Large building curtain wall physical property detection system

Abstract

The invention discloses a large-scale building curtain wall physical performance detection system, and relates to the technical field of building curtain wall detection, the large-scale building curtain wall physical performance detection system comprises a double-zone pressure difference sealing assembly used for forming a detection zone and a buffer zone which are mutually independent on the outer side of a curtain wall, and the detection zone is in sealing cooperation with the outer surface of the curtain wall through an expandable sealing cabin; the buffer area surrounds the detection area through a flexible film and an adsorption belt and is isolated from the outside; the spraying matrix assembly comprises a nozzle array and a drop spectrum monitoring unit arranged on a spraying path, and the nozzle array is used for simulating different rain patterns according to preset spatial distribution and time sequence; according to the large building curtain wall physical performance detection system provided by the invention, the double-area pressure difference sealing assembly is matched with a buffer area to stabilize pressure first and a detection area to adjust pressure later, and volume change compensation of a boundary displacement signal is combined, so that the influence of external wind interference on the pressure difference stability is effectively reduced, the pressure stabilizing time is shortened, and the pressure maintaining precision is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building curtain wall detection, and particularly relates to a large building curtain wall physical performance detection system. BACKGROUND

[0002] In the prior art including the above patent, curtain wall physical performance detection mainly relies on large detection devices in the laboratory, and most detection systems can only complete a single project in one detection process, so that complete physical performance indexes can be obtained through multiple sets of equipment and multiple tests. During single project testing, in order to ensure that the parameters remain consistent during testing, the parameters need to be adjusted and checked for multiple times, which leads to low testing efficiency.

[0003] In the prior art including the above patent, curtain wall physical performance detection mainly relies on large detection devices in the laboratory, and most detection systems can only complete a single project in one detection process, so that complete physical performance indexes can be obtained through multiple sets of equipment and multiple tests. During single project testing, in order to ensure that the parameters remain consistent during testing, the parameters need to be adjusted and checked for multiple times, which leads to low testing efficiency. SUMMARY

[0004] The present application aims to provide a large building curtain wall physical performance detection system to solve the above problems in the prior art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions: A large building curtain wall physical performance detection system, comprising a double-zone differential pressure sealed assembly for forming a detection zone and a buffer zone independent of each other on the outside of the curtain wall, the detection zone being sealed to the outer surface of the curtain wall through an inflatable sealed cabin, and the buffer zone surrounding the detection zone and being isolated from the outside through a flexible film and an adsorption belt; A spray matrix assembly, comprising a spray head array for simulating different rain types according to a preset spatial distribution and timing sequence, and a drop spectrum monitoring unit arranged on a spray path, the drop spectrum monitoring unit being used to collect drop diameter and flow density data in real time and feed back to a control assembly to adjust the spray mode; An adaptive support mechanism is installed in the detection area of the curtain wall, a micro-force sensor is arranged at the support contact point to limit the single-point contact force and decouple the support reaction force from the deformation measurement result; a field self-calibration component includes a controllable leakage reference path and a displacement zero-point calibration unit, which performs reference leakage and zero-point calibration processes before detection starts and working condition switches, and is used to correct sensor drift and environmental interference; an intelligent diagnosis and control component is used to collect various detection data, which predicts the life cycle by establishing a curtain wall BIM model to map physical detection data.

[0006] In the above technical solution, the large building curtain wall physical performance detection system provided by the application effectively reduces the influence of external wind interference on the differential pressure stability, shortens the pressure stabilization time and improves the pressure retention precision, through the double-zone differential pressure sealing assembly cooperating with the buffer zone first pressure stabilization, the detection zone second pressure regulation strategy, and the volume change compensation of the boundary displacement signal.

[0007] It should be understood that the foregoing general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the disclosure.

[0008] The present application file provides an overview of various implementations or examples of the technology described in this disclosure, and is not a comprehensive disclosure of the full scope or all features of the disclosed technology. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0010] Figure 1 A flowchart of a large building curtain wall physical performance detection system is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0011] In order to make the purpose, technical scheme and advantages of the embodiments of the present disclosure clearer, the technical scheme of the embodiments of the present disclosure will be described clearly and completely below in conjunction with the drawings of the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present disclosure.

