A method and system for intelligent deformation detection of spatial grid structures

Abstract

This invention relates to the field of spatial grid structure detection technology, providing an intelligent deformation detection method and system for spatial grid structures. The invention employs multi-UAV self-planning monitoring and control; identifies building monitoring data, determining multiple target grid structures when anomalies are observed; monitors vibration at multiple grid nodes; calculates multiple deformation sensitivity indices and performs spatial neighborhood smoothing filtering; calculates adaptive thresholds, compares them, and selects multiple candidate deformation nodes. This system enables multi-UAV self-planning monitoring and control and identification, selecting multiple target grid structures with anomalies, then monitoring vibration at multiple grid nodes on these target structures, calculating, filtering, and comparing sensitivity indices to select multiple candidate deformation nodes. This comprehensive approach reflects the true deformation of the structure, providing more accurate support for subsequent structural safety assessments and risk warnings.

Description

Technical Field

[0001] This invention belongs to the field of spatial grid structure detection technology, and particularly relates to an intelligent deformation detection method and system for spatial grid structures. Background Technology

[0002] Spatial grid structure inspection is a technology for detecting and evaluating the geometry, component status, connection nodes, and overall stress performance of large-span spatial grids, shells, trusses, and other spatial steel structure systems. It is used to verify whether spatial grid structures meet safety, stability, and reliability requirements during the design, construction, operation, and maintenance phases, and to promptly detect structural defects, abnormal deformations, and potential safety hazards.

[0003] In spatial grid structure inspection, deformation detection is one of the most basic inspection contents.

[0004] In existing technologies, deformation detection of spatial grid structures is often carried out by manually setting up monitoring points. Due to limitations in detection costs, equipment quantity, and site conditions, only a small number of key nodes can be selected for monitoring. For spatial grid structures with a large number of nodes and complex spatial distribution, it is often difficult to fully reflect the actual deformation of the structure, resulting in certain limitations in detection and affecting the accuracy of structural safety assessment and risk warning. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent deformation detection method and system for spatial grid structures, aiming to solve the technical problems existing in the prior art mentioned in the background.

[0006] The embodiments of the present invention are implemented as follows:

[0007] A method for intelligent deformation detection of spatial grid structures, the method specifically includes the following steps:

[0008] Multi-UAV self-planning monitoring and control of spatial grid buildings to acquire building monitoring data;

[0009] Based on preset structural anomaly features, the building monitoring data is identified to determine whether there are appearance anomalies, and when there are appearance anomalies, multiple target grid structures are determined.

[0010] Vibration monitoring is performed on multiple grid nodes on multiple target grid structures to collect vibration response data, output the current mode shape matrix, and match the corresponding reference mode shape matrix;

[0011] Based on the current mode shape matrix and the reference mode shape matrix, the deformation sensitivity index of multiple mesh nodes is calculated, and spatial neighborhood smoothing filtering is performed to calculate multiple optimized sensitivity indices;

[0012] Based on multiple optimized sensitivity indices, an adaptive threshold is calculated, and the multiple optimized sensitivity indices are compared with the adaptive threshold to select multiple candidate deformation nodes.

[0013] As a further limitation of the technical solution of this invention, the step of performing self-planning monitoring and control of multiple UAVs on spatial grid buildings to obtain building monitoring data specifically includes the following steps:

[0014] Identify the spatial grid buildings to be inspected and acquire their BIM data;

[0015] Choose the first drone and the second drone;

[0016] Based on the BIM data and preset planning and monitoring parameters, a coarse monitoring route is planned;

[0017] According to the coarse monitoring route, the first UAV is subjected to coarse monitoring control, and the coarse monitoring data transmitted by the first UAV is acquired;

[0018] Based on preset actual monitoring parameters, the coarse monitoring data is analyzed in detail, and multiple detailed monitoring spatial points are planned.

[0019] The second UAV is monitored and controlled according to the actual monitoring parameters and multiple detailed monitoring spatial points to acquire building monitoring data.

[0020] As a further limitation of the technical solution of this embodiment of the invention, the step of identifying the building monitoring data based on preset structural anomaly features, determining whether there are appearance anomalies, and identifying multiple target grid structures when appearance anomalies are present specifically includes the following steps:

[0021] Import preset structural anomaly features;

[0022] Based on the structural anomaly characteristics, the building monitoring data is identified and matched to determine whether there are any appearance anomalies.

[0023] When there are abnormal appearances, acquire abnormal location data;

[0024] Based on the aforementioned abnormal location data, multiple target grid structures are identified within the spatial grid building.

[0025] As a further limitation of the technical solution of this embodiment of the invention, the step of monitoring the vibration of multiple grid nodes on multiple target grid structures, collecting vibration response data, outputting the current mode shape matrix, and matching the corresponding reference mode shape matrix specifically includes the following steps:

[0026] Determine multiple grid nodes on multiple target grid structures;

[0027] Vibration monitoring was performed on multiple of the aforementioned grid nodes, and vibration response data was collected.

[0028] The vibration response data is identified, and the current mode shape matrix is ​​output.

[0029] Based on the preset reference mode data, the reference mode matrix corresponding to multiple grid nodes is matched.

