Fabricated building detection method and system

Abstract

The invention relates to the technical field of building structure health monitoring, in particular to an assembly type building detection method and system, and the method comprises the steps: collecting a strong interference signal generated by an external impact source of a building, and obtaining the sound source positioning information of the strong interference signal, converging the sound source positioning information of the strong interference signal on a preset three-dimensional geometric model aligned with a coordinate system to generate a real-time interference source distribution map; in the conventional detection mode, after a to-be-distinguished target acoustic signal is detected, obtaining a theoretical acoustic signal arrival time sequence based on sound source positioning information of at least one strong interference signal in the real-time interference source distribution diagram; comparing the actual acoustic signal arrival time sequence of the target acoustic signal with the theoretical acoustic signal arrival time sequence to obtain a comparison result; and judging whether the target acoustic signal is damaged in the building or not according to a comparison result. Real damage events from the interior of the structure can be effectively separated and confirmed, and the accuracy and reliability of detection are improved.

Description

Technical Field

[0001] The present application relates to the technical field of building structure health monitoring, and more specifically, to a prefabricated building detection method and system. Background Art

[0002] Large transportation hubs built using existing prefabricated construction techniques typically feature a steel-concrete composite box girder structure. This structure leverages the tensile strength of steel and the compressive strength of concrete, but also completely encases the most critical mechanical connections—the welds and bolts at the steel structure joints—in thick concrete. After the hubs are put into operation, they are subjected to daily cyclical vibration loads. Under this long-term, high-frequency fatigue load, initial fatigue microcracks may initiate at the weld roots of the steel box girder joints or in the threads of the high-strength bolts, encased within the concrete. Each tiny expansion of these microcracks releases strain energy locally and instantaneously, generating a high-frequency, short-lived stress wave, or acoustic emission signal. This signals structural damage. From the outside, the concrete surface may appear intact, making this incipient damage hidden deep within invisible to conventional visual or displacement detection methods. To capture these weak signals, existing technologies typically deploy an acoustic emission monitoring system on the structure. Sensors are densely arranged on the concrete outer surface of the composite box girder, especially in the areas close to the internal steel structure nodes.

[0003] When the external environment changes, such as an unexpected downpour accompanied by strong winds, high-speed raindrops, driven by the wind, strike the windward curtain wall and roof of a building obliquely with enormous kinetic energy. Each raindrop impact represents a point-like, instantaneous release of energy, generating impact sound waves with a rapid rise and short duration. This is virtually indistinguishable in the time domain from typical acoustic emissions from material cracking. Consequently, in prefabricated steel-concrete composite structures, the impact acoustic signals generated by wind and rain impacting external protective structures (such as glass curtain walls), with their clear propagation paths and high energy, overlap significantly in signal morphology and timing with the weaker acoustic emissions generated by fatigue crack initiation within the steel components, which must travel across the steel-concrete interface and suffer severe attenuation and distortion. Consequently, how to prevent standard sound source localization methods from being misled by strong interference signals and effectively separate and identify the true damage event originating within the structure from a data stream overwhelmed by these strong interference signals remains an urgent challenge. Summary of the Invention

[0004] In order to address the deficiencies of the prior art, the present application provides a method and system for detecting prefabricated buildings, which can effectively separate and confirm real damage events originating from inside the structure, thereby improving the accuracy and reliability of detection.

[0005] The present application provides a method for detecting prefabricated buildings, the method comprising: Collect strong interference signals generated by external impact sources of the building, obtain the sound source positioning information of the strong interference signals, converge the sound source positioning information of the strong interference signals on a preset three-dimensional geometric model aligned with the coordinate system, and generate a real-time interference source distribution map; In the conventional detection mode, after the target acoustic signal to be distinguished is detected, the theoretical acoustic signal arrival time sequence is obtained based on the sound source positioning information of at least one strong interference signal in the real-time interference source distribution map; Compare the actual acoustic signal arrival timing of the target acoustic signal with the theoretical acoustic signal arrival timing to obtain a comparison result; Based on the comparison results, determine whether the target acoustic signal indicates damage inside the building.

[0006] The above scheme allows to effectively distinguish internal damage signals from external strong interference signals by utilizing the spatial and temporal characteristics of known external interference, thereby improving the accuracy of internal damage detection in complex environments.

[0007] To further solve the problem, this application also proposes to collect strong interference signals generated by external impact sources of the building, obtain sound source positioning information of the strong interference signals, converge the sound source positioning information of the strong interference signals on a preset three-dimensional geometric model aligned with the coordinate system, and generate a real-time interference source distribution map, including: Monitor the impact frequency of high-energy impact events from external impact sources of buildings; When the impact frequency exceeds the preset frequency threshold, the conventional detection mode is switched to the reverse discrimination mode to collect strong interference signals generated by impact sources outside the building, obtain the sound source positioning information of the strong interference signals, and converge the sound source positioning information of the strong interference signals on the preset three-dimensional geometric model aligned with the coordinate system to generate a real-time interference source distribution map.

[0008] Through the above scheme, the collection process of strong interference signals is optimized. By switching modes based on the impact frequency, it is ensured that interference data is collected only when the external impact is significant, thereby improving the efficiency and relevance of data collection.

[0009] To further solve the problem, the present application further proposes that, in a conventional detection mode, after detecting the target acoustic signal to be distinguished, the following steps are also included: Perform sound source localization calculation based on the target acoustic signal to obtain a positioning result containing discrete coordinate points; Extract the spatial distribution characteristics of the positioning results to determine whether the target acoustic signal is generated by a linear interference source. The spatial distribution characteristics include the center of mass position, the direction of the main axis, and the ratio of the major and minor axes. If the target acoustic signal is generated by a linear interference source, part of the target acoustic signal generated by the linear interference source is eliminated.

[0010] Through the above solution, the differentiation process is further refined. By identifying and eliminating the signals generated by linear interference sources, the misjudgment of specific types of external noise is effectively avoided.

[0011] To improve the solution, this application also proposes to extract the spatial distribution characteristics of the positioning results and determine whether the target acoustic signal is generated by a linear interference source, including: Extract the spatial distribution characteristics of the positioning results. When the spatial distribution characteristics of the positioning results show multi-center characteristics, cluster the discrete coordinate points of the positioning results into multiple coordinate point subsets; The spatial distribution features of each subset of coordinate points are extracted and compared independently with the linear interference source; If the spatial distribution characteristics of all coordinate point subsets successfully match the linear interference source, the corresponding target acoustic signal is determined to be a composite external interference signal generated by the linear interference source.

[0012] Through the above scheme, the ability to identify linear interference sources is enhanced. Through multi-center spatial distribution analysis and clustering processing, complex external interference signals can be identified more robustly and accurately.

[0013] To improve the solution, this application also proposes to determine whether the target acoustic signal is damage inside the building based on the comparison results, including: If the comparison result shows that the actual arrival timing of the target acoustic signal does not match the theoretical arrival timing of the target acoustic signal, the target acoustic signal is judged to be a suspected internal damage signal; Calculate the spatial distance between the discrete coordinate points obtained by the sound source localization calculation of the suspected internal damage signal and the geometric boundary of the three-dimensional geometric model; According to preset rules, the number of discrete coordinate points of suspected internal damage signals belonging to the three-dimensional geometric model is counted. If the attribution ratio exceeds a preset threshold, the suspected internal damage signal is determined to be an external interference signal.

[0014] Through the above scheme, a re-evaluation mechanism for suspected internal damage signals is provided, which effectively filters out external interference signals that may be initially misjudged by checking their spatial proximity to the building geometric boundaries.

[0015] To improve the solution, this application also proposes that the method further includes: Based on the acoustic propagation characteristics in building media, combined with the three-dimensional geometric model and sensor layout information, the propagation path deflection results of the strong interference signal at the multi-media interface are calculated; According to the propagation path deflection result, the spatial boundary of the three-dimensional geometric model is corrected to form a spatial offset area associated with the propagation path compensation; The number of discrete coordinate points of suspected internal damage signals belonging to the 3D geometric model is counted according to preset rules, including: The number of discrete coordinate points of suspected internal damage signals belonging to the spatial offset area is counted according to preset rules.

