WO2025156554A1 - 一种tbm隧道岩爆段支护方案智能设计方法 - Google Patents

一种tbm隧道岩爆段支护方案智能设计方法

Info

Publication number
WO2025156554A1
WO2025156554A1 PCT/CN2024/100145 CN2024100145W WO2025156554A1 WO 2025156554 A1 WO2025156554 A1 WO 2025156554A1 CN 2024100145 W CN2024100145 W CN 2024100145W WO 2025156554 A1 WO2025156554 A1 WO 2025156554A1
Authority
WO
WIPO (PCT)
Prior art keywords
tunnel
rockburst
surrounding rock
tbm
support scheme
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
PCT/CN2024/100145
Other languages
English (en)
French (fr)
Inventor
何本国
林波
冯夏庭
张恒源
付洪源
李红普
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Northeastern University China
Original Assignee
Northeastern University China
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Northeastern University China filed Critical Northeastern University China
Publication of WO2025156554A1 publication Critical patent/WO2025156554A1/zh
Anticipated expiration legal-status Critical
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model

Definitions

  • the present application relates to the field of tunnel engineering, and specifically to an intelligent design method for a support scheme for a rockburst section of a TBM tunnel.
  • Tunnels as engineering structures along transportation routes, offer significant social and economic benefits. Tunnel construction is a crucial component of every country's development. With the rapid development of the national economy, my country's underground resource development and infrastructure development are rapidly extending to deeper locations. Rockbursts occur occasionally in tunnels such as the Erlangshan Tunnel on the Sichuan-Tibet Highway, the Qinling Railway Tunnel, and the Sichuan-Tibet railway. This is because the geological environment in which the rock mass resides becomes more complex with increasing burial depth and ground stress levels, making rockburst hazards caused by excavation or mining more prominent and severe. This presents unprecedented challenges to the design, construction, and production of deep underground projects. Timely and effective tunnel support solutions can prevent and control rockbursts, playing a decisive role in tunnel safety and long-term stability.
  • TBM Transport-Borne Machine
  • TBM tunnel support schemes are primarily based on the experience of designers and construction technicians and engineering analogies. This leads to problems such as severe subjective influence and low design efficiency. With increasing demands for efficiency, quality, and safety, and the need for effective rockburst prevention and control, manual experience and engineering analogies are no longer sufficient for the development of tunnel support. Therefore, an intelligent design method for rockburst section support schemes in TBM tunnels is urgently needed.
  • the present invention provides an intelligent design method for a support scheme for a rock burst section of a TBM tunnel, which can provide on-site construction personnel with an efficient support scheme for the rock burst section of the tunnel.
  • An intelligent design method for a rockburst support scheme in a TBM tunnel comprises the following steps:
  • Step 1 Based on TBM operating data, surrounding rock feedback evaluation, tunnel engineering geological survey reports, and tunnel area topographic and geological maps, basic surrounding rock parameters are obtained to determine the surrounding rock classification for the initial support tunnel section and the stress distribution of the surrounding rock during tunnel excavation. Rockburst intensity assessment methods are then used to predict rockburst levels at different radial locations along the tunnel's trajectory.
  • Step 1.1 Based on TBM host operation data, surrounding rock feedback evaluation, tunnel engineering geological survey report, and tunnel area topography and geological map, obtain basic surrounding rock parameters. Based on these basic surrounding rock parameters, determine the surrounding rock classification for the tunnel section.
  • the basic parameters of surrounding rock mainly include burial depth, stratum lithology, surrounding rock classification, hydrogeological conditions, uniaxial compressive strength of rock, tunnel cross-section shape, and distribution of surrounding rock structural surfaces near the tunneling face;
  • Step 1.2 Based on the tunnel engineering geological survey report and the topographic and geological map of the tunnel area, numerical simulation is used to invert the initial in-situ stress of the tunnel rock and the stress field distribution caused by excavation, thereby determining the stress distribution of the surrounding rock during tunnel excavation.
  • Step 1.3 Based on the basic parameters of the surrounding rock and the stress distribution of the surrounding rock during tunnel excavation, a comprehensive judgment table of rockburst intensity at different radial positions along the tunnel direction is drawn to comprehensively analyze and predict the rockburst level.
  • rockburst levels specifically include no rockburst, slight rockburst, moderate rockburst and severe rockburst;
  • Step 2 Based on the surrounding rock classification and rockburst level, use computer deep learning technology to determine the initial support plan for the surrounding rock of the TBM tunnel;
  • Step 2.1 Construct a sample library that includes surrounding rock classification, rockburst grade, and corresponding initial support schemes for surrounding rock of completed or under-construction TBM tunnels.
