An urban building structure health state online monitoring and early warning management system and method

Abstract

This invention discloses an online monitoring and early warning management system and method for the health status of urban building structures, relating to the field of building structure health monitoring technology. It includes a building structure acquisition module, a building data analysis module, a structural analysis module, and an early warning management module. The system collects building structure data, environmental data, and operational status data to form time-series data; it performs joint analysis of current and historical data to identify data drift points and calculate drift differences; it assesses building health status through drift compensation and an isolated forest anomaly detection model; it integrates health status levels, trend coefficients, and drift differences for comprehensive risk assessment to determine early warning levels; and it classifies structural units into safety point categories based on risk propagation values ​​and evolution rates, establishing a safe passage network and dynamically planning optimal evacuation paths. This invention achieves effective compensation for data drift and accurate identification of structural anomalies, improving early warning accuracy and emergency response capabilities.

Description

Technical Field

[0001] This invention relates to the field of building structural health monitoring technology, specifically to an online monitoring and early warning management system and method for the health status of urban building structures. Background Technology

[0002] With the acceleration of urbanization, the number of high-rise buildings, large-scale public infrastructure, and old buildings is constantly increasing. Over long periods of service, these buildings are affected by factors such as material aging, environmental erosion, load variations, and natural disasters, leading to a gradual degradation of their structural performance and even sudden damage. Therefore, real-time monitoring of building structural health, timely detection of anomalies, and early warning are of great significance for safeguarding people's lives and property.

[0003] Currently, various monitoring systems exist in the field of building structural health monitoring, typically employing the deployment of various sensors to collect data and setting thresholds for alarms. However, existing systems and methods still have the following shortcomings in practical applications: Different physical quantities are often analyzed in isolation, lacking a unified method for characterizing structural state, making it difficult to comprehensively reflect the overall changes in structural health status, resulting in one-sided analysis results.

[0004] Long-term operation of sensors can cause data drift due to zero-point drift, environmental interference, etc. Existing methods mostly use simple filtering or fixed threshold correction, which fail to effectively identify data drift points and perform accurate compensation, thus affecting the accuracy of anomaly detection.

[0005] Commonly used methods such as thresholding and statistical process control are poorly adapted to complex nonlinear structural states, making it difficult to detect early minor anomalies and unable to effectively distinguish between data drift and real structural anomalies, which can easily lead to false alarms or missed alarms.

[0006] Most systems only provide alarms and lack the ability to classify and identify safety points in the building's internal structure and plan dynamic evacuation routes. They cannot provide scientific guidance for personnel evacuation in emergencies and have limited emergency response capabilities. Summary of the Invention

[0007] The purpose of this invention is to address the problems of low multi-source data fusion, difficulty in effectively identifying and compensating for data drift, and high missed rate of early minor anomalies due to single anomaly detection methods in existing building structure health monitoring systems. Therefore, this invention proposes an online monitoring and early warning management system and method for the health status of urban building structures.

[0008] The objective of this invention can be achieved through the following technical solution: This invention provides an online monitoring and early warning management system for the health status of urban building structures, including a building structure acquisition module, a building data analysis module, a structure analysis module, and an early warning management module; The building structure acquisition module collects urban building data, which includes building structure data, environmental data, and operational status data, and forms time-series urban building data through multi-source sensors. The building data analysis module performs joint analysis of current urban building data and historical urban building data to obtain data drift points, and calculates the drift difference based on the data drift points to obtain the corresponding drift difference value. The structural analysis module obtains the drift difference value and the corresponding building structure data for comprehensive analysis to obtain drift compensation, and performs building health status monitoring to obtain the health status of the building structure; it also performs comprehensive analysis of building operation risks to obtain the building structure risk value; and compares the risk value with the preset risk threshold to obtain the warning level. The early warning management module performs safety early warning and emergency management of building structures based on the early warning judgment results. It conducts a comprehensive analysis of the safety status of each building structure to obtain the safe points, potential risk points, risk points, and high-risk points of the building structure. Based on this, it establishes an internal safe passage network, calculates the optimal evacuation route, and generates a safe path.

[0009] As a preferred embodiment of the present invention, the specific process of jointly analyzing current urban building data and historical urban building data is as follows: A unified data representation is used for time-series urban building data to establish a multi-dimensional structural response state vector; information entropy weight allocation is set to establish a weighted structural state sequence; discrete Fourier transform is performed on the structural state sequence to obtain the frequency domain response function, and the frequency domain energy distribution function is calculated. A structural state network is established using complex network modeling methods. The network centrality index and its change are calculated to determine whether the structural state has shifted. After detecting the structural state shift, a long short-term memory network prediction model is used to train the historical structural state sequence to obtain the prediction function. The prediction residual is calculated and the cumulative deviation is calculated to determine the candidate points of data drift. Dynamic time-warped distance calculation is performed on the structural state sequence before and after the drift candidate point to verify and determine the real data drift point.

