Optimal Transient Real-Time Engine-Generator Control in the Series-Hybrid Vehicle
Paper in proceedings, 2019

We study the dynamic engine-generator optimal control problem with a goal of minimizing fuel consumption while delivering a requested average electrical power. By using an infinite-horizon formulation and explicitly minimizing fuel consumption, we avoid issues inherent with penalty-based and finite-horizon problems. The solution to the optimal control problem, found using dynamic programming and the successive approximation method, can be expressed as instantaneous non-linear state-feedback. This allows for trivial real-time control, typically requiring 10–20 CPU instructions per control period, a few bytes of RAM, and 5–20 KiB of nonvolatile memory. Simulation results for a passenger vehicle indicate a fuel consumption improvement in the region of 5–7% during the transient phase when compared with the class of controllers found in the industry. Bench-tests, where the optimal controller is executed in native hardware, show an improvement of 3.7%, primarily limited by unmodeled dynamics. Our specific choice of problem formulation, a guaranteed globally optimal solution, and trivial real-time control resolve many of the limitations with the current state of optimal engine-generator controllers.

Optimal control

Transients (Dynamics)

Simulation results

Control equipment

Dynamic programming

Engines

Generators

State feedback

Vehicles

Dynamics (Mechanics)

Fuel consumption

Real-time control

Approximation

Electricity (Physics)

Hardware

Author

Jonathan Lock

Chalmers, Electrical Engineering, Signalbehandling och medicinsk teknik, Signal Processing

Rickard Arvidsson

Chalmers, Electrical Engineering, Signalbehandling och medicinsk teknik, Signal Processing

Tomas McKelvey

Chalmers, Electrical Engineering, Signalbehandling och medicinsk teknik, Signal Processing

Dynamic Systems and Control Conference

Vol. 2 V002T12A001

ASME 2019 Dynamic Systems and Control Conference, DSCC 2019
Park City, Utah, USA,

Driving Forces

Sustainable development

Areas of Advance

Transport

Energy

Subject Categories

Computational Mathematics

Vehicle Engineering

Control Engineering

DOI

10.1115/DSCC2019-8964

More information

Latest update

1/9/2020 3