JMEPEG (1993) 2:265-270 9 International

Changes in Mechanical Properties

in Metal-Forming Processes
J. Datsko and W.J. Mitchell

This article presents a paradigm for the mechanical properties of a metal that result from the plastic de-
formation that occurs in forming processes, including multiple-step or cyclic processes. Some of the new
concepts included are a new strength designation that includes the sense (tensile or compressive), the di-
rection (longitudinal, transverse, etc.), the specific type of deformation (tensile, compressive, etc.); a con-
cept of an equivalent strain; and the apparent rules of strain strengthening. Examples of strength
analyses are also included.

they have a high mechanical reliability along with minimum

Keywords strength [
equivalent strain, plastic deformation, strain strengthening, cost or weight, it is necessary for the engineer to know the sense
and direction of the strength in the part in addition to its magni-
tude.
Figure 1 is an example of the need for a strength designation
1. Introduction
that includes the sense of the strength, the direction of the
strength, and the type of previous plastic deformation. This fig-
THE tedious and time-consuming work of the stress analyst of ure summarizes one of hundreds of tests on a variety of ferrous
the past century is rapidly being replaced by very efficient, user and nonferrous metals and a variety of deformation processes
friendly finite-element modeling software programs. The com- performed by the author and his students.
puter displays of these programs show not only the magnitude Figure 1(a) represents a 2-in. cube of annealed brass (No.
of the stress at any point in the part being analyzed, but they 260), from which three tensile and three compression speci-
also indicate the sense (tensile or compressive) of the stress and
the direction in which it acts.
Unfortunately, the mechanical reliability of the fabricated
components is not solely dependent on the accuracy of the
stress analysis. Of equal importance is the accuracy of the de-
termination of the strength of the material at the locations of the
calculated stresses. In other words, the reliability of the part de-
pends on the accuracy of the strength-to-stress ratio. Although
the stress analysis may be accurate to six significant figures, the
strength determination may be very unreliable. In fact, it is pos-
sible for the value listed in some materials handbooks for a
given alloy to be in error by 100%. The author has in his files
several examples of errors of this magnitude. [ ~ Z..O00"
One of the main reasons for the presence of such large errors (a) Co)
in published values of strength is the poor documentation of the Spee.Yield strenglh
No. (Mlh)
particular strength in question. Whereas the stress is specified Tens. C o m p r .
with respect to magnitude, sense, and direction, the strength is
"~....... "~i-;-'"-'-----
specified only in terms of its magnitude. Although this may be 2 --- 74.5
3 76.5 ---
sufficient for noncold worked metals, it is not satisfactory for 4 - - 80.7
5 77.9 ~-
plastically deformed metals. A metal that has been cold worked
has values of strength that are different in the sense of the
7 170 ---
strength (tensile or compressive), the direction within the part, 8 .... 316
9 272 ---
and the sense of the deforming strain during the fabrication of 10 .... 245
11 263 ---
the part.
Formed parts are being used more extensively each year in
12 .... 239
_t_
the manufacture of automobiles, airplanes, home appliances,
and all other mechanical devices. To design such parts so that (c)
Fig. 1 Directional strength in a cold worked part. (a) Tensile
J. Datsko, Emeritus Professor, The University of Michigan, Ann Ar- and compressive tests on an annealed block of alloy No. 260
bor, Michigan, and W.J. Mitchell, Assistant Professor, The University brass. (b) Second block identical to (a). (c) Block (b) after 19%
of Michigan, Dearborn, Michigan. cold work by compression in the Z-direction.