[0010] In a preferred embodiment of the present invention, the process by which the building data analysis module calculates the drift difference based on the identified data drift points is as follows: A data evolution analysis sequence is established with the data drift point as the center. The sequence morphology is rearranged to obtain the sequence morphology distribution vector. The sequence complexity index is calculated to characterize the degree of change in the arrangement structure of the data sequence near the drift point. The probability distribution function of the sequence before and after the drift point is established, and the degree of difference between the two sets of data distributions is calculated through the distribution deviation function. The data change intensity index is calculated, and the data drift difference is calculated by integrating the sequence complexity change, distribution deviation degree and change intensity.

[0011] As a preferred embodiment of the present invention, the specific process of obtaining the drift difference value and the corresponding building structure data for comprehensive analysis is as follows: Obtain building structure status data and corresponding drift differences at different times; Each state parameter is adjusted by the drift compensation coefficient to obtain the compensated structural state vector; The compensated structural state vector is input into the isolated forest anomaly detection model to calculate the path length and average path length of the sample in the random isolated tree, and anomaly index is obtained. The health status of a building structure is determined by comparing abnormal indicators with health status assessment thresholds.

[0012] As a preferred embodiment of the present invention, the specific process for comprehensively analyzing building operation risks is as follows: Obtain the current building structure health status level parameters and the corresponding structural status data set; extract the time dimension trend of the structural status data, and obtain the change trend coefficient through linear regression model fitting; The risk value of the building structure is calculated by a risk assessment function that combines health status level parameters, trend coefficients, and drift differences. The risk value is compared with a preset risk threshold to determine the warning level and generate a warning message which is sent to the warning management module.

[0013] As a preferred embodiment of the present invention, the specific process for comprehensively analyzing the safety status of each building structure is as follows: Receive the building structure health status, drift compensation results and building structure data, establish a structural state expression matrix and perform normalization transformation; Structural potential energy conversion is performed to obtain structural potential energy values, spatial coupling effects of risks are identified, a structural spatial connection matrix is ​​established, and risk propagation calculations are performed. The risk propagation value is converted to a time trend, the risk evolution rate is calculated, and the combined risk index is obtained. Based on the comprehensive risk index, the internal structural units of the building are classified and identified to obtain the building structural safety points, potential risk structural points, risk structural points and high-risk structural points. Establish a safe passage network within the building based on safe points, calculate the optimal evacuation route, generate safe routes, and display the safe routes and issue evacuation instructions through the management terminal when an early warning is triggered.

[0014] In a preferred embodiment of the present invention, after generating a safe path, the early warning management module generates a safe path if it detects that the optimal evacuation path does not contain any potential risk structural points, risk structural points, or high-risk structural points; otherwise, it repairs the risk points in the path and re-corrects the safe route.

[0015] Another aspect of the present invention provides an online monitoring and early warning management method for the health status of urban building structures, the method comprising: S1: Continuously collect building structure data, environmental data and operational status data of the target building through multi-source sensors to form time-series urban building data; S2: The building data analysis module performs joint analysis on current urban building data and historical urban building data, identifies data change trends, obtains data drift points, and calculates the drift difference based on the identified data drift points to obtain the corresponding drift difference. S3: Obtain the drift difference and corresponding building structure data through the structural analysis module, perform drift compensation and building health status monitoring, and obtain the health status of the building structure; S4: After obtaining the health status of the building structure, the structural analysis module performs a comprehensive analysis of the building operation risk to obtain the building structure risk value. The risk value is compared with the preset risk threshold, and the warning level is obtained based on the comparison result. When the warning level reaches the preset trigger condition, the warning information is sent to the warning management module. S5: The early warning management module performs safety early warning and emergency management of building structures based on the early warning judgment results.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention utilizes multi-source data fusion and drift compensation to monitor structural health status. During monitoring, it collects real-time structural data such as building vibration, stress, displacement, and cracks, as well as environmental and operational status data, to establish a multi-dimensional structural response state vector. A weighted structural state sequence is then created through information entropy weight allocation. If a trend shift is detected in the data, the drift difference is obtained through permutation complexity analysis, distribution deviation calculation, and change intensity fusion. Drift compensation is then applied to each monitoring parameter, making the compensated structural state vector more accurately reflect changes in the building itself. After implementation, zero-point drift and environmental interference caused by long-term sensor operation are effectively suppressed, reducing the false alarm rate of anomaly detection. In the isolated forest anomaly identification stage, because the input data has eliminated drift interference, the model can accurately capture early minor structural anomalies, making the health status classification more reliable.