[0016] Through the above scheme, the acoustic propagation characteristics in complex media are taken into account. By calculating and compensating for propagation path deflections, the accuracy of spatial boundary definition in damage assessment is improved and false alarms are reduced.

[0017] To improve the solution, this application also proposes to calculate the propagation path deflection results of the strong interference signal at the multi-media interface based on the acoustic propagation characteristics of the building medium, combined with the three-dimensional geometric model and sensor layout information, and also includes: Get the current ambient temperature data of the building; Update the corresponding acoustic propagation characteristics based on the ambient temperature data and the material's sound velocity-temperature relationship; Based on the updated acoustic propagation characteristics combined with the three-dimensional geometric model and sensor layout information, the propagation path deflection results of strong interference signals at the multi-media interface are updated.

[0018] Through the above scheme, the accuracy of propagation path calculation is improved, and the acoustic propagation characteristics are dynamically updated by introducing real-time ambient temperature data, making the compensation more accurate and adaptive.

[0019] To improve the solution, this application also proposes to modify the spatial boundaries of the 3D geometric model based on the propagation path deflection results to form a spatial offset region associated with propagation path compensation, including: Identifying at least one area of ​​sudden deviation change according to a deviation vector change amount of the propagation path deflection result; According to the overlap between the three-dimensional geometric model and the sudden shift region, the three-dimensional geometric model is divided into a first entity located in the sudden shift region and a second entity located outside the sudden shift region; applying a first correction process to the first entity and applying a second correction process to the second entity, wherein the first correction process has a higher information density than the second correction process; All modified first entities are combined with the second entities to generate a spatial offset region.

[0020] Through the above scheme, the correction strategy of spatial boundaries is optimized. By identifying and performing differentiated and high-density corrections on areas with drastic offset changes, higher accuracy is ensured in key areas.

[0021] To improve the solution, this application also proposes identifying at least one area of ​​sudden deviation change based on the deviation vector change of the propagation path deflection result, including: According to the variation of the offset vector of the propagation path deflection result, a variation rate data set is formed; Partitioning the change rate data set according to the distribution relationship of the data points in the change rate data set to form a high change rate data subset and a low change rate data subset; The spatial range corresponding to the high-change rate data subset is determined as the area of ​​​​rapid offset change.

[0022] Through the above scheme, a systematic method for identifying areas with sudden offset changes is provided. By analyzing the rate of change of the offset vector, the three-dimensional model can be accurately segmented for differential correction.

[0023] To improve the solution, this application also proposes an assembled building detection system, which includes: The acquisition and positioning module is used to collect strong interference signals generated by external impact sources of the building, obtain the sound source positioning information of the strong interference signals, converge the sound source positioning information of the strong interference signals on a preset three-dimensional geometric model aligned with the coordinate system, and generate a real-time interference source distribution map; A detection calculation module is used to, in a conventional detection mode, after detecting the target acoustic signal to be distinguished, derive the theoretical acoustic signal arrival time sequence based on the sound source localization information of at least one strong interference signal in the real-time interference source distribution map; The timing comparison module is used to compare the actual arrival timing of the target acoustic signal with the theoretical arrival timing of the acoustic signal to obtain a comparison result; The damage judgment module is used to judge whether the target acoustic signal is damage inside the building based on the comparison results.

[0024] Through the above solution, a system for implementing the above detection method is provided, which realizes automated and integrated prefabricated building damage detection and improves detection efficiency and reliability.

[0025] In summary, the prefabricated building detection method and system provided in this application effectively solves the technical problem of accurately distinguishing internal damage signals from external interference signals in a complex external interference environment by introducing real-time monitoring and positioning of strong interference signals, time-series comparison of target acoustic signals, and spatial attribution judgment and propagation path compensation of suspected damage signals. It can effectively separate and confirm real damage events originating from inside the structure, thereby improving the accuracy and reliability of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A schematic flow chart of a method for inspecting prefabricated buildings according to one embodiment of the present application.

[0027] Figure 2 One of the flow charts of a prefabricated building inspection method provided in another embodiment of the present application.

[0028] Figure 3 The second flowchart of a prefabricated building inspection method provided in another embodiment of the present application.

[0029] Figure 4 The third flowchart of a prefabricated building detection method provided in another embodiment of the present application.

[0030] Figure 5 This is a fourth flow chart of a method for detecting prefabricated buildings provided in another embodiment of the present application.

[0031] Figure 6 This is a fifth flow chart of a method for detecting prefabricated buildings provided in another embodiment of the present application.

[0032] Figure 7 FIG6 is a flowchart of a method for detecting prefabricated buildings provided in another embodiment of the present application.

[0033] Figure 8 FIG7 is a flowchart of a method for detecting prefabricated buildings provided in another embodiment of the present application.

[0034] Figure 9 FIG8 is a flow chart of a method for detecting prefabricated buildings provided in another embodiment of the present application.

[0035] Figure 10 A flowchart of a prefabricated building inspection system provided in another embodiment of the present application.

[0036] Figure 11 A real-time interference source distribution diagram in a specific implementation of a prefabricated building detection method provided in another embodiment of the present application.

[0037] In the figure: 1. Acquisition and positioning module; 2. Detection and calculation module; 3. Timing comparison module; 4. Damage judgment module. DETAILED DESCRIPTION

[0038] The technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. The components of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0039] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0040] Traditional acoustic inspection methods for prefabricated buildings are prone to being misled by strong external interference signals when localizing sound sources for internal structural damage. These interference signals can have highly similar time-domain characteristics to internal damage signals, making it difficult for the inspection system to effectively distinguish them. Consequently, it is impossible to accurately isolate and confirm true damage events originating from within the structure from massive data streams. This directly impacts the accuracy and reliability of inspection results.

[0041] Reference Figure 1 , this application proposes a method for detecting prefabricated buildings, the method comprising: S1000: Collect strong interference signals generated by impact sources outside the building, obtain sound source positioning information of the strong interference signals, converge the sound source positioning information of the strong interference signals on a preset three-dimensional geometric model aligned with the coordinate system, and generate a real-time interference source distribution map; S2000: In a conventional detection mode, after detecting a target acoustic signal to be distinguished, deriving a theoretical acoustic signal arrival time sequence based on sound source localization information of at least one strong interference signal in a real-time interference source distribution map; S3000: comparing the actual arrival timing of the target acoustic signal with the theoretical arrival timing of the target acoustic signal to obtain a comparison result; S4000: Determine whether the target acoustic signal indicates damage inside the building based on the comparison result.

[0042] In this embodiment, a strong interference signal refers to an acoustic signal with a relatively clear propagation path, generated by external impact sources (such as wind, rain, and traffic vibration). This signal can be collected using an acoustic sensor array, such as a piezoelectric sensor or a fiber optic sensor. Sound source localization information refers to the location of the acoustic signal, calculated using acoustic signal processing techniques (such as time difference of arrival (TDOA) and direction of arrival (DOA). This information can be obtained using a multi-sensor array combined with algorithms (such as least squares and maximum likelihood estimation). This information is primarily used to determine the specific spatial source of the strong interference signal, providing foundational data for subsequent interference source modeling. The pre-set 3D geometric model is a digital 3D representation of the actual building structure, aligned with the coordinate system of the sensor collecting the strong interference signal. The model can be constructed using a BIM model, a CAD model, or laser scanning point cloud data. Its primary purpose is to map the sound source localization information of the strong interference signal to the actual building space, creating a visual representation of the interference source distribution and serving as the basis for subsequent calculations. Convergence means that the sound source data from different times and locations are uniformly placed on the spatial coordinates of the three-dimensional geometric model to form a clear and visual "interference source distribution map".

