A real-time ground settlement monitoring device for buildings and its usage method

Abstract

This application discloses a real-time building ground settlement monitoring device and its usage method, belonging to the field of building structure safety monitoring. The monitoring device includes: a hierarchical sensor network for real-time data acquisition at multiple time scales, comprehensively acquiring deformation data and related environmental parameters of the building structure; a data processing and analysis module for real-time processing and intelligent analysis of the acquired data, timely identifying and classifying abnormal deformation characteristics of the building structure; a deep learning prediction module for quantitatively predicting the probability state and evolution trend of building settlement by constructing a multi-scale time series prediction model; a multi-factor analysis module for coupled modeling and comprehensive analysis of environmental factors, structural characteristics, and abnormal evolution processes to identify key influencing factors and their mechanisms of action; and a risk assessment and early warning module for classifying and assessing building settlement risk based on prediction and analysis results, and generating corresponding early warning information and decision support schemes.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of building structure safety monitoring, more particularly, to a building ground real-time settlement monitoring device and a method thereof. BACKGROUND

[0002] Building ground settlement is an important factor affecting structural safety and functional use, and its abnormal change may cause building cracks, structural damage and even serious safety accidents. Therefore, accurate monitoring of building ground settlement is of great significance to ensure building safety. In the prior art, ground settlement monitoring mainly relies on manual measurement or periodic observation, such as leveling, total station measurement and GPS positioning. Although these methods can obtain settlement data, they have problems such as long observation period, data update lag, difficulty in covering key parts and large interference from environmental factors, which cannot meet the needs of real-time and high-precision settlement monitoring of modern buildings, especially large or high-rise buildings.

[0003] In recent years, with the development of sensor technology and data acquisition technology, automatic monitoring means has been gradually applied to building settlement monitoring. For example, a sensor array is arranged at key parts of the building foundation and load-bearing structure to realize continuous data acquisition, and a data processing module is used for foundation analysis and trend judgment. However, the existing methods still have some deficiencies in practical application: first, a single type of sensor or a simple sensor network cannot simultaneously capture the instantaneous abnormality, short-term fluctuation and long-term evolution trend of building settlement; second, the data analysis method relies on empirical formula or simple statistical model, and it is difficult to fully utilize multi-source, multi-dimensional and high-frequency data to comprehensively analyze the coupling relationship between environmental factors, structural characteristics and settlement evolution; third, there is a lack of reliable risk classification evaluation and dynamic early warning means, which cannot provide real-time and operable decision basis for management and emergency.

[0004] In summary, how to realize the multi-time scale real-time monitoring, abnormality identification, evolution trend prediction and risk classification early warning of building ground settlement has become a technical problem to be solved. SUMMARY

[0005] In order to overcome the series of defects existing in the prior art, the purpose of the present application is to provide a building ground real-time settlement monitoring device, which comprises the following modules.

[0006] Hierarchical sensor network for carrying out real-time data acquisition of multi-time scale and comprehensively obtaining deformation data of building structure and related environmental parameters.

[0007] Data processing and analysis module for real-time processing and intelligent analysis of collected data, and timely identification and classification of abnormal deformation characteristics of building structure.

[0008] The deep learning prediction module quantitatively predicts the probabilistic state and evolution trend of building settlement by constructing a multi-scale time series prediction model.

[0009] The multi-factor analysis module is used to perform coupled modeling and comprehensive analysis of environmental factors, structural characteristics and abnormal evolution processes in order to identify key influencing factors and their mechanisms of action.

[0010] Risk assessment and early warning module: Based on the prediction and analysis results, the risk of building settlement is classified and assessed, and corresponding early warning information and decision support solutions are generated.

[0011] Furthermore, the hierarchical sensor network includes a base layer sensor array, structural layer sensor nodes, and an environmental monitoring sensor group. The base layer sensor array combines high-precision laser displacement sensors with digital tilt sensors, and is installed in key parts of the building foundation according to a grid layout principle. The sensor spacing is set within the range of 5 to 15 meters based on the building span to ensure the capture of minute deformations at the 0.1 mm level. The structural layer sensor nodes use wireless strain sensors and triaxial accelerometers, deployed along the main load-bearing structure of the building to achieve real-time monitoring of structural stress state changes and dynamic responses. The environmental monitoring sensor group includes soil moisture sensors, groundwater level monitoring sensors, temperature and humidity sensors, and ground vibration sensors, used to acquire external environmental parameters affecting building settlement.

[0012] Furthermore, the data processing and analysis module includes a data preprocessing unit, a feature extraction unit, and an anomaly detection unit. Specifically: the data preprocessing unit performs noise filtering and signal enhancement on the raw sensor data, while correcting sensor drift and suppressing environmental interference; the feature extraction unit extracts multi-frequency domain features based on wavelet transform and combines principal component analysis to reduce data dimensionality, thereby extracting key feature parameters that characterize the building's deformation state; the anomaly detection unit establishes dynamic control limits using statistical process control methods, identifies abnormal deformation patterns by comparing with historical benchmark data, triggers an anomaly warning when three consecutive sampling points exceed the control limits, and automatically adjusts the detection sensitivity according to different settlement development stages to meet monitoring requirements.

[0013] Furthermore, the deep learning prediction module adopts a hybrid architecture combining a long short-term memory neural network and an attention mechanism. It captures the temporal dependencies of settlement data by constructing a multi-layer LSTM network, while the attention mechanism is used to adaptively allocate weights for data at different time steps, thereby improving the prediction accuracy for key time nodes. The deep learning prediction module includes a data normalization layer, a feature encoding layer, a temporal modeling layer, and an output decoding layer. Specifically, the data normalization layer standardizes the input multi-dimensional temporal data, the feature encoding layer converts sensor data into high-dimensional feature vectors, the temporal modeling layer learns the dynamic pattern of settlement evolution through a bidirectional long short-term memory neural network, and the output decoding layer generates the predicted settlement probability distribution within the future time window.

[0014] Furthermore, the multi-factor analysis module includes an environmental factor impact assessment unit, a structural characteristic analysis unit, and a settlement mechanism identification unit. Specifically: the environmental factor impact assessment unit is used to quantitatively analyze the effects of various environmental factors on settlement and establish their response functions and sensitivity coefficients; the structural characteristic analysis unit assesses the mechanical state and safety reserve level of the building structure through structural mechanics modeling and measured data correction; and the settlement mechanism identification unit analyzes the spatiotemporal evolution characteristics of settlement through data mining and pattern recognition, identifies dominant factors and triggering conditions, extracts typical settlement patterns, and establishes discrimination criteria.

[0015] Furthermore, the multi-factor analysis module also includes a dynamic weight allocation mechanism implementation unit, which is used to adaptively adjust the weight coefficients of each influencing factor according to different stages of settlement evolution. Its specific implementation includes the following steps.

[0016] By calculating the covariance matrix between each influencing factor and the settlement response in real time, the instantaneous influence intensity of the factors is dynamically assessed.

[0017] A sliding window mechanism is used to perform real-time statistics on the relationship between factors and sedimentation over the past 72 hours, calculate the contribution variance and stability index of each factor, and automatically increase the weight coefficient of a factor when the contribution variance of a factor exceeds a set threshold, so as to ensure that key factors can receive priority attention at critical moments.

[0018] By calculating the marginal contribution of each factor to reducing the settlement prediction error, the importance ranking of factors is updated in real time, and the weights of the top three factors are increased by 20-30%, while the weights of the bottom three factors are reduced accordingly, thereby realizing the dynamic reallocation of weight resources.

[0019] Taking into account the three objectives of prediction accuracy, computational efficiency and model stability, the Pareto optimal solution set is used to select the best weight configuration scheme to ensure that the overall performance is not affected while increasing the weight of key factors.

[0020] The purpose of this application is also to provide a method for using a real-time building ground settlement monitoring device, including the following steps.

[0021] The sampling frequency configuration of each sensor is determined based on the initial state of the settlement monitoring device, the key locations of the building, and the preset monitoring accuracy requirements. The multi-timescale data acquisition scheme of the hierarchical sensor network is then generated by combining the key locations of the building and the sampling frequency configuration.

[0022] The hierarchical sensor network is controlled to collect building structure deformation data and environmental parameter data in real time according to the multi-timescale data acquisition scheme.

[0023] Whenever the hierarchical sensor network completes a round of data acquisition, the currently acquired structural deformation data is marked as target monitoring data. Based on the target monitoring data and a preset adaptive threshold, the anomaly detection parameters of the data processing and analysis module are configured to perform real-time anomaly identification on the target monitoring data and identify instantaneous abnormal deformation of the building structure.

[0024] In the real-time anomaly identification process, the continuous features of the instantaneous abnormal deformation are statistically analyzed in real time through duration analysis. The abnormal deformation is classified and labeled by combining the instantaneous abnormal deformation and the continuous features, and the corresponding feature fingerprint data is extracted at the same time.

[0025] If the duration of the instantaneous abnormal deformation does not reach the preset classification standard, the real-time abnormal identification is maintained until the abnormal classification is completed.

[0026] If the instantaneous abnormal deformation has been classified and labeled, the abnormal identification process of the target monitoring data ends, and the classification result is transmitted to the deep learning prediction module.

[0027] By fusing feature data from different time windows and introducing an attention mechanism, we can achieve quantitative prediction of the probabilistic state of building abnormal deformation evolving into continuous settlement.

[0028] The contribution weights of each influencing factor to the abnormal evolution are calculated based on multivariate time series analysis, and the initial analysis parameters of the multivariate analysis module are obtained.

[0029] Spatial correlation analysis is performed based on the initial analysis parameters, the environmental parameter data, and the quantitative prediction results of the probability state to identify the multi-factor coupling relationship between environmental factors, structural characteristics, and abnormal evolution processes.

[0030] The anomaly classification results and probability state quantitative prediction results are mapped to a two-dimensional risk matrix, and the evaluation benchmark parameters of the risk assessment and early warning module are obtained.

[0031] Based on the aforementioned assessment benchmark parameters and the aforementioned two-dimensional risk matrix, the system automatically assesses and classifies the risks of building structure evolution, and generates comprehensive decision support information including settlement trend curves, risk analysis reports, and emergency response suggestions.

[0032] During the generation of the comprehensive decision support information, early warning notifications are pushed out in real time, and the risk analysis report and emergency response suggestions are combined to guide the on-site emergency response.

