A Case-Based Reasoning Method for Tunnel Boring Machine (TBM) Selection

Abstract

This invention discloses a shield tunneling machine (TBM) selection method based on case-based reasoning, relating to the field of TBM engineering selection technology; it includes the following steps: S01. Collecting past TBM engineering information, extracting feature parameters affecting TBM selection from past TBM engineering information, and storing them in a case library; S02. Extracting feature parameters affecting TBM selection from newly constructed TBM projects; S03. Searching the case library based on the feature parameters extracted from the newly constructed TBM projects, and selecting similar past TBM projects as recommended solutions based on the similarity of the search results; S04. Modifying and adjusting the selected recommended selection solutions according to the actual construction needs of the newly constructed TBM projects, and outputting the final solution; S05. Constructing according to the final solution, and storing the project instance in the case library after construction is completed.

Description

Technical Field

[0001] This invention relates to the field of shield tunneling engineering selection technology, and more specifically, to a shield tunneling selection method based on case-based reasoning. Background Technology

[0002] Shield tunneling is one of the most advanced construction methods for building underground tunnels. Shields are "tailor-made" based on specific characteristics such as engineering geology, hydrogeology, topography, surface buildings, and underground pipelines and structures. Unlike other equipment, the core technology of shield tunneling lies not only in the electromechanical industrial design of the equipment itself but also in how the equipment is adapted to various engineering geological conditions. The success rate of shield tunneling construction depends primarily on the selection of the shield type, which determines whether the shield is suitable for the on-site construction environment. The correctness of the shield selection directly determines the success or failure of the shield tunneling project.

[0003] Traditionally, the selection of tunnel boring machine (TBM) equipment involved collecting data such as geological survey reports, hydrogeological information, design documents, and surrounding environmental conditions. Experts in the field would then discuss the selection, determining the key parameters of the appropriate TBM based on past engineering experience and scientific reasoning. This traditional method was labor-intensive, time-consuming, and required human intervention. The evaluation process lacked scientific judgment and was prone to introducing subjective factors, hindering the accurate assessment of the TBM selection and potentially leading to engineering risks. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies, this invention provides a shield tunneling selection method based on case-based reasoning. By storing previous shield tunneling projects, a construction method can be quickly and accurately prepared by retrieving similar projects from the past when constructing a new project.

[0005] The technical solution adopted by this invention to solve its technical problem is: a shield tunneling machine selection method based on case-based reasoning, the improvement of which includes the following steps:

[0006] S01. Collect information on past tunnel boring machine (TBM) projects, extract characteristic parameters that affect TBM selection from the past TBM project information, and store them in the case library;

[0007] S02. Extract characteristic parameters that affect shield selection in newly constructed shield tunneling projects;

[0008] S03. Based on the feature parameters extracted from the newly constructed shield tunneling project, a search is conducted in the case library, and similar previous shield tunneling projects are selected as recommended solutions based on the similarity of the search results.

[0009] S04. Based on the actual construction needs of the new shield tunneling project, modify and adjust the selected recommended selection schemes, and output the final scheme;

[0010] S05. Carry out the construction according to the final plan, and store the project example in the case library after the construction is completed.

[0011] In step S03 of the above technical solution, during the retrieval process, it is determined whether the construction plan of the newly built shield tunneling project and similar shield tunneling projects needs to be adjusted. If the construction plan needs to be adjusted, it is adjusted according to the needs of the newly built shield tunneling project and stored as a new case in the case library; if no adjustment is needed, it is adjusted according to the actual construction needs of the newly built shield tunneling project and then stored in the case library.

[0012] In step S03 of the above technical solution, the search is performed according to three categories: soil layer, rock layer, and soil-rock composite layer of the construction site.

