Computational Intelligence Nonmodel-Based Calibration Approach for Internal Combustion Engines

He MA, Ziyang LI, Mohammad TAYARANI, Guoxiang LU, Hongming XU, Xin YAO

Research output: Journal PublicationsJournal Article (refereed)peer-review

13 Citations (Scopus)

Abstract

Over the past 20 years, with the increase in the complexity of engines, and the combinatorial explosion of engine variables space, the engine calibration process has become more complex, costly, and time consuming. As a result, an efficient and economic approach is desired. For this purpose, many engine calibration methods are under development in original equipment manufacturers and universities. The state-of-the-art model-based steady-state design of experiments (DOE) technique is mature and is used widely. However, it is very difficult to further reduce the measurement time. Additionally, the increasingly high requirements of engine model accuracy and robust testing process with high data quality by high-quality testing facility also constrain the further development of model-based DOE engine calibration. This paper introduces a new computational intelligence approach to calibrate internal combustion engine without the need for an engine model. The strength Pareto evolutionary algorithm 2 (SPEA2) is applied to this automatic engine calibration process. In order to implement the approach on a V6 gasoline direct injection (GDI) engine test bench, a simulink real-time based embedded system was developed and implemented to engine electronic control unit (ECU) through rapid control prototyping (RCP) and external ECU bypass technology. Experimental validations prove that the developed engine calibration approach is capable of automatically finding the optimal engine variable set which can provide the best fuel consumption and particulate matter (PM) emissions, with good accuracy and high efficiency. The introduced engine calibration approach does not rely on either the engine model or the massive test bench experimental data. It has great potential to improve the engine calibration process for industries. Copyright © 2018 by ASME.
Original languageEnglish
Article number41002
JournalJournal of Dynamic Systems, Measurement and Control, Transactions of the ASME
Volume140
Issue number4
Early online date2 Sept 2017
DOIs
Publication statusPublished - 1 Apr 2018
Externally publishedYes

Bibliographical note

This work was conducted in the Future Engines and Fuels Lab at the University of Birmingham. The authors would like to acknowledge the support from Jaguar Land Rover Ltd., Shell Global Solutions UK, and various research assistants. 3-year Engineering and Physical Sciences Research Council (EPSRC) research project (Grant Nos. EP/J00930X/ 1 and EP/J01043X/1).

Keywords

  • automated engine calibration
  • computational intelligence
  • multi-objective evolutionary algorithm
  • rapid control prototyping simulation

Fingerprint

Dive into the research topics of 'Computational Intelligence Nonmodel-Based Calibration Approach for Internal Combustion Engines'. Together they form a unique fingerprint.

Cite this