Efficient analysis method for the security of a BIM model assembly

Abstract

This invention belongs to the field of building safety data analysis technology, and provides an efficient method for analyzing the safety of components in a BIM model. This invention analyzes the impact of each associated component on the stability of the target component by combining the correlation and positional relationship of monitoring data between the target component and its associated components. Furthermore, it segments the monitoring data and analyzes the differences between the detection data of the target component and its associated components within corresponding time periods, evaluating the safety characteristic expression factor of the target component in each time period. Based on the safety characteristic expression factor, it obtains time periods with low safety information density, thereby obtaining predictive data for the target component and providing safety early warnings. This invention enables dynamic analysis of the safety of target components, improves the efficiency of BIM component safety analysis, and increases the accuracy of predictive data output, thus ensuring construction safety.

Description

Technical Field

[0001] This invention relates to the field of path planning data analysis technology, and in particular to an efficient method for analyzing the safety of components in a BIM model. Background Technology

[0002] Components refer to the parts of a building structure, such as beams, columns, slabs, supports, hoisting components, and connectors. They play a decisive role in the stability and safety of the structure during construction and operation. As the scale and complexity of building projects continue to increase, the number of structural components becomes larger, their types more complex, and the construction environment more variable. Especially in the construction of high-rise buildings, long-span bridges, and complex structures, the construction quality and stress state of the components directly affect the safety and stability of the overall structure.

[0003] BIM (Building Information Modeling) is a comprehensive information platform that integrates geometric information, physical information, construction progress, cost, operation and maintenance data of a building throughout its entire life cycle, based on a digital three-dimensional model. Currently, BIM technology is widely used both domestically and internationally to assist in building design, construction and operation management.

[0004] However, existing BIM applications are mostly focused on geometric modeling, schedule control, and clash detection, with limited support for real-time safety analysis and dynamic early warning of components. During construction, installation deviations, abnormal stress, and the reliability of temporary supports of components are closely related to construction safety. However, traditional safety assessment methods rely on manual inspection, single-point monitoring, or static analysis, which are inefficient, cannot reflect the actual situation on site in real time, and have lag and blind spots. Summary of the Invention

[0005] To address the above technical problems, this invention provides an efficient method for analyzing the safety of components in a BIM model.

[0006] According to the present invention, an efficient method for analyzing the safety of components in a BIM model is provided, the method comprising:

[0007] Based on BIM models and sensors, obtain location information and monitoring data of components;

[0008] Based on the location information, obtain the associated components of the target component;

[0009] The correlation and positional relationship of monitoring data between the target component and its associated components are analyzed to obtain the stability influence factor of each associated component on the target component.

[0010] The monitoring data of the target component is analyzed and segmented to obtain several time series segments.

[0011] The differences between the detection data of the target component and its associated components in the corresponding time period are analyzed to obtain the data fluctuation transmission coefficient of the target component and its associated components in the corresponding time period;

[0012] By analyzing the correlation between the stability influence factor and the data fluctuation transmission coefficient, the safety characteristic expression factor of the target component in each time period is obtained;

[0013] Based on the security feature expression factor, a time series with low security information density is obtained, and then the prediction data of the target component is obtained for security early warning.

[0014] In some embodiments of the present invention, the correlation and positional relationship of monitoring data between the target component and its associated components are analyzed to obtain the stability influence factor of each associated component on the target component, including:

[0015] The correlation between the monitoring data of the target component and its associated components is analyzed to obtain the direct influence coefficient of the stability of the associated components on the target component.

[0016] Calculate the distance between the centroid of the target component and the centroid of its associated component to obtain the first positional relationship between the target component and its associated component;

[0017] Obtain the construction location, and based on the construction location, obtain the construction-affecting components among the associated components of the target component;

[0018] Analyze the second positional relationship between the target component and the components that affect construction;

[0019] By combining the stability direct influence coefficient, the first positional relationship, and the second positional relationship, the stability influence factor of each associated component on the target component is obtained.

[0020] In some embodiments of the present invention, based on the construction location, the construction-affecting components among the associated components of the target component are obtained, including:

[0021] Connect the center of the construction location with the centroid of the target component to obtain the first construction influence line segment;

[0022] Draw a plane perpendicular to the target construction influence line segment through the two endpoints of the target construction influence line segment, and take the associated component between the two planes as the construction influence component of the target component.

[0023] In some embodiments of the present invention, analyzing the second positional relationship between the target component and the construction-affecting component includes:

[0024] Connect the center of the construction location with the centroid of the construction-affected component to obtain the second construction-affected line segment;

[0025] Calculate the distance between the centroid of each of the construction-affected components and the center of the construction location, and calculate the cosine of the angle between the first construction-affected line segment and each of the second construction-affected line segments, and the second positional relationship between the target component and the construction-affected components.

[0026] In some embodiments of the present invention, the changes in monitoring data of the target component are analyzed, and the monitoring data is segmented to obtain several time segments, including:

[0027] Set a short-time window length, perform a short-time Fourier transform on the monitoring data of the target component, decompose the monitoring data in each short-time window into several frequency components, and obtain the relative energy of each frequency component;

[0028] The differences in the relative energy of the frequency components corresponding to the current short-term window and the adjacent short-term windows are analyzed, as well as the differences in the phase values ​​of the frequency components corresponding to all short-term windows and the adjacent short-term windows, to obtain the segment feature values ​​of the monitoring data influencing factors for each short-term window.