  • Step 2.2 Construct a TBM tunnel initial support scheme selection model based on a neural network, and substitute the sample library into the TBM tunnel initial support scheme selection model for training;
  • Step 2.3 Input the surrounding rock classification and rockburst grade of the tunnel to be built into the trained TBM tunnel initial support scheme selection model to obtain the corresponding TBM tunnel initial support scheme.
  • the present invention provides an intelligent design method for a support scheme for a rockburst section of a TBM tunnel.
  • the intelligent design method for a support scheme for a rockburst section of a TBM tunnel is based on TBM host operation data, surrounding rock feedback evaluation, and a tunnel engineering geological survey report to determine the surrounding rock classification, and use a rockburst intensity assessment method to predict the rockburst grade.
  • the method uses computer deep learning technology to determine the initial surrounding rock support method.
  • the method effectively avoids the influence of subjective factors, actively utilizes TBM host operation data and surrounding rock feedback evaluation, combines big data processing and artificial intelligence technology, and proposes an intelligent design support scheme for the rockburst section of a TBM tunnel.
  • the method can provide on-site construction personnel with an efficient support scheme for the rockburst section of the tunnel.
  • FIG1 is a flow chart of an intelligent design method for a rockburst support scheme for a TBM tunnel provided by an embodiment of the present invention
  • FIG2 is a schematic diagram of a monitoring page of a TBM host at a certain excavation moment of a TBM tunnel provided by an embodiment of the present invention
  • FIG3 is a schematic diagram of a TBM tunnel initial support scheme selection model based on an existing TBM tunnel initial support scheme library using deep neural network technology, provided by an embodiment of the present invention
  • FIG4 is a schematic diagram of a plane expansion of a rockburst level prediction for a TBM tunnel provided by an embodiment of the present invention.
  • FIG5 is a flowchart of the selection of the initial support scheme for the surrounding rock of a TBM tunnel provided by an embodiment of the present invention.
  • An intelligent design method for a rockburst support scheme in a TBM tunnel specifically includes the following steps:
  • Step 1 Based on TBM operating data, surrounding rock feedback evaluation, tunnel engineering geological survey reports, and tunnel area topographic and geological maps, basic surrounding rock parameters are obtained to determine the surrounding rock classification for the initial support tunnel section and the stress distribution of the surrounding rock during tunnel excavation. Rockburst intensity assessment methods are then used to predict rockburst levels at different radial locations along the tunnel's trajectory.
  • Step 1.1 Based on TBM host operation data, surrounding rock feedback evaluation, tunnel engineering geological survey reports, and tunnel area topographic and geological maps, obtain basic surrounding rock parameters. These parameters mainly include burial depth, stratum lithology, surrounding rock classification, hydrogeological conditions, uniaxial compressive strength of rock, tunnel cross-section shape, and distribution of surrounding rock structural surfaces near the tunnel face. Based on these basic surrounding rock parameters, determine the surrounding rock classification for the tunnel section.
  • FIG. 2 shows a schematic diagram of the TBM host monitoring page at a certain moment of tunnel excavation in a TBM tunnel. It reflects the TBM host status (TBM excavation footage data, cutterhead torque, total thrust, penetration, excavation speed, pressure and displacement of the left gripper, gripper, top shield, left shield, and right shield) and the tunnel face rock mass evaluation.
  • TBM host status TBM excavation footage data, cutterhead torque, total thrust, penetration, excavation speed, pressure and displacement of the left gripper, gripper, top shield, left shield, and right shield
  • the tunnel engineering geological survey report reflects the surrounding rock status during the tunnel design stage. The former results should be given priority, and the latter results should be supplemented.
  • Step 1.2 Based on the tunnel engineering geological survey report and the topographic and geological map of the tunnel area, numerical simulation is used to invert the initial in-situ stress of the tunnel rock and the distribution of the excavation disturbance stress field, and determine the stress distribution of the surrounding rock during the tunnel excavation project.
  • Step 1.3 Based on the basic parameters of the surrounding rock and the stress distribution of the surrounding rock during tunnel excavation, a comprehensive judgment table of rockburst intensity at different radial positions along the tunnel direction is drawn.
  • the rockburst level is comprehensively analyzed and predicted.
  • the rockburst level includes no rockburst, slight rockburst, moderate rockburst, and severe rockburst.
  • the stress distribution of the surrounding rock during tunnel excavation is often asymmetric, stress concentration occurs in certain areas of the surrounding rock. Predicted rockbursts often occur in these areas. Therefore, the stress distribution of the surrounding rock during tunnel excavation is used to determine the location of stress concentration, and accordingly, more effective support measures are required at these locations. Therefore, the predicted rockburst level not only indicates the rockburst level of a specific section, but also indicates the stress concentration area at a specific location in the surrounding rock, as shown in Figure 4.
