Incorporating Drivability Metrics into Optimal Energy Management Strategies for Hybrid Vehicles Part 2: Real-World Robustness and Constraint Implementation

Hybrid vehicle fuel economy and drive quality are coupled through the “Energy Management” controller that regulates power flow among the various energy sources and sinks. Most analytical studies have evaluated closed-loop performance on government test cycles, and there are few results that compare optimal control algorithms to the controllers employed in industry. This second of a two-part paper studies controllers designed using Shortest Path Stochastic Dynamic Programming (SPSDP), a stochastic optimal control design method which can respect constraints on drivetrain activity while minimizing fuel consumption. Part 1 described the problem formulation, models, and simulation results on government test cycles for a prototype vehicle. In Part 2, controllers are evaluated for robustness through simulation on large numbers of real-world drive cycles and compared to a baseline industrial controller developed by Ford. On real-world driving data, the SPSDP-based controllers yield 10% better fuel economy than the baseline controller, for the same engine and gear activity. SPSDP controllers are further evaluated for robustness to the drive cycle statistics used in their design. Simplified drivability metrics introduced in Part 1 are validated. Looking ahead, production implementations of the SPSDP method will likely require finer control over drivetrain behavior. Several options for achieving this are discussed, along with their relative benefits for performance versus computational tractability.