2019

CONTACT
FORMULATIONS – PART 1

CONTACT FORMULATIONS AND TYPES OF CONTACT IN ANSYS- PART 1

SKILL-LYNC | support@skill-lync.com
Contents
INTRODUCTION: ................................................................................................................................ 1
TYPES OF CONTACTS IN ANSYS: ..................................................................................................... 1
CONTACT FORMULATIONS: ........................................................................................................... 2
Pure Penalty: ............................................................................................................................. 3
Augmented Lagrange: ................................................................................................................ 3
Normal Lagrange: ...................................................................................................................... 3
Multi Point Constraint(MPC) Formulation: ................................................................................. 4
Beam Contact Formulation: ....................................................................................................... 4
CONTACT AND TARGET BODIES: .................................................................................................... 5
TRIM CONTACT:............................................................................................................................. 6

INTRODUCTION:
Contact is said to occur when the surfaces of two separate objects touch each
other such that the surfaces are mutually tangent to each other. The surfaces in contact do not
interpenetrate and can transmit forces from one body to the other. In order to prevent the
interpenetration between the surfaces, the ANSYS program or any FEA software has to
establish relations between the surfaces and this is done using CONTACT
FORMULATIONS. When penetration is prevented by the program, it is said to have
enforced Contact Compatibility. When the program fails to enforce contact compatibility
between two surfaces in contact, interpenetration occurs to a large extent and the solver will
face convergence issues and be forced to stop.

TYPES OF CONTACTS IN ANSYS:

The types of contacts that are present in ANSYS Mechanical are:

1) Bonded: As the name suggests, it allows the user to define a contacting surface like a
glued surface. This type of contact does not allow any sliding or separation to take
place.
2) No Separation: This type of contact allows only sliding to take place. The contact
area will not separate.

1|P ag e
 3) Frictionless: In this type of contact, separation and sliding are both allowed. The
coefficient of friction is taken as zero in this case. The user should check the model if
it is properly constrained whenever this contact model is used. Weak springs are often
added for easier convergence with this model.
4) Rough: This is similar to the frictionless type of contact except that the coefficient of
friction in this case is taken as infinite.
5) Frictional: We can specify any non-negative value as the frictional coefficient. So,
the contact will handle shear stresses up to what is specified by the frictional
coefficient. After this value, sliding will start to take place.

CONTACT FORMULATIONS:

“Formulation” is one of the Advanced setting in ANSYS Structural used to define the
behaviour of the contacts that we specify. There are different algorithms that are used to
specify the type of contact pair computation. Depending on the type of algorithm used, the
system can behave differently. There are 5 algorithms used in ANSYS but 6 settings to
choose from. The last one is a “Program Controlled” option where the ANSYS solver
chooses the algorithm.
The different algorithms used are:
1. Pure Penalty
2. Normal Lagrange
3. Augmented Lagrange
4. MPC( Multi Point Constraints)
5. Beam

Fig 1: Formulation methods available in ANSYS

Penalty methods are algorithms that are used to solve constrained optimization
problems. It basically replaces the initial constrained problem with a series of unconstrained
optimization problems. A “Penalty Function” is defined out of the constraints on the
optimization problem and this function is then added to the function to make it an
unconstrained optimization problem.
Physically, two bodies do not interpenetrate upon contact. The solver uses different
algorithms between surfaces so that they do not pass through each other. The different
algorithms are mentioned below:

2|P ag e
Pure Penalty:
In this Formulation, the Pure Penalty method for solving constrained optimization problem is
used to solve the relationship between two surfaces.
𝑭𝒏𝒐𝒓𝒎𝒂𝒍 = 𝒌𝒏𝒐𝒓𝒎𝒂𝒍 × 𝒙𝒑𝒆𝒏𝒆𝒕𝒓𝒂𝒕𝒊𝒐𝒏

