Micro-Measurements

Strain Gages and Instruments Tech Note TN-503

Measurement of Residual Stresses

by the Hole-Drilling* Strain Gage Method
I. Residual Stresses and Their Measurement specimens, and are not readily adaptable to real test objects
of arbitrary size and shape.
Residual (locked-in) stresses in a structural material or
component are those stresses that exist in the object without X-ray diffraction strain measurement, which does not
(and usually prior to) the application of any service or re­quire stress relaxation, offers a nondestructive alternative
other external loads. Manufacturing processes are the to the foregoing methods, but has its own severe limitations.
most common causes of residual stress. Virtually all Aside from the usual bulk and complexity of the equipment,
manufacturing and fabricating processes — casting, welding, which can preclude field application, the technique is limited
machining, molding, heat treatment, etc. — introduce to strain measurements in only very shallow surface layers.
residual stresses into the manufactured object. Another Although other nondestructive techniques (e.g., ultrasonic,
common cause of residual stress is in-service repair or electromagnetic) have been developed for the same purposes,
modification. In some instances, stress may also be induced these have yet to achieve wide acceptance as standardized
later in the life of the structure by installation or assembly methods of residual stress analysis.
procedures, by occasional overloads, by ground settlement The Hole-Drilling Method
effects on underground structures, or by dead loads which
may ultimately become an integral part of the structure. The most widely used modern technique for measuring
residual stress is the hole-drilling strain-gage method of
The effects of residual stress may be either beneficial
stress relaxation, illustrated in Figure 1.
or detrimental, depending upon the magnitude, sign,
and distribution of the stress with respect to the load- Briefly summarized, the measurement procedure involves
induced stresses. Very commonly, the residual stresses are six basic steps:
detrimental, and there are many documented cases in which
these stresses were the predominant factor contributing • A special three- (or six-) element strain gage rosette is
to fatigue and other structural failures when the service installed on the test part at the point where residual
stresses were superimposed on the already present residual stresses are to be determined.
stresses. The particularly insidious aspect of residual stress • The gage grids are wired and connected to a multi-
is that its presence generally goes unrecognized until after channel static strain indicator, such as the Micro-
malfunction or failure occurs. Measurements Model P3 (three-element gage), or
Measurement of residual stress in opaque objects cannot be System 5000 (six-element gage).
accomplished by conventional procedures for experimental * Drilling implies all methods of introducing the hole (i.e., drilling,
stress analysis, since the strain sensor (strain gage, milling, air abrasion, etc).
photoelastic coating, etc.) is totally insensitive to the history
of the part, and measures only changes in strain after Strain Gage
installation of the sensor. In order to mea­sure residual stress Rosette
with these standard sensors, the locked-in stress must be
relieved in some fashion (with the sensor present) so that the
sensor can register the change in strain caused by removal of
Drilled Hole
the stress. This was usually done destructively in the past —
by cutting and sectioning the part, by removal of successive
Tech Note

surface layers, or by trepanning and coring.

With strain sensors judiciously placed before dissecting the
part, the sensors respond to the deformation produced by
relaxation of the stress with material removal. The initial
residual stress can then be inferred from the measured
strains by elasticity considerations. Most of these techniques
are limited to laboratory applications on flat or cylindrical Figure 1. Hole-Drilling Strain Gage Method
 TN-503-6
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

• A precision milling guide (Model RS-200, shown in location. This occurs because every perpendicular to a
Figure 1) is attached to the test part and accurately free surface (the hole surface, in this case) is necessarily a
centered over a drilling target on the rosette. principal axis on which the shear and normal stresses are
• After zero-balancing the gage circuits, a small, shallow zero. The elimination of these stresses on the hole surface
hole is drilled through the geometric center of the changes the stress in the immediately surrounding region,
rosette. causing the local strains on the surface of the test object to
change correspondingly. This principle is the foundation for
• Readings are made of the relaxed strains, corresponding the hole-drilling method of residual stress measurement,
to the initial residual stress. first proposed by Mathar.2
• Using special data-reduction relationships, the In most practical applications of the method, the drilled
principal residual stresses and their angular orientation hole is blind, with a depth which is: (a) about equal to
are calculated from the measured strains. its diameter, and (b) small compared to the thickness of
The foregoing procedure is relatively simple, and has been the test object. Unfortunately, the blind-hole geometry is
standardized in ASTM Standard Test Method E 837.1 sufficiently complex that no closed-form solution is available
Using commercially available equipment and supplies, and from the theory of elasticity for direct calculation of the
adhering to the recommendations in the ASTM standard, residual stresses from the measured strains — except by
the hole-drilling method can be applied routinely by the introduction of empirical coefficients. A solution can
any qualified stress analysis technician, since no special be obtained, however, for the simpler case of a hole drilled
expertise is required for making the measurements. The completely through a thin plate in which the residual stress
method is also very versatile, and can be performed in is uniformly distributed through the plate thickness. Because
either the laboratory or the field, on test objects ranging of this, the theoretical basis for the hole-drilling method
widely in size and shape. It is often referred to as a “semi- will first be developed for the through-hole geometry, and
destructive” technique, since the small hole will not, in many subsequently extended for application to blind holes.
cases, significantly impair the structural integrity of the
part being tested (the hole is typically 1⁄32 to 3⁄16 in [0.8 to Through-Hole Analysis
4.8 mm] in both diameter and depth). With large test objects,
Depicted in Figure 2a (following) is a local area within a thin
it is sometimes feasible to remove the hole after testing is
plate which is subject to a uniform residual stress, x. The
completed, by gently blending and smoothing the surface
initial stress state at any point P (R, ) can be expressed in
with a small hand-held grinder. This must be done very
polar coordinates by:
carefully, of course, to avoid introducing residual stresses in
the process of grinding. x
 r′ = (1 + cos 2) (1a)
NOTE 1: In its current state of development, the hole- 2
drilling method is intended primarily for applications in x
which the residual stresses are uniform throughout the ′ = (1 ± cos 2) (1b)
drilling depth, or essentially so. While the procedures for 2
data acquisition and reduction in such cases are well- x
established and straightforward, seasoned engineering  r′ = − sin 2 (1c)
judgment is generally required to verify stress uniformity and 2
other criteria for the validity of the calculated stresses. This Figure 2b represents the same area of the plate after a
Tech Note contains the basic information for understanding small hole has been drilled through it. The stresses in
how the method operates, but cannot, of course, encompass the vicinity of the hole are now quite different, since r
the full background needed for its proper application in all and rθ must be zero everywhere on the hole surface. A
cases. An extensive list of technical references is provided in solution for this case was obtained by G. Kirsch in 1898, and
the Bibliography as a further aid to users of the method. yields the following expressions for the stresses at the point
P (R, ):3
NOTE 2: Manual calculation of residual stresses from the
measured relaxed strains can be quite burdensome, but there 
 x  11   x  33 44 
is available a specialized computer program, H-DRILL, that  rr′′′′ == 2x 11 −− r 22  ++ 2x 11 ++ r 44 ±± r 22  cos
 cos 22
 (2a)
2  r  2  r r 
Tech Not

completely eliminates the computational labor.

