Construction engineering waterproof roll paving method

Abstract

A construction engineering waterproof roll paving method comprises the steps that S1, base plane three-dimensional point data are collected to build a base plane model, obstacles are recognized in combination with the environment, and an optimal paving path is generated; s2, the base surface is cleaned, meanwhile, surface defects of the base surface are recognized, and hot air softening and joint filling repairing are conducted; s3, the coiled material is controlled to be unwound, the bonding layer is softened through infrared heating, and real-time position compensation is conducted based on machine vision; and S4, recording data of the whole construction process, and performing dynamic optimization through a reinforcement learning model based on the data. Accurate paving can be achieved, open fire operation risks can be completely avoided, and safety and controllability are achieved; the paving position can be automatically identified, a corresponding paving model can be constructed, and improved learning can be carried out to correct the paving method.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of building waterproof construction, and particularly relates to an intelligent building engineering waterproof roll material paving method for replacing traditional manual roll material paving operation. BACKGROUND

[0002] In current building waterproof engineering, roll material paving completely relies on manual operation, and systematic defects exist, specifically as follows: (1) manual positioning deviation causes roll material lap misalignment error to generally exceed 5 mm, and the continuity of the waterproof layer is damaged; the uneven rolling pressure of workers causes a bubble residual rate of up to 15%-20%, and causes leakage and hollowing; the temperature fluctuation of a flame gun when softening the roll material is more than +50℃, the adhesive strength is reduced by more than 30%, and the construction quality is poor. (2) 2-3 people are needed to carry and pave a roll of roll material of more than 30 kg, and the daily average construction area is less than 200 square meters, and the invalid waiting time accounts for more than 40% due to segmented operation, and the construction efficiency is low. (3) open-type gas flame gun operation causes fire accidents to account for 25% of construction site accidents; the concentration of benzopyrene in hot melt asphalt smoke gas is 8-10 times higher than the occupational limit value, and seriously endangers the health of workers, and the safety risk is large. (4) the skill difference of workers causes the adhesive strength to fluctuate by 50%, a large amount of quality inspection manpower needs to be invested for full inspection, the authenticity of parameters such as temperature and rolling times recorded manually is doubtful, quality problems cannot be traced back to the source, and the management cost is high. These defects are caused by the originality of the process, the discretization of data and the extensive control, and it is urgent to systematically solve the problems through the integration of intelligent equipment, high-precision positioning, closed-loop temperature control and real-time quality feedback. SUMMARY

[0003] The application provides a building engineering waterproof roll material paving method, which solves the technical problems of low precision, high safety risk and low efficiency of the existing manual paving operation.

[0004] To solve at least one of the above technical problems, the technical scheme adopted by the application is:

[0005] A building engineering waterproof roll material paving method, comprising the following steps:

[0006] S1, collecting three-dimensional point data of a base surface to build a base surface model, identifying obstacles in combination with the environment, and generating an optimal paving path;

[0007] S2, cleaning the base surface, identifying surface defects of the base surface, and performing hot air softening and joint repair;

[0008] S3, controlling roll material unwinding, infrared heating softening of the adhesive layer, and real-time position compensation based on machine vision;

[0009] S4, recording data of the whole construction process, and dynamically optimizing based on the data through a reinforcement learning model.

[0010] Further, in S2, a 1200 rpm rotating brush is started and a -5 kPa negative pressure cleaner is connected to remove floating dust on the base surface; then the repair robot performs heating and joint glue injection on the cracks.

[0011] Further, in S3, the control of the unwinding of the roll material is performed, and the adhesive layer is softened by infrared heating. Specifically, the roll material is placed in the unwinding bin by the multi-axis robot arm, the roll material is unwound by the servo motor at a constant tension of 45-55 N, and the roll material is heated by infrared radiation and then output for paving.

[0012] Further, the position compensation includes capturing the offset of the roll material edge from the planned path in real time by a binocular stereo camera, and driving the multi-axis robot arm to perform 0.1 mm level precision fine tuning in X / Y / Z three-dimensional space.

[0013] Further, S3 also includes a rolling and exhausting step, in which the roll material is rolled 3-5 times at a pressure of 0.5-0.8 MPa according to the height of the base surface, and the volume of air bubbles between the roll material and the base surface is detected in real time by an ultrasonic sensor.

[0014] Preferably, when the volume of air bubbles is greater than 0.3 mm 3 , the number of rolling is automatically increased by 2-4 times and the pressure value is increased by 0.1-0.3 MPa.