[0012] EMBODIMENTS The system comprises a double-zone differential pressure sealing assembly for forming a detection zone and a buffer zone on the outer side of the curtain wall, the detection zone being sealed to the outer surface of the curtain wall by an inflatable sealing cabin, and the buffer zone surrounding the detection zone and being isolated from the outside by a flexible film and an adsorption belt. The double-zone differential pressure sealing assembly comprises a sealing cover for forming a closed detection space on the surface of the curtain wall unit or structure, which comprises a rigid frame, an inflatable flexible film and a peripheral adsorption belt. The rigid frame serves as the support framework of the sealing cover and has a rectangular or polygonal frame structure, and is installed at the peripheral position of the outer surface of the curtain wall.

[0013] During installation, the adsorption belt directly contacts the outer surface of the curtain wall and fixes the edge position of the sealing cover through vacuum adsorption, magnetic adsorption or electrostatic adsorption, thereby forming a double-zone sealed space on the surface of the curtain wall.

[0014] The outer edge of the inflatable flexible film is provided with a reinforcing edge strip, which is fixed in the inner sealing groove of the rigid frame by bolt clamping, pressing or bonding, thereby achieving circumferential sealing connection.

[0015] The inflatable flexible film expands or shrinks when inflated or deflated, thereby forming the required sealed space.

[0016] One end of the adsorption belt is fixedly connected to the outer edge of the rigid frame, and the other end extends to the contact interface between the inflatable flexible film and the surface of the curtain wall and is sealed to the surface of the curtain wall.

[0017] The outer edge reinforcing edge strip of the flexible film and the inner edge of the adsorption belt are sealingly connected by a clamping ring or a pressing strip, thereby forming a continuous sealing boundary to prevent gas leakage from the connection between the flexible film and the adsorption belt.

[0018] The differential pressure generating unit consists of a variable frequency fan and a pressure stabilizing component. The variable frequency fan is located in the detection zone, and the pressure stabilizing component is located in the buffer zone. The variable frequency fan can continuously supply or exhaust air into the sealed space to establish a set positive or negative pressure environment. The pressure stabilizing component includes multiple proportional valves and multiple check valves installed on the pipeline.

[0019] A differential pressure sensor array is used to monitor the pressure difference between the inside and outside of the curtain wall in real time. Multiple static pressure taps and multiple differential pressure sensors are arranged in the detection area and the buffer area, respectively. Displacement sensor plates are arranged at key positions of the frame for volume change compensation. In this embodiment of the invention, the critical position of the frame refers to the area where stress concentration occurs after the sealing cover is installed on the curtain wall surface under the combined action of pressure difference and the expansion force of the flexible membrane, specifically including: Four-corner structural sections: Located at the four corners of the rigid frame, due to the geometric constraints of the frame and the change in the stretching direction of the flexible membrane, local deformation occurs under positive or negative pressure, affecting the actual volume of the enclosed space.

[0020] The middle section of the long side structure is located at the midpoint of the long side of the rigid frame. Under the action of internal pressure difference, it is prone to slight outward bulging or inward deformation along the normal direction, which affects the effective volume of the detection area and the buffer zone.

[0021] By placing displacement sensors in key areas of the frame, structural deformation can be acquired in real time. This deformation can then be combined with pressure data measured by differential pressure sensors to automatically compensate for volume changes, ensuring the accuracy of differential pressure measurements.

[0022] A flow meter, installed in the inlet or outlet passage, is used to measure the gas flow rate required to maintain a set pressure differential; A leakage reference unit, used to calibrate the zero-point drift of a differential pressure or flow link, includes a known flow resistance line located on the detection side and a controllable microvalve connected to the known flow resistance line; The controller, connected to the differential pressure generating unit, pressure sensor, and flow meter, is used to automatically adjust the fan speed and the opening of the pressure stabilizing device, and record the detection data. The spray matrix component includes a nozzle array and a droplet spectrum monitoring unit set along the spray path. The nozzle array is used to simulate different rain patterns according to a preset spatial distribution and time sequence. The droplet spectrum monitoring unit is used to collect droplet diameter and stream density data in real time and feed them back to the control component to adjust the spray mode. The nozzle density in the nozzle array is 4 nozzles / m. 2 The sprinklers are arranged in an alternating pattern, with each nozzle producing a certain volume of water. 4L / (m 2 •min), the coverage angle of each nozzle is 120°; The drop spectrum monitoring unit includes a monitoring window installed between the spray array downstream and the measured surface as a window for drop spectrum imaging and optical collection, and the monitoring window is provided with a fog-proof heating film or a micro-airflow purging port to prevent water mist from affecting the imaging quality. The imaging collection unit adopts a high-speed camera or a high-speed CMOS imaging module to capture the motion trajectory of water droplets at the cross section of the monitoring window. The illumination and background light source includes an LED array or a laser sheet light source arranged on the opposite side of the water droplet motion direction to make the water droplets form a high-contrast profile in the imaging. The data processing and analysis unit is a processor connected with the imaging module, which is provided with image segmentation and profile extraction algorithms to calculate the particle size distribution, number density and velocity of water droplets in real time. The adaptive support mechanism is used to adapt to the outer surface shape of different forms of curtain walls, and a micro-force sensor is arranged at the support contact point to limit the single-point contact force and decouple the support reaction force from the deformation measurement result. The adaptive support mechanism includes a support frame composed of multiple adjustable telescopic rods, and the rod ends are connected with the mounting base through spherical hinge joints to enable the support end to adaptively rotate in three-dimensional directions. The contact assembly is provided with an arc-shaped or spherical contact pad at the support end, which is matched with the curved surface member such as a curtain wall, and a micro-stroke slider or a flexible connecting piece is arranged between the contact pad and the telescopic rod end to absorb local micro-displacement and assembly errors. The locking unit includes an adsorption belt or a mechanical locking piece with adjustable pre-tightening force arranged on each telescopic rod, which can stably press the support end on the measured surface. The mechanical decoupling detection unit is provided with a micro-force sensing sheet on the back surface of the contact pad for real-time monitoring of the contact force.