  • rockburst severity prediction methods Based on the tunnel's characteristics, an appropriate rockburst severity prediction method is selected to predict the tunnel's rockburst severity.
  • the main rockburst severity prediction methods include engineering geological analysis, rock mechanics criteria, RVI index, neural networks, and rockburst microseismic early warning.
  • a tunnel was used as an example.
  • Four criteria applicable to the tunnel were selected, as shown below.
  • a comprehensive rockburst severity judgment table was then developed to comprehensively predict the surrounding rockburst severity. If the four rockburst severity prediction results for a particular section of the tunnel disagree, the majority prediction will be used as the final prediction.
  • Step 2 Based on the surrounding rock classification and rockburst level, use computer deep learning technology to determine the initial support plan for the surrounding rock of the TBM tunnel.
  • Neural networks are a newly emerging discipline, inspired by the widespread adoption of computers. They are a complex computational method that simulates the structure of neurons and their connections, inspired by the human nervous system. Neural network technology is primarily designed based on the human neural workflow, utilizing the way humans process information. Neural networks do not require a predefined mathematical equation for the mapping between input and output. Instead, they train themselves to learn certain rules and generate the closest possible output for a given input.
  • a classic neural network model usually consists of an input layer, a hidden layer, and an output layer. The layers are fully interconnected, and the nodes in each layer are not connected. There can be multiple hidden layers. The neural network is reshaped by continuous self-deduction and then the final result is obtained.
  • neural network models with various structures, such as BP neural network, convolutional neural network, recurrent neural network, RBF neural network, etc.
  • a neural network structure suitable for the on-site tunnel can be selected. This application takes the following neural network structure as an example.
  • Step 2.1 Construct a sample library, which includes surrounding rock classification, rockburst grade, and corresponding surrounding rock initial support schemes for completed or under-construction TBM tunnels.
  • Step 2.2 Construct a TBM tunnel initial support scheme selection model based on a neural network, and substitute the sample library into the TBM tunnel initial support scheme selection model for training.
  • the deep neural network used in the model for selecting the initial support scheme for TBM tunnels is set to five layers: an input layer, three hidden layers, and an output layer.
  • the input layer corresponds to two parameters: surrounding rock classification and rockburst severity. These parameters are selected based on their characteristics. For example, there are six types of surrounding rock classification: I, II, III, IV, V, and VI, and four types of rockburst: no rockburst, mild rockburst, moderate rockburst, and severe rockburst, resulting in two nodes.
  • the output layer corresponds to the initial support scheme for TBM tunnels, resulting in one node.
  • the number of nodes in the hidden layer is determined through research; it is generally recommended that the number of nodes be less than twice the number of characteristic items (the number of input layer nodes).
  • a tangent function or a logarithmic function is used as the transfer function (activation function of the hidden layer), as shown in Figure 3.
  • the trained deep learning artificial network selection model is obtained, which is a usable TBM tunnel initial support scheme selection model.
  • the accuracy rate continues to improve as more sample library data is generated during construction, resulting in high accuracy and intelligence.
  • the sample library data will contain no fewer than 100 groups.
  • Step 2.3 Input the surrounding rock classification and rockburst level of the tunnel to be built into the trained TBM tunnel initial support scheme selection model to obtain the corresponding TBM tunnel initial support scheme, as shown in Figure 5.
  • the initial support scheme for the surrounding rock stress concentration area in the initial support scheme of TBM tunnels needs to be optimized to a more effective support method, such as shortening the spacing between steel bars and arranging anchor bolts densely, so as to better prevent the occurrence of rock bursts.
  • the initial support scheme selection model for TBM tunnels it can be understood that the surrounding rock classification, rock burst level, and initial support method of the surrounding rock stress concentration area of the tunnel section are used as inputs, and the initial support scheme for the surrounding rock stress concentration area is used as the output node.
  • the initial support scheme for the surrounding rock stress concentration area is obtained through training as the output node.
  • the initial support measures adopted in a tunnel section with a surrounding rock of medium rock burst level III are as follows:

