Surrogate assisted robust optimization balances mechanical performance and workability in fiber reinforced self compacting concrete
Year: 2026
Type: article
Abstract: Mechanical performance, workability, fiber efficiency, and uncertainty must all be taken into account when designing a mix for fiber-reinforced self-compacting concrete (FRSCC), and adding more fiber does not always translate into a corresponding improvement in performance. In order to assess and optimize FRSCC using Forta, glass, ceramic, binary-hybrid, and triple-hybrid reinforcement systems, this study created a surrogate-assisted robust multi-objective framework. The analysis used 274 experimental observations to construct surrogate models, which were subsequently evaluated across a design space of 274 measured and model-derived mixture–curing combinations with total fiber dosages ranging from 0% to 9% and curing ages between 7 and 180 days. Statistical analysis, response surfaces, Pareto analysis, and bootstrap-based robustness assessment were used to measure compressive strength (f c ), tensile strength (f t ), flexural strength (f f ), slump flow, T 500 , mechanical efficiency index (MEI), and a total performance score. Compressive-to-tensile strength balance, fracture energy, characteristic length, inferred ITZ bonding, pore structure, and fiber dispersion were further explored using mechanically derived proxy indicators. With f c =21.68 MPa, f t =5.68 MPa, f f =9.66 MPa, a total score of 14.48, and MEI of 5.52, Forta demonstrated the best mean mechanical performance. With a total score of 13.53, Forta+Glass was the most effective hybrid system. Forta-D5-T180 was the robust optimum, ranking first in 63.6% of bootstrap resamples and obtaining a cumulative score of around 17.71. Within the surrogate-evaluated design space, Forta-D5-T180 was the highest-ranked model-predicted candidate, ranking first in 63.6% of the bootstrap iterations. However, because the original experiments covered curing ages only up to 90 days, this candidate requires independent experimental validation at 180 days. R 2 = 0.631 and 0.535 were found in the f c -f t and f c -f f correlations, respectively. Mechanically derived indicators were used to explore differences in strength balance and estimated fracture-related behavior among reinforcement systems. These indicators provide supplementary comparisons but do not directly establish interfacial bonding, pore structure, fiber dispersion, crack-bridging mechanisms, or post-cracking performance. Overall, the framework shows that rather than just optimizing fiber dosage or hybridization, reliable FRSCC design necessitates objective-specific optimization.
Source: Scientific Reports
Institution: Islamic Azad University of Kermanshah
Language: English
Subfield: Civil and Structural Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,690
OpenAlex ID: W7221545168