A reduced-order mechanics framework for uncertainty-aware operational assessment of mechanically steered heavy-duty intralogistics chassis

This study reformulates the operational assessment of a mechanically steered heavy-duty intralogistics chassis as an uncertainty-aware reduced-order mechanics problem. Rather than replacing high-fidelity finite-element analysis (FEA), the framework provides a transparent, computationally inexpensive screening layer for identifying operating combinations under which simplified structural assumptions remain credible and conditions requiring higher-fidelity review. An extended four-support analytical model propagates vertical dynamic amplification, longitudinal and transverse load eccentricity, steering-induced lateral demand, unilateral floor unevenness, wheel/support compliance, and equivalent chassis torsional compliance. The carrier is represented by an average roll mode and a differential front-to-rear twist mode coupled through an equivalent torsional stiffness, enabling support-model discrepancy E bc and frame-rigidity discrepancy E frame to be quantified separately. An 8192-point Latin hypercube data set is evaluated directly from the analytical equations, and global sensitivity is assessed using a scrambled Sobol/Jansen protocol with 4096 base samples and eight uncertain inputs. Across the explored domain, the robustness demand index Ω has a median of 2.060 and a 95th percentile of 2.808. Longitudinal and transverse load eccentricities dominate Ω, with total-order Sobol indices of 0.620 and 0.328, followed by vertical dynamic amplification at 0.137. E bc is governed mainly by floor disturbance and support stiffness, with indices of 0.709 and 0.288, whereas E frame is controlled by floor disturbance, support stiffness, and normalized frame torsional stiffness, with indices of 0.528, 0.407, and 0.232. Steering angle and friction remain weak contributors to Ω but dominate the steering-specific demand S *. The formulation therefore separates support compliance from chassis torsional compliance and provides an uncertainty-aware screening bridge between nominal engineering calculations and targeted future high-fidelity verification. The operating-envelope logic supports model-governance and escalation decisions rather than direct prediction of local stress, fatigue life, or certified structural limits before physical calibration.