288
W. W. WlLLMARTH AND S. S. LU
mean Reynolds stress was produced during the bursting or ejection of low
speed fluid from the wall region.
Grass (197 1) also used the hydrogen bubble flow visualization technique
to study instantaneous p r o b of streurnwise a d normal velocities all
across the boundary layer developed on smooth and rough walk. Grass used
a digital computer to select pairs of instantaneous stnarnwisc, u, and
normal, u, velocity fluctuation profiles under the codtion that the streamwise velocity was a maximum or a minimum at various specifled distances
from the wall. The profile pairs were then wed to compute instantaneous
profiles of the product uu (note that in this paper the notation is that
ii = ti = 0). The results for smooth (or rough walls) showed clearly that
when u was a minimum at any distance from the wall, uu was negative in that
region and therefore contributions to the Reynolds stress occurred. Grass
interpreted this to mean that bursts or eruptions of low momentum fluid
exert a continued influence throughout the boundary layer. O n the other
hand, when Grass selected u and u profile pain (measured on smooth walls)
on the basis that the straamwisc velocity waa a m i m u m at a given height,
he found that contributions to Reynolds stress, uv < 0, occurred only at
points near the wall.
Using concepts obtained from the above flow visualization experiments as
a guide we hove devised a number of experiments using hot wire
anemometers. These expetiments are designad to provide additional quantitative information about the Reynolds stress during bursting which is
difficult to obtain with flow visudlization metbuds. In addition to the results
we have obtained, a number of other investigations of Reynolds stress and
bursting phenomena have been reported in which hot wire or hot film
anemometers were used. Thesc results appear m the papers of Blackwelder
and Kaplan (1971) Wallace et al. (1972), and Rao et a/. (1971). When
appropriate ths above p~pers will be d k d and compared with the
present measurements 8s they arc described in the body of the paper.
The present measurements have becur andyed witb the aid of a digital
computer which was uacd to detect the OQCUIFBI#X near the wall of fluid
eruptions or “bursts ” and of inrush or usweep * events. It har been possible
to measure the convection speed and scak of t b events neat the wall. In
addition, measurements of uu throughout tha boundary layer have been
made in which the outward flow of fluid with low stFGamwiSe momentum is
shown to contribute more to the Reynolds etregg than does the inward fiow
of fluid with high streamwise momentum. A unique method was devised to
identify burst and swaap contributions to Reynolds s t r a throughout the
boundary layer. This mcth’od b used to confina that burst and swmp related
events occur and scale with outer variabtes throughout the boundary layer.
W. W. WlLLMARTH AND S. S. LU
mean Reynolds stress was produced during the bursting or ejection of low
speed fluid from the wall region.
Grass (197 1) also used the hydrogen bubble flow visualization technique
to study instantaneous p r o b of streurnwise a d normal velocities all
across the boundary layer developed on smooth and rough walk. Grass used
a digital computer to select pairs of instantaneous stnarnwisc, u, and
normal, u, velocity fluctuation profiles under the codtion that the streamwise velocity was a maximum or a minimum at various specifled distances
from the wall. The profile pairs were then wed to compute instantaneous
profiles of the product uu (note that in this paper the notation is that
ii = ti = 0). The results for smooth (or rough walls) showed clearly that
when u was a minimum at any distance from the wall, uu was negative in that
region and therefore contributions to the Reynolds stress occurred. Grass
interpreted this to mean that bursts or eruptions of low momentum fluid
exert a continued influence throughout the boundary layer. O n the other
hand, when Grass selected u and u profile pain (measured on smooth walls)
on the basis that the straamwisc velocity waa a m i m u m at a given height,
he found that contributions to Reynolds stress, uv < 0, occurred only at
points near the wall.
Using concepts obtained from the above flow visualization experiments as
a guide we hove devised a number of experiments using hot wire
anemometers. These expetiments are designad to provide additional quantitative information about the Reynolds stress during bursting which is
difficult to obtain with flow visudlization metbuds. In addition to the results
we have obtained, a number of other investigations of Reynolds stress and
bursting phenomena have been reported in which hot wire or hot film
anemometers were used. Thesc results appear m the papers of Blackwelder
and Kaplan (1971) Wallace et al. (1972), and Rao et a/. (1971). When
appropriate ths above p~pers will be d k d and compared with the
present measurements 8s they arc described in the body of the paper.
The present measurements have becur andyed witb the aid of a digital
computer which was uacd to detect the OQCUIFBI#X near the wall of fluid
eruptions or “bursts ” and of inrush or usweep * events. It har been possible
to measure the convection speed and scak of t b events neat the wall. In
addition, measurements of uu throughout tha boundary layer have been
made in which the outward flow of fluid with low stFGamwiSe momentum is
shown to contribute more to the Reynolds etregg than does the inward fiow
of fluid with high streamwise momentum. A unique method was devised to
identify burst and swaap contributions to Reynolds s t r a throughout the
boundary layer. This mcth’od b used to confina that burst and swmp related
events occur and scale with outer variabtes throughout the boundary layer.