability in the residual domain calculations due to excessively large amplitudes. After injection, the signal in the residual domain can be considered as an energy reserve that needs further dissipation; its existence is isolated from the main channel, preventing it from directly affecting the nonlinear response of the building structure.

[0042] In the residual domain, a stepwise dissipation operation is performed to dissipate energy in segments. Specifically, the residual signal is divided into multiple frequency bandwidth intervals, and different dissipation functions are applied to each. For example, an exponential decay function can be used for the high-frequency residual signal to release its energy rapidly; for the low-frequency residual signal, a slowly decreasing logarithmic decay function can be used to gradually and smoothly dissipate it. The advantage of segmented dissipation is that it ensures that the residual energy is not released all at once, but rather that the dissipation rate is distributed in an orderly manner according to the characteristics of different frequency bands. Simultaneously, during the stepwise dissipation process, a time-weighted decay factor can be superimposed, so that energy segments approaching the peak of the buckling evolution sensitivity factor are preferentially dissipated, thereby minimizing the impact on the main structural response.

[0043] Feedback monitoring is performed on the stepwise dissipation results in the residual domain, and the results are cross-corrected with the response state of the main channel. During this process, the energy dissipation amount in each time segment of the residual domain is compared with the evolution trend of the buckling evolution sensitivity factor in the main channel to determine whether the residual energy is effectively absorbed. If the buckling evolution sensitivity factor value in the main channel tends to stabilize after residual dissipation, it indicates a significant energy buffering effect. If abnormal amplitude fluctuations still exist in the main channel, further adjustments to the dissipation function parameters in the residual domain are needed, such as increasing the attenuation rate of the high-frequency component or extending the dissipation period of the low-frequency component. Through this feedback mechanism, dynamic coupling between residual energy processing and the main channel response can be achieved, ensuring that the improvement in structural stability due to residual buffering is real-time and controllable.

[0044] After completing residual dissipation and feedback monitoring, a closed-loop verification of the entire residual energy buffering process is required to ensure its long-term effectiveness. Specifically, the main channel response curve is overlaid with the residual domain energy dissipation curve for analysis, observing the energy transfer and release at different stages. If the residual energy can be completely dissipated within a preset time range, and the main channel response no longer exhibits drastic abrupt changes, it indicates that the residual energy buffering mechanism has achieved its intended goal. Simultaneously, the parameters of the residual domain energy dissipation process can be archived as a reference standard for subsequent monitoring, enabling the rapid application of appropriate dissipation strategies when similar disturbances recur. Through this long-term accumulation, the residual energy buffering mechanism can be continuously optimized through multiple runs, gradually forming a stable and mature energy management process.

[0045] By implementing the above steps, incompletely decayed amplitude anomaly segments in the disturbance signal can be effectively processed after modal response transfer, and dissipated step by step in the virtual residual domain, thereby avoiding drastic changes in the main structure response channel due to transient energy concentration. This implementation not only ensures the dispersion and controllability of energy dissipation, but also improves the stability and robustness of the overall monitoring method.

[0046] Constructing a disturbance frequency band deconstruction mechanism: Based on the output of the residual energy buffer mechanism, the fractal disturbance decomposition method is used to divide the real-time monitoring data into two types of structures: high-frequency micro-disturbance and low-frequency trend. In the coupled calculation with the BIM model, the feedback intensity of the low-frequency trend is enhanced, while a nonlinear scaling suppression strategy is applied to the high-frequency micro-disturbance signal, thereby improving the stability and credibility of the structural health assessment results. To implement the perturbation frequency band deconstruction mechanism, based on the residual energy buffering mechanism, the buffered output signal needs further hierarchical frequency band decomposition and targeted processing to prevent residual perturbations from interfering with the stability of building structural health monitoring in different frequency ranges. By employing a fractal perturbation decomposition method, real-time monitoring data is divided into two structural categories: high-frequency perturbations and low-frequency trends. Combined with different processing strategies, this ensures that the low-frequency trend signal receives enhanced feedback in the coupled calculation with the building's 3D information model, while simultaneously controlling high-frequency perturbations through a nonlinear scaling suppression strategy. This improves the stability and reliability of the overall structural health assessment results. This process can be achieved through the following steps: Based on the output of the residual energy buffering mechanism, fractal perturbation decomposition is performed on real-time monitoring data. Specifically, the signal after residual energy buffering is input into a fractal decomposition framework. This framework utilizes a combination of fractal dimension analysis and multi-scale reconstruction to decompose the overall signal's shape at multiple levels. In this way, the signal can be automatically separated into high-frequency and low-frequency components according to its frequency characteristics. The high-frequency components mainly exhibit small-amplitude but rapidly changing transient perturbations, while the low-frequency components exhibit slowly evolving and clearly trending structural response trajectories. During the decomposition process, to ensure the accuracy of the separation results, a hybrid method of wavelet packet decomposition and empirical mode decomposition is used to unfold the signal at multiple scales, and the optimal decomposition level is selected based on the energy distribution law. The result is a set of high-frequency perturbation sequences and a low-frequency trend curve, which are independent yet together constitute a complete representation of the original signal.

