MSE 2016 - Full Program

Back to overview

Lecture

Simulation and analysis of fatigue lifetimes

Wednesday (28.09.2016)
09:45 - 10:00
Part of:


We present a practicle workflow of estimating fatigue lifetimes with the help

of a quite general and adjustable estimation model which provides

easy-to-use tools for model-driven analysis of different composite

material structures such as for instance (ceramic particle or fibre reinforced)

metall matrix composites. Given the data of the metallographic analysis we

shortly describe the process of construction and simulation of appropriate

random spatial geometric particle or fibre models. Two different

competitive methods are looked at in more detail, the so called (stereological)

``unfolding'' procedure of planar intersected objects and the

Quasi-likelihood parameter estimation method. We discuss

the ``pros'' and ``cons'' of each approach from a statistical point of view and

give some ad-hoc advise when to use either of the methods. Further we propose a

fatigue lifetime estimation model as implemented in the R-package

``simLife" for different spatial particle or fibre

configurations with an additional secondary disturbing phase. The lifetime model

allows for the inclusion of predefined densely clustered regions of objects

which can be simulated beforehand. The user can also provide particular material

constants such as Vickers hardness and others according to the phenomenological

properties of specific material structures and for appropriateness of the

lifetime model simulations. We end up with results of a simulation study on how

good the parameters of the lifetime model can be reproduced in order to assess

the accuracy of the estimated model parameters as a preliminary step towards

fitting the model to real world experimental data.

Speaker:
Markus Baaske
TU Bergakademie Freiberg