Abstract
Most of the available pulse-and-glide (PnG) studies focus on internal combustion engine driven and hybrid-electric vehicles. PnG has been recently extended to electric vehicles (EVs), but without detailed consideration of the powertrain losses during the glide phase. Moreover, control allocation (CA) implementations for EVs focus on the optimal torque allocation among multiple motors to reduce energy consumption, while meeting vehicle dynamics constraints. However, PnG has not been used in conjunction with CA so far, since the available optimal CA formulations target the minimization of instantaneous power consumption for an assigned total wheel torque demand, and are thus incompatible with PnG. This paper reformulates the CA problem such that PnG becomes a viable optimization solution. The formulation also includes powertrain power losses during glide. The result is a linear programming (LP) optimization problem, which can be efficiently solved. An alternative convex hulls graphical method is also discussed, generating identical solutions to the LP. From the graphical approach, the conclusion is that, for PnG to be beneficial, the battery power consumption (i.e., DC link side of the inverter) vs. torque curve must present non-convexities. When an EV has multiple identical motors, an infinite number of energy-optimal solutions exist. Rules are developed to select the best among the possible solutions, with heuristic consideration of drivability and passenger comfort. Simulation results, evaluating multiple PnG options for a sport utility EV with onboard and in-wheel powertrains, show up to 20% and 8% energy consumption reductions, respectively, at speeds below 50 km/h, compared to non-PnG operation.