Refine Your Search

Search Results

Viewing 1 to 2 of 2
Technical Paper

Model-Based Fault Diagnosis of Spark-Ignition Direct-Injection Engine Using Nonlinear Estimations

2005-04-11
2005-01-0071
In this paper, the detection and isolation of actuator faults (both measured and commanded) occurring in the engine breathing and the fueling systems of a spark-ignition direct-injection (SIDI) engine are described. The breathing system in an SIDI engine usually consists of a fresh air induction path via an electronically controlled throttle (ECT) and an exhaust gas recirculation (EGR) path via an EGR valve. They are dynamically coupled through the intake manifold to form a gas mixture, which eventually enters the engine cylinders for a subsequent combustion process. Meanwhile, the fueling system is equipped with a high-pressure common-rail injection for a precise control of the fuel quantity directly injected into the engine cylinders. Since the coupled system is highly nonlinear in nature, the fault diagnosis will be performed by generating residuals based on multiple nonlinear observers.
Technical Paper

A Methodology for Virtual Engine Mapping Test of CIDI Engine with Arbitrary Fuel Injection Schedule for Control Purpose

2005-04-11
2005-01-0230
With the introduction of common rail fuel injection system enabling multiple injections per stroke and stringent pollutant emission standards, the optimization and calibration of modern compression ignition direct injection (CIDI) engines become more complex. Thus, a simple and efficient tool for CIDI combustion simulation with arbitrary fuel injection profile is required today. A crank-angle resolved combustion model was developed and validated by using a fuel injection rig and an engine dynamometer for the parameterization. With the calibrated models, accurate prediction of the in-cylinder pressure and NOx and also a virtual dynamometer mapping are possible. The results from these virtual mappings can be used to calibrate the black-box combustion models in control-oriented, dynamic Mean Value Models (MVM).
X