[0030] As a further limitation of the technical solution of this embodiment of the invention, the step of calculating the deformation sensitivity index of multiple grid nodes based on the current mode shape matrix and the reference mode shape matrix, and performing spatial neighborhood smoothing filtering, specifically includes the following steps:

[0031] Select from the current mode shape matrix and the reference mode shape matrix The dominant mode is determined by the first dominant mode, and the current dominant matrix and the baseline dominant matrix are obtained.

[0032] Based on the current dominant matrix and the baseline dominant matrix, calculate the deformation sensitivity index of multiple mesh nodes;

[0033] Adjacency identification is performed on multiple grid nodes to obtain a set of adjacent nodes for each grid node;

[0034] Based on the multiple sets of adjacent nodes, spatial neighborhood smoothing filtering is applied to the multiple deformation sensitivity indices to calculate the optimized sensitivity indices of the multiple grid nodes.

[0035] As a further limitation of the technical solution of this embodiment of the invention, the calculation formulas for the plurality of deformation sensitivity indices are as follows:

[0036] ;

[0037] in, For the first Deformation sensitivity index of each grid node; The first in the current dominant matrix The first grid node The mode shape components of the first mode; The first in the baseline dominant matrix The first grid node The mode shape components of the first mode;

[0038] The formulas for calculating the various optimization sensitivity indices are as follows:

[0039] ;

[0040] in, For the first The optimization sensitivity index of each grid node; For the first The set of adjacent nodes of a grid node; The first in the set of adjacent nodes Deformation sensitivity index of adjacent nodes; Representing the The number of nodes in the set of neighboring nodes of a grid node.

[0041] As a further limitation of the technical solution of this embodiment of the invention, the step of calculating an adaptive threshold based on multiple optimized sensitivity indices and comparing the multiple optimized sensitivity indices with the adaptive threshold to select multiple candidate deformation nodes specifically includes the following steps:

[0042] Based on multiple optimized sensitivity indices, an adaptive threshold is calculated;

[0043] Compare the multiple optimized sensitivity indices with the adaptive threshold and record the comparison results;

[0044] Based on the comparison results, multiple candidate deformation nodes with optimization sensitivity indices greater than the adaptive threshold are selected from the multiple mesh nodes.

[0045] As a further limitation of the technical solution of this embodiment of the invention, the calculation formula of the adaptive threshold is:

[0046] ;

[0047] in, For adaptive thresholds; there are a total of 1 grid node, and ; These are the preset sensitivity parameters.

[0048] A spatial grid structure intelligent deformation detection system, the system comprising a self-planning monitoring and control module, an appearance anomaly identification and judgment module, a vibration monitoring and processing module, a sensitivity index calculation module, and a deformation node selection module, wherein:

[0049] The self-planning monitoring and control module is used to perform self-planning monitoring and control of multiple UAVs on spatial grid buildings and acquire building monitoring data;

[0050] The appearance anomaly identification and judgment module is used to identify the building monitoring data based on preset structural anomaly features, determine whether there is an appearance anomaly, and identify multiple target grid structures when there is an appearance anomaly.

[0051] The vibration monitoring and processing module is used to monitor the vibration of multiple grid nodes on multiple target grid structures, collect vibration response data, output the current mode shape matrix, and match the corresponding reference mode shape matrix.

[0052] The sensitivity index calculation module is used to calculate the deformation sensitivity index of multiple grid nodes based on the current mode shape matrix and the reference mode shape matrix, and to perform spatial neighborhood smoothing filtering to calculate multiple optimized sensitivity indices.

[0053] The deformation node selection module is used to calculate an adaptive threshold based on multiple optimized sensitivity indices, compare the multiple optimized sensitivity indices with the adaptive threshold, and select multiple candidate deformation nodes.

[0054] As a further limitation of the technical solution of this embodiment of the invention, the sensitivity index calculation module specifically includes:

[0055] The dominant mode selection unit is used to select from the current mode shape matrix and the reference mode shape matrix. The dominant mode is determined by the first dominant mode, and the current dominant matrix and the baseline dominant matrix are obtained.

[0056] The deformation sensitivity index calculation unit is used to calculate the deformation sensitivity index of multiple mesh nodes based on the current dominant matrix and the reference dominant matrix.

[0057] The adjacent identification unit is used to identify the adjacent nodes of multiple grid nodes and obtain the set of adjacent nodes of multiple grid nodes;

[0058] An optimized sensitivity index calculation unit is used to perform spatial neighborhood smoothing filtering on multiple deformation sensitivity indices based on multiple sets of adjacent nodes, and to calculate the optimized sensitivity index of multiple grid nodes.

