Objective
The future European emission standards (EURO IV and EURO V) for HD diesel engines demand new truck engine technologies such as an exhaust gas recirculation (EGR), variable nozzle turbocharger (VNT) and exhaust gas after treatment systems. EGR and VNT do influence one another via the exhaust flow and exhaust backpressure. Therefore, within transient operation, a high precision control of EGR and VNT is required to avoid smoke and/or NOx emission peaks. The objective of the proposed project is a research program focusing on a new Advanced Truck Engine Control System ATECS including new strategies for engine control and ECU (electronic control unit) calibration as well as an uncoupling control strategy for EGR and VNT. To achieve this, two European truck engine manufacturers, two suppliers, four university institutes and one combustion engine development company, as co-ordinator will co-operate close together. The expected results are advanced truck engine control strategies for reduced engine out-emissions and improved fuel consumptions as well as an improved dynamic engine response and a reduced transient truck engine application effort.
The objective of the ATECS project is a new Advanced Truck Engine Control System to reduce application effort and improve engine control within transient operation. Starting from the base of a state-of-the-art EURO 3 HD Diesel engine concept the potential of combined EGR and VNT-Turbocharger system application is investigated with focus on EURO 4 and 5 exhaust emission requirements. Special effort will be laid on the investigation of a simulation model for improved engine ECU (Electronic Control Unit) calibration under transient conditions including control algorithms for uncoupling control of EGR and VNT-turbocharger. A controller for uncoupled control of EGR and VGT was researched and tested for Heavy Duty Application. One important part was the generation of a time based non-linear model as development platform for this advanced controller. The non-linear model is programmed in Matlab/Simulink and consists on physically based formulas, which means the usage of maps or neural networks was avoided. By this, real time models are available which later can be implemented in the engine control unit and can be parameterised faster than maps calibrated.
The parameterisation of the model was done by taking engine mappings and turbocharger measurements and optimising the parameters with the least square method. The resulting overall model accuracy is within the range +/-5%, which is sufficient for controller design. Important for controller testing was to achieve a realistic dynamic of the system. Therefore not dynamically measurable values, like the exhaust temperature, were determined by process simulation. By this a reasonable transient model tuning was possible. The structure programmed in Simulink was then implemented in the ECUs of a 1L/Cyl. and a 2L/Cyl. engine. For the 1L/Cyl. engine the controller code was implemented in ASCET and for the 2L/Cyl. engine via target link directly implemented in the ECU code. On the test cell it could be shown, that the uncoupling of VGT and EGR is very effectively possible with the model based predictive controller. This was shown for a 1L/Cyl. engine with the control values air mass and boost, and for a 2L/Cyl. engine with the control values EGR mass and boost. The impact of the uncoupled control on emissions was investigated at load steps and within the ETC. It could be shown that under transient test conditions nearly the same NOx and PM emissions can be realised as they are measured under steady state conditions, this means the normally offset between transient and steady state emissions are negligible using the uncoupled control.