Substrate deformation measurement system adaptive fitting method and system based on fiber bragg grating

Abstract

The invention provides a self-adaptive fitting method and system for a substrate deformation measurement system based on a fiber bragg grating (FBG), and the method comprises the steps: reading an output signal of an FBG sensor network, respectively collecting the wavelength data of a grating sensor and the stress data of a stress sensor at the same position as the grating sensor, and constructing a stress-wavelength data set; fBG data points with most sensitive deformation response are automatically identified, and a dynamic sensitive point set is formed and then preprocessed; a set needing to be continuously monitored is analyzed and determined, a specific point is selected to execute an adaptive segmentation fitting algorithm, segmentation points and fitting polynomial coefficients of all subintervals are output, and a final fitting model is constructed; according to the sub-interval of the wavelength value measured by the FBG sensor, calling a sub-interval fitting model, calculating a corresponding stress value, and further obtaining the deformation quantity of the measured object. According to the invention, the contradiction between the monitoring efficiency and precision of the traditional uniform layout in a complex stress scene is solved, and the technical path and the application target have remarkable industry uniqueness.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of optical fiber sensing, in particular to a substrate deformation measurement system adaptive fitting method and system based on fiber Bragg grating. BACKGROUND

[0002] Optical fiber sensing technology is continuously driving the intelligent process of various industries. Among them, fiber Bragg grating (FBG) sensors, with their unique advantages of easy connection, low loss, excellent spectral characteristics and high reliability, play a crucial role in the structural health monitoring of infrastructure and industrial equipment. From the stability monitoring of large building structures such as bridges, tunnels, dams and power stations, to the operation state monitoring of key facilities such as oil tanks and mechanical equipment, FBG sensors accurately capture the drift of Bragg wavelength, accurately interpret the changes of temperature, strain, vibration, pressure and other physical quantities, and provide core data support for ensuring the safe and stable operation of facilities.

[0003] At the same time, optical measurement technology, with its non-contact, high sensitivity and high precision, has become an indispensable means in scientific research and industrial production, and the pursuit of high-precision and high-reliability measurement methods continues to deepen. However, the application of FBG sensors faces significant challenges: the nonlinear relationship between wavelength and stress requires high-precision nonlinear fitting during the measurement process to meet the stringent precision requirements under complex working conditions.

[0004] The rapid development of aerospace technology has made satellite systems increasingly critical in the fields of Earth observation, communication, navigation, etc. The extreme space environment faced by satellites (such as high vacuum, strong radiation, and severe temperature changes) puts high demands on device performance. In this context, applying FBG sensors to satellite structural health monitoring and environmental parameter measurement is of great significance. However, the complexity of the space environment significantly amplifies the impact of noise interference on FBG sensors - FBG spectrum is extremely susceptible to distortion, leading to measured wavelength deviation, severely damaging the stress-wavelength fitting accuracy, and threatening the normal operation of satellite systems and the accuracy of scientific data.

[0005] Traditional substrate deformation measurement systems mostly use fixed sensor networks, which can provide comprehensive data, but have inherent problems such as complex calculation and low processing efficiency, and are difficult to meet the application requirements of real-time monitoring and rapid response. The current mainstream FBG stress-wavelength fitting algorithm is mainly based on least square polynomial piecewise fitting. However, this kind of method has significant limitations in application: the determination of the segmentation point usually depends on artificial experience or intuitive judgment of the output characteristic curve, which lacks objectivity and universality, and it is difficult to achieve optimal segmentation. Especially when facing complex and variable scenes such as extreme space environment of satellite systems, the accuracy and stability of traditional fitting methods face severe challenges. Therefore, it is urgent to develop a more intelligent, accurate and adaptive fitting algorithm to ensure the reliable application of FBG sensors in complex environments and promote the in-depth development of optical fiber sensing technology in satellite systems and other key fields. SUMMARY

[0006] In view of the defects in the prior art, the purpose of the present application is to provide a substrate deformation measurement system adaptive fitting method based on fiber Bragg grating.