[0015] Further, S3 also includes a hot melting joint step, in which the hot air gun is adjusted to spray hot air at an angle of 30-90° to the surface of the roll material, forming a melting zone of 400±10℃ on the surface of the roll material; the process parameters are dynamically adjusted according to the melting layer thickness feedback by the thermocouple; when the thickness is less than 1.2 mm, the temperature is increased by 10℃ and the heating time is extended by 0.5s.

[0016] Preferably, the dynamic adjustment is based on a temperature-speed coupling control model, when the roll material conveying speed is increased, the output power of the hot air gun is increased by PD algorithm to maintain the stability of the melting zone temperature.

[0017] Further, the air bubble line is detected by an infrared thermal imaging sensor; the adhesive coverage is detected by ultrasonic flaw detection (threshold≥98%); the defect area is sprayed with a fluorescent marker of 20-25 mm in diameter and the secondary processing is automatically triggered.

[0018] Preferably, the secondary processing includes local secondary rolling of the air bubble defect area and reactivation of the hot air gun for supplementary melting at the area with insufficient adhesive coverage.

[0019] Further, it also includes a safety protection step for real-time monitoring of the inclination angle of the equipment, and a three-level response safety mechanism is triggered when the inclination angle is greater than 5°, including:

[0020] Level 1 response: Cut off lithium battery power within 0.2 seconds;

[0021] Level 2 Response: Activate automatic fire suppression system;

[0022] Level 3 response: Upload positioning coordinates and operating condition data to the cloud platform via 5G network.

[0023] Furthermore, the optimal paving path in S1 adopts a reinforcement learning model, with input parameters including base slope, roll material brightness and obstacle density, and output as the smoothness coefficient η (0 < η ≤ 1) of the robotic arm motion trajectory.

[0024] Furthermore, S4 includes:

[0025] Record temperature, pressure, and positioning coordinate parameters during the construction process, and bind them to the positioning system for spatiotemporal information;

[0026] By running a reinforcement learning model through an edge computing terminal, temperature, pressure, and positioning deviation parameters are recorded in real time to generate a three-dimensional construction quality map.

[0027] Based on historical data, the hot melt temperature-speed matching relationship and the correlation function between compaction pressure and base surface roughness are dynamically optimized. The optimization formula is as follows:

[0028]

[0029] In the formula, P new The optimized rolling pressure is expressed in MPa; σ is the yield strength of the waterproof membrane material, expressed in MPa; Ra is the surface roughness, expressed in μm; v is the rolling speed, expressed in m / min; k1 and k2 are proportionality coefficients, where k1 is expressed in μm and k2 is expressed in MPa.

[0030] This application presents a method for laying waterproof membrane in building engineering, which can achieve precise laying to achieve zero-defect bonding; it can automatically identify the laying position, safely avoid the risks of open flame operations, and eliminate the hazards of toxic gas exposure and fire; it can also eliminate the problem of reduced bonding strength caused by misalignment, residual air bubbles, and temperature runaway; it can build a laying construction model to achieve full-process traceability, and can also be improved and learned to optimize the laying method. Attached Figure Description

[0031] Figure 1 This is a structural schematic diagram of a waterproof membrane laying device for building engineering according to this application;

[0032] Figure 2 This is a top view of the structure of a waterproof membrane laying device for building engineering according to this application;

[0033] Figure 3is a flow chart of a waterproofing membrane laying method for building engineering in the present application.

[0034] In the figure:

[0035] 10, mobile chassis 20, base surface processing device 21, laser scanner

[0036] 22, rotating brush 23, negative pressure cleaner 24, repair robot arm

[0037] 241, rotating shaft one 242, main rod one of robot arm 243, end effector one

[0038] 25, repair control module 30, membrane storage device 31, unwinding bin

[0039] 32, conveying roller group 33, deviation correction roller group 34, heating cavity

[0040] 35, membrane conveying port 36, membrane 40, laying execution device

[0041] 41, multi-axis robot arm 411, rotating shaft two 412, main rod two of robot arm

[0042] 413, end effector two 42, positioning system 43, roller pressing group

[0043] 431, extrusion wheel 432, hydraulic lifting shaft 44, laying control module

[0044] 50, hot melt and quality inspection device 51, hot air spray gun 52, cooling and shaping mechanism

[0045] 53, multi-spectral sensor 54, marking head 55, hot melt control module

[0046] 56, defect marking control module 60, control system 61, lithium battery unit

[0047] 62, human-computer interaction panel 63, vertical frame 64, handrail

[0048] 65, wire slot 66, emergency stop switch DETAILED DESCRIPTION

[0049] The present application will be described in detail below in conjunction with the drawings and specific embodiments.