[0043] A real-time interference source distribution map is a dynamic visualization of the sound source location information of strong interference signals collected over a specific time period on a three-dimensional geometric model. This can be achieved using data visualization software or a custom graphical interface. Its primary purpose is to provide an overview of the spatial location and activity level of interference sources outside the building, assisting in identifying the source of the target acoustic signal. The theoretical acoustic signal arrival time sequence refers to the expected time sequence of acoustic signals propagating from the interference source to each sensor, calculated based on the known interference source location and sensor location, combined with the propagation speed of sound waves in the medium. This can be achieved using acoustic wave propagation models (such as ray tracing or finite element methods) combined with geometric distance calculations. Its primary purpose is to establish a benchmark for comparison with the actual arrival time sequence of the target acoustic signal, thereby identifying the signal source. The actual acoustic signal arrival time sequence refers to the sequence of signal arrival time points recorded by each sensor when the sensor array actually receives the target acoustic signal. This can be obtained using techniques such as high-precision timestamp recording and signal trigger detection. Its primary purpose is to provide the actual temporal characteristics of the target acoustic signal at the sensor end, serving as a basis for comparison with the theoretical time sequence.

[0044] This application continuously collects strong interference signals generated by impact sources outside the building, obtains their sound source positioning information, and maps it to a three-dimensional geometric model aligned with the building coordinate system to dynamically generate a real-time interference source distribution map. When the system receives a target acoustic signal to be distinguished in conventional detection mode, it does not directly determine its source. Instead, it calculates the theoretical arrival timing of each sensor based on the positioning information of at least one interference source in the distribution map, combined with the propagation characteristics of sound waves in the building medium. This theoretical timing is then compared with the actual arrival timing of the target signal. If the two match, it is determined to be an external interference signal; if there is a significant difference, it is determined to be a damage signal suspected to originate from inside the building. This method effectively improves the accuracy and reliability of identifying weak damage signals inside structures in complex environments through a reverse exclusion mechanism with clear logic and rigorous steps.

[0045] In some specific implementations of this embodiment, first, a high-sensitivity acoustic sensor array is deployed on the external surface of the prefabricated building, such as curtain walls, roofs and other areas. These sensors can be piezoelectric accelerometers or acoustic emission sensors. These sensors continuously monitor and collect strong interference signals generated by external impact sources such as wind, rain, and traffic noise. By processing these signals with a time difference localization (TDOA) algorithm, the sound source positioning information of each strong interference signal can be accurately obtained, such as its X, Y, and Z coordinates in three-dimensional space. At the same time, a BIM (Building Information Model) that accurately matches the building structure is pre-established as a three-dimensional geometric model, which contains the precise dimensions and material information of the building. The acquired sound source positioning information of the strong interference signal is mapped to this BIM model in real time and dynamically updated to generate a visual real-time interference source distribution map (refer to Figure 11), which displays the location, intensity, and temporal variation of interference sources. When the system receives any target acoustic signal to be distinguished in normal detection mode, such as an acoustic emission signal from a suspected crack, it first extracts the acoustic source location information of strong interference signals close in time to the current target signal from the real-time interference source distribution map. Next, using this interference source location information and the sound velocity parameters of building materials (such as steel and concrete), it calculates the theoretical time it would take for the target signal to propagate to each sensor if it were generated by these external interference sources. This results in the theoretical acoustic signal arrival time series. For example, a ray tracing algorithm is used to simulate the propagation path and time of sound waves in the building structure. The system then compares the actual arrival time series of the target acoustic signal received by the sensor with the previously calculated theoretical arrival time series (see Table 1 below). This comparison can be achieved by calculating the correlation or the root mean square error of the time difference between the two. If the match between the actual and theoretical time series falls below a preset threshold, a mismatch is detected. Finally, based on this mismatch comparison, the system determines that the target acoustic signal is not caused by a strong external interference source, but rather a possible damage signal within the building. This allows the system to effectively distinguish and identify internal damage signals that truly reflect the health of the structure, even in the presence of high external ambient noise.

[0046]

[0047] Table 1 Data Log Description External interference: The majority of high-amplitude events (>80dB) in the table have Y coordinates around 6000mm, accurately corresponding to simulated raindrop interference applied to the top of the box girder. The system successfully matched their actual arrival timing with the theoretical timing of the 3D geometric model, accurately identifying them as external interference.

[0048] Background noise: Events with amplitudes below 60 dB and discrete locations (e.g., EVT-1226, EVT-1238) were classified as background noise and excluded from the main analysis.

[0049] Core Verification Points (EVT-1234): Location: The Y coordinate of the event is 4500 mm, which is inside the structural web and does not match the location of any external interference source.

[0050] Judgment: After the system calculated its arrival timing, it was found that there was a significant difference exceeding the preset threshold from the theoretical timing calculated based on the top three-dimensional geometric model.

[0051] Result: Therefore, the system was not confused by the large number of surrounding, higher-energy interference signals and uniquely identified this event as "suspected internal damage."

[0052] Reference Figure 2 In another embodiment of the present application, it is further proposed that step S1000 includes: S1100: Monitors the impact frequency of high-energy impact events from external impact sources of buildings; S1200: When the impact frequency exceeds the preset frequency threshold, the conventional detection mode is switched to the reverse discrimination mode to collect strong interference signals generated by impact sources outside the building, obtain the sound source positioning information of the strong interference signals, and converge the sound source positioning information of the strong interference signals on a preset three-dimensional geometric model aligned with the coordinate system to generate a real-time interference source distribution map.

[0053] Among them, in this embodiment, the impact frequency of high-energy impact events refers to the density or repetition rate of acoustic impact signals generated by external impact sources of the building, such as wind and rain, vehicle traffic or construction activities, per unit time. It can use an acoustic sensor array to continuously collect environmental sound waves, and use signal processing algorithms such as fast Fourier transform (FFT) or short-time Fourier transform (STFT) to analyze the spectral characteristics of the signal, identify the characteristic frequency components of high-energy impact events, and count the number of occurrences of these events within a specific time window. The purpose is to quantify the activity level of external interference; the frequency threshold refers to a pre-set numerical limit used to determine whether the external impact frequency has reached the level that requires the activation of a specific processing mode. It can use empirical values, results based on statistical analysis of historical data, or through machine The purpose is to provide a judgment basis for mode switching; the conventional detection mode refers to the operating state in which the system, under normal operating conditions, focuses on and processes damage signals that may be generated inside the building, and at the same time performs basic collection and preliminary filtering of external interference signals. It can adopt continuous monitoring, low-power data acquisition and basic signal classification processing. Its purpose is to maintain efficient internal damage detection in a non-strong interference environment; the reverse discrimination mode refers to the operating state in which the system, under strong external interference conditions, focuses on identifying, locating and analyzing external strong interference signals in order to effectively distinguish them from internal damage signals. It can adopt high-sampling rate data acquisition, advanced sound source localization algorithms and real-time interference source modeling. Its purpose is to improve the recognition and utilization efficiency of interference signals in a strong interference environment.

[0054] The solution of the present application monitors the impact frequency of high-energy impact events generated by external impact sources of the building in real time, and compares the frequency with a preset threshold. When the impact frequency exceeds the threshold, it indicates that the external interference has significantly increased, and the system automatically switches from conventional detection mode to reverse discrimination mode. In reverse discrimination mode, the system actively focuses on the precise collection and positioning of strong interference signals, maps its sound source information to a three-dimensional geometric model aligned with the coordinate system, and generates a real-time interference source distribution map. This mechanism ensures that the system maintains conventional detection when external interference is weak, reducing resource consumption; and responds quickly when interference is strong, improving the accuracy of interference analysis. By conditionally switching the working mode, the system can dynamically adapt to changes in the external environment, avoid wasting resources due to invalid analysis, and at the same time improve the efficiency and accuracy of distinguishing internal damage signals in a high-interference background, significantly enhancing the practicality and robustness of the detection system.

[0055] In some specific implementations of this embodiment, the real-time and accuracy of strong interference identification are further improved by establishing a high-precision external acoustic monitoring and mode switching mechanism. The system deploys highly sensitive acoustic sensors at key locations on the building's exterior (such as curtain walls and roof edges) to continuously collect environmental acoustic data, which is then received in real time by a data processing unit. This unit is equipped with an impact frequency monitoring module that identifies high-energy impact events through short-term energy analysis or power spectrum density calculation. When the number of energy surges detected in a high-frequency band (e.g., above 5kHz) exceeds a preset frequency threshold (e.g., more than 10 times per second) per unit time, the system is triggered to switch from conventional detection mode to reverse discrimination mode. In reverse discrimination mode, the system increases the sampling rate and activates a more sophisticated sound source localization algorithm (such as TDOA least squares or a neural network model) to accurately obtain sound source localization information for strong interference signals. This information is then mapped to the BIM or CAD model in real time, dynamically updating the real-time interference source distribution map and providing high-precision spatial context support for subsequent signal discrimination.

