Decoration project tapestry brick hollowing glue injection repair optimization control method and system

Abstract

The invention discloses a decoration engineering tapestry brick hollowing glue injection repair optimization control method and system, and particularly relates to the technical field of decoration engineering.The method comprises the steps that firstly, coordinate mapping is conducted on a facade CAD / BIM map, sound waves, infrared thermal images and laser ranging data in a unified time-space reference, and a hollowing judgment list is generated in real time; then establishing a hollowing model through three-dimensional voxelization, planning a mother hole along the lowest potential energy line, and checking and patching the hole in combination with a radial reachable radius, brick joint offset and appearance limitation; in the multi-stage glue injection process, the system constructs a millisecond-level closed loop by using pressure-flow velocity-flow three-path data, adaptive parameter learning and compensation are implemented according to volume deviation, flow velocity deviation and residual pressure slope, smooth connection of a primer film period, a blind cavity period and a compact period is ensured, and finally glue injection repair with high coverage rate, low residual hollowing and minimum veneer damage is realized.

Description

Technical Field

[0001] This invention relates to the field of decoration engineering technology, and more specifically, to an optimized control method and system for repairing hollow tiles in decoration engineering. Background Technology

[0002] Facing bricks are prone to hollowing defects after long-term service or due to insufficient initial construction quality. If not treated in time, rainwater seepage and freeze-thaw expansion and contraction will cause the surface to fall off and endanger pedestrian safety. The current mainstream repair method is to manually judge the range of hollowing based on the sound of tapping, and then inject adhesive in one go according to the "equidistant hole layout - constant pressure injection" approach.

[0003] This approach has three limitations: First, acoustic auscultation is highly subjective, and the boundaries of voids are easily masked by high-frequency noise or echoes from the surface structure. Second, constant-pressure injection ignores the coupling effect of cavity thickness, viscosity, and ambient temperature on the flow pattern, which can lead to blind pressure application before the primer film has formed, resulting in new cracks or residual blind cavities on the brick back. Third, during construction, sensor signals for pressure, flow rate, and volume are often transmitted via wireless links, making it difficult to identify false anomalies caused by jitter, packet loss, and timestamp misalignment, leading to delayed compensation commands. Although some scholars have proposed composite detection schemes such as superimposed infrared thermography, acoustic arrays, or ultrasonic scanning in recent years, the injection execution stage still lacks closed-loop control logic that can adjust the pulse beat, pause duration, and target pressure in real time based on feedback. Furthermore, a three-dimensional drilling planning system deeply integrated with CAD / BIM models has not been established, thus limiting the overall repair efficiency and quality.

[0004] To address the above shortcomings, there is an urgent need for an intelligent repair solution that integrates multi-source sensing, 3D modeling, dynamic hole layout, and adaptive closed-loop adhesive injection to improve the accuracy of hollow area identification, reduce adhesive injection blind spots, and achieve rapid, safe, and traceable treatment of building facades in service. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide an optimized control method and system for repairing hollow tiles in decoration projects by injecting adhesive to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The optimized control method for repairing hollow tiles in decoration projects includes the following steps: First, a spatial coordinate system is established in the CAD facade environment using a brick grid. Each facing brick is numbered according to the rule of left to right and bottom to top and then imported into the detection software. Subsequently, at the construction site, acoustic wave detection technology and infrared thermal imaging technology are combined to extract acoustic wave and thermal image data and generate a three-dimensional distribution model of hollow areas. Within the three-dimensional distribution model of the hollow area, the lowest potential energy line of the hollow area is automatically identified. Based on the formula of the extreme value of the hollow thickness and the radial achievable radius, the first stage of glue injection holes are arranged, and the offset optimization is performed in combination with the brick surface grouting position. After the first stage of glue injection is completed and the glue has initially solidified, the second stage of glue injection holes and the third stage of glue injection holes are adaptively generated based on the infrared or ultrasonic scanning results. The three-dimensional coordinates and glue injection hole number of each glue injection hole are recorded. Automatically extract the number and three-dimensional coordinates of each dispensing hole; based on the real-time detected viscosity of the colloid and the on-site temperature, match the corresponding volume tolerance, flow rate-pressure coupling coefficient and wireless integrity lower limit from three preset templates: "standard, high viscosity and high temperature, low viscosity and low temperature"; capture the dispensing pump start-up edge as the cycle origin, uniformly add nanosecond-level timestamps to the three signals of pressure, instantaneous flow rate and cumulative flow, and generate an "injection-pause" pulse sequence; calculate the weights according to the link jitter factor, synchronization quality factor and stage factor, and use the weights to scale the initial backtracking time to construct an alignment window, and interpolate and smooth the gap data; Within each pulse cycle, the volume deviation, flow velocity deviation, residual pressure slope, and variance are calculated. Based on the time-series drift assumption, flow velocity-pressure coupling verification, and residual pressure trend hierarchical judgment of error causes, a compensation instruction containing adjustment amounts for start-up duration, target pressure, or pause duration is generated. The parameters of the next pulse are modified in real time according to the compensation instruction, and a verification pulse is initiated.

