Effects of heat release and buoyancy on flow structure and entrainment in turbulent nonpremixed flames

https://doi.org/10.1016/S0010-2180(01)00253-XGet rights and content

Abstract

Particle image velocimetry was used to examine the velocity fields in the near- and far-field of the axisymmetric jet in co-flow under nonreacting and reacting conditions. The jets studied include nonreacting jets at Reo = 10,000 and 2,000 and flames at Reo = 10,000 and 37,500. Qualitatively, heat release was seen to impact the instantaneous flow structure over the first third of the flame length by restricting the large-scale movement of the jet. The instantaneous flow structure over the remainder of the flame length was less affected by heat release, a possible consequence of the lower density gradients. Heat release was seen to impact the mean characteristics of the flow by reducing the local Reynolds number by a factor of 10 over the length of the flame and reducing the global entrainment rate by ∼50%. In addition, heat release acted to narrow the jet width up to 20%, reduce the turbulence intensities up to 40%, and increase the centerline-velocities by factors of 2 to 3. The effects of buoyancy resulting from heat release were also examined and tended to counteract the effects of heat release by promoting entrainment and increasing turbulence intensity.

Introduction

The understanding of entrainment in turbulent nonpremixed flames is important in all aspects of combustor design from physical sizing to NOx prediction. A description of the mass entrainment behavior is available for free nonreacting turbulent jets in which the mass flux of the jet increases linearly with distance from the nozzle. Ricou and Spalding [1] measured the entrainment of jets with different densities and formulated the general, well-known expression /o = Ceo)1/2x/do, where Ce = 0.32 is the entrainment coefficient. In flames, however, the entrainment behavior is less straightforward due to heat release and buoyancy.

In the context of this paper, heat release is the local release of energy from exothermic reactions when a flammable mixture of fuel and oxidizer are ignited. The chemical energy of the reactants is released and hot products are formed. The products have a lower density and higher viscosity than their surroundings, which impacts the mixing dynamics. In the current context, buoyancy is the global phenomenon in which buoyancy forces act to assist the upward movement of the heated jet in its cooler surrounding. Similar global effects can be achieved isothermally by issuing a less dense gas into a denser medium. Although the isothermal analogy can be useful in understanding the global effects of buoyancy, the local effects of heat release at the jet periphery do not have a similar analog as illustrated by the velocity and temperature profile development in a reacting jet (or flame) where the temperature profile has a radial maximum that reaches the centerline at the flame tip. Because the density varies inversely with the temperature, the density will have a corresponding radial minimum. This differs from a purely diluting jet (e.g., helium into air or hot combustion products into air), where the density minimum remains on the centerline for the entire jet development.

In the absence of buoyancy, the momentum flux of a turbulent jet (defined as the momentum crossing a plane perpendicular to the jet axis) remains constant with axial distance and equal to the jet-exit momentum flux, Jo = ouo, where o and uo are jet-exit mass flux and velocity, respectively. As a turbulent jet entrains ambient fluid, the mass flux of the jet increases, and the velocity of the jet decreases due to conservation of momentum. It has been demonstrated 2, 3, 4, 5 that flames have higher centerline velocities and concentrations than those of a nonreacting jet at the same axial location with the same exit condition. Because the momentum flux is held constant, it can be inferred that the total mass flux is less in a flame than in a nonreacting jet with the same exit condition. Therefore, the mass flux due to entrainment is less because the initial mass fluxes are the same. This reduction of mass entrainment due to heat release has been measured experimentally by Ricou and Spalding [1] and Becker and Yamazaki [6].

Becker and Yamazaki [6] observed that in the momentum-dominated regime, flames have an entrainment coefficient, Ce = 0.16 compared to the Ce = 0.32 for nonreacting jets. They also defined a dimensionless streamwise coordinate, ξ, which incorporates the influence of buoyancy and observed that buoyancy acts to dramatically increase entrainment. Buoyancy, because of increasing entrainment and its effect on the temperature field, has also been found to strongly influence NOx production in turbulent nonpremixed flames [7].

Of practical and fundamental interest are the effects of heat release and buoyancy upon the flow structure. Chigier and Strokin [4] and Takagi et al.[5] extensively examined the turbulence properties of nonreacting jets and flames in co-flow to investigate the influence of the flame on the flow. However, they primarily compared nonreacting jets to flames with the same exit Reynolds number and did not address the influence of buoyancy. For example, Chigier and Strokin [4] studied methane nonreacting jets and flames at Reo = 6,600, and later Takagi et al.[5] studied hydrogen/nitrogen nonreacting jets and flames at Reo = 4,200 and 11,000. These previous authors used point-wise measurement techniques to examine the mean velocity, concentration, and turbulence intensities.

