Page 6 | University of Utah Partnerships | J. Willard Marriott Digital Library

American Flame Research Committee

Contents | 6 of 10

Page 6

Download PDF | | Reference URL | Gallery View | Parent Record

Update Item Information

Title	Multivariate Study of the Effects of Geometric Design Parameters on Furnace Performance
Creator	Chang, A.; Samuelsen, G. S.
Publisher	Digitized by J. Willard Marriott Library, University of Utah
Date	1997
Spatial Coverage	presented at Chicago, Illinois
Abstract	In low - NOx burner designs, small changes in burner geometry can significantly increase, or reduce, NOx and CO emissions. Of interest is the optimal design for a given configuration. Rather than testing all possible burner configurations, a design of experiments (DoE) can be used to optimize the testing process by systematically varying the parameters. The statistics-based analysis of variance (ANOVA) can then be used to analyze the results to identify effects of parameters on the responses or interactions between parameters. In the present case, a specially designed (for parametric flexibility) practical burner was utilized. The parameters and their ranges include: a) flame holder angle (30°, 90°); b) fuel injection hole diameter (0.1", 0.3"); c) fuel injection hole location (1.25", 4.25" from base of flame holder); and d) swirl angle (30°, 60°). Results show that CO concentrations and combustion efficiency are not affected by the hardware changes. However, NOx emissions are affected by an interaction between the fuel injection hole diameter and the swirl angle. In particular, in the present case, lower NOx emissions are achieved by either a combination of smaller fuel injection hole diameter and larger swirl angle or a combination of larger fuel injection hole diameter and smaller swirl angle.
Type	Text
Format	application/pdf
Language	eng
Rights	This material may be protected by copyright. Permission required for use in any form. For further information please contact the American Flame Research Committee.
Conversion Specifications	Original scanned with Canon EOS-1Ds Mark II, 16.7 megapixel digital camera and saved as 400 ppi uncompressed TIFF, 16 bit depth.
Scanning Technician	Cliodhna Davis
ARK	ark:/87278/s6v98bph
Setname	uu_afrc
ID	13876
Reference URL	https://collections.lib.utah.edu/ark:/87278/s6v98bph

Page Metadata

Title	Page 6
Format	application/pdf
OCR Text	Show D E S I G N - E X P E R T Plot C O (corrected to 302l A : Flan Holder \|an\|lei B : Fuel Inicctnn Diameter C : Fuel Inieciion Location D Swirl Angle Normal % probability 99~ 90"" 10" 70^ 50_ J0^ 20"^ IOH J- '" Normal plot No tarma ara (elected ^<* - x 'A • ^ ^ ^m l -1 J7 1 1 1 •0.71 -0.04 0.63 Effect • • 1 1 29 Figure 2. Normal probability plot for effect of hardware changes on C O level DESIGN-EXPERT Plot Combustion Efficiency, r\| A: Flame Holder (angle) B: Fuel Injection Diameter C: Fuel Injection Location D: Swirl Angle Normal % probability 99- 95- 9Q_ 80_ 70- 50- 30- 20- 10- 5- 1- Normal plot No terms ara aalactad. y/* i i i i 0 0005 -0.0003 -0.0001 0 0002 Effect • 1 0 0004 Figure 3. Normal probability plot for effect of hardware changes on combustion efficiency Figure 4 reveals that the interaction between the fuel injection diameter and swirl angle has an effect on N O x emissions because it deviates from a normal distribution (the straight line). This deviation indicates that the effect of the interaction between the two parameters on N O x emissions is not random. Figure 5 reveals that the effect of this interaction lies outside the I-beam symbols which represent the 9 5 % (a = 0.05) confidence interval for the plotted points. A 100(1-a)% confidence interval for any parameter 0 is a random interval [Ll, L2] such that the probability of the value of 9 occurring within that 6
Setname	uu_afrc
ID	13871
Reference URL	https://collections.lib.utah.edu/ark:/87278/s6v98bph/13871