Journal of Materials Engineering and Performance Volume 2(2) April 1993--265

mens were tested, one in each of the three principal directions. After After tension
Original
The tensile and compressive yield strengths of all six were ap- bar compression
proximately the same--slightly more than 76 MPa (11 ksi).
Figure l(b) represents an identical annealed cube that was
subsequently compressed (19% cold work) in the Z-direction to
the size shown in Fig. 1(c). Three tensile and three compression
specimens were machined from the deformed block in the same
directions as those from the annealed cube. Note the great
range of values for the 0.2% offset yield strengths, The com-
pressive yield strength in the Z-direction is nearly 100% greater
(a)
than the value of the tensile yield in that direction, and the yield Sense of Direction Prior
Spec:Lme.~. Strength in the bar strain Desisnation
strengths in the transverse direction are intermediate to those in
1 t L c (s-hLc
the longitudinal direction. This variation in strength is not an 2 c L c (S-) ~
3 t fS_) cLc
exception. It occurs with all metals to some degree. It is most 4 c T t
t r ~tTt
" -'cTt
severe with many stainless steels, copper alloys, cobalt alloys,
5 t L t (S-)
etc. Even the ordinary steels exhibit a significant variation. 6 c L t ~( S_~ tLt
7 t c i S - itTc~cLt
The materials handbooks list only one value of yield 8 e
T
T c - cTc
strength for each amount of cold work. For example, the Re-
vere Copper and Brass booklet lists the yield strength of alloy
No. 260 as 345 MPa, with 20% cold work. It does not document
whether the value is for the tensile or the compressive yield, nor
does it indicate the orientation within the part, or the type of de-
formation to which the material was subjected. Designs based
on such poorly specified data can certainly produce unreliable
(b)
parts that result in unexplained failures. As Fig. 1 shows, the ac- Sense of Direction Prior
tual tensile yield strength in one direction can be as low as 50% Specimen Strength :ha the bar strain DesiEnation
of the value listed in the handbooks. i t L t (S-) .

3 t D e S cLt
4 c D c (S-) r~c
C-D~
S t s o (s-)
0 o ~ o CSm)~r~
c~o
2. A New Strength Designation
Fig. 2 Strength designation after (a) uniaxial deformation and
(b) rolling of a plate.
To avoid this great confusion and unreliability of mechani-
cal property data, it is necessary to have a clear and precise
method of designating the strength of a plastically deformed
metal. The author created[l[ such a designation in 1966. During men whose axis is parallel to the direction of the principal de-
the ensuing years, he has found it to be extremely beneficial and forming strain. For symmetrical cross sections such as rounds
reliable. It is now included in the new Standard Handbook of or squares, it may be a T for transverse; an R for radial; or a C
Mechanical Design published by McGraw Hill. for circumferential. These latter three are all equivalent. In the
The symbol used to denote strength is the capital letter S. case of plates or rectangular bars, two transverse directions
This distinguishes it from stress, for which the Greek letter must be specified: the long transverse can be indicated by a B
sigma (o) is commonly used. The S must be followed by four (breadth); the short transverse can be designated with an H
subscripts to identify the (1) type of strength, (2) sense of the (thickness).
strength, (3) direction of the strength, and (4) sense of the last The fourth subscript refers to the sense of the last prior strain
plastic deformation strain. It is extremely beneficial to the read- in the axial direction of the test specimen, but not in the direc-
ability of this notation to place parenthesis in front of the letter tion of the deforming force. The notation here is the same as for
S and after the first subscript. A typical example of this notation the second subscript; that is, t, for tension; c, for compression;
is (Su)tL~ which represents the ultimate tensile strength in the or s, for shear. However, if there was no prior strain, then the
longitudinal direction after prior compression in that direction. subscript o should be used. For a non-cold-worked or heat
The first subscript specifies the type of strength under inves- treated isotropic material, the second, third, and fourth sub-
tigation, which may be any of the following: y, for yield; u, for scripts can be deleted.
ultimate;f, for fracture; e, for endurance limit; and p, for pro- To generalize, the strength of a cold worked material can be
portional limit. represented as (S-) w and that of the original non-cold-worked
The second subscript refers to the sense of the strength, material as (S-) o.
which may be any of the following: t, for tensile; c, for com- A second example is given to illustrate this designation.
pression; and s, for shear. Consider the notation (Sy)cTc, which means the compressive
The third subscript indicates the direction or orientation of yield strength in the transverse direction after compressive
the test specimen within the formed part. For clarity, capital let- strain in the transverse direction (equivalent to an extension in
ters are used for the third subscript. It may be an L for a speci- the longitudinal direction).