[0017] 2. In the risk warning stage, this invention achieves dynamic hierarchical early warning of building safety status by integrating health status level, trend coefficient, and drift difference for comprehensive risk assessment. When an early warning is triggered, the system classifies the safety points of each structural unit inside the building based on structural risk propagation value and risk evolution rate, establishes a safe passage network, and dynamically plans the optimal evacuation route. In emergencies, because potential risk structural points, high-risk structural points, and emergency dangerous structural points have been identified in advance, the generated evacuation routes can proactively avoid dangerous areas, improving personnel evacuation efficiency and reducing the risk of secondary accidents. Attached Figure Description

[0018] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0019] Figure 1 This is a schematic diagram of the principle of the present invention; Figure 2 This is a flowchart of the method steps of the present invention. Detailed Implementation

[0020] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0021] It should be understood that the terms “comprising” and “including” used in this disclosure and claims indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0022] It should also be understood that the terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to limit the disclosure. As used in this disclosure and claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this disclosure and claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations.

[0023] Please see Figure 1 As shown, an online monitoring and early warning management system for the health status of urban building structures includes: a building structure acquisition module, a building data analysis module, a structure analysis module, and an early warning management module; The building structure acquisition module collects and acquires urban building data, which includes: building structure data, environmental data, and operational status data; among which: Building structural data includes: building vibration data, stress data, displacement data, tilt data, and crack monitoring data; environmental data includes: wind speed data, temperature data, humidity data, rainfall data, and seismic vibration data; operational status data includes: building usage load data, personnel density data, and equipment operation vibration data, etc.

[0024] The above data is continuously collected by multiple source sensors to form time-series urban building data.

[0025] The building data analysis module performs joint analysis of current urban building data and historical urban building data to identify data change trends and obtain data drift points; The drift difference is calculated based on the identified data drift points to obtain the corresponding drift difference value.

[0026] The structural analysis module acquires the drift difference and performs comprehensive analysis with the corresponding building structural data to obtain drift compensation. It also monitors the building's health status to determine the overall health condition of the structure. Health status includes: normal, slightly abnormal, moderately abnormal, and severely abnormal.

[0027] After obtaining the building structure health status, a comprehensive analysis of the building operation risks is conducted to obtain the building structure risk value; the risk value is compared with a preset risk threshold; and an early warning level is obtained based on the comparison result; the early warning level includes: Level 1 early warning, Level 2 early warning, Level 3 early warning, and no early warning state; when the risk level reaches the preset trigger condition, the early warning information is sent to the early warning management module.

[0028] The early warning management module performs safety early warning and emergency management of the building structure based on the early warning judgment results. The specific process is as follows: The system acquires the health status of building structures, drift compensation results, and building structure data. It then conducts a comprehensive analysis of the safety status of each building structure to obtain the safe points, potential risk points, risk points, and high-risk points of the building structure.

[0029] An internal safe passage network is established based on the building's structural safety points, potential risk structural points, risk structural points, and high-risk structural points. The optimal evacuation route is calculated to form a safe path. If the optimal evacuation route does not contain any potential risk structural points, risk structural points, or high-risk structural points, then a safe path is generated. Alternatively, structural repairs are performed on the potential risk structural points, risk structural points, or high-risk structural points that exist on the optimal evacuation route. If a safe path is generated, the safe route consisting of the safe exits is then revised.

[0030] When a building is impacted or experiences structural abnormalities, corresponding early warning information is triggered, a safe path is displayed on the management terminal, evacuation instructions are issued to personnel, and personnel are guided to evacuate along the safe path.

[0031] The specific process of jointly analyzing current urban building data and historical urban building data is as follows: The acquired time-series urban building data is presented in a unified data representation; let the currently collected urban building data be D. c (t), historical city building data is D h (t), where t represents the sampling time.

[0032] Based on the different dimensions of building vibration data, stress data, displacement data, and environmental data, a unified structural state vector is established, mapping different types of monitoring data into a multidimensional structural response state vector: X(t)=[x1(t), x2(t), ..., x m [(t)], where: X(t) represents the building structure state vector at time t; x i (t) represents the data value of the i-th type of monitoring data at time t; i is the data type number of the monitoring data, i=1,2,...,m; m represents the total number of monitoring data types.

[0033] Subsequently, the information entropy weight allocation process is set up for each dimension of data to calculate the weight coefficient w. i And establish a weighted structure state sequence: Where: S(t) represents the structural response value at time t; w i x represents the weighting coefficient corresponding to the i-th type of monitoring data; i (t) represents the data value of the i-th type of monitoring data at time t.

[0034] Subsequently, the structural state sequence S(t) is transformed into the frequency domain for feature reconstruction; specifically, a discrete Fourier transform is performed on the structural state sequence to obtain the frequency domain response function: Where: F(ω) represents the frequency domain response function at frequency ω; S(t) represents the comprehensive structural response value at time t; N represents the length of the data sequence involved in the transformation; ω represents the frequency variable; e represents the natural constant; and j represents the imaginary unit.

[0035] Further calculation of the frequency domain energy distribution function: E(ω)=|F(ω)| 2 , where: E(ω) represents the energy distribution value at frequency ω; |F(ω)| represents the amplitude of the frequency domain response function.