[0013] The soil layer retrieval described in the above technical solution is divided into three parts: particle size distribution classification, construction indicators, and geotechnical indicators. The soil layers are classified into multiple categories according to particle size distribution. The construction indicators are the shield diameter and the shield burial depth, where the shield burial depth is the depth of the shield below the ground. The geotechnical indicators are first classified into different categories based on the particle size distribution curve, and then the geotechnical indicators are analyzed based on the typical adaptability problems of shield construction in various types of strata to extract stratum characteristic parameters. Among these, the stratum characteristic parameters are consistency index, excavation face water head, permeability coefficient, equivalent quartz content, limiting particle size, and maximum particle size.

[0014] The strata retrieval described in the above technical solution is divided into three parts: compressive strength classification, construction indicators, and geotechnical indicators. The compressive strength classification divides the strata into multiple categories according to the surrounding rock firmness coefficient. The construction indicators are the shield diameter and the shield burial depth, where the shield burial depth is the depth of the shield below the ground. The geotechnical indicators are first divided into different categories based on the compressive strength, and then the geotechnical indicators are analyzed based on the typical adaptability problems of shield tunneling in various strata to extract stratum characteristic parameters. Among these, the stratum characteristic parameters are the integrity coefficient, karst development, and groundwater conditions expressed as the excavation face water head. The karst development is expressed as the linear dissolution rate.

[0015] The retrieval of the soil-rock composite layer in the above technical solution is based on the classification of soil and rock layers according to particle size distribution and compressive strength, respectively, to determine the type of soil and rock layers in the composite stratum, and then to retrieve the composite layer according to the composite ratio of soil and rock layers in the composite stratum.

[0016] The similarity of the shield diameters is expressed by the following formula: Where a and b represent the shield diameters of newly constructed shield tunneling projects and previous shield tunneling projects in the case library, respectively, and the intervals are... This indicates the range of values ​​for the shield diameter, which depends on the specific circumstances.

[0017] The shield tunneling depth mentioned in the above technical solution is an interval value. The similarity between the shield tunneling depth of a newly constructed shield tunneling project and previous shield tunneling projects in the case library is expressed by the following formula: , where x1 and x2 represent the minimum and maximum values ​​of the x interval, and y1 and y2 represent the minimum and maximum values ​​of the y interval.

[0018] In the above technical solution, after obtaining the similarity between the shield tunneling depth and the shield diameter, the global similarity of the construction indicators between the newly built shield tunneling project and previous shield tunneling projects in the case library is determined by the following formula: ,in, To build global similarity of indicators, This represents the local similarity of instance attributes.

[0019] In the above technical solution, when calculating the similarity of feature parameters, the magnitudes of the components in the parameter set differ significantly, resulting in a large difference in the dispersion of the two variables. Mahalanobis distance is used to correct this scale inconsistency. Where S is the covariance matrix, and when the covariance matrix is ​​the identity matrix, the Mahalanobis distance is simplified to the Euclidean distance.

[0020] The beneficial effects of this invention are as follows: The method provided by this invention collects information on previous shield tunneling projects and extracts their characteristic parameters. When constructing new shield tunneling projects, the characteristic parameters of the new projects are extracted and compared in a case library, which can quickly and accurately select the selection scheme. If there is no similar scheme in the case library, the project information will be stored in the case library after the new shield tunneling project is completed. The case library is a dynamic case library and keeps incrementally learning, which provides convenience for the selection of shield tunneling projects in subsequent new projects. Attached Figure Description

[0021] Figure 1 This is a flowchart of a shield tunneling machine selection method based on case-based reasoning according to the present invention.

[0022] Figure 2 This is a flowchart of a shield tunneling machine selection method based on case-based reasoning according to the present invention.

[0023] Figure 3 This is a flowchart illustrating the soil layer retrieval process in a shield tunneling selection method based on case-based reasoning, as described in this invention.

[0024] Figure 4 This is an example diagram of particle gradation curves in a shield tunneling selection method based on case-based reasoning according to the present invention.

[0025] Figure 5 This is a flowchart of rock strata retrieval in a shield tunnel selection method based on case-based reasoning according to the present invention.

