Damage Diagnosis of Bleacher Based on an Enhanced Convolutional Neural Network with Training Interference  

作　　者：Chaozhi Cai Xiaoyu Guo Yingfang Xue Jianhua Ren 

机构地区：[1]School of Mechanical and Equipment Engineering,Hebei University of Engineering,Handan,056038,China

出　　处：《Structural Durability & Health Monitoring》2024年第3期321-339,共19页结构耐久性与健康监测（英文）

基　　金：the Nature Science Foundation of Hebei Province Grant No.E2020402060;Key Laboratory of Intelligent Industrial Equipment Technology of Hebei Province(Hebei University of Engineering)under Grant 202206.

摘　　要：Bleachers play a crucial role in practical engineering applications, and any damage incurred during their operationposes a significant threat to the safety of both life and property. Consequently, it becomes imperative to conductdamage diagnosis and health monitoring of bleachers. The intricate structure of bleachers, the varied types ofpotential damage, and the presence of similar vibration data in adjacent locations make it challenging to achievesatisfactory diagnosis accuracy through traditional time-frequency analysis methods. Furthermore, field environmentalnoise can adversely impact the accuracy of bleacher damage diagnosis. To enhance the accuracy and antinoisecapabilities of bleacher damage diagnosis, this paper proposes improvements to the existing ConvolutionalNeural Network with Training Interference (TICNN). The result is an advanced Convolutional Neural Networkmodel with superior accuracy and robust anti-noise capabilities, referred to as Enhanced TICNN (ETICNN).ETICNN autonomously extracts optimal damage-sensitive features from the original vibration data. To validatethe superiority of the proposed ETICNN, experiments are conducted using the bleacher model from Qatar Universityas the subject. Comparative studies under identical experimental conditions involve TICNN, Deep ConvolutionalNeural Networks with wide first-layer kernels (WDCNN), and One-Dimensional ConvolutionalNeural Network (1DCNN). The experimental findings demonstrate that the ETICNN model achieves the highestaccuracy, approximately 99%, and exhibits robust classification abilities in both Phases I and II of the damagediagnosis experiments. Simultaneously, the ETICNN model demonstrates strong anti-noise capabilities, outperformingTICNN by 3% to 4% and surpassing other models in performance.

关 键 词：Bleacher vibration signal damage diagnosis convolutional neural network anti-noise ability 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]