Robust active control of combustion stability and emissions performance in a fuel-staged natural-gas-fired industrial burner

Abstract

Urban air pollution regulations, climate change concerns, and energy conservation efforts are placing strict constraints in the design and operation of advanced, stationary combustion systems for heat and power generation. To ensure minimal pollutant emissions and maximal efficiency at every instant of operation while preventing reaction blowout, combustion systems need to react and adapt in real time to external changes. This study describes the development, demonstration, and evaluation of an active combustion control system, designed to maximize the performance of natural-gas-fired, industrial boiler burners. A feedback sensor array is developed, consisting of (1) a dynamic combustion stability sensor based on CH* and CO2* UV chemiluminescence and (2) bulk emission sensors for NOx, CO, and O2 using solidstate electrochemical cells, a conventional continuous emissions monitoring system, and a chemiluminescence-based NOx predictor. Next, a dual time-scale controller is designed to atively optimize operating conditions by maximizing a multivariable performance function J using a linear direction set search algorithm. procedures for defining combustion performance, specifying input control variables, and determining optimization parameters are established. The system is successfully demonstrated on a scaled model (120 kW) commercial boiler burner and evaluated for flexibility, repeatability, and robustness by optimizing (1) for different J function, (2) with different emission and stability sensors, (3) over a load cycle between 100 and 50%, and (4) after a core fuel injector misalignment. The controller locates a global performance peak that simultaneously minimizes emissions and maximizes system efficiency, while preventing reaction blowout.