Utilization of Crumb Rubber and High-Volume Fly Ash in Concrete for Environmental Sustainability: RSM-Based Modeling and Optimization

• Main Authors: Mugineysh Murali, Bashar S. Mohammed, Isyaka Abdulkadir, M. S. Liew, Wesam Salah Alaloul
• Format: Article
• Language: English
• Published: MDPI AG 2021-06-01
• Series: Materials
• Subjects: crumb rubber (CR); high-volume fly ash (HVFA); response surface methodology (RSM); optimization
• Online Access: https://www.mdpi.com/1996-1944/14/12/3322

Waste tire and fly ash (FA) are two waste materials whose disposal and rapid rate of accumulation are among the pressing sources of concern and threat to the environment. Although much research exists on the use of these materials in cementitious composites, very little literature is available on the effectiveness of combining them in high volumes for concrete production. This work aimed to utilize crumb rubber (CR) from waste tires as a partial replacement of fine aggregate at 15%, 22.25%, and 30% by volume, and high-volume fly ash (HVFA) replacement of cement at 50%, 60%, and 70% (by weight of cementitious materials) to produce high-volume fly ash–crumb rubber concrete (HVFA–CRC). Using the central composite design (CCD) option of the response surface methodology (RSM), 13 mixes were produced with different combinations and levels of the CR and FA (the input factors) on which the responses of interest (compressive, flexural, and tensile strengths) were experimentally investigated. Furthermore, the composite influence of CR and HVFA on the workability of the concrete was assessed using the slump test. The results showed a decline in the mechanical properties with increasing replacement levels of the CR and HVFA. However, up to 22.25% and 60% of CR and HVFA replacements, respectively, produced a structural HVFA–CRC with a compressive strength of more than 20 MPa at 28 days. Response predictive models were developed and validated using ANOVA at a 95% confidence level. The models had high R² values ranging from 95.26 to 97.74%. Multi-objective optimization was performed and validated with less than 5% error between the predicted and experimental responses.