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Geometry (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Computer Hardware Design (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Architecture (AREA)
  • Civil Engineering (AREA)
  • Structural Engineering (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Investigation Of Foundation Soil And Reinforcement Of Foundation Soil By Compacting Or Drainage (AREA)
  • Excavating Of Shafts Or Tunnels (AREA)

Abstract

一种TBM隧道岩爆段支护方案智能设计方法，涉及隧道工程领域；该TBM隧道岩爆段支护方案智能设计方法基于TBM主机运行数据与围岩反馈评价、隧道工程地质勘察报告，确定围岩分级，运用岩爆烈度评估方法，预测岩爆等级；基于围岩分级与岩爆等级，运用计算机深度学习技术，确定围岩初期支护方法；该方法有效避免了主观因素影响，积极利用了TBM主机运行数据与围岩反馈评价，结合大数据处理与人工智能技术，提出TBM隧道岩爆段智能设计支护方案，可以为现场施工人员牯弟隧道岩爆段提供一种高效曲主护方案。

Description

一种TBM隧道岩爆段支护方案智能设计方法 技术领域
本申请涉及隧道工程领域，具体涉及一种TBM隧道岩爆段支护方案智能设计方法。
背景技术
隧道作为交通运输线路上的工程构筑物，具有重大的社会、经济效益。隧道建设是各个国家发展进程中的重要组成部分，随着国民经济的快速发展，我国地下资源开发、地下基础建设迅速向深部延伸，如川藏公路二郎山隧道、秦岭铁路隧道、川藏铁路等隧道岩爆时而发生，这是由于随着埋深和地应力水平的增加，岩体所处的地质环境变得更为复杂，开挖或开采引发的岩爆灾害变得更加突出和严重，这给深部地下工程的设计、施工和生产等带来了前所未有的挑战。而及时有效的隧道支护方案能预防与控制岩爆，对隧道安全建设以及长期稳定性起到决定性作用。
TBM隧道在深部硬岩掘进隧道中得到广泛应用，而TBM隧道支护方案的选取主要依据是设计单位和施工单位技术员的经验及工程类比法，这导致隧道支护存在受主观影响严重、设计效率低等问题。随着对效率、质量、安全要求越来越高，对岩爆的有效预防及控制的要求也越来越高，人工经验与工程类比法已无法满足隧道支护的发展要求，亟需提出一种智能化的TBM隧道岩爆段支护方案智能设计方法。
发明内容
针对现有技术的不足，本发明提供一种TBM隧道岩爆段支护方案智能设计方法，可以为现场施工人员针对隧道岩爆段提供一种高效的支护方案。
一种TBM隧道岩爆段支护方案智能设计方法，具体包括以下步骤：
步骤1：基于TBM主机运行数据与围岩反馈评价、隧道工程地质勘察报告、隧道区域地形地质图，获取围岩基本参数，确定初期支护隧道段落的围岩分级以及隧道开挖工程中围岩的应力分布情况，进而运用岩爆烈度评估方法，预测隧道走向不同位置径向不同位置岩爆等级；
步骤1.1：基于TBM主机运行数据与围岩反馈评价、隧道工程地质勘察报告、隧道区域地形地质图，获取围岩基本参数，根据围岩基本参数，确定该隧道段落的围岩分级；
所述围岩基本参数主要包括埋深、地层岩性、围岩分级、水文地质条件、岩石单轴抗压强度、隧道断面形状、掘进掌子面附近围岩结构面分布情况；
步骤1.2：基于隧道工程地质勘察报告与隧道区域地形地质图，通过数值模拟的方法反演出隧道原岩初始地应力和开挖扰动应力场分布，确定隧道开挖工程中围岩的应力分布情况；
步骤1.3：根据围岩基本参数与隧道开挖工程中围岩的应力分布情况，绘制隧道走向不同位置径向不同位置隧道岩爆强度综合判断表，综合分析预测岩爆等级；
所述岩爆等级具体包括无岩爆、轻微岩爆、中等岩爆以及强烈岩爆；
步骤2：基于围岩分级与岩爆等级，运用计算机深度学习技术，确定TBM隧道围岩初期支护方案；
步骤2.1：构建样本库，该样本库包括围岩分级、岩爆等级与对应的已建成或者在建的TBM隧道围岩初期支护方案；
步骤2.2：基于神经网络构建TBM隧道初期支护方案选取模型，将样本库代入所述TBM隧道初期支护方案选取模型进行训练；
步骤2.3：向训练后的TBM隧道初期支护方案选取模型输入待建隧道的围岩分级、岩爆等级获得对应得TBM隧道初期支护方案。
本发明的有用效果是：