II. Principle and Theory of the 

 x  11   x  33 
Hole-Drilling Strain Gage Method ′′′′ == 2x 11 ++ r 22  ±± 2x 11 ++ r 44  cos
 cos 22
 (2b)
2  r  2  r 
The introduction of a hole (even of very small diameter)

into a residually stressed body relaxes the stresses at that  rr′′′′ == ±±  xx 11 −− 34 ++ 22  sin
3 2
sin 22
 (2c)
22  rr 4 rr 2 

rr = R (R ≥ R )
R
= R ≥ Roo
R
Roo
 TN-501-2
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

(a)  x (1 + ν )  1 3 4 
Z r = −  2 − 4 cos 2 + 2 cos 2  (4a)
x 2E r r r (1 + ν ) 
 x (1 + ν )  1 3 4ν 
R θ =−  − 2 + 4 cos 2 − 2 coss 2  (4b)
2E  r r r (1 + ν ) 

 r′ = x (1P(R, )
+ cos2)
 2  r′ = x (1 + cos 2) The preceding equations can be written in a simpler
2
Y Xx form, demonstrating that along a circle at any radius
′ = (1 ± cos2) R (R ≥ Ro) the relieved radial and tangential strains vary in a
2  ′ = x (1 ± cos 2)
2 sinusoidal manner:
 x x  r =  x ( A + B cos 2 )
 r′ = − sin 2  (5a)
2 r′ = − x sin 2

2 θ =  x ( − A + C cos 2 ) (5b)
Comparison of Equations (5) with Equations (4)
(b) demonstrates that coefficients A, B, and C have the following
x definitions:
Ro 1 +ν  1 
 
 r′′ = Rx 1 − 2  + x 1 + 4 ± 2 Acos = 2− (6a)
1 3 4  
2  P (R, r  ) 2  r r  2 E  r2 
 
 r′′ = x 1 − 2  + x 1B+ = 4− ±1 + ν2 cos
 1 3 4 3
 24  1
x  Y 1  x  2  r  2  r r   2 – 4  (6b)
3 X4  2 E  1 + ν r r 
 r′′ = 1− 2 + 1 + 4 ± x2  cos 21  x  3
2  r  2  r′′ = r 1 + 2  ± 1 + 4 cos 2 1 + ν   4ν  1 3
2  r 
 x  1   x  C =3− 2 E  −  1 + ν  r 2 + r 4 
2 r
(6c)
′′=x 1+ 2 ± 1 + 4 cos 2
2  r  2  r 
 
′′ = x 1 + 2  ± x 1 + 4  cos
1 3
2x 3 2
2  r  2   r′′r = ± 1 − 4 + 2 sin 2
2  r r  Thus, the relieved strains also vary, in a complex way, with
Figure 2. Stress states at P (R, ), x  3 
2distance
 r′′ = ± 1 − 4 + 2 sin 2 from the hole surface. This variation is illustrated in
before and after the introduction of a hole. 2  r r 
Figure 3 on page 22, where the strains are plotted along

 r′′ = ± x 1 − 4 + 2  sin 2r= R ( R ≥ R )
3 2
the principal axes, at  = 0° and  = 90°. As shown
2  r r  o
where:
Ro by the figure, the relieved strains generally decrease as
( R ≥ Ro )
R
r =
Ro = hole radius distance from the hole increases. Because of this, it is
Ro desirable to measure the strains close to the edge of the
( R ≥ Ro )
R
r = R = arbitrary radius from hole center
Ro Ro = hole radius hole in order to maximize the strain gage output signal.
On the
R = arbitrary radius from holeother
center hand, parasitic effects also increase in the
Ro = hole radius
immediate vicinity of the hole. These considerations, along
R = arbitrary radius from hole center with practical aspects of strain gage design and application,
necessitate a compromise in selecting the optimum radius
Subtracting the initial stresses from the final (after drilling) (R) for gage location. Analytical and experimental studies
stresses gives the change in stress, or stress relaxation at point have established a practical range of 0.3 < r < 0.45 where
P (R, ) due to drilling the hole. That is: r = Ro /R and R is the radius to the longitudinal center of
the gage.
 r =  r′′ −  r′ (3a)
It can be noticed from Figure 3 that for  = 0° (along the
  =  ′′−  ′ (3b) axis of the major principal stress) the relieved radial strain,
r, is considerably greater than the tangential strain, θ, in
 r  =  r′′ −  r′ (3c) the specified region of measurement. As a result, commercial
Tech Not

strain gage rosettes for residual stress analysis are normally

Substituting Equations (1) and (2) into Equations (3) yields designed with radially oriented grids to measure the relieved
the full expressions for the relaxed (or relieved) stresses. If radial strain, r. This being the case, only Equation (5a) is
the material of the plate is homogeneous and isotropic in its directly relevant for further consideration in this summary.
mechanical properties, and linear-elastic in its stress/strain It is also evident from the figure that the relieved radial
behavior, these equations can then be substituted into the strain along the major principal axis is opposite in sign to
biaxial Hooke’s law to solve for the relieved normal strains at the initial residual stress. This occurs because the coefficients
the point P (R, ). The resulting expressions are as follows:
 TN-503-6
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

 x ( A +different
 ra=slightly
Or, in ) +  y ( A − B cos 2 )
B cos 2form,
+1 0 –1
5 ( ) (
 r = A  x +  y + B  x −  y cos2) (9b)

4 Equations (9) represent the basic relationship underlying
the hole-drilling method of residual stress analysis. This
 = 0°
relationship must be inverted, of course, to solve for the two
R /Ro

3 principal stresses and the angle  in terms of the measured

 = 90° strains that accompany stress relaxation. Since there are
three unknown quantities, three independent measurements
r of the radial strain are required for a complete solution.
2 
These three measurements can be substituted successively
into Equation (9a) or Equation (9b) to yield three equations
1 +1 which are then solved simultaneously for the magnitudes
 and directions of the principal stresses.

The common procedure for measuring the relieved strains is
0 to mount three resistance strain gages in the form of a rosette
Ro r
around the site of the hole before drilling. Such a rosette is
shown schematically in Figure 4, where three radially oriented
–1 strain gages are located with their centers at the radius R from
1 2 3 4 5
R /Ro the center of the hole site. Although the angles between gages
can be arbitrary (but must be known), a 45-degree angular
increment leads to the simplest analytical expressions, and
Figure 3. Variation of relieved radial and tangential
thus has become the standard for commercial residual stress
strains with distance (along the principal axes) from the
rosettes. As indicated in Figure 4, 1 is the acute angle from
center of the drilled hole — uniaxial residual stress.
the nearer principal axis to gage no. 1, while 2 = 1 +45° and
3 = 1 + 90°, with positive angles measured in the direction
A and B in Equation (5a) are always negative, and (for of gage numbering. It should be noted that the direction of
 = 0°) cos 2 = +1. gage numbering for the rosette type sketched in Figure 4 is
clockwise, since gage no. 2, although physically at position
The preceding treatment considered only the simplest case,
2a, is effectively at position 2b for gage numbering purposes.
uniaxial residual stress. In practice, however, residual stresses
Equations (9) can be used to verify that both locations for
are more often biaxial, with two nonzero principal stresses.
This condition can readily be incorporated in the analysis by
employing the superposition principle, which is applicable to – +
linear-elastic material behavior. Referring to Figure 2 again,
it is apparent that had the uniaxial residual stress been along Y
only the Y axis instead of the X axis, Equations (1) and (2)
would still apply, with cos 2 replaced by cos 2( + 90°), or 1
by the equivalent, –cos 2. Thus, the relieved radial strain at
45°
the point P(R, ) due to uniaxial residual stress in only the
Y direction can be written as: 1

 ry =  y ( A − B cos2 ) (7) 2b 45°

And, employing the corresponding notation, Equation (5a)
becomes:
Tech Note

 rx =  x ( A + B cos2) (8) 3
Ro R
When both residual stresses are present simultaneously,
the superposition principle permits algebraic addition of 2a
Equations (7) and (8), so that the general expression for the x
relieved radial strain due to a plane biaxial residual stress Figure 4. Strain gage rosette arrangement
state is: for determining residual stress.
 r =  x ( A + B cos 2 ) +  y ( A − B cos 2 ) (9a)

