Régulation par recherche d'extremums des systèmes à compression de vapeur alimentés par batterie pour les véhicules commerciaux

This paper investigates the real-time energy optimization of battery powered vapor compression systems (VCS) for commercial vehicles. Battery powered VCS are critical for maintaining passenger comfort in engine-off situations, such as for long-haul truck drivers who sleep inside their vehicle overnight. This paper proposes the use of extremum seeking control (ESC), a class of real-time, model-free optimization algorithms, to determine the combination of system inputs that minimizes VCS power consumption while meeting desired cooling requirements. We examine the implementation of three different ESC algorithms established in literature: perturbation-ESC (P-ESC), least squares-ESC (LS-ESC), and recursive least squares-ESC (RLS-ESC) on an experimental setup consisting of a battery powered VCS cooling a vehicle cabin. Experimental results demonstrate significant increases (up to 33%) in energy efficiency and battery life through algorithm use, with RLS-ESC being the most effective of the three.