[0029] Based on the characteristic values ​​of the segment points of the influencing factors of the monitoring data, the monitoring data is segmented to obtain several time segments of the monitoring data of the target component.

[0030] In some embodiments of the present invention, the differences between the detection data of the target component and its associated components within a corresponding time period are analyzed to obtain the data fluctuation transmission coefficient of the target component and its associated components within the corresponding time period, including:

[0031] Based on the results of the short-time Fourier transform, the high-frequency bands within all short-time windows in each time segment corresponding to the target component are obtained, thereby obtaining the short-time fluctuation coefficient of each short-time window;

[0032] Based on the short-time fluctuation coefficient, the short-time fluctuation data segment within each time period is obtained;

[0033] Analyze the temporal characteristics of short-time fluctuation data segments within the time period of the target component and its associated components to obtain the corresponding associated short-time fluctuation data segments in the corresponding time period of the associated components for each short-time fluctuation data segment within the time period of the target component.

[0034] The difference between the relative energy magnitudes of all short-term fluctuation data segments within each time period of the target component and the corresponding short-term fluctuation data segments in the corresponding time period of the associated component is analyzed to obtain the data fluctuation transmission coefficient between each time period of the target component and the corresponding time period of its associated component.

[0035] In some embodiments of the present invention, the correlation between the stability influence factor and the data fluctuation transmission coefficient is analyzed to obtain the safety characteristic expression factor of the target component in each time series, including:

[0036] The stability influence factors corresponding to each associated component of the target component are sorted in descending order to obtain the stability influence factor sequence.

[0037] Based on the order of the stable influencing factor sequence, the data fluctuation transmission coefficient corresponding to each associated component is obtained, thus obtaining the data fluctuation transmission coefficient sequence;

[0038] By analyzing the correlation between the stability influence factor sequence and the data fluctuation transmission coefficient sequence, the safety characteristic expression factor of the target component in each time period is obtained.

[0039] In some embodiments of the present invention, based on the security feature expression factor, a time series segment with low security information density is obtained, thereby obtaining predictive data for the target component and performing security early warning, including:

[0040] Based on the security feature expression factor, K-means is used to cluster each time segment to obtain several time segment clusters.

[0041] Calculate the mean of all security feature expression factors in each time segment cluster, and take the time segment in the time segment cluster with the smallest mean as the time segment with low security information density;

[0042] The monitoring data of the low security information density time series is input into the trained LSTM model, and the predicted data of the target component is output.

[0043] When the predicted data contains data exceeding the warning value of the target component, a safety risk warning is issued in the BIM model.

[0044] In some embodiments of the present invention, the location information includes spatial coordinates and associated nodes; the monitoring data includes force, displacement, deformation, vibration and environmental load.

[0045] In some embodiments of the present invention, the associated component includes other components that have the same associated node as the target component or have contact surfaces that coincide in spatial coordinate position.

[0046] As can be seen from the above embodiments, the efficient analysis method for the safety of components in a BIM model provided by the embodiments of the present invention has the following beneficial effects:

[0047] This invention utilizes sensors deployed in building components during construction to collect relevant data in real time. By combining the correlation and positional relationships of monitoring data between the target component and its associated components, a stability impact factor for each associated component on the target component is obtained. The changes in the monitoring data of the target component are analyzed, and the data is segmented. Based on the differences in the detection data between the target component and its associated components within corresponding time periods, a data fluctuation transmission coefficient is obtained for each time period. The correlation between the stability impact factor and the data fluctuation transmission coefficient is analyzed to evaluate the safety characteristic expression factor of the target component in each time period. Based on the safety characteristic expression factor, time periods with low safety information density are identified, thereby obtaining predictive data for the target component and providing safety warnings. This invention, based on monitoring data and positional information of different components in a BIM model, achieves dynamic analysis of the safety of the target component by analyzing the positional relationships and monitoring data characteristic relationships between the target component and its associated components. This improves the efficiency of BIM component safety analysis and increases the accuracy of predictive data output, thus ensuring construction safety.

[0048] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0049] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A schematic diagram of the basic process of an efficient analysis method for the safety of components in a BIM model provided in an embodiment of the present invention;

[0051] Figure 2 A schematic diagram of the basic process of a method for analyzing the stability influencing factors of a target component provided in an embodiment of the present invention. Detailed Implementation

[0052] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an efficient analysis method for the safety of components in a BIM model based on the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0053] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Terms such as “comprising,” “including,” or any other variations thereof are intended to cover a non-exclusive inclusion, such that a circuit structure, article, or device comprising a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such article or device. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of additional identical elements in the article or device that includes the element.