(
 r = A  x +  y + B  x −  y cos2 )
 TN-501-2
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

gage no. 2 produce the same result providing the residual 3 > 1:  refers to max
stress is uniform over the area later occupied by the hole.
For general-purpose applications, location 2a is usually 3 < 1:  refers to min
preferred, because it minimizes the possible errors caused by   3 = 1:  = ±45°
any eccentricity of the drilled hole. When space for the gage
is limited, as in measuring residual stresses near a weld or 2 < 1: max at +45°
abutment, location 2b permits positioning the hole closest to 2 > 1: max at –45°
the area of interest.
Careful consideration must also be given to determining
Equation (9b) can now be written three times, once for each the appropriate values for coefficients A and B. As defined
gage in the rosette: algebraically in Equations (6), they apply only when
1 = A(x + y) + B(x – y) co 2 (10a) the conditions imposed by the Kirsch solution are met.
This solution gives the stress distribution at points with
2 = A(x + y) + B(x – y) co 2(+ 45°) (10b) coordinates (r, ) around a circular hole through a thin,
3 = A(x + y) + B(x – y) co 2(+ 90°) (10c) wide plate subjected to uniform plane stress. However,
comparison of Figures 3 and 4 illustrates that, since the
strain gage grids in the rosette have finite areas, they sense
When Equations (10) are solved simultaneously for the varying strain distributions such as those plotted in Figure 3.
principal stresses and their direction, the results can be Thus, the output of each gage tends to represent the average
expressed as: strain over the area of the grid. Moreover, because the grids
 + 1
 max = 1 3 –
4A 4B
( 3 −  1 ) + ( 3 +  1 – 2 2 ) (11a)
2 2 are usually composed of parallel lines, those lines which are
not directly on the centerline of a radially oriented grid are
1 +  3 1 not radial. Therefore, the gages are slightly sensitive to the
 min =
4A
+
4B
( 3 −  1 )2 + ( 3 +  1 – 2 2 )2 (11b) tangential strain, as well as the radial strain. As a result,
more accurate values for the coefficients can be obtained
 1 − 2 2 +  3 by integrating Equations (4) over the areas of the respective
tan2 =
1 −  3 gage grids. The coefficients thus determined, which account
for the finite strain gage area, are designated here by A and
where  is the angle from the nearer principal axis to gage B to distinguish them from the values at a point as defined
no. 1 (in the direction of gage numbering, if positive; or by Equations (6). An alternative method for obtaining
opposite, if negative). A and B is to measure them by experimental calibration.
Reversing the sense of  to more conveniently define the The procedure for doing so is described in Section III,
angle from gage no. 1 to the nearer axis, while retaining the “Determining Coefficients A and B.” When performed
foregoing sign convention, correctly, this procedure is potentially the most accurate
means for evaluating the coefficients.
 1 − 2 2 +  3
tan2 = (11c) When employing conventional strain gage rosettes for
 3 − 1
exper­imental stress analysis, it is usually recommended that
The following important comments about Equations the strain measurements be corrected for the transverse
(11) should be carefully noted. These equations are very sensitivity of the gages. Correction relationships for this
similar in appearance to the data-reduction relationships purpose are given in Tech Note TN-509. These relationships
for conventional strain gage rosettes, but the differences are are not directly applicable, however, to the relieved strains
significant. The coefficients A and B not only incorporate mea­sured with a residual stress rosette by the hole-drilling
the elastic properties of the test material, but also reflect method.
the severe attenuation of the relieved strains relative to In the residual stress case, the individual gages in the rosette
the relaxed stress. It can be observed, in addition, that are effectively at different locations in a spatially varying
the signs between terms in Equations (11a) and (11b) are strain field. As a result, the relieved axial and transverse
opposite to those in the conventional rosette equations. This strains applied to each gage are not related in the same
Tech Not

occurs because A and B are always negative; and thus, since manner as they are in a uniform strain field. Rigorous
Equation (11a) is algebraically greater than Equation (11b), correction would require evaluation of the coefficient C
the former must represent the maximum principal stress. [actually, its integrated or calibrated counterpart, C — see
Equation (11c) is identical to that for a conventional Equations (6)], for both the through-hole and blind-hole
three-element rectangular rosette, but must be interpreted geometries. Because of the foregoing, and the fact that the
differently to determine which principal stress is referred to transverse sensitivities of Micro-Measurements residual
gage no. 1. The following rules can be used for this purpose: stress rosettes are characteristically very low (approximately
1%), it is not considered necessary to correct for transverse
 TN-503-6
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

sensitivity. Kabiri 4, for example, has shown that the error 120
due to ignoring transverse sensitivity (in the case of uniaxial GAGE #1
residual stress) is negligible compared to the remaining
1
uncertainties in the measurement and data-reduction 100
3
procedures. 2

PERCENT STRAIN RELIEVED

Blind-Hole Analysis 80
The theoretical background for the hole-drilling method 5 )
f.
was developed in the preceding treatment on the basis of a

e
(R
small hole drilled completely through a thin, wide, flat plate 60

ss
6)

ne
subjected to uniform plane stress. Such a configuration is far

ef.

Vig
(R
from typical of practical test objects, however, since ordinary

r&
ey
machine parts and structural members requiring residual 40

dle
e ls
D Gage

en
0K
stress analysis may be of any size or shape, and are rarely Do

9R
0 .4
thin or flat. Because of this, a shallow “blind” hole is used in

0.2
Z
most applications of the hole-drilling method. 20

The introduction of a blind hole into a field of plane stress

Do /D
produces a very complex local stress state, for which no exact 0
solution is yet available from the theory of elasticity. Fortu­- 0 0.1 0.2 0.3 0.4
n­ately, however, it has been demonstrated by Rendler and Z/D
Vigness5 that this case closely parallels the through-hole Figure 5. Relieved strain versus ratio of hole depth to gage circle
condition in the general nature of the stress distribution. diameter (strains normalized to 100% at Z/D = 0.4).
Thus, the relieved strains due to drilling the blind hole still
vary sinusoidally along a circle concentric with the hole, in The general variation of relieved strain with depth is illustrated
the manner described by Equations (9). It follows, then, that in Figure 5, where the strains have been normalized, in this
these equations, as well as the data-reduction relationships case, to 100% at Z/D = 0.4. The data include experimental
in Equations (11), are equally applicable to the blind-hole results from two different investigators demonstrating the
implementation of the method when appropriate blind-hole manner in which the relieved-strain function is affected by the
coefficients A and B are em­ployed. Since these coefficients ratio of hole diameter to gage circle diameter (Do/D). Both
cannot be calculated directly from theoretical considerations, cases involve uniform uniaxial (plane) stress, in specimens
they must be obtained by empirical means; that is, by that are thick compared to the maximum hole depth. The
experimental calibration or by numerical procedures such as curves plotted in the figure are considered representative
finite-element analysis. of the response to be expected when the residual stress is
uniform throughout the hole depth.
Several different investigators [e.g., (20)–(23)] have published
finite-element studies of blind-hole residual stress analysis. An important contribution of the Rendler and Vigness
The most recently developed coefficients by Schajer are work is the demonstration that, for any given set of material
incorporated in ASTM standard E 837, and are shown properties, E and  , coefficients A and B are simply
graphically for the case of uniform stress in Figure 8 of this geometric functions, and thus constants for all geometrically
Tech Note. The computer program H-DRILL uses these similar cases. This means that once the coefficients have
coefficients. been determined for a particular rosette configuration, the
rosette size can be scaled upward or downward and the same
Compared to the through-hole procedure, blind-hole
coefficients will still apply when the hole diameter and depth
analysis involves one additional independent variable;
are similarly scaled (assuming, of course, the same material).
namely, the dimensionless hole depth, Z/D (see Figure 5).
Several different approaches have been taken in attempting
Thus, in a generalized functional form, the coefficients can
to remove the material depen­dency from A and B, leaving
be expressed as:
only the geometric dependence. One of these, proposed by
A = fA (E, ν, r, Z/D) (12a) Schajer,7 is adopted in this Tech Note. Schajer introduced
Tech Not