[0036] After obtaining the structural energy spectrum features, a complex network modeling method is used to establish the structural energy spectrum features within different time windows as a structural state network. Specifically, the energy spectrum feature vector within each time window is denoted as node V. k The connection relationship between nodes is established based on the feature similarity between nodes to form a structural state network: G=(V,E), where: G represents the structural state network; V represents the set of nodes, and each node corresponds to the structural state of a time window; E represents the set of connection edges between nodes.

[0037] Subsequently, the network centrality index C of the computing nodes k ΔC k =C k C k 1, where: C k C represents the network centrality of the node in the k-th time window; k 1 represents the network centrality of the node in the previous time window; ΔC k This represents the change in centrality.

[0038] When |ΔC k | greater than the preset network stability threshold T c If the current structural state deviates from its position within the overall network structure, further time-series drift detection is required; where, for example, a preset network stability threshold T is used. c =0.15.

[0039] After detecting structural state shifts, a prediction function is obtained by training a long short-term memory network prediction model on historical structural state sequences. =f(S(t 1), S(t) 2), ..., S(t) n)), where: S(t) represents the predicted structural response value at time t; f(·) represents the prediction function model obtained through training; S(t) represents the predicted structural response value at time t. 1), S(t) 2), ..., S(t) (n) represents the structural response values ​​at the first n historical moments; n represents the time step of the prediction model.

[0040] The prediction residual is then calculated: R(t) = S(t) Where: R(t) represents the prediction residual at time t; S(t) represents the actual structural response value; This represents the predicted structural response value.

[0041] Further calculation of the cumulative bias of the residual sequence: Where: A(t) represents the cumulative residual value from the initial time to the current time t; R( ) indicates the first The prediction residual at each time point.

[0042] When the cumulative residual A(t) is greater than the preset drift judgment threshold T d When the current structural state deviates from the historical operating mode, this moment t is marked as a candidate point for data drift; where the preset drift judgment threshold T d For example, T d =0.05.

[0043] Subsequently, a reverse similarity verification was performed on the structural state sequences before and after the drift candidate point. Specifically, this was done by calculating the dynamic time warping distance between the two time window sequences before and after the candidate point: D dtw (i1, j1), where: D dtw (i1, j1) represents the dynamic time warping distance between the i1th time series and the j1th time series; i1 and j1 represent the two time series numbers involved in the matching.

[0044] When D dtw (i1, j1) is greater than the preset similarity threshold T s If so, then the candidate point is determined to be the actual data drift point P. d (t).

[0045] The process of calculating the drift difference based on the identified data drift points is as follows: After confirming the data drift point P d After (t), a data evolution analysis sequence is established with the data drift point as the center; assuming n1 sampling points are selected before and after the drift point, a data sequence set Ω={S(t)} is formed. n1), S(t) n1+1), ..., S(t+n1)}, where: Ω represents the set of drift analysis data sequences; n1 represents the data length parameter involved in the analysis; Subsequently, the sequence set Ω is sorted and mapped by the sequence morphology rearrangement process to obtain the sequence morphology distribution vector U(t) = rank(Ω), where: U(t) represents the sequence morphology sorting vector; rank(point) represents the index vector after sorting and mapping the sequence data according to the numerical size.

[0046] Next, the permutation complexity of the sorted morphological distribution vector is calculated to obtain the sequence complexity index. Where: C(t) represents the sequence complexity index; Show the first The probability of a given permutation pattern occurring in the sequence; k1 represents the number of all possible permutations; ln represents the natural logarithm function. It should be noted that the sequence complexity index C(t) characterizes the degree of change in the permutation structure of the data sequence near the drift point; its permutation complexity changes when the data state changes.

[0047] Subsequently, to further characterize the spatial distribution differences of the data states, a probability distribution function is established for the sequence subsets before and after the drift point; specifically, the data set before the drift point is denoted as Ω1={S(t n1), ..., S(t) 1)}; Let the set of data after the drift point be Ω2={S(t+1),...,S(t+n1)}; and calculate its probability density distribution, P1(x) and P2(x) respectively; where P1(x) represents the probability density function of the sequence before the drift point; P2(x) represents the probability density function of the sequence after the drift point; and x represents the data value variable.

[0048] Subsequently, the degree of difference between the distributions of the two sets of data was calculated using the distribution deviation function; , where D(t) represents the degree of deviation between the distributions of the two sets of data.

[0049] Next, an index of the intensity of data change is calculated for the sequence set. , where: G(t) represents the data change intensity index; S(j1) represents the j1-th data state value in the sequence; |·| represents the absolute value operation.

[0050] Finally, by incorporating changes in sequence complexity, distribution deviation, and the intensity of these changes, the final data drift difference is calculated: δ(t) = α × |C(t) C(t 1) |+β×D(t)+γ×G(t), where: δ(t) represents the data drift difference at time t; α, β and γ represent weighting coefficients used to adjust the influence of each index on the drift difference, where α=0.3, β=0.4 and γ=0.3.