[0050] The present embodiment proposes a waterproofing membrane laying method for building engineering, such as Figures 1-2As shown, including the bottom of the chassis 10 and control system 60 is provided with walking wheels, wherein the top plane of the chassis 10 constitutes the installation platform. In the direction of the coil material paving process, the base surface treatment device 20, the coil material storage and conveying device 30, the paving execution device 40 and the hot melting and quality inspection device 50 are arranged in sequence from front to back on the chassis 10; and the base surface treatment device 20, the coil material storage and conveying device 30, the paving execution device 40, the hot melting and quality inspection device 50 are electrically connected with the control system 60.

[0051] The chassis 10 is made of high-strength alloy steel and is welded into shape, and the upper end face is provided with longitudinal and transverse interlaced wiring grooves, and the bottom is provided with an electromagnetic interference shielding layer. The walking wheels of the chassis 10 include four steering wheels, two front steering wheels and two rear steering wheels with independent intelligent functions, and the four steering wheels are equipped with multi-modal fusion navigation controllers to achieve ±1° steering accuracy.

[0052] A vertical frame 63 is arranged at the tail of the chassis 10, a handrail 64 is arranged at the upper section of the vertical frame 63, a wire groove 65 is arranged at the connection between the vertical frame 63 and the chassis 10, and an emergency stop switch 66 is arranged at the side of the vertical frame 63, which is electrically connected with the chassis inclination sensor and cuts off the power output within 0.2 seconds after triggering. The vertical frame 63 can be a C-shaped hollow fence type frame structure, or a flat plate structure, which is designed with a modular quick release structure, and the wire groove 65 is lined with a bending-resistant shielding cable, and the wiring meets the IP67 protection level.

[0053] The control system 60 is integrated on the chassis 10, including a lithium battery unit 61 for powering the device and a human-machine interaction panel 62 for centralized control and display, the lithium battery unit 61 is a high-density lithium battery pack supporting fast charging and wireless charging, and one full charge can work for ≥10h. Among them, the human-machine interaction panel 62 is a touch screen with built-in AI chip, and integrates a Beidou / GNSS dual-mode positioning module, which can display device working condition data in real time, including temperature, pressure, positioning coordinates and quality detection report. The human-machine interaction panel 62 is fixed between the handrail 64 and the vertical frame 63.

[0054] The base surface treatment device 20 is installed at the front of the mobile chassis 10, including a laser scanner 21 for scanning the base surface, a rotating brush 22 and a negative pressure cleaner 23 for cleaning the base surface, a repair mechanical arm 24 for repairing defects of the base surface, and a repair control module 25 for controlling the repair mechanical arm 24. Among them, the rotating brush 22 is arranged close to the end, the negative pressure cleaner 23 is located on the inner side of the rotating brush 22, and the rotating brush 22 and the negative pressure cleaner 23 are arranged on the lower bottom surface of the mobile chassis 10, which constitutes a base surface dust removal system, that is, first clean the garbage on the base surface, and then process it through the negative pressure cleaner 23. The laser scanner 21 is arranged on the upper end surface of the mobile chassis 10 and close to the front end side, which is vertically suspended for collecting base surface patterns and generating a base surface three-dimensional point cloud model.

[0055] The base surface treatment device 20 further comprises a defect marking control module 56 arranged on the mobile chassis 10, which is connected with the repair mechanical arm 24 and controls the work of the repair mechanical arm 24; and the defect marking control module 56 is arranged on the central axis of the mobile chassis 10 and close to the end of the laser scanner 21. The repair mechanical arm 24 is a four-axis mechanical arm, including a rotating shaft one 241, a mechanical arm main rod one 242 and an end effector one 243, the end effector one 243 integrates a hot air gun and a caulking glue gun, and the hot air temperature control range is 50-300℃.

