Gas turbine combustion modeling for a Parametric Emissions Monitoring System

by Honegger, Ueli

Oxides of nitrogen (NOx), carbon monoxide (CO) and other combustion by-products of gas turbines have long been identified as harmful atmospheric pollutants to the environment and humans. Various government agencies place restrictions on emissions and often require some sort of emissions monitoring even for new low emission gas turbines. Predicting actual emissions from operating parameters that affect the formation of pollutants, called parametric emissions monitoring system (PEMS), has potential economic advantages compared to a continuous emissions monitoring system (CEMS). The problem is that a simple applicable PEMS does not exist.

During this study, a gas turbine combustor model applying first engineering principles was developed to predict the emission formation of NOx and CO in a gas turbine. The model is based on a lean-premixed combustor with a main and pilot burner including the function of a bleeding air valve. The model relies on ambient condition and load. The load is expressed as a percentage of the target speed of the gas producer turbine. Air flow and fuel flow for the main and pilot burner are calculated by the model based on the load through a set of measured input data for a specific gas turbine. To find the combustion temperature characteristics, the combustor is divided into several zones. The temperature for each zone is calculated by applying an energy balance. To predict NOx and CO, several correlations explored by various researchers are used and compared against each other, using the calculated temperatures, pressures and equivalence ratios.

A comparison between collected emissions examples from a turbine test cell data spreadsheet and predicted emissions by the developed model under the same conditions show a highly accurate match for NOx emission at any load. Because of the high variation of CO at part load, the model predictions only match the CO data set at full load.