Experimental and Statistical Evaluation of Mechanical Properties of Green Cement Concretes – Taguchi Integrated Supervised Learning Approach
Abstract
Globally, rapid infrastructure development and environmental challenges associated with the higher carbon footprints of ordinary Portland cement (OPC) based concretes have increased the usage of green cement-based concrete (GCC) to reduce energy consumption and provide a sustainable option. Even though GCC is a superior alternative to OPC, only a few publications have addressed optimizing process parameters in GCC manufacturing to optimize mechanical properties. The Taguchi method is well-known as one of the most effective methods for optimizing predictors to get the desired level of response. Additionally, in the modern era, data-driven supervised machine learning approaches have been used extensively to develop mathematical models to establish relationships between the variables. As a result, the Taguchi method was used in this study to obtain the best mix design targeting a compressive strength of greater than 40 MPa. Numerous design combinations have been tested, and a process for selecting the most effective combination has been established. The analysis aided in comprehending the individual contributions of the major components to the mechanism of strength gain. The observations confirmed the Taguchi method's ability to predict the design mix proportions of the GCC and the ability of machine learning to relate the variables mathematically.