[0056] The coiled material storage and conveying device 30 is installed in the middle of the mobile chassis, including a coiled material unwinding bin 31 for accommodating coiled material, a conveying roller group 32 for conveying coiled material, a heating cavity 34 for heating coiled material, a heating module arranged in the heating cavity 34, and a deviation correction roller group 33 for guiding. The input port of the coiled material unwinding bin 31 is provided with the conveying roller group 32; and the outlet is sequentially provided with the deviation correction roller group 33, the heating cavity 34 and the coiled material transmission port 35. Among them, the deviation correction roller group 33 and the coiled material transmission port 35 form a slope surface to guide the coiled material 36 obliquely, and the deviation correction accuracy of the deviation correction roller group 33 is ±0.5mm.

[0057] The coiled material 36 is placed in the coiled material unwinding bin 31, and the paving transmission direction is as shown in the figure. Figure 1The middle black solid line. The intelligent unwinding bin 31 is driven by the servo motor to drive the roll conveying roller set 32 to output the roll material 36. The roll material 36 is drawn out from the unwinding bin 31, passes through the conveying roller set 32, the deviation correction roller set 33, and is transmitted to the lower surface of the mobile chassis 10 through the roll material transmission port 35, and then is laid on the base surface through the extrusion wheel 431 on the roller pressing group 43. In this transmission direction, after being heated and softened by the infrared radiation heating module in the heating cavity 34, it is guided to the roll material transmission port 35 through the inclined slope, and is blown by the hot air spray gun 51, and is offset towards the side close to the extrusion wheel 431, which is more conducive to the rolling of the extrusion wheel 431 on the roll material 36, so that it is laid on the base surface. The temperature control accuracy of the heating module is ±3℃, and the heated roll material is guided to the laying position through the roll material transmission port 35 after passing through the slope. Preferably, a cutter for cutting the roll material 36 is also provided at the position of the roll material transmission port 35, which is a conventional structure and the attached drawings are omitted here.

[0058] The laying execution device 40 is installed on the mobile chassis 10 and located behind the roll material storage and conveying device 30, and includes a multi-axis mechanical arm 41 for grabbing and laying the roll material, a positioning system 42 for positioning, and a roller pressing group 43 for rolling the roll material 36. The laying control module 44 is provided on one side of the mobile chassis 10 in the length direction, and the multi-axis mechanical arm 41 is fixedly installed on the laying control module 44. It is a six-axis collaborative mechanical arm, which includes six mechanical arm main rods two 412, a rotation shaft two 411 for connecting adjacent mechanical arm main rods two 412, and an end effector two 413 provided at the end thereof. The end effector two 413 includes a vacuum chuck and a pressure feedback system. The suction force of the vacuum chuck can be adjusted in the range of 0.05-0.5MPa. The positioning system 42 is a three-dimensional visual positioning system, which is provided at the front and rear ends of the mobile chassis 10, and includes a binocular stereo camera and an environment perception camera. It combines the data of the environment perception camera and the laser radar, and the positioning error is ≤1mm. Combined with the AI image recognition algorithm, the laying track is positioned in time.

[0059] The roller pressing group 43 is a self-adaptive roller pressing group, which includes an extrusion wheel 431 coated with a silica gel layer and a hydraulic lifting shaft 432 for driving the extrusion wheel 431 to lift. The air pressure of the extrusion wheel 431 can be adjusted and the rolling pressure can be dynamically adjusted. The roller pressing group 43 is symmetrically arranged on both sides of the mobile chassis 10 in width, and is symmetrically arranged, so as to complete the rolling and laying of the roll material 36. The self-adaptive roller pressing group 43 adjusts the rolling height through the hydraulic lifting shaft 432, and the rolling pressure is 0.2-0.8MPa and continuously adjustable. The surface of the roller pressing group 43 is coated with a high-elasticity silica gel layer with a Shore hardness of 60±5HA, and the uniformity of the rolling contact surface pressure distribution is ≥95%. The multi-spectrum sensor 53 can detect the rolling area in real time, and the data is fed back to the central controller.

[0060] The hot melt and quality inspection device 50 is installed behind the paving execution device 40, and includes a hot air spray gun 51 for joint bonding, a cooling and shaping mechanism 52 for accelerating curing, a multi-spectral sensor 53 for quality inspection, a marking head 54 for marking defects, and a hot melt control module 55. Among them, the hot melt control module 55 is two, symmetrically arranged on both sides of the width of the mobile chassis 10; and a hot air spray gun 51 is arranged on each hot melt control module 55, which is installed on the hot melt control module 55 through a rotatable folding shaft, and the spraying direction of the hot air spray gun 51 is adjustable. The driving end hot air spray gun 51 performs joint fusion, and the spray gun realizes 180° azimuth adjustment through the rotatable folding shaft, and then realizes roll joint fusion bonding through adjustable angle hot air spraying. The cooling and shaping mechanism 52 is a semiconductor refrigeration and air cooling composite structure, which can accelerate the curing of the bonding layer, and its cooling rate is ≥10℃ / s. Among them, the distance between the hot air spray gun 51 and the cooling and shaping mechanism 52 is adjustable, the minimum working distance is 50mm, and the maximum extension distance is 300mm.