[0054] This invention addresses the following scenario: BIM (Building Information Modeling) is a comprehensive information platform based on a digital three-dimensional model, integrating data across the entire lifecycle of a building, including geometric information, physical information, construction progress, cost, and operation and maintenance. Components refer to the parts of a building structure, such as beams, columns, slabs, supports, hoisting components, and connectors, collectively known as components. They play a decisive role in structural stability and safety during construction and operation. With the increasing demands for precision and safety in deep foundation pits and large-span structures, and based on the concept of preventing risks from the design stage, aiming to reduce construction rework and operation and maintenance accidents, an efficient method for analyzing the safety of construction components is needed. Therefore, the main objective of this invention is to integrate the entire lifecycle data of components using BIM technology, combined with automated tools and real-time monitoring, to solve the problems of data fragmentation and insufficient dynamic monitoring in existing technologies, ultimately improving the efficiency and accuracy of safety analysis.

[0055] The following section, in conjunction with the accompanying drawings, will provide a detailed description of an efficient method for analyzing the safety of components in a BIM model, as provided in this embodiment.

[0056] Please see Figure 1 This illustrates the basic flow of an efficient analysis method for the safety of components in a BIM model, provided by an embodiment of the present invention.

[0057] like Figure 1As shown in the figure, an efficient method for analyzing the safety of components in a BIM model, provided by one embodiment of the present invention, specifically includes the following steps:

[0058] S100: Based on BIM models and sensors, acquire location information and monitoring data of components.

[0059] By deploying various types of sensors on-site, key parameters such as stress, displacement, deformation, vibration, and environmental loads of components are monitored in real time, obtaining monitoring data for the components. This data includes stress, displacement, deformation, vibration, and environmental loads. Furthermore, the location information of each component is extracted from the component attribute table of the BIM model, including spatial coordinates (X, Y, Z) and associated nodes (connection point numbers). The collected monitoring data is dynamically linked with the BIM model, providing a data foundation for dynamic analysis and early warning of the component's safety status. Specific monitoring indicators, sensor types, and installation locations are shown in the table below:

[0060]

[0061]

[0062] Specifically, based on the construction schedule and the location numbers of the components within the BIM model, a monitoring point layout diagram is created. The spatial coordinates, monitoring number, and corresponding component ID of each sensor are pre-bound to the BIM model attributes. On-site installation is completed according to the layout diagram, and signal stability is tested. It is crucial to ensure a one-to-one correspondence between the sensor locations in the BIM model and the spatial coordinates of the physical sensors.

[0063] Each sensor records sensor data at a preset frequency of 1Hz. Each record includes the sensor number, component ID (BIM model GUID), timestamp, and the actual monitoring data obtained by the sensor. Real-time sensor data is transmitted to the BIM platform database via wireless communication technology (such as WLAN), and the sensor data undergoes noise reduction and cleaning processing.

[0064] At this point, the location information and monitoring data of the components were obtained.

[0065] In a BIM model, components are connected by nodes. Loads are transferred from top to bottom, from beams to columns, and from walls to the ground. The safety of an individual component is directly affected by adjacent and connected components. Furthermore, neighboring components are influenced by construction sequence, stress state, and deformation trends. If certain critical components exceed their limits and become unstable, it will affect the stability of adjacent components. Therefore, it is necessary to assess the stability impact of each component in the BIM model on other components and improve the accuracy of analyzing the potential impact of other components on the target component. This specifically includes steps S200 and S300.

[0066] S200: Based on the location information, obtain the associated components of the target component.

[0067] Associated components that are in contact with or connected to the target component are the main source of influence on its stability changes. Although these associated components present differentiated monitoring data (such as stress, deformation, etc.) due to their different structural functions, they all reflect changes in their physical or mechanical properties and act on the target component through mechanical connection paths, thereby affecting its stability changes.

[0068] Therefore, in the embodiments of the present invention, the associated components of the target component are first obtained based on the location information. Specifically, based on the spatial coordinates (X, Y, Z) and associated nodes (connection point numbers) of each component in the BIM component attribute table, other components that have the same associated nodes as the target component or have contact surfaces with overlapping spatial coordinate positions are selected as the associated components of the target component.

[0069] S300: Analyze the correlation and location relationship of monitoring data between the target component and its associated components to obtain the stability influence factor of each associated component on the target component.

[0070] The correlation between the monitoring data of the associated component and the monitoring data of the target component is one of the important factors affecting the stability of the target component: the closer the monitoring data changes of the associated component and the target component are, the stronger the impact on the stability of the target component.

[0071] Meanwhile, the influence of associated components on the stability of the target component is also related to the distance between the centroids of the associated components and the target component: the distance between the centroids of the components is an important factor affecting the transmission of mechanical effects. The closer the distance, the more significant the impact of changes in the monitoring values ​​of the associated components on the mechanical effects and stability of the target component. In addition, the influence of the current construction location also needs to be considered: during construction, the current construction location has a significant impact on the stability of other components, especially associated components that have connections or load transfer paths with the target component. Due to the additional loads and deformation amplification effects, they often have a greater impact on the stability of the target component.

[0072] Based on the above analysis, in the embodiments of the present invention, by analyzing the correlation and positional relationship of monitoring data between the target component and its associated components, the stability influence factor of each associated component on the target component is obtained.

[0073] Please see Figure 2 This illustrates the basic flow of a method for analyzing the stability influencing factors of a target component, provided by an embodiment of the present invention.