[0007] In a preferred embodiment, the lowest potential energy line is the three-dimensional fold line inside the hollow area where the adhesive flows naturally under the action of gravity and eventually converges to form a "bottom adhesive film", that is, the bottom boundary of the hollow area.

[0008] In a preferred embodiment, the radially achievable radius can be calculated using the following formula: Where ΔP is the injection design pressure, h is the average thickness of the hollow cavity, μ is the dynamic viscosity of the adhesive, and k is the field experience coefficient.

[0009] In a preferred embodiment, the weights are calculated by weighted summation based on the link jitter factor, synchronization quality factor, and stage factor.

[0010] In a preferred embodiment, the link jitter factor is first calculated by taking the average of the actual sampling periods of the three sensors in the original records within the most recent unit as the average sampling interval; then, the segment with the longest continuous packet loss time within the same time period is found, which is called the maximum packet loss interval; finally, the maximum packet loss interval is divided by three times the average sampling interval to obtain the dimensionless ratio, and the dimensionless ratio is used as the link jitter factor; if the dimensionless ratio exceeds two, it is recorded as two.

[0011] In a preferred embodiment, the synchronization quality factor is used to measure whether the overall patching of the wireless link over multiple pulse cycles has approached the system's minimum integrity requirement.

[0012] In a preferred embodiment, the stage factor is automatically assigned a value according to the different stages it is in.

[0013] In a preferred embodiment, the optimized control system for repairing hollow tiles in decorative engineering includes the following modules: The data acquisition and coordinate mapping module is used to establish a spatial coordinate system with brick grid in the CAD facade environment, and to number each facing brick according to the rule of left to right and bottom to top; The 3D hollow modeling and hole planning module is used to extract acoustic and thermal image data and generate a 3D distribution model of hollow areas; and based on the 3D distribution model of hollow areas, it generates injection holes for each stage and records the 3D coordinates and injection hole number of each injection hole. The multi-stage dispensing scheduling and execution module writes the pulse cycle and three-stage injection formula into the task sequence according to the hole position table, issues pump start, pump stop and stage switching instructions for each hole, and collects pressure, flow rate and flow rate feedback at the end of each pulse. The adaptive closed-loop control and parameter learning module calculates volume deviation, flow rate deviation and residual pressure slope in real time during the dispensing process, and uses the three-stage volume tolerance and flow rate-pressure coupling coefficient as the metric benchmark to sequentially complete wireless frame loss verification, coupling blockage verification and residual pressure trend diagnosis; generates compensation instructions that include adjustment of start-up duration, target pressure or pause duration; and modifies the next pulse parameters and initiates a verification pulse in real time according to the compensation instructions.

[0014] In a preferred embodiment, the adaptive closed-loop control and parameter learning module includes a window adjustment module, which is used to dynamically construct an aligned window.

[0015] The technical effects and advantages of this invention are as follows: By integrating CAD / BIM facade models, acoustic wave detection, infrared thermal imaging, and laser ranging data within a unified time and space reference, this invention achieves high-precision identification and three-dimensional modeling of hollow locations in facing bricks. During the hollow repair process, guided by the lowest potential energy line and combined with radial reachable radius and grout offset optimization, the drilling position is scientifically planned, so that the adhesive is injected and diffused evenly along the optimal path, significantly reducing blind spots and omissions. By controlling the primer film stage, blind cavity filling stage, and sealing hole stage in stages, and dynamically adjusting the beat, pauses, and pressure based on real-time closed-loop feedback of pressure, flow rate, and flow data, high coverage, low residual voids, and minimal surface damage are achieved. At the same time, a weighting mechanism formed by link jitter factor, synchronization quality factor, and stage factor is introduced to fill and smooth data gaps, ensuring the integrity and real-time nature of feedback data. Overall, the level of intelligence and precision of the repair is significantly improved, construction efficiency and repair quality are improved, and subsequent safety hazards are reduced. Attached Figure Description