[0061] The multi-spectral sensor 53 is located between the cooling and shaping mechanism 52 and the defect marking control module 56, and the multi-spectral sensor 53 and the defect marking control module 56 are fixedly arranged on the middle axis of the mobile chassis 10. The multi-spectral sensor 53 is an infrared thermal imaging and ultrasonic flaw detection composite sensor array, which detects bubbles through infrared thermal imaging and detects bonding density through ultrasonic flaw detection; and then analyzes and generates quality atlas through edge AI processor in real time. The marking head 54 is a fluorescent paint spraying head, which is installed at the defect marking control module 56 through a rotatable folding shaft, and its spraying direction is adjustable and controlled by the defect marking control module 56, which can automatically spray fluorescent marking to locate unqualified areas. The defect marking control module 56 can control the marking head 54 to spray fluorescent marks according to the detection result, and the positioning accuracy is ±2mm.

[0062] The working of the paving equipment, first through the laser scanner 21 in the base surface treatment device 20 collects the base surface three-dimensional point data to build a base surface model, combines with the environment to identify obstacles and generate the optimal paving path; then start the rotating brush 22 and the negative pressure cleaner 23 to clean the base surface, at the same time, through the repair mechanical arm 24 to identify the surface defects of the base surface and perform hot air softening and caulking repair. Then through the coiled material storage and conveying device 30, the coiled material 36 is placed into the unwinding bin 31 by the multi-axis mechanical arm 41, the coiled material 36 is driven by the servo motor with constant tension, and is conveyed through the conveying roller group 32 and the deviation correction roller group 33, and is output after being heated and softened by infrared radiation in the heating cavity 34. The paving execution device 40 places the coiled material in the unwinding bin 31 by the multi-axis mechanical arm 41, and captures the offset amount of the coiled material edge and the planned path in real time based on the binocular stereo camera in the positioning system 42, drives the mechanical arm to perform millimeter-level position compensation in three-dimensional space; at the same time, the roller group 43 implements multiple rolling according to the base surface undulation with variable pressure, and detects the bubble volume through the ultrasonic sensor to dynamically adjust the rolling parameters. The hot melt and quality inspection device 50 sprays a molten belt to complete the joint bonding by the adjustable angle of the hot air spray gun 51, the cooling and shaping mechanism 52 accelerates the curing, and the multi-spectral sensor 53 detects bubbles and bonding coverage in real time, sprays fluorescent markers on the defect area and automatically triggers secondary processing. The whole paving process is coordinated by the control system 60 to record the construction parameters, dynamically optimize the paving process through the reinforcement learning model, and realize the full-process integrated operation of digitalization and precision operation.

[0063] The building engineering waterproof coiled material paving equipment designed in the application constructs a full-automatic waterproof coiled material paving system, realizes the full-process integrated operation from intelligent base surface processing, coiled material constant tension conveying and preheating, millimeter-level accurate positioning paving, to self-adaptive rolling and exhaust and closed-loop temperature control hot melt joint. The device can fundamentally solve the quality problems such as paving misplacement, bubble residue and poor bonding existing in traditional manual construction, improve the construction efficiency, and completely eliminate the safety hazards caused by open fire operation; the construction parameters in the whole process are recorded in real time and can be traced on the cloud, realizing digitalization and precision quality management.

[0064] A building engineering waterproof coiled material paving method, the flow chart is as Figure 3 The steps include:

[0065] S1, collect base surface three-dimensional point data to build a base surface model, combine with the environment to identify obstacles and generate the optimal paving path.

[0066] The laser scanner 21 installed on the front of the mobile chassis 10 scans the construction base surface (the scanning density is greater than or equal to 500 points / m 2), three-dimensional point data of the base surface is collected, and an accurate three-dimensional model of the base surface is built using the data. During the modeling process, the system simultaneously collects on-site image data through the environmental perception camera, and identifies the obstacles existing on the base surface in combination with the environmental information. Finally, based on the constructed base surface model and the identified obstacle information, a reinforcement learning model is used for calculation to generate an optimal motion path suitable for the multi-axis robot 41 to perform the tiling task. The input parameters of the reinforcement learning model include the base surface slope, the roll brightness, and the obstacle density, and the output result is a smoothness coefficient η (0 < η < 1) of the robot motion trajectory.