In the present study, PIV was used to examine the velocity fields of nonreacting jets and flames. The advantage of PIV over point-wise measurements such as LDV is the ability to examine the instantaneous flow structure of the jet and surrounding fluid. Both techniques, however, allow the examination of mean velocities and turbulence intensities. The present study aims to expand upon the work of Ricou and Spalding [1], Becker and Yamazaki [6], and Takagi et al.[5] by investigating the influence of heat release and buoyancy on the entrainment and turbulence behavior of nonpremixed flames in a co-flowing air stream.

Section snippets

Experimental system

A schematic of the experimental setup is shown in Fig. 1. The fuel jet is centered within a vertical, in-draft wind tunnel (30 × 30 cm cross-section). Honeycomb and screens are placed at the tunnel inlet (4:1 area contraction) to provide uniform tunnel velocity (6% spatial variation, 1.3% turbulence intensity). The tunnel test section is 80-cm long, and optical access is available on two orthogonal walls via Pyrex windows. A pilot flame is required to stabilize the jet flame at the nozzle. The

Results

The flow conditions examined in this study are listed in Table 1. The exit velocity, uo, is calculated based on the bulk flow rate, Q, exit diameter, d = 4.57 mm, and jet fluid density, ρo = 0.943 kg/m3 such that uo = Q/πρod2 and the exit Reynolds number is defined as Reouodo, where νo = 1.63 × 10−5 m2/s. The first part of the study examines the velocity of a reacting jet (flame) and a nonreacting jet at the same exit conditions (Reo = 10,000), with a brief comparison to a nonreacting jet

Effects of heat release

We embark now on a fuller discussion of the effects of heat release on turbulent flow structure and entrainment. Previous studies have measured higher centerline velocities and concentrations in turbulent jet flames compared to isothermal jets 2, 3, 4, 5, which implies that jet flames have reduced entrainment rates. The connection between these two trends is discussed below.

As a turbulent jet entrains ambient fluid, the mass flux of the jet increases while the velocity of the jet decreases due

Conclusions

The primary objective of this study was to examine the effects of heat release and buoyancy on turbulent flow structure and entrainment by using PIV. Trends of centerline properties such as velocity, temperature, and turbulence intensities agree well with the point-wise measurements of previous studies. Furthermore, PIV can be used to examine the instantaneous velocity structure of nonreacting and reacting jets, which has lead to insights into the effects of heat release and buoyancy.

The

Acknowledgements

This work is sponsored by the National Science Foundation and the Gas Research Institute, R. V. Serauskas, technical monitor. L. Muñiz acknowledges the support of the Lucent Technologies Cooperative Research Fellowship Program. The authors would like to gratefully acknowledge the meaningful interactions and support of J. E. Broadwell, D. Han, E. F. Hasselbrink, and W. D. Urban.

References (32)

  • P.D Sunavala et al.

    Combust. Flame

    (1957)
  • T Takagi et al.

    Combust. Flame

    (1981)
  • H.A Becker et al.

    Combust. Flame

    (1978)
  • J.E Broadwell et al.

    Combust. Flame

    (1998)
  • H.A Becker et al.

    Combust. Flame

    (1978)
  • D Han et al.

    Combust. Flame

    (2001)
  • N.T Clemens et al.

    Combust. Flame

    (1995)
  • F.P Ricou et al.

    J. Fluid Mech.

    (1961)
  • Hawthorne, W. R., Weddell, D. S., and Hottel, H. C., Third Symposium on Combustion, Flame and Explosions Phenomenon,...
  • N.A Chigier et al.

    Combust. Sci. Technol.

    (1974)
  • L Muñiz et al.

    Proceedings of the Eighth International Symposium on Applications of Laser Techniques to Fluid Mechanics

    (1996)
  • F.R Fristrom et al.

    Flame Structure

    (1965)
  • S Biringen

    J. Fluid Mech.

    (1978)
  • Dahm, W. J. A., and Dibble, R. W., Twenty-Second International Symposium on Combustion, The Combustion Institute,...
  • M.G Mungal et al.

    Combust. Sci. Technol.

    (1991)
  • L Muñiz

    Ph. D. thesis

    (2001)
  • Cited by (0)

    1

    Present address: Ford Research Laboratories, P.O. Box 2053, MD2629, Dearborn, MI 48121.

    View full text