Search Results

Nowadays requirements towards a reduction in fuel consumption and pollutant emissions of Internal Combustion Engines (ICE) keep on pushing manufacturers to improve engines performance through the enhancement of existing subsystems (e.g.: electronic fuel injection, air systems) and the introduction of specific devices (e.g.: exhaust gas recirculation systems, SCR, …). Modern systems require a combined design and application of different after-treatment devices. Mathematical models are useful tools to investigate the complexity of different system layouts, to design and to validate (HIL/SIL testing) control strategies for the after-treatment management. This study presents a mean value model of an exhaust system with SCR; it has been coupled with a common rail diesel engine combustion black box model (Neural Network based). So, dedicated models for exhaust pipes, oxidation catalyst, diesel particulate filter and selective catalytic converter are developed.

In the field of passenger car engines, recent research advances have proven the effectiveness of downsized, turbocharged and direct injection concepts, applied to gasoline combustion systems, to reduce the overall fuel consumption while respecting particularly stringent exhaust emissions limits. Knock and turbocharger control are two of the most critical factors that influence the achievement of maximum efficiency and satisfactory drivability, for this new generation of engines. The sound emitted from an engine encloses many information related to its operating condition. In particular, the turbocharger whistle and the knock clink are unmistakable sounds. This paper presents the development of real-time control functions, based on direct measurement of the engine acoustic emission, captured by an innovative and low cost acoustic sensor, implemented on a platform suitable for on-board application.

The new driving cycles require a greater focus on a wider engine operative area and especially in transient conditions where a proper air path control is a challenging task for emission and drivability. In order to achieve this goal, turbocharger speed measurement can give several benefits during boost pressure transient and for over-speed prevention, allowing the adoption of a smaller turbocharger, that can further reduce turbo-lag, also enabling engine down-speeding. So far, the use of turbocharger speed sensor was considered expensive and rarely affordable in passenger car applications, while it is used on high performance engines with the aim of maximizing engine power and torque, mainly in steady state, eroding the safe-margin for turbocharger reliability. Thanks to the availability of a new cost effective turbocharger speed technology, based on acoustic sensing, turbocharger speed measurement has become affordably also for passengers car application.

The trend of CO2 emission limits and the fuel saving due to the oil price increase are important drivers for engines development. The involved technologies have the aim to improve the global engine efficiency, improving combustion and minimizing energy losses. The engine auxiliary devices electrification (i.e. cooling pump or lubricating pump) is a way to reduce not useful energy consumption, because it becomes possible to control them depending on engine operating point. This kind of management can be applied to the electric low pressure fuel pump. Usually the fuel delivery is performed at the maximum flow rate and a pressure regulator discharges the exceeding fuel amount inside the rail (i.e. gasoline engine) or upstream of the high pressure pump (i.e. common rail diesel engine). At part load, especially in diesel application, the electric fuel pump flow is higher than needed for engine power generation.

The paper focuses on the experimental identification and validation of different neural networks for virtual sensing of NOx emissions in combustion compression ignition engines (CI). A comparison of several neural network architectures (NN, TDNN and RNN) has been carried out in order to evaluate precision and generalization in dynamic prediction of NOx formation. Furthermore the model complexity (number and types of inputs, neuron and layer number, etc.) has been considered to allow a future ECU implementation and on line training. Suited training procedures and experimental tests are proposed to improve the models. Several measurements of NOx emissions have been performed through different devices applied to the outlet of a EURO 5 Common Rail diesel engine with EGR. The accuracy of the developed models is assessed by comparing simulated and experimental trajectories for a wide range of operating conditions.