Adaptive Cabin Climate Control for Battery Electric Vehicles Using TD3 Reinforcement Learning

Sammanfattning

The transition from combustion engine vehicles to Battery Electric Vehicles (BEVs) hasincreased the importance of efficient thermal management in trucks. Conventional cabinclimate control methods prioritize transparency and safety but lack adaptability, limitingpotential energy savings. This thesis addresses this challenge using the Twin-DelayedDeep Deterministic Policy Gradient (TD3) combined with a novel surrogate model, theTwin Fourier Neural Operator (Twin FNO), designed to capture cabin thermodynamicsthrough partially enforced physics and dual output heads for improved accuracy. TheTwin FNO predicts average cabin temperature with a mean absolute error of 0.55 °Cwhile being 180 times faster than numerical solvers. Integrated into the TD3 framework,the agent effectively balances energy efficiency and thermal comfort. It maintainsthe setpoint temperature when energy is abundant and deliberately creates an offsetwhen energy is limited, achieving up to 40% reduction in Heating Ventilation and AirConditioning (HVAC) energy consumption with a 6 °C temperature deviation. Theseresults highlight the potential of reinforcement learning and surrogate modeling to enableenergy-adaptive thermal control strategies in BEVs while raising questions on acceptablethermal comfort trade-offs.