**1. Introduction**

In Computational Fluid Dynamics (CFD), many simplifications and assumptions are made to the mathematical models to make them computationally affordable. The rapid development and progress in the field of computer processors and enhancement of computer memory has enabled scientists and engineers to review and revisit some of the assumptions. This has helped to improve and enhance the predictive capabilities of computer modeling. This approach has also been implemented in engine simulation codes and tools. The objective of this study and research is to identify the reactions that govern the chemical kinetics of fuel oxidization and move from multi-step global reactions to a reduced number of elementary reactions that can better model the combustion and engine performance parameters [1].

For the development of future fuels and the optimization of automotive engines, computer modeling and simulation has proved itself as an inseparable tool alongside experimental study. Due to computational limitations, the traditional approach has been to utilize simplified global reactions to simulate and evaluate the combustion and performance parameters in internal combustion engines. Due to increased interest in various advanced engine configurations with various combustion modes and injection strategies, this approach could reduce credibility of the predictions for advanced concepts since it depends on arbitrary adjustment of model parameters.

Oxidation of hydrocarbons is shown by detailed chemical kinetic mechanisms. These detailed mechanisms are very large and comprised of a large number of species and reactions. As the size of the hydrocarbon increases, the length (number of species and reactions) of the mechanism increases along with it. Due to the construction from smaller hydrocarbons (HC) progressing to larger HC and intermediate radicals, detailed mechanisms of large hydrocarbons consist of many species and reactions that are redundant. Those species do not have significant impact on simulating ignition and combustion phenomenon and needlessly raise the computational and memory requirements. These reactions and species can be identified and eliminated from the detailed reaction mechanism without compromising the accuracy and integrity of the detailed reaction mechanism. The detailed and large mechanisms cannot be employed in present solvers because they are time expensive. For example, gasoline and diesel fuels consist of thousands of reactions and hundreds of hydrocarbons [2].

Fuels, such has gasoline and diesel, are composed of hundreds of hydrocarbons. These hydrocarbons include alkanes, alkenes, aromatics and naphthenes. Surrogate fuel mechanisms that contain limited number of hydrocarbons from all the abovementioned hydrocarbon types are important in this regard. There is a necessity to use reduction techniques to produce reduced order mechanisms that can replicate the predictions of detailed mechanisms [3]. The reduction techniques decrease the number of species and reactions.

Redundant reactions were identified by two methods [1–3]:


In a reaction mechanism, there are two subsets of reactions; slow reactions and fast reactions. The reaction rates analysis divides a reaction mechanism into abovementioned two subsets of slow and fast reactions. The sensitivity analysis is performed to divide the reaction mechanism into two subsets of rate limiting and non-rate limiting reactions. When combined, both analyses identify redundant reactions. The redundant reactions identified by this method are non-rate limiting slow reactions. As the redundant reactions are eliminated, the species taking part in those reactions gets automatically eliminated, hence, a reduced mechanism is obtained [1–3]. Sensitivity analysis is discussed further in detail in the next section with detailed references.

*Numerical Simulations and Validation of Engine Performance Parameters Using Chemical… DOI: http://dx.doi.org/10.5772/intechopen.106536*

Using the computational singular perturbation (CSP) technique [4, 5], reduction of iso-octane/n-heptane reaction mechanism by Soyhan et al. [4] and Valorani et al. [5] has been performed that has resulted in reduced and skeletal mechanisms.

Using various computational codes [6–15] and experimental tools, various researchers performed multiple studies to review surrogate fuel mixtures [8], reduced PRF mechanisms [9], with variable intake parameters including an operating range of equivalence ratios, intake pressures and temperatures while considering various engine performance parameters such as heat release rate (HRR) analysis, in-cylinder pressure data [10] and emissions on various engine geometries operating at various operational ranges [11, 12]. Using a mixture of iso-octane, n-heptane and toluene as gasoline surrogate fuel predicts engine performance parameters correctly, especially in HCCI and SI engines. Further addition of di-isobutylene and methylcyclohexane is also recommended. Under stoichiometric and lean conditions, significant number of small particulate formation occurs while large particulate formation shows existence with increasing equivalence ratio. Decrease in the peak in-cylinder pressure can be achieved by mass and temperature reduction. This phenomenon occurs due to heat loss to the chamber walls [13].

A reaction mechanism reduction through sensitivity analysis of a skeletal reaction mechanism for the compression and power strokes by utilizing computational singular perturbation (CSP) method and using the low temperature reaction pathway analysis leads to a reaction mechanism that predicts accurate results for computational studies. Detailed chemistry in conjunction with fluid dynamics enhances the ability of a computational code to correctly predict the engine performance parameters. This is proven in benchmarking the global and quasi-global mechanisms [1] which provided necessary data and confidence in the use of detailed chemistry to correctly predict the engine performance parameters. Also, using 90% iso-octane and 10% n-heptane as surrogate fuel for gasoline helps in correct prediction and best modeling the engine performance parameters. Along with a correct reduced mechanism, mesh independent study is a key to correctly predict and validate the engine performance parameters against the experimental data for a range of equivalence ratios in premixed spark ignition engines.
