Intelligent 3D printing method for complex tunnel lining

Abstract

The invention discloses an intelligent 3D printing method for a complex tunnel lining, and relates to the technical field of tunnel construction. Aiming at complex tunnel lining construction requirements, intelligent printing is realized through multi-module cooperation: a sensing module is utilized to obtain a tunnel contour and construct a digital twinborn model, and a lining completion contour and a to-be-printed area are determined; the material supply module is used for preparing printing materials based on the multi-objective optimization ratio; the walking module advances according to intelligent planning, a mechanical arm performs autonomous adjustment, dynamic adjustment and control are matched with a printing nozzle, performance adjustment and control are performed on a vertical wall, an arch shoulder and an arch crown through a regional self-adaptive algorithm, meanwhile, configuration optimization is performed on a printed region, and a complete process from data perception and material supply to regional printing and optimization is formed. The intelligent and precise level of tunnel lining construction can be improved, and the construction quality and efficiency are guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tunnel construction, in particular to an intelligent 3D printing method for complex tunnel lining. BACKGROUND

[0002] In the field of underground engineering, tunnel lining is the core structure to resist the pressure of surrounding rock, and the construction quality and efficiency of the lining determine the overall performance and construction period of the tunnel project.

[0003] At present, the construction of complex tunnel lining still mostly adopts the method of formwork support and concrete casting. Because the irregular lining structure needs to be customized with formwork, the construction period is prolonged, the cost is increased, and the reusability is poor; at the same time, the formwork is installed by workers, and the precision is affected, thereby affecting the flatness of the lining surface. During the concrete pouring process, due to insufficient concrete vibration, defects such as hollowing and honeycomb may occur in the crown and haunch parts, thereby easily producing cracks, increasing the maintenance cost in the later period, and affecting the safety and durability of the tunnel project.

[0004] In recent years, many enterprises have begun to use 3D printing technology for tunnel lining construction, but for the scene of complex tunnels, 3D printing technology still has not broken through the bottleneck.

[0005] The existing 3D printing equipment still uses unified printing parameters, and cannot configure different printing parameters (such as printing speed, nozzle height, etc.) according to different partitions and different scenes, so the structural differences of each partition are ignored, which reduces the printing efficiency and also reduces the printing quality.

[0006] At present, 3D printing technology has not established the logic of "deviation detection-parameter adjustment-effect verification", so there is still a lack of precision control. When the tunnel contour deviates due to external disturbance, the printing system cannot be corrected in time, thereby causing the gap between the lining and the surrounding rock to be too large or too small, increasing the construction cost and the construction period.

[0007] Although the current 3D printing technology can improve automation and reduce the number of construction personnel during tunnel lining construction, due to the above defects, the construction efficiency is slightly improved, and if the temperature fluctuation of the construction environment is too large, it is easy to cause abnormal situations such as segregation of the printing material and too long or too short setting time.

[0008] In the face of the multiple challenges of unstable quality, low efficiency, insufficient intelligence and automation, and low adaptability to complex scenes in the current complex tunnel lining construction, an intelligent 3D printing technology that can realize regional adaptive regulation and control, high-precision closed-loop control, and multi-module collaboration is urgently needed to break through the bottleneck of traditional processes and existing technologies, and to realize high-quality, high-efficiency, high-intelligence, and high-automation of complex tunnel construction. SUMMARY

[0009] The application aims to provide an intelligent 3D printing method for complex tunnel lining to improve the quality, efficiency, intelligence and automation of complex tunnel construction.

[0010] To solve the above technical problems, the application provides the following technical solutions:

[0011] An intelligent 3D printing method for complex tunnel lining, comprising the following steps:

[0012] S1. Tunnel excavation profile: a digital twin of the tunnel is constructed by using a physical model and sensors, comprising the following steps:

[0013] The laser scanner, structured light camera are used to obtain the tunnel excavation profile coordinates; the stress and strain sensor and displacement sensor are used to obtain the tunnel lining stress and surrounding rock displacement; the temperature and humidity sensor is used to obtain the temperature and humidity data in the tunnel; the encoder and flow sensor mounted on the mechanical arm are used to obtain the position coordinates of the mechanical arm nozzle and the nozzle flow data;

[0014] The dynamic data such as displacement and coordinates are processed by using discrete Kalman filtering to reduce noise and standardize coordinates of the original data;

