NDTnet 1998 Aug, Vol.3 No.8

Abstract

The problems of fatigue and fracture prediction for construction materials practically always run into difficulties connected with low cost effectiveness of experiments, which cannot create the full spectrum of real loads and actions for the given test period. Furthermore, the main experiments of fatigue and fracture analysis use short time testing for simulation of real life processes. In this case many basic characteristics of materials, such as plastic flow, micro relaxation with redistribution of tension inside the material, are not able to manifest fully. That is why fatigue and fracture prediction based on the data, which was received during these tests do not correspond well with the in-service structure performance. It is specially significant for brittle and elasto-plastic materials, where thermodynamic balance between elastic and plastic properties in time has considerable differences for short and long time actions. Moreover, the majority of known approaches, which are used for fatigue physical tests, have deterministic character, and prognosis models are limited by analysis of material's response for physical action. In this case it may be extremely tideous to estimate other types of loads and actions, which are left beyond the frames of the experiments. These problems are characteristic of the following brittle and elasto-plastic materials: ceramics, metal ceramics, fibre glass, concrete, and some other types of composites. To overcome these problems the authors made an attempt to apply the stochastic modelling methods using data of fatigue short term physical tests, and data about cracking process for adjusting of stochastic models (computation of coefficients) with further simulation of long term testing for estimation of fatigue and fracture. Different types of transfer functions (Box-Jenkins models) are used for building stochastic models of brittle and elasto-plastic materials' behavior. Spectrum density of short term testing loads and spectrum characteristics of cracking processes (based on acoustic emission data) are combined for analysis of transfer function (stochastic model). Then computer simulation is used for prediction of material behavior under long term loads. Some experiments were carried out on a number of concrete specimens for modelling of brittle and elasto-plastic materials using special composition and plasticizing. Various types of cyclic loads (spectrum density) were used to study their influence on correctness of building stochastic models as well as on estimation of material fatigue taking Wohler curve for comparison. The presented approach may serve as a base for research and development of new materials and control of in-service materials and structures. It would also give an opportunity to make tests more time and cost effective thanks to combination of short term testing for adjusting of stochastic models of materials' behavior with computer on of the models for estimation of fatigue and fracture.

Abstract Source:
Book of Abstracts, 7th European Conference on Non-Destructive Testing, 26-29 May 1998, ISBN: 87-986898-0-00

Full-Text Source:
Proceedings of the 7th European Conference on Non-Destructive Testing, 26-29 May 1998, ISBN:

Publication Contact:
7th ECNDT, Park Alle 345, DK-2605 Broendby, Denmark, Fax: +45 46 26 70 11, Email: 101373.3414@compuserve.com