[0051] The process of obtaining the drift difference and performing a comprehensive analysis with the corresponding building structure data is as follows: Two types of data are acquired: building structure status data and drift difference at corresponding time points. Current building structure state data X(t) = [x 12 (t), x 22 (t), ..., x m2 [(t)], where each x i2(t) is the physical quantity actually measured by the sensor; the drift difference δ(t) at the corresponding time; the physical quantity includes vibration, stress, displacement, etc. To eliminate the impact of data drift on structural health assessment, a drift compensation coefficient λ is used. i Adjust each state parameter: x i2 ′(t)=x i2 (t) λ i2 ·δ(t), where: x i2 ′(t) represents the state parameters after drift compensation; λ i2 The influence weight of structural parameters on drift difference is obtained by statistically analyzing the sensitivity of each parameter in historical data; for example, it is defined as λ. i2 =0.8.

[0052] The compensated structural state vector is obtained through the above compensation process, X′(t)=[x 12 ′(t), x 22 ′(t), ..., x m2 X′(t)], where X′(t) represents the compensated structural state vector; Subsequently, the compensated structural state vector is input into the isolated forest anomaly detection model for comprehensive analysis; Establish an isolated forest model set F: F = {T} 12 T 22 , ..., T L2}, where: F represents the set of isolated forest models; T 2 indicates the first Two random isolation trees; L2 represents the number of random isolation trees.

[0053] The compensated structural state vector X′(t) is input into each random isolation tree for path partitioning, and the path length h of the sample in each tree is calculated. 2(X′(t)), where: h 2(X′(t) represents the sample X′(t) at the th... The path length between two isolated trees.

[0054] Then calculate the average path length of the sample throughout the entire isolated forest: , where H(t) represents the average path length of a sample in the isolated forest model.

[0055] Subsequently, anomaly indicators were calculated based on the average path length. Where: Q(t) represents the structural state anomaly index; c(m) represents the normalization factor of the path length; and m0 represents the scale of the data used in training. The anomaly index Q(t) represents the degree to which the current structural state deviates from the normal state. The more abnormal the structural state, the closer the anomaly index is to 1.

[0056] After obtaining the abnormal indicators, they are compared with the health status judgment thresholds. The thresholds are determined based on the abnormal indicators and historical experience data, including: health status thresholds T1, T2, and T3, respectively, and satisfying 0 < T1 < T2 < T3 < 1. For example, T1 = 0.3, T2 = 0.6, and T3 = 0.85. If Q(t) < T1, the current building structure health status is determined to be normal; if T1 ≤ Q(t) < T2, the current building structure health status is determined to be slightly abnormal; if T2 ≤ Q(t) < T3, the current building structure health status is determined to be moderately abnormal; if Q(t) ≥ T3, the current building structure health status is determined to be severely abnormal.

[0057] The specific process for conducting a comprehensive analysis of building operation risks is as follows: Obtain the current building structural health status H s (t) and the corresponding structural state data set X′(t)=[x 12 ′(t), x 22 ′(t), ..., x m2 H′(t)], where H s (t) is the health status level parameter, and its value is assigned according to the health status: when the health status is normal, H s (t)=1; when the health status is a slightly abnormal state, H s (t)=2; when the health status is a moderately abnormal state, H s (t)=3; when the health status is a severely abnormal state, H s (t)=4.

[0058] Subsequently, trend extraction is performed on the structural state data along the time dimension. Assuming a historical data window of length k2 is selected up to the current time t, a structural state time series is established, Ψ={E(t... k2+1), E(t) k2+2), ..., E(t), where: E(t) represents the comprehensive state index obtained in the structural health assessment stage; k2 represents the length of the time window involved in the trend calculation.

[0059] The trend coefficient is obtained by fitting the sequence Ψ to the trend using a linear regression model. Where: θ(t) represents the structural state change trend coefficient; Represents the first in a time series Several status indicators.

[0060] The trend coefficient θ(t) represents the direction and rate of change of the structural state over time. When θ(t) is positive and has a large value, it indicates that the structural state is developing in an abnormal direction.

[0061] Subsequently, the building structure risk value is calculated using the risk assessment function: J(t) = α1·Hs(t) + β1·|θ(t)| + γ1·δ(t), where J(t) represents the building structure risk value at time t; α1, β1, and γ1 are risk weight coefficients used to adjust the degree of influence of different factors on the risk value, and are defined as α1 = 0.5, β1 = 0.3, and γ1 = 0.2, respectively. After obtaining the risk value J(t), it is compared with a preset risk threshold. The risk thresholds are J1, J2, and J3, and satisfy J1 < J2 < J3. For example, J1 = 0.2, J2 = 0.5, and J3 = 0.8. If J(t) < 1, it is determined to be a state without warning; if J1 ≤ J(t) < J2, it is determined to be a level 3 warning; if J2 ≤ J(t) < J3, it is determined to be a level 2 warning; if J(t) ≥ J3, it is determined to be a level 1 warning.

