This study focuses on a structural health monitoring (SHM) using case-based reasoning (CBR). Structural condition is
diagnosed using propagation patterns of ultra-sonic waves. Firstly, emitted pseudo-AE waves are measured in pencil lead
fracture experiments. Then, the AE signals are classified into 90 types according to location, magnitude and structural
condition. Secondly, pattern identification is conducted using feature parameters extracted from the signals for damage
pattern recognition. Finally, feasibility of the method to real structures using CBR is studied. Results showed that the
damage patterns could be determined with 82% accuracy. If only the damage location is needed, of the accuracy was
higher with 95%. The proposed method using ultra-sonic waves and CBR is thus feasible for practical applications.