REYNOLDS STRESS STRUCTURE IN TURBULENT BOUNDARY LAYER 297
sweep events is represented by the dashed line in Fig. 6. The sweep convection speed Ua was found to bc nearly the same as the burst convection
speed. Thus, U d W = U d U Y 0.425 and Ucr/U, 1 U,#, '5 0.8.
Rough estimates of the spced of convection of the burst events were also
made at various larger distances from the wall, Thc bunt convection speed
Urn, was found to increase with the distance from the wall, however we have
not bacn able to obtain accurate tneasumnents of convection sped at
greater distances from the wan. It ap2wan from o h mcasument (Willmarth and Wooldridgt, 1962) that wall pressure disturbances, for example,
are convected at speeds ranging h m 0.58 e V,/V, < 0.83. The lower convection speeds are associated with small scak cddics near the wall. The
present measurements therefore s u e t that the srnalf scale burst pattern
emanates from the wall region, travels outward, a d grows larger as it is
carried downstream. As it enlarges it also is sheared and distorted because
the convection velocity is higher farther from the wall. The evolving and
enlarging burst pattern soon loses coherence with the detection criterion
near the wall (u, = -uk with negative slope) so that the sampled values of
( u v 2 ) can no longer reveal the burst structure. The bursts still exist,
however, as will become apparent below.
Before the detection criterion fails, the peak in the (uu)/iE value which
represents the region of disturbance cawed by the burst that is coherent with
the sampling criteria increases from a size of y/S* = 0.506 at x/S* = 0 to a
size of y/6* = 0.912 at x/6* = 1.686 or more a$ one travels further downstream. There is still some contribution to even at a station of x/S* = 2.53
downstream of the u, detector wire. Tbe spanwise extent of the region of
disturbance is confined to a llpzrow swept back rbgioa with an included
angle of approximately 20" centered upon the free stream direction.
3.4. Large Individual Contributions to Reynolds Stress near the Wall
In another attempt to improve the detection of bursts (and swaeps) two
hot wires mounted side by side near the wall at y+ 3 15 were used to
measure u,, and uW2. The distance between the wire centers was z + % 20
and the wire length was 1 ' 3 20. Note that Kim et uI. (1968) report that the
typical spacing between streaks cawcd by the low and high velocity in the
sublayer is of the order of z+ cy 100. A wire spacing of z+ 1 20 should be
small enough to allow the detection of the center of a streaky region. To find
the central region of a burst, two or more simultaneous criteria are required.
After a number of trials of different detection schema the following criteria
were successful. A computer prqpam was written that searched the u,, and
uw,, data for those few events in which both u , , , aad ha reached the level
-%, and -uL2 at approximatsly the =me t h e with nqative slope,
sweep events is represented by the dashed line in Fig. 6. The sweep convection speed Ua was found to bc nearly the same as the burst convection
speed. Thus, U d W = U d U Y 0.425 and Ucr/U, 1 U,#, '5 0.8.
Rough estimates of the spced of convection of the burst events were also
made at various larger distances from the wall, Thc bunt convection speed
Urn, was found to increase with the distance from the wall, however we have
not bacn able to obtain accurate tneasumnents of convection sped at
greater distances from the wan. It ap2wan from o h mcasument (Willmarth and Wooldridgt, 1962) that wall pressure disturbances, for example,
are convected at speeds ranging h m 0.58 e V,/V, < 0.83. The lower convection speeds are associated with small scak cddics near the wall. The
present measurements therefore s u e t that the srnalf scale burst pattern
emanates from the wall region, travels outward, a d grows larger as it is
carried downstream. As it enlarges it also is sheared and distorted because
the convection velocity is higher farther from the wall. The evolving and
enlarging burst pattern soon loses coherence with the detection criterion
near the wall (u, = -uk with negative slope) so that the sampled values of
( u v 2 ) can no longer reveal the burst structure. The bursts still exist,
however, as will become apparent below.
Before the detection criterion fails, the peak in the (uu)/iE value which
represents the region of disturbance cawed by the burst that is coherent with
the sampling criteria increases from a size of y/S* = 0.506 at x/S* = 0 to a
size of y/6* = 0.912 at x/6* = 1.686 or more a$ one travels further downstream. There is still some contribution to even at a station of x/S* = 2.53
downstream of the u, detector wire. Tbe spanwise extent of the region of
disturbance is confined to a llpzrow swept back rbgioa with an included
angle of approximately 20" centered upon the free stream direction.
3.4. Large Individual Contributions to Reynolds Stress near the Wall
In another attempt to improve the detection of bursts (and swaeps) two
hot wires mounted side by side near the wall at y+ 3 15 were used to
measure u,, and uW2. The distance between the wire centers was z + % 20
and the wire length was 1 ' 3 20. Note that Kim et uI. (1968) report that the
typical spacing between streaks cawcd by the low and high velocity in the
sublayer is of the order of z+ cy 100. A wire spacing of z+ 1 20 should be
small enough to allow the detection of the center of a streaky region. To find
the central region of a burst, two or more simultaneous criteria are required.
After a number of trials of different detection schema the following criteria
were successful. A computer prqpam was written that searched the u,, and
uw,, data for those few events in which both u , , , aad ha reached the level
-%, and -uL2 at approximatsly the =me t h e with nqative slope,