[0015] The data collected by the laser scanner is generated into a grid model by using the Poisson reconstruction algorithm, and the three-dimensional geometric model of the tunnel excavation surface, design surface and printing surface is outputted;

[0016] An interaction model of surrounding rock and lining is constructed, and the core formula is as follows:

[0017] ① Lining bending moment:

[0018] ,

[0019] (q is the uniform pressure of surrounding rock, and x is the distance along the lining axis)

[0020] ② Lining displacement:

[0021] ,

[0022] According to the encoder data of the mechanical arm and the flow sensor, a motion trajectory equation and a material accumulation thickness model are constructed;

[0023] The data of each sensor is bound to the above formula to realize the initialization of the digital twin;

[0024] The update period of the data is determined according to the sampling frequency of the sensor to ensure that the digital twin in the display is consistent with the tunnel construction state.

[0025] S2. Region division: Import the digital twin into the system to obtain various parameters, and divide the tunnel model into three regions: vertical wall (A region), arch shoulder (B region), and arch top (C region), each region is assigned an independent parameter control identifier;

[0026] S3. Multi-source implementation sensing: The 3D printing robot plans the printing path according to the digital twin intelligence, and then collects various parameters according to the sensing modules on the mechanical arm and the surrounding tunnel, including the following steps:

[0027] Select the sensors involved in data fusion, such as: to fuse the lining stress data, select the mechanical arm force sensor and the tunnel stress and strain sensor;

[0028] Use variance to evaluate the reliability of data, the smaller the variance, the higher the reliability;

[0029] According to the proportion of the reliability of each data, the weight of each data is obtained;

[0030] Finally, multiply each weight by each data to obtain the fused data;

[0031] Subsequent data can be continuously collected, and the weight can be continuously updated for data fusion optimization.

[0032] S4. Regional adaptive control: The data collected by S3 are integrated to generate printing parameter instructions for each region, including mechanical arm motion trajectory, printing nozzle flow rate, feed ratio, and printing thickness;

[0033] S5. Precision closed-loop control: Compare the printed area profile data with the target profile data, if a certain data exceeds the set threshold, repeat the adaptive control of step S4 until the data is controlled within the threshold;

[0034] S6. Configuration optimization: Use sensors to detect the printed area and implement configuration optimization according to the tunnel mechanical environment, including the following steps:

[0035] Use printed area sensors to collect data and identify defects, such as using a structured light camera and a laser scanner to detect lining flatness, uniformity, and profile fit; use stress and strain sensors, displacement sensors, etc. to detect the maximum principal stress of the lining, the displacement rate of the surrounding rock, etc., and mark high stress and high displacement risk areas;

[0036] Compare the sensor data with the mechanical model of the digital twin to quantify the mechanical environment characteristics of the tunnel and determine the configuration optimization direction, such as:

[0037] First, calculate the foundation beam stress in the digital twin:

[0038] ,

[0039] Then the error between the calculated stress and the actual stress is obtained by the least square method, and the optimization priority is determined according to the specified grading standard;

[0040] Finally, the strategy is adjusted according to the priority, for example, if the lining stress needs to be reduced and the displacement needs to be inhibited, the wall thickness can be locally thickened, the cross section shape can be optimized, and the embedded steel bars can be arranged.

[0041] As a further technical solution of the application: the perception module includes a laser ranging sensor, a displacement sensor, a stress and strain sensor, a structured light camera, a temperature and humidity sensor, and a 5G private network as a transmission carrier to realize real-time transmission of multi-source data.

[0042] As a further technical solution of the application: the sub-area adaptive algorithm is realized by a deep learning model.

[0043] As a further technical solution of the application: the multi-target optimization proportioning of the feeding module is realized by importing data from various sensors, and the most suitable proportion is obtained after deep learning model analysis.

[0044] As a further technical solution of the application: if the sensor detects that the hollow gap in the vault area is greater than 1.5mm, the printing speed is reduced to 35-40mm / s, and the nozzle height is increased.

[0045] As a further technical solution of the application: the flatness deviation threshold of the A area is 5mm, and the profile deviation threshold of the C area is 2mm.

[0046] As a further technical solution of the application: the feeding device is associated with the flow sensor and the liquid level sensor, and the proportioning of each material of the concrete is adjusted according to the instructions.