[0026] Figure 6This diagram illustrates an embodiment of information storage in the case library of a shield tunneling machine selection method based on case reasoning, as described in this invention. Detailed Implementation

[0027] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0028] The following will clearly and completely describe the concept, specific structure, and technical effects of the present invention in conjunction with embodiments and accompanying drawings, so as to fully understand the purpose, features, and effects of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. Other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are all within the scope of protection of the present invention. Furthermore, all connections / linkages involved in the patent do not simply refer to direct contact between components, but rather to the ability to form a better connection structure by adding or reducing connecting accessories according to specific implementation conditions. The various technical features in this invention can be combined interactively without contradicting each other.

[0029] Reference Figure 1 As shown in the figure, this invention provides a shield tunneling machine selection method based on case-based reasoning, including the following steps:

[0030] S01. Collect information on past tunnel boring machine (TBM) projects, extract characteristic parameters affecting TBM selection from this information, and store them in a case library. (Refer to...) Figure 6 As shown, Figure 6 This document provides an example of an information storage system for a case study library. Specifically, engineering information is categorized into basic information, tunnel boring machine (TBM) selection information, geological conditions, construction data, test information, and construction failures. Each type of information is further subdivided into subcategories. The core technology of a TBM lies not only in the electromechanical industrial design of the equipment itself but also in its suitability for various engineering geological conditions. The success rate of TBM construction largely depends on the selection of the TBM, which in turn determines whether the TBM is suitable for the on-site construction environment. The correctness of the TBM selection directly determines the success or failure of the construction. Therefore, the information is categorized according to the characteristic parameters affecting TBM selection, as shown in the diagram. This includes the TBM's cutterhead system, propulsion system, and muck removal system, as well as the geological conditions in the TBM case study project, such as soil type and consistency index. Construction data is categorized into soil chamber pressure, cutterhead rotation speed, and propulsion speed.

[0031] S02. Similarly, following the above method, extract the feature parameters that affect the selection of shield tunneling in newly constructed shield tunneling projects. The feature parameter categories are consistent with the feature parameters in the previous shield tunneling project information stored in the case library. The consistency of the feature parameter categories between newly constructed shield tunneling projects and those in the case library is beneficial to the accuracy of the retrieval.

[0032] S03. Based on the feature parameters extracted from the newly constructed shield tunneling project, a search is conducted in the case library, and similar previous shield tunneling projects are selected as recommended solutions based on the similarity of the search results.

[0033] S04. Based on the actual construction needs of the new tunnel boring machine (TBM) project, modify and adjust the selected recommended selection schemes, and output the final scheme. Since no two projects are exactly the same, and many factors affect the selection of TBMs, the recommended selection schemes also need to be adjusted according to the actual construction needs. The recommended selection schemes can basically meet the selection requirements.

[0034] S05. Carry out construction according to the final plan. Record all adjustments made during construction based on site requirements and construction needs. After construction is completed, store the project examples in the case library.

[0035] The method provided by this invention can quickly and accurately select shield tunneling machines, and has the characteristics of dynamic knowledge base and incremental learning. It can continuously learn new construction cases, expand the case library, and provide learning cases for shield tunneling machine selection in subsequent new shield tunneling projects.

[0036] Reference Figure 2 As shown, Figure 2 This is a flowchart of the shield tunneling machine selection method based on case-based reasoning of the present invention. In step S03, during the retrieval process, it is determined whether the construction plan for the newly constructed shield tunneling project and similar projects needs adjustment. If adjustment is required, the adjusted plan, based on the needs of the new shield tunneling project, is stored as a new case in the case library. If no adjustment is needed, the plan is adjusted according to the actual construction requirements of the new shield tunneling project and then stored in the case library. While providing convenient and accurate recommended shield tunneling machine selection schemes, it can also dynamically learn new selection schemes, adding new shield tunneling machine selection schemes and providing retrieval cases for the selection of shield tunneling machines for subsequent new projects.

[0037] The specific method for selecting shield tunneling machines (TBMs) in TBM engineering is as follows: In step S03, the search is conducted according to three categories: soil layers, rock layers, and soil-rock composite layers at the construction site. The parameters of the soil layers, rock layers, and soil-rock composite layers at the construction site are used as the search criteria to provide suitable TBM selection schemes.

