Linear Analysis of Bolted Joints
Linear Analysis of Bolted Joints: Theoretical Foundations
Mechanics of Bolted Joints
Professor, how do you perform FEM analysis of bolted joints?
Bolted joints are the most common joining method in structural analysis, but FEM modeling is surprisingly difficult. Bolt pre-tension (initial tension), compression of the clamped parts, axial force fluctuation due to external loads, slippage of contact surfaces... all contain nonlinear elements.
Basic Mechanics of Bolted Joints
The foundation of bolted joint mechanics is the bolt-clamped part series spring model.
Bolt axial stiffness: $k_b = E_b A_b / L_b$
Clamped part compression stiffness: $k_c$
After applying pre-tension (initial tension) $F_i$, when an external force $F_{ext}$ acts:
Here, $\Phi = k_b / (k_b + k_c)$ is the internal force coefficient (load introduction factor).
Only a portion ($\Phi$ times) of the external force enters the bolt as additional load. The remainder is absorbed by the relaxation of the clamped part compression.
In typical bolted joints, $\Phi = 0.1 \sim 0.3$. This means only 10-30% of the external force enters the bolt. This is the effect of pre-tension, which significantly improves the bolt's fatigue life.
So pre-tension is important for fatigue reasons.
Yes. Without pre-tension, the external force directly enters the bolt, leading to fatigue fracture. Appropriate pre-tension can significantly reduce the stress fluctuation amplitude in the bolt.
VDI 2230 Guideline
Are there design standards for bolted joints?
VDI 2230 (German Association of Engineers guideline) is the global reference standard for bolted joint design. It systematically addresses pre-tension, load introduction, clamping force, fatigue, and temperature effects.
VDI 2230 hand calculations are limited to the axial problem of a single bolt. For eccentric loads or interactions of multiple bolts, FEM is required.
Modeling Levels in FEM
There are three levels of FEM modeling for bolted joints:
| Level | Modeling | Accuracy | Cost |
|---|---|---|---|
| Level 1 | Spring element + RBE2/RBE3 | Low (rough estimate) | Low |
| Level 2 | Beam element + contact | Medium | Medium |
| Level 3 | Solid element + contact + thread geometry | High | High |
Level 3 includes the thread geometry?
It's necessary if you want to precisely evaluate stress concentration at the thread root or bearing pressure. However, Level 1-2 is usually sufficient. Level 3 is only for special cases (fatigue evaluation, high-temperature bolts, etc.).
Summary
Let me organize the theory of bolted joints.
Key points:
- Series spring model — The ratio of bolt stiffness $k_b$ to clamped part stiffness $k_c$ is key
- Internal force coefficient $\Phi = k_b/(k_b+k_c)$ — Only 10-30% of external force enters the bolt
- Pre-tension improves fatigue life — Reduces stress fluctuation amplitude
- VDI 2230 is the design standard — Hand calculations are for single bolt axial problems
- Three modeling levels — Appropriate level of detail for the application
So the value of $\Phi$ determines everything in bolt design.
If you know $\Phi$ and the initial tension $F_i$, you can calculate the bolt's maximum axial force, bearing pressure, and fatigue life. In FEM, the goal is to correctly evaluate these two parameters.
History of Bolt Tightening Theory
The concept of bolt axial force (pre-tension) was first systematized in the VDI 2230 standard issued by VDI (German Engineers Association) in 1938. This standard has been revised many times and VDI 2230:2015 is used as a global standard in the automotive and aerospace industries. The core concept of the "tightening factor" present from the first edition remains unchanged even 80 years later.
Computational Methods for Linear Analysis of Bolted Joints
Pre-tension Setting in FEM
How do you set bolt pre-tension in FEM?
Each solver has dedicated functions for pre-tension.
Nastran
```
$ Pre-tension on bolt beam element
TEMP(INIT) = 100
TEMP(LOAD) = 200
```
Simulating pre-tension with thermal loads is a classic method. Recently, SOL 400's *BOLT PRELOAD function is used.
Abaqus
```
*STEP
*STATIC
*BOLT LOAD
bolt_section, bolt_mid_surface, 50000.
*END STEP
```
*BOLT LOAD directly applies initial tension to the bolt cross-section. It's the most intuitive.
Ansys
```
! Workbench: Bolt Pretension load
! APDL: PRETS179 element
SLOAD, bolt_section, PRETEN, 50000.
```
Abaqus's *BOLT LOAD seems the easiest to understand. You cut the bolt cross-section and apply initial tension.
In Abaqus implementation, a "pre-tension node" is added to the bolt's mid-surface, and an axial force is applied to that node. By switching from force to lock displacement, the pre-tension is maintained even after external forces act.
2-Step Analysis
Do you apply pre-tension and external forces simultaneously?
It's standard to analyze in 2 steps:
Step 1: Pre-tension
- Apply initial tension to the bolt
- Clamped parts are compressed
- Bolt axial force = $F_i$ (design value)
Step 2: External Force
- "Lock" the pre-tension (fix the length)
- Add external forces (internal pressure, self-weight, wind load, etc.)
- Bolt axial force = $F_i + \Phi \cdot F_{ext}$
Why "lock" it?
If the pre-tension force is simply maintained as-is in Step 2, the pre-tension force remains constant even when deformed by external forces (physically incorrect). By switching to lock (displacement fixation), the bolt axial force naturally fluctuates due to deformation from external forces. This aligns with the VDI 2230 model.
Necessity of Contact
Is contact necessary for bolted joint analysis?
It's essentially a nonlinear contact problem. The clamping force on the clamped surfaces may decrease due to external forces, potentially leading to separation.
However, for linear analysis within elastic range where no separation occurs, contact can be approximated as tied (glued) surfaces. This simplifies the computation significantly.
So we use tied contact only in linear analysis.
Correct. Once you want to evaluate actual contact pressure distribution, slippage, or separation behavior, you must switch to nonlinear contact analysis. That's a different and more complex topic.
Summary
So linear analysis of bolted joints is a simplified approach?
Yes. But it's practical. In engineering, we often use linear analysis first to quickly evaluate bolt stress and check fatigue life. Only when detailed accuracy is required—or when separation/slippage is suspected—do we perform nonlinear analysis.
Typical workflow:
- Create CAD model with bolt and clamped parts
- Mesh using beam (Level 2) or solid elements (Level 3)
- Define bolt pre-tension in Step 1
- Lock pre-tension and apply external forces in Step 2
- Extract bolt stress using $\sigma = F_b / A_b$ and check fatigue
- If results are marginal, refine with nonlinear contact analysis