266---Volume 2(2) April 1993 Journal of Materials Engineering and Performance

 Four strength designations, as illustrated in Fig. 2, are value of strain during cycle No. i, disregarding the sense ( - or
needed to completely specify the strength in bar or plate that is +) or direction of the strain; n is the total number of cycles of
upset (compressed in the longitudinal direction). The tensile strain; a = 0 for the strengths that have the same sense as the last
and compressive strengths in the longitudinal direction are de- prestrain. That is, when the subscripts 2 and 4 in the strength
noted as (S-)tLc and (S-)cLc, respectively. The dash can be re- designation have the same sense, or when the 4th subscript is
placed by any of the symbols discussed above that can be used zero; and a = l for strengths having the opposite sense at the last
for the first subscript. The tensile and compressive strengths in prestrain. That is, when the subscripts 2 and 4 in the strength
the transverse direction are expressed as (S-)tT t and (S-)cT t, re- designation are of the opposite sense.
spectively. In this case, the fourth subscript t indicates that the The value of ew for any one deformation step that should be
prior strain in the transverse direction was tensile during the used to calculate equ is the largest of the three principal strains.
compressive deformation. A cycle of strain occurs whenever there is a reversal or change
Four strength designations are also needed for a bar that has of the sense of the strain in any one of the three principal direc-
been axially (longitudinally) stretched. In the longitudinal tions.
direction, the tensile and compressive strengths are (S-)tLt and
(S-)cLp respectively. Similarly, in the transverse direction, they
are (S-)tTc and (S-)cTc. Again, the last subscript c indicates that
the prior strain in the transverse direction was compressive. 4. The Apparent Rules of Strain Strengthening
In the case of rolling or bending of a plate (width >8 • thick-
ness), six designations are necessary. There are both tensile and The term "apparent rules" is used because they portray what
compressive strengths in each of the three perpendicular direc- the strength of a cold worked material appears to be on the basis
tions: longitudinal, L; long transverse, B; and short transverse, of a broad range of experimental studies. They have been de-
H. Thus, the six designations are (S-)tLt and (S-)cL{, (S-)tHc and veloped by the author, with the help of students, during the past
(S-)cH~ (S-)tBo, and (S-)cBo. The last subscripts are t, for the 25 years, while attempting to present to engineering students,
longitudinal direction because the length increases; c, for in a systematic manner, the results of many individual research
the short transverse because the thickness is reduced; and o, projects in the fields of plastic deformation and mechanical
for the long transverse because the width remains constant. properties. Included among these projects are cyclic axial de-
When the desired strength is at some angle other than 0 ~ or formation of cylinders, cyclic deformation of large cubes in the
90 ~ that is other than longitudinal or transverse, then a simple three perpendicular directions, bending and unbending of flat
linear interpolation of the strength on the basis of the 0 ~ and 90 ~ specimens, cyclic torsional deformation of cylinders, shearing
values is required. Thus, if a flat bar is uniaxially deformed in of blanks and strips, deep drawing of channel sections and cy-
tension and if the tensile yield strength at an angle of 30 ~ to the lindrical cups, wire drawing, forward and back extrusion, and
longitudinal direction is desired, then the strength is calculated cold rolling. A variety of tensile, compressive, fatigue, and
as follows: hardness tests were conducted on the cold worked metals.
(Sy) t = 2/3(Sy)tL t + 1/3(Sy)tTc The strength in any direction of a cold worked part is calcu-
Likewise, the compressive yield strength is calculated as lated by integrating the equivalent strain into the plastic mono-
follows: tonic tensile stress-strain equation. For purposes of clarity and
(Sy) c = 2/3(Sy)cL t + 1/3(Sy)cTc preciseness in presenting this subject to engineering students,
The techniques for calculating the numerical values for the exponential strain-strengthening equation is expressed as:
these strengths are presented in the following section. 13 = 13oEm

