Overview#
When a simulation includes more than one body, the solver has to know how they interact. Contacts define that relationship. Without a defined contact, Autodesk notes, the bodies could pass through each other during the simulation. With the wrong contact, load travels through paths that do not exist, or fails to travel through paths that do.
Fusion generates automatic contacts during the solve (and when you generate the mesh), using the Bonded contact type. Automatic contacts are a convenience, not a guarantee; always review them. Constraints do not connect two parts: all loads go through the constraints into the ground rather than being passed to the next part, so you must define contact between parts.
Where to find it#
- Setup tab > Contacts panel: Automatic Contacts (sets the Contact Detection Tolerance and generates), Manual Contacts, and Contact Manager. You can also open the Contacts Manager from the Contacts node in the browser using Edit.
- Event simulation studies use Global Contacts and Manage Contacts in the Contacts panel instead.
- Contacts are not applicable to the Electronics Cooling study.
- Which contact types are offered depends on the study type (see the table).
Step by step#
- Create your study and make sure all bodies of interest are included.
- Click Automatic Contacts, set the Contact Detection Tolerance, and click Generate. Autodesk generates a contact set where the distance between entities does not exceed the tolerance and the entities are parallel within a fixed 15 degree tolerance. If two separate bodies start farther apart than the tolerance, no automatic contact is created. Automatic contact sets use Contact Type Bonded and Penetration Type Symmetric.
- Open the Contact Manager. Choose List by Contact Set (default) or List by Bodies. Select a row to see the participating bodies and entities. Use the drop-down lists in the Contact Type and Penetration Type cells to change a pair (you can multi-select rows and change them together). Manual sets are marked with [M].
- Right-click an entry and choose Edit Contact Set to change other parameters such as Maximum Activation Distance, Thermal Conductance or Coefficient of Static Friction. You can suppress, rename or isolate contact sets; you cannot delete automatic ones, only suppress them.
- For pairs automatic detection misses, use Manual Contacts: pick one Primary Body, one Secondary Body, the faces (or edges) on each, the Contact Type and the Penetration Type, then OK. For self-contact within one part, pick the same body as Primary and Secondary; self-contact must be defined manually.
- For bolted joints, add a Bolt Connector and then update contacts manually: the automatic contact is Bonded, but contact between bodies held by bolts should usually be Separation.
- Run Pre-check and solve. Use the Contact Pressure and Contact Force results (computed for all contact types) to check how load passes through the interface.
Options reference#
Contact types you can define manually (Autodesk lists six):
| Contact type | Behavior | Notes |
|---|---|---|
| Bonded | Bodies perfectly bonded where they meet; no relative movement for the whole analysis | Default for automatic contacts |
| Offset Bonded | Bodies not touching are bonded as if in direct contact | Set Max. Activation Distance large enough; recommended minimum is the gap plus about half the mesh size |
| Separation | No penetration; bodies can separate; free to slide tangentially when touching | Optional Coefficient of Static Friction; not available for Modal Frequencies |
| Sliding | No penetration and no separation; surfaces slide freely tangentially | Not available for Modal Frequencies |
| Rough | No penetration; can separate; cannot slide while in contact | Like Separation with very high friction; not available for Modal Frequencies |
| No Contact | Prevents contact interaction between selected entities | The nonlinear static stress topic calls it Free; automatic routines will not generate contact between them |
Contact types apply to all structural analyses except event simulation, which has its own global contact settings. Self-contact is not a type: choose a type and assign the same body as Primary and Secondary.
Other contact settings:
| Setting | What it does | Notes |
|---|---|---|
| Contact Detection Tolerance | Maximum distance for automatic contact detection | Set in the Automatic Contacts dialog |
| Penetration Type | Symmetric (default, prevents interference either way, slowest) or Unsymmetric (only nodes on the Secondary Body are prevented from penetrating the Primary Body) | Use Unsymmetric only if you are sure loads will not cause interference; choice of primary and secondary then matters |
| Automatic Max. Activation Distance / Max. Activation Distance | Distance within which contact is enforced | Recommended for Offset Bonded and Separation; if automatic is off and no value is entered, no contact elements form |
| Thermal Conductance | Heat flow per area per degree across the interface (thermal and thermal stress only) | Blank by default, meaning infinite conductance (no resistance) |
| Coefficient of Static Friction | Tangential force equals the coefficient times the normal force | Separation only; blank means frictionless |
| Advanced Options | Extra contact parameters based on Nastran contact | Need in-depth knowledge of Nastran contact |
Event simulation global contacts: Global Separation Contact is on by default with a Default Friction Coefficient; Proximity Bonded Contact is off by default (bodies within the Proximity Distance are bonded and override Global Separation); Allow Self Contact is on by default and its Stiffness Factor governs load transfer. In Manage Contacts, use Suppress All Self-Contacts and Suppress Contact to remove unneeded pairs and cut analysis time.
Tips and pitfalls#
Warning: Bonded contacts can transmit tension across a joint that in reality would separate. In a bolted flange under bending, bonded contact makes the joint behave like a single solid. If separation matters, use Separation or Rough contact (and consider a nonlinear study).
Tip: Reduce unnecessary contacts. Each pair adds work for the solver. Suppress contacts between parts that only appear to touch.
Tip: For thermal studies, a contact with no Thermal Conductance entered conducts heat perfectly. If the interface resistance matters, enter a conductance value from your own data.
Warning: If a body penetrates another after the analysis, Autodesk suggests increasing the Maximum Activation Distance or trying a nonlinear static stress analysis.
