Predictive Analytical Framework for Industrial Steel Pallet Rack Connection Behavior

The development of a reliable analytical model for industrial steel pallet rack connections has long been a goal for researchers, as practitioners still rely primarily on costly and time-consuming experimental testing to assess structural behavior. The present study addresses this gap by developing an analytical framework that enables the prediction of moment–rotation behavior without extensive testing. The framework is calibrated for a specific upright profile within defined parameter ranges and is intended as a configuration-specific predictive tool to reduce reliance on testing within the validated geometric space. Eighteen specimens with varying connection configurations commonly used in practice were tested under monotonic loading in accordance with EN 15512. Major failure modes and the influence of configuration on behavior, particularly in terms of strength and stiffness, were identified. A four-parameter Richard Abbott analytical model was successfully calibrated using experimental data, achieving exceptional predictive accuracy, with R2 > 0.94 across all configurations. Parametric equations expressing model parameters as functions of beam depth and tab count were developed, enabling direct prediction of behavior. A local parametric sensitivity analysis identified initial stiffness and reference moment as the most influential parameters, with average sensitivities of 16.3% and 18.7%. The validated model significantly outperforms previous approaches, offering a reliable design tool that removes the need for configuration-specific testing.