Detlef Kretschmer

Generalized Formulation for Droplet Size Distribution in a Spray

Based on experimental data in the open literature, a generalized formulation is proposed for droplet size distribution in liquid sprays. The methodology is based on the use of the Pearson system of distribution curves. Experimental data demonstrated that the beta family (type I) distribution in the Pearson system was able to fully characterize the whole data. It was also verified that the gamma family (type III) could be used as a simplified model with a comparable accuracy to a Nukiyama-Tanasawa distribution or better than a Rosin-Rammler.

Prediction of wall heat transfer for a gas turbine combustor

Abstract This paper describes an improved method for predicting wall temperatures in gas turbine combustors in terms of critical parameters such as scale and pressure. It is demonstrated that existing prediction techniques are not necessarily applicable to small combustors and that a good part of it is due to an imprecise knowledge of the hot gas temperatures within the recirculation, primary and secondary zones. From the analysis of new experimental data, obtained recently in the Laval Combustion Laboratory, with wall temperature measurements at three different scales of the same combustor, a new volume function was implemented in the formulation to account for scaling effects.

Prediction of Soot Emissions in Gas-Turbine Combustors

From a large number of experimental data, a formulation was developed for predicting the smoke number (SN) as measured in gas-turbine exhausts according to the well-established Society of Automotive Engineers standard. Three different scales of the same combustor were used with inlet temperature and pressure ranging from 300 to 600 K and from 0.1 to 0.9 MPa, respectively. The formulation is based on the residence time that is calculated from the mass e ow rate, density, and the volumes of the primary and secondary zones of the combustor. The reaction rate has an Arrhenius form with the equivalence ratio to take into consideration the air and fuel e ow rates. All of the required parameters can be evaluated from desired operating conditions. Nineteen different types of fuel were used, varying from a parafe nic mixture to a pure aromatic compound. The fuel is characterized by its calorie c value and the hydrogen mass fraction. With this wide range of fuels burned in the experiments, giving a SN variation from 0 to 100, the accuracy of the prediction (standard deviation of 40% on the relative error to experimental values for each scale and 60% when all scales are combined ) is acceptable for most purposes. Measured SN values already have a 20% error because of the commonly accepted variability of the technique. The formulation should be particularly useful in assessing the efe ciency of new systems for smoke reduction or in calculating the SN from older experimental data where it was not measured.