The increased production of natural gas harvested from unconventional sources, such as shale, has led to fluctuations in the species composition of natural gas moving through pipelines. These variations alter the chemical properties of the bulk gas mixture and, consequently, affect the operation of pipeline compressor engines which use the gas as fuel. Among several possible ramifications of these variations is that of unacceptably high engine-out NOx emissions. Therefore, engine controller enhancements which can account for fuel variability are necessary for maintaining emissions compliance. Having the means to predict NOx emissions from a field engine can inform the development of such control schemes.
There are several types of compressor engines; however, this study considers a large bore, lean-burn, two-stroke, integral compressor engine. This class of engine has unique operating conditions which make the formation of engine-out NOx different from typical automotive spark-ignited engines. For this reason, automotive-based methods for predicting NOx emissions are not sufficiently accurate. In this study, an investigation is performed on the possible NO and NO2 formation pathways which could be contributing to exhaust emissions. Additionally, a modeling method is proposed to predict engine-out NOx emissions using a 0-D/1-D model of a Cooper-Bessemer GMWH-10C compressor engine. Predictions are achieved with GRI-Mech3.0, a natural gas combustion mechanism, which allows for simulated formation of NOx species. The implemented technique is tuned using experimental data from a field engine to better predict emissions over a range of engine operating conditions. Tuning the model led to acceptable agreement across operating points varying in both load and trapped equivalence ratio.