[0059] Compared with the prior art, the beneficial effects of the present invention are:

[0060] This invention enables self-planning monitoring and control of multiple unmanned aerial vehicles (UAVs); identifies building monitoring data and determines multiple target grid structures when anomalies are observed; monitors vibration at multiple grid nodes; calculates multiple deformation sensitivity indices and performs spatial neighborhood smoothing filtering; calculates adaptive thresholds, compares them, and selects multiple candidate deformation nodes. This comprehensive approach, capable of identifying and controlling multiple UAVs, selecting multiple target grid structures with anomalies, monitoring vibration at multiple grid nodes on these target structures, calculating, filtering, and comparing sensitivity indices, and selecting multiple candidate deformation nodes, fully reflects the true deformation of the structure and provides more accurate support for subsequent structural safety assessments and risk warnings. Attached Figure Description

[0061] Figure 1 A flowchart of the intelligent deformation detection method for spatial grid structures provided in an embodiment of the present invention is shown;

[0062] Figure 2 The following is an application architecture diagram of the intelligent deformation detection system for spatial grid structures provided in an embodiment of the present invention. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0064] Understandably, deformation detection is one of the most basic aspects of spatial grid structure inspection. Currently, deformation detection of spatial grid structures is often carried out by manually setting up monitoring points. Due to limitations in inspection costs, equipment quantity, and site conditions, only a small number of key nodes can be selected for monitoring. For spatial grid structures with a large number of nodes and complex spatial distribution, it is often difficult to fully reflect the actual deformation of the structure, resulting in certain limitations in the inspection and affecting the accuracy of structural safety assessment and risk warning.

[0065] To address the aforementioned issues, this invention discloses an intelligent deformation detection method and system for spatial grid structures. This method involves acquiring building monitoring data through multi-UAV self-planning monitoring and control of a spatial grid building; identifying the building monitoring data based on preset structural anomaly features to determine if there are any appearance anomalies, and identifying multiple target grid structures if such anomalies are present; monitoring the vibration of multiple grid nodes on the target grid structures, collecting vibration response data, outputting the current mode shape matrix, and matching it with the corresponding reference mode shape matrix; calculating the deformation sensitivity index of multiple grid nodes based on the current and reference mode shape matrices, performing spatial neighborhood smoothing filtering, and calculating multiple optimized sensitivity indices; calculating an adaptive threshold based on the multiple optimized sensitivity indices, comparing the optimized sensitivity indices with the adaptive threshold, and selecting multiple candidate deformation nodes. This method enables multi-UAV self-planning monitoring and control and identification, selecting multiple target grid structures with appearance anomalies, and then monitoring the vibration of multiple grid nodes on these target grid structures. The calculation, filtering, and comparison of sensitivity indices to select multiple candidate deformation nodes comprehensively reflects the true deformation of the structure, providing more accurate support for subsequent structural safety assessments and risk warnings.

[0066] Specifically, Figure 1 A flowchart of the intelligent deformation detection method for spatial grid structures provided in an embodiment of the present invention is shown.

[0067] In a preferred embodiment of the present invention, a method for intelligent deformation detection of a spatial grid structure specifically includes the following steps:

[0068] Step S101: Perform self-planning monitoring and control of multiple UAVs on the spatial grid building to obtain building monitoring data.

[0069] In this embodiment of the invention, a spatial grid building to be detected is determined, and its BIM data is acquired. Simultaneously, a first UAV for global monitoring and a second UAV for refined monitoring are selected. Based on preset planning monitoring parameters, a global coarse monitoring planning analysis is performed on the BIM data to plan a coarse monitoring route capable of long-distance inspection and monitoring of all spatial grid structures within the spatial grid building. Following this route, the first UAV is controlled for coarse monitoring, enabling it to perform long-distance, full-area inspection of all spatial grid structures and transmit the acquired coarse monitoring data back. By receiving the coarse monitoring data transmitted by the first UAV and based on preset actual monitoring parameters, a refined monitoring analysis is performed on the coarse monitoring data, focusing on all spatial grid structures within the spatial grid building. Multiple fine monitoring spatial points are planned. Then, based on the actual monitoring parameters and these fine monitoring spatial points, the second UAV is controlled for actual monitoring, enabling it to perform close-range actual monitoring of multiple spatial grid structures within the spatial grid building at these fine monitoring spatial points and transmit the acquired building monitoring data back. This process is repeated to obtain the building monitoring data transmitted by the second UAV.

[0070] It is understandable that the BIM data of a spatial grid building consists of the three-dimensional model data and node distribution data of the spatial grid building.

[0071] Understandably, each fine monitoring point corresponds to one or more spatial grid structures, ensuring that the second UAV can perform fine monitoring of the corresponding spatial grid structure at the optimal observation distance and angle.

[0072] It is understandable that long-range monitoring and short-range monitoring refer to the distance between the first and second UAVs and the spatial grid structure in the spatial grid building. The first UAV is farther away from the spatial grid structure, while the second UAV is closer to the spatial grid structure.

[0073] It is understandable that the first UAV's long-distance monitoring of the spatial grid structure is not to monitor the anomalies of the spatial grid structure itself, but to monitor the spatial position of the spatial grid structure and the spatial position of other structures attached to the spatial grid structure, thereby providing data support for the subsequent planning of close-range actual monitoring of the spatial grid structure itself.

[0074] It is understood that, in the embodiments of the present invention, other structures attached to the spatial grid structure may include roof panels, sunshades, protective nets, billboards, support structures, lighting equipment, monitoring equipment, broadcasting equipment, etc., and the positions of these other structures may be dynamically adjusted with time and weather changes. Therefore, it is necessary to collect coarse monitoring data through the first UAV to provide data support for the planning of fine monitoring spatial points that conform to the current state.

