Masonry wall damage by restraint to shrinkage

Restrained shrinkage has been identified as a major source of damage to buildings in The Netherlands. Several numerical studies have been performed and reported. Yet, successful experimental simulation and quantification have not yet been performed. In this paper, the successful design, setup, and execution of such an experiment are described. The specimen was a solid masonry wall of 2 m length, 1.3 m height, and 220 mm thickness, with a central opening of 420 mmX420 mm. An inverse approach to restrained shrinkage was followed by heating the constraining boundaries, in this case stiff, solid aluminum beams simulating rigidly connected upper and lower floors. To simulate upper wall and floor bearing weight, precompression was applied through six pairs of steel rods spanning between the aluminum beams. At a temperature difference of about 12degreesC a crack was initiated in the wall at the opening, which instantly and audibly propagated up and down over the full height of the wall. This was in good agreement with the predicted behavior, based on computational modeling with the recently developed finite element models for masonry, with parameters extracted from separate compression, shear, and tensile tests on small specimens of the same masonry.