Damage Detection Method of Ancient Timber Structure Based on BP Neural Network and Total Wavelet Energy Rate. Issue 3 (March 2020)

• Record Type: Journal Article
• Authors: Hu, Weibing; Yang, Jia; Hou, Yanfang
• Is Part Of: IOP conference series. Volume 780:Issue 3(2020)
• Publication Date: 2020-03
• Subjects: Materials science -- Periodicals; 620.1105
• DOI: 10.1088/1757-899X/780/3/032002
• Languages: English
• ISSNs: 1757-8981

Abstract

Abstract: In order to detect the damage of beams and mortise joints of ancient buildings under environmental excitation, a damage identification method combining wavelet transform with improved BP neural network is proposed. The acceleration signal of the structure under the environment excitation is extracted and the total energy change rate of wavelet is obtained by wavelet discrete reconstruction, which is taken as the damage index. A finite element model of a timber frame of Xi'an bell tower is established to verify the proposed method. When the damage extent is less than 50%, 96% of the test samples in the neural network output the desired value; This method provides a theoretical basis for the study of damage prediction of timber structures in ancient buildings under environmental incentives.