[0033] If the risk level of the building structure evolution does not reach the warning standard, normal monitoring will be maintained and data collection will continue.

[0034] If the risk level of the building structure evolution has reached the warning standard, the emergency plan will be activated immediately, and the real-time ground settlement monitoring device will be controlled to enter the high-frequency monitoring mode.

[0035] Furthermore, by analyzing the duration of the transient abnormal deformation in real time, the continuous characteristics of the transient abnormal deformation are statistically analyzed, and the abnormal deformation is classified and labeled by combining the transient abnormal deformation and the continuous characteristics, including the following steps.

[0036] A time window is established for each instantaneous abnormal deformation event, and its start time and duration are tracked in real time to quantify the continuous characteristics of abnormal deformation.

[0037] Calculate the key persistence parameters of abnormal deformation events, including duration, cumulative deformation amplitude, and trend changes.

[0038] The intensity of the anomaly is assessed and classification is provided based on the degree of deviation of key persistence parameters of the abnormal deformation from the normal range.

[0039] By combining the results of continuous characteristics and strength assessment, abnormal deformations are classified and it is determined whether they have reached a preset threshold.

[0040] The classification results are labeled in a structured manner, anomaly types and feature parameters are recorded, and synchronized to the deep learning prediction module and multi-factor analysis module to update the anomaly deformation feature library.

[0041] Furthermore, the steps for quantitative prediction of probabilistic states are as follows.

[0042] The Monte Carlo method is used to randomly sample the parameters of the prediction model, generating multiple combinations of potential parameters to construct different prediction trajectories and quantify prediction uncertainty.

[0043] The parameters obtained by random sampling are input into the prediction model to generate multiple settlement prediction trajectories covering different future states.

[0044] Statistical analysis is performed on the predicted trajectory to calculate the probability of occurrence of different settlement levels, form a probability distribution, and extract key indicators, including mean prediction, confidence interval, and probability of extreme events.

[0045] The prediction process comprehensively considers model uncertainty, parameter uncertainty, and observation uncertainty, and updates the posterior distribution of model parameters through Bayesian inference.

[0046] The prediction results in the form of probability distribution are used as output to provide quantitative reference for decision-making, including average state, confidence interval range and probability of extreme settlement events.

[0047] Furthermore, spatial correlation analysis is performed based on the initial analysis parameters, the environmental parameter data, and the quantitative prediction results of the probability state to identify the multi-factor coupling relationship between environmental factors, structural characteristics, and abnormal evolution processes, including the following steps.

[0048] A spatial weight matrix is ​​constructed based on the spatial location and structural connectivity of various parts of the building to describe the spatial relationship between different monitoring points.

[0049] The initial parameters, environmental parameters, and probabilistic state prediction results are uniformly organized and coded to enable unified processing of multiple factors in spatial correlation analysis.

[0050] Spatial autocorrelation indicators are used to quantify the aggregation characteristics of settlement in different parts of a building, identify abnormal settlement hotspots and cold spots, and provide a reference for causal relationship analysis.

[0051] Based on the constructed spatial weight matrix, a regression model considering spatial effects is established to analyze the influence of environmental factors and structural characteristics on the evolution of abnormal settlement, while eliminating the interference of spatial dependence on parameter estimation.

[0052] Through main effect analysis and interaction effect analysis, we quantitatively assess the direct and synergistic effects of various environmental factors and structural characteristics on the anomalous evolution process, revealing the multi-factor coupling mechanism.

[0053] The results of spatial correlation analysis and coupling relationship identification are output in a structured manner to provide quantitative basis for the multi-factor analysis module and risk assessment module, and to update the analysis parameters of the monitoring device.

[0054] Compared with the prior art, this application has the following beneficial effects: This application combines hierarchical sensor network, multi-timescale data acquisition, data processing and analysis, deep learning prediction, multi-factor coupling analysis and risk assessment and early warning to realize the abnormal identification, probabilistic evolution prediction and automated risk classification and early warning of building structure settlement, thereby providing quantitative, intelligent and real-time decision support for structural safety management. Attached Figure Description

[0055] Figure 1 This is a module communication timing diagram for the real-time ground settlement monitoring device of this application. In the diagram, a hierarchical sensor network collects building structural deformation data and environmental parameters, and sends the data to the data processing and analysis module for real-time processing and anomaly identification. The processed data is simultaneously transmitted to the deep learning prediction module and the multi-factor analysis module. The deep learning prediction module predicts the probabilistic state of abnormal deformation evolving into settlement, while the multi-factor analysis module performs a coupled analysis of environmental factors, structural characteristics, and the abnormal evolution process. Finally, the risk assessment and early warning module summarizes the prediction and analysis results, performs risk classification assessment, and generates comprehensive decision support information such as settlement trends, early warning information, and emergency response suggestions.

[0056] Figure 2 This is a flowchart illustrating the usage method of the real-time ground settlement monitoring device for buildings as described in this application. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be described in more detail below with reference to the accompanying drawings. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are some embodiments of this invention, but not all embodiments.

[0058] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] The embodiments and directional terms described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0060] like Figure 1 As shown, a real-time ground settlement monitoring device for buildings includes the following modules.

[0061] A hierarchical sensor network is used to conduct real-time data acquisition at multiple time scales, comprehensively acquiring deformation data of building structures and related environmental parameters.

[0062] The data processing and analysis module is used to process and intelligently analyze the collected data in real time, and to identify and classify abnormal deformation characteristics of building structures in a timely manner.

[0063] The deep learning prediction module quantitatively predicts the probabilistic state and evolution trend of building settlement by constructing a multi-scale time series prediction model.

[0064] The multi-factor analysis module is used to perform coupled modeling and comprehensive analysis of environmental factors, structural characteristics and abnormal evolution processes in order to identify key influencing factors and their mechanisms of action.

[0065] Risk assessment and early warning module: Based on the prediction and analysis results, the risk of building settlement is classified and assessed, and corresponding early warning information and decision support solutions are generated.

[0066] This real-time building ground settlement monitoring device achieves dynamic perception, intelligent analysis, trend prediction, and risk warning of the entire building settlement process through the coordinated operation of multiple modules in its overall design. Firstly, through a hierarchical sensor network, the device can monitor the building ground and related environment in real time at different time scales. This multi-level, multi-dimensional data acquisition method not only improves the completeness and accuracy of the data but also ensures the capture of minute deformation signals in the early stages of building settlement, thus avoiding the lag and omissions caused by excessively large sampling intervals or insufficient data collection dimensions in traditional monitoring methods. Secondly, in the data processing and analysis stage, the device's real-time data processing module can denoise, correct, and intelligently classify the multi-source data collected by the sensors, enabling the complex settlement process to be broken down into different feature types and identified. For example, when uneven settlement, localized abnormal subsidence, or deformation related to changes in environmental loads occur in the building foundation, clear analysis results can be quickly generated, reducing the ambiguity and errors of manual interpretation.

[0067] At the prediction level, the device incorporates deep learning methods, treating the settlement process as a typical time series problem. It simulates the dynamic evolution of settlement by constructing a multi-scale time series prediction model. Compared to traditional prediction methods based on empirical formulas or single regression models, deep learning models can simultaneously capture short-term fluctuations and long-term trends, and can automatically learn potential nonlinear relationships. This means it can not only judge the current state but also predict the settlement trend and its uncertainty range over a future period, thus providing a forward-looking reference for building safety management. This proactive and quantitative feature of prediction significantly enhances the initiative and scientific nature of risk management.

[0068] Furthermore, the introduction of a multi-factor analysis module further enhances the comprehensiveness and explanatory power of the device. During building settlement, influencing factors are often multifaceted, including changes in geological conditions, fluctuations in groundwater levels, differences in construction load distribution, and the stiffness and flexibility of the building structure itself. By coupling these environmental factors, structural characteristics, and settlement anomaly processes into a model, we can go beyond simply identifying phenomena and delve deeper into key influencing factors and their mechanisms of action. Such analysis helps determine whether settlement problems are caused by external environmental disturbances or by design and construction defects in the building itself, thus enabling targeted solutions in subsequent treatment and remediation, improving the pertinence and effectiveness of management and maintenance.

[0069] In terms of risk management and early warning, this device establishes a tiered risk assessment and early warning mechanism by combining prediction results with multi-factor analysis results. Unlike traditional monitoring systems that only issue alarms after reaching a fixed threshold, this device can dynamically calculate the risk level and generate tiered early warning information based on the comprehensive weight of settlement trends, abnormal evolution speed, and influencing factors. For low-risk situations, it can prompt daily monitoring and inspections; for medium-risk situations, it can suggest further testing and local reinforcement; and for high-risk situations, it can quickly trigger emergency warnings and decision support plans, reminding relevant management personnel to take immediate preventive measures. This tiered early warning system not only avoids false alarms and missed alarms caused by overly simplistic threshold settings, but also enables differentiated response strategies at different risk stages, thereby achieving rational resource allocation and efficient emergency response.

[0070] In summary, the advantages of this real-time ground settlement monitoring device are mainly reflected in the following aspects: First, it can accurately capture and continuously track settlement in its early stages; second, it can clearly analyze complex settlement processes; third, it can provide reasonable trend predictions before problems appear; fourth, it can reveal the essential mechanism of settlement through multi-factor modeling; and fifth, it can ensure that risks at different levels can be responded to and managed accordingly.

[0071] In summary, the real-time building ground settlement monitoring device proposed in this embodiment achieves a complete chain from information collection to risk management through the collaborative work of multiple modules, including sensor networks, data analysis, deep learning prediction, multi-factor coupled modeling, and hierarchical risk early warning. Its innovation lies not only in the significant improvement in monitoring capabilities but also in the scientific rigor and practicality of the prediction and early warning mechanisms, thereby significantly enhancing building safety assurance capabilities in practical engineering applications.

[0072] Furthermore, the hierarchical sensor network includes a base layer sensor array, structural layer sensor nodes, and an environmental monitoring sensor group. The base layer sensor array combines high-precision laser displacement sensors with digital tilt sensors, and is installed in key parts of the building foundation according to a grid layout principle. The sensor spacing is set within the range of 5 to 15 meters based on the building span to ensure the capture of minute deformations at the 0.1 mm level. The structural layer sensor nodes use wireless strain sensors and triaxial accelerometers, deployed along the main load-bearing structure of the building to achieve real-time monitoring of structural stress state changes and dynamic responses. The environmental monitoring sensor group includes soil moisture sensors, groundwater level monitoring sensors, temperature and humidity sensors, and ground vibration sensors, used to acquire external environmental parameters affecting building settlement.

