Anda di halaman 1dari 3

ANSYS ADVANTAGE | 2012

MODELING THREADED
BOLTED JOINTS IN
ANSYS WORKBENCH
Although bolted joints are extremely common, they can be difcult to model
accurately without using some best practices from an industry specialist.
B
olted joints are extremely com-
mon fasteners in construction
and machine design. However,
creating a fnite element
model (FEM) of a threaded
bolted joint is a complicated task but
well worth the efort. The steps are:
develop a solid model, create contacts,
mesh the model and then establish
boundary conditions. Each of these tasks
has its own set of challenges and issues;
these all impact the ability of the software
not only to converge, but to give accurate
results. The model is used to determine
not only what happens when torque is
applied and clamp load developed, but
what happens when external loads are
placed on the joint: Does the joint remain
tight, or is there failure? Here are some
suggestions for best practices.
Using a 3-D modeling approach is pre-
ferred for most applications that are
deemed critical or that involve tightening
against or into soft material (in this case,
aluminum). Torque must be applied to
the hex faces of the bolt to provide rota-
tion. This develops clamp load in the joint
and mimics reality. Avoid using simulated
nonthreaded fasteners (place keepers) or
pretension elements due to lack of rota-
tion. The rationale for both approaches
will become evident.
GEOMETRY CREATION
The authors practice is to create the 3-D
model (Figure 1) within ANSYS APDL
using a specifcally created .mac or macro
fle. (Alternatively, the geometry could be
created in a CAD tool or in ANSYS
DesignModeler.) This .mac fle is an inter-
active program that allows creation of a
joint of almost any thread size and pitch.
The joint could consist of a standard hex,
heavy hex or socket head bolts using a nut
or internal thread (all per applicable
industry specifcations). The program
allows adjustment of the ft for both the
internal and external threads within the
limits of geometric size per industry
standards. This is an important concept
from a manufacturing standpoint because
the threads have geometric size toler-
ances and allowances. Figure 2 shows the
maximum and minimum fank-to-fank
contact that can exist within the specifca-
tion limits. Create the fanks surface
areas as coincident, because a gap here
will likely cause a bisection in the frst
substep of the solve.
FINITE ELEMENT MODELING
Once the 3-D model has been created,
export it to SolidWorks design software
via IGES in which the individual volumes
of each part of the joint are combined. The
critical point in this operation is that
when the volumes are combined, the area
segments disappear to create one contact
surface instead of four. Once edited,
import the solid model into the ANSYS
Workbench environment directly using
the geometry interface module.
This model contains only three contact
zones: bolt underhead, threads and bot-
tom of the joint. Figure 3 shows a cross
section of the FEM of Figure 1, with its
three contact zones. Note that there are
contact and noncontact areas in the exter-
nal and internal threaded region. The con-
tact parameters for each zone are the
same (except for the frictional values), as
shown in Table 1. The values input for
friction are critical and determine whether
the analysis is correct. The values used in
By Michael Oliver, Oliver Testing and Consulting, Warner Robins, U.S.A.
Figure 1. 3-D model of threaded bolted
joint using standard hex bolt into internal
thread
BEST PRACTICES
Figure 2. Cross section of external and
internal threads. Threads shown are for
maximum fank-to-fank condition (top) and
minimum fank-to-fank condition (bottom)
allowed per industry specifcation.
ANSYS.COM 2012 ANSYS, INC. ANSYS ADVANTAGE
Figure 5. Meshed model from Figure 1 Figure 4. Cross section of internal thread
showing developed mesh. The fanks on
the lower surface are loaded. The bolt
head is up.
this example were determined from
experiments performed in a fastener test
lab. (See more below.) Tabular values
from industrial specifcations or hand-
books will work, but be aware that friction
in the underhead region and threads are
rarely the same. If experimental data is
not available, you can establish a range of
values and execute multiple solves. In
reality, changes made to the underhead
region will have a bigger impact on clamp
load development than that of the
threaded region.
MODELING
Consistency of size and shape of mesh ele-
ments on the thread fanks is important
for evaluation of load and stress distribu-
tions. Figure 4 shows the meshed internal
threads of the model from within
Workbench. A fairly dense mesh on the
contact fanks and root is recommended,
then a less-dense mesh on the noncontact
fanks. The top fank of the frst internal
thread requires a fne mesh as well,
because this thread is not complete and
can actually plastically deform. The mesh
on the bearing surface is controlled
through contact sizing; in this example,
element sizing was used on the threads of
the bolt. Figure 5 shows the entire meshed
model. Default settings were used with all
mesh sizing defnitions with a relevance
set at zero. The analysis settings are
shown in Table 2 for this static structural
solution. This example model had
365,427 elements (88,178 contact ele-
ments) and 607,233 nodes.
Table 1. Connection and Contact Settings

