High performance continuously variable transmission control through robust-control-relevant model validation

T.A.E. Oomen, S.H. Meulen, van der

Research output: Contribution to journalArticleAcademicpeer-review

1 Citation (Scopus)
5 Downloads (Pure)

Abstract

Optimal operation of continuously variable transmissions (CVTs) is essential to meet tightening emission and fuel consumption requirements. This is achieved by accurately tracking a prescribed transmission ratio reference and simultaneously optimizing the internal efficiency of the CVT. To reduce the power losses in a CVT, the absolute pressure levels are lowered, which increases the sensitivity to torque disturbances and increases the importance of disturbance feedforwards. This requires a high performance feedback controller for the hydraulic actuation system in a CVT. The aim of this paper is to develop a multivariable feedback controller for the hydraulic actuation system that is robust with respect to the varying system dynamics that are induced by the varying operating conditions, including transmission ratio changes. Hereto, new connections between system identification and robust control are exploited to achieve high performance. As a result, the varying system dynamics are directly evaluated in terms of closed-loop performance objectives. Subsequent robust control design reveals an increase of the control performance of almost a factor two in terms of the criterion value. This leads to improved simulated and measured closed-loop step responses, including a decrease in settling time from 0.4¿s to 0.2¿s. Finally, the designed robust controller is successfully validated in a standardized driving cycle experiment.
Original languageEnglish
Pages (from-to)061018-
Number of pages14
JournalJournal of Dynamic Systems, Measurement and Control : Transactions of the ASME
Volume135
Issue number6
DOIs
Publication statusPublished - 2013

Fingerprint Dive into the research topics of 'High performance continuously variable transmission control through robust-control-relevant model validation'. Together they form a unique fingerprint.

Cite this