Molding Process Design for Asphalt Mixture Based on Response Surface Methodology

Abstract

The values of key parameters, asphalt aggregate ratio, mixing/compacting temperature, striking/compacting times, and fiber content are usually determined according to bulk properties or empirical values when molding asphalt mixture, which is inefficient and subjective. In this study, a modified response surface methodology (RSM) method was introduced to comprehensively optimize design of asphalt mixture molding process with the optimal value of pavement performance as a response variable. The single-factor tests were first conducted to find the specific ranges of key factors in the response surface models. Then the response surface models between effect factors and three different pavement performance parameters were established based on the Box-Behnken design (BBD) theory. Lastly, the analytic hierarchy process (AHP) was applied to determine weights distribution of the effect factors, and the optimal values of effect factors and pavement performances were obtained. Results indicate that fitting errors of this modified RSM method is lower than 1.2%, and that the average prediction error is 0.87%. This modified model was proven to possess favorable fitting accuracy, predictive capability, and stability. Meanwhile, the method can ensure volumetric parameters are in compliance with regulatory requirements under the premise of improving the overall asphalt pavement performance.