[0074] like Figure 2 As shown, an embodiment of the present invention provides a method for analyzing the stability influencing factors of a target component, comprising:

[0075] S301: Analyze the correlation between monitoring data of the target component and its associated components to obtain the direct influence coefficient of the associated components on the stability of the target component.

[0076] The correlation between monitoring data of the target component and its associated components is analyzed to obtain the direct influence coefficient of the associated components on the stability of the target component. Specifically, the absolute value of the correlation coefficient between the monitoring data obtained by each sensor on each associated component of the target component and the monitoring data obtained by each sensor on the target component is calculated. That is, there are multiple correlation coefficients (number of monitoring indicators of the target component) between the monitoring data of a sensor on an associated component (e.g., force data) and the monitoring data of the target component. The maximum value of the absolute values ​​of the multiple correlation coefficients between the monitoring data of each sensor on the associated component and the monitoring data of the target component is selected, and the mean of the maximum absolute values ​​of the correlation coefficients corresponding to all sensors of the associated component is calculated as the direct influence coefficient of the associated component on the stability of the target component.

[0077] S302: Calculate the distance between the centroid of the target component and the centroid of its associated component to obtain the first positional relationship between the target component and its associated component.

[0078] Calculate the distance between the centroid of the target component and the centroid of its associated components to obtain the first positional relationship between the target component and its associated components. Specifically, use the API of BIM software (such as Revit API, Grasshopper) to develop a plugin, call the geometric core function to obtain the centroid coordinates of each component; calculate the Euclidean distance between the centroid of the target component and the centroid of each associated component, and normalize it, which is recorded as the first positional relationship between the target component and its associated components.

[0079] S303: Obtain the construction location, and based on the construction location, obtain the construction-affected components among the associated components of the target component.

[0080] Obtain the construction location, and based on the construction location, identify the construction-affecting components among the related components of the target component. Specifically, obtain the construction location according to the construction plan. Connect the center of the construction location with the centroid of the target component to obtain the first construction influence line segment. Draw a plane perpendicular to the target construction influence line segment through its two endpoints. The related components between the two planes (including those on the plane) are considered as the construction-affecting components of the target component; that is, the construction-affecting components belong to the related components.

[0081] S304: Analyze the second positional relationship between the target component and the components that affect construction.

[0082] The analysis focuses on the second positional relationship between the target component and the components affecting construction. Specifically, the center of the construction location is connected to the centroid of the components affecting construction to obtain the second construction influence line segment. The Euclidean distance between the centroid of each component affecting construction and the center of the construction location is calculated and normalized to obtain the normalized distance between the component affecting construction and the construction location. Furthermore, the cosine value of the angle θ (ranging from [0, π)) between the first construction influence line segment and each second construction influence line segment is calculated and expressed as cosθ. By eliminating the influence of negative numbers and setting the value range to [0,1], the angular relationship between the target component and the component affected by construction is obtained; by combining the normalized distance and the angular relationship, the second positional relationship between the target component and the component affected by construction is obtained.

[0083] S305: Combining the stability direct influence coefficient, the first positional relationship, and the second positional relationship, the stability influence factor of each associated component on the target component is obtained.

[0084] Combining the stability direct influence coefficient, the first positional relationship, and the second positional relationship, the stability influence factor of each associated component on the target component is obtained. Specifically, the formula for calculating the stability influence factor of each associated component in constructing the target component on the target component is as follows:

[0085]

[0086] In the formula, S i D represents the stability influence factor of the i-th associated component on the target component; i Indicates the first positional relationship between the target component and the i-th associated component; C ci This represents the angular relationship between the target component and the c-th construction-affected component (the i-th associated component); r i D represents the direct influence coefficient of the i-th associated component on the stability of the target component; ci This represents the normalized distance between the c-th construction-affected component and the construction location; This represents the second positional relationship between the target component and the c-th construction-affecting component (the i-th associated component); adding 0.1 is to prevent the denominator from being 0. It should be noted that if the i-th associated component is not a construction-affecting component, then D... ci and C ci Set to 1.

[0087] r iThe larger the value, the greater the direct influence coefficient of the associated component on the stability of the target component, indicating that the monitoring data changes of the associated component and the target component are more similar, and the stronger the influence on the stability of the target component; D i The smaller the value, the closer the Euclidean distance between the center of mass of the target component and the center of mass of the associated component, indicating that the change in the monitoring value of the associated component has a more significant impact on the mechanical effect and stability of the target component; C ci This indicates the relationship between the direction of influence of the construction location on the affected component and the direction of influence of the construction location on the target component. The smaller the angle between the construction influence segments, i.e., the larger the value of the construction influence component, and the closer the direction of influence of the construction location on the affected component is to the direction of influence of the construction location on the target component, the more significant the mechanical effect transmitted from the construction location to the target component. (D) ci The smaller the value, the closer the construction-affected component is to the construction location, indicating that the construction location has a significant impact on the stability of the construction-affected component. In particular, related components that have connections or load transfer paths with the target component often have a greater impact on the stability of the target component due to the additional load and deformation amplification effect. The larger the value of S, the greater the degree of influence of the related component on the stability of the target component.

