Influence of the number of acoustic emission descriptors on the accuracy of destructive process identification in concrete structures

摘要

The use of multi-parameter analysis of acoustic emission signals allows identifying destructive processes not only in prestressed concrete elements but also in classical concrete and reinforced concrete members. The analysis is not based on single descriptors but on destructive processes, which may be used to evaluate structural integrity of the elements. Therefore the key question is the minimum number of parameters needed to develop a database of reference signals. To determine the effect of particular descriptors on the correctness of the database being created, the authors applied dendograms and three statistical criteria, Wilks' Lambda, Rij and Tou criteria.