Economic MPC for transient Diesel engine applications

Patrick Klaffenböck

Research output: ThesisMaster's / Diploma thesis

Abstract

In times of more and more stringent emission regulations and heightened public awareness of the health risks of pollutants like nitrogen oxide and particulate matter as well as the negative influence of carbon dioxide on climate car manufacturers all around the world have an incentive to find improved control strategies to reduce the emissions generated by the combustion process. The industrial standard approach attempts to capture the behaviour of internal combustion engines for a multitude of operating conditions in steady state realizing a feedforward control that is augmented by separate feedback loops to account for mismatches between development and real driving. This approach disregards both the dynamic behaviour and the coupled nature of a turbocharged Diesel engine with exhaust gas recirculation. The focus of this thesis is the development of an economic model predictive control based on a nonlinear model of the engine that considers the system dynamics as well as the cross-influence of the multiple-input-multiple-output structure. In order to solve the optimal control problem on which the model predictive control is based, a control oriented model, i.e. a simplified system description comprising the major influences on the airpath and combustion, is derived by combining relations based on first principles and empirical submodels. Measurements from the test bench in steady state and during transients are used to identify the model parameters and validate the resulting model. Based on this model, an economic model predictive control is formulated that attempts to minimize the economic objective, i.e. a cost consisting of fuel consumption and amount of emissions, by predicting the future system trajectories and control actions necessary to achieve this goal, while complying with the constraints imposed by the model dynamics and limited actuator actions. The performance of the economic model predictive control was then compared with the strategy applied by the engine control unit for the Federal Test Procedure on a provided simulation environment. A reduction of over 30% in the nitrogen oxide emissions during transient manoeuvres was achieved, while hardly increasing the fuel consumption.
Original languageEnglish
Supervisors/Reviewers
  • Del Re, Luigi, Supervisor
  • Waschl, Harald Siegfried, Co-supervisor
Publication statusPublished - Mar 2020

Fields of science

  • 206002 Electro-medical engineering
  • 207109 Pollutant emission
  • 202 Electrical Engineering, Electronics, Information Engineering
  • 202027 Mechatronics
  • 202034 Control engineering
  • 203027 Internal combustion engines
  • 206001 Biomedical engineering

JKU Focus areas

  • Digital Transformation
  • Sustainable Development: Responsible Technologies and Management

Cite this