National Technical University of Athens (NTUA)
Development and evaluation of a semi-empirical multi-zone thermodynamic model for the estimation of nitric oxide emissions and formation history in diesel engines using the measured cylinder pressure trace
Abstract
dc:descriptionThe objective of the present dissertation, is the development and evaluation of a simplified, semi–empirical, multi–zone, thermodynamic model for the prediction of tail–pipe NOx emissions and history inside the combustion chamber, during the closed cycle of DI Diesel engines. The model is implementable on different engine configurations by using the measured cylinder pressure trace.The motivation for this work, emerges from the requirement of the internal combustion engines industry to control pollutant emissions. This requirement is attributed to the strict legislation limits that the governments worldwide have adopted. One of the most affected engine type from this legislation is the DI Diesel engine because it is widely used in the transportation sector (maritime and automotive), in non–road application (heavy–duty industry), in power generation in distant/isolated areas and in the passenger car sector. Specifically, in the last sector, Diesel engines have gained important share against gasoline engines the last years. Moreover, one of the most important controlled pollutant, in these sectors, is NOx. For these reasons, in the present work, the author focuses on the field of NOx control in DI Diesel engines, targeting on the development and validation of a simplified model that can contribute to the deployment and application of NOx control strategies.A variety of simulation models exist which can predict performance and emissions of Diesel engines. These models are divided into three major categories:1. Three dimensional models (i.e. CDF, spray phenomenological models).2. Zero dimensional models (i.e. physical–thermodynamic models).3. Statistical/empirical models.The models of the first category, provide a detailed and strong physical (fundamental) description of the processes taking place inside the combustion chamber, but suffer from high complexity and computational cost. Because of these, this kind of models is inadequate for practical and real–time applications. The models of the second category make use of the basic thermodynamic laws and chemistry following a theoretical approach rather than a detailed spatial description of the combustion chamber processes. Their philosophy is to divide the in–cylinder charge into zones. Thus there exist models that make use one to multiple zones. Single zone models are not appropriate for NOx emission prediction because they provide only the mean in–cylinder temperature which is too low to cause NOx formation. The use of zero dimensional multi–zone models is a solution to this drawback. Nevertheless, the existing models of this kind, make use of semi–empirical or phenomenological correlations for the prediction of fuel distribution inside the zones, air entrainment, etc. which increase model’s complexity, calibration effort and computational cost. The third category comprises models that make use of totally empirical correlations derived from experimental data processing. These models are usually applicable only on the engine type for which they have been developed and validated. Their implementation range, regarding engine operating points, is limited due to the lack of physical base. However, these modes are very simple and provide directness (since they use actual engine operating data, e.g. from ECU and sensors) and very low computational cost.Considering the advantages and disadvantages of the models of the aforementioned categories, it was decided to develop a model that combines the merits of the zero–dimensional multi–zone and empirical models in conjunction with the use of the measured pressure trace on which the most important phenomena taking place inside the combustion chamber are reflected on.Thus, a new model for NOx prediction for DI Diesel engines was developed and evaluated. The proposed model makes use of the engine’s geometry data, measured in–cylinder pressure trace and basic engine operating parameters. Initially, it utilizes the measured pressure trace to estimate the fuel combustion rate via heat release rate analysis. From this, a temporal distribution of the burnt fuel is created in the engine cycle. Then, the resulting elementary fuel amounts burnt consecutively at each time instance and attributed to individual combustion zones, following a multi–zone approach. The amount of the requisite unburnt in–cylinder charge in each zone, is determined from the zone fuel–air equivalence ratio. This parameter is calculated from an empirical correlation developed in the present work and is assumed constant for all zones. This correlation involves engine operating parameters and data derived from the pressure trace processing and heat release rate analysis. The aforementioned parameter is considered constant during the engine cycle but varies with operating conditions, engine settings and engine type. Each zone, after its generation, behaves as a closed thermodynamic system and evolves inside the combustion chamber having its own history of chemical composition, temperature and volume, while maintaining the in–cylinder measured pressure prevailing at each crank angle step. The amount of NOx formed inside each zone is calculated using the extended Zeldovich mechanism. The total amount of NOx emissions at the engine exhaust is calculated from the sum of zone NOx at exhaust valve opening. According to this concept, the model provides the time history (zonal and total) of NOx formation inside the combustion chamber.The proposed model has been validated and evaluated on various types of supercharged DI Diesel engines such as large–scale, 2–stroke engines used for ship propulsion and stationary engines used for power generation, large–scale 4–stroke engines used as auxiliary engines on ships or at power plants and a heavy and a light–duty 4–stroke automotive engine, utilizing a total of 221 test cases. These cases correspond to various engine operating points and engine adjustments/settings. Thus, the ability of the model to predict the absolute NOx values for each test case and capture the effect of the variation of these parameters on NOx formation, has been examined. The parameters considered are the following: Engine load Engine speed. Fuel type. Injection timing. Exhaust gas recirculation (EGR) rate. Fuel injection pressure. Intake manifold pressure.As revealed from the implementation of the model on the aforementioned extensive range of engine types and operating points, it is capable to predict NOx emissions with adequate accuracy for the majority of the test cases examined. The most important is its ability to capture the NOx variation (i.e. trend) as engine parameters vary. The last is accomplished for various engine configurations without significant calibration effort, since only a single multiplier for each engine is used to correct the model’s results. Another significant advantage of the proposed model is the low computational cost. These encourage its use for engine research, development and optimization in the scope of NOx reduction, real–time/model–based and close–loop NOx control, NOx monitoring and verification.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2015
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Savva, Nikolaos
- Σάββα, Νικόλαος
Subjects
dc:subject × 12Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/39466