two new coefficients, denoted here as a and b, and defined

B = fB (E, ν, r, Z/D) (12b) as follows:
For any given initial state of residual stress, and a fixed
hole diameter, the relieved strains generally increase (at 2 EA
a decreasing rate) as the hole depth is increased. Therefore, a=− (13a)
1 +ν
in order to maximize the strain signals, the hole is normally
drilled to a depth corresponding to at least Z/D = 0.4 (ASTM b = −2 EB (13b)
E 837 specifies Z/D = 0.4 for the maximum hole depth).
 TN-501-2
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

By comparison with Equations (6), it can be seen that the initial residual stress must be low enough to avoid the
for the through hole, at least a is material-independent, risk of local yielding due to the stress-concentrating effect
and b depends only weakly on Poisson’s ratio. Schajer of the hole.
has determined from finite-element calculations that for
Basically, the calibration procedure involves measuring the
blind holes, a and b vary by less than 2% for the range of
rosette strains under the same applied load or calibration
Poisson’s ratio from 0.25 to 0.35.
stress, σc , both before and after drilling the hole. Such a
procedure is necessary in order to eliminate the effect of
III. Determining Coefficients A and B the strain relief that may occur due to the relaxation of
Whether the residual stress analysis application involves any initial residual stress in the calibration specimen. With
through-hole or blind-hole drilling, the coefficients A and this technique, the observed strain difference (before and
B (or a and b ) must be determined to calculate the stresses after hole drilling) is caused only by the applied calibration
from the relieved strains. In the case of the through hole, stress, and is uniquely related to that stress. The steps in the
reasonably accurate values of the coefficients can be obtained calibration procedure can be summarized briefly as follows,
for any particular case by analytical means, if desired. This noting that the strains in only grid no. 1 and grid no. 3 need
is done by integrating, over the area of the gage grid, the to be measured, since these grids are known to be aligned
component of strain parallel to the primary strain-sensing with the principal axes of the specimen.
axis of the gage. Given the details of the grid geometry 1. Zero-balance the strain gage circuits.
(line width and spacing, number of lines, etc.), Slightly
greater accuracy may be obtained by integrating along the 2. Apply a load, P, to the specimen to develop the desired
individual grid lines. This method cannot be applied to calibration stress, c.
blind-hole analysis because closed-form expressions relating 3. Measure strains ′1 and ′3 (before drilling).
the relieved strains to the residual stress, in terms of hole
4. Unload the specimen, and remove it from the testing
depth, are not available.
machine.
Experimental Calibration 5. Drill the hole, as described in Section V, “Experimental
The needed coefficients for either through-hole or blind- Techniques”.
hole analysis can always be determined by experimental 6. Replace the specimen in the testing machine, re-zero
calibration. This procedure is particularly attractive since it the strain gage circuits, and then reapply exactly the
automatically accounts for the mechanical properties of the same load, P.
test material, strain gage rosette geometry, hole depth and 7. Measure strains ′1′ and  ′3′ (after drilling).
diameter, and the strain-averaging effect of the strain gage
The calibration strains corresponding to the load, P, and the
grid. When performed correctly, with sufficient attention
stress, σc , are then:
to detail, it is potentially the most accurate means for
determining the coefficients. Its principal disadvantage is c1 = ′1′ – ′1
that the calibration must be repeated each time a different set
of geometric parameters is involved. c3 = ′3′ – ′3

Calibration for A and B is accomplished by installing a Calibration reliability can ordinarily be improved by loading
residual stress strain gage rosette on a uniaxially stressed the specimen incrementally and making strain measurements
tensile specimen made from the same material as the test at each load level, both before and after drilling the hole.
part. The rosette should be oriented to align grid no. 1 This permits plotting a graph of c versus c1 and c3, so that
parallel to the loading direction, placing grid no. 3 along best-fit straight lines can be constructed through the data
the transverse axis of the specimen. Care must be taken that points to minimize the effect of random errors. It will also
the tensile stress is uniform over the cross section of the test help identify the presence of yielding, if that should occur.
specimen; i.e., that bending stress is negligible. To minimize The resulting relationship between the applied stress and
edge and end effects, the specimen width should be at least the relieved strain is usually more representative than that
ten times the hole diameter, and the length between machine obtained by a single-point determination.
Tech Not

grips, at least five times the width. When determining A Since the calibration is performed with only one nonzero
and B for blind-hole applications, a specimen thickness of principal stress, Equation (5a) can be used to develop
five or more times the hole diameter is recommended. For expressions for the calibrated values of A and B. Successively
through-hole calibration, the thickness of the calibration substituting  = 0° (for grid no. 1) and  = 90° (for grid no.
specimen is preferably the same as that of the test part. It 3) into Equation (5a):
is also important that the maximum applied stress during
calibration not exceed one-half of the proportional limit c1 = c [A + B cos (0°)] = c (A + B)
stress for the test material. In any case, the applied stress plus c3 = c [A + B cos (2 x 90°)] = c (A – B)
 TN-503-6
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

Solving for A and B, • same hole form (through hole or full-depth blind hole);
• uniform stress with depth;
 c1 +  c 3
A= (14a) • nominally uniform in-plane stress at the hole.
2 c
 c1 −  c 3 Coefficients for Micro-Measurements
B= (14b)
2 c Residual Stress Rosettes
Micro-Measurements supplies special strain gage rosettes
The procedure described here was applied to a through- for residual stress analysis in four basic configurations,
hole specimen made from Type 304 stainless steel, and the illustrated and described in Figure 7. Among other features,
calibration data are plotted in Figure 6. It can be seen from the these rosette designs incorporate centering patterns for
figure that for this geometry (Do /D = 0.35) and material, c1 positioning the boring tool precisely at the center of the gage
and c3 are –90 and +39, respectively, when c is 10 000 psi circle. All RE and UL designs have geometrically similar
[69 MPa]. Substituting into Equations (14), grid configurations, with the gage-circle diameter equal to

A = –0.25 × 10–8 psi–1 [–0.36 × 10–12 Pa–1] EA-XX-062RE-120

This geometry conforms to the early Rendler
B = –0.65 × 10–8 psi–1 [–0.94 × 10–12 Pa–1] and Vigness design5 and has been used in
most reported technical articles (see refer-
Although the preceding numerical example referred to the ences). It is available in a range of sizes to
through-hole coefficients, the same procedure is followed accommodate applications requiring different
in calibrating for full-depth blind-hole coefficients. Once A hole diameters or depths.
and B have been obtained in this manner, the corresponding
material-independent coefficients, a and b, can be calculated CEA-XX-062UL-120
from Equations (13) if the elastic modulus and Poisson’s ratio This rugged, encapsulated de­sign incorpo-
of the test material are known. If desired, the procedure can rates all practical advantages of the CEA
then be repeated over the practical range of Do /D to permit strain gage series (integral copper solder-
plotting curves of a and b for all cases of interest. ing tabs, conformability, etc.). Installation
time and expense are greatly reduced, and
It should be noted that the values for the basic coefficients A all solder tabs are on one side of the gage to simplify leadwire
and B obtained from a particular calibration test are strictly routing from the gage site. It is compatible with all methods of
applicable only for residual-stress measurement conditions introducing the hole, and the strain gage grid geometry is identi-
that exactly match the calibration conditions: cal to the 062RE pattern.