[0088] Thus, the stable influence factor of each associated component on the target component was obtained.

[0089] The main purpose of safety analysis for each component is to identify and mitigate potential safety risks during construction or use. Therefore, it is necessary to analyze the changing trends of monitoring data for each dimension of the component and identify any abnormal trends to issue early safety warnings. Currently, the industry primarily uses machine learning methods to predict the changing trends of structural monitoring data, especially for real-time monitoring data of components. Utilizing Long Short-Term Memory (LSTM) network models to model the nonlinear characteristics and long-term dependencies in time-series data typically yields relatively ideal and stable prediction results.

[0090] However, the monitoring data of components may be affected by factors irrelevant to safety, such as non-structural vibrations during construction (e.g., short-term fluctuations in vibration peaks from pedestrians, vehicles, and small machinery, which are insufficient to affect safety) and meteorological factors (e.g., strong winds, thunderstorms, and rainfall, which cause short-term vibration anomalies or minor load changes, and generally pose no safety risk if they do not exceed the specified limits). These factors can cause fluctuations in sensor monitoring data, but do not necessarily indicate a safety problem with the components. When using an LSTM model for prediction, these fluctuations may increase the convergence difficulty of the LSTM and affect the accuracy of the prediction. Therefore, it is necessary to analyze the factors affecting the fluctuations in monitoring data to eliminate factors irrelevant to safety. This specifically includes steps S400 to S600.

[0091] S400: Analyze the changes in monitoring data of the target components, segment the monitoring data, and obtain several time series segments.

[0092] To assess the characteristics of data changes caused by factors unrelated to security, it is necessary to analyze the characteristics of data changes in local time series. The data should be divided into multiple time series according to the differences in influencing factors to achieve homogeneous data classification and ensure the consistency and effectiveness of the analysis.

[0093] During periods when the same influencing factor dominates, the patterns of change in monitoring data usually exhibit similarity or homogeneity. However, when the influencing factor changes, the data change patterns often undergo abrupt changes. These abrupt changes are mainly manifested in the significant differences in spectral characteristics between the abrupt change period and adjacent periods, such as changes in energy distribution at different frequencies and differences in phase within the same frequency band. These differences in spectral characteristics can effectively reflect the effects of different influencing factors on the monitoring data of components. Therefore, based on abrupt change pattern analysis and external event marking, the monitoring data time series can be divided into several time series segments.

[0094] Based on the above analysis, in this embodiment of the invention, by analyzing the changes in the monitoring data of the target component, the monitoring data is segmented to obtain several time series segments. Further, this includes:

[0095] First, set the short-time window length (which can be 5 minutes), perform a short-time Fourier transform (STFT) on the monitoring data of the target component, and decompose the monitoring data in each short-time window into several frequency components; normalize the energy of each frequency component in each short-time window to obtain the relative energy of each frequency component.

[0096] Then, the degree of difference in the relative energy magnitude of the frequency components corresponding to the current short-term window and adjacent short-term windows is analyzed, as well as the degree of difference in the phase values ​​of the corresponding frequency components of all short-term windows and adjacent short-term windows, to obtain the segment feature values ​​of the monitoring data influencing factors for each short-term window. Specifically, the mean square error of the relative energy magnitude of the frequency components of the current short-term window and adjacent short-term windows, and the mean square error of the phase values ​​of the corresponding frequency components of all short-term windows and adjacent short-term windows are calculated to obtain the segment feature values ​​of the monitoring data influencing factors for each short-term window:

[0097] p = MSE e ×MSE p

[0098] In the formula, p represents the characteristic value of the segment point of the monitoring data influencing factors for each short-time window; MSE eThis indicates the degree of difference in the relative energy magnitude of the frequency components corresponding to the current short-time window and the adjacent short-time windows; that is, the mean square error (MSE) of the relative energy magnitude of the frequency components between the current short-time window and the adjacent short-time windows. p It represents the degree of difference between the phase values ​​of the corresponding frequency components of all short-time windows and adjacent short-time windows, that is, the mean square error of the phase values ​​of the corresponding frequency components of all short-time windows and adjacent short-time windows.

[0099] Finally, based on the characteristic values ​​of the influencing factors in the monitoring data, the monitoring data is segmented to obtain several time segments of the monitoring data for the target component. Specifically, the characteristic values ​​p of the influencing factors in each short-term window are sorted in descending order, and the difference between adjacent elements in the sort is calculated. The two elements with the largest difference are selected, and the short-term windows where the characteristic value p of the influencing factors in the monitoring data is greater than or equal to the larger element are taken as time segment intervals, thus dividing the monitoring data into several time segments. Each time segment includes the previous time segment interval but does not include the next time segment interval. That is, p can divide each short-term window into two time segment intervals. Thus, a time segment includes the time segment interval of the later segment within the previous short-term window and the time segment interval of the earlier segment within the current short-term window, or a time segment includes the time segment interval of the later segment within the current short-term window and the time segment interval of the earlier segment within the next short-term window.

[0100] It should be noted that, at the same time, the monitoring data of each associated component is divided into time segments corresponding to each time segment of the target component, and these time segments are denoted as the corresponding time segments of the associated component and the target component; that is, the corresponding time segments of the associated component and the time segments of the target component have a one-to-one time correspondence.