[0047] After completing the fractal perturbation decomposition, the processing method for low-frequency trend signals needs to be optimized. Since low-frequency trends often correspond to the overall structural stress evolution and buckling approximation process, they should be given higher feedback intensity in the coupled calculations with the building's 3D information model. Specifically, the low-frequency trend curve is matched with the mechanical response parameters of the corresponding components in the building model, and the feedback value of the low-frequency trend signal is weighted using an amplification function. For example, when the low-frequency signal reflects a continuous increase in strain or displacement, its influence weight should be amplified in the model calculation to highlight the potential critical state of the structure. This enhancement not only allows the model to better reflect the actual structural evolution trend in calculations but also provides early warnings when the critical state approaches, thereby improving the foresight of health assessments.

[0048] While strengthening low-frequency trend feedback, it is also necessary to apply a nonlinear scaling suppression strategy to the high-frequency perturbation signal to avoid random fluctuations in the high-frequency component interfering with the overall health assessment. Specifically, the amplitude distribution of the high-frequency perturbation signal is first statistically modeled to identify peak segments and random noise components. Then, a nonlinear scaling function, such as a hyperbolic tangent function or a piecewise power function, is introduced to the peak segments to compress large transient perturbations into a smaller range, thereby significantly reducing their impact on the overall calculation results. For random noise components, an energy threshold-based suppression method can be used to directly zero out portions below a certain threshold, preventing their repeated accumulation in the structural response calculation. During this process, it is also necessary to maintain the time series continuity of the high-frequency signal to avoid disrupting the overall temporal correlation of the signal during suppression. In this way, the interference effect of high-frequency perturbations can be effectively suppressed, making the model calculation results more stable.

[0049] After enhancing the low-frequency trend and suppressing the high-frequency perturbation, the processed results are re-inputted into the building's 3D information model for coupled calculation, and the health assessment results are verified. Specifically, the enhanced low-frequency trend signal and the suppressed high-frequency perturbation signal are synthesized into a new input sequence, which drives the dynamic response calculation of the building structure. After the calculation, the changes in the structural buckling evolution sensitivity factors before and after the coupled calculation are compared to verify whether the processing can effectively eliminate false instability indicators and smooth the numerical curve. If the response in the high excitation potential region shows a clear trend and stable amplitude, and no abnormal peaks appear in non-critical regions, it indicates that the perturbation band deconstruction mechanism has played its expected role. Simultaneously, the energy release in the residual domain and the evolution curve of the low-frequency trend can be jointly analyzed to evaluate the stability and adaptability of the mechanism under long-term monitoring. If the results show that the high-frequency component is successfully suppressed and the low-frequency component is significantly enhanced, this method not only improves the credibility of structural health assessment but also provides reliable data support for subsequent early warning and maintenance decisions.

[0050] By implementing the above steps, frequency band deconstruction and hierarchical processing of disturbance signals can be achieved based on residual energy buffering. This allows for full utilization of the early warning value of low-frequency trend signals during structural health monitoring, while minimizing the interference effect of high-frequency micro-disturbances. This method not only makes the assessment results more stable and reliable but also ensures the credibility of the building's 3D information model in coupled calculations with real-time monitoring data.

[0051] This invention effectively solves the problem of nonlinear amplification of small-amplitude disturbance signals in building structures approaching the critical buckling state by constructing a disturbance response identification and control mechanism centered on buckling evolution sensitive factors. It achieves precise control of the disturbance signal throughout its temporal rhythm, modal path, and energy transfer processes. Before the disturbance enters the main structural response channel, disturbance source identification, amplification trend assessment, and energy guidance processing are completed, fundamentally avoiding misjudgments of structural safety due to false triggering under high sensitivity conditions. By identifying highly flexible modal paths within high excitation potential regions and setting response threshold attenuation conditions, disturbance energy is effectively confined to specific local areas, preventing diffusion to non-critical components, thereby significantly improving the robustness and response stability of the health monitoring system under critical conditions.

