Statistical and probabilistic techniques in modeling of epoxy cracking phenomena

Tomasz Nowak¹, Mateusz Ambroziński²

¹ABB Corporate Research Center, Krakow, Poland.
²AGH University of Science and Technology, Krakow, Poland.

DOI:

https://doi.org/10.7494/cmms.2013.1.0410

Abstract:

Epoxy resins filled with mineral particles, such as silica and alumina, are well known and extensively used as insulating materials in power product applications. They are relatively cheap and easy to process, and their dielectric, thermal and mechanical properties are appropriate to maintain their functionality in different electrical devices. However, the mechanical performance of particle filled epoxy systems can be often influenced by limited resistance to cracking. It is frequently experienced during the lifetime of the epoxy cast products that some members of the production lot will crack during verification tests, while the others will remain healthy, even if the test loads are seriously increased. It is especially seen during low-temperature tests, when the thermal conditions are changed from +120°C to -60°C, when a few products can crack at +10°C, while others may easily survive till end of the tests. This phenomenon prohibits of using the classical material strength as a failure criterion. For this reason, some statistical and probabilistic approaches to describe the cracking resistance of the epoxy-based material have to be considered. This work is focused on a study of process-induced residual stresses in epoxy components and their effect on material cracking likelihood. The proposed failure probability model incorporates Weibull distribution. The paper provides the experimental procedure for estimation of the Weibull model parameters (, ), as well as implementation remarks. In addition, the Design of Experiment (DoE) method was used to support FEM calculations, and to find out the correlation between product design parameters and the failure likelihood. Finally, the article outlines some recommendations for applying statistical and probabilistic methods into numerical procedures of FEM analysis.

Cite as:

Nowak, T., & Ambroziński, M. (2013). Statistical and probabilistic techniques in modeling of epoxy cracking phenomena. Computer Methods in Materials Science, 13(1), 49 – 55. https://doi.org/10.7494/cmms.2013.1.0410

Article (PDF):