| OCR Text |
Show hardware development costs. Data obtained from small-scale burner tests must, however, be properly "scaled up" to account for the differences in size and thermal loadings in order to yield the performance at full industrial scales. Current methods for scale-up of burner performance are typically based on preserving one, or at most a few, physical burner parameters such as fuel or air velocities, velocity ratios, or geometrical relationships. However none of these classical methods can at present reliably insure similarity in performance. General theoretical treatments on combustion system scaling can be found, for example, in Spalding [1] and Beer [2]. Several scaling methods have been reviewed by Salvi & Payne [3], Lawn et. al [4], and more recently by Smart [5]. These involve considerations of fluid flow, chemical kinetics, and heat transfer in the vicini ty of the burner and in the furnace. In general, the number of scaling parameters that must be matched to achieve perfect similarity between small-scale tests and full-scale performance are prohibitive. However, imperfect but nevertheless adequate similarity may be achievable, since not all of the flow, chemistry, and heat transfer processes involved in practical burners and furnaces are equally important. This is termed "partial modeling" by Spalding, and is commonly used in a wide range of other engineering fields, such as wind tunnel testing of aircraft performance, where accurate scale-up of results from small-scale tests must routinely be done. Development of comparably reliable scaling methods for gas burners and furnaces requires data over a wide range of sizes and thermal inputs, from small-scale laboratory burners to full industrial scales, for which carefully selected aspects of the burner and furnace configuration are fixed. To date, few experiments have been performed on industrial burners with the specific objective of examining performance at different scales. The results of Fricker et. al [6] and Salvi & Payne [3] are notable in this regard, where industrial natural gas flames over a range of thermal inputs were analyzed in the context of flame scaling. The latter study was part of a larger collaboration between the Central Electricity Generating Board of the United Kingdom and IFRF; see Lawn et. al [4]. That scaling study focused primarily on oil flames in the thermal input range 0.5 M\tV to 33 :NI\tV; the natural gas flames studied were limited to the 0.5 to 2.3 MW thermal input range. An effort termed the SCALING 400 study has been underway during the past five years to specifically assess existing scaling methods and develop improved methods for scale-up of combustion performance and emissions of industrial gas burners. The testing portion of this study was completed two years ago, and has provided the first comprehensi ve scaling data for gas burners over the thermal range from 30 k W to 12 ~l\tV, representing a factor of 400 in thermal scale. This was a cooperative effort involving the Gas Research Institute, The University of Michigan, The International Flame Research Foundation, The John Zink Co., and Sandia National Laboratories. Five scaled versions of a generic, swirl-stabilized, gas burner (see Fig. 1 and Table 1) designed by IFRF using the constant velocity scaling principle were tested in separate furnaces under conditions of near-uniform geometric, aerodynamic, and thermal similarity with uniform measurement protocols and operating conditions. The attention to uniformity among tests allowed meaningful comparisons of data from all five scales. 2 |
