Abstract
The public and industrial transport impose a heavy burden on the environment. Because of this, there is increasing focus on alternative fuels such as natural gas, LPG and hydrogen. Their emission characteristics are considerably better than those of the fossil fuels gasoline and diesel.
The research at the Laboratory of Transporttechnology, Ghent University, is focussed on the use of these gaseous alternative fuels in spark-ignited engines. Extensive experimental research has been done on the combustion of natural gas, LPG, hydrogen and mixtures of hydrogen and natural gas.
However, these experiments demand considerable amounts of time and money, especially for hydrogen. Thus, the attention is also directed towards simulation. A simulation programme for the total cycle of a four stroke SI engine is written. A first part of this work looked at appropriate algorithms for the calculation of the gasdynamics (1-D) in the inlet- and exhaust pipes and the filling and emptying process of the cylinders. In this stage, a quasi-dimensional twozone thermodynamic model, based on the work by Tabaczynski et al and Brehob, is implemented for the simulation of the thermodynamic cycle. Results are presented for propane and methane, and are compared with measurements in a single cylinder CFR engine.
To simulate the combustion processes in a SI engine, one can follow different paths. The simplest option is a zerodimensional model. On the other side of the spectrum of complexity, one has the multi-dimensional models. These consider the flow processes in the cylinder, the chemical kinetics and their interactions. The computation of these turbulent flows and their interaction with the combustion processes demand considerable computing power and CPU time, thus losing some of the advantages of computer simulation. A compromise can be taken with a quasi-dimensional model. This is a model that takes some geometrical parameters into consideration and uses phenomenological models for the description of the turbulence and its interaction with the combustion process. Thus, a reasonable accuracy can be combined with fast computation. It is to mention that the existing models for the simulation of the combustion process have been designed and validated for gasoline. SI engines working with alternative fuels have different characteristics and thus, the simulations with the existing models give less accurate results.
In a second stage, several quasi-dimensional combustion models will be compared for the simulation of the combustion of hydrogen. Special attention is given to the choice of the laminar flame speed formula because for hydrogen there exists uncertainty about the proposed formulas in the literature. Initial simulation results are given for variations in airfuel ratio, ignition timing and compression ratio, and compared with measurements. Key elements for a hydrogen combustion simulation code are discussed. The most challenging requirement for an accurate simulation of the combustion of hydrogen is the large variation in air-fuel ratio used in hydrogen engines (throttleless operation).