[0073] As can be seen from the above, multi-dimensional collaborative monitoring of the foundation layer, structural layer, and environmental layer achieves high-precision and comprehensive perception of building settlement. The foundation layer sensor array can capture minute deformations with sub-millimeter accuracy, ensuring timely identification of early anomalies; the structural layer nodes acquire stress and vibration information in real time, reflecting the dynamic response of the building under load; and the environmental monitoring sensor group comprehensively collects external factors such as hydrology, climate, and ground disturbance, providing sufficient data support for settlement cause analysis. Through this layered deployment and data fusion approach, the building settlement status can be accurately reflected at multiple scales from micro to macro, significantly improving the reliability of monitoring and the scientific nature of early warning.

[0074] Furthermore, the data processing and analysis module includes a data preprocessing unit, a feature extraction unit, and an anomaly detection unit. Specifically: the data preprocessing unit performs noise filtering and signal enhancement on the raw sensor data, while correcting sensor drift and suppressing environmental interference; the feature extraction unit extracts multi-frequency domain features based on wavelet transform and combines principal component analysis to reduce data dimensionality, thereby extracting key feature parameters that characterize the building's deformation state; the anomaly detection unit establishes dynamic control limits using statistical process control methods, identifies abnormal deformation patterns by comparing with historical benchmark data, triggers an anomaly warning when three consecutive sampling points exceed the control limits, and automatically adjusts the detection sensitivity according to different settlement development stages to meet monitoring requirements.

[0075] The data preprocessing unit uses a Kalman filter as the core noise removal algorithm, by setting the process noise covariance Q = 1 × 10⁻⁶. -5 Measurement noise covariance R = 1 × 10 -2A recursive optimal estimation method is used to achieve a complete filtering process, including state prediction, error covariance calculation, gain coefficient update, and state correction, resulting in a signal-to-noise ratio improvement of over 25dB and a processing latency of less than 10 milliseconds. Simultaneously, the filtering parameters are dynamically adjusted based on real-time data statistical characteristics, updated every 100 sampling points. Signal enhancement processing employs a fourth-order Butterworth bandpass filter with a passband range of 0.01-50Hz to cover settlement characteristics across different frequency bands, including long-term trend settlement (0.01-1Hz), diurnal temperature effect (1-10Hz), and structural vibration response (10-50Hz). An adaptive gain control algorithm normalizes signals from different sensors to a uniform power level, achieving a signal quality improvement rate of over 95%. Sensor drift correction employs a recursive least squares algorithm, setting a 24-hour sliding window to linearly fit the data within the window to y(t)=a×t+b. The average drift rate across multiple windows is used for correction, establishing a three-level drift warning mechanism: a drift rate >0.05mm / day triggers an alert, >0.10mm / day triggers a warning, and >0.20mm / day triggers an alarm and recommends sensor inspection. After correction, the drift error can be controlled within ±0.02mm. For multi-sensor monitoring networks, a collaborative correction mechanism is implemented, utilizing redundant information from adjacent sensors to cross-validate drift estimates. Sensors with abnormal drift are automatically marked and their data quality is downgraded, and sensor health reports are generated periodically. Environmental interference suppression employs an adaptive notch filter. For interference sources such as power frequency interference (50Hz / 60Hz), higher harmonics (100Hz, 150Hz, 120Hz, 180Hz), and mechanical vibration, a notch filter with a quality factor of 30 and a notch depth ≥40dB is set, using zero-phase filtering. Through spectrum analysis, the main interference frequency components are automatically identified, and the corresponding notch filters are dynamically configured, reducing environmental interference power to 1 / 100 of its original value, achieving a suppression effect of over -40dB. After a complete data preprocessing workflow, the quality of the raw sensor data is significantly improved: signal-to-noise ratio increased by over 25dB, drift error within ±0.02mm, environmental interference suppression by over -40dB, processing latency <10ms, and data integrity rate over 99.8%, providing a high-quality data foundation for subsequent feature extraction and anomaly detection.

[0076] The feature extraction unit uses Daubechies db8 wavelet as the transform basis function to perform 6-level wavelet decomposition, decomposing the preprocessed settlement signal into 7 frequency bands: cD1 (25-50Hz) corresponds to high-frequency noise and vibration, cD2-cD3 (6.25-25Hz) corresponds to structural vibration response and short-term disturbances, cD4-cD6 (0.78-6.25Hz) corresponds to intraday fluctuations, temperature effects, and groundwater level changes, and cA6 (0-0.78Hz) corresponds to long-term settlement trend. Four statistical features—mean, standard deviation, maximum amplitude, and energy—are extracted from the wavelet coefficients of each layer, forming 28 basic features. Further, time-frequency joint features (correlation coefficient between adjacent layers, skewness and kurtosis, energy distribution, and wavelet entropy, totaling 32 dimensions), nonlinear features (approximate entropy, sample entropy, and fractal dimension, totaling 20 dimensions), frequency domain features (dominant frequency components and band energy ratio, totaling 16 dimensions), and time-domain extended features (zero-crossing rate, peak factor, margin factor, and waveform factor, totaling 24 dimensions) are constructed, ultimately forming a complete 128-dimensional feature vector that comprehensively characterizes the time-frequency characteristics of the settlement signal. Principal component analysis (PCA) reduces the 128-dimensional high-dimensional features to 16-dimensional principal components through eigenvalue decomposition of the covariance matrix, achieving an information retention rate of 98.5% and a dimensionality reduction ratio of 8:1. The 16 principal components have clear physical meanings: PC1-PC3 mainly come from low-frequency wavelet coefficients reflecting long-term subsidence trends (approximately 40% contribution); PC4-PC7 mainly come from mid-frequency wavelet coefficients reflecting seasonal and diurnal periodic variations (approximately 30% contribution); PC8-PC12 mainly come from higher-frequency wavelet coefficients reflecting short-term fluctuations and structural responses (approximately 20% contribution); and PC13-PC16 mainly come from higher-order statistical features reflecting the randomness and complexity of the data (approximately 8.5% contribution). Simultaneously, an incremental PCA algorithm is implemented to support online updates of the principal component model. An update is triggered every 1000 new samples, using incremental eigenvalue decomposition to update the covariance matrix, ensuring that the PCA model always reflects the statistical characteristics of the current data. In addition, key parameters such as displacement amplitude features (peak-to-peak value, effective value, absolute average value, peak factor), frequency component features (identifying the dominant frequency, calculating the frequency band energy distribution and frequency centroid through FFT analysis), phase difference features (extracting instantaneous phase through Hilbert transform and calculating the phase synchronization index to analyze multi-point coordination), and energy density features (time-domain energy density, frequency-domain energy density, wavelet energy density, and normalized energy entropy) are extracted. The importance of features is evaluated through variance analysis, correlation analysis, and information gain to achieve adaptive feature selection. Redundant features with importance below the threshold are eliminated, and the top 16 core features in terms of importance are retained. Z-score standardization is used to ensure that different features are comparable.The performance metrics of the feature extraction unit are as follows: the feature dimension is reduced from 120-dimensional sensor data to 16-dimensional principal components after extracting 128-dimensional wavelet features; the information retention rate is 98.5%; the processing latency is <5ms; the feature discrimination (inter-class distance / intra-class distance ratio under different settlement states) is >5; and the feature stability (relative standard deviation RSD of repeatability test) is <2%. It provides high-quality, low-dimensional, and highly representative feature data for anomaly detection and predictive analysis.

[0077] The anomaly detection unit uses XR control charts (mean-range control charts) for anomaly detection, establishing control limits based on the 3σ criterion: upper control limit UCL = μ + 3σ, lower control limit LCL = μ - 3σ. Since the probability of data falling within the range of μ ± 3σ under a normal distribution is 99.73%, the false alarm rate for normal data exceeding the limits is only 0.27%, demonstrating high reliability. A dynamic baseline update mechanism is implemented: an initial baseline is established using the first 1000 sampling points, calculating the initial mean μ0 and standard deviation σ0. The baseline is updated every 100 new data points using the Exponential Weighted Moving Average (EWMA) algorithm, with the update formula being μ0 = μ0 + 3σ. new =α·μ old +(1-α)·x newThe weighting coefficient α=0.95 maintains the memory of historical data; at the same time, the seasonal cycle of the data is identified, and a seasonal baseline μ (season) is established. The deseasonalized data is used for control chart analysis to adapt to the non-stationary characteristics of the settlement process. At the same time, six types of control chart anomaly criteria are implemented: (1) a single point exceeds the 3σ control limit, (2) nine consecutive points fall on the same side of the center line, (3) six consecutive points increase or decrease monotonically, (4) fourteen consecutive points fluctuate alternately, (5) two out of three consecutive points fall in the 2σ-3σ region, and (6) fifteen consecutive points fall within the μ±σ range (overly stable). Meeting any one of these criteria is considered an anomaly. Anomaly pattern recognition uses Mahalanobis distance to measure the degree of anomaly. A baseline mean vector and covariance matrix are established based on a 36-month historical database. The Mahalanobis distance between the current data and the baseline is calculated. Based on the Mahalanobis distance threshold, 12 anomaly patterns are classified: normal fluctuation (MD < 2), trend drift (2 ≤ MD < 4), periodic anomaly (4 ≤ MD < 6), sudden jump (MD ≥ 6), slow decay, oscillation enhancement, step change, noise surge, seasonal deviation, nonlinear drift, intermittent anomaly, and compound anomaly. An anomaly pattern database is established to store typical anomaly samples for rapid matching. The recognition time is < 5 seconds, and the accuracy rate reaches 96.8%. The early warning triggering mechanism adopts a continuous over-limit triggering rule: it detects the three most recent sampling points, and triggers an early warning when three consecutive points exceed the control limit; it automatically adjusts the detection sensitivity according to the settlement development stage, and determines the stage by calculating the data variance: in the initial stage (data variance > 1.0), it uses high sensitivity to adjust the control limit to 2.5σ; in the stable stage (variance 0.5-1.0), it uses medium sensitivity to maintain 3.0σ; and in the convergence stage (variance < 0.5), it uses low sensitivity to relax to 3.5σ, ensuring that anomalies can be detected in a timely manner while avoiding excessive alarms. A tiered early warning mechanism is implemented: Level I (low risk) normal monitoring, Level II (lower risk) intensive observation, Level III (medium risk) key monitoring, Level IV (higher risk) early warning response, and Level V (high risk) immediate handling. Upon triggering an early warning, comprehensive decision support information is automatically generated, including anomaly location and time, anomaly type and severity, possible cause analysis (based on similarity matching from a historical case database), recommended handling measures (divided into four levels: immediate response 0-24 hours, short-term handling 1-7 days, medium-term improvement 1-3 months, and long-term prevention 3-12 months), and impact assessment. Decision support information is pushed to the management terminal in real time, with a response time of <1 second, supporting multi-terminal synchronization and confirmation feedback mechanisms. The anomaly detection unit's performance indicators are: detection accuracy 96.8%, false alarm rate <0.5%, false negative rate <1.0%, response time <1 second, and real-time dynamic sensitivity adjustment cycle, providing timely and reliable anomaly early warning and decision support for building safety monitoring.