[0062] When the risk level reaches the preset trigger condition, a corresponding early warning message is generated and sent to the early warning management module; the early warning message includes at least the building number, the current risk value, the early warning level, and the corresponding time information.

[0063] The comprehensive analysis of the safety status of each building structure is carried out in the following process: The system receives the building structure health status Hs(t), drift compensation X′(t), and the corresponding building structure data set output from the structural analysis module, and performs unified state representation processing on the data. Specifically, different types of monitoring parameters are mapped to a unified structural response vector, and a structural state representation matrix is ​​established, M4(t)=[m i4j4 (t)] n4×k4 Where: M4(t) represents the structural state representation matrix at time t; m i4j4 (t) represents the state value of the i4th structural unit after compensation under the j4th type of monitoring parameter; n4 represents the number of structural units; k4 represents the number of monitoring parameter types.

[0064] Next, a state normalization transformation is performed on the matrix to obtain the dimensionless response matrix. , where: Y i4j4 (t) represents the normalized structural response value; σ represents the historical mean of the j4th type parameter; j4This represents the historical standard deviation of the j4th type parameter.

[0065] Subsequently, a state potential energy transformation is performed on the normalized response matrix; specifically, the degree of deviation of the structural state is mapped to the structural risk potential energy to obtain the structural potential energy value. , where: E i4 (t) represents the structural potential energy value of the i-th structural unit; After obtaining the structural potential energy value, the spatial coupling effect of the risk is further identified through the structural risk transmission and transformation process. A structural spatial connection matrix is ​​established, A4=[a i5j5 ] n5×n5 , where: a i5j5 =1 indicates that structural unit i5 and structural unit j5 have a structural connection relationship; a i5j5 =0 indicates that there is no direct connection between the two.

[0066] Based on the connectivity matrix, the spatial transmission of structural potential energy is calculated to obtain the structural risk propagation value. , where: P i5 (t) represents the risk propagation value of structural unit i; d i5j5 The spatial distance between structural units is shown; λ5 represents the risk coupling coefficient, with an example λ5=0.3.

[0067] After obtaining the structural risk propagation value, a time trend transformation is performed on the risk propagation value; specifically, within the time window [t]... Establish a risk propagation sequence within k5, t], Ω i5 ={P i5 (t k5+1), P i5 (t k5+2), ..., P i5 (t)}, and calculate the risk evolution rate, , where: V i5 (t) represents the rate of structural risk evolution, used to reflect the trend of risk growth.

[0068] Subsequently, the structural risk propagation value and risk evolution rate are fused together to obtain the comprehensive structural risk index, Z. i5 (t)=P i5 (t)×(1+μ5V i5 (t)), where: Z i5 (t) represents the comprehensive risk index of the structural unit; μ5 represents the trend adjustment coefficient, with μ5=0.1 in the example, which is used to adjust the impact of the risk growth trend on the result.

[0069] Based on the calculated comprehensive risk index Z i5(t) is used to classify and identify the internal structural units of the building. Risk thresholds R1, R2, and R3 are set, and if R1 < R2 < R3, four types of structural points are obtained: If the comprehensive risk index Z i5 When (t) < R1, the structural unit is determined to be a safe point in the building structure; if R1 ≤ comprehensive risk index Z i5 When (t) < R2, it is identified as a potential risk structure point; if R2 ≤ comprehensive risk index Z i5 When (t) < R3, it is judged as a high-risk structural point; if the comprehensive risk index Z i5 When (t)≥R3, it is determined to be an emergency dangerous structural point. In this embodiment, R1 is 0.3, R2 is 0.6, and R3 is 0.9.

[0070] After completing the risk classification of structural nodes, the early warning management module will spatially map various nodes to form a risk distribution map of the building's internal structure.

[0071] Subsequently, the structural nodes are screened for safety, and only the set of safe points of the building structure Vs={i5|Zi5(t)<R1} is retained. The safe points are used as nodes to establish the safe passage network Gs=(Vs, Es), where Es represents the passable connection relationship between safe nodes.

[0072] Furthermore, the path planning process fully considers the risk warning results, performs risk modulation and transformation on the path weights, and establishes a path risk cost function, C. i5j5 =d i5j5 (1+α5Z j5 +β5V j5 ), where: C i5j5 This represents the path cost from node i5 to node j5; d i5j5 Z represents the path distance; j5 V represents the node risk index; j5 α5 represents the risk evolution rate; β5 and β5 are adjustment coefficients. In this embodiment, α5 is 0.5 and β5 is 0.3.

[0073] By performing path optimization calculations on the safe passage network, the optimal evacuation path, P, is obtained. =argmin∑C i5j5 , where: P This represents the optimal safe evacuation route.

[0074] When an impact or structural abnormality is detected in the building and the warning level reaches the trigger condition, the warning management module immediately activates the warning mechanism, displaying a building risk distribution map and the optimal safe path P in real time on the management terminal. .