[0075] Understandably, the planned monitoring parameters include data such as the flight speed, hovering distance, sampling frequency, and image resolution of the first UAV; the actual monitoring parameters include data such as the flight speed, hovering distance, sampling frequency, and image resolution of the second UAV.

[0076] Specifically, in another preferred embodiment provided by the present invention, the step of performing self-planning monitoring and control of multiple UAVs on spatial grid buildings to obtain building monitoring data specifically includes the following steps:

[0077] Identify the spatial grid buildings to be inspected and acquire their BIM data;

[0078] Choose the first drone and the second drone;

[0079] Based on the BIM data and preset planning and monitoring parameters, a coarse monitoring route is planned;

[0080] According to the coarse monitoring route, the first UAV is subjected to coarse monitoring control, and the coarse monitoring data transmitted by the first UAV is acquired;

[0081] Based on preset actual monitoring parameters, the coarse monitoring data is analyzed in detail, and multiple detailed monitoring spatial points are planned.

[0082] The second UAV is monitored and controlled according to the actual monitoring parameters and multiple detailed monitoring spatial points to acquire building monitoring data.

[0083] Furthermore, the intelligent deformation detection method for spatial grid structures also includes the following steps:

[0084] Step S102: Based on preset structural anomaly features, identify the building monitoring data, determine whether there are appearance anomalies, and when there are appearance anomalies, determine multiple target grid structures.

[0085] In this embodiment of the invention, preset structural anomaly features are imported, and then the structural anomaly features are used as identification references to perform feature identification and matching on building monitoring data to determine whether there are appearance anomalies. If appearance anomalies are determined, anomaly location data is obtained. Then, according to the anomaly location data, multiple target grid structures with appearance anomalies are determined in the spatial grid building. Among them, structural anomaly features include rust features, dent features, crack features, coating anomaly features, etc.

[0086] It is understandable that abnormal location data includes data such as structure number and spatial coordinates.

[0087] Specifically, in another preferred embodiment provided by the present invention, the step of identifying the building monitoring data based on preset structural anomaly features, determining whether there are appearance anomalies, and identifying multiple target grid structures when appearance anomalies are present specifically includes the following steps:

[0088] Import preset structural anomaly features;

[0089] Based on the structural anomaly characteristics, the building monitoring data is identified and matched to determine whether there are any appearance anomalies.

[0090] When there are abnormal appearances, acquire abnormal location data;

[0091] Based on the aforementioned abnormal location data, multiple target grid structures are identified within the spatial grid building.

[0092] Furthermore, the intelligent deformation detection method for spatial grid structures also includes the following steps:

[0093] Step S103: Vibration monitoring is performed on multiple grid nodes on the multiple target grid structures, vibration response data is collected, the current mode shape matrix is ​​output, and the corresponding reference mode shape matrix is ​​matched.

[0094] In this embodiment of the invention, multiple grid nodes on multiple target grid structures are determined, MEMS triaxial acceleration sensors are deployed on multiple grid nodes, vibration monitoring is performed on multiple grid nodes, vibration response data is collected, and then the vibration response data is identified, the corresponding current mode shape matrix is ​​output, and the reference mode shape matrix corresponding to multiple grid nodes is matched from the preset reference mode shape data.

[0095] It is understood that, in the embodiments of the present invention, a mesh node can be a connection point on the target mesh structure where they intersect, such as a bolt ball node, a welded ball node, etc.

[0096] It is understandable that the reference mode shape matrix is ​​the mode shape matrix corresponding to the vibration response of multiple grid nodes in a healthy state.

[0097] Specifically, in another preferred embodiment provided by the present invention, the step of monitoring the vibration of multiple grid nodes on multiple target grid structures, collecting vibration response data, outputting the current mode shape matrix, and matching the corresponding reference mode shape matrix specifically includes the following steps:

[0098] Determine multiple grid nodes on multiple target grid structures;

[0099] Vibration monitoring was performed on multiple of the aforementioned grid nodes, and vibration response data was collected.

[0100] The vibration response data is identified, and the current mode shape matrix is ​​output.

[0101] Based on the preset reference mode data, the reference mode matrix corresponding to multiple grid nodes is matched.

[0102] Furthermore, the intelligent deformation detection method for spatial grid structures also includes the following steps:

[0103] Step S104: Calculate the deformation sensitivity index of multiple grid nodes based on the current mode shape matrix and the reference mode shape matrix, and perform spatial neighborhood smoothing filtering to calculate multiple optimized sensitivity indices.

[0104] In this embodiment of the invention, by selecting from both the current mode shape matrix and the reference mode shape matrix... The dominant mode is identified to obtain the corresponding current dominant matrix and baseline dominant matrix. Then, based on the current dominant matrix and baseline dominant matrix, the deformation sensitivity index of multiple grid nodes is calculated, and neighbor identification is performed on multiple grid nodes to obtain the neighbor node set of multiple grid nodes. Then, based on the multiple neighbor node set, spatial neighborhood smoothing filtering is applied to the deformation sensitivity index of multiple grid nodes to calculate the optimized sensitivity index of multiple grid nodes. Specifically, the calculation formula of multiple deformation sensitivity indices is as follows:

[0105] ;

[0106] in, For the first Deformation sensitivity index of each grid node; The first in the current dominant matrix The first grid node The mode shape components of the first mode; The first in the baseline dominant matrix The first grid node The mode shape components of the first mode;

[0107] The formulas for calculating multiple optimization sensitivity indices are as follows:

[0108] ;

[0109] in, For the first The optimization sensitivity index of each grid node; For the first The set of adjacent nodes of a grid node; The first in the set of adjacent nodes Deformation sensitivity index of adjacent nodes; Representing the The number of nodes in the set of neighboring nodes of a grid node.

