Ultra-low-NOx natural gas burner: Development and performance

In a climate of increasing concern for the environment, restrictions are being placed on allowable emissions of NO x from natural gas furnaces. The Canadian Gas Research Institute initiated development of a new, very-low-NO x burner, work that CAGCT subsequently took over. The result is a burner that produces less than 30 ppm NO x under maximum firing rate and air-preheat conditions. The burner consists of a ring of alternating fuel and air ports with the angle that the air ports make with the burner axis being less than that of the fuel ports. This arrangement realizes FODI (Fuel/Oxidant Direct Injection) strategy by postponing the mixing of oxidant and fuel streams until they have both been significantly diluted by entrained furnace gases that have been cooled by furnace heat load. An extensive set of tests was performed on a CAGCT research furnace, a 2 MW, multi-burner facility, with access for detailed probe measurements, especially in the burner near-field region. The concentrations of NO x , O 2 , CO, CO 2 and unburned hydrocarbons were measured continuously from a matrix of positions in the furnace. Additionally, gas temperatures in the furnace were measured using a micro-thermocouple and gas velocities were determined by a 2D LDV. The furnace was operated under a wide range of firing rates, air preheat levels, heat-sink exposures and equivalence ratios. The details of these results presented here reveal that the burner generally produced less than 20 ppm NO x , more than exceeding expectations. More controlled, lab-scale experiments were conducted, which suggest that the main reason for the very low NO x emissions is a dilution effect reducing combustion temperatures and greatly expanding the reaction zone. In conjunction with the experimental work, a commercial CFD package was used to simulate the furnace transport phenomena in order to determine how well the performance of such complex mixing and combustion field can be predicted numerically. The velocity and temperature fields were predicted reasonably well, however, no useful NO x estimates could be obtained due to a very simple combustion model implemented.