[0052] This invention introduces a virtual residual domain and fractal perturbation decomposition method to achieve energy buffering and multi-scale structural deconstruction of incompletely attenuated perturbation signals, enabling the building structural health monitoring system to possess stronger adaptive interference suppression and trend perception capabilities. In the residual domain, anomalous amplitude segments are dissipated step-by-step, effectively reducing the impact risk caused by the superposition of multi-source perturbation peaks. Simultaneously, by dividing real-time monitoring data into two structural categories—high-frequency micro-perturbations and low-frequency trends—the feedback strength of low-frequency trends is specifically enhanced, suppressing numerical interference from high-frequency components, significantly improving the smoothness and reliability of health assessment results. In the coupled calculation with Building Information Modeling (BIM), a deep integration of data-driven and physical modeling is further achieved, providing more stable and accurate support for subsequent structural safety early warning and operation and maintenance decisions.

[0053] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A method for monitoring the health of building structures by coupling BIM models with real-time monitoring data, characterized in that, Includes the following steps: Real-time monitoring data of building structures are acquired, and strain growth rate, stiffness degradation magnitude and displacement evolution trajectory in the monitoring data are integrated to extract key characteristic parameters that reflect the nonlinear evolution trend of the structure and generate buckling evolution sensitive factors for dynamically characterizing the structure's approach to buckling state. The buckling evolution sensitive factor is mapped to the perturbation gain domain. The high excitation potential region corresponding to the buckling evolution sensitive factor is identified by the preset response amplification function, and the local component region that is easy to trigger perturbation amplification is located in the BIM model of the building structure. Based on the temporal evolution characteristics of the buckling evolution sensitive factors in the high excitation potential region, the phase of the disturbance signal in the real-time monitoring data is fine-tuned to regulate its coupling rhythm with the nonlinear response of the structure. Based on the phase-modulated perturbation signal, it is guided to a highly flexible local modal path in the high excitation potential region, and a response threshold attenuation condition is set to suppress modal crosstalk and cascade amplification effects in non-critical regions. After completing the modal response transfer, extract the amplitude anomaly segments that are not fully decayed in the disturbance signal, inject them into the virtual residual domain, and perform a step-by-step dissipation operation in the residual domain. Based on the output results of the residual energy buffering mechanism, the fractal perturbation decomposition method is used to divide the real-time monitoring data into two types of structures: high-frequency perturbation and low-frequency trend. In the coupled calculation with the BIM model, the feedback strength of the low-frequency trend is enhanced, while a nonlinear scaling suppression strategy is applied to the high-frequency perturbation signal.

2. The method for monitoring the health of building structures by coupling BIM models with real-time monitoring data according to claim 1, characterized in that, Generating buckling evolution sensitivity factors includes the following steps: Real-time monitoring data of building structures is acquired, and the monitoring data is denoised, smoothed, and interpolated to obtain a time-aligned and continuous multi-source response dataset. Based on the multi-source response dataset, the strain growth rate, displacement evolution trajectory, and stiffness degradation amplitude are calculated separately, and the strain growth rate, displacement evolution trajectory, and stiffness degradation amplitude are normalized and combined into a multi-dimensional feature vector. The dynamic time warping method is used to compare the multidimensional feature vector with the historical feature trajectory, and the Kullback-Leibler divergence and time gradient are combined to calculate and identify key feature parameters to characterize the nonlinear evolution trend of the structure approaching the buckling state. The key feature parameters are weighted and superimposed, and combined with the geometric information, component boundary conditions and connection relationships in the building information model to generate buckling evolution sensitivity factors.

3. The method for monitoring the health of building structures by coupling BIM models with real-time monitoring data according to claim 2, characterized in that, In the process of generating the buckling evolution sensitive factor, a hyperbolic tangent function or an exponential function is used as a nonlinear amplification function to make the output value of the key characteristic parameter rise rapidly when it approaches the buckling critical state, thereby enhancing the buckling evolution sensitive factor's ability to characterize the critical approach trend.

4. The method for building structural health monitoring that couples BIM model with real-time monitoring data according to claim 2, characterized in that, The steps for mapping buckling evolution sensitivity factors to the perturbation gain domain include: The buckling evolution sensitive factor sequence is input into the dynamic normalization stage, and the difference between the extreme values ​​and normal values ​​of the buckling evolution sensitive factor is amplified by combining the statistical distribution under historical operating conditions. The normalized buckling evolution sensitive factor is then mapped into the perturbation gain domain. The perturbation gain domain is calculated using a preset response amplification function. When the buckling evolution sensitivity factor continues to rise within a certain time window and the amplified response value exceeds the threshold, the region is determined to be a high excitation potential region. The coordinate information of the high excitation potential region is registered with the geometric topology of the building information model to establish the correspondence between sensor coordinates and model nodes. The local component regions that are prone to triggering disturbance amplification are determined by combining the stiffness distribution and boundary conditions of the components.

