为提高黏性盾构泥渣的资源化利用率，提出一种利用废弃黏性盾构泥渣部分替代膨润土制备渣土改良剂的方案。以膨润土掺量、黏性盾构泥渣掺量及碳酸钠掺入比为影响因素，以漏斗黏度、体积质量、酸碱度和滤失量为响应值，基于响应面法设计试验并构建材料配比与改良剂性能间的预测模型。采用自主研制的承压渗透仪，克服常规试验的尺寸效应，系统验证不同配比的泥渣基改良剂对圆砾、砾砂及中粗砂3种典型粗颗粒地层的抗渗改良效果。结果表明： 1）基于响应面法建立的回归模型与试验结果吻合良好，能够有效预测不同材料配比下泥渣基改良剂的性能； 2）泥渣基改良剂通过引入黏土颗粒来优化土体颗粒级配，当浆液注入率大于25%时，3种粗颗粒土样的稳定渗透系数均降低至工程安全标准10^-5 m/s以下。

The large amount of muck generated during earth pressure balance shield tunneling often causes environmental pollution and resource waste due to improper disposal. Meanwhile, bentonite, as a key conditioning agent, is an expensive nonrenewable resource. To improve the resource utilization of cohesive shield muck, this study proposes a green approach in which waste cohesive shield muck is used to partially replace bentonite for the preparation of soil conditioners. Bentonite content, cohesive shield muck content, and the sodium carbonate ratio are selected as the three influencing factors, while funnel viscosity, density, pH, and fluid loss are selected as response values. Experiments are designed based on the response surface methodology, and predictive models are established to describe the relationship between material proportions and conditioner performance. A self-developed pressurized permeameter is used to overcome the size effect of conventional tests, and the permeability improvement provided by the muck-based conditioner is systematically validated for three typical coarse-grained soils: cobble, gravelly sand, and medium coarse sand. The results show that: 1)the regression models established via the response surface methodology agree well with the experimental data and can effectively predict the performance of muck-based conditioners under different material proportions. 2)By introducing clay particles, the muck-based conditioner optimizes soil particle gradation. When the slurry injection rate exceeds 25%, the stabilized permeability coefficients of all three coarse-grained soil samples are reduced below the engineering safety threshold of 10^－5 m/s.