[0078] The data processing and analysis module employs a pipelined parallel processing architecture consisting of a preprocessing unit (10ms), a feature extraction unit (5ms), and an anomaly detection unit (10ms). This architecture enables end-to-end processing from raw sensor data to anomaly detection results, with total processing latency controlled within 25ms and a data throughput of 1.2M records / day, meeting real-time monitoring requirements. A comprehensive quality control and monitoring mechanism is established to monitor the operational status of each unit in real time, recording key performance indicators such as processing success rate, latency distribution, and error types. A data quality scoring mechanism comprehensively evaluates signal-to-noise ratio, completeness, consistency, and timeliness, automatically marking and isolating low-quality data, and periodically generating module performance reports to provide a basis for optimization. The module implements an adaptive optimization mechanism, automatically optimizing algorithm parameters, including the Q / R parameters of the Kalman filter, the number of wavelet decomposition layers, the number of PCA principal components, and control chart sensitivity, based on changes in the statistical characteristics of long-term monitoring data. A reinforcement learning strategy is used to optimize detection accuracy and false alarm rate, achieving intelligent adaptive adjustment of parameters. Simultaneously, a parameter evolution history record is established to track parameter change trends, providing data support for algorithm improvement and upgrades. Seamless integration between processing units is achieved through standardized data interfaces: the preprocessing unit outputs clean time-series data, the feature extraction unit outputs dimensionality-reduced feature vectors, and the anomaly detection unit outputs anomaly determination results and risk levels. All data includes complete timestamps, sensor IDs, data quality tags, and other metadata for easy traceability and auditing. A fault-tolerant handling mechanism is implemented to automatically activate backup algorithms or degradation processing strategies when a unit fails, ensuring continuous availability. A data caching mechanism is also established to cache data to be processed during network outages or maintenance, and automatically compensate for data loss upon recovery, ensuring data continuity and integrity. The module's overall performance indicators are: total processing latency of 25ms, data processing success rate of 99.2%, feature extraction accuracy of 98.5%, anomaly detection accuracy of 96.8%, availability of 99.9%, and data throughput of 1.2M records / day. Through the collaborative work of the three-level architecture of preprocessing, feature extraction, and anomaly detection, it achieves high-quality, high-efficiency, and high-reliability settlement monitoring data processing, providing a solid data foundation and technical support for the deep learning prediction module, multi-factor analysis module, and risk assessment and early warning module, thus forming the core data processing engine of the real-time building ground settlement monitoring device.

[0079] As can be seen from the above, the organic combination of preprocessing, feature extraction, and anomaly detection achieves efficient processing and accurate identification of settlement data. The data preprocessing unit effectively filters out noise and environmental interference, ensuring data reliability; the feature extraction unit uses multi-frequency domain analysis and dimensionality reduction methods to extract key parameters, improving the ability to express structural deformation characteristics; and the anomaly detection unit identifies abnormal settlement in a timely manner by dynamically controlling limits and comparing with historical data, and has an adaptive sensitivity adjustment function. Overall, this data processing and analysis module significantly improves monitoring accuracy and the real-time performance and reliability of anomaly early warning.

[0080] Furthermore, the deep learning prediction module adopts a hybrid architecture combining a long short-term memory neural network and an attention mechanism. It captures the temporal dependencies of settlement data by constructing a multi-layer LSTM network, while the attention mechanism is used to adaptively allocate weights for data at different time steps, thereby improving the prediction accuracy for key time nodes. The deep learning prediction module includes a data normalization layer, a feature encoding layer, a temporal modeling layer, and an output decoding layer. Specifically, the data normalization layer standardizes the input multi-dimensional temporal data, the feature encoding layer converts sensor data into high-dimensional feature vectors, the temporal modeling layer learns the dynamic pattern of settlement evolution through a bidirectional long short-term memory neural network, and the output decoding layer generates the predicted settlement probability distribution within the future time window.

[0081] As can be seen from the above, this deep learning prediction module effectively improves the modeling and prediction capabilities of building settlement time-series characteristics by combining a long short-term memory network with an attention mechanism. Its hierarchical structure first normalizes and encodes high-dimensional features from multidimensional sensor data, then uses a bidirectional LSTM to capture long-term and short-term dynamic patterns, and highlights the contribution of key time nodes to the prediction results through an attention mechanism. Finally, the decoding layer outputs the probability distribution of future settlement. This process not only improves the accuracy and stability of predictions but also provides a reliable basis for early identification of settlement risks and trend assessment.

[0082] Furthermore, the multi-factor analysis module includes an environmental factor impact assessment unit, a structural characteristic analysis unit, and a settlement mechanism identification unit. Specifically: the environmental factor impact assessment unit is used to quantitatively analyze the effects of various environmental factors on settlement and establish their response functions and sensitivity coefficients; the structural characteristic analysis unit assesses the mechanical state and safety reserve level of the building structure through structural mechanics modeling and measured data correction; and the settlement mechanism identification unit analyzes the spatiotemporal evolution characteristics of settlement through data mining and pattern recognition, identifies dominant factors and triggering conditions, extracts typical settlement patterns, and establishes discrimination criteria.

[0083] As can be seen from the above, this multi-factor analysis module achieves an in-depth analysis of the causes and evolution of building settlement through comprehensive modeling of environmental, structural, and mechanistic aspects. The environmental factor assessment unit can quantify the degree of influence of external conditions on settlement and reveal sensitive factors and their response relationships; the structural characteristic analysis unit combines theoretical modeling with experimental correction to accurately reflect the mechanical state and safety margin of the building; and the settlement mechanism identification unit extracts typical settlement patterns and establishes discrimination criteria through data mining and pattern recognition, thereby clarifying the dominant factors and triggering conditions.

[0084] Furthermore, the multi-factor analysis module also includes a dynamic weight allocation mechanism implementation unit, which is used to adaptively adjust the weight coefficients of each influencing factor according to different stages of settlement evolution. Its specific implementation includes the following steps.

[0085] By calculating the covariance matrix between each influencing factor and the settlement response in real time, the instantaneous influence intensity of the factors is dynamically assessed.

[0086] A sliding window mechanism is used to perform real-time statistics on the relationship between factors and sedimentation over the past 72 hours, calculate the contribution variance and stability index of each factor, and automatically increase the weight coefficient of a factor when the contribution variance of a factor exceeds a set threshold, so as to ensure that key factors can receive priority attention at critical moments.

[0087] By calculating the marginal contribution of each factor to reducing the settlement prediction error, the importance ranking of factors is updated in real time, and the weights of the top three factors are increased by 20-30%, while the weights of the bottom three factors are reduced accordingly, thereby realizing the dynamic reallocation of weight resources.

[0088] Taking into account the three objectives of prediction accuracy, computational efficiency and model stability, the Pareto optimal solution set is used to select the best weight configuration scheme to ensure that the overall performance is not affected while increasing the weight of key factors.

[0089] As can be seen from the above, by introducing a dynamic weight allocation mechanism, adaptive adjustments to influencing factors at different stages are achieved, improving the accuracy and stability of settlement prediction. Its core lies in dynamically evaluating the contribution of factors based on real-time covariance analysis and sliding window statistics, prioritizing the increase of important factors' weights at critical stages while reducing the influence of irrelevant factors. Simultaneously, the marginal contribution of prediction errors is combined to optimize factor ranking, and the optimal weight configuration is selected through a Pareto optimal strategy. This ensures that while enhancing the role of key factors, overall computational efficiency and prediction performance are maintained, thereby achieving accurate modeling and risk warning of the settlement evolution process.

[0090] Furthermore, the settlement mechanism identification unit also includes a multi-scale spatiotemporal coupling modeling subunit, which is used to deeply analyze the coupling mechanism of influencing factors at different spatiotemporal scales, and its implementation includes the following methods.

[0091] Microscale coupling modeling: For a single sensor monitoring point, a local factor coupling model is established. Partial least squares regression analysis is used to analyze the direct coupling relationship of environmental factors such as groundwater level, soil moisture content, and temperature. The coupling strength is quantified by a weighted combination of Pearson correlation coefficient and mutual information. When the coupling strength exceeds 0.7, it is marked as a strong coupling relationship.

[0092] Mesoscale coupling modeling: A regional factor coupling network is established based on structural partitioning. The spatial transmission effect between adjacent monitoring points is captured by graph convolutional neural network. At the same time, a time delay function is introduced to model the spatial propagation characteristics of factor effects, so as to realize dynamic modeling of cross-regional factor effects.

[0093] Macro-scale coupling modeling: Construct a factor coupling matrix at the whole building system level, use principal component analysis to extract the system-level dominant factor combination, identify the dominant coupling mode between factors through eigenvalue decomposition, and establish the time evolution equation of factor coupling strength.

[0094] Cluster analysis is used to classify historical coupling patterns into four categories: stable, fluctuating, mutational, and hybrid. Feature templates are established for each category. The matching degree between the current coupling state and historical patterns is monitored in real time. When the matching degree is lower than 0.8, a new pattern learning program is triggered to automatically update the coupling pattern library.