[0056] Reference Figure 3 In another embodiment of the present application, it is further proposed that sub-step S2100 of step S2000: in the conventional detection mode, after detecting the target acoustic signal to be distinguished, further includes: S2200: Perform sound source localization calculation based on the target acoustic signal to obtain a positioning result including discrete coordinate points; S2300: Extracting spatial distribution features of the positioning results to determine whether the target acoustic signal is generated by a linear interference source. Spatial distribution features include centroid position, principal axis direction, and major-minor axis ratio. S2400: If the target acoustic signal is generated by a linear interference source, remove part of the target acoustic signal generated by the linear interference source.

[0057] In this embodiment, sound source localization calculation refers to determining the position of the sound source in three-dimensional space by analyzing information such as the arrival time difference or signal strength of the target acoustic signal at different sensors, combined with the geometric layout of the sensor array, and applying a specific algorithm. Its purpose is to convert the received acoustic signal into geometric information that can be used for spatial analysis; discrete coordinate points refer to a series of independent data points obtained by the sound source localization calculation that represent the possible position of the sound source in three-dimensional space. These points together constitute the distribution pattern of the sound source in space, and their purpose is to provide basic data for subsequent spatial feature extraction; linear interference sources refer to interference sources in building structures that can generate acoustic signals that are continuously or quasi-continuously distributed along a specific direction, such as wind and rain hitting long curtain wall keels, fluid flow in pipes, or friction of structural connectors. Its characteristic is that the acoustic signal it generates shows an obvious linear or band distribution in spatial positioning, and its purpose is to clearly identify and eliminate specific types of interference signals; spatial distribution characteristics refer to quantitative indicators used to describe the geometric shape of a set of discrete coordinate points in three-dimensional space, specifically including the center of mass position, main axis direction and major-minor axis ratio, among which the center of mass position refers to the geometric center of the discrete coordinate point set, reflecting the average spatial position of the sound source positioning result; the main axis direction refers to the direction in which the discrete coordinate point set extends longest in space, reflecting the dominant direction of the sound source distribution; the major-minor axis ratio refers to the ratio of the length of the discrete coordinate point set in the main axis direction to the maximum length perpendicular to the main axis direction, reflecting the flatness or linearity of the sound source distribution. Its purpose is to identify linear interference sources by quantitatively analyzing the spatial shape of the sound source.

[0058] This application improves the recognition accuracy of internal building damage signals by introducing spatial distribution feature analysis of target acoustic signals in conventional detection mode. After receiving the target acoustic signal, the system first locates the sound source, obtains a set of discrete coordinate points, and calculates its center of mass, main axis direction, and major-minor axis ratio to determine whether the signal has the spatial characteristics of a linear interference source. If it is confirmed to be from a linear interference source, it will be eliminated. This analysis based on spatial geometry complements the previous discrimination mechanism based on arrival time sequence comparison, significantly enhancing the ability to identify external interference, avoiding misjudgment, and improving overall detection reliability.

[0059] In some specific implementations of this embodiment, and in preferred embodiments, this application effectively identifies and eliminates target acoustic signals generated by linear interference sources by constructing a discrimination mechanism based on the spatial distribution characteristics of the sound source. After receiving the target acoustic signal in conventional detection mode, the system uses a multi-channel acoustic sensor array to collect data. Using a time difference of arrival (TDOA) localization algorithm (such as generalized cross-correlation combined with least squares), combined with sensor placement information, the system calculates the three-dimensional coordinates of the sound source, generating a set of discrete coordinate points. To extract the spatial distribution characteristics, the system performs principal component analysis on this point set to determine its principal axis direction, major-minor axis ratio, and center of mass position. These features are then compared with a pre-defined linear interference source feature model. If the major-minor axis ratio exceeds a set threshold and the principal axis direction matches a known linear structure, the signal is identified as a linear interference signal. The system then performs a rejection operation, removing such signals from the data stream or marking them as non-impairment signals, thereby preventing them from participating in subsequent time series comparison and impairment determination, effectively improving the accuracy and robustness of the detection process.

[0060] Reference Figure 4 In another embodiment of the present application, step S2300 includes: S2310: Extracting spatial distribution features of the positioning results. When the spatial distribution features of the positioning results exhibit multi-center features, clustering the discrete coordinate points of the positioning results into multiple coordinate point subsets. S2320: Extract spatial distribution features for each subset of coordinate points and compare them independently with linear interference sources; S2330: If the spatial distribution characteristics of all coordinate point subsets successfully match the linear interference source, the corresponding target acoustic signal is determined to be a composite external interference signal generated by the linear interference source.

[0061] In this embodiment, the multi-center feature refers to the fact that the discrete coordinate points of the positioning results show two or more obvious clustering areas or density peaks in space, and each clustering area can correspond to an independent linear interference source. Clustering is a data analysis technique used to divide the discrete coordinate points in a data set into several groups or clusters based on their similarity or proximity. It can be implemented using the K-means algorithm, DBSCAN algorithm, hierarchical clustering algorithm, etc. Its purpose is to separate the positioning results of multiple linear interference sources that are mixed together to form multiple independent coordinate point sets. The coordinate point subset refers to a set of one or more points with similar spatial distribution characteristics that are separated from the discrete coordinate points of the original positioning results after clustering. Each subset represents the positioning trajectory of a potential independent interference source. Independent comparison refers to the extraction of spatial distribution characteristics of each coordinate point subset separately, and comparing and matching them with the preset linear interference source feature model. Its purpose is to ensure that each potential linear interference source meets the linear characteristics, rather than misjudging multiple non-linear sources as composite linear sources. A composite external interference signal refers to a target acoustic signal generated by the combined action of two or more independent linear interference sources. These linear interference sources may be close to or overlap each other in space, resulting in a multi-center distribution in the positioning results.

[0062] This application effectively solves the problem that traditional methods are difficult to identify complex linear interference sources through refined cluster analysis of sound source positioning results. In conventional detection mode, the system performs sound source positioning on the detected target acoustic signal. If the positioning result shows a multi-center distribution, the system clusters the discrete coordinate points and divides them into multiple coordinate point subsets, each of which represents a potential linear interference source trajectory. The system extracts the spatial distribution characteristics of each subset, including the center of mass, main axis direction, and major-minor axis ratio, and independently compares them with the preset linear interference source feature model. Only when all subsets are matched successfully, the corresponding target acoustic signal is judged to be generated by a complex linear interference source. Compared with the traditional overall judgment method, this method is more targeted and robust. It can accurately identify complex interference signals in complex interference environments and avoid misjudging them as internal damage signals, effectively improving the accuracy and reliability of prefabricated building detection.

[0063] In some specific implementations of this embodiment, after the system receives a target acoustic signal and localizes the sound source, obtaining a set of discrete three-dimensional coordinate points, it first extracts the principal axis direction, major-minor axis ratio, and centroid through principal component analysis (PCA). The system then combines the covariance matrix with clustering validity indicators (such as the silhouette coefficient) to determine whether there are multi-center features. If there are multiple density peaks or the point set is not suitable for classification as a single cluster, the system uses the DBSCAN algorithm to cluster the coordinate points, automatically dividing them into multiple coordinate point subsets, each representing a potential linear interference source trajectory. Subsequently, the spatial distribution features of each coordinate point subset are extracted and independently compared with a preset linear interference source feature template. If all subsets meet the linear distribution characteristics, the system determines that the target acoustic signal is a composite external interference signal composed of multiple linear interference sources and performs a removal process. This method effectively improves the recognition accuracy of complex interference signals, preventing them from being mistakenly included in the subsequent damage assessment process.