[0075] Evacuation instructions are issued to people inside the building through broadcasting systems, mobile terminals, or building display terminals. Based on the real-time location of the people, the nearest safe node is matched, and individualized evacuation routes are dynamically generated to guide people to avoid potential risk structural points, high-risk structural points, and emergency dangerous structural points, and to evacuate in an orderly manner along the safe routes.

[0076] Please see Figure 2 As shown, a method for online monitoring, early warning, and management of the structural health status of urban buildings includes the following steps: S1: The building structure acquisition module uses multi-source sensors to continuously collect data on the target building to obtain urban building data; the collected data is then organized into time-series urban building data.

[0077] S2: The building data analysis module performs joint analysis on the urban building data collected at the current moment and the historical urban building data, identifies the data change trend, obtains the data drift point, and calculates the drift difference based on the identified data drift point to obtain the corresponding drift difference value.

[0078] S3: Obtain the current building structure state vector and the corresponding drift difference, adjust each state parameter through the drift compensation coefficient to obtain the compensated structure state vector; input the compensated structure state vector into the isolated forest anomaly detection model, calculate the average path length of the sample in the isolated forest, and then obtain the anomaly index; compare the anomaly index with the preset health status judgment threshold to determine whether the current building structure health status is normal, slightly abnormal, moderately abnormal, or severely abnormal.

[0079] S4: Obtain the current building structure health status level parameters and the compensated structural status data set; extract the trend of the structural status data over time; obtain the trend coefficient by fitting a linear regression model; calculate the building structure risk value using a risk assessment function that comprehensively considers the health status level parameters, the absolute value of the trend coefficient, and the drift difference, and assigns corresponding risk weight coefficients; compare the calculated risk value with a preset risk threshold; if the risk value is less than the first risk threshold, it is determined to be in a no-warning state; if the risk value is greater than or equal to the first risk threshold and less than the second risk threshold, it is determined to be a level 3 warning; if the risk value is greater than or equal to the second risk threshold and less than the third risk threshold, it is determined to be a level 2 warning; if the risk value is greater than or equal to the third risk threshold, it is determined to be a level 1 warning; when the warning level reaches the preset triggering condition, generate a warning message containing the building number, current risk value, warning level, and corresponding time information, and send the warning message to the warning management module.

[0080] S5: Acquire the structural health status, drift compensation results, and structural data of the building structure; conduct a comprehensive analysis of the safety status of each structural unit, including mapping monitoring parameters to a unified structural response vector, establishing a structural state expression matrix and performing normalization transformation, and then performing state potential energy conversion to obtain the structural potential energy value; next, identify the spatial coupling effect of risk, calculate the spatial transmission of structural potential energy through the structural spatial connection matrix to obtain the structural risk propagation value, then perform time trend conversion on the risk propagation value to obtain the risk evolution rate, and integrate the risk propagation value and the risk evolution rate to obtain the comprehensive structural risk index; based on the comparison between the comprehensive risk index and the preset risk threshold, classify each structural unit into a building structure. The system identifies safe points, potential risk structural points, high-risk structural points, and emergency hazardous structural points. Based on these structural safety points, a safe passage network is established within the building. The optimal evacuation route is calculated using a path risk cost function, which comprehensively considers path distance, node risk index, and risk evolution rate. When the building is impacted or experiences structural anomalies and the warning level reaches the trigger condition, an early warning mechanism is activated. The management terminal displays a risk distribution map of the building's internal structure and the optimal safe path. Evacuation instructions are issued to personnel inside the building via a broadcast system, mobile terminal, or building display terminal. Based on the personnel's real-time location, the system matches the nearest safe node, dynamically generates individualized evacuation paths, and guides personnel to evacuate in an orderly manner along these safe paths.

[0081] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. An online monitoring and early warning management system for the health status of urban building structures, comprising a building structure acquisition module, a building data analysis module, a structural analysis module, and an early warning management module; characterized in that: The building structure acquisition module collects urban building data, which includes building structure data, environmental data, and operational status data, and forms time-series urban building data through multi-source sensors. The building data analysis module performs joint analysis of current urban building data and historical urban building data to obtain data drift points, and calculates the drift difference based on the data drift points to obtain the corresponding drift difference value. The structural analysis module obtains the drift difference and performs a comprehensive analysis with the corresponding building structure data to obtain drift compensation, and monitors the building health status to obtain the health status of the building structure. A comprehensive analysis of building operation risks is conducted to obtain building structural risk values; The building structure risk value is compared with a preset risk threshold to obtain the warning level; The early warning management module performs safety early warning and emergency management of building structures based on the early warning judgment results. It conducts a comprehensive analysis of the safety status of each building structure to obtain the safe points, potential risk points, risk points, and high-risk points of the building structure. Based on this, it establishes an internal safe passage network, calculates the optimal evacuation route, and generates a safe path.