本发明提供了一种TBM隧道岩爆段支护方案智能设计方法，该TBM隧道岩爆段支护方案智能设计方法基于TBM主机运行数据与围岩反馈评价、隧道工程地质勘察报告，确定围岩分级，运用岩爆烈度评估方法，预测岩爆等级；基于围岩分级与岩爆等级，运用计算机深度学习技术，确定围岩初期支护方法；该方法有效避免了主观因素影响，积极利用了TBM主机运行数据与围岩反馈评价，结合大数据处理与人工智能技术，提出TBM隧道岩爆段智能设计支护方案，可以为现场施工人员针对隧道岩爆段提供一种高效的支护方案。
附图说明
图1是本发明实施案例提供的TBM隧道岩爆段支护方案智能设计方法的流程图；
图2是本发明实施案例提供的某TBM隧道某个掘进时刻TBM主机的监控页面示意图；
图3是本发明实施案例提供的利用深度神经网络技术，基于已有TBM隧道初期支护方案库的TBM隧道初期支护方案选取模型示意图；
图4是本发明实施案例提供的预测TBM隧道岩爆等级平面展开示意图；
图5是本发明实施案例提供的TBM隧道围岩初期支护方案选取流程图。
具体实施方式
下面结合附图和实施例，对本发明做进一步说明；
一种TBM隧道岩爆段支护方案智能设计方法，如图1所示，具体包括以下步骤：
步骤1：基于TBM主机运行数据与围岩反馈评价、隧道工程地质勘察报告、隧道区域地形地质图，获取围岩基本参数，确定初期支护隧道段落的围岩分级以及隧道开挖工程中围岩的应力分布情况，进而运用岩爆烈度评估方法，预测隧道走向不同位置径向不同位置岩爆等级；
步骤1.1：基于TBM主机运行数据与围岩反馈评价、隧道工程地质勘察报告、隧道区域地形地质图，获取围岩基本参数，围岩基本参数主要包括埋深、地层岩性、围岩分级、水文地质条件、岩石单轴抗压强度、隧道断面形状、掘进掌子面附近围岩结构面分布情况，根据围岩基本参数，确定该隧道段落的围岩分级。
需要说明的是TBM主机运行数据与围岩反馈评价体现的是隧道掘进过程中实时的TBM主机状态与围岩状态，如图2为某TBM隧道某个掘进时刻TBM主机的监控页面示意图，体现了TBM主机状态(TBM掘进进尺数据，刀盘扭矩，总推力，贯入度，掘进速度，左撑靴、撑靴、顶护盾、左侧护盾、右侧护盾压力和位移)与掌子面岩体评价，而隧道工程地质勘察报告体现的是隧道设计阶段围岩状态，应以前者结果为主，后者结果为辅；
步骤1.2：基于隧道工程地质勘察报告与隧道区域地形地质图，通过数值模拟的方法反演出隧道原岩初始地应力和开挖扰动应力场分布，确定隧道开挖工程中围岩的应力分布情况。
步骤1.3：根据围岩基本参数与隧道开挖工程中围岩的应力分布情况，绘制隧道走向不同位置径向不同位置隧道岩爆强度综合判断表，综合分析预测岩爆等级，其中岩爆等级具体包括无岩爆、轻微岩爆、中等岩爆以及强烈岩爆。
因为隧道开挖工程中围岩的应力分布情况常常不对称，导致隧道围岩某个部位应力集中，而预测的岩爆常常从此处发生，所以根据隧道开挖工程中围岩的应力分布情况知道围岩应力集中部位，而相应的在此处的支护手段需要更为有效。故预测的岩爆等级不仅显现某个段落的岩爆等级，而且要显现围岩某个位置的应力集中区域，如图4所示。
国内外专家针对岩爆预测理论提出了多种岩爆等级预测方法，依据隧道自身特性选择适用的岩爆等级预测方法，对隧道的岩爆等级进行预测，目前对于岩爆等级预测方法主要有工程地质分析法、岩石力学判据法、RVI指标法、神经网络法和岩爆微震预警法。本实施案例中，以某隧道为实例，该隧道选取适用于该隧道的4个判断依据，如下所示，并绘制岩爆强度综合判断表，对围岩进行岩爆等级的综合预测。若该隧道某段4个岩爆等级预测结果不一致，将采用多数预测结果作为最终预测结果。
1)埋深；
2)围岩级别；
3)岩石应力强度比法σθ/σc，其中σθ开挖面洞周最大切向应力(MPa，与岩石饱和单轴抗压强度(MPa)；
4)岩石强度应力比法σc/σmax，其中σc为岩石饱和单轴抗压强度(MPa)，σmax为围岩的最大地应力(MPa)；
步骤2：基于围岩分级与岩爆等级，运用计算机深度学习技术，确定TBM隧道围岩初期支护方案。
神经网络是计算机大面积普及后的一门新兴学科，它是受到人脑神经系统的启发，模拟人脑神经元及神经元连接结构的一种复杂的计算方法，神经网络技术主要以人神经的工作流程为设计的范本，利用人神经处理相关内容的方式进行计算。神经网络无需事先确定输入输出之间映射关系的数学方程，仅通过自身的训练，学习某种规则，在给定的输入值时得到最接近期望输出值的结果。
经典的神经网络模型通常由输入层、隐含层和输出层组成，层与层之间全互连，每层节点之间不相连，隐含层可以有多个，其通过不断自我反复推演重塑神经网络，然后得到最终结果，随着国内外专家学者对神经网络的研究与应用，有多种结构的神经网络模型，如BP神经网络、卷积神经网络、循环神经网络、RBF神经网络等，为了达到TBM隧道岩爆段支护方案智能设计的目的，可以选择适合现场隧道的神经网络结构，本申请以如下神经网络结构示例。
步骤2.1：构建样本库，该样本库包括围岩分级、岩爆等级与对应的已建成或者在建的TBM隧道围岩初期支护方案。
某隧道的对应特征如下表所示：