[0101] S500: Analyze the differences between the detection data of the target component and its associated components in the corresponding time period to obtain the data fluctuation transmission coefficient of the target component and its associated components in the corresponding time period.

[0102] For monitoring data of components, especially long-term monitoring data obtained under complex operating conditions and multi-factor interference scenarios, properly screening and processing unimportant or low-value data segments is a core prerequisite for improving the efficiency and accuracy of LSTM prediction. After dividing the monitoring data into several time series segments, it is necessary to analyze the level of safety feature expression in different time series segments of the target component by combining the changes in the monitoring data of related components. The level of safety feature expression refers to the density and intensity of important safety-related data features in the monitoring data of the target component within a certain time series segment.

[0103] To reduce the interference of time series segments with low levels of safety characteristic expression on subsequent predictions, it is necessary to analyze the variation characteristics of monitoring data for each time series segment. Among these, data changes caused by factors irrelevant to safety are typically short-term fluctuations, such as forces transmitted to components by construction workers or vehicles passing by on nearby roads, resulting in short-term changes in monitoring data. These fluctuations manifest as short-term fluctuations within the time series segments.

[0104] Based on the above analysis, in an embodiment of the present invention, by analyzing the differences between the detection data of the target component and its associated components within the corresponding time period, the data fluctuation transmission coefficient of the target component and its associated components within the corresponding time period is obtained. Further, it includes:

[0105] First, based on the results of the short-time Fourier transform (STFT), the high-frequency bands within all short-time windows of each time segment corresponding to the target component are obtained, thus yielding the short-time fluctuation coefficient for each short-time window. Specifically, the median of the frequencies of all sinusoidal decomposition components in the STFT results within each short-time window of each time segment is selected, and decomposition components with frequencies greater than this median are defined as the high-frequency bands within the current short-time window. Using the normalized amplitude values ​​of the high-frequency bands within the short-time window as weights, the frequency-weighted average of the high-frequency bands within each short-time window is calculated, and this frequency weight is used as the short-time fluctuation coefficient for the current short-time window.

[0106] Then, based on the short-time fluctuation coefficients, short-time fluctuation data segments within each time period are obtained. Specifically, the short-time fluctuation coefficients in each short-time window are arranged in ascending order to obtain a short-time fluctuation coefficient sequence. The difference between adjacent elements in the sequence is calculated, and the two elements with the largest difference are selected. The short-time window with a short-time fluctuation coefficient greater than or equal to the larger element is defined as the short-time fluctuation data segment of the current time period.

[0107] Next, the temporal characteristics of short-term fluctuation data segments within the time series of the target component and its associated components are analyzed to obtain the corresponding associated short-term fluctuation data segments in the corresponding time series of its associated components for each short-term fluctuation data segment within the time series of the target component. Specifically, the short-term fluctuation data segment in the corresponding time series of each associated component that is temporally closest to the short-term fluctuation data segment within the time series of the target component is selected without repetition, and this segment is taken as the corresponding associated short-term fluctuation data segment in the corresponding time series of its associated components.

[0108] Finally, the differences in relative energy magnitudes between the detected data of all short-term fluctuation data segments within each time segment of the target component and the corresponding related short-term fluctuation data segments within the corresponding time segment of the associated component are analyzed to obtain the data fluctuation transmission coefficient between each time segment of the target component and the corresponding time segment of its associated component. Specifically, the differences in relative energy magnitudes (normalized result of the total energy of all frequency components within the short-term fluctuation data segments within each time segment of the target component) are analyzed between the detected data of the corresponding related short-term fluctuation data segments within the corresponding time segment of the associated component and the corresponding time segment of the associated component are obtained to obtain the data fluctuation transmission coefficient between each time segment of the target component and the corresponding time segment of its associated component. Specifically, the formula for calculating the data fluctuation transmission coefficient between the f-th time segment of the target component and its associated component within the corresponding time segment is as follows:

[0109]

[0110] In the formula, This represents the data fluctuation transmission coefficient between the target component and its associated components in the corresponding time segment of the f-th time segment; m represents the number of short-term fluctuation data segments with corresponding associated short-term fluctuation data segments in the target component's f-th time segment (since the number of short-term fluctuation data segments in the target component's f-th time segment may be greater than the number of short-term fluctuation data segments in the corresponding time segment of the associated components, the short-term fluctuation data segments in the target component's f-th time segment may not have corresponding associated short-term fluctuation data segments). E represents the relative energy magnitude of the detected data of the corresponding associated short-time fluctuation data segment within the f-th time segment of the target component; j The value represents the relative energy of the j-th short-term fluctuation data segment within the f-th time segment of the target component; sig represents the sigmoid function.

[0111] The larger the value, the greater the data fluctuation energy of the associated component, and the more it has transmission characteristics.

[0112] Similarly, the data fluctuation transmission coefficients of each time segment of the target component and all associated components within the corresponding time segment are obtained.

[0113] S600: Analyze the correlation between the stability impact factor and the data fluctuation transmission coefficient to obtain the safety characteristic expression factor of the target component in each time series.