[0067] Through high-precision three-dimensional modeling and environmental perception, an accurate path planning basis is provided for subsequent tiling operations, and automatic identification and accurate positioning of the tiling position are realized. This avoids the problem of roll lap misplacement (error exceeding 5mm) caused by positioning deviation in traditional manual construction, ensures the continuity of the waterproof layer, and lays a key foundation for realizing zero-defect bonding. At the same time, the generated optimized path also improves the smoothness and efficiency of the robot motion.

[0068] S2, clean the base surface, identify surface defects, and perform hot air softening and joint filling repair.

[0069] The rotating brush 22 and the negative pressure cleaner 23 in the base surface treatment device 20 are started, wherein the rotating brush rotates at a speed of 1200 rpm to sweep away loose debris and floating dust on the base surface; at the same time, the negative pressure cleaner 23 with a linkage negative pressure value of -5 kPa is operated to completely suck away the swept floating dust, realizing synchronous cleaning and dusting to ensure the cleanliness of the base surface.

[0070] After cleaning, the base surface defects are handled by the repair robot arm 24, which is a four-axis robot arm, and its end effector 243 integrates a hot air gun and a joint filling gun. For the identified defects such as base surface cracks, the repair robot arm 24 repairs the base surface cracks. First, the hot air gun is used to heat and soften it, and the hot air temperature can be accurately controlled in the range of 50-300℃, aiming to improve the activity of the base surface material and enhance the adhesion of the filling material. Then, the repair robot arm 24 switches to the joint filling gun, and under the injection pressure of 0.2 MPa, the filling glue is accurately injected into the softened crack to complete the filling repair.

[0071] Through the linkage of the high-speed rotating brush 22 and the negative pressure cleaner 23, the influence of floating dust on the adhesion strength of the roll and the base surface is completely eliminated, avoiding quality hidden dangers such as hollowing and debonding caused by unclean base layer. Through the hot air softening and pressure injection repair of the repair robot arm 24 to the cracks, the defects of the base surface are pre-filled to prevent the waterproof layer from being pulled apart or stress concentrated due to base surface cracking, and to fundamentally eliminate the risk of subsequent leakage.

[0072] S3, control the unwinding of the roll material, soften the adhesive layer by infrared heating, and compensate for the real-time position based on machine vision.

[0073] The roll material 36 is grabbed by the multi-axis mechanical arm 41 and placed into the unwinding bin 31. Then, the servo motor drives the roll material 36 to unwind at a constant tension of 45-55 N, ensuring smooth and wrinkle-free unwinding. During the conveying process, the roll material passes through the heating chamber 34 and receives infrared radiation heating, causing the adhesive layer to soften in advance, preparing for subsequent adhesive laying and paving.

[0074] During the paving process, the positioning system 42 based on machine vision uses a binocular stereo camera to capture the offset between the edge of the roll material 36 and the optimal planning path generated in step S1 in real time, and immediately drives the multi-axis mechanical arm 41 to make fine adjustments in the X / Y / Z three-dimensional space at the level of 0.1 mm, ensuring accurate paving of the roll material along the predetermined path.

[0075] Immediately after paving, rolling and degassing are performed. The roller group 43 applies 3-5 passes of rolling to the roll material at a pressure of 0.5-0.8 MPa according to the undulating height of the base surface. At the same time, the ultrasonic sensor integrated in the device detects the bubble volume between the roll material 36 and the base surface in real time. When the detected bubble volume is >0.3 mm 3 , the system automatically triggers the optimization mechanism, increasing the number of rolling passes by 2-4 and increasing the rolling pressure by 0.1-0.3 MPa to completely eliminate the bubbles.

[0076] Step S3 also includes a hot melt seam step. For the lap joint of the roll material, a hot air gun is used for processing. The hot air gun 51 is adjusted to spray hot air at an angle of 30-90° to the surface of the roll material, forming a molten band on the surface of the roll material at a temperature of 400±10℃. The system dynamically adjusts the process parameters according to the molten layer thickness feedback by the thermocouple; when the thickness is <1.2 mm, the temperature is automatically increased by 10℃ and the heating time is extended by 0.5s. This dynamic adjustment is based on a temperature-speed coupling control model. When the roll material conveying speed increases, the output power of the hot air gun is increased by PD (Proportional-Derivative) algorithm, etc. to maintain the stability of the molten band temperature.