[0095] As can be seen from the above, multi-scale spatiotemporal coupling modeling enables a deep analysis of the interrelationships of factors in the settlement mechanism. At the microscopic level, it identifies the strong coupling characteristics between individual factors; at the mesoscopic level, it captures the spatial transmission and temporal delay effects of regional factors; and at the macroscopic level, it extracts the system-level dominant factor combinations and constructs evolutionary equations, comprehensively presenting the synergistic effects of factors at different scales. Simultaneously, by combining historical pattern clustering with real-time matching, the coupled pattern library is dynamically updated to ensure the model's adaptability to environmental and structural changes, thereby improving the accuracy of settlement prediction and risk identification capabilities.

[0096] like Figure 2 As shown, a method for using a real-time building ground settlement monitoring device includes the following steps.

[0097] The sampling frequency configuration of each sensor is determined based on the initial state of the settlement monitoring device, the key locations of the building, and the preset monitoring accuracy requirements. The multi-timescale data acquisition scheme of the hierarchical sensor network is then generated by combining the key locations of the building and the sampling frequency configuration.

[0098] The hierarchical sensor network is controlled to collect building structure deformation data and environmental parameter data in real time according to the multi-timescale data acquisition scheme.

[0099] Whenever the hierarchical sensor network completes a round of data acquisition, the currently acquired structural deformation data is marked as target monitoring data. Based on the target monitoring data and a preset adaptive threshold, the anomaly detection parameters of the data processing and analysis module are configured to perform real-time anomaly identification on the target monitoring data and identify instantaneous abnormal deformation of the building structure.

[0100] In the real-time anomaly identification process, the continuous features of the instantaneous abnormal deformation are statistically analyzed in real time through duration analysis. The abnormal deformation is classified and labeled by combining the instantaneous abnormal deformation and the continuous features, and the corresponding feature fingerprint data is extracted at the same time.

[0101] If the duration of the instantaneous abnormal deformation does not reach the preset classification standard, the real-time abnormal identification is maintained until the abnormal classification is completed.

[0102] If the instantaneous abnormal deformation has been classified and labeled, the abnormal identification process of the target monitoring data ends, and the classification result is transmitted to the deep learning prediction module.

[0103] By fusing feature data from different time windows and introducing an attention mechanism, we can achieve quantitative prediction of the probabilistic state of building abnormal deformation evolving into continuous settlement.

[0104] The contribution weights of each influencing factor to the abnormal evolution are calculated based on multivariate time series analysis, and the initial analysis parameters of the multivariate analysis module are obtained.

[0105] Spatial correlation analysis is performed based on the initial analysis parameters, the environmental parameter data, and the quantitative prediction results of the probability state to identify the multi-factor coupling relationship between environmental factors, structural characteristics, and abnormal evolution processes.

[0106] The anomaly classification results and probability state quantitative prediction results are mapped to a two-dimensional risk matrix, and the evaluation benchmark parameters of the risk assessment and early warning module are obtained.

[0107] Based on the aforementioned assessment benchmark parameters and the aforementioned two-dimensional risk matrix, the system automatically assesses and classifies the risks of building structure evolution, and generates comprehensive decision support information including settlement trend curves, risk analysis reports, and emergency response suggestions.

[0108] During the generation of the comprehensive decision support information, early warning notifications are pushed out in real time, and the risk analysis report and emergency response suggestions are combined to guide the on-site emergency response.

[0109] If the risk level of the building structure evolution does not reach the warning standard, normal monitoring will be maintained and data collection will continue.

[0110] If the risk level of the building structure evolution has reached the warning standard, the emergency plan will be activated immediately, and the real-time ground settlement monitoring device will be controlled to enter the high-frequency monitoring mode.

[0111] The real-time ground settlement monitoring device for buildings utilizes a systematic and hierarchical process to achieve closed-loop management from data acquisition, anomaly identification, trend prediction to risk assessment and emergency response. First, in the data acquisition phase, the sampling frequency of sensors is rationally configured based on the building's key locations and preset accuracy requirements, forming a multi-timescale data acquisition scheme. This approach not only ensures that monitoring needs at different locations and stages are met but also effectively balances monitoring accuracy and data processing efficiency, avoiding monitoring blind spots or data redundancy caused by a single sampling frequency. Second, the hierarchical sensor network acquires structural deformation and environmental parameters in real time according to the acquisition scheme, providing rich and multi-dimensional data support for subsequent analysis. In this way, the building's settlement status can be captured at the millimeter level or even smaller, while also incorporating external environmental factors into the monitoring scope, making the data more comprehensive and relevant.

[0112] In the data processing and anomaly identification stages, adaptive threshold configuration and real-time anomaly detection mechanisms are fully utilized to analyze and identify the collected target data in real time, enabling rapid response when instantaneous abnormal deformations occur in the building structure. Compared with traditional offline analysis or periodic sampling, this real-time processing method significantly shortens the time from anomaly occurrence to identification, effectively avoiding potential risks caused by delayed response. Furthermore, it not only identifies individual anomalies but also statistically analyzes the evolutionary characteristics of anomalies based on duration analysis, and distinguishes different types of anomalies through classification labels and feature fingerprint extraction. This refined anomaly classification not only improves the accuracy of anomaly identification but also provides more precise input conditions for subsequent prediction and factor analysis.

[0113] In the prediction phase, a hybrid architecture of deep learning was introduced, combining a long short-term memory network with an attention mechanism to fuse feature data from different time windows. This allows the model to capture both long-term subsidence trends and highlight the significant impact of key time points on subsidence development, thereby generating more accurate and dynamic probabilistic predictions. Compared to traditional linear prediction methods, this approach better addresses the nonlinear changes and uncertainties inherent in the subsidence process.

[0114] Meanwhile, by using multivariate time series analysis to calculate the contribution weights of each influencing factor to the abnormal evolution and combining it with the initial analysis parameters to identify spatial correlations, the coupling relationship between environmental factors, structural characteristics, and settlement evolution can be revealed. This analysis goes beyond simply judging the results; it can also delve into the mechanism of settlement formation, thereby helping managers accurately determine whether the settlement is caused by geological conditions, water level fluctuations, or structural design problems, and thus providing a solid basis for formulating targeted remediation measures.

[0115] In the risk assessment and early warning stages, an intuitive and scientific risk grading system was established by mapping anomaly classification results and probability prediction results to a two-dimensional risk matrix. The assessment results can automatically trigger tiered early warnings, generating comprehensive decision support information including settlement trend curves, risk analysis reports, and emergency response recommendations. This decision information is not merely a warning, but a comprehensive action plan that provides practical guidance for on-site management and emergency response. For example, when the risk level is low, it can be recommended to strengthen daily monitoring and localized detection; when the risk level increases, emergency response recommendations will be pushed out and the corresponding emergency plans will be activated on-site, ensuring that risks are controllable from the outset and that responses are timely and effective.

[0116] Furthermore, when the identified risk level does not meet the warning criteria, normal monitoring can be maintained, thus avoiding over-response and waste of resources. However, once the risk level reaches or exceeds the warning criteria, the emergency plan is immediately activated, and the system automatically switches to a high-frequency monitoring mode to continuously track settlement status with higher sampling density and real-time performance. This phased dynamic management strategy ensures both the economy and stability of the monitoring process, while also guaranteeing rapid response in critical moments, significantly improving the initiative and safety of building settlement management.

[0117] Overall, this integrated process of data collection, analysis, prediction, assessment, and response transforms building settlement monitoring from a single-point detection or passive response into a dynamic, intelligent, and comprehensive management system covering the entire lifecycle. This not only significantly improves the real-time performance and reliability of settlement monitoring but also effectively reduces safety risks and economic losses during building operation.

[0118] Furthermore, the implementation of the adaptive threshold configuration includes the following steps.

[0119] The settlement data and related environmental parameters for the most recent 30 days are divided into statistical analysis windows in chronological order.

[0120] Based on statistical analysis window data, key statistical indicators of building settlement and related environmental parameters are calculated to quantify the fluctuation characteristics of the building's current state.

[0121] Anomaly detection thresholds are set based on the calculated key statistical indicators to achieve hierarchical monitoring of abnormal deformations of different degrees.

[0122] Based on the building's current settlement status and seasonal variation characteristics, the initially set threshold range is dynamically adjusted to adapt to changes in the building under different operating environments and time conditions.

[0123] Perform statistical analysis on historical abnormal events to identify abnormal patterns and threshold adaptability issues, and optimize threshold configuration accordingly.

[0124] As can be seen from the above, this embodiment achieves hierarchical monitoring by dynamically extracting key indicators to set anomaly detection thresholds through statistical analysis of settlement and environmental data over the past 30 days. Based on the initial threshold setting, it dynamically adjusts the thresholds by combining the real-time status and seasonal variation characteristics of building settlement, making the thresholds more adaptable and flexible. Simultaneously, by retrospectively analyzing historical anomaly events, the threshold configuration is continuously optimized to avoid false alarms and missed alarms, thereby effectively improving the accuracy and robustness of anomaly monitoring and maintaining stable and reliable monitoring results under different environmental and time periods.

[0125] Furthermore, the duration analysis includes the following steps.

[0126] Continuous tracking of transient anomalies is performed, with a minimum duration threshold of 10 minutes and a maximum tracking time of 24 hours.

[0127] Calculate the degree of deviation of the abnormal deformation amplitude from the normal deformation range, assess the abnormal intensity, and determine the abnormality level.

[0128] By fitting abnormal deformation data and analyzing its development trend, it can be determined whether the deformation is increasing, decreasing or stable, in order to assist in the assessment of potential risks.

[0129] Analysis of variance is used to assess the degree of fluctuation in abnormal deformation, thereby determining the stability and persistence of abnormal signals.

[0130] When abnormal deformation lasts for more than 30 minutes and shows an increasing trend, it is automatically marked as a persistent anomaly, triggering a deep analysis program and increasing the sampling frequency of sensors in the relevant area to obtain more detailed deformation information.

[0131] As can be seen from the above, this embodiment achieves a comprehensive assessment of the intensity, development trend, and stability of abnormal deformations by continuously tracking and limiting the duration of instantaneous anomalies, combined with anomaly amplitude deviation, trend fitting, and fluctuation variance analysis. When an anomaly continuously exceeds a set threshold and shows an increasing trend, it can be automatically marked as a persistent anomaly and trigger in-depth analysis. At the same time, the sampling frequency of local sensors is increased to obtain more refined data, thereby effectively avoiding misjudgments caused by short-term disturbances and improving the accuracy of anomaly identification and the foresight of risk assessment.