[0047] One or more technical solutions provided in the embodiments of the application have at least the following technical effects or advantages:

[0048] (1) The amplitude and phase errors of the entire nested array are estimated by estimating the amplitude and phase errors of each subarray of the nested array respectively.

[0049] (2) The influence of the array amplitude and phase errors is eliminated by respectively correcting the nested array sample covariance matrix, the pseudo sample covariance matrix and the augmented covariance matrix using the estimated nested array amplitude and phase errors.

[0050] (3) The interpolation difference sum co-array becomes a virtual uniform linear array, and the virtual linear vector of the interpolation difference sum co-array is restored, thereby increasing the number of estimated signals and improving the degree of freedom of satellite communication interference suppression. BRIEF DESCRIPTION OF DRAWINGS

[0051] Figure 1 Flow chart of intelligent 3D printing method for complex tunnel lining of the application;

[0052] Figure 2 Schematic diagram of intelligent 3D printing method for complex tunnel lining of the application;

[0053] Figure 3 Schematic diagram of mechanical arm of the application;

[0054] Figure 4 Schematic diagram of feeding system of the application. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the application will be clearly and completely described below; obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the application.

[0056] Reference is made to the accompanying drawings Figure 1 The application discloses an intelligent 3D printing method for complex tunnel lining, comprising the following steps: constructing a digital twin of the tunnel by using a physical model and a sensor; importing the digital twin into a system to obtain various parameters, and dividing the tunnel model into three regions of a vertical wall (region A), a spandrel (region B) and a vault (region C), and giving each region an independent parameter control identifier; a 3D printing robot intelligently plans a printing path according to the digital twin, and then collects various parameters according to a perception module carried on the mechanical arm; generating printing parameter instructions for each region by integrating the collected data, including a mechanical arm motion trajectory, a printing nozzle flow rate, a feeding ratio and a printing thickness; precision closed-loop control: comparing the profile data of the printed region with the target profile data, and if a certain data exceeds a set threshold, repeatedly performing adaptive regulation and control until each data is controlled within the threshold; detecting the printed region by using a sensor, and optimizing the tunnel configuration according to the tunnel mechanical environment.

[0057] The inclination sensor, the miniature structured light camera and the temperature and humidity sensor are installed at the joints of the mechanical arm and the fuselage, which avoids printing layer slippage, prevents equipment collision, and can also identify small defects, thereby ensuring stable performance of the concrete.

[0058] In the application, the sensors are deployed according to regions, a displacement sensor and a visual camera are arranged every 5m in region A, a structured light camera and an inclination sensor are arranged every 3m in region B, and a group of laser profile sensors and stress and strain sensors are arranged every 2m in region A.

[0059] The data obtained by each part is transmitted to the control system through industrial Ethernet to ensure low transmission delay of construction. 5G private network can be deployed in important areas to further reduce the delay.

[0060] The application adopts a lightweight decision tree model to analyze the correlation between the data transmitted by each area and mechanical arm and the 3D printing effect in real time, and synchronously update to the rule base.

[0061] The 3D printing robot used in the application selects a 6-axis industrial mechanical arm; the printing nozzle adopts a variable caliber nozzle.

[0062] The BIM format of the tunnel lining is imported into the system, and the coordinate points of the key parts of each area of the tunnel are extracted as the accuracy control reference.

[0063] Every 10s, the profile data transmitted by the sensor is compared with the reference profile data, if the data of a certain area exceeds the threshold value for three times, secondary regulation is carried out until the deviation is reduced to within the threshold value.

[0064] In addition, it should be understood that although the present specification is described in terms of embodiments, not every embodiment contains only one independent technical solution, and the description of the specification is only for the sake of clarity, and those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment have been appropriately combined to form other embodiments easily understood by those skilled in the art.