[0023] The on-site self-calibration assembly includes a controllable leakage reference path and a displacement zero-point calibration unit, which performs reference leakage and zero-point calibration processes before detection starts and working condition switches to correct sensor drift and environmental interference. The controllable leakage reference path includes a plurality of capillary tubes arranged in the detection area, and the two ends of the capillary tubes are connected with the gas path of the detection area and the external environment gas path respectively to simulate a repeatable micro-leakage. The control valve unit is provided with an adjustable valve opening, an electromagnetic valve or a micro-needle valve installed on the reference pipeline to open or close the leakage path according to the control signal. The interface and detection port are provided with quick couplings at the pipeline access points to facilitate replacement of flow resistance elements. The displacement zero-point calibration unit comprises a calibration reference element, which is fixedly connected to a magnetic patch or a reflective target on the curtain wall detection area, and has a position accuracy of better than ±0.1 mm; A sensor end, which comprises a magnetic induction sensor, a photoelectric switch or a laser range finder mounted on the displacement measurement link, and can detect the accurate position of the calibration reference element; A position adjusting mechanism for adjusting the position of each sensor and the calibration reference element during calibration; A signal interface connected to the control component, which automatically triggers the zero-point setting instruction when the reference position is detected; An intelligent diagnosis and control component for collecting various detection data, which predicts the service life by mapping physical detection data through a curtain wall BIM model.

[0024] The intelligent diagnosis and control component comprises: A multi-sensor data acquisition module, which synchronously acquires data in real time through pressure, displacement or temperature and humidity sensors, with a sampling frequency of 100 Hz; A computer vision-based automatic crack identification module for quickly detecting crack positions and comparing crack width and length with historical detection data to determine crack propagation rate; A water seepage path simulation algorithm for predicting the water seepage time, path set, water seepage amount per unit time and high-risk positions of water entering gaps or joints from the outer surface under given rain type, wind pressure and structural details, and outputting a water seepage risk map that can be used for rectification.

[0025] The modules used in the automatic crack identification include: An image acquisition module for acquiring high-quality images of the structure surface at the monitoring points; In this embodiment, high-quality images mean: Based on resolution: the pixel resolution is not less than 1920×1080, or the single-pixel physical size is not greater than a certain value (μm); Based on clarity / sharpness: the center area contrast measured by the modulation transfer function (MTF) is greater than a certain percentage (such as 30%) at a certain frequency.

[0026] Based on signal-to-noise ratio (SNR): the image signal-to-noise ratio is not less than a certain threshold (such as 35 dB) to ensure that details can be distinguished.

[0027] Based on integrity: the image edge is complete without missing, and the target area coverage rate is ≥90%.

[0028] Based on contrast and brightness balance: the gray histogram is distributed within a certain range, and there are no obvious overexposed or underexposed areas.

[0029] An image preprocessing module for marking the crack positions; A model training module for constructing a crack model; A crack detection and positioning module for sending the collected curtain wall images of the site into the trained model training module; A crack feature analysis module for comparing historical detection data to determine the crack propagation rate; A result visualization and storage module for superimposing detection boxes, category labels, width or length values on the original image to generate a crack detection report and upload it to the cloud.

