OCR Text |
Show Neither external flue gas recirculation nor the natural gas fuel staging through the 8 jet nozzles shown in Figure 2 were applied to flame 258. The burner (Fig. 2) was fired with natural gas and combustion air mass flow rates of 163kg/h, respectively 2810kg/h. The gas-air ratio corresponded with an excess air level of 14%. The inlet temperatures of the natural gas and the preheated combustion air were 25°C, respecitvely 331°C. The mean Figure 1: IFRF furnace No. 1 from: Peters and Weber (1995) axial and tangential air mlet conditions were 76m/s and 52m/s with a calculated inlet swirl number S2 of 0.6 (Weber and Dugue, 1992). The radial natural gas inlet velocity at the little injection holes was 152m/s. The flame body was observed to be blue and non-sooty. The availability of appropriate in-flame measurements and the relative simple burner design made flame 258 suited to validate mathematical models. NC - Na*ralG» FC - F I M CM Figure 2: M F B - 1 natural gas burner from: Peters and Weber (1995) Recent modelling of the MFB-1 flame. In a recent study of Peters and Weber (1995) predicted profiles of tune-averaged temperature, species concentration, velocity components and mean local velocity fluctuations are compared against the experimental data. The predictions are obtained by means of a 2-reaction E B U combustion model and the k- £ turbulence model. Focusing on the near burner zone (NBZ), it is shown that the overall agreement between predicted and measured values of these, so-called, "main-flame-properties" is matching the accuracy level of second-order information. Details of geometry, boundary conditions and the validation can be found in the paper by Peters and Weber (1995). The theoretical formulation of the combustion model used by Peters and Weber will be presented and discussed in the section "The Mathematical Model". Accuracy of Predictions of Local Turbulent Reaction Rates in the Near Burner Zone A combustion model for turbulent diffusion flames generally calculates local turbulent reaction rates, which can be used in C F D software packages as source terms for the species mass and the enthalpy transport equation. Unfortunately, turbulent reaction rates are difficult to measure directly. Therefore "main-flame-properties" are commonly used to validate the models by comparison of predicted results with experimental data. In principal it m a y be stated that "The better the agreement, the better the combustion model used". However, while investigating modern combustion designs, engineers are less interested in "where something is measured, but where something is reacting". Hence, the analysis of local reaction rates generated by C F D predictions may help to improve the understanding of burner and/or reactor performance. Assumed that accuracy of predictions is matching or exceeding the level of second-order information, the accuracy of the predicted local reaction rates should be in the same order of information. In other words: Since reaction rates are sources for temperature and species concentration fields and if there is a second-order information agreement between predictions and measurements, this agreement is deemed satisfactory for the predictions of local turbulent reaction rates. Therefore the present study postulates that, if second-order information accuracy is met, CFD-predictions may deliver additional information m terms of local turbulent reaction rates. The Mathematical Model The formulation of transport equations for momentum, continuity, species and enthalpy, including models for physical properties such like radiation properties and turbulent flow models are subject of a number of textbooks and studies (Patankar, 1980; Tennekes and Lumley, 1972; Weber et al., 1993; Kremer, 1993) and will not be repeated here. In this study only the combustion models relevant for the analysis of the natural gas flame predictions will be outlined. In the following equations all physical properties are considered to represent stationary, time-mean values. Combustion Model of the MFB-1 flame prediction. The prediction of the M F B 1 flame number 258 are obtained by considering six gaseous species: hydrocarbons (CxHy), oxygen, carbon monoxide, carbon dioxide, water vapor and nitrogen. Although the E B U model originally assumes a single-step irreversible reaction, it has been successfully applied to the following two-step mechanism (Visser, 1991, Philipp, 1991; Kjaldman, 1993; Breithaupt et al. 1994): 3 |