OCR Text |
Show INTRODUCTION Flue gas recirculation (FGR) is a well-known technique for oxides of nitrogen (NOx) control in industrial burner applications. With this control technique, the recirculated flue gases reduce flame temperatures, which in turn results in decreased thermal NO production rates. Recently, it has been demonstrated in a small-scale boiler that introducing the recirculated flue gas with the fuel results in a much greater reduction in NOx, per unit mass of recirculated gas, when compared to mixing the flue gas with the air [1]. This technique is referred to as fuel injection recirculation (FIR). For example, NOx emissions were reduced from 90 to 30 ppm with 5% FIR, while 23% conventional windbox FGR was required to achieve the same reduction. The qualitative behavior of FIR versus wind box FGR was the same with or without overfrre air. Based on the success of the 2 MMBTU/hr boiler demonstration, Southern California Edison Company also equipped a 45 MW utility boiler with FIR. Field trials of this large-scale unit show results consistent with the small-scale boiler [2]. That FIR is more effective than wind box FGR is quite remarkable. On a simple heatcapacity basis, the reduction in flame temperature resulting from recirculated flue gas should not depend on whether the gas is mixed with the air or fuel, but only on the quantity recirculated. Hopkins et al. [1] speculate that Fenimore prompt NOx is affected. At present, there is no fundamental understanding of how FIR produces the effects demonstrated. The basic question we are addressing is why indeed does FIR have a greater effectiveness, per unit mass recirculated, than ordinary FGR. To separate issues of a chemistry from issues of turbulent mixing, both of which may be significant contributors to the FIR effect in practical applications, we have chosen to study laminar, nonpremixed flame systems. Studying such flames allows as to understand how NOx emissions and detailed flame chemistry and flame structure are impacted by the location of diluent addition, i.e., either to the fuel stream (simulating FIR) or to the air stream (simulating FGR). This report discusses, first, computer modeling studies of counterflow diffusion flames employing detailed chemical kinetics for methane combustion and NOx formation, and, second, experimental studies of laminar, Cf4-air, jet flames. |