Real Time Modeling of Engine Coolant Temperature

An engine with double cooling circuits operating at two temperature levels has been developed at Volvo Cars. In this thesis the cooling circuit at the lower temperature level is studied and a model estimating the coolant temperature in this circuit has been developed in Simulink/TargetLink. The model is to be implemented in the engine control unit and used for function based diagnosis of the cooling system. The cooling system consists of the following components: a water cooled air cooler which is used to cool the charged air entering the engine, turbochargers in two-stages with cooled bearings systems and a cooled compressor house in one of the turbochargers, an inlet throttle (ETM) cooled for component protection, an SCR injector also cooled for component protection, and a radiator used to cool the coolant. After an investigation of the average heat transfer rate from each component, the ETM was excluded from the model. The SCR injector was also excluded since this component had not yet been installed in the studied engine. The model was formulated with a physical foundation, using energy balances of the system as well as experimentally obtained heat transfer relationships. An overall energy balance was used to calculate the coolant temperature in each discrete time step, based on heat transfer from the modeled components in the system (the water cooled air cooler, turbochargers and radiator). Model evaluation was performed using vehicle data obtained from real time measurements in a four cylinder diesel engine with extra measurement sensors installed in the air system and cooling system. The developed model estimates the temperature with a total mean error of −0.2 C. The 95 th and 5 th percentile for all simulated data is 1.4 C respectively −2.8 C. The model was also shown to be robust against input errors in a sensitivity analysis done for a representative test case