Non-destructive testing of concrete layer adhesion by means of vibration measurement

Abstract

A common practice in the construction of the garage floors of residential and office buildings is the construction of the concrete slab in two steps: one thick layer at the beginning of the work, a second thin layer at the end of it. This procedure saves time but also brings in the problem of a cold joint between the two layers which, eventually, may get loose or even split. However it is difficult to control the status of the concrete layers until they are completely split unless one make use of destructive tests such as the extraction of test samples.In this paper we demonstrate the use of vibrational data to assess and control the status of the joint between the two layers. To that end we have carried out two series of experiments.The first one is an experimental procedure in the laboratory with the goal to see if vibrational data are able to distinguish among different levels of fixation between the two layers. We use a concrete slab as the thick layer and a MDF board as the thin layer, bonded by means of gypsum plaster. Different amounts of the surface covered by the gypsum plaster give different levels of fixation between the layers. For each configuration we produced vibration with a standard tapping machine and measured it with the aid of a vibrometer. Our results are positive; with a simple test vibrational data can distinguish whether two layers are fixed by 25% or less, by 50% or by 100% of their surface. With the laboratory results we have designed a non-destructive test to be carried on in-situ measurements so as to ascertain the level of fixation of two layers of a concrete slab. The test is based on the installations described in Annex A of ISO 16251-1, so that it does not require large investments or facilities. It is able to cope with different kind of data and is independent of the characteristics of the slab thanks to the statistics tools applied to the data.The second one is the application of the test to data measured in an actual building with the aforementioned problem in its garage floor. We measured vibration induced by a different source, a car passing by, and in several locations, both with loose and well fixed layers. The vibration was recorded by the same vibrometer as above. Notwithstanding the differences in the data source, the test can be applied to the data and the results are compatible with the known status of the slab: we obtain a 77.0% of sensitivity and 54.5% of specificity, while only a 1.0% of false positives and none false negatives. The only drawback of the test is its high number of undecidable cases, which amounts to 24.1%.