Structural damage identification method combining convolution and recurrent neural network

Abstract

The invention discloses a structural damage identification method combining convolution and recurrent neural networks. The method comprises the steps of collecting vibration response acceleration dataof different positions of a to-be-detected target by utilizing a plurality of sensors; preprocessing the vibration response acceleration data to form a time series data matrix; extracting spatial correlation features and short time scale dependence features from the time series data matrix by using a convolutional neural network; extracting a long-time scale dependence feature based on the spatial correlation feature and the short-time scale dependence feature by using a gating loop network; and classifying the damage state of the to-be-detected target by using the long-time scale dependencecharacteristic. According to the invention, the relation of monitoring data in time and space is considered; the problems of insufficient feature extraction, poor robustness and relatively weak mode classification capability existing in damage state identification are avoided, structure damage identification precision is substantially improved, more calculation space consumption can be saved underreasonable training conditions, and relatively good calculation time and space balance are realized.

Description

technical field [0001] The invention relates to the field of structural damage recognition, in particular to a structural damage recognition method of joint convolution and cyclic neural networks. Background technique [0002] With the rapid development of roads, bridges and other infrastructure construction, my country has entered the ranks of the road and bridge powers, and has become the country with the largest number of bridges in service in the world. However, during the long-term operation of the structure, it will inevitably be affected by factors such as natural environment erosion, human activities, and material aging. In addition, bridge management generally has problems such as "reconstruction and light maintenance", which will make the fatigue damage of the bridge structure worse. , resulting in the structure failing to meet the requirements of long-term operation safety, durability, maintainability and sustainability. Due to the large scale of the structure, o...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the deficiencies of the above-mentioned prior art, the problems actually solved by the present invention include: how to fully consider the connection of monitoring data in time and space, and avoid the situation of insufficient feature extraction, poor robustness, and weak pattern classification ability in damage state identification

Images