[0077] After rolling and hot melting, real-time quality detection and automatic repair are performed. An infrared thermal imaging sensor is used to detect bubble residues, and an ultrasonic flaw detection technology is used to detect the adhesive coverage (set to ≥98% as the threshold for acceptance). For the identified defect areas (such as bubbles or insufficient adhesion), the marking nozzle sprays a fluorescent marker with a diameter of 20-25 mm, and automatically triggers secondary processing. Secondary processing includes automatically performing local secondary rolling on the bubble area and reactivating the hot air gun for supplementary melting on the area with insufficient adhesive coverage.

[0078] The step S3 further comprises a safety protection step, that is, the system monitors the inclination angle of the device in real time; when the inclination angle of the device is > 5°, a three-level response safety mechanism is triggered immediately. The three-level response safety mechanism comprises: a first level response, that is, cutting off the lithium battery power supply within 0.2 seconds; a second level response, that is, starting an automatic fire extinguishing device; and a third level response, that is, uploading the device positioning coordinates and working condition data to a cloud platform through a 5G network.

[0079] Real-time position compensation based on machine vision controls the laying positioning accuracy to the level of 0.1 mm, completely solves the problem of traditional manual construction lap misplacement error exceeding 5 mm, and ensures the continuity and integrity of the waterproof layer. At the same time, through constant tension conveying, self-adaptive pressure rolling, and ultrasonic real-time detection of bubbles and linkage feedback adjustment, air can be efficiently removed, and the residual rate of bubbles is reduced to a very low level, thereby fundamentally avoiding the leakage hidden danger caused by air bubbles and hollows. Infrared preheating and closed-loop temperature control hot air welding are also used to stably control the temperature of the adhesive layer in the optimal range of 400±10℃, avoid the temperature fluctuation caused by traditional open flame operation, which leads to a decrease in adhesive strength (more than 30%), and ensure the uniformity and high reliability of the adhesive quality. Throughout the process, infrared radiation and controllable hot air are used to completely avoid the fire and health hazards caused by open gas flame guns. Quality detection is also integrated into the laying process to find defects in real time and automatically mark and repair them immediately, realizing the detection, judgment and processing at the same time, replacing the traditional full number inspection relying on manual work, and ensuring the traceability of quality.

[0080] S4, record the whole construction process data, and dynamically optimize based on the data through a reinforcement learning model.

[0081] Record the temperature, pressure and positioning coordinate parameters during the construction process, and bind the space-time information with the positioning system; run the reinforcement learning model through the edge computing terminal to record the temperature, pressure and positioning deviation parameters in real time, and generate a three-dimensional construction quality map.

[0082] According to the historical data, the hot melting temperature-speed matching relationship and the rolling pressure and base surface roughness correlation function are dynamically optimized, and the optimization formula is:

[0083]

[0084] In the formula, P new is the optimized rolling pressure, unit: MPa; σ is the yield strength of the waterproof roll material, unit: MPa; Ra is the base surface roughness, unit: um; v is the rolling speed, unit: m / min; k1 and k2 are proportional coefficients, wherein the unit of k1 is um, and the unit of k2 is MPa.

[0085] The building engineering waterproof roll paving method designed by the application can realize accurate paving to achieve zero defect bonding, can automatically identify the paving position, safely avoid the risk of open flame operation, eliminate toxic gas exposure and fire hazards, can also eliminate the problem of reduced bonding strength caused by lap misplacement, bubble residue and temperature out of control, can build a paving construction model, realize full-process traceability, and can also be improved and learned to optimize the paving method.

[0086] The above describes the embodiments of the application in detail, and the content is only the preferred embodiments of the application and cannot be considered as limiting the scope of the application. Any equivalent changes and improvements made within the scope of the application should still belong to the patent coverage of the application.

Claims

1. A method of applying a waterproofing membrane to a construction project, the method comprising the steps of Comprise: S1, collect the base surface three-dimensional point data to build the base surface model, identify the obstacles in combination with the environment and generate the optimal paving path; S2, clean the base surface, identify the surface defects of the base surface and perform hot air softening and crack repair at the same time; S3, control the unwinding of the coiled material, soften and bond the layer by infrared heating, and compensate the position in real time based on machine vision; S4, record the whole process data, and dynamically optimize based on the data through reinforcement learning model.

