Measurements are reported of premixed hydrogen-air
turbulent burning velocities, made by the double kernel
method during explosions. Turbulence was created by four
high speed fans driven by electric motors within the
explosion vessel. This arrangement created a central
region of uniform, isotropic turbulence in which all
measurements were made.
The ratio of turbulent to laminar burning velocity
correlates well with both the turbulent Reynolds number of
the reactants and the ratio of laminar burning velocity to
r. m. s. turbulent velocity. The use of hydrogen-air
mixtures has extended the data on premixed turbulent
combustion to regimes with higher values of the last
dimensionless ratio. At high values of the ratio there is
evidence of a wrinkled laminar flame structure, but at
lower values a small scale eddy structure seems to be
dominant.
A two eddy theory of turbulent combustion is
presented. This rests-upon the assumption, supported by a
good deal of experimental evidence, that two scales of eddy
are particularly important. One is associated with the
integral scale of turbulence, the other with the Kolmogorov
microscale. It is assumed that all the material in the
large eddies is used in the formation of the smaller dissipative eddies. It is assumed that laminar flame propagation occurs through the large eddies, whilst two approaches are considered in the case of dissipative eddies. In the first approach, laminar flame propagation across a vortex tube is employed, whilst in the second the concept of reaction time in the vortex tube is used.
It is shown that the rate of burning in small eddies
can be many times greater than that in large eddies.
Theoretical values are obtained for the ratio of turbulent
to laminar burning velocity, in terms of turbulent Reynolds
number and the ratio of laminar burning velocity to r. m. s.
turbulent velocity. These are in fair agreement with
experimental values, but more data are required on the
intermittency and chemical lifetimes of small eddies.
Experiments are reported on the effect of turbulence
upon flammability limits. These are narrowed as
turbulence increases, but counter-action may be taken by
increasing the spark iginition energy and by establishing
the initial flame in a shielded region where the turbulence
is reduced. The relevance of the theory to these results
is discussed.
Finally, the application of these findings in practical combustion chambers is discussed