Warning: Overlapping bodies make contact behave unpredictably and can create very high initial stress. Fix overlaps in the design, or treat them deliberately as an interference in a nonlinear study.
Tip: The Max. Activation Distance must be set up before the solve, before the model deforms. Autodesk recommends setting it manually and gives the example of a value about 1.1 to 1.2 times a 10 mm gap.
Practice exercise#
Make two blocks, one on top of the other, 10 mm thick each, with a sideways force at the top edge of the top block and the bottom block fixed. Solve once with Bonded contact. Note the stress at the interface. Then clone the study, change the contact set to Separation (with a downward weight added using gravity) or Sliding, and solve again. Expected result: the bonded stack behaves as one body, the Separation or Sliding model lets the blocks move relative to each other, and the two results differ markedly.
Worked example: bonded versus sliding in a stacked plate#
The dimensions are illustrative values for the exercise. Two 10 mm steel plates are stacked and loaded in bending at one end.
- With a bonded contact over the whole interface, the plates act as a single 20 mm thick plate. Bending stiffness is proportional to thickness cubed: 20^3 = 8000 against 2 x 10^3 = 2000 for two independent plates, so a bonded model is four times stiffer in bending than free-sliding layers (hand calculation).
- With a Sliding contact, the plates bend independently. Displacement is about four times larger.
- Reality sits between the two, depending on friction and on bolt preload: a preloaded joint stays clamped until the external load exceeds the clamp force, then opens.
Use the bonded model for a first look, and use Separation, Sliding or Rough contact (nonlinear studies when needed) when the joint carries bending, when a gap can open, or when you want the true bolt load.
Automatic contact detection in detail#
| Behavior | Detail |
|---|---|
| Generated when | You click Automatic Contacts, Generate Mesh or Solve |
| Prior settings | Settings for manual sets and for edited automatic sets are retained when the routine is rerun |
| Criteria | Distance within Contact Detection Tolerance; entities parallel within 15 degrees (fixed) |
| Self-contact | Not detected automatically; define it manually |
| Deleting | Automatic sets cannot be deleted, only suppressed |
If a part rests on another with a small gap (for example from a fit allowance), the default tolerance may miss it and leave the part free to float. Increase the Contact Detection Tolerance to just above the largest gap, regenerate contacts, and check the Contact Manager. As a general sanity check (not an Autodesk rule), a simple chain of four parts should have about three contact sets; many more suggests duplicate or false matches.
Contact choices by situation#
General modelling guidance, not Autodesk data:
| Situation | Contact | Study |
|---|---|---|
| Welded or glued parts | Bonded | Static stress |
| Plate resting on a table | Separation (frictionless or with friction) | Static or nonlinear |
| Pin in an oversized hole | Separation with a friction coefficient | Nonlinear |
| Bolt clamping two plates | Bolt Connector with preload, contact set to Separation | Static or nonlinear |
| Parts sliding along each other without separating | Sliding | Static or nonlinear |
| Parts in contact that stick but can separate | Rough | Static or nonlinear |
| Heat flowing across a joint | Contact with a Thermal Conductance value | Thermal |
| Moving assemblies and impacts | Global contacts | Event simulation |
Friction coefficients are material and surface dependent; take them from your own references or tests, and run a range to test sensitivity. Autodesk's help does not publish a table of friction coefficients.
Thermal contact conductance#
With Thermal Conductance left blank, a contact is infinitely conductive and temperatures are identical on both faces. When you enter a value, the results show a temperature difference across the interface. A general calculation with hypothetical numbers: 10 W through a 100 mm^2 (1e-4 m^2) interface at an assumed 5000 W/(m^2 K) gives 10 / (5000 x 1e-4) = 20 K. The conductance must come from your own data, as Autodesk does not list typical values.
Failure modes and fixes#
| Symptom | Cause | Fix |
|---|---|---|
| Part flies away in the result | No contact detected, so the part is unconstrained | Increase the Contact Detection Tolerance; add a manual contact |
| Joint behaves too stiffly | Bonded where separation would occur | Change the pair to Separation or Rough |
| Body penetrates another | Activation distance too small | Increase Max. Activation Distance or try nonlinear static stress |
| Nonlinear solve does not converge with friction | Friction adds nonlinearity | Start frictionless, add friction afterwards |
| Stress spikes at the edge of a contact patch | Contact edge singularity | Judge stress away from the edge; refine locally |
| Heat does not flow between parts | No contact between them | Add the pair or a manual contact |
| Offset bonded pair not connected | Max. Activation Distance smaller than the gap | Use at least the gap plus about half the mesh size |
Practice exercise#
Quantify the effect of bonded contact on stiffness.
- Model two aluminium plates 100 x 30 x 5 mm, stacked, fixed at one end.
- Apply 20 N at the free end of the top plate only. Solve in static stress with bonded contact and record tip displacement.
- Predict the sliding case: the load goes into the top plate alone, which bends as a 5 mm plate; deflection = F L^3 / (3 E I) with I = 30 x 125 / 12 = 312.5 mm^4, so 20 x 1e6 / (3 x 68900 x 312.5) = 0.31 mm (E of 68.9 GPa is a typical aluminium value; use the one in your Fusion material).
- Change the contact to Sliding (the load transfer is idealised) and compare. Expect the sliding model near 0.31 mm and the bonded model near one eighth of that (about 0.04 mm), because the bonded stack bends as a single 10 mm plate (10^3 / 5^3 = 8 times stiffer).
- Switch to Separation with a Coefficient of Static Friction of 0.3 and record the change.
Related#
- Loads and Constraints
- Nonlinear Static Stress
- Static Stress Study