[0038] Reference Figure 3The soil layer retrieval is divided into three parts: particle size distribution classification, construction indicators, and geotechnical indicators. Soil layers are classified into four categories (I-IV) according to particle size distribution. For example: Category I consists of more than 60% particles smaller than 200 mesh, primarily composed of highly plastic clay and silt particles, with a small amount of silt particles. Engineering characteristics include high water content, low shear strength, and high sensitivity in soft clay, while hard clay exhibits relatively high strength and strong cohesion. The main construction risk is the adhesion of clay to metal, manifesting as mud cake formation and blockage during construction. Category II consists of 30%-60% particles smaller than 200 mesh, consisting of silty clay, silt, and fine sand. Engineering characteristics include low cohesion, poor self-stability, and low strength. The main construction risks are excavation face instability, water and sand inrush, and a lower risk of blockage or tool wear. Category III consists of 12%-30% particles smaller than 200 mesh, with more than 50% particles smaller than 4 mesh. The stratum is mainly composed of coarse-grained sand and gravel, characterized by non-cohesiveness, high permeability, and high hardness. The construction risk lies in the medium to high risk of wear and tear on the tunnel boring machine (TBM) caused by this stratum. Class IV strata consist of particles smaller than 200 mesh (less than 12%) and particles smaller than 4 mesh (15%-50%), primarily composed of gravel, pebbles, and some boulders. It is characterized by a lack of fine particles and a predominance of large particles. Construction risks include cutter wear, difficulty in muck removal, and subsequent settlement. Figure 4 As shown, gradation curves are plotted according to the particle size distribution in the soil layer.

[0039] The construction indicators are the diameter of the shield and the depth of the shield burial, which refers to the depth of the shield below the ground.

[0040] The geotechnical indices are first determined by classifying the strata into different categories based on particle size distribution curves. Then, based on typical adaptability issues for shield tunneling in various strata, the geotechnical indices are analyzed to extract stratum characteristic parameters. These parameters include: consistency index Ic and plasticity index. The consistency index is Where wl represents the liquid limit, wp represents the plastic limit, and w represents the natural water content. Excavation face water head He (m), permeability coefficient K (cm / s), and equivalent quartz content EQC are also included. Where: Vi is the percentage content of minerals, Ri is the ratio of minerals to Rosiwal hardness of quartz, the limit particle size d60 (mm), and the maximum particle size dmax (mm).

[0041] Reference Figure 5The rock strata retrieval is divided into three parts: compressive strength classification, construction indicators, and geotechnical indicators. Compressive strength classification categorizes rock strata into four classes (I-IV) based on the surrounding rock's firmness coefficient. For example: Class I has a uniaxial compressive strength <30MPa and a firmness coefficient <3, indicating extremely soft rock, which does not present any rock-breaking issues during construction. Class II has a uniaxial compressive strength between 30MPa and 60MPa and a firmness coefficient between 3 and 6, indicating soft rock, which also generally does not present rock-breaking issues during construction. Class III has a uniaxial compressive strength between 60MPa and 100MPa and a firmness coefficient between 6 and 10, indicating moderately hard rock, which can be broken relatively easily under normal construction conditions. Class IV has a uniaxial compressive strength between 100MPa and 150MPa and a firmness coefficient between 10 and 15, indicating hard rock, requiring strict control of construction parameters and consideration of deep-hole blasting assistance during construction.

[0042] The construction indicators are the diameter of the shield and the depth of the shield burial, which refers to the depth of the shield below the ground.

[0043] The geotechnical index is first classified into different categories based on compressive strength. Then, the geotechnical index is analyzed based on the typical adaptability problems of shield tunneling in various types of strata, and the stratum characteristic parameters are extracted. Among them, the stratum characteristic parameters are integrity coefficient Kv, karst development Cr, and groundwater conditions represented by the excavation face water head He. Karst development is represented by linear dissolution rate.

