CN-121994383-A - Ear plate local stress high spatial resolution monitoring method and system based on ultra-weak grating spectrum shape demodulation

CN121994383ACN 121994383 ACN121994383 ACN 121994383ACN-121994383-A

Abstract

The invention discloses an ear plate local stress high spatial resolution monitoring method and system based on ultra-weak grating spectrum shape demodulation, the system comprises an optical fiber sensing array, a light source module, a detection module, a data acquisition and physical information neural network inversion module and a display alarm module. The optical fiber sensing array consists of N ultra-weak gratings distributed in an ear plate stress concentration area and used for sensing local strain, the light source and detection module adopts a time-division and wavelength-division multiplexing strategy to realize synchronous monitoring of multiple ear plates, the data acquisition and inversion module acquires stress distribution along the length direction of the ultra-weak gratings based on a physical information neural network inversion model, and the display alarm module realizes local stress threshold alarm of the ear plates. The invention can realize high-precision and real-time monitoring of local stress of the metal connecting nodes such as the ear plates, is suitable for long-term on-line health monitoring of large-scale structure connecting areas such as bridge guy ropes, steel truss girders, cable towers and the like, and provides technical support for structural fatigue early warning and safety evaluation.

Inventors

LIU WEIKANG
GE HUIBIN
LUO YAOZHI
SHEN YANBIN

Assignees

浙江大学长三角智慧绿洲创新中心

Dates

Publication Date: 20260508
Application Date: 20251224

Claims (8)

1. The method for monitoring the local stress of the ear plate with high spatial resolution based on the demodulation of the ultra-weak grating spectrum shape is characterized by comprising the following steps of: (1) Arranging an optical fiber sensing array in a stress concentration area of the ear plate, wherein the optical fiber sensing array comprises N ultra-weak gratings; (2) Injecting broadband incident light into the optical fiber sensing array, and accessing each ultra-weak grating through a time division multiplexing and wavelength division multiplexing strategy to obtain a reflection distortion spectrum of the ultra-weak grating; (3) Finite element simulation is carried out on the ear plate, strain distribution data of the surface of the ear plate are obtained and analyzed, and a typical strain gradient mode is obtained; (4) Based on a typical strain gradient mode of the surface of the ear plate, a modified transmission matrix model is adopted, grating discrete unit lengths are divided according to preset spatial resolution, and an ultra-weak grating distortion spectrum shape and a corresponding high spatial resolution strain data set under different strain gradient modes are obtained through calculation; (5) Constructing a neural network inversion model based on physical information, constructing a loss function containing strain gradient constraint, and training and parameter tuning the model by adopting the data set obtained in the step (4) to obtain a trained neural network inversion model; (6) Inputting the real ultra-weak grating reflection distortion spectrum obtained in the step (2) into a trained neural network inversion model to obtain strain distribution along the length direction of the ultra-weak grating; (7) And inputting the strain distribution into a constitutive model of the material used by the ear plate to obtain the stress distribution of the ear plate, and realizing high-spatial resolution stress monitoring of the stress concentration area of the ear plate.
2. The method for monitoring the local stress of the ear plate with high spatial resolution based on the demodulation of the ultra-weak grating spectrum according to claim 1, wherein in the step (4), the modified transmission matrix model considers the non-uniform distribution of the ultra-weak grating period, the effective refractive index gradient change and the fiber cladding strain transfer effect to improve the distortion spectrum fitting precision, wherein the local coupling relation along the axial direction of the ultra-weak grating satisfies the following modified transmission matrix equation: ; Wherein, the And Respectively a forward envelope field amplitude and a reverse envelope field amplitude along the light propagation direction; an amount of propagation constant mismatch for position dependence; as a result of the coupling coefficient, Modulating the web for an effective refractive index; representing a non-uniform ultra-weak grating period affected by the strain field; Is the effective photoelastic coefficient of the optical fiber; In order to consider the strain transfer effect of the fiber cladding, the strain along the axial direction of the fiber satisfies the following conditions: Wherein In order to be a coefficient of strain transfer, Is the true strain of the ear plate surface.
3. The method for monitoring the local stress of the ear plate with high spatial resolution based on the demodulation of the ultra-weak grating spectrum shape according to claim 1, wherein in the step (5), the neural network inversion model based on the physical information comprises a double-branch structure, wherein a first branch is a convolution neural network and is used for extracting the multi-scale spectrum shape characteristics of the ultra-weak grating distortion spectrum, and a second branch is a physical constraint layer based on a strain gradient rule and is used for realizing the real physical mapping of the spectrum shape and the strain.
4. The method for monitoring the local stress of the ear plate with high spatial resolution based on the demodulation of the ultra-weak grating spectrum shape according to claim 1, wherein in the step (5), a loss function containing a strain gradient constraint is constructed, specifically: ; Wherein, L rec is a strain reconstruction error term, which is used for constraining the deviation between the output strain value of the model and the actually measured strain value, and the calculation formula is as follows: ; the method is used for limiting the spatial change rate of a model predicted strain field for a strain gradient constraint term, improving the smoothness and physical consistency of an inversion result, and the calculation formula is as follows: ; for regularization term, used for suppressing model overfitting, controlling network parameter complexity, its calculation formula is: ; For the predicted strain value of the neural network, As the measured strain value of the corresponding location, As a trainable parameter of the neural network, 、、 The weight coefficients of the loss terms are respectively obtained.
5. An ear plate local stress high spatial resolution monitoring system based on ultra-weak grating spectrum shape demodulation, which is characterized by comprising: The optical fiber sensing array comprises N ultra-weak gratings, and each ultra-weak grating is arranged in an ear plate stress concentration area and used for sensing local stress; the light source module is used for injecting broadband incident light into the sensing optical fiber array; The detection module is matched with the light source module, and accesses each ultra-weak grating through a time division multiplexing and wavelength division multiplexing strategy so as to acquire the reflection distortion spectrum of the ultra-weak grating; The data acquisition and physical information neural network inversion module is used for acquiring ultra-weak grating spectrum shape data and obtaining local stress distribution at ultra-weak grating measuring points on the earplates based on inversion of the ultra-weak grating spectrum shape data and the earplate material constitutive model; And the display alarm module is used for displaying stress distribution curves of the surfaces of the earplates and carrying out abnormal alarm when the local stress indexes of the earplates exceed a threshold value.
6. The ultra-weak grating spectral shape demodulation-based ear plate local stress high spatial resolution monitoring system according to claim 5, wherein in the optical fiber sensing array, the reflectivity of the ultra-weak grating is smaller than-45 dB, the large capacity multiplexing on a single optical fiber is realized, and a variable space multiplexing design is adopted to adapt to different spaces of adjacent ear plates.
7. The ultra-weak grating spectrum demodulation-based ear plate local stress high spatial resolution monitoring system according to claim 5, wherein the light source module adopts a broadband high-stability ASE light source, and the detection module adopts an integratable InGaAs photoelectric detector.
8. The ultra-weak grating spectral shape demodulation-based high spatial resolution monitoring system for local stress of an ear panel according to claim 5, wherein in the display alarm module, the local stress index of the ear panel is defined as: ; Wherein, the For maximum stress within a single ultra-weak grating gauge length, For the stress of each sub-segment grating within a single ultra-weak grating gauge, i=1, 2, 3..n, n is the number of single ultra-weak grating segments.

