Development of PEMS Models for Predicting NOx Emissions from Large Bore Natural Gas Engines 2001-01-1914
In this work two different Parametric Emissions Monitoring System (PEMS) models are developed, an algebraic, semi-empirical model and a neural network model. The semi-empirical model is based on general relationships between oxides of nitrogen (NOx) emissions and engine parameters. The neural network model utilizes a similar set of input parameters, but relies on the neural network code to determine the relationships between input parameters and measured NOx emissions. Two sets of data are used for model development. The first set is composed of typical engine parametric variations and is used to train the models. The second set is used to test the models and is composed of changes to engine operation associated with engine degradation, termed Operations and Maintenance (O&M) issues.