步骤2.2：基于神经网络构建TBM隧道初期支护方案选取模型，将样本库代入所述TBM隧道初期支护方案选取模型进行训练。
对于TBM隧道初期支护方案选取模型的深度神经网络，设置其网络层数为5层，即输入层，3层隐含层和输出层；其中输入层对应2个参数，即围岩分级、岩爆等级，选其特征做为参数，比如围岩级别分为6种：Ⅰ、Ⅱ、Ⅲ、Ⅳ、Ⅴ、Ⅵ级，岩爆等级分为4种：无岩爆、轻微岩爆、中等岩爆、强烈岩爆，故有2个节点，而输出层对应TBM隧道初期支护方案，故有1个节点，隐含层节点数通过研究确定，一般情况下建议节点数应该小于特征项个数(输入层节点数)的2倍。选用正切函数或者对数函数作为传递函数(隐含层的激活函数)，如附图3所示。
神经网络建立后，输入大量的样本库对深度学习人工网络选取模型训练直至收敛，达到设定误差标准，获得训练完成的深度学习人工网络选取模型即获得可用的TBM隧道初期支护方案选取模型。
通过大量的样本库数据训练，且TBM隧道初期支护方案选取模型具有自学习功能，随着施工过程中产生的样本库数据的增多，准确率不断提高，具有准确性高、智能化程度高的特点。在具体实施时，样本库的数据不少于100组。
步骤2.3：向训练后的TBM隧道初期支护方案选取模型输入待建隧道的围岩分级、岩爆等级获得对应得TBM隧道初期支护方案。如附图5所示。
TBM隧道初期支护方案中围岩应力集中区域的初期支护方案需优化为更有效的支护方法，如缩短钢筋排间距，加密布置锚杆手段，以此来更好的预防岩爆的发生，在TBM隧道初期支护方案选取模型中可理解为以该隧道段落的围岩应力集中区域的围岩分级、岩爆等级、TBM隧道初期支护方作为输入，围岩应力集中区域的初期支护方案作为输出节点，通过训练得到围岩应力集中区域的初期支护方案作为输出节点，如某个预测为中等岩爆Ⅲ级围岩隧道段落采用的初期支护措施如下表：
则该隧道段落围岩应力集中区域的初期支护措施应优化可如下表：

Claims (7)