• material with the same elastic properties; CEA-XX-062UM-120

Another CEA-Series strain gage, the
• same rosette geometry (but rosette orientation is 062UM grid widths have been reduced to
arbitrary); facilitate positioning all three grids on one
• same hole size; side of the measurement point as shown.
With this geometry, and appropriate trim-
100 ming, it is possible to position the hole closer to welds and other
GAGE #1 GAGE #3 irregularities. The user should be reminded, however, that the
data reduction equations are theoretically valid only when the
80 holes are well removed from free boundaries, discontinuities,
sc = 10 000 psi
STRESS (1000 psi)

10 abrupt geometric changes, etc. The UM design is compatible

STRESS [MPa]

with all methods of introducing the hole.

60

N2K-XX-030RR-350/DP
c3 = +39
c1 = –90

The K-alloy grids of this open-faced six-

1
40
5 element rosette are mounted on a thin,
3 high-performance laminated polyimide
Tech Not

2 20 film backing. Solder tabs are duplex cop-

per plated for ease in making solder joints
for lead attachment. Diametrically op-
posed circumferential and radial grids are
–100 –80 –60 –40 –20 0 20 40
to be wired in half-bridge configurations.
CALIBRATION MICROSTRAIN (c )
Figure 7. Residual stress strain gage
Figure 6. Stress versus relieved strain for calibration of
rosettes (shown approximately 2X).
coefficients A and B on 304 Stainless Steel (through-hole).
 TN-501-2
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

3.25 times the active gage length. The 062RE rosette, for IV. Measuring Nonuniform Residual Stresses
example, has a gage-circle diameter of 0.202 in [5.13 mm].
Because of this similitude, the same material-independent The coefficients given in this Tech Note and in ASTM 
coefficients a and b apply to all sizes of the RE rosette, E 837-99 are strictly applicable only to situations in which
and to the UL rosette, for geometrically similar holes (i.e., the residual stresses do not vary in magnitude or direction
for the same Do/D and Z/D ratios). The 062UM rosette with depth from the test-part surface. In reality, however,
configuration has the same ratio of gage circle to grid length, residual stresses may often vary significantly with depth,
but the grids are narrower to permit their close grouping on due, for example, to different manufacturing processes such
one side of the hole. As a result, the sensitivity of the gage to as casting, forging, heat treatment, shot peening, grinding,
the relieved strains is slightly greater, and coefficients specific etc.
to the 062UM are required for data reduction. Numerous finite-element studies have been made in attempts
The 030RR rosette is fundamentally different from the other to treat this situation [see, for instance, references (20)
rosettes illustrated in Figure 7. To begin with, this rosette through (23)]. The results of the finite element work by
includes both radially and circumferentially oriented grids Schajer have been incorporated in a computer program,
which are to be connected as half-bridge pairs. The 030RR H-DRILL, for handling stress variation with depth. When
rosette incorporates a number of features that contribute the measured strains from hole drilling do not fit the
to its greater output and higher accuracy compared to reference curves in Figure 10, or when there is any other
conventional three-element rosettes: (a) the individual basis for suspecting significant nonuniformity, the program
gridlines in the radial elements are purely radial, instead H-DRILL or some other finite-element-based program
of being simply parallel to the central gridline as in the is necessary to accurately determine the stresses from the
other rosettes; (b) for a given maximum hole diameter, measured relaxed strains.
the outermost radius of the grids is considerably less than
for the corresponding conventional rosettes, and thus the V. Experimental Techniques
grids sense slightly greater average released strains; and (c)
since the radial and circumferential grids get connected in As in all experimental methods, proper materials, application
a half-bridge configuration, the bridge output is augmented procedures, and instrumentation are essential if accurate
correspondingly, and the circuit is intrinsically self- results are to be obtained. The accuracy of the hole-drilling
temperature compensating. As a result of these features, the method is dependent chiefly upon the following technique-
030RR rosette yields about twice the output of the three- related factors:
element rosettes for a given state of residual stress, while
• strain gage selection and installation.
displaying better stability and accuracy.
• hole alignment and boring.
Since the sign of the residual stress is of primary importance
in determining its effect on the structural integrity of any • strain-indicating instrumentation.
mechanical component, the user of the six-element rosette • understanding the mechanical properties of the test
(030RR) must exercise care in connecting the rosette grids material.
into Wheatstone bridge circuits. To obtain the correct sign
in the instrument output signal, the radially oriented grids Strain Gage Selection and Installation
should always be connected between the positive excitation
Installing three individual strain gages, accurately spaced
and the negative signal terminals, while the tangentially
and oriented on a small circle, is neither easy to do nor
oriented grids are to be connected between the negative
advisable, since small errors in gage location or orientation
signal and negative excitation terminals.
can produce large errors in calculated residual stresses. The
The a and b coefficients for Micro-Measurements residual rosette con­figurations shown in Figure 7 have been designed
stress rosettes are provided graphically in Figure 8 on page and developed by Micro-Measurements specifically for
28, where the solid lines apply to full-depth blind holes and re­sid­ual stress measurement. The rosette designs incorporate
the dashed lines to through holes assuming, in both cases, centering marks for aligning the boring tool precisely at the
that the initial residual stress is uniform with depth. Both center of the gage circle, since this is critical to the accuracy
Tech Not

the through-hole and full-depth blind-hole coefficients of the meth­od.9,10,11 All configurations are available in a
plotted in Figure 8 were determined by a combination of range of temperature compensations for use on common
finite-element analysis and experimental verification. These structural metals. However, only the RE design is offered
coefficients are also supplied numerically in tabular form in in different sizes (031RE, 062RE, and 125RE), where the
ASTM E 837-99, where RE/UL rosettes are designated as three-digit prefix represents the gage length in mils (0.001 in
Type A, UM rosettes as Type B, and RR rosettes as Type [0.0254 mm]). The RE design is available either open-faced
C. For the blind-hole coefficients in the ASTM standard, or with Option SE (solder dots and encapsulation).
“full depth” corresponds to a value of 0.40 for the depth to
rosette-mean-diameter-ratio, Z/D.
Tech Not

(a) RE AND UL ROSETTES (b) UM ROSETTE (c) RR ROSETTE

0.8 0.8 1.3
1 1
1 2 b
1 1
TN-503-6

D 1 Do
Do D 1
0.7 0.7
1 1.2
3 3
Do D
2 
b

0.6 0.4D 0.6 0.4D 1.1

0.4D


b
0.5 0.5 1.0

 AND b
Micro-Measurements

0.4 0.4 0.9

 AND b
blind hole blind hole
through hole through hole

COEFFICIENTS a
0.3 0.3 0.8

COEFFICIENTS a AND b
a

a
blind hole
through hole a
0.2 0.2 COEFFICIENTS a 0.7

0.1 0.1 0.6

SUGGESTED LIMITS SUGGESTED LIMITS

0 0 0.5
0.30 0.35 0.40 0.45 0.50 0.30 0.35 0.40 0.45 0.50
Do /D Do /D

0.4
SUGGESTED LIMITS

0.3
0.40 0.45 0.50 0.55 0.60
Do /D

Figure 8. Full-depth data-reduction coefficients a and b versus dimensionless hole diameter (typical) for
Micro-Measurements residual stress rosettes, in accordance with ASTM E 837.
Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method
 TN-501-2
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