[0110] Understandably, multiple deformation sensitivity indices can reflect the salience of deformation at multiple mesh nodes. Specifically, the value range of the deformation sensitivity index is... Furthermore, the larger the deformation sensitivity index, the more significant the deformation of the corresponding mesh nodes.

[0111] Understandably, by performing spatial neighborhood smoothing filtering on multiple deformation sensitivity indices and calculating the optimized sensitivity indices of multiple grid nodes, false positives caused by single-point noise can be suppressed, thereby preserving the clumping characteristics of the true deformation region.

[0112] In a specific embodiment of the present invention, among the nine candidate deformation nodes, six correspond to preset deformation locations (all three node subsidence points were detected, two of the three member bending points were detected, and one bending point was not detected due to its small amplitude (5mm); the detection rate is 5 / 6 = 83.3%). Of the remaining three candidate nodes, two are nodes below the coating peeling area (abnormal appearance but no structural deformation), and one is a node whose vibration mode shift was caused by the addition of a new equipment support. Traditional manual monitoring of all nodes would require at least 80-120 monitoring points to achieve the same spatial coverage. This invention requires only 57 MEMS triaxial accelerometers (deployed in 5 areas with abnormal appearance), significantly reducing the sensor installation area. Specifically, the effect of spatial neighborhood smoothing filtering was evaluated through comparison: before filtering, there were 17 nodes with DSI > 0.68 (including 7 isolated noise points); after filtering, there were 9 nodes with OSI > 0.68 (all isolated noise points were suppressed), reducing the number of candidate nodes by approximately 47%, and improving the consistency with the preset deformation location from 64.7% to 83.3%. The specific effect comparison evaluation is shown in the table below:

[0113] index Before filtering (DSI) After filtering (OSI) Number of nodes exceeding the threshold 17 9 Wherein the corresponding preset deformation 11（64.7%） 8 (88.9%, including 2 areas with abnormal appearance) Isolated noise points 7 0 Preset deformation detection rate 83.3% 83.3%

[0114] Specifically, in another preferred embodiment provided by the present invention, the step of calculating the deformation sensitivity index of multiple grid nodes based on the current mode shape matrix and the reference mode shape matrix, and performing spatial neighborhood smoothing filtering, and calculating multiple optimized sensitivity indices specifically includes the following steps:

[0115] Select from the current mode shape matrix and the reference mode shape matrix The dominant mode is determined by the first dominant mode, and the current dominant matrix and the baseline dominant matrix are obtained.

[0116] Based on the current dominant matrix and the baseline dominant matrix, calculate the deformation sensitivity index of multiple mesh nodes;

[0117] Adjacency identification is performed on multiple grid nodes to obtain a set of adjacent nodes for each grid node;

[0118] Based on the multiple sets of adjacent nodes, spatial neighborhood smoothing filtering is applied to the multiple deformation sensitivity indices to calculate the optimized sensitivity indices of the multiple grid nodes.

[0119] Furthermore, the intelligent deformation detection method for spatial grid structures also includes the following steps:

[0120] Step S105: Calculate an adaptive threshold based on multiple optimized sensitivity indices, and compare the multiple optimized sensitivity indices with the adaptive threshold to select multiple candidate deformation nodes.

[0121] In this embodiment of the invention, an adaptive threshold is calculated based on multiple optimized sensitivity indices. The optimized sensitivity indices are then compared with the adaptive threshold, and the comparison results are recorded. Following this, based on the comparison results, grid nodes with optimized sensitivity indices greater than the adaptive threshold are selected from multiple grid nodes and identified as candidate deformation nodes. This achieves automatic selection of multiple candidate deformation nodes, providing a basis for subsequent structural safety assessment and risk warning. Specifically, the formula for calculating the adaptive threshold is:

[0122] ;

[0123] in, For adaptive thresholds; there are a total of 1 grid node, and ; These are the preset sensitivity parameters.

[0124] Specifically, in another preferred embodiment provided by the present invention, the step of calculating an adaptive threshold based on a plurality of optimized sensitivity indices, and comparing the plurality of optimized sensitivity indices with the adaptive threshold to select a plurality of candidate deformation nodes specifically includes the following steps:

[0125] Based on multiple optimized sensitivity indices, an adaptive threshold is calculated;

[0126] Compare the multiple optimized sensitivity indices with the adaptive threshold and record the comparison results;

[0127] Based on the comparison results, multiple candidate deformation nodes with optimization sensitivity indices greater than the adaptive threshold are selected from the multiple mesh nodes.