[0132] Furthermore, the extraction of the feature fingerprint data includes the following steps.

[0133] Multidimensional feature fusion is performed on the abnormal deformation data of buildings to extract time-domain, frequency-domain, and spatial distribution features. The time-domain features include abnormal peak values, duration, and rate of change parameters. The frequency-domain features obtain the spectral characteristics of abnormal deformation through fast Fourier transform. The spatial distribution features analyze the propagation patterns of anomalies in different parts of the building.

[0134] The extracted multidimensional features are integrated into a 128-dimensional feature vector to achieve a structured representation of abnormal deformations and subsequent processing.

[0135] Principal component analysis was used to reduce the dimensionality of the 128-dimensional feature vector and extract 32-dimensional core features.

[0136] The dimensionality-reduced core feature vectors are stored in the abnormal deformation feature library, and historical abnormal events are classified and stored based on feature similarity.

[0137] By utilizing feature libraries and classification results, we can provide a reference for analyzing the causes of abnormal changes and predicting their development trends, thereby enabling precise monitoring of abnormal building conditions.

[0138] As can be seen from the above, this embodiment constructs a 128-dimensional feature vector by fusing time-domain, frequency-domain, and spatial distribution features, and then reduces the dimensionality to 32-dimensional core features through principal component analysis, achieving an efficient and structured representation of abnormal deformations. The core features are stored in a feature library and archived as historical events using similarity classification, thus forming a systematic management of abnormal information. This not only improves the ability to express and identify abnormal deformations but also provides reliable data support for causal analysis and trend prediction, significantly enhancing the accuracy and traceability of building anomaly monitoring.

[0139] Furthermore, by analyzing the duration of the transient abnormal deformation in real time, the continuous characteristics of the transient abnormal deformation are statistically analyzed, and the abnormal deformation is classified and labeled by combining the transient abnormal deformation and the continuous characteristics, including the following steps.

[0140] A time window is established for each instantaneous abnormal deformation event, and its start time and duration are tracked in real time to quantify the continuous characteristics of abnormal deformation.

[0141] Calculate the key persistence parameters of abnormal deformation events, including duration, cumulative deformation amplitude, and trend changes.

[0142] The intensity of the anomaly is assessed and classification is provided based on the degree of deviation of key persistence parameters of the abnormal deformation from the normal range.

[0143] By combining the results of continuous characteristics and strength assessment, abnormal deformations are classified and it is determined whether they have reached a preset threshold.

[0144] The classification results are labeled in a structured manner, anomaly types and feature parameters are recorded, and synchronized to the deep learning prediction module and multi-factor analysis module to update the anomaly deformation feature library.

[0145] As can be seen from the above, this embodiment, by establishing a time window for transient abnormal events and tracking them in real time, can accurately quantify the sustained characteristics of abnormal deformations. Simultaneously, it assesses the intensity of the abnormality by combining key parameters such as duration, cumulative amplitude, and trend changes. Based on this, the degree of deviation between the abnormal and normal states is used as the classification criterion to ensure the scientific validity and comparability of the classification results. After classification, the abnormality type and feature parameters are recorded in a structured manner and simultaneously updated to the prediction and multi-factor analysis modules, achieving dynamic improvement of the feature library.

[0146] Furthermore, the implementation of the attention mechanism includes the following steps.

[0147] Data from different time windows and different sensors are mapped to a unified feature space to achieve feature alignment.

[0148] Self-attention calculation is performed on a single sequence in the time dimension, and weights are assigned based on the correlation between different times in the sequence to highlight the impact of key time nodes.

[0149] Cross-attention calculation is performed on the features of different sensors in the spatial dimension to analyze the correlation between sensors and identify spatial correlation patterns.

[0150] Attention weights are generated through matrix operations on query vectors, key vectors, and value vectors, and the results are normalized.

[0151] The importance of each feature is adjusted based on the normalized attention weights, and a weighted aggregation method is used to generate the fused feature representation.

[0152] As can be seen from the above, this embodiment achieves dynamic weighting and fusion of key features by performing self-attention and cross-attention calculations on sensor data in both temporal and spatial dimensions. In the temporal dimension, it highlights the key time nodes in the sequence that have the greatest impact on settlement prediction; in the spatial dimension, it analyzes the correlation between sensors and identifies important spatial patterns. Normalized weights are generated through matrix operations on query, key, and value vectors, and the importance of each feature is adjusted accordingly. Finally, a fused feature representation is generated through weighted aggregation, thereby significantly enhancing the sensitivity and expressive power of the prediction model to key spatiotemporal information, and improving the accuracy and reliability of settlement trend prediction.

[0153] Furthermore, the steps for quantitative prediction of probabilistic states are as follows.

[0154] The Monte Carlo method is used to randomly sample the parameters of the prediction model, generating multiple combinations of potential parameters to construct different prediction trajectories and quantify prediction uncertainty.

[0155] The parameters obtained by random sampling are input into the prediction model to generate multiple settlement prediction trajectories covering different future states.

[0156] Statistical analysis is performed on the predicted trajectory to calculate the probability of occurrence of different settlement levels, form a probability distribution, and extract key indicators, including mean prediction, confidence interval, and probability of extreme events.

[0157] The prediction process comprehensively considers model uncertainty, parameter uncertainty, and observation uncertainty, and updates the posterior distribution of model parameters through Bayesian inference.

[0158] The prediction results in the form of probability distribution are used as output to provide quantitative reference for decision-making, including average state, confidence interval range and probability of extreme settlement events.

[0159] As shown above, this embodiment generates multiple sets of prediction model parameters through Monte Carlo random sampling, constructing multiple potential settlement trajectories to quantify the uncertainty of future states. Through statistical analysis of the trajectories, the probability of different settlement levels is calculated, forming a probability distribution and extracting key indicators such as mean prediction, confidence intervals, and extreme event probabilities. In this process, model, parameter, and observational uncertainties are considered simultaneously, and Bayesian inference is used to update the posterior distribution of the model parameters. The final probabilistic prediction results provide a quantifiable reference for building settlement risk assessment and decision-making.

[0160] Furthermore, multivariate time series analysis includes the following steps.

[0161] The augmented Dickey-Fuller test was used to test the stationarity of the time series data of each influencing factor to ensure that the data met the stationarity requirements.

[0162] Perform cointegration tests on stationary or differencing time series to identify long-term equilibrium relationships among variables.

[0163] A multivariate time series model of influencing factors and settlement response was constructed, and the Granger causality test was used to analyze the strength of causal relationships and identify the dynamic impact of each factor on settlement changes.

[0164] The maximum likelihood method was used to estimate the parameters of the multivariate time series model, and the regression coefficients of each influencing factor and the residual characteristics of the model were obtained.

[0165] The variance decomposition method was used to calculate the proportion of each influencing factor that explains the sedimentation variation, so as to comprehensively reflect the direct and indirect effects of the factors.

[0166] The system outputs the dynamic causal relationships, contribution weights, and settlement prediction results of each factor, providing a quantitative basis for subsequent settlement risk assessment and management.

[0167] As can be seen from the above, this embodiment ensures that the influencing factor data meets the modeling requirements and identifies the long-term equilibrium relationship among variables through stationarity and cointegration tests. Based on this, a multivariate time series model is constructed, and the dynamic impact of each factor on settlement is analyzed using Granger causality tests. Maximum likelihood estimation is used to estimate parameters and obtain factor regression coefficients and residual characteristics. Variance decomposition is used to calculate the direct and indirect contributions of each factor to settlement variation, ultimately outputting dynamic causal relationships, contribution weights, and settlement prediction results. This provides a quantitative basis for settlement risk assessment, achieves a systematic analysis of the multi-factor mechanism, and improves the accuracy and reliability of monitoring and management.

[0168] Furthermore, after the multivariate time series analysis is completed, intelligent factor screening and weight optimization are implemented, specifically including...

[0169] The information entropy and conditional information entropy of each influencing factor are calculated. The information contribution of the factor to the sedimentation is quantified by mutual information. Redundant factors with mutual information less than 0.1 are automatically eliminated to avoid overcomplicating the model.

[0170] A three-tiered factor importance assessment system is established. The first tier is the assessment of direct impact, which quantifies the intensity of direct effects through linear regression coefficients. The second tier is the assessment of indirect impact, which identifies the indirect effects of factors through other variables through path analysis. The third tier is the assessment of interactive impact, which uses regression tree methods to identify nonlinear interaction patterns between factors.

[0171] An improved particle swarm optimization algorithm is used to optimize the factor weights in real time. The objective function is a weighted combination of minimizing prediction error and optimizing model complexity.

[0172] When the weight of a factor exceeds 1.5 times the average weight for three consecutive time steps, it is automatically marked as a current key factor. In the following 6 hours, the sampling frequency of the factor is increased to twice the normal frequency. At the same time, an additional 20% weight gain is assigned to the factor in the coupled modeling to ensure that key impacts are monitored.

[0173] As can be seen from the above, intelligent factor screening and weight optimization significantly improve the accuracy and efficiency of settlement prediction. First, low-contribution factors are eliminated based on mutual information, reducing redundancy and avoiding excessive model complexity. Then, a three-layer importance assessment system is constructed to comprehensively analyze the direct, indirect, and interactive effects of factors, ensuring the comprehensiveness of influence relationship identification. Finally, an improved particle swarm optimization is used to dynamically adjust weights, and after key factor identification, their sampling frequency and coupling weights are increased to achieve focused monitoring of important factors, thus balancing prediction accuracy, real-time performance, and system stability.

[0174] Furthermore, spatial correlation analysis includes the following steps.

[0175] A spatial weight matrix is ​​constructed based on the spatial location and structural connectivity of each part of the building to describe the spatial relationships between the parts.

[0176] Moran's I index was used to perform spatial autocorrelation analysis on settlement data to quantify the spatial clustering of settlement and identify hot and cold settlement areas.

[0177] Establish a regression model that considers spatial effects, analyze the relationship between influencing factors and settlement, and eliminate the interference of spatial correlation on parameter estimation.

[0178] Main effect analysis was used to assess the independent impact of a single factor on sedimentation, and interaction effect analysis was used to identify synergistic or antagonistic effects between factors, thereby revealing the complex coupling relationship of multiple factors.

[0179] The results of integrated spatial autocorrelation analysis, spatial regression and multi-factor coupling analysis provide a basis for the analysis of building settlement mechanism, risk identification and control strategy formulation.