The UL and UM configurations are supplied in 1⁄16 in precisely in the center of the strain gage rosette.” More recent
[1.6 mm] gage length, and both types are fully encapsulated. studies have quantified the error in calculated stress due to
Both configurations have integral, copper-coated eccentricity of the hole. For example, with a hole that is 0.001
solder tabs, and offer all advantages of the popular in [0.025 mm] off-center of the 062RE or 062UL rosette, the
C-Feature strain gage series. These residual stress rosettes error in calculated stress does not exceed 3% (for a uniaxial
are constructed with self-temperature-compensated stress state).9,10,11 In practice, the required alignment precision
constantan foil, mounted on a flexible polyimide carrier. to within 0.001 in [0.025 mm] is accomplished using the
Gage resistance is 120 ohms ±0.4%. The 030RR six- RS-200 Milling Guide shown in Figure 9. The milling
element rosette incorporates self-temperature-compensated guide is normally secured to the test object by bonding its
K-alloy (modified Karma) foil on a laminated polyimide three foot pads with a quick-setting, frangible adhesive. A
film backing. Solder tabs are duplex copper plated for microscope is then installed in the unit and visual alignment
ease in making solder connections. Gage resistance is is achieved with the aid of the four X-Y adjustment screws
350 ohms ±0.4%. on the exterior of the guide.
Surface preparation for installing the rosettes is basically Boring
standard, as described in Application Note B-129. Caution
should be observed, however, in abrading the surface of Numerous studies on the effects of hole size and shape
the test object, since abrasion can alter the initial state of and machining procedures have been published. Rendler
residual stress.12 In general, it is important that all surface- and Vigness5 specified a specially dressed end mill which
preparation and gage-installation procedures be of the is compatible with the residual stress rosettes of Figure 7.
highest quality, to permit accurate measurement of the small The end mill is ground to remove the side cutting edges,
strains typically registered with the hole-drilling method. As and then relieved immediately behind the cutting face to
evidenced by the calibration data in Figure 6, the relieved avoid rubbing on the hole surface. It is imperative that the
strains corresponding to a given residual stress magnitude milling cutter be rigidly guided during the drilling operation
are considerably lower than those obtained in a conventional so that the cutter pro­gresses in a straight line, without side
mechanical test at the same stress level. Because of the small pressure on the hole, or friction at the noncutting edge.
measured strains, any drift or inaccuracy in the indicated These end mills generate the desired flat-bottomed and
gage output, whether due to improper gage installation, square-cornered hole shape at initial surface contact, and
unstable instrumentation, or otherwise, can seriously affect maintain the appropriate shape until the hole is completed.
the calculated residual stresses. In doing so, they fulfill the incremental drilling requirements
as stipulated in ASTM E 837. Specially dressed end mills
Strain-Measurement Instrumentation offer a direct and simple approach when measuring residual
stresses on readily machinable materials such as mild steel
The residual stress rosettes described in Figure 7 impose no
and some aluminum alloys. Figure 9b shows the RS-200
special instrument requirements. When measurements are
Milling Guide with the microscope removed and the end-
to be made in the field, a portable, battery-operated static
mill assembly in place. The end mill is driven through the
strain indicator, supplemented by a precision switch-and-
universal joint at the top of the assembly, by either a hand
balance unit, is ordinarily the most effective and convenient
drill or variable-speed electric drill.
instrumentation. The Model P3 Strain Indicator and Recorder
is ideally suited for this type of application. In the laboratory In 1982, Flaman13 first reported excellent results for residual
it may be convenient to use a computerized automatic data stress measurement using a high-speed (up to 400 000 rpm)
system such as System 5000, which will rapidly acquire and air turbine and carbide cutters. This technique maintains
record the data in an organized, readily accessible form. all of the advantages (good hole shape, adaptability to
A special offline, Windows ®-based computer program incremental drilling, etc.) of the specially dressed end mill
H-DRILL is also available to perform the calculations and while providing for easier operation and more consistent
determine the residual stress magnitudes in accordance with results. Further, the air turbine is highly recommended
ASTM E 837. The database for the program includes values for use with test materials that are difficult to machine, such
of the coefficients a and b for blind holes, and covers the full as Type 304 stainless steel. Carbide cutters are not effective
range of recommended hole dimensions applicable to all for penetrating glass, most ceramics, very hard metals, etc.;
Tech Not

Micro-Measure­ments residual stress rosettes. but diamond cutters have shown promise on these kinds of
test materials. Figure 9c shows the air turbine/carbide cutter
Alignment assembly installed in the same basic RS-200 Milling Guide.
Rendler and Vigness observed that “the accuracy of the Bush and Kromer14 reported, in 1972, that stress-free holes
(hole-drilling) method for field applications will be directly are achieved using abrasive jet machining (AJM). Modifi­
related to the operator’s ability to position the milling cutter cations and improvements were made to AJM by Procter
and Beaney,10 and by Bynum.15 Wnuk16 experienced good
Windows is a registered trademark of Microsoft, Inc.
 TN-503-6
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

(a) Alignment Setup results by mechanically adapting AJM apparatus for use
in the RS-200 Milling Guide. The principal advantage of
Eyepiece AJM is its reported ability to generate stress-free holes in
Microscope
virtually all materials. Its chief limitations center about
Tube
the considerable changes in hole shape as a function of
Locking
hole depth. The initial shape is saucer-like, and the final is
Collar cylindrical with slightly rounded corners. During drilling
there is also uncertainty as to the actual hole depth at any
X-Y
Adjustments stage. These factors make AJM a less practical technique for
Illuminator (4) determining the variation of relieved strain with hole depth,
as recommended in ASTM E 837.
Vertical Height
Adjustments Mechanical Properties
(3)
Cap As in any form of experimental stress analysis, the accuracy
Locking of residual stress measurement is limited by the accuracies to
Pad Nuts which the elastic modulus and Poisson’s ratio are known. But
typical uncertainties in the mechanical properties of common
(b) End-Mill Drilling Setup steel and aluminum alloys are in the neighborhood of 1 to
3% and, as such, are minor contributors to potential errors in
residual stress analysis. Much larger errors can be introduced by
Milling Bar with Locking
Universal Joint At- Collar
deviations from the assumptions involved in the basic theory,
tached as described in Section II. A key assumption, for instance,
Micrometer is linear-elastic material behavior. If the stress/­s train
Lock relationship for the test material is nonlinear (as is the case
Micrometer for cast iron), due to yielding or other causes, the calculated
Adjustment residual stresses will be in error.
Micrometer
When the initial residual stress is close to the yield strength
of the test material, the stress concentration caused by the
presence of the hole may induce localized yielding. It is
therefore necessary to establish a threshold level of residual
stress below which yielding is negligible. This problem
End Mill has been studied both experimentally and analytically,
and there is substantial agreement among the different
investigations.10,17,18 That is, errors are negligible when the
residual stress is less than 70% of the proportional limit of
(c) Hi-Speed Drilling Setup the test material — for both blind holes and through holes.
On the other hand, when the initial residual stress is equal
To Air Supply
to the proportional limit, errors of 10 to 30% (and greater)
Air Turbine have been observed. The error magnitude obviously depends
Spring Assembly on the slope of the stress/strain diagram in the post-yield
Assembly region; and tends to increase as the curve becomes flatter,
Grooved Nylon approaching the idealized perfectly plastic behavior.18
Collar
VI. Data Reduction and
Anti-Rotation
Ring Adapter Interpretation — Blind Hole
Mount
As recommended in ASTM E 837, it is always preferable
Basic RS-200
Tech Not

to drill the hole in small increments of depth, recording

Milling Guide
the observed strains and measured hole depth at each
increment. This is done to obtain data for judging whether
Carbide Cutter
the residual stress is essentially uniform with depth, thus
validating the use of the standard full-depth coefficients a
Figure 9. RS-200 Milling Guide, used for and b for calculating the stress magnitudes. If incremental
machining a precisely located flat-bottomed hole. measurements are not taken, there is no means for making
 TN-501-2
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

inferences about stress uniformity, and the calculated residual RE AND UL ROSETTE
stress may be considerably in error. In such cases, when
the stress varies with depth, it should be realized that the 100