[0030] The process of automatic crack recognition based on computer vision includes the following steps S1, data preparation and collection: use the image acquisition module to shoot or collect images containing cracks, then use the image preprocessing module to ensure that different materials, lighting conditions, and crack shapes are covered, then use professional labeling tools such as LabelImg to draw labeling boxes on the crack positions, and generate annotation files supported by YOLOv5; S2, data preprocessing: enhance the brightness, contrast, rotation, flip, scaling, etc. of the above processed images containing cracks to improve the robustness of the model, then uniformly scale the images to the required input size of YOLOv5, then divide the data set into training set, validation set and test set according to the proportion; S3, model construction and training: input data into the model and monitor the loss curve and mAP indicators in real time; S4, model verification and testing: use the validation set to detect mAP, precision, recall, and F1-score, then check the false positives and false negatives, then randomly select test pictures to check the position of the prediction box and the classification confidence; S5, model deployment and application: can be deployed on servers, embedded devices or mobile terminals, then input the camera video stream into the model to realize online automatic detection of cracks, then output the crack position coordinates, detection confidence, and compare historical detection data to determine the crack propagation rate, and generate a detection report uploaded to the cloud.

[0031] The water seepage path simulation algorithm includes the following steps A1, geometry acquisition and correction: use three-dimensional scanning to import the models of the components in the curtain wall, clean up small features, and repair joint topologies; A2, material or interface property assignment: assign roughness, static or dynamic contact angle, and hysteresis interval to the material surfaces of the components in the curtain wall; A3, region division and physical model selection: External air zone: incompressible or weakly compressible air; Free liquid surface: VOF (volume fraction) or similar multiphase model; Cracks, holes or slits: automatic switching between laminar flow and transitional flow; Turbulence: k-w SST near the gap, gravity and surface tension if necessary wherein in the k-w model: k: represents the turbulent kinetic energy w: represents the specific dissipation rate, i.e. the rate of dissipation of the turbulent kinetic energy The k-w model can better handle the boundary layer and near-wall flow, and capture small-scale flow characteristics more accurately. A4, mesh generation and encryption strategy: globally generate unstructured mesh, initially encrypt in the gap, corner, hole edge, and expected liquid surface area, and set adaptive refinement trigger conditions; A5, working condition boundary mapping: map the drop spectrum parameters of the spray matrix component to the inlet boundary, then give the outside flow and turbulence intensity, and apply the time series of the inner side according to the DP(t) provided by the two-zone pressure difference module, and then set the time anchor point perturbation for multi-channel data alignment; A6, initial conditions and time step control: first, steady-state external wind field, switch to transient multiphase flow, based on CFL adaptive time step, enable smaller time step near the liquid surface; turn on interface compression / amplitude limiting to avoid numerical diffusion; A7, multi-scale coupled solution: first, steady-state external wind field, switch to transient multiphase flow, then based on CFL adaptive time step, enable smaller time step near the liquid surface, turn on interface compression or amplitude limiting to avoid numerical diffusion, i.e. external zone -> gap zone -> porous zone partitioned solution and weakly coupled exchange of fluxes, and enable weak deformation correction for the contact area of the sealing strip / pressure line (adjust the equivalent gap width according to the contact force threshold) In the embodiment, before performing the transient multiphase flow calculation of the water seepage path, a three-dimensional calculation model containing the curtain wall external flow region and the local refined region of the gap is first established. The k-w SST turbulence model is used to solve the steady-state external flow field until the convergence criteria (residual less than 10-5 and wind speed change rate of monitoring points less than 0.1%) are met, and the stable flow velocity distribution and pressure field are obtained. The steady-state result is used as the initial flow field condition for transient solution, which can effectively shorten the development time of transient calculation, improve the calculation efficiency and stability.

[0032] After the steady-state solution is completed, the calculation is switched to the transient multiphase flow solution mode, the fluid phase includes air phase and water phase, and the VOF (Volume of Fluid) interface capturing method is used to track the water-gas interface. The gap and its surrounding area are locally encrypted on the mesh to improve the analysis accuracy of droplet impact, liquid film formation, and seepage into the curtain wall interior.

[0033] In order to balance the calculation accuracy and efficiency, the adaptive time step strategy based on CFL (Courant-Friedrichs-Lewy) number is used in transient calculation: The initial value of the global time step is determined according to the flow velocity in the main flow region and the grid size, so that the CFL number is controlled to be 0.5-1.0; In the region near the liquid surface (the VOF phase fraction gradient is greater than a set threshold value), the time step is automatically reduced, so that the local CFL number is kept below 0.25, so as to improve the interface tracking accuracy; During the calculation process, when the flow field changes sharply (such as the first impact of a droplet on a curtain wall or the appearance of rapid accelerating flow at a gap), the system automatically triggers time step refinement to prevent numerical oscillation.

