A Comparative Analysis of Three Optimisation Approaches to Free Swell Characterisation of Particulate Coconut Shell Reinforcement Composite Material

Abstract

Abstract. There is paucity of data relating to the free swell properties of composite reinforcement materials in outdoor applications despite the widely-debated need for more investigations in this area. In this paper, a comparative performance analysis of three optimisation models -Taguchi method, simplex algorithm and Taguchi-simplex method -is pursued. We used experimental data of 0.600 mm particulate coconut shell with the theories behind L27 (3 11 ) orthogonal arrays, to achieve our aim for the Taguchi method. The simplex algorithm required modelling the free swell problem as the objective function subject to parametric constraints defined by the factors and their respective levels. For the Taguchi-simplex method, the S/N ratios and the optimal parametric setting of the Taguchi method were used to model the objective function and the new constraints that the simplex algorithm optimised. Parameters include initial and final volumes of coconut shell particulates (CSPs), mass of water, volumes of CSPs and water as well as mass of CSPs. The optimised initial and final volumes were 10 and 34 cm 3 (Taguchi), 20 and 34 cm 3 (simplex) and 16.6 and 24.82 cm 3 (Taguchi-simplex). The optimal masses of water were 141.54, 146.81 and 89.12 g while the volumes of CSPs and water were 152, 152 and 106.4 cm 3 , and the mass of CSPs were 12.07, 12.07 and 6.28 g for Taguchi, simplex and Taguchi-simplex methods, respectively. Confirmatory test validated the results. In conclusion, the optimum parametric setting of the Taguchi-simplex method was the best.