Claims

1. A smart 3D printing method for complex tunnel lining, characterized in that, Includes the following steps: S1. Tunnel Excavation Outline: A digital twin of the tunnel is constructed using physical models and sensors; S2. Region Division: Import the digital twin into the system to obtain various parameters, and divide the tunnel model into three regions: vertical wall, arch shoulder, and arch top, which are defined as region A, region B, and region C, respectively. Each region is assigned an independent parameter control identifier. S3. Multi-source perception: The 3D printing robot intelligently plans the printing path based on the digital twin, and then collects various parameters based on the sensing modules mounted on the robotic arm and around the tunnel. S4. Regional adaptive control: The data collected by S3 is integrated to generate printing parameter instructions for each region, including the robotic arm motion trajectory, printing nozzle flow rate and feed ratio and printing thickness; S5. Precision closed-loop control: Compare the printed area contour data with the target contour data. If any data exceeds the set threshold, repeat step S4 for adaptive adjustment until all data are controlled within the threshold. S6. Configuration Optimization: Utilize sensors to detect the printed area and optimize the tunnel configuration based on the tunnel's mechanical environment.

2. The intelligent 3D printing method for complex tunnel lining according to claim 1, characterized in that, The sensing module includes a laser rangefinder, a displacement sensor, a stress-strain sensor, a structured light camera, and a temperature and humidity sensor. It uses a 5G private network as the transmission carrier to achieve real-time transmission of multi-source data.

3. The intelligent 3D printing method for complex tunnel lining according to claim 2, characterized in that, The regional adaptive algorithm is implemented by a deep learning model.

4. The intelligent 3D printing method for complex tunnel lining according to claim 3, characterized in that, The multi-objective optimization ratio of the feeding module is obtained by importing data from various sensors and analyzing it through a deep learning model to determine the most suitable ratio.

5. The intelligent 3D printing method for complex tunnel lining according to claim 4, characterized in that, If the sensor detects a gap greater than 1.5mm in the arch area, the printing speed is reduced to 35-40mm / s, and the nozzle height is increased.

6. The intelligent 3D printing method for complex tunnel lining according to claim 5, characterized in that, The flatness deviation threshold for area A is 5mm, and the contour deviation threshold for area C is 2mm.

7. The intelligent 3D printing method for complex tunnel lining according to claim 6, characterized in that, The feeding device is linked to flow sensors and level sensors, and adjusts the proportions of each concrete material according to instructions.

8. The intelligent 3D printing method for complex tunnel lining according to claim 1, characterized in that, Step 1 specifically involves: Laser scanners and structured light cameras are used to acquire the tunnel excavation outline coordinates; stress and strain sensors and displacement sensors are used to acquire the tunnel lining stress and surrounding rock displacement; temperature and humidity sensors are used to acquire the temperature and humidity data inside the tunnel; encoders and flow sensors mounted on the robotic arm are used to acquire the position coordinates of the robotic arm nozzles and the nozzle flow data. Discrete Kalman filtering is used to process displacement and coordinate dynamic data to reduce noise and standardize coordinates of the original data; The Poisson reconstruction algorithm is used to generate a mesh model from the data collected by the laser scanner, and outputs a three-dimensional geometric model of the tunnel's excavation face, design face, and printing face; The core formulas for constructing the interaction model between the surrounding rock and the lining are as follows: ① Lining bending moment: ， Where q is the uniformly distributed pressure of the surrounding rock, and x is the distance along the lining axis; ② Lining displacement: ， Based on the encoder data of the robotic arm and the flow sensor, the motion trajectory equation for printing and the material deposition thickness model are constructed. Bind the data from each sensor to the formulas in the above steps to initialize the digital twin; The data update cycle is determined based on the sensor sampling frequency to ensure that the digital twin displayed is consistent with the tunnel construction status.

9. The intelligent 3D printing method for complex tunnel lining according to claim 1, characterized in that, Step 3 specifically involves: Select the sensors to participate in data fusion; Variance is used to assess the reliability of data; the smaller the variance, the higher the reliability. The weight of each data point is determined based on its reliability percentage. Finally, multiplying each weight by each data point yields the fused data. Subsequently, based on continuous data collection, the weights are continuously updated, and data fusion optimization is carried out.

10. The intelligent 3D printing method for complex tunnel lining according to claim 1, characterized in that, Step 6 specifically involves: using sensors in the printed area for data acquisition and defect identification, such as using structured light cameras and laser scanners to detect the flatness, uniformity, and contour fit of the lining; using stress-strain sensors and displacement sensors to detect the maximum principal stress of the lining and the displacement rate of the surrounding rock, while marking high-stress and high-displacement risk areas; comparing the sensor data with the mechanical model of the digital twin to quantify the characteristics of the tunnel's mechanical environment and determine the direction of configuration optimization.

Images