An Integrated Physical–Mechanistic Model for Predicting Fuel Consumption on High-Category Roads: Incorporating Pavement Friction and Infrastructure Constraints

Road transport fuel consumption is strongly affected by pavement conditions and road infrastructure, whereas conventional normative approaches primarily consider vehicle characteristics and operating conditions while neglecting pavement friction and infrastructure-induced driving cycles. This study develops and validates an integrated physical–mechanistic model for predicting actual fuel consumption on high-category roads by combining vehicle energy-balance equations with pavement diagnostic parameters and infrastructure constraints. Unlike existing mechanistic or empirical approaches, the proposed framework explicitly incorporates pavement friction as a quantitative indicator of both skid resistance and road-related energy losses, enabling the simultaneous assessment of traffic safety and fuel efficiency within a unified engineering model. The methodology was applied to the A-350 Almaty–Taldykorgan highway using telemetry from 150 independent vehicle-route observations, 840 complete corridor transits, approximately 4.8 million CAN-GPS records, and 340 pavement friction measurements. The results indicate that actual fuel consumption exceeded normative values by an average of 9.2%, while a reduction in pavement friction below the regulatory threshold was associated with a substantial increase in fuel consumption, further amplified by pedestrian crossings and urban road sections. Model calibration and independent validation demonstrated that the proposed model achieved strong predictive performance (R2 = 0.94; MAE = 1.8%) under the observed operating conditions. This study extends existing fuel consumption modeling by reinterpreting pavement friction as a dual-purpose engineering indicator for both safety assessment and energy-efficiency diagnostics. The proposed methodology provides a scientific basis for corridor-level pavement management and infrastructure-related fuel-efficiency assessment and provides a conceptual foundation for future integration into intelligent transportation systems, subject to broader multi-corridor and seasonal validation.