A Reduced Order Powertrain Model Concept for Model Based Development Process with a Driver Agent 2012-01-1628

Reduced models are preferred for real time application such as HILS. Driving is interaction between a driver and a vehicle. Cut off frequency criteria for models to be constructed is determined by driver's operation response time to information recognition. Response property is selected based on the frequency. A mean value engine model is taken for base of above mentioned models, because output power and exhaust gas volume are discretized by intake and exhaust valves in each stroke. A torque index is calculated from intake air mass for a gasoline engine and injected fuel for a diesel engine. Particular nonlinearity in torque generation exists in compression and combustion strokes. The nonlinearity can be handled as an amplitude modulation and demodulation like an AM radio system. Based on the AM concept, nonlinearity problem in modeling is solved. And torque profile in a cycle can be generated with the torque index and a basis function of combustion torque. A transmission model represents torque transfer property, i.e. response and gain, including non-linearity. A vehicle system model was constructed based on the proposed concept and combined with a driver agent (model) with learning capability. Adapted driving operation to compensate response delay in powertrain and speed dependent gain was observed.