Structural damage identification method and device based on parallel convolutional neural network

Abstract

The invention provides a structural damage identification method and device based on a parallel convolutional neural network. The method comprises the following steps: S11, data sampling: collecting acceleration data recorded by an acceleration sensor; s12, data preprocessing: preprocessing the acceleration data recorded by each acceleration sensor to obtain an acceleration data sequence and a time-frequency diagram thereof; s13, damage identification: inputting the input data set corresponding to each acceleration sensor into a trained damage identification model based on a parallel convolutional neural network, and performing feature extraction and damage category prediction on the input data set by the damage identification model to obtain a probability conforming to each damage working condition data feature; and obtaining a corresponding structural damage identification result according to the probability conforming to each damage working condition feature. Wherein the input data set comprises an acceleration time sequence and a time-frequency diagram thereof. According to the invention, the recognition effect of the structural damage can be improved.

Description

technical field [0001] The invention relates to the field of bridge health monitoring, in particular to a structural damage identification method and device based on a parallel convolutional neural network. Background technique [0002] During the long-term service of civil engineering structures, different degrees of structural damage will inevitably occur due to external loads, material performance degradation, and unexpected events, which will lead to a reduction in the structural bearing capacity and affect the reliability of the structure. sex and safety. The civil engineering structural health monitoring system (SHM) collects long-term monitoring data from sensors, monitors and evaluates the structural status, releases early warning information in a timely manner, and makes decisions on inspection, repair and reinforcement. Cutting-edge technology at the level of intelligent maintenance and management, in order to realize a structural health monitoring (SHM) system th...

AI Technical Summary

Problems solved by technology

[0006] At present, many structural damage identification methods based on deep learning have been developed, but there are still some limitations, which limit the application of deep learning in structural health monitoring.

Although 1D-CNN has powerful signal classification capabilities, it is only suitable for one-dimensional signals. In the face of vibration signals collected by multiple sensors in actual engineering, some scholars treat multiple channel data in series as a single data processing. However, this will Increasing the dimension of the input signal affects the performance of CNN. At the same time, the structural vibration signal is distributed in a wide frequency range and is affected by noise. It contains a lot of redundant information, and the damage features in the signal are easily submerged; in the prior art The damage recognition model based on 1D-CNN cannot meet the needs of modern civil engineering structure health monitoring

Images