Single fiber grating and neural network based structural damage position identification system

Abstract

The invention discloses a single fiber grating and neural network based structural damage position identification system which comprises a broadband light source, a fiber isolator, a tunable F-P filter, an erbium-doped fiber amplifier, a fiber circulator, a fiber grating sensing array, an unbalanced M-Z interferometer, a photoelectric detector array, a data collection card, a data processing unitand a signal generation module. One fiber grating sensor in the fiber grating sensing array is mounted in each structure area to be detected, is used to collect a vibration response signal under excitation of an active signal of the structure; and a wavelet decomposition technology is used to decompose the collected signal to obtain a characteristic vector including damage position information, the characteristic vector is input to the neural network for training, and the structural damage position can be identified accurately via the trained neural network. The system cost and the sensor network complexity are reduced, and the correct identification rate can reach 96%.

Description

technical field [0001] The invention relates to a system for identifying and monitoring the damaged position of a plate structure, in particular to a system for identifying and monitoring the damaged position of a plate structure where traditional electronic sensors are not suitable. Background technique [0002] The healthy and stable operation of the aerospace equipment structure and the safety and stability of the power system equipment structure are of great significance. In order to reduce economic losses and avoid casualties, a complete structural health monitoring system must be established. There are many shortcomings based on traditional electromagnetic sensors, and the structural damage identification system based on fiber gratings has many advantages that traditional sensors cannot match. The best choice for domain structural damage identification. However, the currently used FBG damage location identification method often requires three or more sensors to determ...

AI Technical Summary

Problems solved by technology

However, the currently used FBG damage location identification method usually requires three or more sensors to determine whether there is damage in an area, which greatly increases the cost of the system and the complexity of the sensor network, and the FBG wavelength demodulation system constitutes Complicated, slow demodulation speed, and complex damage location recognition algorithm

Images