abaqus-fatigue-analysis

疲劳与耐久性分析的工作流程——进行循环计数、损伤累积，并预测疲劳寿命。

SKILL.md

--- frontmatter

name: abaqus-fatigue-analysis
description: Workflow for fatigue and durability analysis - cycle counting, damage accumulation, and fatigue life prediction.
allowed-tools:
  - Read
  - Write
  - Edit
  - Glob
  - Grep
  - Bash(abaqus:*)
  - Skill

Abaqus Fatigue Analysis Skill

Predict fatigue life from FEA stress results using S-N curves and damage accumulation.

When to Use This Skill

Route here when user mentions:

•"fatigue", "how many cycles", "fatigue life"
•"durability", "S-N curve", "cycles to failure"
•"rainflow counting", "Miner's rule"
•"high-cycle fatigue", "low-cycle fatigue"

Route elsewhere:

•Just stress analysis → /abaqus-static-analysis
•Crack propagation → specialized fracture tools
•Static strength check → /abaqus-static-analysis

Important: Abaqus Fatigue Limitations

Abaqus has limited native fatigue capabilities. The typical workflow is:

•Run structural analysis in Abaqus (stress/strain results)
•Extract stress history from ODB
•Apply fatigue criteria externally (Basquin, Miner's rule)

For full fatigue analysis, consider external tools: fe-safe, nCode, FEMFAT.

Prerequisites

Before fatigue analysis:

•✅ Completed static or dynamic analysis with converged results
•✅ Material fatigue data (S-N curve or Coffin-Manson parameters)
•✅ Stress output at critical locations

Workflow Steps

Step 1: Run Stress Analysis

Use /abaqus-static-analysis for constant loads or /abaqus-dynamic-analysis for time-varying.

Ensure output requests include:

•S - Stress components (principal, Mises)
•E - Strain components
•PEEQ - Equivalent plastic strain (for low-cycle)

Step 2: Identify Critical Location

Find the maximum stress location:

•Use /abaqus-odb to extract peak stress
•Check stress concentrations (fillets, holes, notches)
•Consider fatigue notch factor (Kf) vs stress concentration (Kt)

Step 3: Extract Stress History

For constant amplitude: single max/min stress values. For variable amplitude: full stress-time history for rainflow counting.

Step 4: Apply Fatigue Criteria

Use appropriate method based on loading and life regime.

Step 5: Calculate Life and Damage

Apply Basquin equation for life, Miner's rule for cumulative damage.

Key Decisions

Fatigue Approach

Approach	When to Use	Data Needed
Stress-life (S-N)	High-cycle (N > 10^4)	S-N curve
Strain-life (e-N)	Low-cycle (N < 10^4)	Coffin-Manson params
Fracture mechanics	Crack growth	da/dN curve

Loading Type

Loading	Analysis Method
Constant amplitude	Single static analysis
Variable amplitude	Multiple loads + rainflow
Proportional	Single load case
Non-proportional	Critical plane method

Mean Stress Correction

Method	Use Case
Goodman	Conservative, tensile mean
Gerber	Less conservative
Soderberg	Very conservative
SWT	Strain-life with mean stress

What to Ask the User

If unclear, ask:

•Material fatigue properties? S-N curve coefficients or test data?
•Loading type? Constant amplitude or variable (spectrum)?
•Mean stress? Fully reversed (R=-1) or with mean stress (R=0)?
•Critical location known? Or need to find max stress?
•Life target? What's the required number of cycles?

Key Parameters

Parameter	Typical Values	Notes
S-N slope (b)	0.08-0.15	Lower = longer life
Endurance limit	40-50% UTS (steel)	Stress below which infinite life
Fatigue notch factor (Kf)	1.0-3.0	Kf = 1 + q(Kt-1)
Notch sensitivity (q)	0.7-0.95	Higher for stronger steels

Troubleshooting

Problem	Cause	Solution
Unrealistically short life	Stress singularity	Use Kf correction, refine mesh away from singularity
Wrong units	MPa vs Pa mismatch	Verify stress units match S-N data
Unconservative prediction	Missing mean stress	Apply Goodman/Gerber correction
Very long calculated life	Stress below endurance limit	Check if stress > endurance limit

Related Skills

•/abaqus-static-analysis - Base stress analysis
•/abaqus-dynamic-analysis - Time-varying loading
•/abaqus-amplitude - Cyclic loading definition
•/abaqus-odb - Extract stress history from results

Code Patterns

For API syntax, equations, and code examples, see: