Tim Deppen

  • Curriculum Vitae
  • tdeppen@illinois.edu
  • Postdoc
  • PhD., University of Illinois at Urbana-Champiagn, 2013
  • Research Interest: energy management, model predictive control, switched systems, optimal control

 

 

Current Research

My current research is applying energy optimization techniques to complex multi-domain systems in which timescale separation of system dynamics is being utilized to enable a hierarchical control design.  Within this control structure, separate optimization problems are being solved at each tier of system resolution.  For example, at the highest tier (system level) the objective function may seek to satisfy performance objectives and minimize overall energy use while using preview of future demand and/or disturbances.  Then at the subsystem tier, each subsystem (electrical, thermal, hydraulic, etc.) will solve an optimization problem to meet the demands of the system level control while minimizing their own energy consumption by coordinating the actions of multiple actuators (generators, distributors, and storage components).  Finally, at the lowest tier, individual component controllers are utilized to enforce reference tracking for each actuator.  This research project is a team effort and in my current role, I am coordinating and advising other students in the lab.

Past Research

Sponsored by the Center for Compact and Efficient Fluid Power

Figure 1: AEVPS at the University of Illinois at Urbana-Champaign

This research focused on developing a methodology for optimizing energy use in mobile power systems which include energy storage.  Mobile power systems represent a significant portion of global energy consumption and present a number of unique challenges.  The energy management challenge in the context of mobile systems is a constrained optimization problem in which one may have complete or limited knowledge of the duty cycle and the control complexity is constrained by the need for fast update rates with limited processing power.  To achieve this goal, a hybrid systems perspective is employed in which each allowable configuration of the power system is a mode of operation and separate convex optimization problems are defined for each operating mode.  A supervisor is then designed to regulate switching between each mode and model predictive control is employed to solve the optimization problems online.  This energy management strategy has been implemented experimentally on an augmented earth moving vehicle powertrain simulator (AEVPS), which is a hydrostatic transmission with a gas charged accumulator for energy storage.  My PhD thesis can be found here.

Figure 2: Schematic of the AEVPS

My masters research focused on modeling and control of the hydraulic hybrid passenger vehicle.  The hydraulic hybrid powertrain uses variable displacement pump/motors to regulate the transfer of energy to/from a high pressure accumulator, which is used to store energy.  This stored energy can be used to assist the engine or completely supply the operator’s power demand.  In addition, energy which is normally dissipated by mechanical brakes can be captured in the accumulator when decelerating.  The model of the powertrain was developed using the Simscape and Simhydraulics toolboxes of Matlab®.  A rule based energy management strategy was developed which regulated the use of the engine and stored energy. A copy of my MS thesis can be found here.

Selected Publications