[0034] S8, connected water body identification and "path diagram" construction: at each time step, the liquid volume fraction field is thresholded and connected domain analysis is performed, the center skeleton of the connected liquid is extracted, the water seepage path diagram is constructed, and the accessibility index of each edge: the penetration rate, the narrowest cross section, the residence time, and the pressure difference stability are calculated; In the present embodiment, in order to realize the visualization and quantitative analysis of the water seepage path, the system identifies and constructs the path of the liquid phase volume fraction field in the calculation domain at each time step, and the specific method is as follows: 1. Liquid phase volume fraction field thresholding The liquid phase volume fraction field (VOF phase fraction, denoted as α(x, y, z, t)) obtained by solving the transient multiphase flow is thresholded:

[0035] Wherein, α_th is the threshold value for determining the existence of liquid phase (for example, 0.5), and M is a binary mask matrix.

[0036] Connected domain analysis The M matrix is labeled by three-dimensional connected domain (in the present embodiment, 6-adjacent is used to ensure the capture of diagonal paths).

[0037] Each connected domain represents an independent water body, and the water body number ID_i is identified by the FloodFill or Union-Find algorithm, and the volume, centroid coordinates, and boundary range are recorded 3. Connectivity determination For each connected domain, it is determined whether it intersects with the outer region boundary B_out of the curtain wall and the inner surface boundary B_in of the detection area at the same time:

[0038] If C_i=1, the water body is considered to be a through water body.

[0039] 4. Water body skeleton extraction The three-dimensional skeletonization is performed on each through water body, and the center path line segment set P={p1, p2,..., p n} of the water body is obtained.

[0040] Node is defined as the intersection of paths or the boundary entry point, and Edge is defined as the continuous path segment between nodes.

[0041] 5. Permeation path graph construction Each water body skeleton is mapped to a graph structure G(V, E): V: node set, containing geometric position and timestamp information; E: edge set, containing path length, minimum cross-sectional area, average liquid phase velocity, residence time, etc.

[0042] Calculate the accessibility index for each edge:

[0043]

[0044]

[0045]

[0046] 6. Output and storage Store the path graph and its attributes associated with the calculation time step t, for subsequent A9 onset determination and permeation calculation.

[0047] This structured path data can be used to generate a permeation risk heat map and a key node analysis table, providing data support for engineering diagnosis and improved design.

[0048] A9, onset determination and permeation calculation: Onset time: the first appearance of a connected path from "curtain outside to detection area"; Instantaneous permeation: Integrate the liquid phase volume fraction flux on the virtual cross-section on the inside; Cumulative permeation / penetration rate: Time integration and can be normalized to unit seam length; A10, adaptive grid and parameter backstepping: Refine the local channel with high accessibility and large error, and use sensitivity analysis to backstep the contact angle, roughness, and porous equivalent parameters in small steps, converging to the interval consistent with the measured value; In this embodiment, "channels with high accessibility and large error" refer to channels in the permeation path graph G(V, E) constructed according to step A8 during the permeation path identification and simulation calculation process, which satisfy the following two conditions: 1. High accessibility The proportion of time that the channel maintains a liquid phase through state during the entire permeation simulation process is high, i.e., the through rate RpR_pRp is higher than the preset threshold (e.g., 0.6), indicating that the path exists effective water flow connection in most calculation time steps.

[0049] Such a channel often locates in the main stream path or the main stream of the branch in the water body skeleton, and has a significant influence on the total amount of seepage and the seepage time.

[0050] 2. Simulation-measurement deviation is large By comparing the simulation results with the measured results, the deviation of the local seepage amount, pressure loss, and seepage time of the channel exceeds the preset allowable range (for example, the deviation rate |E| = 10% or the absolute deviation Δ = 0.051 L / (m 2 ·min).

[0051] The sources of the deviation can include: The material surface roughness is not accurately set; The contact angle and hysteresis parameter do not conform to the actual situation; The local grid resolution is insufficient, and the small geometric features cannot be resolved; The seepage characteristics of the local porous structure are not considered.

[0052] The following processes are performed in the A10 step: A101: Trigger adaptive grid refinement in the channel and a certain range upstream and downstream thereof, so that the minimum cell size is reduced to 30% of the original size, to improve the resolution of local flow details (such as thin liquid film, small vortex, capillary effect); A102: Select the material or interface attribute parameters (roughness k s , static / dynamic contact angle θ S / θ d , porous equivalent permeability K, etc.) related to the channel as variables, sort them based on the sensitivity of the local seepage amount error, and preferentially adjust the parameters with high sensitivity by using small step iteration (the parameter step can be set to 5% of the initial value), until the deviation between the simulation results and the measured values is below the set threshold.

