Additives optimization for expansive soil subgrade modification based on Taguchi grey relational analysis

Abstract

Taguchi approach, which is an offline quality control optimization technique, was combined with grey relational analysis (GRA) in this study to evaluate its effectiveness to optimize additives for expansive soil subgrade improvement. Sawdust ash (SDA), quarry dust (QD) and Portland cement (PC), which were the controlled parameters, were the additives utilized in the study. Soil properties evaluated included California bearing ratio (CBR), unconfined compressive strength (UCS) and differential free swell (DFS). Taguchi mixed level orthogonal array, L18 (6^1 × 3^2), was selected for the experiment design. Optimal results obtained using only the Taguchi optimization method were found at 8% SDA, 20% QD and 8% PC (for the UCS); 20% SDA, 20% QD and 8% PC (for the CBR) and 20% SDA, 20% QD and 3% PC (for the DFS). While the optimal result obtained using a combination of GRA and Taguchi method was found at 20% SDA, 20% QD and 8% PC. Confirmatory tests and scanning electron microscope analysis performed on the expansive soil treated with the optimal level of additives obtained from the combination of Taguchi method and GRA showed the formation of cementation compounds in the soil-additive mixtures and that clearly revealed that Taguchi GRA can be employed to optimize additives for expansive soil improvement. Lastly, statistical analysis performed with the designed experiment and evaluated using analysis of variance (ANOVA) showed the significance of the parameters used in this study.