Operability of Dry-Low-NOx (DLN) combustors in heavy-duty gas turbines is dependent in part on satisfying boundaries or constraints associated with lean blow out (LBO), NOx, and combustion dynamics. To satisfy each of these boundaries or constraints, the determination of combustor airflow is critical. Since combustor airflow is difficult to measure directly, a surrogate or an estimator of combustor airflow that can be implemented in the control system is required. Presented in this paper is a methodology for determining the combustor airflow surrogate based on measurable parameters. The approach relies on developing a nonlinear equation of combustor airflow based on cycle simulations over a wide range of ambient (temperature, pressure and humidity) and parameter (e.g., inlet guide vane and load) variations. To understand the robustness of the combustor airflow estimate, sensitivity analysis with regard to sensor uncertainty, machine degradation, and machine-to-machine variation is presented. In each of these cases, the uncertainty/parameter variation is investigated via Monte Carlo simulations to quantify performance in terms of mean shift (i.e, bias) and standard deviation. It is shown that the resulting surrogate, which is comprised of compressor discharge pressure, compressor discharge temperature and specific humidity, is capable of addressing the variations listed above and is capable of representing a compressor map at different IGV settings via a simple relationship.

This content is only available via PDF.
You do not currently have access to this content.