2. The online monitoring and early warning management system for the health status of urban building structures according to claim 1, characterized in that, The specific process of jointly analyzing current urban building data and historical urban building data is as follows: A unified data representation is used for time-series urban building data to establish a multi-dimensional structural response state vector; information entropy weight allocation is set to establish a weighted structural state sequence; discrete Fourier transform is performed on the structural state sequence to obtain the frequency domain response function, and the frequency domain energy distribution function is calculated. A structural state network is established using complex network modeling methods. The network centrality index and its change are calculated to determine whether the structural state has shifted. After detecting structural state shifts, the historical structural state sequence is trained using a long short-term memory network prediction model to obtain a prediction function. The prediction residuals are calculated and the cumulative deviation is calculated to determine candidate data drift points. Dynamic time-warped distance calculation is performed on the structural state sequence before and after the drift candidate point to verify and determine the real data drift point.

3. The online monitoring and early warning management system for the health status of urban building structures according to claim 2, characterized in that, The process by which the building data analysis module calculates the drift difference based on the identified data drift points is as follows: A data evolution analysis sequence is established with the data drift point as the center. The sequence morphology is rearranged to obtain the sequence morphology distribution vector. The sequence complexity index is calculated to characterize the degree of change in the arrangement structure of the data sequence near the drift point. The probability distribution function of the sequence before and after the drift point is established, and the degree of difference between the two sets of data distributions is calculated through the distribution deviation function. The data change intensity index is calculated, and the data drift difference is calculated by integrating the sequence complexity change, distribution deviation degree and change intensity.

4. The online monitoring and early warning management system for the health status of urban building structures according to claim 3, characterized in that, The specific process of obtaining the drift difference and performing a comprehensive analysis with the corresponding building structure data is as follows: Obtain building structure status data and corresponding drift differences at different times; Each state parameter is adjusted by the drift compensation coefficient to obtain the compensated structural state vector; The compensated structural state vector is input into the isolated forest anomaly detection model to calculate the path length and average path length of the sample in the random isolated tree, and anomaly index is obtained. The health status of a building structure is determined by comparing abnormal indicators with health status assessment thresholds.

5. The online monitoring and early warning management system for the health status of urban building structures according to claim 4, characterized in that, The specific process for conducting a comprehensive analysis of building operation risks is as follows: Obtain the current building structure health status level parameters and the corresponding structural status data set; extract the time dimension trend of the structural status data, and obtain the change trend coefficient through linear regression model fitting; The risk value of the building structure is calculated by a risk assessment function, which combines health status level parameters, change trend coefficients, and drift differences. The risk value is compared with the preset risk threshold to determine the warning level, and the warning information is generated and sent to the warning management module.

6. The online monitoring and early warning management system for the health status of urban building structures according to claim 5, characterized in that, The specific process for conducting a comprehensive safety status analysis of each building structure is as follows: Receive the building structure health status, drift compensation results and building structure data, establish a structural state expression matrix and perform normalization transformation; Structural potential energy conversion is performed to obtain structural potential energy values, spatial coupling effects of risks are identified, a structural spatial connection matrix is ​​established, and risk propagation calculations are performed. The risk propagation value is converted to a time trend, the risk evolution rate is calculated, and the combined risk index is obtained. Based on the comprehensive risk index, the internal structural units of the building are classified and identified to obtain the building structural safety points, potential risk structural points, risk structural points and high-risk structural points. Establish a safe passage network within the building based on safe points, calculate the optimal evacuation route, generate safe routes, and display the safe routes and issue evacuation instructions through the management terminal when an early warning is triggered.

7. The online monitoring and early warning management system for the health status of urban building structures according to claim 6, characterized in that, After generating a safe path, the early warning management module will generate a safe path if it detects that the optimal evacuation path does not contain any potential risk structure points, risk structure points, or high-risk structure points. Otherwise, structural repairs should be carried out on the risk points in the path, and the safe route should be revised.

8. A method for online monitoring, early warning, and management of the structural health status of urban buildings, characterized in that, The method for implementing the online monitoring and early warning management system for the health status of urban building structures as described in any one of claims 1-7 includes: S1: Continuously collect building structure data, environmental data and operational status data of the target building through multi-source sensors to form time-series urban building data; S2: The building data analysis module performs joint analysis on current urban building data and historical urban building data, identifies data change trends, obtains data drift points, and calculates the drift difference based on the identified data drift points to obtain the corresponding drift difference. S3: Obtain the drift difference and corresponding building structure data through the structural analysis module, perform drift compensation and building health status monitoring, and obtain the health status of the building structure; S4: After obtaining the health status of the building structure, the structural analysis module performs a comprehensive analysis of the building operation risk to obtain the building structure risk value. The risk value is compared with the preset risk threshold, and the warning level is obtained based on the comparison result. When the warning level reaches the preset trigger condition, the warning information is sent to the warning management module. S5: The early warning management module performs safety early warning and emergency management of building structures based on the early warning judgment results.

Images