[0016] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings; Figure 1 This is a flowchart illustrating the optimized control method for repairing hollow tiles in decorative engineering according to the present invention. Figure 2 This is a flowchart illustrating the real-time closed-loop sub-process of the present invention. Figure 3 This is a schematic diagram of the structure of the optimized control system for repairing hollow tiles in decoration engineering according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1: The present invention provides an optimized control method for repairing hollow tiles in interior decoration projects using adhesive injection, such as... Figure 1 As shown, it includes the following steps: First, a spatial coordinate system needs to be established for the inspection. Take out the previously prepared elevation drawings. If no drawings are available, use a tape measure or laser rangefinder to measure the height of each floor, the dimensions of each brick, and the location of the grout lines. Then, draw a detailed brick grid diagram in CAD or Revit. Number the bricks, for example, label each brick from left to right and from bottom to top as B001, B002, B003, etc. Import this elevation grid diagram into the software on your computer for subsequent real-time positioning.

[0019] Before commencing testing, a work platform or suspended basket should be erected at the construction site to ensure the testing instruments can safely contact the surface of the facing bricks. The digital acoustic sensor should then be securely attached to the surface of the brick using a flexible clip or suction cup, ensuring full contact between the sensor and the wall without looseness or gaps to avoid signal interference. The acoustic sensor is connected to a portable data acquisition system via a signal cable and then to a laptop computer. The computer should be pre-loaded with acoustic wave acquisition and analysis software. This software should possess real-time waveform acquisition, spectrum analysis, and automatic classification and recognition algorithms, capable of automatically distinguishing the acoustic differences between solid and hollow bricks using a machine learning model.

[0020] During formal testing, the operator uses a rubber mallet to strike designated test points on each tile with constant force, typically at the center and four corners of the tile surface. The sound waves generated by the strikes are collected by attached sound sensors and transmitted in real time to the software interface. The software automatically analyzes the waveform of each strike, extracting waveform feature values ​​through a trained hollow sound wave model, including the dominant frequency range, energy distribution, pulse decay time, number of peaks, and waveform envelope changes. This data is then compared with the stored waveform features of solid and hollow tiles in the database to automatically determine whether the tile is solid or hollow. The determination result is displayed instantly on the software interface, with the tile number marked in a different color. Simultaneously, all original waveform data and determination results are automatically stored in the system database to ensure traceability.

[0021] After inspecting one wall, the software imports all brick inspection results into the facade CAD or BIM facade model and generates an inspection diagram through automated procedures. In this diagram, all hollow bricks are highlighted in red, solid bricks are highlighted in green, and bricks with a probability between the two are highlighted in yellow, indicating that further observation is needed. Based on the hollow brick detection results of adjacent bricks, the software automatically delineates the boundaries of the hollow areas through cluster analysis or image segmentation algorithms, forming closed polygon outlines, and automatically calculates the area, perimeter, and location coordinates of the hollow areas.

[0022] To further verify the shape of the hollow area, an infrared thermal imager should be used to scan the detection area simultaneously. The thermal imager should have a thermal sensitivity with a temperature measurement accuracy within 0.05℃ and be connected to the accompanying analysis software. During operation, the thermal imager should be moved slowly and evenly from above or below the facade to record the thermal distribution image of the facade. The thermal imager software will automatically identify areas of abnormal temperature through image processing algorithms. Because of the air layer behind the hollow area, the heat conduction is significantly different from the surrounding area, which will appear as hot spots or cold spots on the thermal image. The detection software can automatically delineate the closed boundary lines of the temperature abnormal area in the infrared image and output it as a vector graphic format.

[0023] After completing the thermal imaging inspection, the vector image exported by the thermal imaging analysis software should be imported into the CAD or BIM model. The system will automatically compare the hollow shape generated by the acoustic wave detection with the infrared thermal image recognition result. If the boundary heights of the two are consistent, the accurate range of the hollow can be directly confirmed. If there are local differences, the system will use an overlap algorithm to take the intersection or union of the two to generate the corrected final hollow area boundary.

[0024] All test results should be automatically generated by the software to create a three-dimensional hollow distribution model. The report should include a list of hollow bricks, polygon coordinates, area, perimeter of the hollow area, a schematic diagram of the hollow distribution on the facade, and the original acoustic data and thermal images. The three-dimensional hollow distribution model should be output in the form of an electronic PDF or CAD file.

[0025] After accurately identifying the hollow areas, a drilling point layout plan is formulated based on the three-dimensional hollow distribution model.

