REYNOLDS STRESS STRUCIZIRE IN TURBULENT BOUNDARY LAYER 289
A mechanism or model that we have previously proposed (Willmarth and
Tu, 1967) for the deterministic structure responsible for bursts and sweeps is
discussed in the light of the p m n t results.
2. EXP@RIWNTAL APPARATUS AND h"HODS
The experiments wen conducted in a thick (b ?5: 5 in.) turbuknt boundary
layer in the 5 x 7 ft wind tunnel of the Department of Aerospace Engineering at The University of Michigan. Most of the measurements were made in
a boundary layer with a thick sublayer that was produced at low free stream
speeds, U, = 20 ft/sec. A few measurements were ma& at higher free
stream speeds U, 'Y 200 ft/sec. The boundary layer parameters arc set forth
in Table I.
A complete description of the experimental apparatus has been given in
Willmarth and Lu (1971, 1972) and Lu and Willmarth (1972). This includes
the mean flow in the boundary layer, the hot wire probes, the electronic
equipment, the frequency modulated analog tape recorder and the analog to
digital converter used to prepam the data (which were reproduced at oneeighth the recording speed) for digital proccseing. Almost all d the data
reduction was performed with an IBM 360/67 computcp using several simple
FORTRAN programs and a few assembly language subroutines. We emphasize that the use of the digital computer for data reduction makes it possible
to measure a great variety of difFcrant statistical parameters inexpensively
and efficiently. Indeed, it would not have been possible to perform dl the
measurements that we shall describe using analog methods, owing to limitations of time, money, and manpower.
3. EXPERIMENTAL MEASUREMENTS
The measurements are described in chronological order, which is naturally related to our increased understanding of the pbenomena during the
course of the investigation and to our increased skill in devising and writing
new data reduction programs for the computer. The first exploratory measurements were made using analog methods to detect eruptions of low speed
fluid near the wall and to selcct samples of the signal utr when eruptions
occurred. It soon became apparent that analog datn gathering and processing methods were e x t m d y inefficient in comparison to digital methods.
Using analog methods om must first construct the special circuitry and then
operate it successfully over the long periods of time required to obtain stable
average values. When sampling random data, the simplest analog averaging
method, a capacitor fitted with a resistive leak, is very wasteful of time and
A mechanism or model that we have previously proposed (Willmarth and
Tu, 1967) for the deterministic structure responsible for bursts and sweeps is
discussed in the light of the p m n t results.
2. EXP@RIWNTAL APPARATUS AND h"HODS
The experiments wen conducted in a thick (b ?5: 5 in.) turbuknt boundary
layer in the 5 x 7 ft wind tunnel of the Department of Aerospace Engineering at The University of Michigan. Most of the measurements were made in
a boundary layer with a thick sublayer that was produced at low free stream
speeds, U, = 20 ft/sec. A few measurements were ma& at higher free
stream speeds U, 'Y 200 ft/sec. The boundary layer parameters arc set forth
in Table I.
A complete description of the experimental apparatus has been given in
Willmarth and Lu (1971, 1972) and Lu and Willmarth (1972). This includes
the mean flow in the boundary layer, the hot wire probes, the electronic
equipment, the frequency modulated analog tape recorder and the analog to
digital converter used to prepam the data (which were reproduced at oneeighth the recording speed) for digital proccseing. Almost all d the data
reduction was performed with an IBM 360/67 computcp using several simple
FORTRAN programs and a few assembly language subroutines. We emphasize that the use of the digital computer for data reduction makes it possible
to measure a great variety of difFcrant statistical parameters inexpensively
and efficiently. Indeed, it would not have been possible to perform dl the
measurements that we shall describe using analog methods, owing to limitations of time, money, and manpower.
3. EXPERIMENTAL MEASUREMENTS
The measurements are described in chronological order, which is naturally related to our increased understanding of the pbenomena during the
course of the investigation and to our increased skill in devising and writing
new data reduction programs for the computer. The first exploratory measurements were made using analog methods to detect eruptions of low speed
fluid near the wall and to selcct samples of the signal utr when eruptions
occurred. It soon became apparent that analog datn gathering and processing methods were e x t m d y inefficient in comparison to digital methods.
Using analog methods om must first construct the special circuitry and then
operate it successfully over the long periods of time required to obtain stable
average values. When sampling random data, the simplest analog averaging
method, a capacitor fitted with a resistive leak, is very wasteful of time and