Road-to-rig-to-desktop: Virtual development using real-time engine modelling and powertrain co-simulation. (September 2019)

• Record Type: Journal Article
• Authors: Andert, Jakob; Xia, Feihong; Klein, Serge; Guse, Daniel; Savelsberg, Rene; Tharmakulasingam, Raul; Thewes, Matthias; Scharf, Johannes
• Other Names: Kamimoto Takeyuki guest-editor.
• Is Part Of: International journal of engine research. Volume 20:Number 7(2019)
• Page Start: 686
• Page End: 695
• Publication Date: 2019-09
• Subjects: Real-time engine modelling -- real driving emission -- virtual development -- Model-in-the-Loop; Engines -- Periodicals; 629.25
• DOI: 10.1177/1468087418767221
• Languages: English
• ISSNs: 1468-0874

Abstract

By front-loading of the conventional vehicle testing to engine test bench or even further forward to offline simulations, it is possible to assess a large variation of powertrain design parameters and testing manoeuvres in the early development stages. This entails a substantial cost reduction compared to physical vehicle testing and hence an optimisation of the modern powertrain development process. This approach is often referred to as road-to-rig-to-desktop. To demonstrate the potential of this road-to-rig-to-desktop methodology as a seamless development process, a crank angle–resolved real-time engine model for a turbocharged gasoline engine was built with the simulation tool GT-POWER ® . The model was validated with measurement data from an engine test bench and integrated into a vehicle co-simulation, which also includes a dual clutch transmission, the chassis, the environment and the automated driver. The most relevant functions of the engine and the transmission control systems were implemented in a Simulink-based software control unit. To verify the engine model in the transient vehicle simulation, two 900-s time windows from a 2-h real driving emission test, representing urban and motorway conditions, are simulated using the developed co-simulation platform. The simulation results are compared with the respective vehicle measurement data. The fuel consumption deviation caused by the combustion engine model is within 5%. The transient system behaviour and the dominant engine operation points could be predicted with a satisfying accuracy.