1 DAMAGE ASSESSMENT OF REINFORCED CONCRETE USING ULTRASONIC WAVE PROPAGATION AND PATTERN RECOGNITION

Abstract. An active damage detection method using Support Vector Machine (SVM) is proposed. The SVM is known as a powerful pattern recognition tool that can be applied to complicated classification problems. A damage detection method using ultrasonic waves and SVM was adopted to characterize damages in concrete structures. AE sensors composed of stacked piezoelectric transducers (PZT) were used to predict damages by using differences of propagation wave signals recorded before and after damages. The wavelet transform was applied for a time-frequency analysis of the propagation waves to characterize the feature vectors for feeding to SVMs. This method can correctly estimate the location of damages. Applicability of proposed damage detection method was successfully demonstrated.