Assessment of a Low-Throughput Predictive Model for Indicated Cycle, Combustion Noise and NOx Calculation in Diesel Engines in Steady-State and Transient Operations

Abstract

A predictive zero-dimensional low-throughput combustion model that was previously developed by the authors has been refined and applied to a EURO V diesel automotive engine. The model is capable of simulating, in real time, the time-histories of the HRR (Heat Release Rate), in-cylinder pressure, in-cylinder temperatures and NOx (nitrogen oxides) concentrations, on the basis of a few quantities estimated by the ECU (Engine Control Unit), such as the injection parameters, the trapped air mass, the intake manifold pressure and temperature. It has been developed for model-based feedforward control purposes in DI (Direct Injection) diesel engines featuring an advanced combustion system or new combustion-mode concepts, such as LTC/PCCI (Low Temperature Combustion/Premixed Charge Compression Ignition) engines. In the present work, the model has been assessed in detail by analyzing a wide set of experimental engine data that were acquired during the engine calibration phase. The experimental data set has been defined according to the DoE (Design of Experiment) methodology currently used for engine calibration purposes, and applied to six ‘key-points’ that are representative of engine working operations during an NEDC (New European Driving Cycle) for a D-class passenger car. Different injection strategies (pilot-main, double pilot-main; pilot-main-after; double pilot-main-after) have been considered for each key point, and all the main engine operating parameters (rail pressure, injected quantities, boost level, intake temperature, EGR rate,…) have been included in the DoE variation list. Therefore, about 1000 steady-state engine operating conditions have been investigated. In addition, several NEDC driving cycles have been realized with the engine installed on a dynamic test rig, and the combustion parameters and emission levels have continuously been measured during the transient operations. The model has been applied to all the investigated conditions. It has shown excellent accuracy in estimating the values of the main combustion parameters, and a good matching between the calculated and predicted NOx concentrations was found, for both steady-state and transient operations.