IN SITU CONCRETE STRENGTH ASSESSMENT: PART II - VALUATION OF EQUIVALENT AGGREGATE HARDNESS BY NEURO-FUZZY TECHNIQUES

In this paper a series of nondestructive and destructive laboratory tests is presented with the purpose of investigating the influence of aggregate hardness on the test results of Windsor probe test system. A series of specimens fabricated by aggregates having various values of Mohs hardness (i.e., aggregates of fluvial origin) and by aggregates with only a class of Mohs hardness (crushed aggregates) was prepared. During the concrete casting, a series of cubical specimens was prepared. Subsequently the Windsor probe system was applied to estimate the in situ strength of specimens. Then, from each specimen two cores were extracted. Finally, a comparison between penetration tests and core strengths was carried out. The selection of the more suitable choice of Mohs hardness was carried out by neuro-fuzzy techniques. Fuzzy surfaces techniques were exploited in order to reduce the computational complexity and to select the input set of our system. The comparison between nondestructive and destructive tests, conducted on cores from the specimens, gave evidence of the reliability of the technique.