Mixture optimisation for cement-soil mixtures with embedded GFRP tendons

• DOI: 10.1016/j.jmrt.2022.02.076
• Publication Date: 2022-02-26T16:37:07Z
• Authors: Genbao Zhang; Changfu Chen; Junbo Sun; Kefei Li; Fan Xiao; Yufei Wang; Mengcheng Chen; Jizhuo Huang; Xiangyu Wang
• Keywords: Rebar; Root mean square

The glass fiber-reinforced polymer (GFRP) rebar reinforced cemented soil is widely employed to solve the weak foundation problem led by sludge particularly. robustness of this structure highly dependent on interface bond strength between GFRP tendon and soils. However, its application obstructed owing deficient studies influence factors. Therefore, study investigates effects water content (Cw: 50%–90%), cement proportion (Cc: 6%–30%), curing period (Tc: 28–90 days) peak residual strengths (Tp Tt), as well unconfined compression (UCS). Results indicated that mechanical properties were positively responded Tc Cc, while negatively correlated Cw. Besides, Random Forest (RF), one machine learning (ML) models, was developed with hyperparameters tuned firefly algorithm (FA) based experimental dataset. pullout predicted ML model for first time. High correlation coefficients low root-mean-square errors verified accuracy established RF-FA models in study. Subsequently, a coFA-based multi-objective optimisation (MOFA) introduced optimise tri-objectives UCS, Tp (or cost. Pareto fronts successfully acquired optimal mixture designs, which contributes practice. In addition, sensitivity input variables evaluated ranked.