Development of a Variable-Frequency Hammering Method Using Acoustic Features for Damage-Type Identification

Abstract

Hammer testing, a nondestructive testing method, has been demonstrated to provide information on structural damage. One of the biggest challenges with this testing method is the simultaneous identification of surface, internal, and composite damage (consisting of both surface and internal damage) in a complex environment, such as post-disaster. A method of identification based on variable-frequency hammering is proposed to solve this problem. The importance and feasibility of using variable-frequency impact hammers and the generated acoustic data to identify multiple types of damage in concrete structures are presented. First, a type of variable-frequency hammering acoustic feature was generated using acoustic feature extraction and selection based on the acoustic data obtained from variable-frequency hammering. Second, a damage recognition model was established using a support vector machine to identify four types of damage occurring simultaneously in the same concrete member specimens, including a type of composite damage with two types of damage occurring simultaneously within 20 mm. Finally, the feasibility of this variable-frequency hammering method was verified experimentally. This method exhibited good performance, with an accuracy of 97.8%; moreover, the method ensures that the feature dimensionality remains unchanged while increasing the effective information of the data.