where m is the strain strengthening exponent, and 13o is the

3. The Equivalent Strain stress coefficient.
The values of strength calculated by means of the following
To calculate the strength of a metal that has been subjected rules agree with experimentally determined values in most
to one or more cycles of plastic deformation, it is necessary to cases to an accuracy of less than 5% and in a few cases to less
know the value of the equivalent strain. In this context, the than 10%. This is true for metals that have been subjected to as
equivalent strain is the value of the accumulated strain during many as five cycles of strain and where the yield strength has
the deformation processes, which when used in conjunction been increased from a value of 8 ksi prior to cold work to 524
with the monotonic plastic tensile stress-strain properties en- MPa after cold work. The calculations made on the basis of
ables the multidirectional strength of the material to be calcu- these rules show the effect of cold work alone on the strength of
lated. the material. They are valid when no other strengthening
Borden [2] developed a reliable model of the equivalent mechanism such as precipitation or recrystallization occurs af-
strain equ, which is defined by the equation: ter the cold working process.
n

4.1 Rule 1
equ = ~ a+l
i=1 Strain strengthening is a bulk mechanism. Even a deforma-
tion load that is applied in only one direction causes deforma-
where i is the order number of a strain cycle when the strains are tion and strengthening in all directions. For example, in the
listed in order of decreasing magnitude; Ei is the maximum cold rolling of a plate, even though there is no increase in the

Journal of Materials Engineering and Performance Volume 2(2) April 1993--267

breadth of the plate, both the tensile and compressive yield Table 1 Equations for calculating yield strength
strengths in the breadth direction are increased by 10 to 100%,
depending on the amount of cold work. Group No. Yield strengths Equation
1 ................................. (Sy)tLt(Sy)cLc = f equs ~m
(Sy)tBo (Sy)cBo (30
4.2 Rule 2 (Sy)cnc L" I
1 + 0.2equs)
Equs Im
The maximum deformation that can be given a material dur-
2 ................................. (Sy)tTt (Sy)cTc = ~o 1 + 0.5equs)
ing the forming of a part is the deformation that induces a ten-
sile strain that is numerically equal to the fracture strain of that
material when tested in the same direction and identical envi- 3 ................................. (Sy)tTc (Sy)cTt = 0.90(Sy)tTt
ronment. In other words, fracture will occur when the induced (+Eq~~ ~
4 ................................. (Sy)cLt (Sy)t~ = Crol 1
tensile strain Et is equal to the fracture strain ep This has pre- (Sy)tHc
viously been referred to by the author as the Failure Theory for
Plastic Working. [3] However, from a practical approach, a few
forming operations are terminated when the induced tensile
strain is equal to the ultimate load strain, which is also numeri- The value of e_ u that should be used in the above two equa-
cally equal to the strain-strengthening exponent. This is true for tions is equ s (a = ~) if the last deforming strain in the direction
those forming operations in which the condition known as of the desired strength was tensile. If the last deforming strain
necking occurs. in the direction of the desired strength was compressive, then
Some operations are a combination of stretching and bend- equo(a = 1) must be used.
ing, and in these cases, the failure occurs at some value of strain
between eu and ej~ depending on the ratio of bending to stretch- 4.6 Rule 6
mg strata.
The yield strength of a cold worked metal is determined by
means of the equation:
4.3 Rule 3 (Sy) w = O o(Eqy)m
The numerical value of ~w in any one deformation cycle to where eq.r is the equivalent yield strain as defined in Table 1.
be used in calculating the equivalent strain e q-. is the largest of The yield strengths are divided into four groups as shown in
the three strains present at any location. It is usually the longi- Table 1. For any given strain sequence, the numerical values of
tudinal or axial strain. For example, if a cube is compressed in all the yield strengths in any one group are equal. The largest
the Z-direction to a strain o f - 0 . 2 , the strains in the X and Y di- strengths are in Group 1, and the smallest are in Group 4.
rections are each +0.1. The numerical value of strain that is
used for ew in this case is 0.2, even when calculating the
strength in the transverse direction.
5. Examples of Strength Analysis
4.4 Rule 4
In determining the value of Equ, the values of Ewi are added The application of the above concepts and relationships can
in order of decreasing numerical value rather than in their best be illustrated by considering three typical strength analysis
chronological order, as defined in Table 1, and without regard problems. Consider first the problem in which a 2.5-in. diame-
to their sign. Thus, for a strain sequence of +0.10, -0.20, +0.15, ter head is cold upset on the end of a 2-in. diameter shaft prior
-0.05, the value of Equs (~qu for a = 0) is 0.20 + 0.15/2 + 0.10/3 to machining splines on the enlarged end. Also, to save some
+ 0.05/4 or 0.32. Th~ valfie of ~quo (E~u for a = 1) is 0.20/2 + machining, a cold drawn bar rather than an annealed one is se-
0.15/3 + 0.10/4 + 0.05/5 or 0.19. Thi~ method of adding the lected. Assume that in the cold drawing operation a 2.25-in. di-
strains gives the greatest effect to the largest strain regardless of ameter bar was reduced to the 2-in. diameter. The material
when it occurred in the actual strain sequence. selected is a type 304 stainless steel with a yield strength of 341
For only one cycle of strain, equs = ew and e~uo = 0.5 ew" MPa, a tensile strength of 593 MPa, o o = 1379 MPa, m = 0.50,
As stated previously when the concept o~the equivalent and ef= 1.7.
strain was presented, the strength of the cold worked metal is Because the force acting on the splines is in the circumferen-
calculated by combining the equivalent strain with the plastic tial, or transverse, direction it is necessary to know both the ten-
stress-strain relationship of the metal, as explained in rules 5 sile and compressive strengths in that direction. The first cold
and 6. drawing strain is calculated as:
E1 = 2 In 2.25/2.00 = 0.24
4.5 Rule 5
The upsetting strain is:
The tensile strength of a cold worked metal is determined by E1 = 2 In 2.00/2.50 = -0.45
means of the two equations:
According to the notations and equations in Table 1, the
(Su) w = (Su)oeequ for Equ < m
equivalent strains are calculated as:
(Su) w = o w = O o(Equ )rn for Equ >_m equs = 0.45 + 0.24/2 = 0.57