[0128] Understandably, the comprehensive evaluation of the intelligent deformation detection method for spatial grid structures is as follows:

[0129] Evaluation Items Traditional manual sampling inspection Method of the present invention Node coverage strategy Fixed, limited deployment points, with a coverage rate of 5-15%. Drone-based preliminary screening → Vibration measurement of suspicious areas, focusing resources on highly suspicious nodes Testing basis Single node coordinate change Multimode shape deviation + spatial neighborhood smoothing False positive control Relies on manual review, without systematic filtering Spatial neighborhood smoothing filter + adaptive threshold statistical screening Sensor deployment The sensors are evenly distributed throughout the area, but some may be monitoring normal areas. Deploy only in areas with abnormal appearance to reduce invalid installations. Adaptability None, all regions are treated the same. Dynamic planning of coarse monitoring data for fine monitoring spatial points Scalability As the number of nodes increases, the number of sensors increases linearly. The cost of the coarse screening stage is independent of the number of nodes, while the fine screening stage only focuses on abnormal areas.

[0130] Furthermore, Figure 2 The following is an application architecture diagram of the intelligent deformation detection system for spatial grid structures provided in an embodiment of the present invention.

[0131] Specifically, in another preferred embodiment provided by the present invention, a smart deformation detection system for a spatial grid structure includes:

[0132] The self-planning monitoring and control module 101 is used to perform self-planning monitoring and control of multiple UAVs on spatial grid buildings and acquire building monitoring data.

[0133] In this embodiment of the invention, the self-planning monitoring and control module 101 determines the spatial grid building to be detected and acquires its BIM data. Simultaneously, it selects a first UAV for global monitoring and a second UAV for refined monitoring. Based on preset planning and monitoring parameters, it performs global coarse monitoring planning analysis on the BIM data, planning a coarse monitoring route capable of long-distance inspection and monitoring of all spatial grid structures within the spatial grid building. Following this route, it performs coarse monitoring control on the first UAV, enabling it to perform long-distance, full-area inspection of all spatial grid structures along the route and transmit the acquired coarse monitoring data back. By receiving the coarse monitoring data transmitted by the first UAV and based on preset actual monitoring parameters, it performs refined monitoring analysis on the coarse monitoring data, closely monitoring all spatial grid structures within the spatial grid building, planning multiple fine monitoring spatial points. Then, according to the actual monitoring parameters and these fine monitoring spatial points, it performs actual monitoring control on the second UAV, enabling it to perform close-range actual monitoring of multiple spatial grid structures within the spatial grid building at these fine monitoring spatial points and transmit the acquired building monitoring data back, thereby acquiring the building monitoring data transmitted by the second UAV.

[0134] The appearance anomaly identification and judgment module 102 is used to identify the building monitoring data based on preset structural anomaly features, determine whether there is an appearance anomaly, and identify multiple target grid structures when there is an appearance anomaly.

[0135] In this embodiment of the invention, the appearance anomaly identification and judgment module 102 imports preset structural anomaly features, and then uses the structural anomaly features as an identification reference to perform feature identification and matching on building monitoring data to determine whether there is an appearance anomaly. If it is determined that there is an appearance anomaly, it obtains anomaly location data, and then, according to the anomaly location data, determines multiple target grid structures with appearance anomalies in the spatial grid building. Among them, structural anomaly features include rust features, dent features, crack features, coating anomaly features, etc.

[0136] The vibration monitoring and processing module 103 is used to monitor the vibration of multiple grid nodes on multiple target grid structures, collect vibration response data, output the current mode shape matrix, and match the corresponding reference mode shape matrix.

[0137] In this embodiment of the invention, the vibration monitoring and processing module 103 determines multiple grid nodes on multiple target grid structures, deploys MEMS triaxial acceleration sensors on multiple grid nodes, monitors the vibration of multiple grid nodes, collects vibration response data, identifies the vibration response data, outputs the corresponding current mode shape matrix, and matches the reference mode shape matrix corresponding to multiple grid nodes from the preset reference mode shape data.

[0138] The sensitivity index calculation module 104 is used to calculate the deformation sensitivity index of multiple grid nodes based on the current mode shape matrix and the reference mode shape matrix, and to perform spatial neighborhood smoothing filtering to calculate multiple optimized sensitivity indices.

[0139] In this embodiment of the invention, the sensitivity index calculation module 104 selects from both the current mode shape matrix and the reference mode shape matrix. The dominant mode is identified to obtain the corresponding current dominant matrix and baseline dominant matrix. Then, based on the current dominant matrix and baseline dominant matrix, the deformation sensitivity index of multiple grid nodes is calculated, and neighbor identification is performed on multiple grid nodes to obtain the neighbor node set of multiple grid nodes. Then, based on the multiple neighbor node set, spatial neighborhood smoothing filtering is applied to the deformation sensitivity index of multiple grid nodes to calculate the optimized sensitivity index of multiple grid nodes. Specifically, the calculation formula of multiple deformation sensitivity indices is as follows:

[0140] ;

[0141] in, For the first Deformation sensitivity index of each grid node; The first in the current dominant matrix The first grid node The mode shape components of the first mode; The first in the baseline dominant matrix The first grid node The mode shape components of the first mode;

[0142] The formulas for calculating multiple optimization sensitivity indices are as follows:

[0143] ;

[0144] in, For the first The optimization sensitivity index of each grid node; For the first The set of adjacent nodes of a grid node; The first in the set of adjacent nodes Deformation sensitivity index of adjacent nodes; Representing the The number of nodes in the set of neighboring nodes of a grid node.

