The development of control strategies for new Engine Management Systems requires a simulation tool to represent a model of the engine such that these strategies can be tested and verified. Current simulation tools for hardware-in-the loop (HiL) testing are limited to mean-value engine models. Although these lumped parameter type models are sufficient for current production Engine Management Systems, new technologies are emerging that would require sophisticated modeling tools to support the development of more complex systems such as closed-loop combustion control and for more thorough testing of onboard diagnostics. However, such hi-fidelity models have traditionally been distributed parameter typed models (e.g. WAVE), reserved for engine design purposes and therefore inappropriate for real-time testing.
A new piece of software has been developed to bridge the gap between distributed and lumped parameter models. This is physically based, with crank resolved in-cylinder models and accurate capture of wave propagation effects in the air path. As such it provides a more accurate transient response than mean-value models. The software is designed to run in real time, enabling it to be used in HiL systems to prove out the control systems surrounding the engine model. Furthermore, the model can be directly and automatically converted from donor WAVE models allowing the engine to be accurately modelled before being converted to the real-time implementation thereby greatly reducing development timescales.
The new tool chain has been validated with respect to a 4-cylinder turbocharged diesel engine with variable geometry turbine, EGR circuit and intercooler. Work is underway to show that this process provides an improvement to control strategy development time on a project.