Combustion Characteristics of Lean Burn and Stoichiometric With Exhaust Gas Recirculation Spark-Ignited Natural Gas Engines

Abstract

Natural gas (NG) has been widely used in reciprocating engines for various applications such as automobile, electricity generation, and gas compression. It is in the public interest to burn fuels more efficiently and at lower exhaust emissions. NG is very suitable to serve this purpose due to its clean combustion, small carbon footprint, and, with recent breakthroughs in drilling technologies, increased availability and low cost. NG can be used in lean burn spark-ignited (LBSI) or stoichiometric EGR spark-ignited (SESI) engines. Selection of either LBSI or SESI requires accommodation of requirements such as power output/density, engine efficiency, emissions, knock margin, and cost. The work described in this paper investigated the feasibility of operating an engine originally built as an LBSI under SESI conditions. Analytical tools and workflow developed by Cummins, Inc., are used in this study. The tools require fundamental combustion properties as inputs, including laminar flame speed (LFS), adiabatic flame temperature (AFT), and auto-ignition interval (AI). These parameters provide critical information about combustion duration, engine out NOx, and relative knock propensity. An existing LBSI engine operating at its as released lambda was selected as baseline. The amount of exhaust gas recirculation (EGR) for the SESI configuration was selected so that it would have the same combustion duration as that of the LBSI at its reference lambda. One-dimensional (1D) cycle simulations were conducted under both SESI and LBSI conditions assuming constant output power, compression ratio, volumetric efficiency, heat release centroid, and brake mean effective pressure (BMEP). The 1D cycle simulations provide peak cylinder pressure (PCP) and peak unburned zone temperature (PUZT) under LBSI and SESI conditions. The results show that the SESI configuration has lower PCP but higher PUZT than that of the LBSI for the same output power. Also, for the same combustion duration, SESI has higher AFT than that of LBSI, resulting in higher engine out NOx emissions. The SESI configuration has shorter AI than that of LBSI engine, or smaller relative knock margin. Reduction of output power and emissions aftertreatment in the form of a three-way catalyst (TWC) is required to operate under SESI engine conditions.

On the use of artificial neural networks to model the performance and emissions of a heavy-duty natural gas spark ignition engine

“…2 In addition, reducing the development time will keep the OEM in line with the continuously changing engine emissions legislation and societal pressure. 3 As engine development is a major portion of the engine cost, the use of computational models for IC engine development is ubiquitous today. 4 For example, three-dimensional (3D) computational fluid dynamics (CFD) engine modeling uses a combination of autonomous meshing, physical models, chemistry solvers, and moving geometry to complement experimental investigations and predict in-cylinder behaviors.…”