Choosing Between Bonded and Frictional Contact for Bolted Joint Preload in Ansys Mechanical

Problem: Predicting realistic preload loss in a bolted joint

When you model a bolted joint in Ansys Mechanical, the contact definition between the bolt head/nut and the clamped parts has a direct influence on how preload is transferred and whether slip occurs. Using an overly stiff contact can hide micro‑slip and give non‑conservative stiffness, while an overly compliant contact may cause convergence difficulties. The choice therefore affects both accuracy and solve time.

Thesis: Bonded contact is appropriate for fully stuck interfaces; frictional contact with a realistic coefficient of friction should be used when slip is expected or when you need to capture load redistribution near the bolt.

Section 1: Understanding the two contact formulations

Bonded contact assumes infinite shear stiffness across the interface. The surfaces are tied together as if they were welded, which eliminates any relative tangential motion. This formulation simplifies the model, reduces the number of contact iterations, and typically yields faster solves. It is suitable when you are confident that the joint will remain fully stuck under the expected load range (e.g., high preload, low external shear). Frictional contact introduces a tangential stiffness governed by a coefficient of friction (μ). The solver allows stick‑slip transitions, enabling micro‑slip and load redistribution. This captures the real‑world behavior of bolted joints where the clamped plates may shift slightly relative to the bolt head/nut, especially under cyclic or off‑axis loading. The trade‑off is increased nonlinearity, longer solution times, and a need for careful stabilization settings.

Section 2: Worked example – M8 bolt through two 10 mm plates

To illustrate the difference, consider a simple axisymmetric model: an M8 bolt (shank diameter 8 mm, length 30 mm) passing through two 10 mm thick steel plates. The goal is to apply a 10 kN tensile preload and examine the bolt axial force after the preload step.

1. Geometry: Create the bolt shank, head, nut, and plates in DesignModeler or SpaceClaim. Use a single bolt‑through‑hole representation; the head and nut can be modeled as rigid bodies if desired.
2. Contacts:
   • Bonded case: Define a bonded contact between the bolt head/nut faces and the adjacent plate faces.
   • Frictional case: Define a frictional contact with μ = 0.2 (typical for dry steel‑steel) on the same face pairs.
3. Preload: Use the Preload load step (Analysis Settings → Preload) to apply a 10 kN axial force on the bolt shank. Enable large deflection and turn on nonlinear controls.
4. Mesh: Apply a refined mesh in the bolt shank and contact zone (e.g., 0.5 mm element size) and a coarser mesh elsewhere (2 mm). Perform a mesh refinement study with sizes 2 mm, 1 mm, 0.5 mm.
5. Solution: Run a static structural analysis for each contact type, capturing the reaction force at the bolt shank after the preload step.

What to look for (do not claim these results are from a tested run):

• In the bonded model, the bolt axial reaction should be very close to the applied 10 kN, indicating negligible preload loss.
• In the frictional model, a small reduction (on the order of a few percent) may appear due to allowed micro‑slip at the interface.
• As the mesh is refined, the reaction force should converge; the curve of bolt force versus element size will flatten, confirming that the contact solution is mesh‑independent.

If the frictional model shows a larger loss than expected, check the stabilization settings (increase the stabilization factor or switch to augmented Lagrangian) and verify that large deflection is active.

Section 3: Trade‑offs and limitations

Bonded contact is computationally efficient but can over‑predict joint stiffness when slip is possible. This may lead to non‑conservative designs, especially for joints subjected to vibration or lateral loads where micro‑slip influences fatigue life.

Frictional contact provides a more realistic representation of load transfer and can predict preload loss accurately, but it requires:

• Careful selection of the friction coefficient (μ). A value that is too low may cause excessive slip; too high may behave almost like bonded contact.
• Appropriate nonlinear solver controls (large deflection, adequate iteration limits, and possibly augmented Lagrangian stabilization).
• A sufficiently refined mesh in the contact region; coarse meshes can artificially stiffen the interface and mask slip.

Both formulations depend on correct definition of the contact surfaces (normal direction, contact tolerance) and proper application of the preload step. Neglecting any of these can cause premature termination or unrealistic results.

Actionable closing

Start your bolted joint analysis with a bonded contact to verify basic load path and obtain a quick baseline. If the joint is expected to experience any relative tangential motion (e.g., due to external shear, thermal cycling, or loosening concerns), switch to a frictional contact with a physically based μ and perform a mesh refinement study. Always verify that the bolt axial reaction after preload matches the applied value within an acceptable tolerance and that the solution converges with mesh refinement. This approach balances computational efficiency with the fidelity needed for reliable preload‑loss predictions.