Optimal control of switched /hybrid systems with applications to the control of hybrid electric vehicles
Abstract
Established at the confluence of control, communication, and computer science, hybrid systems, where discrete (decision making) events interact with continuous or discrete-time processes, are present in many real world applications. For example, in the automotive area, hybrid electric vehicles are equipped with, besides an internal combustion engine, an electric motor-generator drive, a storage battery and a supervisory controller. The dynamic evolution, and therefore the performance, of this hybrid system significantly depends on the discrete decisions of the supervisor such as: the instants of changing the mode of operation of the electric motor-generator, the instants of coupling (decoupling) the internal combustion engine to (from) the drive shaft, etc. The hybrid system considered in this research consists of a family of dynamical systems and a rule that selects, at each time instant, a member of this family (mode of operation). This research studies a hybrid optimal control problem: determine, if they exist, a sequence of modes of operation and control inputs for each mode that minimize a given performance measure subject to state and input constraints. For example, the optimal control problem for a hybrid electric vehicle is to switch modes for minimizing fuel consumption and pollution subject to constraints on the state of charge of the battery, generated power, and drivability. The proposed approach is to reformulate the hybrid optimal control problem for a larger family of systems constructed through a continuous parameterization of the original hybrid system. The set of trajectories of the original system being dense in the set of trajectories of the more general system validates the approach. The reformulation allows the application of the classical results of optimal control theory which are not applicable to the original problem. This research shows that the optimal solution of the more general problem solves the hybrid control problem of interest, except in a possible small number of cases when suboptimal trajectories can be constructed by appropriate mode switches. The technique developed in this research is used at the supervisory level of a hierarchical control strategy for a hybrid electric vehicle. At this level, the optimal control problem uses a hybrid power flow model capturing the two modes of operation of the electric motor-generator; the results are reference power inputs for each subsystem of the vehicle.
Degree
Ph.D.
Advisors
DeCarlo, Purdue University.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.