Mode-Selecting Nonlinear Model Predictive Control for Enhanced Thermal Management in Electric Refrigerated Vans

Refrigerated vans play a critical role in cold chain logistics, ensuring the preservation of temperature-sensitive goods. However, traditional thermal management systems, often based on relay-operated on-off control, suffer from significant temperature fluctuations and energy inefficiencies. This study introduces a mode-selecting Non-linear Model Predictive Control (NLMPC) strategy designed to enhance thermal management in electric refrigerated vans by integrating both discrete and continuous control variables. The controller for the van refrigeration system used here includes two binary switches, resulting in four distinct modes of operation. The proposed controller evaluates optimal costs across the four operating modes, enabling coordinated decisions that improve energy efficiency and temperature regulation. Simulation results over a standard van drive cycle demonstrate that the mode-selecting NLMPC reduces temperature constraint violations by 44.84% compared to a traditional relay controller and by 29.07% compared to an NLMPC with relay control while consuming marginally less energy. Additionally, the mode-selecting NLMPC maintains the reference temperature within the cooling chamber for longer durations, highlighting its superior performance in thermal management. Despite its higher computational cost, the modeselecting NLMPC offers a promising solution for optimizing energy efficiency and temperature control in ERVs.