Connections Auto detection No
Tolerance slider 0
Tolerance value 0.15 mm
Face/face Yes
Face/edge Yes
Edge/edge Yes

Contacts Type Frictional
Frictional coefcient 0.059 / 0.147/ 0.1
Scope method Manual
Behavior Auto symmetric
Advanced Formulation Augmented Lagrange
Interface treatment Add ofset, no ramping
Ofset 0 mm
Normal stifness Manual
Normal stifness factor 0.1
Update stifness Each iteration
Pinball region Radius
Pin radius 0.1 mm
Under-head
contact regions
Bottom of joint
contact regions
Threaded
contact region
No thread contact
Bolt and nut
thread contact
A
Figure 3. Cross section of threaded joint
model showing three contact regions.
Area A shows a magnifed view of the
threaded contact region.
Table 2. Analysis Settings
Step Controls Number of steps 1
Step end time 1
Scope method 1
Auto time stepping On
Defne by Substeps
Initial substeps 30
Minimum substeps 1
Maximum substeps 100
Advanced Solver type Direct
Weak springs Program controlled
Large defection On
Nonlinear Controls Force convergence On
- Value Calculated by solver
- Tolerance 0.5 percent
- Minimum reference 0.01 N
ANSYS ADVANTAGE | 2012
Figure 8. Deformation in Z direction for
top of internal threaded region
BEST PRACTICES
This model used experimental data:
thread torque, clamp load and total torque
(torque in) with the aid of a torque/ten-
sion load cell and a torque transducer.
Underhead torque as well as thread and
underhead coefcients of friction were
calculated using the following equations
(which can either be solved or obtained
from any good mechanical engineering
reference book).
The material for the joint and internal
threads was A356 aluminum. Thread size
was M10 x 1.5 6g6g for the external
threads and M10 x 1.5 6H6H for the inter-
nal threads. The measured clamp load
value of the actual joint, at 40 N-m of
torque, was 27,105 N, while the reaction
force from the simulation was clampload
value of 27,367 N. This represents a 0.1
percent error.
Figure 6. Equivalent stress on the joint
bearing surface of the torqued solve.
Maximum stress is 192.12 MPa.
The above model was also used in
a comparison study with pretension
elements (conducted in Workbench)
applied to the shank of the bolt, using
the clampload value of 27,105 N. Figures
6 and 7 show the bearing surface from
both the torqued and pretension solu-
tions, respectively. Note the magnitude of
the pressure on both solves (A356 yields
at 179 MPa). The rationale for the higher
Figure 7. Equivalent stress on joint
bearing surface of pretension solve.
Maximum stress is 177 MPa.
pressure on the rotational model is obvi-
ously because of rotational shear forces.
Figure 8 shows the top of the internal
threaded region from the rotational solve.
The high amount of deformation is a
direct result of the partial thread that
exists. The amount of deformation was
measured in the lab, matching that of
the model.
Where:
d
p
= Pitch diameter Pitch = Thread pitch
D
km
= Average of dw and dh T
in
= Total torque or torque in
d
w
= External fange diameter of T
thd
= Thread torque
bolt or nut T
und

= Under head torque
d
k
= Joint clearance hole diameter
thd

= Thread coefcient of friction
P = Clamp load
und
= Under head coefcient of friction
SUMMARY
The goal of simulation is to simulate the
part or assembly under actual operating
conditions. But even with a 3-D model,
results can be inaccurate and misleading
if you do not follow best practices. This
may not be an issue with steel on steel,
but if the joint has soft material, or
even a gasket within it, a rotating 3-D
model is a must, and frictional values
must be carefully considered.
REFERENCE
moliverconsulting.com
The author thanks Mai Doan from
ANSYS for help in transitioning the .mac
fle from solving in ANSYS Parametric
Design Language to solving in the
ANSYS Workbench platform.

Anda mungkin juga menyukai