Back to Projects

Shell Eco-Marathon Vehicle Simulation

Physics-based MATLAB & Simulink vehicle model for ultra-low energy prototype vehicles

MATLAB Simulink Stateflow Vehicle Dynamics Energy Modeling GPS Track Integration Systems Engineering

This project develops a physics-based MATLAB and Simulink vehicle model used to analyze energy consumption and optimize drivetrain and driver control strategies for a Shell Eco-Marathon prototype vehicle. Key vehicle parameters including mass, aerodynamic drag, rolling resistance, gear ratio, and control thresholds were iteratively tuned to study efficiency tradeoffs and realistic operating behavior. The model evolved through multiple versions, progressing from a simple force-balance model into a Stateflow-controlled burn-coast strategy validated using experimental dyno data and scored using the Shell competition km/L metric.

Model Development Timeline

Version 1 — Baseline Physics Model
Simple force-balance vehicle model with constant throttle and basic energy calculations.
Version 2 — Modeling Platform
Structured MATLAB parameter system, improved powertrain model, automated analysis scripts, and design sweeps.
Versions 3–5 — Driver Strategy Development
Closed-loop throttle control, dyno-based torque maps, and PID-based speed regulation development.
Version 6 — Burn-Coast Driver Strategy
Realistic pulse-and-coast throttle strategy with torque-map-driven powertrain modeling and validated burn-coast behavior. PID controller replaced with threshold-based on/off switching.
Version 8 — Accuracy Overhaul & Sweep Optimization
Corrected wheel radius, RPM bounds, and fuel accounting. Stateflow hysteresis replaces single-threshold driver strategy. Shaft-power fuel model and four-parameter km/L sweep with design sensitivity ranking completed.
Version 9 — Real Track Integration & Race Simulation (Current)
IMS Infield Road Course GPS + elevation profile (767 breakpoints, 5 m resolution). Zone-aware three-mode driver strategy (burn-coast / cruise-hold / corner-hold). Grade force in vehicle dynamics. Lap-based race simulation with Article 226 pass/fail. Five-parameter sweep including BSFC sensitivity. Baseline result: 621.3 km/L, 2016.2 s, attempt valid.

Select a Version

Click any version below to expand its full technical writeup.

← All Projects Home Resume GitHub