[0145] Specifically, in another preferred embodiment provided by the present invention, the sensitivity index calculation module 104 specifically includes:

[0146] The dominant mode selection unit is used to select from the current mode shape matrix and the reference mode shape matrix. The dominant mode is determined by the first dominant mode, and the current dominant matrix and the baseline dominant matrix are obtained.

[0147] The deformation sensitivity index calculation unit is used to calculate the deformation sensitivity index of multiple mesh nodes based on the current dominant matrix and the reference dominant matrix.

[0148] The adjacent identification unit is used to identify the adjacent nodes of multiple grid nodes and obtain the set of adjacent nodes of multiple grid nodes;

[0149] An optimized sensitivity index calculation unit is used to perform spatial neighborhood smoothing filtering on multiple deformation sensitivity indices based on multiple sets of adjacent nodes, and to calculate the optimized sensitivity index of multiple grid nodes.

[0150] Furthermore, the intelligent deformation detection system for the spatial grid structure also includes:

[0151] The deformation node selection module 105 is used to calculate an adaptive threshold based on multiple optimized sensitivity indices, and compare the multiple optimized sensitivity indices with the adaptive threshold to select multiple candidate deformation nodes.

[0152] In this embodiment of the invention, the deformation node selection module 105 calculates an adaptive threshold based on multiple optimized sensitivity indices, compares the multiple optimized sensitivity indices with the adaptive threshold, records the comparison results, and then selects grid nodes from multiple grid nodes whose corresponding optimized sensitivity indices are greater than the adaptive threshold, determining them as candidate deformation nodes. This achieves automatic selection of multiple candidate deformation nodes, providing a basis for subsequent structural safety assessment and risk warning. Specifically, the formula for calculating the adaptive threshold is:

[0153] ;

[0154] in, For adaptive thresholds; there are a total of 1 grid node, and ; These are the preset sensitivity parameters.

[0155] It is understood that, in the embodiments of the present invention, the key parameters and uses of some devices in the intelligent deformation detection system for spatial grid structures are as follows:

[0156] Equipment type Key parameters use First drone (inspection) Hexacopter, 42-megapixel wide-angle camera, RTK positioning accuracy ±2cm, 40-minute flight time. Comprehensive coarse monitoring Second UAV (Precision) Quadcopter, 61-megapixel high-resolution camera, laser ranging accuracy ±1mm, 30-minute flight time. Target area fine monitoring MEMS triaxial accelerometer Triaxial, sensitivity 1000mV / g, frequency range 0.1Hz—500Hz, weight 8g Nodal vibration response acquisition Vibration monitoring and processing module 16-channel, 24-bit ADC, 256Hz sampling rate, supports synchronous acquisition. Multi-node synchronous vibration recording Self-planning monitoring and control module Revit+Dynamo automation scripts Coarse monitoring route planning and fine monitoring spatial point generation Sensitivity Index Calculation Module 64GB RAM, RTX 4080, Python + MATLAB Modal analysis, sensitivity index calculation and visualization