268---Volume 2(2) April 1993 Journal of Materials Engineering and Performance

 By referring to Fig. 3, it is apparent that the outside surface
of the leg is the inner fiber during the first bending of the leg.
Therefore, the bending strain ~1 = In [(0.125 +0)/(0.125 +
o~ ~ J~-~7 ~ o ~ SPRING 0.0625)] = -0.405. As the drawing operation continues, the
bent metal is unbent with an equal tensile strain of 0.405 be-
cause the finished leg is straight.
:. , - The tensile yield strength in the long transverse direction
during bending is designated (Sy)tBo because there is no strain
in the 30-in. direction. The value of equ s is 0.405 + 0.405/2 =
\\ DIE DIE 0.608. The tensile yield strength at the outside surface in the 30-
in. direction can be calculated as follows:
(Sy)tBo = 200 [0.608/(l + 0.2 x 0.608)]0"30= 166 MPa. This
V///////// // / / / is an increase of 300% over the original yield strength.
The calculation of the yield strength at the midthickness is
Fig. 3 Draw die to a hat (channel) section. slightly more complicated. Although the neutral axis remains at
the midthickness, some metal originally below the midthick-
hess is displaced to a position above the midthickness. For ex-
ample, the metal that ends up at the midthickness (0.0625-in.)
equo = 0.24/2 + 2/24/3 = 0.30 was originally 0.0521-in. above the bottom surface of the flat
The transverse tensile yield strength is calculated as: strip. Also, when the instantaneous radius is 0.310-in., this par-
ticular element is subjected to a compressive strain of-0.014.
( Sy) tTt = t~o( s qys)m
Then as the radius is further reduced to 0.125-in., a tensile
where strain of 0.014 is induced in this element, which is now at the
midthickness with a final net strain of 0. All of this occurs dur-
s = 0.57(1 + 0.5 • 0.57) = 0.44 ing the bending of the flat strip to a 0.125-in. radius.
As the drawing operation continues, the bent metal is unbent
(Sy)tTt = 1379(0.44) 0.5 = 917 MPa because the leg, in its final shape, is straight. During the un-
The transverse compressive yield strength is calculated as: bending step, this same element of metal is again compressed to
a strain o f - 0 . 0 1 4 and finally stretched to a strain of 0.014 for a
(Sy)cTt = 0.90((Sy)tTt) = 0.90 x 917 = 827 MPa net total strain of 0. The final thickness is 0.125-in. However,
However, if the properties were measured experimentally in the metal has received a total of four cycles of strain, each equal
the longitudinal direction, they would more closely resemble to 0.014. The value of the equivalent strain is determined as fol-
the following calculated values: lows:
(Sy)cLc = 137910.57/(1 + 0.2 • 0.57)] 0.5 = 986 MPa equs = 0.014 + 0.014/2 + 0.014/3 + 0.014/4 = 0.029
The tensile strength in the transverse direction is:
which is approximately 3% cold work.
(Su)tTt = 1379(0.57)0"5= 1041 MPa Finally, the tensile yield strength at the midthickness in the
In the longitudinal direction, the tensile strength is: 30-in. direction can be calculated as follows:
(Su)tL c = 5930"30= 800 MPa
(Sy)tBo = 20010.029/1 + 0.2 x 0.029)0"30= 69 MPa