5. The method for monitoring the health of building structures by coupling BIM models with real-time monitoring data according to claim 4, characterized in that, The steps for fine-tuning the phase of disturbance signals in real-time monitoring data include: Based on the time series characteristics of the buckling evolution sensitive factor in the high excitation potential region, the phase information of the main energy concentration frequency band is extracted by fast Fourier transform, and the phase difference is calculated by combining the actual phase characteristics of the disturbance signal, and this is used as the phase reference value. The disturbance signal is segmented within a time window, and the Hilbert transform is used to obtain the continuous change function of the instantaneous phase, thereby realizing time sequence reconstruction. The instantaneous phase of the disturbance signal is compared point by point with the phase reference value, and phase correction is achieved by introducing time delay, advance, or using a nonlinear phase offset function; Multiple phase-corrected disturbance signals are superimposed and analyzed to calculate phase correlation. When the peaks tend to overlap, a small phase delay is applied to make them staggered. The disturbance signal after comprehensive regulation is input into the nonlinear calculation of the building information model to verify whether the peak amplitude has decreased and the change has become stable, and the result is compared with that without regulation to quantify the improvement effect.

6. The method for monitoring the health of building structures by coupling BIM models with real-time monitoring data according to claim 1, characterized in that, The steps of guiding the phase-modulated perturbation signal to a highly flexible local modal path and setting a response threshold attenuation condition include: Finite element vibration analysis is performed on the overall modal distribution of the building structure to identify the flexibility modes with large displacement amplitude in the low frequency range. The displacement participation coefficient and strain energy distribution of the component elements in different modes in the high excitation potential region are calculated to determine the local modal path of large flexibility. The perturbation signal after phase modulation is subjected to frequency domain spectrum analysis to extract its main energy components and compare them with the frequency of the flexibility mode. Only the components of the corresponding flexibility mode path are retained by modal orthogonality projection. In modal projection calculation, an energy threshold is set, modal components below the threshold are attenuated to zero, and nonlinear attenuation functions that gradually saturate with increasing amplitude are applied to components exceeding the threshold. Combined with the component boundary conditions in the building information model, local components that can accommodate energy are selected. The disturbance signal, after projection and attenuation processing, is input into the overall structural dynamic response calculation. The response amplitude changes between the high excitation potential region and the non-critical region are compared, and dynamic feedback correction is achieved by adjusting the energy threshold or optimizing the attenuation function parameters.

7. The method for monitoring the health of building structures by coupling BIM models with real-time monitoring data according to claim 6, characterized in that, After completing the modal response transfer, the steps of extracting incompletely decayed amplitude anomaly segments from the disturbance signal and performing stepwise dissipation operations in the residual domain include: The perturbation signal after modal response transfer processing is scanned in its entire time series. An amplitude anomaly segment exceeding the threshold is extracted using a sliding window, and the threshold is dynamically adjusted according to the changes in the buckling evolution sensitivity factor. Amplitude anomaly segments are extracted from the main channel signal and injected into the virtual residual domain, maintaining temporal continuity and scaling to the standard energy scale of the residual domain during the injection process. In the residual domain, the amplitude anomalous segments are frequency-divided, an exponential decay function is applied to the high-frequency part, a logarithmic decay function is applied to the low-frequency part, and a time-weighted decay factor is superimposed to achieve stepwise dissipation. The residual domain dissipation results are compared with the evolution trend of the buckling evolution sensitive factor in the main channel. When the buckling evolution sensitive factor tends to be stable, the dissipation is deemed effective; otherwise, the dissipation function parameters are adjusted. The main channel response curve is superimposed with the residual domain energy dissipation curve for analysis to verify whether the residual energy is completely eliminated and archived as parameters within a preset time.

8. The method for monitoring the health of building structures by coupling BIM models with real-time monitoring data according to claim 7, characterized in that, The steps involved in dividing the real-time monitoring data into two types—high-frequency perturbations and low-frequency trends—based on the output of the residual energy buffering mechanism using the fractal perturbation decomposition method and processing them separately are as follows: The signal after residual energy buffering is input into the fractal decomposition framework, and multi-scale decomposition is performed by combining wavelet packet decomposition and empirical mode decomposition methods to obtain high-frequency perturbation sequence and low-frequency trend curve. The low-frequency trend curve is matched with the mechanical response parameters of the components in the building 3D information model, and the low-frequency trend feedback is weighted by an amplification function to highlight the structural evolution trend. Amplitude distribution modeling is performed on high-frequency perturbation signals, peak segments are compressed using a nonlinear scaling function, and low-amplitude random noise is suppressed using an energy thresholding method while maintaining the continuity of the time series. The enhanced low-frequency trend signal and the suppressed high-frequency perturbation signal are combined into a new input sequence, which is then re-input into the building 3D information model for coupled calculation, and it is verified whether it eliminates false instability indicators.

Images