[0044] The integrity coefficient is divided into five categories: I to V. Category I is Kv < 0.15, indicating fragmented integrity. Category II is 0.15 ≤ Kv < 0.35, indicating poor integrity. Category III is 0.35 ≤ Kv < 0.55, indicating moderate integrity. Category IV is 0.55 ≤ Kv ≤ 0.75, indicating relatively complete integrity. Category V is Kv > 0.75, indicating structural integrity.

[0045] The development of karst is represented by linear dissolution rate, and is divided into four categories (I-IV) according to the linear dissolution rate. Category I is characterized by a linear dissolution rate of Cr > 15%, indicating strong development. Category II is characterized by a linear dissolution rate of 10% ≤ Cr < 15%, indicating moderate development. Category III is characterized by a linear dissolution rate of 5% ≤ Cr < 10%, indicating weak development. Category IV is characterized by a linear dissolution rate of Cr < 5%, indicating very weak development.

[0046] For soil-rock composite layers, the retrieval of soil-rock composite layers is based on the classification of soil layers and rock layers according to the particle size distribution and compressive strength mentioned above, to determine the type of soil and rock layers in the composite stratum, and then to retrieve the composite layer based on the composite ratio of soil and rock layers in the composite stratum.

[0047] By searching the soil layers, rock layers, and soil-rock composite layers mentioned above, and using the information mentioned in these layers as features, a shield tunneling machine selection scheme can be quickly and accurately recommended. Furthermore, if the selection scheme is modified, the case study is stored in a case library based on the information mentioned above, providing a reference for shield tunneling machine selection in future new shield tunneling projects.

[0048] For shield tunnel diameter retrieval in soil and rock strata, the similarity of shield tunnel diameters is expressed by the following formula: Where a and b represent the shield diameters of newly constructed shield tunneling projects and previous shield tunneling projects in the case library, respectively, and the intervals are... This indicates the range of values ​​for the shield diameter, which depends on the specific circumstances.

[0049] For the search of shield tunneling depth in soil and rock strata, since the shield tunneling depth is an interval value, the similarity between the shield tunneling depth of newly built shield tunneling projects and previous shield tunneling projects in the case library is expressed by the following formula: , where x1 and x2 represent the minimum and maximum values ​​of the x interval, and y1 and y2 represent the minimum and maximum values ​​of the y interval.

[0050] After obtaining the similarity between shield tunneling depth and diameter using the methods described above, the global similarity of construction indicators between the newly constructed shield tunneling project and previous shield tunneling projects in the case library is then determined using the following formula. ,in, To build global similarity of indicators, This represents the local similarity of instance attributes.

[0051] When calculating the similarity of feature parameters, the components of the parameter set differ significantly in magnitude. For example, the consistency index typically ranges from 0 to 1, while the equivalent quartz content ranges from 0 to 100. Therefore, the dispersion of the two variables differs greatly. Mahalanobis distance is used to correct this scale inconsistency. Where S is the covariance matrix, and when the covariance matrix is ​​the identity matrix, the Mahalanobis distance is simplified to the Euclidean distance.

[0052] In summary, this invention retrieves information influencing shield tunneling project selection according to three main categories: soil layers, rock layers, and soil-rock composite layers. These three main categories are further subdivided into multiple subcategories of features. Features influencing shield tunneling project selection from past projects are extracted according to these categories and stored for each project. For subsequent new shield tunneling projects, feature parameters are extracted using the same classification method. By comparing the feature parameters of new and past projects, similar shield tunneling project selection schemes are recommended. After a new shield tunneling project is completed, its information is stored in a case library according to the referenced shield tunneling project selection scheme and adjustments and modifications made during actual construction, categorized by feature parameters, to provide a reference for future new shield tunneling projects.