[0053] A103: When the local deviation converges and is no longer above the threshold, apply the corrected parameter set to the local grid area where the channel is located; recalculate the transient flow in the key period to update the seepage path diagram and accessibility index, and ensure the accuracy of subsequent seepage judgment and cumulative seepage calculation.

[0054] A11, reference leakage closed-loop calibration: according to the working condition process, a controllable leakage reference path is turned on for a short time, and then the ΔP-Q measured curve is obtained and compared with the simulation results. If the deviation exceeds the threshold (if the deviation exceeds the preset deviation threshold), the flow conversion coefficient or the porous parameter is automatically corrected, and the key period is recalculated in the A7 step.

[0055] A12, output and criterion: output: water seepage risk thermal map, most likely path, seepage time, cumulative seepage, sensitive parameter ranking, treatment suggestion, such as thickening the joint, changing the drainage aperture / position, adding a water retaining edge, if the seepage time is 15 min or the seepage per unit area is 2.2 L / (m 2 ·min), it is judged as "qualified".

[0056] The foregoing merely illustrates some exemplary embodiments of the present application, and it is needless to say that the described embodiments can be modified in various ways without departing from the spirit and scope of the present application for those skilled in the art. Therefore, the foregoing drawings and descriptions are illustrative in nature, and should not be construed as limiting the scope of the claims of the present application.

Claims

1. A physical performance testing system for large building curtain walls, characterized in that, It includes a dual-zone differential pressure sealing assembly for forming an independent detection zone and a buffer zone on the outside of the curtain wall. The detection zone is sealed to the outer surface of the curtain wall through an expandable sealing chamber, and the buffer zone is surrounded by a flexible membrane and an adsorption strip and isolated from the outside. The spray matrix component includes a nozzle array and a droplet spectrum monitoring unit set on the spray path. The nozzle array is used to simulate different rain patterns according to a preset spatial distribution and time sequence. The droplet spectrum monitoring unit is used to collect droplet diameter and stream density data in real time and feed them back to the control component to adjust the spray mode. An adaptive support mechanism is installed in the detection area of ​​the curtain wall, and micro-force sensors are set at the support contact points to limit the single-point contact force and decouple the support reaction force from the deformation measurement results; The on-site self-calibration component, including the controllable leakage reference path and displacement zero-point calibration unit, performs the reference leakage and zero-point calibration process before the start of detection and before the change of operating conditions, in order to correct sensor drift and environmental interference; The intelligent diagnostic and control component is used to collect various types of test data. It predicts the life cycle by mapping physical test data to establish a curtain wall BIM model.

2. The physical performance testing system for large building curtain walls according to claim 1, characterized in that, The dual-zone differential pressure sealing assembly includes: A sealing cover is used to form a closed testing space on the surface of the curtain wall unit or structure under test. It consists of a rigid frame, an expandable flexible membrane, and a peripheral adsorption strip. The rigid frame surrounds the outer periphery of the expandable flexible membrane, and the peripheral adsorption strip is arranged circumferentially along the outer edge of the expandable flexible membrane and is sealed and fitted to the curtain wall surface. A testing area and a buffer zone are formed on the outer surface of the curtain wall, and a connecting micro valve is provided between the testing area and the buffer zone. The differential pressure generating unit consists of a variable frequency fan and a pressure stabilizing component. The variable frequency fan is located in the detection zone, and the pressure stabilizing component is located in the buffer zone. The variable frequency fan can continuously supply or exhaust air into the sealed space to establish a set positive or negative pressure environment. The pressure stabilizing component includes multiple proportional valves and multiple check valves installed on the pipeline. A differential pressure sensor array is used to monitor the pressure difference between the inside and outside of the curtain wall in real time. Multiple static pressure taps and multiple differential pressure sensors are arranged in the detection area and the buffer area, respectively. Displacement sensor plates are arranged at key positions of the frame for volume change compensation. A flow meter, installed in the inlet or outlet passage, is used to measure the gas flow rate required to maintain a set pressure differential; A leakage reference unit, used to calibrate the zero-point drift of a differential pressure or flow link, includes a known flow resistance line located on the detection side and a controllable microvalve connected to the known flow resistance line; The controller, connected to the differential pressure generating unit, pressure sensor, and flow meter, is used to automatically adjust the fan speed and the opening of the pressure stabilizing device, and record the detection data.