The Fnormal is the contact force and it is a function of the contact stiffness. The higher
the contact stiffness, the lower is the xpenetration. In an ideal situation, in order to get no
penetration, there should be infinite stiffness but it is not possible while solving the Penalty
based methods. So, as long as the value of xpenetration is very small, the results will be accurate.
The Pure Penalty formulation is more sensitive to the contact stiffness chosen by the
solver. As a result, the penetration between bodies is present and is uncontrolled. This
formulation usually shows good convergence behaviour with a few iterations itself. Pure
penalty formulation can be used with any type of contact.
Augmented Lagrange:
This method is also a penalty-based method. It uses the same equation for solving but has an
extra augmentation constant, λ. Because there is an extra term in the equation, the contact
force becomes less sensitive to the stiffness. So, there will be lesser penetration at the
contacts. Hence, this formulation is preferred more.
𝑭𝒏𝒐𝒓𝒎𝒂𝒍 = 𝒌𝒏𝒐𝒓𝒎𝒂𝒍 × 𝒙𝒑𝒆𝒏𝒆𝒕𝒓𝒂𝒕𝒊𝒐𝒏 + 𝝀
If penetration present is large, the solver will require additional iterations to converge.
Contact penetration is allowed in this formulation but it is more controlled than it was
in pure penalty method. This method can be used with any type of contact

Normal Lagrange:
This method is not a penalty based method. In this method, the contact stiffness is not
considered here. Instead of the contact stiffness, an additional contact pressure that acts
normal to the surfaces in contact is used. Since there is no contact stiffness in this method,
almost zero penetration can be achieved here. But using this formulation will often lead to an
increase in the computational time.

Normal Lagrange method has an inherent problem. It allows what is called as

“Chattering” to take place.

3|P ag e
 Normal Penalty and
Lagrange Augmented Open
Open Lagrange

Penetration Gap Penetration Gap

Closed Closed

Refer the image above, it can help you understand what takes place. In Normal Lagrange
formulation (Which is shown to the left of the image), at a particular area during a particular
time, the contact points can have both an open and a closed behaviour at a particular time.
You can see that it behaves like a step function here. At that particular time, the contact will
oscillate between open and closed status and this can sometimes lead to convergence issues.
On the other hand, Penalty and Augmented Lagrange allows a slight penetration when the
contact status changes.
Multi Point Constraint(MPC) Formulation:
This method is only available when you use Bonded or No Separation type of contact.
Without using any penalty or Lagrange based method, this formulation directly relates the
displacement between two nodes in a contacting interface which is bonded. This method
provides excellent convergence behaviour.
Beam Contact Formulation:
This method is only available to use when we define a bonded contact. It can be used to
reduce computational time whenever we want to find out beam mechanics. Bending of a
beam can be accurately captured in a 1D simulation. As you can see in the image below, you
can model a 3D beam as a 1D line. If you use any other contact formulation other than Beam,
you will encounter an error or get an inaccurate result.

4|P ag e
In conclusion, Pure penalty, Augmented Lagrange and Normal Lagrange can be used with
any type of contact. Out of these three, Augmented Lagrange is generally preferred because
it leads to better convergence out of the three and can be applied to any type of contact
available in ANSYS mechanical. MPC and Beam contact formulations cannot be used when
there is relative motion between the interfaces and beam contact is restricted to be used only
for bonded type of contact.

CONTACT AND TARGET BODIES:

In an interface, we define the contact and target. While deciding between these two,
we should keep in mind that ANSYS considers the contact body as the one which is not
allowed to penetrate into the target body. But the target is allowed to penetrate into the

5|P ag e
contact body. If you are unsure of the behaviour of your system, then you can choose
Symmetric type of contact. Otherwise, an Asymmetric type of contact can be used.

ANSYS Mechanical provides three types of behaviour for contacts and they are:
1) Asymmetric: The user has to choose the contact and target body while defining the
type of contact. If one of the contacting surface has a finer mesh or is a smaller body
than the other, this will be chosen as the contact and the other surface with a coarser
mesh will be the target. This is a general rule of thumb.
2) Symmetric: In symmetric type of behaviour, both the contacting surfaces are taken as
contact and target. This behaviour sometimes aids the convergence.
3) Auto Asymmetric: In this type of behaviour, ANSYS chooses the contacts and
targets. Essentially, ANSYS automatically tries to create an Asymmetric behaviour
between the contacts but you are free to edit the contacts and targets.

TRIM CONTACT:
This setting speeds up the solution time by “trimming” off the number of contact
elements sent to the solver while calculating the solution. You get three options while
selecting Trim Contact.
1) Program Controlled: With this setting, Trim contact is usually set to “On” condition.
If there are manual contact conditions that were used, then trim option will be set to
“Off”.
2) Off: Trimming of contacts will not be performed.
3) On: If Trim contact is set to ON, then the user has to specify a “Trim Tolerance”.
This allows the solver to choose only elements from both the contact and target side
and ignore the rest. Trim tolerance will take the nodes within a radius of the value
specified by the user. Trim contact can reduce computation time by around 15% in
most cases.

6|P ag e