Data-Driven Estimation of Fatigue Parameters in Concrete: A Minimal-Input Approach Based on Compressive Strength

Abstract

Given the key role of concrete in civil infrastructure, assessing its performance under fatigue loading remains a persistent challenge, largely due to the reliance of fatigue parameters on case-specific experimental data, which are often unavailable in real-world engineering practice. To address this limitation, this study developed an empirical framework for estimating fatigue parameters of concrete based on an integrated literature analysis, eliminating the need for costly and extensive laboratory testing. By employing statistical analysis and machine-learning techniques, the research proposed predictive relationships for Paris' law parameters ( $C$ and $m$), threshold stress intensity factor ( $∆K_{\mathrm{th}}$), and critical stress intensity factor or fracture toughness ( $K_c$) of concrete. In this approach, compressive strength ( $f_c^{\prime }$) serves as the primary predictor to reflect its fundamental influence on fatigue resistance. The proposed relationships enable efficient and practical evaluation of fatigue behavior of concrete, thereby facilitating the assessment and design of concrete infrastructure subjected to fatigue loads.