[0114] During construction, such as during large-scale hoisting installation or dismantling, the transmitted force in the construction area will also cause short-term changes in the monitoring data of the target components. However, this change is usually more drastic compared to factors unrelated to safety, and it is related to the transmission direction in the construction area. Due to the influence of force attenuation during the transmission process, the data fluctuations of related components with larger stability factors among the target components are greater than those of the target components at the same time. Such short-term data fluctuations have a greater impact on the safety of the components.

[0115] Based on the above analysis, in the embodiments of the present invention, the safety characteristic expression factor of the target component in each time period is obtained by analyzing the correlation between the stability influence factor and the data fluctuation transmission coefficient. Specifically, the stability influence factors corresponding to each associated component of the target component are sorted in descending order to obtain a stability influence factor sequence; according to the order of the stability influence factor sequence, the data fluctuation transmission coefficient corresponding to each associated component and each time period is obtained to obtain a data fluctuation transmission coefficient sequence for each time period; the correlation between the stability influence factor sequence and the data fluctuation transmission coefficient sequence is analyzed to obtain the safety characteristic expression factor of the target component in each time period. The larger the safety characteristic expression factor value, the more the current short-term fluctuation has the characteristics of construction force transmission between components, and the higher the safety information density of the short-term fluctuation.

[0116] S700: Based on the security feature expression factor, it obtains time series segments with low security information density, and then obtains predictive data of target components to provide security warnings.

[0117] The security feature information density of monitoring data is evaluated by security feature expression factor. In order to avoid prediction lag or inaccuracy caused by LSTM remembering irrelevant details, it is necessary to smooth out the local redundancy or low information density areas in the data and compress redundant information according to the security feature expression factor.

[0118] Therefore, firstly, based on the security feature expression factors, time series with low security information density are obtained. Specifically, based on the security feature expression factors, K-means is used to cluster each time series, and the clustering parameter K is obtained through the elbow method, resulting in K time series clusters. The mean of all security feature expression factors in each time series cluster is calculated, and the time series in the cluster with the smallest mean is taken as the low security information density time series. Gaussian filtering is used to smooth the monitoring data of the low security information density time series, where the Gaussian filtering parameter is the reciprocal of the security feature expression factor of the corresponding time series, resulting in preprocessed monitoring data of the low security information density time series. By preprocessing the monitoring data of components, LSTM can focus more on learning the core trends and periodic patterns in the data. In addition, non-stationary data (such as trend abrupt changes or sequences with unstable variance) will increase the convergence difficulty of LSTM, while stabilization preprocessing can shorten the training time and reduce gradient oscillations.

[0119] Then, the monitoring data for the low security information density time series is input into the trained LSTM model, which outputs prediction data for the target components. Specifically, the preprocessed low security information density monitoring data for the target components and environmental monitoring data from the past 8 hours are input into the trained LSTM model, which outputs prediction data for the target components for the next 10 minutes. The LSTM model training process is as follows:

[0120] A large number of historical monitoring data samples were obtained. Each sample contained the data sequence of structural components and environmental monitoring data in the past 8 hours (480 minutes) and the corresponding monitoring value for the next 10 minutes as a label. Time series samples were constructed using the sliding window method, with each window spanning 480 minutes to predict the next 10 minutes. The sliding step size was 1 minute, forming a large number of continuous samples.

[0121] The network structure is as follows:

[0122] Input layer: 480 minutes × n_features

[0123] LSTM hidden layers:

[0124] First layer: Units 64-128, activation function tanh;

[0125] Dropout layer: 0.2-0.5 to prevent overfitting;

[0126] Fully connected layer (Dense): Outputs predicted values ​​for the next 10 minutes × n_targets;

[0127] Activation function: Linear (regression task).

[0128] The training data is divided into three sets according to time sequence: training set: 70%, validation set: 15%, and test set: 15%. The loss function is defined as mean squared error. The Adam optimizer, batch size is set to 32, and the number of epochs is set to 100. The training is stopped when the MSE of the validation set does not decrease for 10 consecutive epochs, thus completing the model training.

[0129] Finally, when the predicted data contains data exceeding the warning value of the target component, a safety risk warning is issued in the BIM model. Specifically, the predicted data of the target component is displayed in the BIM model in real time, facilitating visualization, dynamic monitoring, and risk warning; when the predicted data contains data exceeding the warning value of the corresponding component, a safety risk warning is issued in the BIM model, realizing the safety analysis of the component.