  1. 一种TBM隧道岩爆段支护方案智能设计方法，其特征在于，具体包括以下步骤：
    步骤1：基于TBM主机运行数据与围岩反馈评价、隧道工程地质勘察报告、隧道区域地形地质图，获取围岩基本参数，确定初期支护隧道段落的围岩分级以及隧道开挖工程中围岩的应力分布情况，进而运用岩爆烈度评估方法，预测隧道走向不同位置径向不同位置岩爆等级；
    步骤2：基于围岩分级与岩爆等级，运用计算机深度学习技术，确定TBM隧道围岩初期支护方案。
  2. 根据权利要求1所述的一种TBM隧道岩爆段支护方案智能设计方法，其特征在于，步骤1具体为：
    步骤1.1：基于TBM主机运行数据与围岩反馈评价、隧道工程地质勘察报告、隧道区域地形地质图，获取围岩基本参数，根据围岩基本参数，确定该隧道段落的围岩分级；
    步骤1.2：基于隧道工程地质勘察报告与隧道区域地形地质图，通过数值模拟的方法反演出隧道原岩初始地应力和开挖扰动应力场分布，确定隧道开挖工程中围岩的应力分布情况；
    步骤1.3：根据围岩基本参数与隧道开挖工程中围岩的应力分布情况，绘制隧道走向不同位置径向不同位置隧道岩爆强度综合判断表，综合分析预测岩爆等级。
  3. 根据权利要求2所述的一种TBM隧道岩爆段支护方案智能设计方法，其特征在于，所述围岩基本参数主要包括埋深、地层岩性、围岩分级、水文地质条件、岩石单轴抗压强度、隧道断面形状、掘进掌子面附近围岩结构面分布情况。
  4. 根据权利要求2所述的一种TBM隧道岩爆段支护方案智能设计方法，其特征在于，所述岩爆等级具体包括无岩爆、轻微岩爆、中等岩爆以及强烈岩爆。
  5. 根据权利要求1所述的一种TBM隧道岩爆段支护方案智能设计方法，其特征在于，步骤2具体为：
    步骤2.1：构建样本库；
    步骤2.2：基于神经网络构建TBM隧道初期支护方案选取模型，将样本库代入所述TBM隧道初期支护方案选取模型进行训练；
    步骤2.3：获得对应得TBM隧道初期支护方案。
  6. 根据权利要求5所述的一种TBM隧道岩爆段支护方案智能设计方法，其特征在于，所述样本库包括围岩分级、岩爆等级与对应的已建成或者在建的TBM隧道围岩初期支护方案。
  7. 根据权利要求5所述的一种TBM隧道岩爆段支护方案智能设计方法，其特征在于，步骤2.3具体为：向训练后的TBM隧道初期支护方案选取模型输入待建隧道的围岩分级、岩爆等级获得对应得TBM隧道初期支护方案。
PCT/CN2024/100145 2024-01-24 2024-06-19 一种tbm隧道岩爆段支护方案智能设计方法 Pending WO2025156554A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202410096786.5A CN117910112A (zh) 2024-01-24 2024-01-24 一种tbm隧道岩爆段支护方案智能设计方法
CN202410096786.5 2024-01-24

Publications (1)

Publication Number Publication Date
WO2025156554A1 true WO2025156554A1 (zh) 2025-07-31

Family

ID=90689415

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2024/100145 Pending WO2025156554A1 (zh) 2024-01-24 2024-06-19 一种tbm隧道岩爆段支护方案智能设计方法

Country Status (2)

Country Link
CN (1) CN117910112A (zh)
WO (1) WO2025156554A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN121033590A (zh) * 2025-10-29 2025-11-28 中铁十四局集团有限公司 基于图像与深度学习的tbm支护等级决策方法与系统
CN121615235A (zh) * 2026-02-03 2026-03-06 江西省交通投资集团有限责任公司 一种隧道岩土结构分析方法及系统

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115758515B (zh) * 2022-10-31 2024-10-25 盾构及掘进技术国家重点实验室 Tbm隧道不良地质段智能支护决策方法
CN117910112A (zh) * 2024-01-24 2024-04-19 东北大学 一种tbm隧道岩爆段支护方案智能设计方法

