An In-Cycle Hardware in the Loop (HiL) Simulator for Future Engine Control Strategy Development

Conventional engine HiL simulators use Mean Value (MV) modelling techniques to represent the plant and provide closed loop feedback parameters to the ECU. Once configured, these models require parameterising with the engine specific data. This data can be obtained from two primary sources:

• Test bed data - running a range of steady state/dynamic speed/load points. Engine required.
• WAVE* simulation model - Physical sizes of engine required for simulation. No Engine required.

The accuracy of the MV model once fully parameterised is in the region of 70 - 80% assuming accurate test data. Another limitation of current techniques is that differences between individual cylinders, for example due to intake system geometrical arrangement, are ignored.

With the increase of sensors that can monitor in-cycle events or cylinder conditions, for example cylinder pressure sensors, and associated control techniques, the HiL models must be capable of supplying more detailed crank synchronous, in-cycle flow information to the ECU. This is unachievable with mean value models. Therefore an in-cycle engine model is required to verify and validate cylinder pressure based control strategies.

This paper describes the Automotive Real-Time Engine Model based on In-cycle Simulation (ARTEMIS). A description is given of the gasoline model and its validation and verification to real engine test bed data. Once the model is validated then the HiL system will be used in closed loop operation to allow control algorithm development of cylinder pressure based control strategies.