Masonry arches crack inexorably after decentering. This phenomenon is well known
to any master builder. For small deformations these cracks do not affect the safety of
the arch. Indeed, the arch with time may show different patterns of cracking, which
lead to different sets of internal forces. Within the frame of modern limit
analysis, developed for masonry structures, mainly by Professor Heyman
since the 1960s, we know that cracking is irrelevant to safety: indeed, it is
the capacity of forming cracks which gives “plasticity” to masonry. Small
deformations do not distort the overall form of the arch. A direct corollary of
the safe theorem states that if it is possible to draw a line of thrust within
the arch, the arch will not collapse, and it is safe. This is independent of
the “actual” state of the arch, manifested by a certain pattern of cracks.
This pattern will change with very small (unpredictable) variations in the
boundary conditions; a tiny spreading of the abutments will produce a complete
change.
However, when the deformations are large, the geometry of the arch is severely
distorted and we cannot study the stability with the original geometry: it is necessary
to proceed step by step, considering the deformed geometry. This phenomenon has
been rarely studied. However, even these studies consider, as a simplification, that
the crack patterns do not vary and the movement is studied under this assumption.
That this is not the case can be seen with tests with small models of arches: the
position of cracks can be altered, and this may influence the validity of the study.
The present contribution proposes a method of analysis which permits us to
study the history of cracking until collapse. This has not only theoretical
interest; it may be used in the analysis of some critical cases which occur in
practice.