[0026] Specifically, firstly, based on the three-dimensional hollow distribution model, the first stage of drilling points is generated. In the CAD or BIM environment, the lowest potential energy line of the hollow area is identified (the three-dimensional broken line inside the hollow area where the adhesive flows naturally under the action of gravity and eventually converges to form the "bottom adhesive film", i.e. the bottom boundary of the hollow area), and preliminary hole positioning is carried out along this potential energy line.

[0027] The rules for the minimum potential energy line hole layout method are as follows: In each segment (e.g., the length is 0.5m to 1.0m) divided along the minimum potential energy line, the extreme point of the hollow thickness must first be identified. If the difference between the extreme value of the hollow thickness and the average value of the surrounding area is greater than the system's preset threshold Tn, then the point where the perpendicular line from that point to the minimum potential energy line intersects should be the first to be set as the glue injection hole, because the glue will flow to the thickest hollow area during the glue injection process, and the thickest hollow point is often the location where the blind cavity is most likely to form later.

[0028] If a certain section has a relatively uniform thickness distribution, and the thickness fluctuation does not exceed the system's preset threshold Bd, then the geometric center of that section should be selected as the main hole placement point to ensure that the adhesive diffuses evenly to both sides and reduce the risk of dead zones. This step involves using a 3D model to determine the first batch of the most core and critical adhesive injection hole positions.

[0029] After completing the initial hole positioning, the hole spacing should be technically checked according to the radial radius R of the adhesive to ensure that the cavity areas covered by adjacent injection holes can effectively overlap, thereby preventing the occurrence of uninjected dead zones.

[0030] The radial reachable radius R can be calculated using the following empirical formula: Where ΔP is the design pressure for glue injection (Pa), h is the average thickness of the hollow cavity (m), μ is the dynamic viscosity of the glue (Pa·s), and k is the field experience coefficient.

[0031] For example, if the measured parameters are ΔP = 0.15 MPa, h = 5 mm, μ = 0.20 Pa·s, and k = 0.28... Therefore, the calculated value is R≈0.15m. Thus, the maximum distance L between adjacent master holes should not exceed 2R, i.e., L≤0.30m, to ensure effective overlap and coverage of the adhesive injection wavefront within the cavity. If the maximum distance L between adjacent master holes in the initial hole layout in the first step is greater than 2R, then additional holes must be made between these two adjacent master holes; the specific number of additional holes is determined by L.

[0032] This step involves radial flow verification of the initial hole layout results to determine whether additional holes need to be added in areas with large void thickness.

[0033] After checking the radial reachability, the third step requires adjusting the hole placement based on the grid pattern of the facing bricks. Since the optimal grouting position often doesn't perfectly coincide with the grout intersection, to maintain the aesthetics of the finish and facilitate future repairs, it's stipulated that if the calculated offset of the theoretical hole position from the nearest grout intersection is within the system's preset range, the hole position can be moved to the grout intersection to reduce surface damage and repair difficulty. If the offset exceeds the system's preset range, priority should still be given to ensuring the hole position is located at the extreme value of the hollow thickness or the geometric center, and drilling on the brick surface should be considered if necessary. This step optimizes the grout alignment and corrects the offset of the hole placement results, ensuring effective grout coverage while also considering the surface appearance and ease of repair.

[0034] Following the aforementioned logical rules, corresponding first-stage drilling points are generated on the CAD image for workers to reference during drilling. The first stage of adhesive injection is then performed, and the adhesive injection control process for this stage calls the real-time closed-loop sub-process described below.

[0035] Approximately two hours after the first stage of adhesive injection is completed and has initially cured, infrared thermography or ultrasonic testing should be used again to scan the repaired area and identify the blind cavity areas cut out due to the formation of the substrate film during the first stage of adhesive injection. The test results should be imported into a CAD or BIM system and automatically overlaid into the three-dimensional void distribution model generated after the first stage of adhesive injection to form a corrected blind cavity distribution map.

[0036] The second-stage injection hole must be precisely located at the geometric center of these blind cavities. If the distance between the geometric center of the blind cavity and the edge seam is less than the system's preset range, and the edge seam is also within the blind cavity range, then the point closest to the geometric center of the blind cavity and on the edge seam is used as the second-stage injection hole; otherwise, the geometric center of the blind cavity is used as the second-stage injection hole.

[0037] The locations of the second-stage injection holes are displayed on the CAD image for workers to refer to when drilling. The second-stage injection process is then performed, and the injection control flow for this stage calls the real-time closed-loop sub-process described below.

