Design of experiments and empirical models for up to date burners design for process industries

conference paper

English

At present period the application of complex models based on computational fluid dynamics (CFD) increases. At the same time, the cost of full experiments with high number of factors grows rapidly. In some cases, however, experimental approach cannot be fully substituted by computational modeling, which applies also to tuning of design of new burners. Due to high cost of experiments therefore grows the importance of properly designed experiments and statistical data analysis. Statistical experimental design and analysis is an indispensable tool for experimenters and one of the core topics in a statistics curriculum. The present work deals with the design of experiment from its preparation, through experiment planning, that is followed by preliminary considerations about the model, implementation of the experiment and subsequent evaluation. Furthermore, we point out the problems of the performing of full experiment with high number of factors and outline the possibilities of creating the basis for fractional experiment. The experiment presented in this work is carried out at an up to date experimental facility for testing diffusion gas burners. The investigation focuses on the influence of design parameters on the formation of nitrogen oxides. Priority is however always given to the functionality of the burner (namely flame stability). This paper shows a convenient way how to set up the plan of expensive and time consuming experiments in testing industrial burners. Importance for industrial practice, equipment and consequently processes is demonstrated.

design of experiments, statistical data analysis, empirical model, formation of nitrogen oxides

BĚLOHRADSKÝ, P.; KERMES, V.; STEHLÍK, P.

2007

16. 9. 2007

Copenhagen

978-87-91435-56-0

Book of abstracts: Volume 1

Book of abstracts

1

649

650

2