[0157] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for intelligent deformation detection of spatial grid structures, characterized in that, The method specifically includes the following steps: Multi-UAV self-planning monitoring and control of spatial grid buildings to acquire building monitoring data; The process involves identifying the spatial grid buildings to be inspected and acquiring their BIM data. Simultaneously, a first UAV for global monitoring and a second UAV for refined monitoring are selected. Based on preset planning and monitoring parameters, a global coarse monitoring planning analysis is performed on the BIM data. A coarse monitoring route is planned to enable long-distance inspection and monitoring of all spatial grid structures within the spatial grid buildings. Following this route, the first UAV is controlled for coarse monitoring, allowing it to perform long-distance, full-area inspection of all spatial grid structures and transmit the acquired coarse monitoring data. By receiving the coarse monitoring data from the first UAV and using preset actual monitoring parameters, a refined monitoring analysis is performed on the coarse monitoring data, focusing on all spatial grid structures within the spatial grid buildings. Multiple fine monitoring points are planned. Then, based on the actual monitoring parameters and these fine monitoring points, the second UAV is controlled for actual monitoring, enabling it to perform close-range, actual monitoring of multiple spatial grid structures within the spatial grid buildings at these fine monitoring points and transmit the acquired building monitoring data. This process is repeated to obtain the building monitoring data transmitted by the second UAV. Based on preset structural anomaly features, the building monitoring data is identified to determine whether there are appearance anomalies, and when there are appearance anomalies, multiple target grid structures are determined. Vibration monitoring is performed on multiple grid nodes on multiple target grid structures to collect vibration response data, output the current mode shape matrix, and match the corresponding reference mode shape matrix; Based on the current mode shape matrix and the reference mode shape matrix, the deformation sensitivity index of multiple mesh nodes is calculated, and spatial neighborhood smoothing filtering is performed to calculate multiple optimized sensitivity indices; By selecting from both the current mode shape matrix and the reference mode shape matrix The dominant mode is identified to obtain the corresponding current dominant matrix and baseline dominant matrix. Then, based on the current dominant matrix and baseline dominant matrix, the deformation sensitivity index of multiple grid nodes is calculated, and neighbor identification is performed on multiple grid nodes to obtain the neighbor node set of multiple grid nodes. Then, based on the multiple neighbor node set, spatial neighborhood smoothing filtering is applied to the deformation sensitivity index of multiple grid nodes to calculate the optimized sensitivity index of multiple grid nodes. Specifically, the calculation formula of multiple deformation sensitivity indices is as follows: ； in, For the first Deformation sensitivity index of each grid node; The first in the current dominant matrix The first grid node The mode shape components of the first mode; The first in the baseline dominant matrix The first grid node The mode shape components of the first mode; The formulas for calculating multiple optimization sensitivity indices are as follows: ； in, For the first The optimization sensitivity index of each grid node; For the first The set of adjacent nodes of a grid node; The first in the set of adjacent nodes Deformation sensitivity index of adjacent nodes; Representing the The number of nodes in the set of adjacent nodes of a grid node; Based on multiple optimized sensitivity indices, an adaptive threshold is calculated, and the multiple optimized sensitivity indices are compared with the adaptive threshold to select multiple candidate deformation nodes; Based on multiple optimized sensitivity indices, an adaptive threshold is calculated. These indices are then compared to the adaptive threshold, and the comparison results are recorded. Following this, based on the comparison results, grid nodes with optimized sensitivity indices greater than the adaptive threshold are selected from multiple grid nodes and identified as candidate deformation nodes. This achieves automatic selection of multiple candidate deformation nodes, providing a foundation for subsequent structural safety assessment and risk warning. Specifically, the formula for calculating the adaptive threshold is: ； in, For adaptive thresholds; there are a total of 1 grid node, and ; These are the preset sensitivity parameters.

2. The intelligent deformation detection method for spatial grid structures according to claim 1, characterized in that, The process of identifying building monitoring data based on preset structural anomaly features, determining whether there are appearance anomalies, and identifying multiple target grid structures when appearance anomalies are present specifically includes the following steps: Import preset structural anomaly features; Based on the structural anomaly characteristics, the building monitoring data is identified and matched to determine whether there are any appearance anomalies. When there are abnormal appearances, acquire abnormal location data; Based on the aforementioned abnormal location data, multiple target grid structures are identified within the spatial grid building.

3. The intelligent deformation detection method for spatial grid structures according to claim 1, characterized in that, The process of monitoring the vibration of multiple grid nodes on multiple target grid structures, collecting vibration response data, outputting the current mode shape matrix, and matching the corresponding reference mode shape matrix specifically includes the following steps: Determine multiple grid nodes on multiple target grid structures; Vibration monitoring was performed on multiple of the aforementioned grid nodes, and vibration response data was collected. The vibration response data is identified, and the current mode shape matrix is ​​output. Based on the preset reference mode data, the reference mode matrix corresponding to multiple grid nodes is matched.

4. A space grid structure intelligent deformation detection system for performing the space grid structure intelligent deformation detection method as described in any one of claims 1-3, characterized in that, The system includes a self-planning monitoring and control module, an appearance anomaly identification and judgment module, a vibration monitoring and processing module, a sensitivity index calculation module, and a deformation node selection module, wherein: The self-planning monitoring and control module is used to perform self-planning monitoring and control of multiple UAVs on spatial grid buildings and acquire building monitoring data; The appearance anomaly identification and judgment module is used to identify the building monitoring data based on preset structural anomaly features, determine whether there is an appearance anomaly, and identify multiple target grid structures when there is an appearance anomaly. The vibration monitoring and processing module is used to monitor the vibration of multiple grid nodes on multiple target grid structures, collect vibration response data, output the current mode shape matrix, and match the corresponding reference mode shape matrix. The sensitivity index calculation module is used to calculate the deformation sensitivity index of multiple grid nodes based on the current mode shape matrix and the reference mode shape matrix, and to perform spatial neighborhood smoothing filtering to calculate multiple optimized sensitivity indices. The deformation node selection module is used to calculate an adaptive threshold based on multiple optimized sensitivity indices, compare the multiple optimized sensitivity indices with the adaptive threshold, and select multiple candidate deformation nodes.

5. The intelligent deformation detection system for spatial grid structures according to claim 4, characterized in that, The sensitivity index calculation module specifically includes: The dominant mode selection unit is used to select from the current mode shape matrix and the reference mode shape matrix. The dominant mode is determined by the first dominant mode, and the current dominant matrix and the baseline dominant matrix are obtained. The deformation sensitivity index calculation unit is used to calculate the deformation sensitivity index of multiple mesh nodes based on the current dominant matrix and the reference dominant matrix. The adjacent identification unit is used to identify the adjacent nodes of multiple grid nodes and obtain the set of adjacent nodes of multiple grid nodes; An optimized sensitivity index calculation unit is used to perform spatial neighborhood smoothing filtering on multiple deformation sensitivity indices based on multiple sets of adjacent nodes, and to calculate the optimized sensitivity index of multiple grid nodes.

Images