Using fatigue software for probabilistic data assessment of S-N and ε-N fields, crack-growth and damage based design under varying load

18.7.2018

Seminar Fraunhofer IWM, 18. July 2018, Seminarraum des Fraunhofer-Instituts für Werkstoffmechanik IWM in Freiburg, Dipl. Ing. Sergio Blasón, University of Oviedo

Details: 

 

Was:                        Seminar Fraunhofer IWM: Using fatigue software for probabilistic data assessment of S-N and ε-N fields, crack-growth and damage based design under varying load

Wann                      18. Juli 2018, 14:00 Uhr

Wo:                         Seminarraum W9a des Fraunhofer Instituts für Werkstoffmechanik IWM in Freiburg

Vortragender:        Dipl. Ing. Sergio Blasón, University of Oviedo

 

Three different methodologies related to fatigue design are presented for data assessment of a) S-N and ε-N fields, b) cumulative damage prediction and c) crack-growth curves. All the three are handled using software programs including probabilistic considerations:

 

a)             P-S-N curves from fatigue test results

Introduction and management of the ProFatigue software

Using ProFatigue to evaluate experimental results

Test strategy: ProFatigue as a support tool for decision making in experimental campaigns

 

The program provides an estimation of the parameters involved in the Weibull regression fatigue model developed by Castillo-Canteli, allowing a probabilistic definition of the S-N field to be applied to the fatigue design of structural and mechanical components using both stress and strain based approaches. An extension to multiaxial and other complex lifetime cases is possible by chosing more general driving forces as reference parameter. The fortresses of the model become apparent in the successive model developments and improvements performed up to date.

 

b)             Assessment of damage rules (linear and non-linear) using a probabilistic approach

Fundamentals and methodology

Implementation example

Operating with ProMiner software

 

As an alternative to deterministic damage rules, a probabilistic Miner cumulative damage approach is proposed for lifetime prediction of components subject to varying loads. The approach requires the probabilistic definition of the S-N field according to the precedent Section to proceed to a parallel calculation of the Miner number and the normalized variable, V=log (N/N0) log (Δσ/Δσ0) where the latter can be conveniently related to probability of failure. The inherent scatter due to the random sequence is taking into account by recursive application of a proportionally reduced load spectrum and. The suitability of this approach is confirmed for different experimental programs under distinct materials and variable loading histories.

 

c)             Fatigue crack growth rate curves:

Main concepts of the model based on Fracture Mechanics

FCGR curve normalization

ProPagation for the evaluation of experimental results

 

This tool allows the entire fatigue crack growth rate to be entirely fitted from experimental data, by normalizing the stress intensity range (ΔK*+) as a dimensionless quantity varying in the 0-1 range and analyzing the da/dN term as a cumulative distribution function. In particular, an estimate of the threshold stress intensity factor range value is provided, as a model parameter. Based on a fundamental definition of the geometric factor, once-through integration of the crack growth rate curve provides a master crack growth curve to be defined, allowing the crack growth curves to be obtained for any other initial crack size or load conditions. In this way the propagation S-N field may be defined in a probabilistic way.