Next, consider a slightly more complicated strength analy-
sis problem. A channel section is made by deep drawing 1/8-in.
or nearly twice the original value.
thick by 6-in. wide by 30-in. long strips over a 1/8-in. die radius
As a final example, consider the experimental data shown in
(see Fig. 3). The finished channel has 2-in. high legs. The origi-
Fig. 1. The average monotonic tensile test data for the brass al-
nal material is 3003-0 aluminum having the following proper-
loy 260 used in that project are proportional limit, 67 MPa;
ties:(Sy)o = 41 MPa; (Sy)o = 103 MPa; t~o = 200 MPa; m = 0.30;
yield strength, 76 MPa; tensile strength, 303 MPa; strength co-
Ef= 1.50. efficient a, 758 MPa; m, 0.60; and fracture strain, 1.21. The
The question is, "What are the numerical values of the ten-
compressive strain for 19% axial cold work is E w = 0.21.
sile yield strength at the outside surface and at the midthickness
Based on the above original properties of alloy 260, the six
of the drawn legs in the direction of the 30-in. length?"
yield strengths of the compressed block can be calculated as
The calculations are based on the fact that the neutral axis
follows:
remains at the midthickness during pure bending. With the
For specimen No. 7:
proper hold-down pressure and good lubrication, thinning of
the metal during bending and unbending is negligible, and (Sy)tLc = 758(0.105/1.21)0"60= 167 MPa
therefore, the inner fiber compressive strain is considerably For specimen No. 8:
greater than the outer fiber tensile strain. (Sy)cLc = 758(0.21/1.042)0'60= 291 MPa
The strain at any distance Y from the neutral axis is calcu-
For specimens No. 9 and 11:
lated from the expression e = In [(R + Y)/(R + N)], where R is
the inside radius and N i s the distance the neutral axis lies above
(Sy)tTt = 758(0.21/1.105) 0.60 = 281 MPa
R. In this case, N = h/2. For specimens No. 10 and 12:

Journal of Materials Engineering and Performance Volume 2(2) April 1993---269

 (Sy)cTt = 0.90 x 281 = 253 MPa strength distribution in solid cylinders after cyclic torsional de-
The agreement between the calculated values and the ex- formation as well as other forming operations.
perimentally determined values is very good. The larger dis-
crepancies are with the compressive yield strengths, and this is References
mostly due to the difficulty in experimentally measuring the 1. J. Datsko, Material Properties and Manufacturing Processes,
compressive yield strength. John Wiley & Sons, 1966, p 342
These three examples illustrate the techniques of calculat- 2. M.P. Borden, "Multidirectional Tensile Properties of Materials
ing the strength of the material in parts made by the cold work- Subjected to Large Cyclic Strains," doctoral thesis, University of
ing processes, including cyclic straining. These rules of strain Michigan, 1975
strengthening have also been found to predict quite reliably the 3. J. Datsko, Material Properties and Manufacturing Processes,
John Wiley & Sons, 1966, p 315

270---Volume 2(2) April 1993 Journal of Materials Engineering and Performance