Search Results

In the last two decades the abilities of neural networks as universal approximation tools of non linear functional relationships as well as identification tools for nonlinear dynamic systems have been recognized and used successfully in many applications areas like modelling, control and diagnosis of technical systems. At the same time an increasing interest in optimal design methods is observed. Design of experiment is used to cope with the growing amount of measurements needed for the calibration of engines due to the rising number of control variables to be considered and the need for more accuracy in the description of engine behaviour to derive the best control strategies. In this paper a strategy for the integration of the concept of D-optimality in the learning process of neural networks is proposed. This leads to an optimal selection of data to be presented to the training procedure of the neural network aiming to a generation of robust neural models using fewer training data.

Stochastic Preignition (SPI) is an abnormal combustion event that occurs in a turbocharged engine and can lead to the loss in fuel economy and engine hardware damage, and in turn result in customer dissatisfaction. It is a significant limiting factor on the use and continued downsizing of turbocharged spark ignited direct injection (SIDI) gasoline engines. Understanding and mitigating all the factors that cause and influence the rate and severity of SPI occurrence are of critical importance to the engine’s continued use and fuel economy improvements for future designs. Previous studies have shown that the heavy molecular weight components of the fuel formulations are one factor that influences the rate of SPI from a turbocharged SIDI gasoline engine. All the previous studies have involved analyzing the fuel’s petroleum hydrocarbon chemistry, but not specifically the additives that are put in the fuel to protect and clean the internal components over the life of the engine.

A main focus of development in modern SI engine technology is variable valve timing, which implies a high potential of improvement regarding fuel consumption and emissions. Variable opening, period and lift of inlet and outlet valves enable numerous possibilities to alter gas exchange and combustion. However, this additional variability generates special demands on the calibration process of specific engine control devices, particularly under cold start and warm-up conditions. This paper presents procedures, based on Hardware-in-the-Loop (HiL) simulation, to support the classical calibration task efficiently. An existing approach is extended, such that a virtual combustion engine is available including additional valve timing variability. Engine models based purely on physical first principles are often not capable of real time execution. However, the definition of initial parameters for the ECU requires a model with both real time capability and sufficient accuracy.

The modern engine design process is characterized by shorter development cycles and a reduced number of prototypes. However, simultaneously exhaust after-treatment and emission testing is becoming increasingly more sophisticated. The introduction of predictive real-time simulation tools that represent the entire powertrain can likely contribute to improving the efficiency of the calibration process. Engine models, which are purely based on physical first principles, are usually not capable of real-time applications, especially if the simulation is focused on cold start and warm-up behavior. However, the initial data definition for the ECU using a Hardware-in-the-Loop (HiL)-Simulator requires a model with both real-time capability and sufficient accuracy. The use of artificial intelligence systems becomes necessary, e.g. neural networks. Methods, structures and the realization of a hybrid real-time model are presented in this paper, which combines physical and neural network models.

Advanced technologies such as direct injection DI, turbocharging and variable valve timing, have lead to a significant evolution of the gasoline engine with positive effects on driving pleasure, fuel consumption and emissions. Today's developments are primarily focused on the implementation of improved full load characteristics for driving performance and fuel consumption reduction with stoichiometric operation, following the downsizing approach in combination with turbocharging and high specific power. The requirements of a relatively small cylinder displacement with high specific power and a wide flexibility of DI injection specifications lead to competing development targets and additionally to a high number of degrees of freedom during optimization. In order to successfully approach an optimum solution, FEV has evolved an advanced development methodology, which is based on the combination of simulation, optical diagnostics and engine thermodynamics testing.