View all news

Seminar Download: Dr John Hart Smith (Boeing) 'Failures in Fibre-Polymer Composite Structures' 26th September, Faculty Board Room.

19 September 2016

'Is there really no need to be able to predict matrix failures in fibre-polymer composite structures?'

A logic-based case is presented that it IS necessary to be able to predict matrix failures in fibre-polymer composites, even though some 50 years after the introduction of the advanced fibre-polymer composites there is still no widely accepted theory capable of actually doing so. lack is due to the widespread malpractice of artificially homogenizing distinct fibre and matrix constituents into an “equivalent” anisotropic solid, to simplify the mathematics. The enormous residual thermal stresses in the matrix, even when the lamina is stress-free, are not widely recognized; these detract appreciably from the ability of the matrix to withstand applied mechanical loads. The extreme thermal mismatch between the fibres and the polymer causes the resin to try to shrink during cool-down after cure at high temperature, but the strong stiff fibres prevent this from happening. Such internal residual stresses cannot possibly exist in truly homogeneous materials, so the consequences are overlooked whenever homogenized material models are used in analyses when they shouldn’t have been. The case is made that it is impossible to predict matrix failures using any theory in which the fibre and resin constituents have been artificially homogenized, despite the myriad of theories claiming to do so that can’t. The proof of this failure of the traditional homogenized composite failure theories to predict failures in the matrix (as distinct from failures of the fibres) is illustrated by several practical situations that led to unanticipated premature failures.


'Is there really no need to be able to predict matrix failures in fibre-polymer composite structures?' John Hart Smith Presentation (PDF, 583kB)

Part 1: Explanation of fatal flaws in existing theories. John Hart Smith Matrix Failures 1 (PDF, 376kB) 

Part 2: Examples of matrix failures preceding fibre failures. John Hart Smith Matrix Failures 2 (PDF, 505kB)


BIO for Dr. John Hart-Smith

Dr Hart-Smith completed his education in his native Australia, obtaining a PhD degree in Engineering before joining Douglas Aircraft in Long Beach, California in 1968. Since then, he has worked extensively on adhesively bonded and mechanically fastened composite joints, particularly in composites, and trying (with great difficulty) to explain that the words “composites” and “homogeneous materials” are mutually exclusive to experts on fibre-polymer composites. Homogenized composite failure theories are intrinsically incapable of predicting matrix failures, despite claims to the contrary, because none of the transverse strains in the matrix are common with those in the lamina. Since retiring in 2008, he has resumed the studies he started as a postgraduate student at Monash University, in the 1960s, to predict the buckling strength of thin shells and explain the real source of the gross differences between test and theory as errors in the classical shell-buckling analyses, not imperfections in the experiments. He has now completed the correct analyses for most of the classical problems for cylindrical and spherical shells, and uncovered a great number of clear scientific errors in the classical works that later shell-buckling researchers overlooked.

Edit this page