[0038] The voids remaining after the first two stages are often small and isolated. After completing the second stage of adhesive injection and waiting for the adhesive to initially cure, a re-inspection must be conducted on-site to confirm the existence of any tiny cavities that were not filled by the first two injections—these are the residual voids that need to be addressed in the third stage. These cavities are usually very small; refined thermal imaging is used to display the residual voids on a CAD image for workers to reference when drilling. The third stage of adhesive injection then proceeds, and the adhesive injection control process in this stage calls upon the real-time closed-loop sub-process described below.

[0039] The first stage is the base film formation period, the initial stage of adhesive injection. The adhesive first spreads into a thin film between the brick back and the base layer. At this time, the gaps are large and the resistance is small. The required flow rate is high and the pressure is low. The purpose is to quickly "base" the material and provide a uniform wetting interface for subsequent filling.

[0040] The second stage is the blind cavity filling period. The colloid continues to advance into the deeper blind cavity and begins to rise. The hollow volume is being largely occupied, and the flow resistance increases significantly. The control strategy shifts to "medium flow rate and medium pressure," focusing on monitoring for signs of blockage such as "decreased flow rate and increased pressure." At the same time, pauses are used to allow the colloid to permeate and release air on its own.

[0041] The third stage is the sealing stage, where the cavity is basically filled, leaving only tiny pores and residual gas. At this stage, a low flow rate and stable high pressure are required to allow the colloid to slowly compact and seal the seepage channel. Emphasis is placed on monitoring the flow meter's zero drift and residual pressure decay curves. Sealing can only proceed after confirming that the volume has been replenished and the pressure dissipation conforms to the exponential model.

[0042] After completing the drilling stages, when the glue injection operation starts, the system issues a command to control the glue injection machine to continuously inject glue into the selected injection hole, maintaining the initially set injection time (e.g., 5 seconds). At the same time, it collects data from the injection pressure, injection flow rate, and flow sensor feedback in real time. After the set injection time is reached, the system immediately closes the solenoid valve of the glue injection machine, stops the glue injection, and enters a brief pause phase (e.g., 2 seconds).

[0043] The real-time closed-loop subprocess is as follows: Figure 2 As shown, the specific steps are as follows: The process involves initializing the hole group and memorizing the stages. After importing the CAD image, the software automatically extracts the number, 3D coordinates, and distance between each injection hole and writes them into the database. Based on the colloid viscosity and the ambient temperature, a parameter set is selected from three templates: "standard," "high viscosity high temperature," and "low viscosity low temperature," to obtain different volume tolerances εV1, εV2, and εV3, flow rate-pressure coupling coefficient kμ, and wireless integrity lower limit Rcom. These three templates are pre-input templates, each corresponding to different volume tolerances εV1, εV2, and εV3, flow rate-pressure coupling coefficient kμ, and wireless integrity lower limit Rcom based on different colloid viscosities and ambient temperatures.

[0044] After the acquisition card captures the start edge of the dispensing pump (i.e., the valve opening signal, which is also the dispensing start signal), it sets the edge as the origin of the cycle. At the same time, it adds a nanosecond-level timestamp to the three sampling channels of pressure, flow rate, and flow rate and writes it into the circular buffer, so that the wired and wireless data are completely aligned.

[0045] Simultaneously with detecting the origin of the cycle, a "start dispensing" command is issued to the dispensing machine: the solenoid valve opens, and the pump continuously injects the adhesive at a set displacement for a certain period of time (e.g., 5 seconds); during this period, pressure, instantaneous flow rate, and cumulative flow data are collected in real time at fixed intervals. At the end of the five-second countdown, the system immediately closes the solenoid valve to stop dispensing and enters a short pause (e.g., 2 seconds) to continuously record the static pressure decay curve. This forms a complete "5-second injection + 2-second pause" pulse cycle.

[0046] After the aforementioned seven-second pulse ends, the backtracking time is determined based on the time-determination process, with the beat origin as the center. Build an alignment window The system locates the nearest valid sample for the three curves. If a gap appears in the window, it is filled by linear interpolation between the two nearest points; when the gap span exceeds the device threshold, a sliding weighted average smoothing slope is added. The filled sample, along with the gap span, hole number, and nearest neighbor hole distance, is written to the cache.

[0047] The backtracking time determination process steps are as follows: First, determine the link jitter factor and the synchronization quality factor.

