Semiempirical predictions and correlations of CO emissions from utility combustion turbines

Abstract

Semiempirical models for gas-turbine combustor emissions distinguish between the dominant subprocesses of combustion related to pollutant emissions. They assign characteristic times, which are based on combustor geometry, fuel characteristics, and operating conditions, to these subprocesses. Linear ratios of these characteristic times, or of sums of the times weighted with empirical constants, are proportional to the emissions indices of oxides of nitrogen and of carbon monoxide (CO). The capability of a semiempirical model to predict and correlate CO emissions from two heavy-duty, dual-fuel (natural gas and fuel oil) diffusion flame combustors without inert injection is assessed. The model is termed the characteristic time model. Since the characteristic time model had not been used to correlate or predict CO emissions from stationary, heavy-duty combustors prior to this study, application of the models to the heavy-duty emissions data was not expected to, and did not yield acceptable predictions or correlations. Thus, modification of the characteristic time model was performed. A new algorithm determines the CO oxidation quench position in the combustor and provides a reasonable correlation of the CO data.