GIDPC.Green Infrastructure · Disaster Prevention
Research & Technology
International Journal of FatigueUnder Review

Probabilistic Eccentric Decay Geometry: Monte Carlo Confidence Intervals on the Wind-Fatigue Life of Urban Street Trees

Relaxing the Mattheck concentric assumption, 10,000 Monte Carlo eccentric geometries show 99.6% of fatigue-life variance is explained by normalized eccentricity — statistical justification for prioritizing sonic-tomography surveys.

Authors
Junsuk Kang
Year
2026

Keywords

wind-fatigueeccentric decayMonte Carlo simulationurban street treesprobabilistic risk assessmentsonic tomographyMattheck threshold

Background

The fatigue-life pipeline of our earlier work (Trees submission) uses Mattheck & Breloer (1994)'s concentric hollow-cylinder model. But field sonic-tomography surveys show real decay cavities are eccentric, irregular, and frequently elongated.

This study quantifies how uncertainty in cavity geometry propagates into uncertainty on fatigue life.

Method

  • Derived closed-form section properties for circular and elliptical cavities of arbitrary eccentricity
  • Verified a vectorisable stress-amplification function that reproduces the Mattheck formula to machine precision in the concentric limit
  • Per scenario, sampled 10,000 cavity geometries: eccentricity ~ Beta(2,2), orientation ~ Uniform
  • Anchored to absolute-life pipeline from prior paper
  • Compared three wind-direction distributions (isotropic, Seoul empirical, Jeju empirical)
  • First-order Sobol decomposition for variance attribution

Key Results

| Metric | Value | |---|---| | Median fatigue life at Mattheck threshold | 25% of concentric estimate | | 5th-percentile lower bound shortening factor | 47× | | Variance contribution — normalized eccentricity | 99.6% | | Effect of wind direction (uniform-orientation prior) | ~0 |

Significance

Statistically proves that cavity eccentricity dominates fatigue-life uncertainty — far above wind climate. Therefore sonic-tomography (SoT) precision is the top priority for inspection-resource investment. We propose risk-based inspection keyed to 5th-percentile lower bounds, tabulated for three Korean wind regimes — replacing deterministic concentric estimates.