[0064] Reference Figure 5 In another embodiment of the present application, step S4000 includes: S4100: If the comparison result shows that the actual arrival timing of the target acoustic signal does not match the theoretical arrival timing of the target acoustic signal, then the target acoustic signal is determined to be a suspected internal damage signal. S4200: Calculate the spatial distance between the discrete coordinate points obtained by performing sound source localization calculation on the suspected internal damage signal and the geometric boundary of the three-dimensional geometric model; S4500: Count the number of discrete coordinate points of the suspected internal damage signal belonging to the three-dimensional geometric model according to a preset rule. If the attribution ratio exceeds a preset threshold, the suspected internal damage signal is determined to be an external interference signal.

[0065] In this embodiment, a suspected internal damage signal refers to an acoustic signal initially determined to be likely originating from internal building damage. Its characteristic is that the arrival timing of the actual acoustic signal does not match the theoretical arrival timing of the acoustic signal. A geometric boundary refers to the spatial outline or surface of a building structure defined by a three-dimensional geometric model; it defines the physical extent of the building structure. Spatial distance calculation refers to a method for calculating the shortest distance between discrete coordinate points and the geometric boundary of a three-dimensional geometric model. This can be implemented using Euclidean distance, Manhattan distance, or distance calculation based on geometric projection. A preset rule refers to a pre-set logical condition or algorithm used in data statistics and judgment. This can be implemented using point attribution judgment rules based on distance thresholds, statistical rules based on region division, or rules based on machine learning classification. The attribution ratio refers to the ratio of discrete coordinate points that meet specific conditions (e.g., being inside or outside the geometric boundary of a three-dimensional geometric model) to the total number of discrete coordinate points during the statistical process. A preset threshold value refers to the numerical limit used to determine whether the attribution ratio meets specific conditions. When the attribution ratio exceeds or falls below this value, a specific judgment result is triggered.

[0066] This application improves the accuracy of damage identification by introducing a spatial information verification mechanism to conduct a secondary screening of signals that are initially judged to be suspected internal damage. When the actual acoustic signal arrival timing of the target acoustic signal does not match the theoretical acoustic signal arrival timing, the system preliminarily determines it as a suspected internal damage signal. Subsequently, the system locates the sound source of the signal, obtains its corresponding discrete coordinate points, and calculates the spatial distance between these coordinate points and the geometric boundary of the three-dimensional geometric model of the building. The system counts the number of discrete coordinate points belonging to the three-dimensional geometric model according to preset rules and calculates the attribution ratio. If the attribution ratio exceeds the preset threshold value, it means that most of the sound source points are located outside the structure, and even if the timing is abnormal, it is more likely to be an external interference signal. This method combines timing analysis and spatial position judgment, makes up for the shortcomings of single timing judgment, effectively avoids external complex propagation signals from being misjudged as internal damage, and enhances the reliability of prefabricated building damage detection.

[0067] In some specific implementations of this embodiment, when the system determines that the actual arrival timing of a target acoustic signal does not match the theoretical arrival timing of the target acoustic signal, it immediately marks it as a suspected internal damage signal. The system then performs sound source localization on these suspected internal damage signals. Specifically, a localization algorithm based on time difference of arrival (TDOA) can be used. This algorithm utilizes the arrival time differences of signals received by multiple acoustic sensors and iteratively solves the problem using the least squares or Gauss-Newton method. This algorithm obtains a series of discrete coordinate points in three-dimensional space, each representing the possible location of the sound source in (X, Y, Z) format. The system then calculates the spatial distance between these discrete coordinate points and the geometric boundary of a pre-imported 3D geometric model represented by a building information model (BIM). Specifically, a spatial index structure, such as a KD-tree, can be used to quickly find the shortest distance between each discrete coordinate point and the surface mesh of the BIM model. The system then counts the number of discrete coordinate points that are suspected internal damage signals belonging to the 3D geometric model according to pre-set rules. For example, a spatial distance threshold can be set. If the shortest distance from a discrete coordinate point to the geometric boundary is less than the threshold, the point is considered to belong to the interior of the 3D geometric model. The system will count how many discrete coordinate points of all suspected internal damage signals meet this attribution condition. Finally, the proportion of the number of discrete coordinate points belonging to the 3D geometric model to the total number of discrete coordinate points is calculated and compared with the preset threshold. If this attribution ratio exceeds the preset threshold, the system will determine that the suspected internal damage signal is actually an external interference signal and remove it from the list of internal damage signals.

[0068] Reference Figure 6 In another embodiment of the present application, it is further proposed that the method further includes: S4300: Based on the acoustic propagation characteristics of building media, combined with the three-dimensional geometric model and sensor layout information, calculates the propagation path deflection results of strong interference signals at multi-media interfaces; S4400: Correcting the spatial boundary of the three-dimensional geometric model based on the propagation path deflection result to form a spatial offset region associated with propagation path compensation; The sub-steps of S4500 are to count the number of discrete coordinate points of suspected internal damage signals belonging to the three-dimensional geometric model according to preset rules: S4510: to count the number of discrete coordinate points of suspected internal damage signals belonging to the spatial offset area according to preset rules.

[0069] Sub-step S4520 of S4500: If the attribution ratio exceeds a preset threshold value, the suspected internal damage signal is determined to be an external interference signal.

[0070] In this embodiment, acoustic propagation characteristics refer to the physical properties exhibited by sound waves when propagating through different building materials (such as concrete, steel, masonry, and air), including sound velocity, acoustic impedance, attenuation coefficient, scattering characteristics, and reflection and refraction patterns at interfaces between different media. Their purpose is to provide basic data and physical models for simulating the actual propagation paths of sound waves in complex building structures. These can be obtained through experimental measurements, material manuals, or numerical simulations. Combining a 3D geometric model with sensor placement information involves integrating a digital 3D representation of the building structure (including material distribution, geometry, and internal components) with the actual installation locations and quantity of acoustic sensors on the structure. This aims to construct a complete spatial information framework that reflects the sound wave propagation environment and monitoring points, facilitating sound wave propagation path analysis and sound source location. The 3D geometric model can be obtained through BIM model import, laser scanning data reconstruction, or manual modeling, while sensor placement information can be obtained through sensor installation records or GPS positioning. Calculating the propagation path deflection results of a strong interfering signal at a multi-media interface involves applying acoustic propagation theory (such as Snell's law and Huygens' principle) to simulate the specific conditions in which the propagation direction and path of sound waves generated by a strong interfering signal bend or change when passing through different building material interfaces (such as steel-concrete and concrete-air). This method aims to quantify the non-linear effects of sound wave propagation in complex media, thereby correcting for sound source localization errors caused by medium inhomogeneities. This can be achieved through numerical calculation methods such as ray tracing, the finite element method, or the boundary element method. The propagation path deflection results use the specific data of the calculated sound wave propagation path deflection (such as deflection angle, offset, propagation time difference, etc.) as input or reference. This method aims to provide a physical model-based basis for subsequent corrections to the three-dimensional geometric model. This method can take the form of a deflection vector field, a deflection angle matrix, or a propagation time correction table. Modifying the spatial boundary of a three-dimensional geometric model refers to geometrically adjusting or deforming the outer contour or boundary of the three-dimensional geometric model representing the internal area of ​​the building structure based on the deflection effect of the sound wave propagation path. Its purpose is to create a new spatial reference area that can reflect the actual propagation range of the sound wave and the judgment of the sound source, so as to compensate for the bending effect of sound wave propagation. This can be achieved through geometric operations such as boundary expansion, contraction, translation, deformation or generating a new envelope.The spatial offset region associated with propagation path compensation refers to a specific spatial range obtained after the spatial boundary of the three-dimensional geometric model is corrected, which has a certain offset or shape change compared to the original model and is used for sound source attribution judgment. This region is formed to compensate for the path deflection caused by medium changes during the propagation of sound waves, making it more accurate to judge the sound source attribution within this region. Its purpose is to provide a reference space that has been corrected for acoustic propagation characteristics and is more in line with the actual sound source position judgment logic, thereby improving the accuracy of the judgment of suspected internal damage signals. It can be an expanded or contracted geometric body, an envelope composed of multiple corrected surfaces, or a fuzzy boundary region generated based on the uncertainty region of sound wave propagation. Counting the number of discrete coordinate points of suspected internal damage signals belonging to a spatially offset region according to preset rules means, after obtaining a corrected spatially offset region, determining the spatial relationship between the discrete coordinate points of the suspected internal damage signal obtained by sound source localization calculation and the spatially offset region. The discrete coordinate points falling within the region are counted according to pre-set logic (for example, whether the point is within the region, whether the distance from the point to the region boundary is less than a threshold, etc.). This aims to utilize a more accurate spatial reference compensated for acoustic propagation characteristics to determine whether the suspected internal damage signal truly originates from within the building structure, thereby improving the reliability of the judgment. This can be achieved through algorithms such as point-in-polygon testing, point-to-surface distance calculation, or voxelized grid judgment.