2. A method of applying a waterproofing membrane to a construction project according to claim 1, wherein, In S2, start 1200rpm rotating brush and link-5kPa negative pressure cleaner to remove dust on the base surface; Then the repair mechanical arm performs heating and crack filling glue injection on the crack.

3. A method of applying a waterproofing membrane to a construction project according to claim 1 or 2, characterised in that, In S3, about controlling the unwinding of the coiled material, softening and bonding the layer by infrared heating, specifically: through the multi-axis mechanical arm to grab the coiled material and place it in the unwinding bin, through the servo motor to drive the coiled material to expand with a constant tension of 45-55N, and at the same time, use infrared radiation to heat the coiled material and output paving.

4. A method of applying a waterproofing membrane to a construction project according to claim 3, wherein, The position compensation includes: capturing the offset amount of the coiled material edge and the planned path in real time through a binocular stereo camera, and driving the multi-axis mechanical arm to perform 0.1mm level precision fine tuning in X / Y / Z three-dimensional space.

5. A method of applying a waterproofing membrane to a construction project according to any one of claims 1-2, 4, wherein, S3 also includes a rolling and exhausting step, which applies 3-5 passes of rolling to the coiled material at a pressure of 0.5-0.8MPa according to the height of the base surface, and detects the bubble volume between the coiled material and the base surface in real time through an ultrasonic sensor; Preferably, when the bubble volume > 0.3 mm 3 is detected, the number of rolls is automatically increased by 2-4 passes and the pressure value is increased by 0.1-0.3 MPa.

6. A method of applying a waterproofing membrane to a construction project according to claim 5, wherein, S3 also includes a hot melting step, which adjusts the hot air gun to spray hot air at an angle of 30-90° to the surface of the coiled material, forming a molten zone on the surface of the coiled material at a temperature of 400±10℃; dynamically adjust the process parameters according to the molten layer thickness feedback by thermocouple; when the thickness is less than 1.2mm, increase the temperature by 10℃ and extend the heating time by 0.5s; Preferably, the dynamic adjustment is based on a temperature-speed coupling control model, when the coiled material conveying speed increases, the output power of the hot air gun is increased by PD algorithm to maintain the stability of the molten zone temperature.

7. A method of applying a waterproofing membrane to a construction project according to claim 6, wherein, Bubble line is detected by infrared thermal imaging sensor; ultrasonic flaw detection (threshold value ≥98%) is used to detect the bonding coverage, and 20-25mm diameter fluorescent marker is sprayed on the defect area to automatically trigger secondary processing; Preferably, the trigger secondary processing includes automatically performing local secondary rolling on the bubble defect area, and reactivating the hot air gun to supplement the melting on the area with insufficient bonding coverage.

8. A method of applying a waterproofing membrane to a construction project according to claim 1, wherein, It also includes a safety protection step for real-time monitoring of the device inclination angle, which triggers a three-level response safety mechanism when the device inclination angle is greater than 5°, including: First level response: cut off the lithium battery power supply within 0.2 seconds; Second level response: start the automatic fire extinguishing device; Third level response: upload the positioning coordinates and working condition data to the cloud platform through 5G network.

9. A method of applying a waterproofing membrane to a construction project according to any one of claims 1-2, 4, 6-8, wherein, The optimal paving path in S1 uses a reinforcement learning model, the input parameters include base surface slope, coiled material brightness and obstacle density, and the output is the smoothness coefficient η(0<η≤1) of the mechanical arm motion trajectory.

10. A method of applying a waterproofing membrane to a construction project according to claim 1, wherein, S4 includes: Record the temperature, pressure, positioning coordinate parameters during construction, and bind the space-time information with the positioning system; The reinforcement learning model is run through the edge computing terminal, real-time records temperature, pressure, positioning deviation parameters, and generates a three-dimensional construction quality atlas; According to historical data, the heat melting temperature-speed matching relationship and the rolling pressure and base roughness correlation function are dynamically optimized, and the optimization formula is: In the formula, P new In order to optimize the post-rolling pressure, the unit is MPa; sigma is the yield strength of the waterproofing membrane material, the unit is MPa; Ra is the surface roughness, the unit is um; v is the rolling speed, the unit is m / min; k1 and k2 are proportional coefficients, wherein the unit of k1 is um, and the unit of k2 is MPa.

Images