3. The physical performance testing system for large building curtain walls according to claim 1, characterized in that, The drop spectrum monitoring unit includes a monitoring window installed downstream of the spray array and between the surface being measured, serving as a viewing window for drop spectrum imaging and optical acquisition. The monitoring window is equipped with an anti-fog heating film or a micro-airflow purging port to prevent water mist from affecting the imaging quality. The imaging acquisition unit uses a high-speed camera or a high-speed CMOS imaging module to capture the motion trajectory of water droplets at the monitoring window section. Illumination and background light sources, including LED arrays or laser sheet light sources, are arranged on the opposite side of the direction of water droplet movement, so that the water droplets form a high-contrast outline in the image; The data processing and analysis unit, a processor connected to the imaging module, has built-in image segmentation and contour extraction algorithms to calculate the particle size distribution, number density and velocity of water droplets in real time. Among them, the nozzle density in the nozzle array is ≥4 nozzles / m 2 The sprinklers are arranged in a staggered pattern, with each sprinkler having a water output of ≥4L / (m²). 2 (min), the coverage angle of each nozzle is 60-120°.

4. The physical performance testing system for large building curtain walls according to claim 1, characterized in that, The adaptive support mechanism includes: a support frame composed of multiple adjustable telescopic rods, the ends of which are connected to the mounting base via ball joints, enabling the support ends to rotate adaptively in three dimensions; each telescopic rod is equipped with a spring adjustment unit, which is used to change the length to adapt to the measured surface with different curvatures or thicknesses. The contact assembly has a replaceable contact pad installed at the support end. The bottom surface of the contact pad is arc-shaped or spherical, which fits into curved components such as curtain walls. A micro-stroke slider or flexible connecting piece is provided between the contact pad and the end of the telescopic rod to absorb local small displacements and assembly errors. The locking unit, including an adjustable preload suction band or mechanical locking component set on each telescopic rod, can stably press the support end onto the surface being measured. The mechanical decoupling detection unit has a micro-force sensor plate installed on the back of its contact pad to monitor the contact force in real time. When the single-point contact force exceeds the set threshold, the controller automatically adjusts the length or preload of the corresponding telescopic rod to prevent the support structure from applying excessive additional force to the component under test.

5. The physical performance testing system for large building curtain walls according to claim 1, characterized in that, The controllable leakage reference path includes multiple capillary tubes arranged in the detection area, with their two ends connected to the gas path of the detection area and the external environment gas path, respectively, to simulate repeatable micro-leakage. The control valve unit, which is an adjustable valve opening or a solenoid valve or a miniature needle valve, is installed on the reference pipeline and is used to open or close the leakage path according to the control signal; The interface and testing port are equipped with quick connectors at the pipe connection point to facilitate the replacement of flow resistance components.

6. The physical performance testing system for large building curtain walls according to claim 5, characterized in that, The zero-point displacement calibration unit includes: The calibration reference component is a magnetic patch or reflective target fixedly connected to the curtain wall inspection area, with a positional accuracy better than ±0.1mm; The sensor end includes a magnetic induction sensor, photoelectric switch, or laser rangefinder installed on the displacement measurement link, which can detect the precise position of the calibration reference component; Position adjustment mechanism, which is used to adjust the position of each sensor and calibration reference during calibration; The signal interface connects to the control components and automatically triggers a zero-point setting command when a reference position is detected.

7. The physical performance testing system for large building curtain walls according to claim 1, characterized in that, The intelligent diagnostic and control component includes: Data is collected by multiple sensors, and data is collected synchronously in real time through pressure, displacement or temperature and humidity sensors, with a sampling frequency ≥100Hz; Automatic crack identification based on computer vision is used to quickly detect crack locations and compare crack width and length with historical detection data to determine crack propagation rate. The seepage path simulation algorithm, given the rainfall pattern, wind pressure, and structural details, predicts the infiltration time, path set, infiltration volume per unit time, and high-risk locations of water entering cracks or nodes from the outer surface, and outputs a seepage risk map that can be used for remediation.

8. The physical performance testing system for large building curtain walls according to claim 7, characterized in that, The module used for automatic crack identification includes: The image acquisition module is used to acquire high-quality images of the structural surface at the monitoring points; The image preprocessing module is used to mark the location of cracks; The model training module is used to build crack models; The crack detection and location module sends the collected on-site curtain wall images into the trained model training module; The crack feature analysis module is used to compare historical detection data to determine the crack propagation rate. The results visualization and storage module overlays detection boxes, category labels, and width or length values ​​onto the original image to generate a crack detection report, which is then uploaded to the cloud.