[0048] The link jitter factor measures the severity of the worst packet loss that occurs in the wireless link within a very short period. The method involves first calculating the average of the actual sampling periods of the three sensors from the most recent 100 milliseconds of raw data, which is then used as the average sampling interval. Next, the longest consecutive packet loss interval within the same time period is identified, termed the maximum packet loss interval. Finally, the maximum packet loss interval is divided by three times the average sampling interval to obtain the dimensionless ratio. A ratio close to one indicates mild jitter, while a larger value indicates more pronounced jitter. However, to avoid overweighting a single extreme anomaly, a result exceeding two is counted as two in the algorithm.

[0049] The synchronization quality factor is used to measure whether the overall patching performance of the wireless link over multiple pulse cycles has approached the system's minimum integrity requirement. Specifically, it involves recording the patching rate of the last five pulses (the ratio of interpolated patching points within the alignment window to the theoretical number of samples), taking their arithmetic average to obtain the short-term average patching rate, and then dividing this average by the system's preset lower limit for wireless integrity. If the result is greater than one, it indicates that the patching rate has reached or exceeded the minimum requirement in the short term, and the synchronization quality is declining; if the result is less than one, it indicates that the link is still within an acceptable range. Similarly, to prevent excessive weighting, if this ratio exceeds two, it is used as two.

[0050] The combined weight W is calculated from the link jitter factor, synchronization quality factor, and stage factor; the formula is as follows: W=α*f1+βf2+γf3; where f1, f2, and f3 represent the link jitter factor, synchronization quality factor, and stage factor, respectively, and α, β, and γ are the weights of the link jitter factor, synchronization quality factor, and stage factor.

[0051] After determining the composition weights, the initial backtrack time (e.g., 5ms) is scaled according to the weight values ​​to obtain the backtrack time. ; Construct an alignment window .

[0052] The stage factor is automatically assigned a value according to the stage it is in, for example, 1.2 for the first stage, 1 for the second stage, and 0.8 for the third stage.

[0053] Within a single pulse cycle, the theoretical volume is calculated based on the pump displacement and operating duration, and the volume deviation ΔV is obtained by integrating the difference with the flow meter. The velocity deviation Δq is obtained by subtracting the measured velocity from the set instantaneous velocity. Logarithmic linear regression is performed on the pressure during the pause to obtain the residual pressure slope κP and variance φP. Within each sampling cycle, the system uses kμ to convert the real-time velocity into predicted pressure, and then subtracts it from the measured pressure to generate a residual sequence. The slope and variance of the residuals together determine the triggering timing of "model divergence" or "upward concave anomaly".

[0054] Six key quantities are identified: First, Rgap represents the proportion of samples that need to be "filled in" within the current alignment window, directly quantifying the severity of the synchronization gap; second, Rcom is the minimum requirement for communication integrity, used to determine whether the wireless link is in an acceptable state; third and fourth, βV and βq are the excess multiples of volume error ΔV and flow velocity error Δq relative to the current stage tolerance (εV1, εV2, εV3), respectively; fifth, σqp reflects the consistency of the first flow velocity peak and pressure trough in sign after the valve is opened, used to capture the typical blockage feature of "flow velocity decreases while pressure increases". If the consistency is met, it is assigned a value of 1, otherwise it is assigned a value of 0; sixth is a static pressure decay curve and the recursive residual of the exponential model, where κP is the slope of the residual over time and φP is the variance of the residual. The two together characterize whether the "residual pressure dissipation is abnormal".

[0055] The discrimination logic executes three layers sequentially. The first layer proposes a "timing drift hypothesis": if the patching rate has reached or exceeded the wireless integrity requirement (Rgap≥Rcom), while the volume and flow rate simultaneously exceed the tolerance (βV, βq≥1), the most likely cause is a "false anomaly" caused by wireless frame loss or timestamp misalignment. At this point, the system does not rush to reduce the pump pressure, but instead issues a compensation instruction to "adjust the time axis"; the confidence level is multiplied by min(βV,βq) with Rgap / Rcom, the purpose of which is to make the compensation action more decisive when the deviation amplitude is large and the patching rate is also high. If this condition is not met, it means that the synchronization quality is acceptable or only one deviation exceeds the limit, and the algorithm enters the second layer of coupling verification.

[0056] The second layer examines the "velocity peak-pressure trough" coupling by combining the physical characteristics of different stages: In the first stage, the colloid flows smoothly, and it is considered normal as long as σqp equals 1; In the second stage, the cavity resistance increases. If the flow velocity drops significantly while the pressure still soars, and the volume and flow velocity deviations are significant, it is determined that a blockage has occurred. The compensation method is to extend the pause and lower the target pressure; After entering the third stage, if the pressure trough value at the end of the pause drops abnormally fast and ΔV has turned negative, it indicates that the flow meter zero drift has caused the integral to be too small. The system issues an online zeroing command and calculates the confidence level using the ratio of κP to φP. The steeper the slope and the smaller the fluctuation, the more determined the zeroing action is.