[0180] As can be seen from the above, this embodiment quantifies the spatial clustering characteristics of settlement in various parts of the building by constructing a spatial weight matrix and applying Moran's I index, identifying settlement hotspots and cold spots. Based on this, a spatial effect regression model is used to analyze the relationship between influencing factors and settlement, eliminating the interference of spatial correlation on parameter estimation. Simultaneously, main effect and interaction effect analyses reveal the independent, synergistic, or antagonistic effects among factors, comprehensively characterizing the coupling relationship of multiple factors. This allows for a systematic analysis of the settlement mechanism, clarification of spatial distribution characteristics, and provides a scientific basis for risk identification and control strategy formulation.

[0181] Furthermore, based on the results of spatial correlation analysis, a mechanism for highlighting the role of key factors is established, and the identification and modeling of important influencing factors are strengthened through the following means.

[0182] A recursive feature elimination algorithm combined with cross-validation is used to dynamically evaluate the importance contribution of each factor to settlement prediction, calculate the factor significance score, and mark the top 30% of factors as key factors and increase their weight in coupled modeling.

[0183] By introducing kernel principal component analysis and manifold learning algorithms, we can explore the nonlinear coupling relationships between factors. By mapping the factor space to a high-dimensional feature space through Gaussian radial basis function kernel mapping, we can identify complex coupling patterns that are difficult to find through linear analysis.

[0184] Establish an adaptive adjustment mechanism for coupling strength, dynamically adjust the coupling coefficient between factors according to the settlement development stage (initial stage, development stage, stable stage, and acceleration stage), and automatically increase the coupling weight of key environmental factors by 40-60% during the settlement acceleration stage to ensure that the role of key factors in key stages is fully reflected.

[0185] The variance decomposition method is used to quantify the independent effects, pairwise interactions, and higher-order interactions of each factor, and a synergistic effect matrix is ​​established. When the synergistic effect exceeds 50% of the individual effect, a factor combination model is automatically established and the combination weights are increased.

[0186] As can be seen from the above, spatial correlation analysis and advanced modeling techniques enhance the identification and representation of key factors. First, recursive feature elimination and cross-validation are used to screen out the factors with the highest contribution, and their weights are increased in the modeling process. Then, kernel principal component analysis and manifold learning are used to uncover nonlinear coupling relationships, revealing complex patterns that are difficult to identify using conventional methods. Simultaneously, the factor coupling coefficients are dynamically adjusted according to the settlement stage, highlighting the role of environmental factors at critical stages. Finally, variance decomposition quantifies the independent and synergistic effects among factors, increasing the weights of combinations with significant synergistic effects to ensure that the model comprehensively and accurately reflects the settlement mechanism.

[0187] Furthermore, spatial correlation analysis is performed based on the initial analysis parameters, the environmental parameter data, and the quantitative prediction results of the probability state to identify the multi-factor coupling relationship between environmental factors, structural characteristics, and abnormal evolution processes, including the following steps.

[0188] A spatial weight matrix is ​​constructed based on the spatial location and structural connectivity of various parts of the building to describe the spatial relationship between different monitoring points.

[0189] The initial parameters, environmental parameters, and probabilistic state prediction results are uniformly organized and coded to enable unified processing of multiple factors in spatial correlation analysis.

[0190] Spatial autocorrelation indicators are used to quantify the aggregation characteristics of settlement in different parts of a building, identify abnormal settlement hotspots and cold spots, and provide a reference for causal relationship analysis.

[0191] Based on the constructed spatial weight matrix, a regression model considering spatial effects is established to analyze the influence of environmental factors and structural characteristics on the evolution of abnormal settlement, while eliminating the interference of spatial dependence on parameter estimation.

[0192] Through main effect analysis and interaction effect analysis, we quantitatively assess the direct and synergistic effects of various environmental factors and structural characteristics on the anomalous evolution process, revealing the multi-factor coupling mechanism.

[0193] The results of spatial correlation analysis and coupling relationship identification are output in a structured manner to provide quantitative basis for the multi-factor analysis module and risk assessment module, and to update the analysis parameters of the monitoring device.

[0194] As can be seen from the above, this embodiment integrates and analyzes initial parameters, environmental data, and probabilistic state prediction results, and constructs a spatial weight matrix by combining the spatial location and structural connectivity of various parts of the building, thereby realizing the description of the spatial relationships between monitoring points. First, various types of data are uniformly organized and coded to achieve unified multi-factor processing in spatial correlation analysis. Then, spatial autocorrelation indicators are used to quantify settlement aggregation characteristics, identify anomalous hotspots and cold spots, and provide a reference for causal relationship analysis. Based on the spatial weight matrix, a regression model considering spatial effects is established to analyze the impact of environmental factors and structural characteristics on the evolution of anomalous settlement, while eliminating the interference of spatial dependence on parameter estimation. Furthermore, through main effect and interaction effect analysis, the direct effects and synergistic or antagonistic effects of each factor are quantitatively evaluated, revealing the multi-factor coupling mechanism in the anomalous evolution process. Finally, the spatial correlation and coupling analysis results are output in a structured manner, providing a reliable quantitative basis for multi-factor analysis and risk assessment modules, and dynamically updating the analysis parameters of the monitoring device to achieve refined analysis and scientific management of building settlement status.

[0195] Furthermore, the two-dimensional risk matrix is ​​constructed with the probability of settlement occurrence as the horizontal axis and the severity of settlement consequences as the vertical axis. The horizontal axis is divided into five levels: extremely low, low, medium, high, and extremely high, while the vertical axis is divided into five levels: slight, moderate, severe, major, and extremely severe, thus forming 25 risk units. The risk mapping process converts the quantitative prediction results into risk levels through fuzzy membership functions. The probability level is determined based on the cumulative probability distribution of the prediction, and the consequence level is determined based on the degree of impact of settlement on the safety, functionality, and economic value of the building structure.

[0196] As can be seen from the above, this embodiment divides building settlement risk into 25 units using the probability of settlement occurrence as the horizontal axis and the severity of settlement consequences as the vertical axis, achieving quantitative risk classification management. The horizontal axis is divided into five levels: extremely low, low, medium, high, and extremely high, while the vertical axis is divided into five levels: minor, moderate, severe, major, and extremely severe, forming a clear risk grid. Through fuzzy membership functions, the probability distribution results obtained from deep learning predictions are mapped to the corresponding risk levels. Simultaneously, the consequence level is assessed based on the impact of settlement on structural safety, functionality, and economic value, thereby effectively transforming quantitative prediction results into actionable risk levels.

[0197] Furthermore, the assessment benchmark parameters include four core parameters: settlement rate benchmark value, cumulative settlement benchmark value, prediction trend coefficient, and environmental impact coefficient. Among them, the settlement rate benchmark value is obtained from the standard specifications based on the building type and geological conditions; the cumulative settlement benchmark value takes into account the building's design allowable deformation value and service life; the prediction trend coefficient reflects the acceleration characteristics of settlement development; and the environmental impact coefficient quantifies the amplification effect of external environmental factors on settlement.

[0198] As can be seen from the above, this embodiment achieves scientific and quantitative management of building settlement risk through four core indicators: settlement rate benchmark value, cumulative settlement benchmark value, prediction trend coefficient, and environmental impact coefficient. The settlement rate benchmark value is formulated based on building type and geological conditions to ensure that the assessment conforms to actual engineering characteristics; the cumulative settlement benchmark value, combined with design allowable deformation and service life, reflects the structure's total settlement capacity; the prediction trend coefficient depicts the acceleration change in settlement development, revealing potential aggravation risks; and the environmental impact coefficient quantifies the amplification effect of external factors on settlement, achieving a comprehensive consideration of environmental coupling effects. This overall parameter system provides a scientific and quantifiable basis for risk assessment and early warning, improving the accuracy and reliability of building settlement management.

[0199] Furthermore, the settlement trend curves are displayed using a multi-scale time axis, including real-time curves, daily trend curves, weekly trend curves, and monthly trend curves. Specifically, the real-time curve displays the settlement changes over the past 24 hours, the daily trend curve displays the average daily settlement over the past 30 days, the weekly trend curve displays the average weekly settlement over the past 12 weeks, and the monthly trend curve displays the average monthly settlement over the past 12 months. The Y-axis scale is automatically adjusted according to the data variation range, highlighting the detailed characteristics of settlement changes. Shaded areas indicate the uncertainty range of the prediction results, and prediction accuracy indicators are numerically labeled on the charts. The risk analysis report uses a structured format, including six parts: monitoring overview, anomaly analysis, risk assessment, development trend, influencing factor analysis, and recommended measures. The report content is automatically generated based on the current settlement status and risk level, ensuring the comprehensiveness and practicality of the decision support information.

[0200] As can be seen from the above, this embodiment provides comprehensive visual management of building settlement through a multi-scale time axis. Real-time curves show settlement changes over the past 24 hours, while daily, weekly, and monthly trend curves reflect the average settlement over the past 30 days, 12 weeks, and 12 months, respectively, combining short-term dynamic monitoring with long-term evolution analysis. The Y-axis automatically adjusts with the data range, refining the details of changes; shaded areas indicate prediction uncertainty, and prediction accuracy indicators are also marked, providing a quantitative reference for the monitoring results. The risk analysis report adopts a structured format, including six modules: monitoring overview, anomaly analysis, risk assessment, development trend, influencing factor analysis, and response recommendations. The content is automatically generated based on the current settlement status and risk level, achieving visualization of the settlement situation, scientific risk analysis, and comprehensive and practical decision support information.

[0201] Furthermore, the generation of the emergency response recommendations includes the following steps.

[0202] Establish an expert knowledge base covering standard handling procedures and best practices for different types of settlement and risk levels; at the same time, build a case reasoning system to store historical emergency response cases for retrieval and comparison.

[0203] Based on the similarity matching method, the most similar treatment plan to the current settlement risk is retrieved from the historical case database, and a set of candidate treatment plans is formed by combining the standard procedures in the knowledge base.

[0204] Taking into account the building's structural characteristics, functional use, surrounding environment, and available resources, a multi-criteria decision analysis method is used to select the best candidate solutions and identify feasible and effective disposal recommendations.

[0205] The optimized emergency response recommendations will be output to form a tiered and actionable decision support result, providing quantitative guidance for emergency response and risk management.

