复合地层大直径泥水平衡盾构掘进效率智能分析与控制

吴学兵

摘 要

盾构法施工因其施工高效等优势，已被广泛运用到市政、水利等领域的地下基础设施工程建造。在实际工程施工中，为保证安全性和经济性，需对盾构掘进效率等指标做预测评估，以便及时调整盾构施工参数保障施工效率。然而，由于影响因素复杂，科学合理预测盾构掘进效率还面临着诸多挑战，现阶段主要还是以人工经验评估调整盾构施工参数为主，存在主观性较大、决策效率不高、操作风险高、自动化程度低等问题。本研究旨在综合分析地质参数和盾构施工参数，探究复杂地层条件下盾构机的盾-土参数映射关系，构建复合地层盾构掘进效率智能预测模型，建立盾构施工参数智能决策系统，为盾构机操作人员调整施工参数提供辅助。主要研究内容如下：

（1）基于某隧道项目，针对软弱地层、上软下硬地层、复合岩层等地质条件，引入了“场切深指数”（Field penetration index, FPI）开展盾构施工参数与地质条件映射关系研究。主要包括：构建复合地层盾构施工参数敏感性分析指标体系；通过多项式混沌克里格法（Polynomial chaos kriging, PCK）构建输入与输出参数间的非线性关系；采用全局敏感性分析方法，探讨在全区段和分区段情况下，输入输出参数间的相关性，识别影响FPI指数的关键盾构施工参数，并给出相应的盾构施工参数控制建议。

（2）结合参数敏感性分析结果，构建盾构掘进效率预测模型的输入输出指标体系，搭建掘进效率智能预测模型。该指标体系包括岩土物理力学参数和敏感性分析中排序较高的盾构施工参数。由于不同地层物理力学性质差异显著，本文针对案例工程的软弱地层、上软下硬地层和复合岩层三种地质条件分别建模和预测，并从多角度评价了预测模型效果，研究了该方法在盾构掘进效率预测方面的适用性，为后续根据地质条件调整盾构施工参数建立基础。

（3）构建了盾构施工参数智能决策模型。建立了在不同地质条件下FPI和盾构施工参数的对应关系，根据预测FPI值和实时FPI值，可实现盾构施工参数的实时决策。该模型在案例工程软土层中进行了试用验证。结果表明，通过智能参数决策模型得到的盾构施工建议值与实际工程记录人工控制值相近，且决策效率更高。

本文遵循“数据分析-效率预测-控制决策”的研究思路，从案例工程盾构机历史记录参数的数据分析出发，研究了盾构施工参数与地质条件的映射关系，结合敏感性高的盾构施工参数和地质参数，构建了不同地层条件下的盾构掘进效率智能预测模型。通过构建盾构施工参数智能决策系统，结合实时和预测的FPI值，进行盾构施工参数的智能选择决策，研究结果对于指导盾构掘进控制与关联工序有着重要工程意义。

关键词：复合地层；泥水平衡盾构；盾构掘进效率；敏感性分析；机器学习

Abstract

Due to the advantages of high efficiency and other advantages, shield construction has been widely used in the construction of underground infrastructure projects in municipal, water conservancy and other fields. In actual engineering construction, in order to ensure safety and economy, it is necessary to predict and evaluate indicators such as shield tunneling efficiency, so as to adjust shield construction parameters in time to ensure construction efficiency. However, due to the complex influencing factors, there are still many challenges to scientifically and reasonably predict the tunneling efficiency of shield tunneling machines. At this stage, manual experience evaluation and adjustment of shield tunneling parameters are mainly used, which are subject to high subjectivity, low decision-making efficiency, operational risks high and low degree of automation. The purpose of this research is to comprehensively analyze the geological parameters and shield construction parameters, explore the shield-soil parameter mapping relationship of the shield machine under complex stratum conditions, build an intelligent prediction model for the excavation efficiency of shield tunneling in composite strata, and establish an intelligent decision-making system for shield control parameters. Providing assistance for shield machine operators to adjust control parameters. The main research contents are as follows:

(1) Based on a tunnel project, for the geological conditions such as soft strata, upper soft and lower hard strata, compound rock strata, the "Field penetration index" (FPI) was introduced to carry out the research on the mapping relationship between shield construction parameters and geological conditions . It mainly includes: constructing a sensitivity analysis index system for composite formation shield construction parameters; constructing the nonlinear relationship between input and output parameters through Polynomial Chaos Kriging (PCK); In the case of the whole section and subsection, the correlation between the input and output parameters, identify the key shield construction parameters that affect the FPI, and give the corresponding shield construction parameter control suggestions.

(2) Combined with the results of parameter sensitivity analysis, construct the input and output index system of the shield tunneling efficiency prediction model, and build an intelligent prediction model of tunneling efficiency. The index system includes geotechnical physical and mechanical parameters and shield construction parameters ranked higher in sensitivity analysis. Due to the significant differences in physical and mechanical properties of different strata, this paper modeled and predicted the three geological conditions of the case project: soft strata, upper soft and lower hard strata, and composite rock strata, and evaluated the effect of the prediction model from multiple perspectives. The applicability of shield tunneling efficiency prediction establishes a foundation for subsequent adjustment of shield construction parameters according to geological conditions.

(3) The intelligent decision-making model of shield control parameters is constructed. The corresponding relationship between FPI and shield construction parameters under different geological conditions is established. According to the predicted FPI value and the real-time FPI value, the real-time decision of the shield construction parameters can be realized. The model is tested and verified in a soft soil layer of a case project. The results show that the suggested value of shield construction obtained by the intelligent parameter decision-making model is similar to the manual control value of the actual engineering record, and the decision-making efficiency is higher.

Following the research idea of " data analysis-efficiency prediction-control decision", this paper starts from the data analysis of the historical record parameters of the shield machine in the case project, studies the mapping relationship between the shield construction parameters and the geological conditions, and combines the construction parameters with high sensitivity and geological parameters, an intelligent prediction model of shield tunneling efficiency under different stratum conditions is constructed. By building an intelligent decision-making system for shield construction parameters, combined with real-time and predicted FPI values, the intelligent selection and decision-making of shield construction parameters is carried out. The research results have important engineering significance for guiding shield tunneling control and associated processes.

Key words: Composite formation, Mud-water balance shield, Shield tunneling efficiency, Sensitivity analysis, Machine learning