[0057] If the first two layers cannot provide a conclusion, the third layer calculates the "overall trend of residual pressure": In the first stage, if κP continues to rise positively, it indicates that the model deviates from the actual measurement on a macroscopic level, i.e., "model divergence," and the compensation method is to adjust the start-up time; In the second stage, if the residual curve shows an upward concave shape while Δq continues to accumulate, it is judged as an "abnormal upward concave shape," and the pause needs to be lengthened and the pressure reduced; In the third stage, if the residual is almost unchanged and ΔV has approached zero, it indicates that the system is operating stably, and the existing parameters can be maintained.

[0058] The discriminator then encapsulates the identified error cause, the calculated confidence level, and the corresponding compensation instruction into a structured record and writes it into the stage error table. After reading this record, the closed-loop control module performs real-time corrections to the duration, pressure, or pause before the next pulse is initiated.

[0059] The controller adjusts the start duration, target pressure, or pause of the next pulse in real time according to the compensation command, or performs flow meter zero-point calibration; then it runs a rapid verification pulse to verify the compensation effect. If ΔV and Δq decrease by no less than the system threshold Th (e.g., 30%), convergence is determined; otherwise, compensation is increased round by round until an alarm is triggered. If the coupling orifice pressure rises abnormally after convergence, the system automatically starts the pump at the coupling orifice with a delay of one to two seconds.

[0060] Example 2: The design of the optimized control system for repairing hollow tiles in decoration projects according to the present invention is based on the method in Example 1, specifically as follows: Figure 3 The following modules are shown: The data acquisition and coordinate mapping module receives CAD elevation drawings, acoustic waveforms, and infrared thermal images within a single time reference and elevation coordinate system. First, it writes the three-dimensional coordinates and center distance of the brick joints of each facing brick into the database through image-grid analysis. Then, it adds millisecond-level timestamps to the two types of real-time detection data and maps them to the same grid node. The system simultaneously performs acoustic-infrared joint feature extraction and threshold discrimination, directly generating a list of hollow areas to provide a precise basis for subsequent processes.

[0061] The 3D hollow modeling and hole planning module quickly generates a 3D hollow model after reading the hollow judgment list; the system extracts candidate points of the mother hole along the lowest potential energy line, and filters each hole according to the radial reachable radius verification, brick joint offset correction and appearance restriction, and finally outputs a complete hole location table with attributes of the base film period, blind cavity period and dense period.

[0062] The multi-stage dispensing scheduling and execution module writes the pulse cycle and three-stage injection formula into the task sequence according to the hole position table, issues pump start, pump stop and stage switching instructions for each hole, and collects pressure, flow rate and flow rate feedback at the end of each pulse, and performs millisecond-level handshake with the closed-loop control module; to ensure smooth process connection between the base film stage, blind cavity stage and densification stage.

[0063] The adaptive closed-loop control and parameter learning module calculates volume deviation, flow rate deviation and residual pressure slope in real time during the dispensing process. It uses the three-stage volume tolerance and flow rate-pressure coupling coefficient as the dimensional benchmark to complete wireless frame loss verification, coupling blockage verification and residual pressure trend diagnosis in sequence. When model divergence or concave abnormality is detected, the pump pressure is adjusted, pause or time axis compensation is performed.

[0064] The adaptive closed-loop control and parameter learning module includes a window adjustment module, which is used to dynamically construct an aligned window.