[0071] This application improves the accuracy of identifying suspected internal damage signals by introducing a modeling and compensation mechanism for acoustic wave propagation path deflections. After initially identifying a target acoustic signal as a suspected internal damage signal, the system calculates the propagation path deflections of strong interfering signals at multi-media interfaces based on the acoustic propagation characteristics of the building medium, combined with a three-dimensional geometric model and sensor placement information. This calculation simulates the refraction and reflection behavior of sound waves propagating through different materials such as steel and concrete. Based on the calculation results, the spatial boundaries of the three-dimensional geometric model are modified to form a spatial offset region associated with propagation path compensation, making the model more closely match the actual propagation trajectory of the sound waves. Based on this, the system then calculates whether the discrete coordinate points of the suspected internal damage signal fall within this spatial offset region according to preset rules, thereby determining the sound source's location. Compared to traditional spatial distance calculation methods based on static geometric boundaries, this solution effectively avoids positioning errors caused by acoustic wave propagation deflections, improves the accuracy of identifying internal damage signals in complex structural environments, reduces the risk of misjudgments and missed detections, and enhances the system's robustness and practicality.

[0072] In some specific implementations of this example, assume that a suspected internal damage signal needs to be identified in a prefabricated steel-concrete composite box girder structure. First, the system uses acoustic propagation characteristic data for the steel, concrete, and their interfaces within the building structure. This data can be obtained in advance through experimental measurements or by consulting a material database. For example, the speed of sound in steel is approximately 5900 m / s, while in concrete it is approximately 3500 m / s, and there are differences in acoustic impedance at the interfaces. Simultaneously, the system combines a three-dimensional geometric model of the box girder structure (e.g., a BIM model derived from CAD software, containing geometric information such as the steel beam, concrete layer, and embedded components) with the coordinates of acoustic sensors (e.g., piezoelectric sensors) placed on the box girder surface. Using this information, a ray tracing algorithm can be used to simulate the path of a strong interference signal (e.g., sound waves generated by raindrops impacting a curtain wall) from its source to each sensor. During the simulation, when an acoustic wave encounters a steel-concrete interface, its refraction angle and energy attenuation are calculated according to Snell's law. This results in the deflection of the acoustic wave's propagation path at the multi-media interface. For example, a series of incident and refraction angles at different interfaces are obtained, along with the resulting spatial offset vector. Based on these calculated propagation path deflection results, the system can perform spatial boundary corrections on the 3D geometric model of the box girder structure. Specifically, if the calculation results indicate that the acoustic wave is deflected inward when passing through the concrete layer and entering the interior of the steel beam, to compensate for this deflection, the boundary representing the interior of the steel beam in the original 3D geometric model can be expanded outward, forming a spatial offset region associated with the propagation path compensation. For example, the outer surface of the original model can be normal-shifted, with the offset amount determined by the cumulative deflection of the acoustic wave along a specific path. This spatial offset region no longer represents the original precise geometric boundary, but rather an "effective" internal space that accounts for the "bending" effect of the acoustic wave. Subsequently, when a suspected internal damage signal is calculated by sound source localization to obtain a series of discrete coordinate points, the system will count the number of discrete coordinate points belonging to the newly formed spatial offset area according to the preset rules. For example, the rule can be set as follows: if a discrete coordinate point is located inside the spatial offset area, or its distance to the boundary of the spatial offset area is less than a preset minimum threshold, then the point is considered to belong to the area. By counting the proportion of discrete coordinate points falling within the spatial offset area, it is possible to more accurately determine whether the suspected internal damage signal is actually derived from damage inside the box girder structure, rather than an external interference signal. For example, if more than 80% of the discrete coordinate points fall within the corrected spatial offset area, the signal can be determined to be an internal damage signal. This method effectively avoids misjudgment caused by sound wave deflection and improves the reliability of damage detection.

[0073] Reference Figure 7In another embodiment of the present application, step S4300 further includes: S4310: Obtain the current ambient temperature data of the building; S4320: Update the corresponding acoustic propagation characteristics based on the ambient temperature data and the sound velocity-temperature relationship of the material; S4330: Based on the updated acoustic propagation characteristics combined with the three-dimensional geometric model and sensor layout information, update the propagation path deflection results of the strong interference signal at the multi-media interface.

[0074] In this embodiment, the material's sound velocity-temperature relationship refers to a function or empirical model that describes how the speed of sound waves in a specific material changes with temperature. This can be implemented using a pre-established lookup table, a polynomial fitting formula, or a theoretical calculation formula based on a physical model. Its purpose is to quantify the effect of temperature on the speed of sound and provide a basis for accurately updating acoustic propagation characteristics. The propagation path deflection result is a quantitative description of the change in direction of a sound wave due to differences in sound speed when crossing the interface of different acoustic media. In this embodiment, it can also be expressed in the form of a deflection angle, an offset vector, or a corrected propagation distance. Its purpose is to accurately reflect the actual propagation trajectory of the sound wave in a complex multi-media environment.

[0075] The solution of the present application optimizes the calculation process of the propagation path of strong interference signals by introducing considerations of ambient temperature. Specifically, first, the current ambient temperature data of the building is obtained, which provides real-time input for subsequent adjustments to the acoustic model. Secondly, based on the acquired ambient temperature data and the pre-established material sound speed-temperature relationship, the system can dynamically update the corresponding acoustic propagation characteristics, which means that parameters such as sound speed and attenuation are no longer fixed values, but are adjusted according to the actual temperature, so that the acoustic model is more in line with the actual operating status of the building. On this basis, using the updated acoustic propagation characteristics, combined with the three-dimensional geometric model of the building and the sensor layout information, the system recalculates and updates the propagation path deflection results of strong interference signals at the multi-media interface. It is precisely because the acoustic propagation characteristics are temperature compensated that the obtained propagation path deflection results can more accurately reflect the actual propagation trajectory of sound waves at different temperatures. This more accurate propagation path deflection result can provide a more reliable input for the subsequent correction of the spatial boundaries of the three-dimensional geometric model based on the propagation path deflection result, thereby forming a spatial offset area associated with the propagation path compensation. When judging whether the target acoustic signal is internal damage to the building, the number of discrete coordinate points of the suspected internal damage signal belonging to the spatial offset area can be more accurately counted, effectively avoiding the misjudgment of external interference signals as internal damage, thereby improving the accuracy and reliability of detection.

[0076] In some specific implementations of this embodiment, to obtain the current ambient temperature data of a building, multiple high-precision temperature sensors can be deployed at key locations inside and outside the building, such as within the concrete structure, on the surface of the steel structure, and in the ambient air. These sensors can periodically collect temperature readings and transmit the data to a central processing unit via a wired or wireless network. Upon receiving this ambient temperature data, the central processing unit can make real-time corrections to the sound speeds of the steel and concrete materials involved in the current building structure based on a pre-stored database of material sound speed-temperature relationships, which can contain experimental data or theoretical calculation curves for the sound speeds of different types of steel and concrete at different temperatures. For example, if the sound speed of a certain type of concrete is 3500 m / s at 20 degrees Celsius and its sound speed temperature coefficient is -2 m / s / degree Celsius, when the ambient temperature rises to 30 degrees Celsius, its sound speed will be updated to 3480 m / s. Once the acoustic propagation characteristics of the material, such as the speed of sound, are updated, the system can rerun the ray tracing algorithm or finite element analysis model. Using the updated acoustic parameters, combined with the building's 3D geometry and the specific placement of sensors within the building, it can recalculate the propagation path deflections of strong interference signals from the external impact source to the sensor at multiple interfaces, such as the steel-concrete interface and the concrete-air interface. This allows for more accurate propagation path deflection results. For example, the system can calculate the slightest change in the deflection angle of a sound wave due to temperature changes when passing through the interface between a steel beam and concrete, and accumulate this change in the final propagation path.

