Optimization of bitumen formulations using mixture design of experiments (MDOE)

摘要

This work addresses the preparation of bitumen blends using asphaltic residue, vacuum residue, and three aromatic extracts (by-products) from the refining process of base oils. A constrained mixture design was adopted to guide bitumen formulation. The measured responses were fitted to Scheffe polynomial models, namely a quadratic model for the penetration value and a linear model for the softening point. Both models were validated using ANOVA. In a second stage the quadratic model was used to design formulations with a precise penetration value. The predicted and measured responses demonstrate the accuracy of both models with average deviations of only 4.67% and 1,53% for penetration and softening point, respectively. In the whole, mixture design has been shown to be a successful and essential tool for bitumen formulation especially if the blends involve a large number of components. (C) 2017 Elsevier Ltd. All rights reserved.