[0065] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0066] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0067] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0068] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0069] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for optimizing the repair of hollowing and glue injection of decorative tiles in finishing works, characterized in that, Includes the following steps: First, a spatial coordinate system is established in the CAD facade environment using a brick grid. Each facing brick is numbered according to the rule of left to right and bottom to top and then imported into the detection software. Subsequently, at the construction site, acoustic wave detection technology and infrared thermal imaging technology are combined to extract acoustic wave and thermal image data and generate a three-dimensional distribution model of hollow areas. Within the three-dimensional distribution model of the hollow area, the lowest potential energy line of the hollow area is automatically identified. Based on the formula of the extreme value of the hollow thickness and the radial achievable radius, the first stage of glue injection holes are arranged, and the offset optimization is performed in combination with the brick surface grouting position. After the first stage of glue injection is completed and the glue has initially solidified, the second stage of glue injection holes and the third stage of glue injection holes are adaptively generated based on the infrared or ultrasonic scanning results. The three-dimensional coordinates and glue injection hole number of each glue injection hole are recorded. Automatically extract the number and three-dimensional coordinates of each dispensing hole; based on the real-time detected viscosity of the colloid and the on-site temperature, match the corresponding volume tolerance, flow rate-pressure coupling coefficient and wireless integrity lower limit from three preset templates: "standard, high viscosity and high temperature, low viscosity and low temperature"; capture the dispensing pump start-up edge as the cycle origin, uniformly add nanosecond-level timestamps to the three signals of pressure, instantaneous flow rate and cumulative flow, and generate an "injection-pause" pulse sequence; calculate the weight according to the link jitter factor, synchronization quality factor and stage factor, and use the weight to scale the initial backtracking time to construct an alignment window, and interpolate and smooth the gap data; Within each pulse cycle, the volume deviation, flow velocity deviation, residual pressure slope, and variance are calculated. Based on the time-series drift assumption, flow velocity-pressure coupling verification, and residual pressure trend hierarchical judgment of error causes, a compensation instruction containing adjustment amounts for start-up duration, target pressure, or pause duration is generated. The parameters of the next pulse are modified in real time according to the compensation instruction, and a verification pulse is initiated.

2. The method according to claim 1, wherein the method is characterized by: The lowest potential energy line is the three-dimensional folded line inside the hollow area, where the adhesive flows naturally under the action of gravity and eventually converges to form the "bottom adhesive film," which is the bottom boundary of the hollow area.

3. The optimized control method for repairing hollow tiles in decorative engineering according to claim 1, characterized in that: The achievable radial radius can be calculated using the following formula: Where ΔP is the injection design pressure, h is the average thickness of the hollow cavity, μ is the dynamic viscosity of the adhesive, and k is the field experience coefficient.

4. The optimized control method for repairing hollow tiles in decorative engineering according to claim 1, characterized in that: The weights are calculated by weighted summation of the link jitter factor, synchronization quality factor, and stage factor.

5. The optimized control method for repairing hollow tiles in decorative engineering according to claim 4, characterized in that: The link jitter factor is first calculated by taking the average of the actual sampling periods of the three sensors in the most recent unit of the original records, which is used as the average sampling interval; then, the segment with the longest continuous packet loss time in the same time period is found, which is called the maximum packet loss interval; finally, the maximum packet loss interval is divided by three times the average sampling interval to obtain the dimensionless ratio, which is used as the link jitter factor; if the dimensionless ratio exceeds two, it is recorded as two.

6. The optimized control method for repairing hollow tiles in decorative engineering according to claim 4, characterized in that: The synchronization quality factor is used to measure whether the overall patching of the wireless link over multiple pulse cycles has approached the system's minimum integrity requirements.

7. The optimized control method for repairing hollow tiles in decorative engineering according to claim 4, characterized in that: The stage factor is automatically assigned a value based on the stage in which it is located.

8. An optimized control system for repairing hollow tiles in interior decoration projects, characterized in that: The control system, based on the method according to any one of claims 1-7, includes the following modules: The data acquisition and coordinate mapping module is used to establish a spatial coordinate system with brick grid in the CAD facade environment, and to number each facing brick according to the rule of left to right and bottom to top; The 3D hollow modeling and hole planning module is used to extract acoustic and thermal image data and generate a 3D hollow distribution model; and based on the 3D hollow distribution model, it generates injection holes for each stage and records the 3D coordinates and injection hole number of each injection hole. The multi-stage dispensing scheduling and execution module writes the pulse cycle and three-stage injection formula into the task sequence according to the hole position table, issues pump start, pump stop and stage switching instructions for each hole, and collects pressure, flow rate and flow rate feedback at the end of each pulse. The adaptive closed-loop control and parameter learning module calculates volume deviation, flow rate deviation and residual pressure slope in real time during the dispensing process, and uses the three-stage volume tolerance and flow rate-pressure coupling coefficient as the metric benchmark to sequentially complete wireless frame loss verification, coupling blockage verification and residual pressure trend diagnosis; generates compensation instructions that include adjustment of start-up duration, target pressure or pause duration; and modifies the next pulse parameters and initiates a verification pulse in real time according to the compensation instructions.

9. The optimized control system for repairing hollow tiles in decorative engineering according to claim 8, characterized in that: The adaptive closed-loop control and parameter learning module includes a window adjustment module, which is used to dynamically construct an aligned window.

Images