[0077] Reference Figure 8 In another embodiment of the present application, step S4400 includes: S4410: Identify at least one area where the deviation changes dramatically based on the deviation vector change of the propagation path deflection result; S4420: dividing the three-dimensional geometric model into a first entity located in the offset sudden change region and a second entity located outside the offset sudden change region according to the overlap between the three-dimensional geometric model and the offset sudden change region; S4430: Apply a first correction process to the first entity and a second correction process to the second entity, wherein the information density of the first correction process is higher than that of the second correction process; S4440: Combine all corrected first entities and second entities to generate a spatial offset region.

[0078] Among them, in this embodiment, the offset vector change refers to the spatial rate of change or gradient of the degree of offset of the sound wave propagation direction or position in the propagation path deflection result, which can be quantified by calculating the Euclidean distance, angle difference or a combination thereof of the offset vectors at adjacent spatial points, with the purpose of identifying areas where the deflection effect of the sound wave propagation path is significantly uneven. The offset sudden change area refers to the spatial range where the offset vector change reaches or exceeds the preset threshold, which can be determined by threshold segmentation or cluster analysis of the offset vector change, with the purpose of delineating key areas that require refined correction. The first entity refers to the part of the three-dimensional geometric model that spatially overlaps with the offset sudden change area, which can be determined by Boolean operations or spatial query algorithms, with the purpose of separating the area that requires high-precision correction from the overall model. The second entity refers to the part of the three-dimensional geometric model that is located outside the offset sudden change area, which can be determined by subtracting the first entity from the overall model, with the purpose of distinguishing the areas that can be relatively roughly corrected. The first correction process refers to a boundary correction method with high precision used for the first entity. It can be performed using methods such as local interpolation, nonlinear fitting, or finite element analysis. Its purpose is to ensure the correction accuracy in areas with severe deviations. The second correction process refers to a boundary correction method with relatively low precision used for the second entity. It can be performed using methods such as global translation, linear interpolation, or simplified geometric transformation. Its purpose is to ensure the overall correction efficiency while taking into account the correction of non-dramatic areas.

[0079] Higher information density means that the first correction processing is more intensive or complex than the second correction processing in terms of the number of data points, computational complexity or richness of model parameters used in the correction process. It can be reflected in a smaller correction step, finer grid division or more complex mathematical model. Its purpose is to achieve higher-precision correction in key areas.

[0080] The solution of the present application identifies at least one area of ​​sudden deviation based on the amount of change in the acoustic wave deviation vector in the propagation path deflection result, that is, a spatial segment where the acoustic wave propagation deflection is significant and the change is uneven. Subsequently, the model is divided into a first entity (located within the area of ​​sudden deviation) and a second entity (located outside the area of ​​sudden deviation) based on the spatial overlap relationship between the three-dimensional geometric model and the area of ​​sudden deviation. For the first entity, the system adopts a first correction process with high information density to provide a finer calculation granularity; while for the second entity, a second correction process with low information density is adopted to improve efficiency and reduce the calculation burden. Finally, by combining the corrected first entity and the second entity, a spatial deviation area corresponding to the propagation path compensation height is constructed. This solution takes into account both correction accuracy and calculation cost, solves the problems of low efficiency and unreasonable accuracy distribution caused by the unified correction of the overall model, and improves the spatial mapping accuracy in the damage signal judgment process.

[0081] In some specific implementations of this embodiment, after obtaining the propagation path deflection results of a strong interference signal at a multi-media interface, the system can first calculate the offset vector at each spatial point and further calculate the spatial rate of change of these offset vectors. For example, the Euclidean distance or direction cosine change of the offset vectors between adjacent grid points can be calculated to form a rate of change distribution graph. Then, by setting a rate of change threshold, for example, when the rate of change exceeds a preset value, the corresponding spatial region is identified as a region of abrupt offset change. These regions may correspond to locations in the structure where complex material interfaces, sudden geometric changes, or stress concentrations lead to drastic changes in sound velocity. Next, a spatial overlap analysis is performed between a preset 3D geometric model and the identified region of abrupt offset change. For example, using a CAD or BIM model, Boolean operations or spatial query algorithms can be used to define the portion of the 3D geometric model that overlaps the region of abrupt offset change as a first entity, while the remaining portion is defined as a second entity. The first entity can then be modified using a method based on local non-uniform rational B-spline (NURBS) surface fitting, or finite element analysis can be performed using smaller grid cells to capture subtle changes in its boundaries. For example, in areas where the offset changes dramatically, each vertex on the original boundary can be fine-tuned according to the trend of change in its local offset vector, and a new boundary can be generated using a high-order interpolation algorithm. For the second entity, a relatively simple correction method can be used, such as uniformly translating or scaling the entire area, or using linear interpolation to adjust its boundary to reduce computational complexity. Finally, the first entity and the second entity that have undergone different corrections are seamlessly combined to form a complete, corrected spatial offset area. This combination process ensures that the corrected boundaries remain continuous between different areas, thereby generating a spatial offset area that maintains high precision in key areas while taking into account computational efficiency overall.

[0082] Reference Figure 9 In another embodiment of the present application, step S4410 includes: S4411: generating a change rate data set based on the offset vector change of the propagation path deflection result; S4412: partitioning the change rate data set according to the distribution relationship of the data points in the change rate data set to form a high change rate data subset and a low change rate data subset; S4413: Determine the spatial range corresponding to the high-change rate data subset as the offset sudden change area.

[0083] In this embodiment, the distribution relationship of data points in the rate of change data set refers to the central tendency, degree of dispersion, peak values, valley values, and relative positions or densities of the data points in terms of numerical value, which can be characterized by statistical methods, cluster analysis, or density estimation. Partitioning processing refers to dividing a complete data set into two or more subsets with specific attributes according to preset rules or algorithms. Specifically, it can be done by setting dynamic thresholds, applying clustering algorithms, or dividing data points into different regions based on statistical quantiles. Its purpose is to effectively distinguish data points with different characteristics in the data set so that targeted processing can be performed.

[0084] The solution of the present application quantifies the change in the offset vector of the propagation path deflection result and converts it into an analyzable rate of change data set. Since sound waves will be deflected when propagating at the interface of different media, and this deflection is not uniformly distributed in space, by capturing the change in the offset vector, the degree of local distortion of the sound wave propagation path can be reflected. On this basis, the solution further partitions the data set according to the distribution relationship of the data points in the rate of change data set, thereby distinguishing the data points with significant changes from the data points with gentle changes, forming a high rate of change data subset and a low rate of change data subset. This partitioning process can realize the identification of the area where the sound wave propagation path is significantly deflected, avoiding indiscriminate correction of the entire three-dimensional geometric model. Ultimately, the spatial range corresponding to the high rate of change data subset is determined as the area of ​​​​drastic offset change.

[0085] In some specific implementations of this embodiment, to first form a rate of change dataset based on the change in the offset vector resulting from the propagation path deflection, the offset vector resulting from the propagation path deflection can be calculated for each discrete grid point or preset sampling point in the three-dimensional geometric model. Subsequently, the local variation in the offset vector can be quantified by calculating the change in the Euclidean distance or angle between adjacent grid points or sampling points. For example, the modulus of the offset vector difference between each point and its nearest neighbor can be calculated, or the gradient within a local region can be calculated. These changes can be aggregated to form a rate of change dataset.

[0086] Reference Figure 10 In another embodiment of the present application, a prefabricated building detection system is further proposed, the system comprising: The acquisition and positioning module 1 is used to collect strong interference signals generated by impact sources outside the building, obtain the sound source positioning information of the strong interference signals, converge the sound source positioning information of the strong interference signals on a preset three-dimensional geometric model aligned with the coordinate system, and generate a real-time interference source distribution map; Detection calculation module 2 is used to, in the conventional detection mode, after detecting the target acoustic signal to be distinguished, derive the theoretical acoustic signal arrival time sequence based on the sound source localization information of at least one strong interference signal in the real-time interference source distribution map; The timing comparison module 3 is used to compare the actual acoustic signal arrival timing of the target acoustic signal with the theoretical acoustic signal arrival timing to obtain a comparison result; The damage determination module 4 is used to determine whether the target acoustic signal is damage inside the building based on the comparison result.

