0
Research Papers: Internal Combustion Engines

Engine Emission Modeling Using a Mixed Physics and Regression Approach

[+] Author and Article Information
Michael Benz

Department of Mechanical and Process Engineering, ETH Zurich, 8092 Zurich, Switzerlandmbenz@alumni.ethz.ch

Christopher H. Onder, Lino Guzzella

Department of Mechanical and Process Engineering, ETH Zurich, 8092 Zurich, Switzerland

J. Eng. Gas Turbines Power 132(4), 042803 (Jan 19, 2010) (11 pages) doi:10.1115/1.3204510 History: Received March 10, 2009; Revised June 04, 2009; Published January 19, 2010; Online January 19, 2010

This paper presents a novel control-oriented model of the raw emissions of diesel engines. An extended quasistationary approach is developed where some engine process variables, such as combustion or cylinder charge characteristics, are used as inputs. These inputs are chosen by a selection algorithm that is based on genetic-programming techniques. Based on the selected inputs, a hybrid symbolic regression algorithm generates the adequate nonlinear structure of the emission model. With this approach, the model identification efforts can be reduced significantly. Although this symbolic regression model requires fewer than eight parameters to be identified, it provides results comparable to those obtained with artificial neural networks. The symbolic regression model is capable of predicting the behavior of the engine in operating points not used for the model parametrization, and it can be adapted easily to other engine classes. Results from experiments under steady-state and transient operating conditions are used to show the accuracy of the presented model. Possible applications of this model are the optimization of the engine system operation strategy and the derivation of virtual sensor designs.

Copyright © 2010 by American Society of Mechanical Engineers
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Figure 2

Comparison between measurement data (black) and results of quasistationary simulations (gray) during a load step at constant engine speed: (a) engine torque and (b) PM

Grahic Jump Location
Figure 3

Comparison between measurement data (black) and results of quasistationary simulations (gray) during a load step at constant engine speed: (a) boost pressure, (b) EGR rate, (c) temperature after intercooler, (d) start of main injection, (e) fuel rail pressure, and (f) injection duration

Grahic Jump Location
Figure 4

Control-oriented NOx emission model formulation

Grahic Jump Location
Figure 5

The wrapper approach as an input selection algorithm

Grahic Jump Location
Figure 6

Bit string manipulation of the input selection algorithm

Grahic Jump Location
Figure 8

Symbolic regression: tree manipulations

Grahic Jump Location
Figure 9

Fitness value for the NOx emissions for the two different data sets: light-duty engine (circle) and heavy-duty engine (cross)

Grahic Jump Location
Figure 10

Regression performance of the NOx emissions for the heavy-duty engine data set for different training data fractions: training data set (black) and validation data set (gray)

Grahic Jump Location
Figure 11

Interactions of the three models: virtual sensor application

Grahic Jump Location
Figure 12

Regression plot of the NOx emissions: (a) LD and (b) HD engine types using in-cylinder pressure sensors

Grahic Jump Location
Figure 13

Regression plot of the PM emissions: (a) LD and (b) HD engines using in-cylinder pressure sensors

Grahic Jump Location
Figure 14

Regression plot of the NOx emission model coupled with the combustion model: (a) LD and (b) HD engine types without using in-cylinder pressure sensors

Grahic Jump Location
Figure 15

Regression plot of the PM emission model coupled with the combustion model: (a) LD and (b) HD engine types without using in-cylinder pressure sensors

Grahic Jump Location
Figure 16

Regression plot of the emission model using the HD engine data set with the LD engine parametrization, NOx: (a) PM and (b) emissions

Grahic Jump Location
Figure 17

Comparison of measurement data and simulation results of the combustion model during a load step from 20% to 80% at 2250 rpm: measurement (black) and combustion model (gray)

Grahic Jump Location
Figure 18

Comparison of simulation results and measurement data of (a) PM and (b) NOx emissions during a load step from 20% to 80% at 2000 rpm: measurement (black), raw emission model (gray), and base map value (dashed)

Grahic Jump Location
Figure 19

Comparison of simulation results and measurement data of (a) PM and (b) NOx emissions during an EGR-rate step from 0% to 10%: measurement (black) and raw emission model (gray)

Grahic Jump Location
Figure 20

Comparison of simulation results and measurement data of (a) PM and (b) NOx emissions during a boost pressure step from 2.1 bars to 1.7 bars: measurement (black) and raw emission model (gray)

Grahic Jump Location
Figure 21

Comparison of simulation results and measurement data of (a) PM and (b) NOx emissions during a fuel rail pressure step from 600 bars to 800 bars: measurement (black) and raw emission model (gray)

Grahic Jump Location
Figure 22

Comparison of simulation results and measurement data of (a) PM and (b) NOx emissions during the start of injection in crank angle (CA) before top dead center (bTDC): measurement (black) and raw emission model (gray)

Grahic Jump Location
Figure 7

Pareto front of the input selection algorithm for the normalized nitrogen oxide emissions without factor f

Grahic Jump Location
Figure 1

Illustration of a standard diesel engine gas path and its components

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In