Description

Ear plate local stress high spatial resolution monitoring method and system based on ultra-weak grating spectrum shape demodulation Technical Field The invention relates to the field of structural health monitoring and optical fiber sensing, in particular to an ear plate local stress high spatial resolution monitoring method and system based on ultra-weak grating spectrum shape demodulation. Background The lug plates are key stress nodes for connecting the inhaul cable and the girder in the steel structure bridge and the cable tower, the stress distribution is complex, the local stress concentration is obvious, and fatigue cracks or early damage easily occur. The traditional monitoring means, such as a resistance strain gauge, a vibrating wire strain gauge and a fiber bragg grating strain gauge, can only acquire the average stress in the gauge length range of the sensor, and cannot reflect the real stress distribution characteristics in the gauge length, so that the high-precision high-spatial resolution monitoring requirement on the local stress concentration area cannot be met. The high spatial resolution optical fiber sensing technology (such as optical frequency domain reflection technology, OFDR) can acquire stress distribution information in a monitoring gauge length with millimeter-level spatial resolution, but the spatial resolution and the sensing distance are limited, so that synchronous monitoring of the multi-ear plate area is difficult to realize. In addition, the technology relies on a high-stability narrow-linewidth light source, the system cost is high, if a low-cost light source is adopted, the signal to noise ratio is required to be improved through multiple times of average, and the monitoring instantaneity is insufficient, so that the practical application of the technology in an engineering structure is limited. In recent years, the ultra-weak fiber Bragg grating provides a new technical path for monitoring the large-scale and dense measuring points of the structure by virtue of the low reflectivity and high-density multiplexing characteristic. However, similar to the traditional sensing means, the ultra-weak grating technology still can only acquire the average stress in the monitoring scale distance, can not analyze the continuous stress distribution in the scale distance, and under the action of strong stress gradient, the reflection spectrum is easy to generate obvious distortion, so that the traditional peak tracking method is invalid, and the real stress distribution along the length direction of the grating is difficult to accurately recover, thereby limiting the inversion and accurate monitoring of the high spatial resolution stress in the complex stress field. In order to solve the problem, scholars at home and abroad propose various strain demodulation methods based on fiber grating spectrum shapes. For example, literature "Simultaneous restoration of a non-smooth strain distribution and temperature of an optical fiber Bragg grating based on its intensity spectrum"（2025） adopts a transmission matrix model and combines a Nelder-Mead nonlinear optimization algorithm to realize simultaneous inversion of strain and temperature distribution of the fiber grating along the length direction, but single demodulation time is long, and real-time monitoring requirements of actual engineering are difficult to meet. In addition, for example, in the chinese patent document with publication No. CN118009903a, it is proposed that an optical fiber grating is equivalent to a plurality of sub-gratings, and non-uniform strain distribution along the length direction of the grating is obtained by combining a transmission matrix with a deep learning or fitting method, but the method is completely dependent on data driving, lacks physical constraint, and has a problem of overfitting; and strain gradient constraint or physical consistency constraint is not introduced, so that the smoothness and rationality of the inversion result are insufficient; the method is only suitable for the field of shape sensing of the conventional reflectivity fiber bragg grating, cannot be expanded to an ultra-weak grating array, and cannot realize high spatial resolution reconstruction of an ear plate stress field. Therefore, a high-precision monitoring technology combining spectral physical modeling and deep learning inversion is needed, continuous strain demodulation and stress field reconstruction integration based on ultra-weak grating spectral patterns can be realized, so that non-uniform stress of key structural nodes such as an ear plate and the like is monitored in a high spatial resolution, real-time and visual manner, and reliable technical support is provided for health assessment and life prediction of a complex structure. Disclosure of Invention In order to solve the problem that the uneven stress of the earplates is difficult to monitor with high spatial resolution in structural health monitoring, the inventio