[0087] Among them, the acquisition and positioning module refers to a functional unit used to collect acoustic signals and determine the location of their sources. It can be implemented by an acoustic sensor array combined with a sound source positioning algorithm. Its purpose is to obtain the sound source positioning information of strong interference signals and provide basic data for subsequent signal identification; wherein, the detection and calculation module refers to a functional unit that derives the theoretical signal arrival timing based on known interference source information under a specific detection mode. It can be implemented by a signal processing unit combined with a preset acoustic propagation model. Its purpose is to provide a theoretical basis for distinguishing between real damage signals and interference signals; wherein, the timing comparison module refers to a functional unit used to compare the difference in arrival timing between actual signals and theoretical signals. It can be implemented by a data processing unit executing a timing matching algorithm. Its purpose is to determine whether the target acoustic signal is consistent with the external interference signal; wherein, the damage judgment module refers to a functional unit that ultimately determines the damage inside the building based on the signal timing comparison results. It can be implemented by a decision logic unit combined with preset judgment rules. Its purpose is to comprehensively consider the interference source information and signal timing information and ultimately make a judgment on the damage inside the building.

[0088] The above description is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, various modifications and variations of the present application are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for detecting assembled buildings, characterized in that: The method comprises: Collect strong interference signals generated by impact sources outside the building, obtain sound source positioning information of the strong interference signals, converge the sound source positioning information of the strong interference signals on a preset three-dimensional geometric model aligned with the coordinate system, and generate a real-time interference source distribution map; In the conventional detection mode, after detecting the target acoustic signal to be distinguished, the theoretical acoustic signal arrival time sequence is obtained based on the sound source positioning information of at least one of the strong interference signals in the real-time interference source distribution map; Comparing the actual acoustic signal arrival timing of the target acoustic signal with the theoretical acoustic signal arrival timing to obtain a comparison result; It is determined whether the target acoustic signal indicates damage inside the building based on the comparison result.

2. The method for detecting prefabricated buildings according to claim 1, wherein: The method includes collecting strong interference signals generated by external impact sources of the building, obtaining sound source positioning information of the strong interference signals, converging the sound source positioning information of the strong interference signals on a preset three-dimensional geometric model aligned with a coordinate system, and generating a real-time interference source distribution map, including: Monitor the impact frequency of high-energy impact events from external impact sources of buildings; When the impact frequency exceeds a preset frequency threshold, the conventional detection mode is switched to the reverse discrimination mode, the strong interference signal generated by the impact source outside the building is collected, the sound source positioning information of the strong interference signal is obtained, and the sound source positioning information of the strong interference signal is converged on a preset three-dimensional geometric model aligned with the coordinate system to generate the real-time interference source distribution map.

3. The method for detecting prefabricated buildings according to claim 1, wherein: In the conventional detection mode, after detecting the target acoustic signal to be distinguished, the method further includes: Performing sound source positioning calculation according to the target acoustic signal to obtain a positioning result including discrete coordinate points; Extracting spatial distribution features of the positioning result to determine whether the target acoustic signal is generated by a linear interference source, the spatial distribution features including the center of mass position, the main axis direction, and the major-minor axis ratio; If the target acoustic signal is generated by the linear interference source, the portion of the target acoustic signal generated by the linear interference source is removed.

4. The method for detecting prefabricated buildings according to claim 3, wherein: Extracting the spatial distribution characteristics of the positioning result and determining whether the target acoustic signal is generated by a linear interference source includes: Extracting spatial distribution features of the positioning results, and when the spatial distribution features of the positioning results present multi-center features, clustering the discrete coordinate points of the positioning results into a plurality of coordinate point subsets; Extracting spatial distribution features from each subset of coordinate points, and independently comparing them with the linear interference source; If the spatial distribution characteristics of all the coordinate point subsets successfully match the linear interference source, the corresponding target acoustic signal is determined to be a composite external interference signal generated by the linear interference source.

5. The method for detecting prefabricated buildings according to claim 3, wherein: The determining, based on the comparison result, whether the target acoustic signal indicates damage inside the building includes: If the comparison result is that the actual acoustic signal arrival timing of the target acoustic signal does not match the theoretical acoustic signal arrival timing, then the target acoustic signal is determined to be a suspected internal damage signal; Performing spatial distance calculation between discrete coordinate points obtained by performing sound source localization calculation on the suspected internal damage signal and the geometric boundary of the three-dimensional geometric model; The number of discrete coordinate points of the suspected internal damage signal belonging to the three-dimensional geometric model is counted according to a preset rule. If the attribution ratio exceeds a preset threshold value, the suspected internal damage signal is determined to be an external interference signal.

6. The method for detecting prefabricated buildings according to claim 5, wherein: The method further comprises: Based on the acoustic propagation characteristics in the building medium, combined with the three-dimensional geometric model and sensor layout information, the propagation path deflection result of the propagation path of the strong interference signal at the multi-medium interface is calculated; According to the propagation path deflection result, the spatial boundary of the three-dimensional geometric model is corrected to form a spatial offset region associated with propagation path compensation; The counting of the number of discrete coordinate points of the suspected internal damage signal belonging to the three-dimensional geometric model according to a preset rule includes: The number of discrete coordinate points of the suspected internal damage signal belonging to the spatial offset region is counted according to a preset rule.

7. The method for detecting prefabricated buildings according to claim 6, wherein: The method of calculating the deflection of the propagation path of the strong interference signal at the multi-medium interface based on the acoustic propagation characteristics in the building medium and combining the three-dimensional geometric model with the sensor layout information further includes: Get the current ambient temperature data of the building; Based on the ambient temperature data and the sound velocity-temperature relationship of the material, updating corresponding acoustic propagation characteristics; Based on the updated acoustic propagation characteristics in combination with the three-dimensional geometric model and sensor layout information, the propagation path deflection result of the strong interference signal at the multi-medium interface is updated.

8. The method for detecting prefabricated buildings according to claim 6, wherein: The step of correcting the spatial boundary of the three-dimensional geometric model according to the propagation path deflection result to form a spatial offset region associated with propagation path compensation includes: identifying at least one area of ​​sudden deviation change according to a variation of a deviation vector of the propagation path deflection result; According to the overlap between the three-dimensional geometric model and the sudden deviation change region, the three-dimensional geometric model is divided into a first entity located in the sudden deviation change region and a second entity located outside the sudden deviation change region; applying a first correction process to the first entity and a second correction process to the second entity, wherein the first correction process has a higher information density than the second correction process; All modified first entities and second entities are combined to generate the spatial offset region.

9. The method for detecting prefabricated buildings according to claim 8, wherein: The step of identifying at least one area of ​​sudden deviation change according to the deviation vector change of the propagation path deflection result includes: forming a change rate data set according to a change amount of a deviation vector of the propagation path deflection result; Partitioning the change rate data set according to a distribution relationship of data points in the change rate data set to form a high change rate data subset and a low change rate data subset; The spatial range corresponding to the high-change-rate data subset is determined as the offset sudden change area.

10. An assembled building detection system, characterized in that: The system comprises: The acquisition and positioning module is used to collect strong interference signals generated by impact sources outside the building, obtain sound source positioning information of the strong interference signals, converge the sound source positioning information of the strong interference signals on a preset three-dimensional geometric model aligned with the coordinate system, and generate a real-time interference source distribution map; A detection calculation module is configured to, in a conventional detection mode, after detecting a target acoustic signal to be distinguished, derive a theoretical acoustic signal arrival time sequence based on the sound source localization information of at least one of the strong interference signals in the real-time interference source distribution map; A timing comparison module is used to compare the actual acoustic signal arrival timing of the target acoustic signal with the theoretical acoustic signal arrival timing to obtain a comparison result; The damage determination module is used to determine whether the target acoustic signal is damage inside the building according to the comparison result.

Images