Uniaxial compressive strength of concrete inversion using machine learning and computational intelligence approach

The uniaxial compressive strength of concrete is a key indicator of its quality and performance, directly impacting the safety and durability of structural systems. Traditional experimental methods are costly and time-consuming, while machine learning approaches often face challenges such as overfitting and high computational costs when applied to large datasets. To address these issues, this study proposes an optimized Kernel Extreme Learning Machine (KELM) model enhanced by the Hybrid Chaotic Crested Ibis Algorithm (HCCIA). By leveraging chaotic mapping and optimal neighborhood disturbance operations, HCCIA avoids local optima and improves both computational efficiency and prediction precision. Experimental results show that the HCCIA-KELM model achieves superior prediction performance compared to traditional methods, demonstrating high precision and practical applicability for large-scale concrete strength inversion. This novel approach offers significant potential for efficient and scalable solutions in concrete strength assessment.