[0130] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0131] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for efficient analysis of the security of a BIM model assembly, characterized in that, The method includes: Based on BIM models and sensors, obtain location information and monitoring data of components; Based on the location information, obtain the associated components of the target component; The correlation and positional relationship of monitoring data between the target component and its associated components are analyzed to obtain the stability influence factor of each associated component on the target component. The monitoring data of the target component is analyzed and segmented to obtain several time segments; The differences between the detection data of the target component and its associated components in the corresponding time period are analyzed to obtain the data fluctuation transmission coefficient of the target component and its associated components in the corresponding time period; By analyzing the correlation between the stability influence factor and the data fluctuation transmission coefficient, the safety characteristic expression factor of the target component in each time period is obtained; Based on the security feature expression factor, a time series with low security information density is obtained, and then the prediction data of the target component is obtained for security early warning. The method for obtaining the stability impact factor is as follows: analyze the correlation between the monitoring data of the target component and its associated components to obtain the direct impact coefficient of the associated components on the stability of the target component; Calculate the distance between the centroid of the target component and the centroid of its associated component to obtain the first positional relationship between the target component and its associated component; Connect the center of the construction location with the centroid of the target component to obtain the first construction influence line segment; draw a plane perpendicular to the first construction influence line segment through the two endpoints of the first construction influence line segment, and take the associated component between the two planes as the construction influence component of the target component; Connect the center of the construction location with the centroid of the construction-affected component to obtain a second construction-affected line segment; calculate the distance between the centroid of each construction-affected component and the center of the construction location, and calculate the cosine of the angle between the first construction-affected line segment and each second construction-affected line segment, and the second positional relationship between the target component and the construction-affected component; Combining the stability direct influence coefficient, the first positional relationship, and the second positional relationship, the stability influence factor of each associated component on the target component is obtained; the formula for calculating the stability influence factor is: ;in, The first component representing the target component The stability influence factors of each related component on the target component; Indicates the target component and the first The first positional relationship between the related components; Indicates the target component and the first The construction affects the angular relationship between components; Indicates the first The coefficient of direct influence of each associated component on the stability of the target component; Indicates the first Normalized distance between construction-affected components and construction location.

2. The efficient analysis method for the safety of components in a BIM model according to claim 1, characterized in that, Analyze the changes in the monitoring data of the target component, and segment the monitoring data to obtain several time series segments, including: A short-time window length is set, and a short-time Fourier transform is performed on the monitoring data of the target component. The monitoring data within each short-time window is decomposed into several frequency components, and the relative energy of each frequency component is obtained. The energy of each frequency component within each short-time window is normalized to obtain the relative energy of each frequency component. The differences in the relative energy of the frequency components corresponding to the current short-term window and the adjacent short-term windows are analyzed, as well as the differences in the phase values ​​of the frequency components corresponding to all short-term windows and the adjacent short-term windows, to obtain the segment feature values ​​of the monitoring data influencing factors for each short-term window. Based on the characteristic values ​​of the segment points of the influencing factors of the monitoring data, the monitoring data is segmented to obtain several time segments of the monitoring data of the target component.

3. The efficient analysis method for the safety of components in a BIM model according to claim 2, characterized in that, Analyzing the differences between the detection data of the target component and its associated components within the corresponding time period, the data fluctuation transmission coefficient between the target component and its associated components within the corresponding time period is obtained, including: Based on the results of the short-time Fourier transform, the high-frequency bands within all short-time windows in each time segment corresponding to the target component are obtained, thereby obtaining the short-time fluctuation coefficient of each short-time window; Based on the short-time fluctuation coefficient, the short-time fluctuation data segment within each time period is obtained; Analyze the temporal characteristics of short-time fluctuation data segments within the time period of the target component and its associated components to obtain the corresponding associated short-time fluctuation data segments in the corresponding time period of the associated components for each short-time fluctuation data segment within the time period of the target component. The difference between the relative energy magnitudes of all short-term fluctuation data segments within each time period of the target component and the corresponding short-term fluctuation data segments in the corresponding time period of the associated component is analyzed to obtain the data fluctuation transmission coefficient between each time period of the target component and the corresponding time period of its associated component.

4. The efficient analysis method for the safety of components in a BIM model according to claim 1, characterized in that, Analyzing the correlation between the stability impact factor and the data fluctuation transmission coefficient yields the safety characteristic expression factors of the target component in each time series, including: The stability influence factors corresponding to each associated component of the target component are sorted in descending order to obtain the stability influence factor sequence. Based on the order of the stable influencing factor sequence, the data fluctuation transmission coefficient corresponding to each associated component is obtained, thus obtaining the data fluctuation transmission coefficient sequence; By analyzing the correlation between the stability influence factor sequence and the data fluctuation transmission coefficient sequence, the safety characteristic expression factor of the target component in each time period is obtained.

5. The efficient analysis method for the safety of components in a BIM model according to claim 1, characterized in that, Based on the security feature expression factor, low security information density time series are obtained, and then predictive data for the target component is obtained for security early warning, including: Based on the security feature expression factor, K-means is used to cluster each time segment to obtain several time segment clusters. Calculate the mean of all security feature expression factors in each time segment cluster, and take the time segment in the time segment cluster with the smallest mean as the time segment with low security information density; The monitoring data of the low security information density time series is input into the trained LSTM model, and the predicted data of the target component is output. When the predicted data contains data exceeding the warning value of the target component, a safety risk warning is issued in the BIM model.

6. The efficient analysis method for the safety of components in a BIM model according to claim 1, characterized in that, The location information includes spatial coordinates and associated nodes; the monitoring data includes force, displacement, deformation, vibration and environmental load.

7. The efficient analysis method for the safety of components in a BIM model according to claim 6, characterized in that, The associated components include other components that have the same associated nodes as the target component or have contact surfaces that coincide with each other in spatial coordinates.

Images