Taguchi regression analysis and constrained particle swarm optimization for amended unconfined compressive strength (UCS) of expansive subgrade soil

In the present study, constrained particle swarm optimization, a metaheuristic algorithm, was integrated into the Taguchi optimization technique to optimize additives for the amendment of the UCS of an expansive subgrade soil. The additives, which included rice husk ash (RHA), quarry dust (QD), and limestone Portland cement (LPC), were blended with an expansive subgrade soil with mix ratios generated from the Taguchi orthogonal array. Next, the mean UCS values for the various mix ratios were obtained and utilized to develop a regression model that was subsequently integrated into the constrained particle swarm optimization algorithm for the determination of the optimal mix ratio of additives that would yield optimum UCS of the expansive subgrade soil. The result obtained from the developed optimization technique revealed that the combination of additives that resulted in significant improvement in the 28 days UCS of the soil was found at R3Q3L1, which denotes 15% RHA, 20% QD, and 2%LPC. The significant improvement in the 28 days UCS is attributed to the cementation compounds (calcium silicate hydrate and calcium aluminate hydrate) formed during the hydration and pozzolanic reactions experienced between the additives and the expansive subgrade soil. Lastly, microstructural analyses of both the natural and amended soil implemented with scanning electron microscopy and Fourier transform infrared spectroscopy clearly revealed the morphology of the structure and the formation of the cementation compounds that led to the improvement of the UCS of the expansive subgrade soil.