Structural Robustness and Regime-Dependent Uncertainty in a Reduced-Order Physics Model for Vehicle Fuel Consumption Using Monte Carlo Analysis

Accurate prediction of vehicle fuel consumption typically relies either on simplified empirical correlations or on high-fidelity simulations that are computationally expensive. However, the structural robustness of reduced-order physics-based models under parametric uncertainty has not been systematically quantified. In particular, the interaction between model simplifications and uncertainty in vehicle and fuel properties across different operating regimes remains insufficiently investigated. This study presents a reduced-order physics-based framework derived from fundamental force and energy balances to estimate fuel consumption in L/100 km. The model includes aerodynamic drag, rolling resistance, inertial effects, drivetrain efficiency, and fuel lower heating value. Unlike purely empirical formulations, the proposed structure preserves physical interpretability while remaining computationally efficient. Monte Carlo simulations are employed to propagate simultaneous uncertainties in vehicle mass, drag coefficient, rolling resistance, engine efficiency, and fuel energy content. Thousands of randomized realizations are executed to quantify output variability, compute confidence intervals, and evaluate robustness indices. In addition, regime-dependent dominance transitions are analyzed by comparing urban and highway operating conditions. Results show that parameter influence is strongly dependent on speed regime: mass and rolling resistance dominate in low-speed conditions, while aerodynamic parameters become dominant at high speeds. Fuel energy content and efficiency exhibit nearly linear inverse relationships with consumption. The reduced-order structure demonstrates stable variance behavior under realistic uncertainty ranges, supporting its suitability for parametric studies and alternative fuel assessment. The proposed framework contributes a systematic evaluation of structural robustness in simplified physics-based fuel consumption models and provides a scalable methodology for uncertainty-aware automotive performance analysis.