PERCENT RELIEVED STRAIN

calculated stress is always lower than the actual maximum. 3 + 1
There is currently no absolute criterion for verifying 80
stress uniformity from the surface of the test piece to the 3 – 1 or
bottom of a full-depth hole. However, the incremental data, 60 3 + 1 – 22
consisting of relieved strain versus hole depth, can be used
in two different ways to aid in detecting a nonuniform stress
40 D
distribution. The first of these is to calculate, for each depth Gage

increment, the sums and differences of the measured strain Do

Z
data, 3 + 1 and 3 – 1 respectively.1 Express each set of data 20
as fractions of their values when the hole depth equals 0.4
times the mean diameter of the strain gage circle. Plot these 0
percent strains versus normalized hole depth. These graphs 0.0 0.1 0.2 0.3 0.4
should yield data points very close to the curves shown in Z/D
Figure 10. Data points which are removed from the curves in
Figure 10 indicate either substantial stress nonuniformity or UM ROSETTE
strain measurement errors. In either case, the measured data
are not acceptable for residual stress calculations using the 100
PERCENT RELIEVED STRAIN

full-depth coefficients a and b.

When a principal residual stress direction is closer to the 80 3 + 1
axial direction of gage no. 2 in Figure 4 than to either gage 3 – 1 or
nos. 1 or 3, the strain sum 3 + 1 – 22 will be numerically 60 3 + 1 – 22
larger than 3 – 1. In such a case, the percent strain data
check should be done using 3 + 1 – 22 instead of 3 – 1.
40 D
NOTE: This graphical test is not a sensitive indicator of Gage
Do
stress field uniformity. Specimens with significantly
20 Z
nonuniform stress fields can yield percentage relieved strain
curves substantially similar to those shown in Figure 10. The
main purpose of the test is to identify grossly nonuniform 0
stress fields. Further, the graphical comparison test using 0.0 –0.1 0.2 0.3 0.4
Z/D
3 – 1 or 3 + 1 – 22, for example, becomes ineffective when
the residual stress field approaches equal biaxial tension or
compression (1 ≅ 2 ≅ 3) as expected in surface blasting and RR ROSETTE
heat treating procedures. Comparison to the 3 + 1 plot is
ineffective when 3 = –1 (pure shear); however, this condition 100
is relatively uncommon in the practical industrial setting.
PERCENT RELIEVED STRAIN

Limitations and Cautions 80 3 + 1

3 – 1 or
Finite-element studies of the hole-drilling method by
60 3 + 1 – 22
Schajer and by subsequent investigators20,21,22,23 have shown
that the change in strain produced in drilling through any
depth increment (beyond the first) is caused only partly 40 D Gage
by the residual stress in that increment. The remainder Do
of the incremental relieved strain is generated by the
Tech Not

Z
20
residual stresses in the preceding increments, due to the
increasing compliance of the material, and the changing
stress distribution, as the hole is deepened. Moreover, the 0
relative contribution of the stress in a particular increment 0.0 –0.1 0.2 0.3 0.4
to the corresponding incremental change in strain decreases Z/D
rapidly with distance from the surface. As a result, the
Figure 10. Percent strain versus normalized hole depth
total relieved strain at full-hole depth is predominantly
for uniform stress with depth for different rosette
influenced by the stresses in the layers of material closest
types, after ASTM E 837.1
 TN-503-6
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

to the surface — say, in the upper third, or perhaps half, drilled (milled) without introducing significant additional
of the hole depth. At hole depths corresponding to residual stresses. To the degree that any of the foregoing
Z/D > 0.2, the stresses in these increments have very little requirements fail to be met, accuracy will be sacrificed
effect on the observed strains. This behavior is confirmed accordingly.
(for uniform stress) by the shape of the normalized strain
Strains relieved by drilling the hole are measured
graph in Figure 5, where about 80% of the total strain
conventionally, with static strain instrumentation. The
relief normally occurs in the first half of the hole depth.
indicated strains are characteristically much smaller,
Because of these characteristics, little, if any, quantitative
however, than they would be for the same stress state in
interpretation can safely be made of the incremental strain
an externally loaded test part. As a result, the need for
data for increments beyond Z/D = 0.2, irrespective of the
stable, accurate strain measurement is greater than usual.
analytical method employed for data reduction.
With incremental drilling, the strains measured in the first
To summarize, the ideal application of the hole-drilling few depth increments can be especially low, and errors of
method is one in which the stress is essentially uniform with a few microstrain can cause large percentage errors in the
depth. For this case, the data-reduction coefficients are well- calculated stresses for those depths.
established, and the calculated stresses sufficiently accurate
Beyond the above, it is also necessary that the underlying
for most engineering purposes — assuming freedom from
theoretical assumptions of the hole-drilling method be
significant experimental errors. Incremental drilling and
reasonably satisfied. In full-depth drilling per ASTM E 837,
data analysis should always be performed, however, to
the stress must be essentially uniform with depth, both in
verify the stress uniformity. If the graph of percent-strain-
magnitude and direction, to obtain accurate results. With
relieved versus Z/D (see Figure 10) suggests that the stress
finite-element and other procedures for investigating stress
is nonuniform with hole depth, then the procedure specified
variation in subsurface layers, it is required only that the
by ASTM E 837 is not applicable, and a program such as
directions of the principal stresses not change appreciably
H-DRILL must be used to calculate the stresses.
with depth. As for all conventional strain-gage rosette
Error and uncertainty are always present, in varying degrees, measurements, the data-reduction relationships assume
in all measurements of physical variables. And, as a rule, that the stress is uniformly distributed in the plane of the
their magnitudes are strongly dependent on the quality of the test surface. However, for residual stress measurements the
experimental technique as well as the number of parameters effective “gage-length” is the hole diameter rather than the
involved. Since residual stress determination by the hole- relatively large dimensions of the overall rosette geometry.
drilling method involves a greater number and variety of Consequently, uncertainties introduced by in-plane surface
techniques and parameters than routine experimental stress strain gradients are generally lower for residual stress
analysis, the potential for error is correspondingly greater. determination than for conventional static load testing. No
Because of this, and other considerations briefly outlined in generalization can currently be made about the effects of
the following, residual stresses cannot usually be determined steeply varying, nonlinear stress distributions in subsurface
with the same accuracy as stresses due to externally applied planes parallel to the rosette.
static loads.
Introduction of the small hole into the test specimen is BIBLIOGRAPHY
one of the most critical operations in the procedure. The
1. “Determining Residual Stresses by the Hole-Drilling
instruction manual for the RS-200 Milling Guide contains
Strain-Gage Method.” ASTM Standard E 837.
detailed directions for making the hole; and these should be
followed rigorously to obtain maximum accuracy. The hole 2. Mathar, J., “Determination of Initial Stresses by
should be concentric with the drilling target on the special Measuring the Deformation Around Drilled Holes.”
strain gage rosette. It should also have the prescribed shape Trans., ASME 56, No. 4: 249-254 (1934).
in terms of cylindricity, flat bottom, and sharp corner at the 3. Timoshenko, S. and J.M. Goodier, Theory of Elasticity,
surface. It is particularly necessary that the requirements on New York: McGraw-Hill (1951).
hole configuration be well-satisfied when doing incremental
drilling to examine stress variation with depth. Under 4. Kabiri, M., “Measurement of Residual Stresses by
the Hole-Drilling Method: Influences of Transverse
Tech Note