[0053] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A shield tunneling machine selection method based on case-based reasoning, characterized in that... Includes the following steps: S01. Collect information on past tunnel boring machine (TBM) projects, extract characteristic parameters that affect TBM selection from the past TBM project information, and store them in the case library; S02. Extract characteristic parameters that affect shield selection in newly constructed shield tunneling projects; S03. Based on the feature parameters extracted from the newly constructed shield tunneling project, a search is conducted in the case library, and similar previous shield tunneling projects are selected as recommended solutions based on the similarity of the search results. In step S03, the search is performed according to three categories: soil layer, rock layer, and soil-rock composite layer of the construction site. The soil layer retrieval is divided into three parts: particle size distribution classification, construction indicators, and geotechnical indicators. The soil layers are classified into multiple categories according to particle size distribution. The construction indicators are the shield diameter and the shield burial depth, which refers to the depth of the shield below the ground. The geotechnical indicators are first classified into different categories based on the particle size distribution curve, and then the geotechnical indicators are analyzed based on the typical adaptability problems of shield tunneling in various types of strata to extract stratum characteristic parameters. Among them, the stratum characteristic parameters are consistency index, excavation face water head, permeability coefficient, equivalent quartz content, limiting particle size, and maximum particle size. The strata retrieval is divided into three parts: compressive strength classification, construction indicators, and geotechnical indicators. The compressive strength classification divides the strata into multiple categories according to the surrounding rock firmness coefficient. The construction indicators are the shield diameter and the shield burial depth, where the shield burial depth is the depth of the shield below the ground. The geotechnical indicators are first divided into different categories based on the compressive strength, and then the geotechnical indicators are analyzed based on the typical adaptability problems of shield tunneling in various strata to extract stratum characteristic parameters. Among them, the stratum characteristic parameters are the integrity coefficient, karst development, and groundwater conditions expressed as the excavation face water head. The karst development is expressed as the linear dissolution rate. The similarity of the shield diameters is expressed by the following formula: Where a and b represent the shield diameters of newly constructed shield tunneling projects and previous shield tunneling projects in the case library, respectively, and the intervals are... This indicates the range of values ​​for the shield diameter, which depends on the specific circumstances. The shield tunneling depth is a range value. The similarity between the shield tunneling depth of a newly constructed shield tunneling project and previous shield tunneling projects in the case library is expressed by the following formula: Where x1 and x2 represent the minimum and maximum values ​​of the x interval, and y1 and y2 represent the minimum and maximum values ​​of the y interval; S04. Based on the actual construction needs of the new shield tunneling project, modify and adjust the selected recommended selection schemes, and output the final scheme; S05. Carry out the construction according to the final plan, and store the project example in the case library after the construction is completed.

2. The shield tunneling machine selection method based on case-based reasoning according to claim 1, characterized in that: In step S03, during the retrieval process, it is determined whether the construction plan of the newly built shield tunneling project and similar shield tunneling projects needs to be adjusted. If the construction plan needs to be adjusted, it is adjusted according to the needs of the newly built shield tunneling project and stored as a new case in the case library. If no adjustment is needed, it is adjusted according to the actual construction needs of the newly built shield tunneling project and then stored in the case library.

3. The shield tunneling machine selection method based on case-based reasoning according to claim 1, characterized in that: The retrieval of the soil-rock composite layer is based on the classification of soil and rock layers according to particle size distribution and compressive strength, respectively, to determine the type of soil and rock layers in the composite stratum, and then to retrieve the composite layer according to the composite ratio of soil and rock layers in the composite stratum.

4. The shield tunneling machine selection method based on case-based reasoning according to claim 1, characterized in that: After determining the similarity between the shield tunneling depth and diameter, the global similarity of the new shield tunneling project with respect to construction indicators is then determined using the following formula: ,in, To build global similarity of indicators, This represents the local similarity of instance attributes.

5. The shield tunneling machine selection method based on case-based reasoning according to claim 4, characterized in that: When calculating similarity for feature parameters, the magnitudes of the components in the parameter set differ significantly, resulting in large differences in the dispersion of the two variables. Mahalanobis distance is used to correct this scaling inconsistency. Where S is the covariance matrix, and when the covariance matrix is ​​the identity matrix, the Mahalanobis distance is simplified to the Euclidean distance.

Images