Design of damage identification algorithm for mechanical structures based on convolutional neural network

摘要

Damage identification, location, and estimation of engineering structures are very popularresearch topics in recent years. Structural damage detection technology has been widely used inaerospace, civil engineering, machinery, and nuclear industry. It is a multi-disciplinary and comprehensivetechnology based on damage mechanism, sensor technology, signal analysis technology,computer technology, and convolutional intelligence technology. Compared with the traditionalstructural damage detection methods, this paper mainly studies the theory and application ofstructural damage detection technology based on convolutional neural network. In this paper, thedamage location and damage degree of a frame structure are numerically simulated by the combinedparameter method,which is suitable for structural damage identification. At the same time,the input parameters of the improved convolutional neural network are constructed by a suitablemethod, and the structure damage detection and identification are carried out by using thetrained convolutional neural network.