prediction of rice husk ash-based scc compressive strength: data-driven framework

The construction and upkeep of concrete structures have posed significant technical and financial challenges over the past decade. in response, self-compacting concrete (scc) has gained attention due to its superior mechanical performance and reduced environmental footprint. this study investigates the application of gene expression programming (gep) in developing a predictive model for estimating the compressive strength (cs) of self-compacting concrete incorporating rice husk ash (rha). to assess the model's reliability, its predictions were benchmarked against those from two well-established machine learning methods: multiple linear regression (mlr) and artificial neural networks (ann). a total of 651 experimental records related to rha-based scc were gathered from trustworthy references. the model's performance was then quantified using key statistical measures, including the correlation coefficient (r), root mean squared error (rmse), and mean absolute error (mae). the gep model outperformed the ann and mlr approaches, delivering greater accuracy and lower error levels. additionally, the study introduced a gene expression-based formula derived from the gep model for estimating compressive strength at different curing ages, achieving a correlation coefficient of 0.49 and error values ranging from 5 to 9 mpa, which highlights its strong predictive ability. this equation provides a practical tool for preliminary mix design and the quick assessment of scc mixtures. sensitivity analysis revealed that binder content was the most significant parameter influencing cs prediction.