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110374636A (zh) * 2019-07-09 2019-10-25 中铁十七局集团第二工程有限公司 一种隧道岩爆安全防护的施工方法
CN110648082A (zh) * 2019-10-08 2020-01-03 东北大学 一种深埋硬岩隧道岩爆等级评估的快速查表方法
US20210209263A1 (en) * 2019-03-08 2021-07-08 Shandong University Tunnel tunneling feasibility prediction method and system based on tbm rock-machine parameter dynamic interaction mechanism
CN115758515A (zh) * 2022-10-31 2023-03-07 中铁隧道局集团有限公司 Tbm隧道不良地质段智能支护决策方法
CN116066108A (zh) * 2023-01-13 2023-05-05 东北大学 一种非对称高应力隧道掌子面超前爆破卸压岩爆控制方法
CN116758343A (zh) * 2023-05-30 2023-09-15 山东大学 一种tbm施工隧道围岩智能自动分级方法及系统
CN117910112A (zh) * 2024-01-24 2024-04-19 东北大学 一种tbm隧道岩爆段支护方案智能设计方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20210209263A1 (en) * 2019-03-08 2021-07-08 Shandong University Tunnel tunneling feasibility prediction method and system based on tbm rock-machine parameter dynamic interaction mechanism
CN110374636A (zh) * 2019-07-09 2019-10-25 中铁十七局集团第二工程有限公司 一种隧道岩爆安全防护的施工方法
CN110648082A (zh) * 2019-10-08 2020-01-03 东北大学 一种深埋硬岩隧道岩爆等级评估的快速查表方法
CN115758515A (zh) * 2022-10-31 2023-03-07 中铁隧道局集团有限公司 Tbm隧道不良地质段智能支护决策方法
CN116066108A (zh) * 2023-01-13 2023-05-05 东北大学 一种非对称高应力隧道掌子面超前爆破卸压岩爆控制方法
CN116758343A (zh) * 2023-05-30 2023-09-15 山东大学 一种tbm施工隧道围岩智能自动分级方法及系统
CN117910112A (zh) * 2024-01-24 2024-04-19 东北大学 一种tbm隧道岩爆段支护方案智能设计方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
CHEN WEIZHONG, XIAO ZHENGLONG; TIAN HONGMING: "RESEARCH ON SQUEEZING LARGE DISPLACEMENT AND ITS DISPOSING METHOD OF WEAK ROCK TUNNEL UNDER HIGH IN-SITU STRESS", CHINESE JOURNAL OF ROCK MECHANICS AND ENGINEERING, vol. 34, no. 11, 15 November 2015 (2015-11-15), pages 2215 - 2226, XP093340537, DOI: 10.13722/j.cnki.jrme.2015.1000 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN121033590A (zh) * 2025-10-29 2025-11-28 中铁十四局集团有限公司 基于图像与深度学习的tbm支护等级决策方法与系统
CN121615235A (zh) * 2026-02-03 2026-03-06 江西省交通投资集团有限责任公司 一种隧道岩土结构分析方法及系统

Also Published As

Publication number Publication date
CN117910112A (zh) 2024-04-19

Similar Documents

Publication Publication Date Title
WO2025156554A1 (zh) 一种tbm隧道岩爆段支护方案智能设计方法
US20240378337A1 (en) Mine stress field twin modeling assimilation system for full space-time mining process, and method
CN111553000A (zh) 钻爆法隧道智能建造方法
CN109934398A (zh) 一种钻爆法施工隧道瓦斯危险等级预测方法和装置
Qiu et al. TBM tunnel surrounding rock classification method and real-time identification model based on tunneling performance
CN109345140B (zh) 一种煤矿突水灾害预警的辅助方法
CN110147875A (zh) 一种基于lstm神经网络的盾构机辅助巡航方法
CN114818493B (zh) 一种隧道岩体完整程度定量评价的方法
CN115481565A (zh) 一种基于lstm和蚁群算法的土压平衡盾构机掘进参数预测方法
CN116797017A (zh) 一种基于模糊数学理论的tbm施工风险评价方法及系统
CN119538359A (zh) 一种地层变形预测方法及系统
CN114943363A (zh) 一种煤矿瓦斯涌出量预测方法
Lv et al. A novel stacking-based ensemble learning model for drilling efficiency prediction in earth-rock excavation
CN118332667A (zh) 基于掌子面信息的隧道稳定性智能判定方法及系统
Hai et al. Multi-element drilling parameter optimization based on drillstring dynamics and ROP model
Zhang et al. Prediction of longitudinal surface settlement in composite formation using large-diameter shield machine based on machine learning techniques
CN120373080A (zh) 深埋隧洞tbm掘进卡机风险预测方法、系统及设备
CN119754787A (zh) 一种富水地层电缆盾构隧道掘进开挖面稳定控制方法
CN117196041A (zh) 一种基于案例推理的tbm卡机脱困方案智能决策方法
CN116976498A (zh) 一种基于人工智能的盾构施工适应性评价方法
CN118468404A (zh) 一种基于bim和改进topsis算法的隧道设计方案评估方法及系统
CN117763959A (zh) 一种改进cnn的爆破设计参数优化方法
Wang et al. Comparative study on prediction models of EPB shield tunnelling parameters in the water-rich round-gravel formation
CN117172113A (zh) 一种旋转导向钻井井眼轨迹预测方法、系统、设备及介质
CN116975623A (zh) 钻爆法隧道施工阶段大变形等级预测方法、装置及介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 24919639

Country of ref document: EP

Kind code of ref document: A1