these same circumstances, it is important that the hole

depth at each drilling increment be measured as accurately Sensitiv ity of the Gage s and Rel ieve d Strai n
as possible, since a small absolute error in the depth can Coefficients,” Experimental Mechanics 25: (252-256)
produce a large relative error in the calculated stress. Because (Sept. 1984).
of practical limitations on measuring shallow hole depths, 5. Rendler, N.J. and I. Vigness, “Hole-drilling Strain-gage
the first depth increment should ordinarily be at least 0.005 Method of Measuring Residual Stresses.” Proc., SESA
in [0.13 mm]. Accurate measurement of the hole diameter XXIII, No. 2: 577-586 (1966).
is also necessary. Finally, it is imperative that the hole be
 TN-501-2
Micro-Measurements

Measurement of Residual Stresses by the Hole-Drilling Strain Gage Method

6. Kelsey, R.A., “Measuring Non-uniform Residual 20. Flaman, M.T., B.E. Mills, and J.M. Boag, “Analysis of
Stresses by the Hole-drilling Method.” Proc., SESA Stress-Variation-With-Depth Measurement Procedures
XIV, No. 1: 181-194 (1956). for the Centre Hole Method of Residual Stress
7. Schajer, G.S., “Appl ication of Fin ite Element Measurements.” Experimental Techniques 11: 35-37
Calculations to Residual Stress Measurements.” (June 1987).
Journal of Engineering Materials and Technology 103: 21. Schajer, G.S., “Measurement of Non-Uniform Residual
157-163 (1981). Stresses Using the Hole Drilling Method,” Journal of
8. Redner, S. and C.C. Perry, “Factors Affecting the Engineering Materials and Technology, 110, No. 4: Part
Accuracy of Residual Stress Measurements Using the I, 338-343; Part II, 344-349 (1988).
Blind-Hole Drilling Method.” Proc., 7th International 22. Ajovalasit, A., “Measurement of Residual Stresses
Conference on Experimental Stress Analysis. Haifa, by the Hole-Drilling Method: Inf luence of Hole
Israel: Israel Institute of Technology, 1982. Eccentricity.” Journal of Strain Analysis 14, No. 4: 171-
9. Sandifer, J.P. and G.E. Bowie, “Residual Stress by Blind- 178 (1979).
hole Method with Off-Center Hole.” Experimental 23. Beaney, E.M. and E. Procter, “A Critical Evaluation
Mechanics 18: 173-179 (May 1978). of the Centre-hole Technique for the Measurement of
10. Procter, E. and E.M. Beaney, “Recent Developments Residual Stresses.” Strain, Journal of BSSM 10, No. 1:
i n C entre -hole Te ch n ique for Re sidu a l-stre s s 7-14 (1974).
Measurement.” Experimental Techniques 6: 10-15 24. Nawwar, A.M., K. McLachlan, and J. Shewchuk, “A
(December 1982). Modified Hole-Drilling Technique for Determining
11. Wang, H.C., “The Alignment Error of the Hole-Drilling Residual Stresses in Thin Plates.” Experimental
Method.” Experimental Mechanics 17: 23-27 (1979). Mechanics 16: 226-232 (June 1976).

12. Prevey, P.S., “Residual Stress Distributions Produced 25. Witt, F., F. Lee, and W. Rider, “A Comparison of
by Strain Gage Surface Preparation.” Proc., 1986 Residual Stress Measurements Using Blind-hole,
SEM Conference on Experimental Mechanics (1986). Abrasive-jet and Trepan-ring Methods.” Experimental
Techniques 7: 41-45 (February 1983).
13. Flaman, M.T., “Brief Investigation of Induced Drilling
Stresses in the Center-hole Method of Residual-stress 26. Schajer, G.S., “Judgment of Residual Stress Field
Measurement.” Experimental Mechanics 22: 26-30 Uniformity when Using the Hole-Drilling Method,”
(January 1982). Proceedings of the Inter­n ational Conference on
Residual Stresses II, Nancy, France. November 23-25,
14. Bush, A.J. and F.J. Kromer, “Simplification of the Hole- 1988, 71-77.
drilling Method of Residual Stress Measurements.”
Trans., ISA 112, No. 3: 249-260 (1973). 27. Flaman, M.T. and J.A. Herring, “SEM/ASTM Round-
Robin Residual-Stress-Measurement Study — Phase
15. Bynum, J.E., “Modifications to the Hole-drilling Tech- 1, 304 Stainless-Steel Specimen,” Experimental
nique of Measuring Residual Stresses for Improved Techniques, 10, No. 5: 23-25.
Ac cur-acy and Reproducibility.” Exper i mental
Mechanics 21: 21-33 (January 1981). 28. Yavelak, J.J. (compiler), “Bulk-Zero Stress Standard
— AISI 1018 Carbon-Steel Specimens, Round Robin
16. Wnuk, S.P.. “Residual Stress Measurements in the Field Phase 1,” Experimental Techniques, 9, No. 4: 38-41
Using the Airbrasive Hole Drilling Method.” Presented (1985).
at the Technical Committee for Strain Gages, Spring
Meeting of SESA, Dearborn, Michigan, June, 1981. 29. Schajer, G.S., “Strain Data Averaging for the Hole-
Drilling Method.” Experimental Techniques. Vol. 15,
17. Delameter, W.R. and T.C. Mamaros, “Measurement of No. 2, pp. 25-28, 1991.
Residual Stresses by the Hole-drilling Method.” Sandia
National Laboratories Report SAND-77-8006 (1977), 30. Schajer, G.S. and E. Altus, “Stress Calculation Error
27 pp. (NTIS). Analysis for Incremental Hole-Drilling Residual Stress
Measurements.” Journal of Engineering Materials and
Tech Not

18. Flaman, M.T. and B.H. Manning, “Determination Technology. Vol. 118, No. 1, pp. 120-126, 1996.
of Residual Stress Variation with Depth by the Hole-
Drilling Method.” Experimental Mechanics 25: 205- 31. Schajer, G.S., “Use of Displacement Data to Calculate
207 (1985). Strain Gauge Response in Non-Uniform Strain Fields.”
Strain. Vol. 29, No. 1, pp. 9-13, 1993.
19. Niku-Lari, A.J. Lu and J.F. Flavenot, “Measurement of
Residual Stress Distribution by the Incremental Hole- 32. Schajer, G.S. and Tootoonian, M., “A New Rosette
Drilling Method.” Experimental Mechanics 25: 175-185 Design for More Reliable Hole-drilling Residual Stress
(1985). Measurements.” Experimental Mechanics. Vol. 37, No.
3, pp. 299-306, 1997.