[0206] As can be seen from the above, this embodiment achieves a scientific response to building settlement risks by organically combining an expert knowledge base and a historical case database. First, it establishes standard handling procedures and best practices covering different settlement types and risk levels, while storing historical emergency cases for retrieval and comparison. Using a similarity matching method, it selects the handling plan most similar to the current risk from historical cases and generates a candidate solution set by combining it with the standard procedures in the knowledge base. Subsequently, considering the building's structural characteristics, usage functions, surrounding environment, and available resources, it uses a multi-criteria decision analysis method to optimize the candidate solutions, forming hierarchical and actionable emergency response recommendations. The final output recommendations are not only scientifically quantifiable and highly operable, but also provide clear guidance for on-site emergency response and risk management, improving the timeliness, accuracy, and reliability of building settlement management decisions.

[0207] Furthermore, based on the assessment benchmark parameters and the two-dimensional risk matrix, the automatic assessment and graded early warning of building structure evolution risk are realized, and comprehensive decision support information including settlement trend curves, risk analysis reports and emergency response suggestions is generated, including the following steps.

[0208] By combining the assessment benchmark parameters, each risk unit in the two-dimensional risk matrix is ​​automatically assessed to determine the corresponding risk level.

[0209] Multi-scale settlement trend curves are plotted based on monitoring data and prediction results to reflect the dynamic evolution of building structures and changes in potential risks.

[0210] Integrate risk assessment results, abnormal event records, and trend information to automatically generate structured risk analysis reports.

[0211] Based on expert knowledge base and historical case reasoning, hierarchical and actionable emergency response recommendations are generated according to the current risk level and structural status.

[0212] The settlement trend curve, risk analysis report and emergency response suggestions are integrated into comprehensive decision support information and pushed to the management terminal in real time to realize a closed loop of risk warning and decision support.

[0213] As can be seen from the above, this embodiment achieves risk quantification and level classification through evaluation benchmark parameters and a two-dimensional risk matrix. Risk units are automatically evaluated, risk levels are clearly defined, and multi-scale settlement trend curves are plotted based on monitoring data and prediction results, comprehensively reflecting the dynamic evolution of the building structure and changes in potential risks. Simultaneously, risk assessment results, abnormal event records, and trend information are integrated to generate a structured risk analysis report. Combining expert knowledge base and historical case reasoning, hierarchical and actionable emergency response suggestions are generated based on the current risk level and structural status, ensuring the feasibility and effectiveness of the emergency response plan. Finally, the settlement trend curves, risk analysis reports, and emergency response suggestions are integrated into comprehensive decision support information and pushed to the management terminal in real time, realizing closed-loop management from monitoring, analysis, early warning to emergency response.

[0214] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A device for real-time monitoring of settlement of a building ground, characterized in that, The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application relates to a building subsidence monitoring system. The application By calculating the covariance matrix between each influencing factor and the settlement response in real time, the real-time influence strength of the factors is dynamically evaluated; Real-time statistics are performed on the relationship between the factors and the settlement in the last 72 hours, and the contribution variance and stability index of each factor are calculated; By calculating the marginal contribution of each factor to the reduction of settlement prediction error, the importance ranking of the factors is updated in real time, and the weights of the top three factors are increased by 20-30%, while the weights of the last three factors are reduced accordingly, realizing the dynamic reallocation of weight resources; Considering the prediction accuracy, calculation efficiency and model stability, the Pareto optimal solution set is used to select the best weight configuration scheme to ensure that the key factor weight is increased without affecting the overall performance.

2. The real-time settlement monitoring device for building ground according to claim 1, characterized in that, The hierarchical sensor network includes a basic layer sensor array, a structural layer sensor node, and an environmental monitoring sensor group, wherein: the basic layer sensor array adopts a combination of high-precision laser displacement sensors and digital tilt sensors, and is installed at key positions of the building foundation according to the grid layout principle; the structural layer sensor node adopts wireless strain sensors and three-axis acceleration sensors, and is arranged along the main load-bearing structure of the building to realize real-time monitoring of the stress state change and dynamic response of the structure; the environmental monitoring sensor group includes soil moisture sensors, underground water level monitoring sensors, temperature and humidity sensors, and ground vibration sensors, which are used to obtain external environmental parameters affecting building settlement.

3. The method of using the real-time building ground settlement monitoring device, based on the real-time building ground settlement monitoring device of claim 2, wherein, The method comprises the following steps: According to the initial state of the settlement monitoring device, the key positions of the building and the preset monitoring accuracy requirements, the sampling frequency configuration of each sensor is determined, and a multi-time scale data acquisition scheme of the hierarchical sensor network is generated in combination with the key positions of the building and the sampling frequency configuration; The hierarchical sensor network is controlled to acquire building structure deformation data and environmental parameter data in real time according to the multi-time scale data acquisition scheme; Whenever the hierarchical sensor network completes a round of data acquisition, the currently acquired structure deformation data is marked as target monitoring data, the abnormality detection parameters of the data processing and analysis module are configured based on the target monitoring data and the preset adaptive threshold, and the target monitoring data is subjected to real-time abnormality identification to identify the instantaneous abnormal deformation of the building structure; During the real-time abnormality identification process, the duration of the instantaneous abnormal deformation is analyzed to obtain the duration characteristics of the instantaneous abnormal deformation, the instantaneous abnormal deformation and the duration characteristics are combined to classify and mark the abnormal deformation, and the corresponding feature fingerprint data is extracted; If the duration of the instantaneous abnormal deformation does not reach the preset classification standard, the real-time abnormality identification is maintained until the classification is completed; If the instantaneous abnormal deformation has been classified and marked, the abnormality identification process of the target monitoring data is ended, and the classification result is transmitted to the deep learning prediction module; By fusing feature data of different time windows and introducing an attention mechanism, the probability state quantitative prediction of the evolution of building abnormal deformation to continuous settlement is realized; Based on multivariate time series analysis, the contribution weights of each influencing factor to abnormal evolution are calculated, and the analysis initial parameters of the multi-factor analysis module are obtained. According to the analysis of the initial parameters, the environmental parameter data and the probability state quantitative prediction results, spatial correlation analysis is performed to realize the identification of the multi-factor coupling relationship between environmental factors, structural characteristics and abnormal evolution process; The abnormal classification results and the probability state quantitative prediction results are mapped to a two-dimensional risk matrix, and the evaluation benchmark parameters of the risk assessment and early warning module are obtained; Based on the evaluation benchmark parameters and the two-dimensional risk matrix, automatic assessment and hierarchical early warning of the building structure evolution risk are realized, and comprehensive decision support information including settlement trend curve, risk analysis report and emergency disposal suggestion is generated; During the generation of the comprehensive decision support information, real-time early warning notification is pushed, and the risk analysis report and the emergency disposal suggestion are combined to guide the on-site emergency response; If the building structure evolution risk level does not reach the early warning standard, the normal monitoring state is maintained and data collection is continued; If the building structure evolution risk level has reached the early warning standard, the emergency plan is immediately started, and the building ground real-time settlement monitoring device is controlled to enter the high-frequency monitoring mode.

4. The method of using a device for monitoring real-time settlement of a building ground surface according to claim 3, wherein, The duration analysis is used to statistically analyze the duration characteristics of the instantaneous abnormal deformation, and the instantaneous abnormal deformation and the duration characteristics are combined to classify and mark the abnormal deformation, including the following steps: A time window is established for each instantaneous abnormal deformation event to track its starting time and duration in real time, thereby quantifying the duration characteristics of the abnormal deformation; The key duration parameters of the abnormal deformation event are calculated, including the duration, cumulative deformation amplitude and trend change; According to the deviation degree of the key duration parameters of the abnormal deformation relative to the normal range, the abnormal intensity is evaluated and a reference is provided for classification; The duration characteristics and the intensity evaluation results are combined to classify the abnormal deformation and determine whether the preset threshold is reached; The classification results are marked in a structured manner, the abnormal type and characteristic parameters are recorded, and are synchronized to the deep learning prediction module and the multi-factor analysis module to update the abnormal deformation feature library.

5. The method of using a device for monitoring real-time settlement of a building ground according to claim 4, wherein, The probability state quantitative prediction implementation steps are as follows: Monte Carlo method is used to randomly sample the prediction model parameters to generate multiple potential parameter combinations, to construct different prediction trajectories and quantify the prediction uncertainty; The parameters obtained by random sampling are input into the prediction model to generate multiple settlement prediction trajectories covering different future states; Statistical analysis is performed on the prediction trajectories to calculate the occurrence probability of different settlement levels, form a probability distribution and extract key indicators, including mean prediction, confidence interval and extreme event probability; In the prediction process, model uncertainty, parameter uncertainty and observation uncertainty are considered comprehensively, and the posterior distribution of the model parameters is updated through Bayesian inference; The prediction results in the form of probability distribution are taken as the output to provide quantitative reference for decision-making, including average state, confidence interval range and extreme settlement event probability.

6. The method of using a device for monitoring real-time settlement of a building ground according to claim 5, wherein, According to the analysis of the initial parameters, the environmental parameter data and the probability state quantitative prediction results, spatial correlation analysis is performed to realize the identification of the multi-factor coupling relationship between environmental factors, structural characteristics and abnormal evolution process, including the following steps: According to the spatial position and structural connectivity of each part of the building, a spatial weight matrix is constructed to describe the spatial relationship between different monitoring points; The initial parameters, environmental parameter data, and probability state prediction results are uniformly sorted and coded to facilitate the uniform processing of multiple factors in the spatial correlation analysis; The spatial autocorrelation index is used to quantify the aggregation characteristics of the settlement of each part of the building, identify abnormal settlement hot and cold spot areas, and provide a reference for causal relationship analysis; Based on the constructed spatial weight matrix, a regression model considering spatial effects is established to analyze the influence of environmental factors and structural characteristics on the evolution of abnormal settlement, while eliminating the interference of spatial dependence on parameter estimation; Through main effect analysis and interaction effect analysis, the direct action and synergistic effect of each environmental factor and structural characteristic on the abnormal evolution process are quantitatively evaluated, and the multi-factor coupling mechanism is revealed; The results of spatial correlation analysis and coupling relationship identification are structured and output to provide quantitative basis for the multi-factor analysis module and risk assessment module, and to update the analysis parameters of the monitoring device.

Images