9. A large-scale building curtain wall physical performance testing system according to claim 8, characterized in that, The computer vision-based automatic crack identification process includes the following steps: S1. Data preparation and acquisition: Use the image acquisition module to capture or collect images containing cracks, and then use the image preprocessing module to ensure that different materials, lighting conditions and crack shapes are covered. Then use professional annotation tools such as LabelImg or Labelme to draw annotation boxes on the crack locations and generate annotation files supported by YOLOv5. S2. Data preprocessing: Enhance the brightness, contrast, rotation, flipping, and scaling of the processed images containing cracks to improve the robustness of the model. Then, scale the images to the required input size for YOLOv5. Finally, divide the dataset into training, validation, and test sets according to the proportions. S3. Model building and training: Input the data into the model and monitor the loss curve and mAP metrics in real time; S4. Model Validation and Testing: Use the validation set to test mAP, precision, recall, and F1-score, then check for false positives and false negatives, and then randomly select test images to check the predicted bounding box positions and classification confidence. S5. Model Deployment and Application: It can be deployed on servers, embedded devices, or mobile devices. The camera video stream is then input into the model to achieve online automatic crack detection. The crack location coordinates and detection confidence are then output. The crack propagation rate is determined by comparing with historical detection data, and a detection report is generated and uploaded to the cloud.

10. A physical performance testing system for large building curtain walls according to claim 7, characterized in that, The seepage path simulation algorithm includes the following steps. A1. Geometric Acquisition and Correction: Import the models of various components inside the curtain wall using 3D scanning, and perform small feature cleaning and joint topology repair; A2. Assigning material or interface properties: Specify the roughness, static or dynamic contact angle, and hysteresis range for the material surfaces of each component within the curtain wall; A3. Regional Division and Physical Model Selection: A4. Mesh generation and refinement strategy: Generate unstructured meshes globally, initially refine the meshes at seams, corners, hole edges, and expected liquid surface areas, and set adaptive refinement trigger conditions; A5. Working condition boundary mapping: Map the droplet spectrum parameters of the spray matrix component to the inlet boundary, then give the outer inflow and turbulence intensity, apply the time series to the inner side according to ΔP(t) provided by the dual-zone pressure difference module, and then set the time anchor point perturbation for multi-channel data alignment. A6. Initial conditions and time step control: First, calculate the external wind field in steady state, then switch to transient multiphase flow, and use CFL adaptive time step. A smaller time step is used near the liquid surface; enable interface compression / limiting to avoid numerical diffusion. A7. Multi-scale coupled solution: First, calculate the external wind field in steady state, then switch to transient multiphase flow, and then use a smaller time step near the liquid surface based on CFL adaptive time step, and enable interface compression or limiting to avoid numerical diffusion. A8. Identification of Connected Water Bodies and Construction of "Path Map": At each time step, thresholding and connected domain analysis are performed on the liquid volume fraction field to extract the central skeleton of connected liquids, construct the seepage path map, and calculate the accessibility index of each edge: penetration rate, narrowest cross section, residence time, and pressure difference stability. A9. Determination of seepage initiation and calculation of seepage volume: The moment of seepage initiation: the first appearance of a connection path from "outer side of the curtain wall to the detection area"; Instantaneous infiltration rate: Integral of liquid volume fraction flux over the inner virtual cross section; Cumulative seepage volume / seepage rate: obtained by integrating over time and normalized to a unit crack length; A10. Adaptive mesh and parameter backtracking: The channels with high accessibility but large errors are further refined locally, and then sensitivity analysis is used to backtrack the contact angle, roughness, and porosity equivalent parameters in small steps, converging to the range consistent with the actual measurement. A11. Reference Leakage Closed-Loop Calibration: The controllable leakage reference path is briefly connected according to the working condition process. The measured curve of ΔP–Q is obtained and then compared with the simulation results. If the deviation exceeds the threshold, the flow conversion coefficient or the orifice parameter is automatically corrected and the critical period is recalculated in S7. A12. Output and Criteria: Output: Seepage risk heat map, most likely path, infiltration time, cumulative seepage volume, ranking of sensitive parameters, and remediation suggestions, such as thickening the sealant joint, changing the drainage hole diameter / location, and adding a water-retaining edge. If the infiltration time is ≤15min or the seepage volume per unit area is >2.0±0.2L / (m²), the output will be considered. 2 (min), is judged as "unqualified".

Images