Optimizing the shield tunnel backfilling grouts with supplementary cementitious materials by response surface methodology

Cement -based grout is widely used for backfill grouting in shield tunnels. However, traditional cement -based grout is prone to segregation, layering, and low retention rates, which leads to a not ideal filling effect of the shield tail gap. This paper innovatively uses three industrial solid wastes as supplementary cementitious materials (SCMs), to replace cement in the composition of backfill grouting materials: silica fume (SF); ground granulated blast -furnace slag (GGBS); and fly ash (FA). Taking the content of SF, GGBS and FA as variables, and fluidity, bleeding rate, concretion rate, setting time and compressive strength as response targets, a regression model was established by the Response Surface Method to explore the effects of the different factors and their interactions on the responses. Results show that the SF content has a significant effect on the characteristics of the grout; the GGBS content is positively correlated with the setting time and compressive strength; the content of FA has a significant effect on the hardening characteristics of the grout, being positively correlated with the setting time and negatively correlated with the compressive strength. Additionally, combined with the desirability function, multi -objective optimization was performed and an optimized mix ratio of SF=5.3 wt%, GGBS=9.0 wt%, FA=60.0 wt%, PC=25.7 wt%, was obtained. The response values of the grout for this composition were verified by experiments and the errors are less than 10%. Compared with the traditional cementbased grout, the grout with the optimized composition uses SCMs to reduce the PC content and significantly improve the grout stability and later strength, indicating that the SCMs used in this paper are effective and the RSM has high accuracy and good reliability. Consequently, RSM can be used to analyze, predict and optimize the mixing ratio of backfill grout in shield tunnelling.