[0007] According to the substrate deformation measurement system adaptive fitting method based on fiber Bragg grating provided by the present application, the method comprises the following steps:

[0008] Data acquisition step: reading the output signal of the FBG sensor network, and using the fiber Bragg grating demodulator and the stress measurement system to respectively collect the wavelength data of the grating sensor and the stress data of the stress sensor pasted at the same position as the grating sensor, and to construct a stress-wavelength data set;

[0009] Dynamic sensitive point identification step: applying an advanced signal processing algorithm to automatically identify the FBG data points most sensitive to deformation response from the output signal to form a dynamic sensitive point set;

[0010] Signal preprocessing step: preprocessing the output signal and the stress-wavelength data set;

[0011] Data analysis and processing step: analyzing the preprocessed data, determining the sensitive point set to be continuously monitored, selecting a specific point in the sensitive point set to execute an adaptive piecewise fitting algorithm, outputting the piecewise point and the fitting polynomial coefficient of each sub-interval, and constructing a final fitting model;

[0012] Deformation measurement and calculation step: according to the sub-interval where the wavelength value measured by the FBG sensor is located, the corresponding fitting model of the sub-interval is called to calculate the corresponding stress value, and then the deformation variable of the measured object is obtained.

[0013] Preferably, the FBG sensor network is composed of a plurality of fiber Bragg gratings (FBGs) distributed on the substrate.

[0014] Preferably, the dynamic sensitive point set reflects the deformation characteristic core region of the structure under different stress states.

[0015] The preprocessing includes smoothing filtering, normalization and gross error rejection.

[0016] The smoothing filtering uses an adaptive filter to suppress environmental noise.

[0017] The gross error rejection uses a stress-wavelength fitting method to reject abnormal data points with residual error greater than a set threshold, wherein an adaptive threshold determination algorithm dynamically adjusts the fitting residual threshold and error threshold according to the characteristics of real-time data to adapt to different measurement conditions and environmental changes.

[0018] Preferably, the analysis includes multi-scale analysis, model fusion, system identification and state estimation, physical model fusion, and multi-sensor data fusion.

[0019] The multi-scale analysis includes combining wavelet transform or Fourier transform to analyze FBG data at different scales, separate different frequency signal components, and scale fitting to improve accuracy.

[0020] The model fusion includes combining various fitting models such as polynomials, exponents, logarithms, etc., dynamically selecting the optimal model for fitting according to data characteristics through a weighting or switching mechanism.

[0021] The system identification and state estimation include treating the stress-wavelength relationship of the FBG sensor as a nonlinear dynamic system, applying state estimation techniques including Kalman filtering to track and predict system state, and realizing real-time fitting in dynamic environment.

[0022] The physical model fusion includes establishing a physical model of the FBG sensor based on material mechanics and optical principles, combining experimental data, and implementing fitting through parameter estimation and optimization algorithms.

[0023] The multi-sensor data fusion includes fusing temperature, vibration and other data, and using multi-sensor information fusion technology for fitting.

[0024] Preferably, the final results are visually displayed through a host computer interface.

[0025] The adaptive piecewise fitting algorithm includes piecewise determination, model selection, parameter initialization, adaptive fitting, and model evaluation.

[0026] The piecewise determination includes determining piecewise points based on stress-wavelength data set distribution characteristics, and dividing the entire stress range into several subintervals.

[0027] The model selection includes selecting or combining appropriate nonlinear fitting models for each subinterval data characteristics.

[0028] The parameter initialization comprises initializing selected model parameters, and the parameters of the model fitted in each sub-interval are initialized respectively;

[0029] The adaptive fitting comprises optimizing and adjusting the model parameters of each sub-interval by using an optimization algorithm to minimize the fitting error;

[0030] The model evaluation comprises evaluating whether the performance of each sub-interval model meets the preset accuracy requirement by using indicators such as mean square error and mean absolute error, and the interval that does not meet the requirement needs to return to the parameter initialization step for re-optimization, and all the intervals that meet the requirement are determined as the final segmented fitting model.

[0031] The adaptive fitting system of the substrate deformation measurement system based on the fiber Bragg grating provided by the application comprises:

[0032] The data acquisition module reads the output signal of the FBG sensor network, and collects wavelength data of the grating sensor and stress data of the stress sensor pasted at the same position as the grating sensor by using a fiber grating demodulator and a stress measurement method, and constructs a stress-wavelength data set;

[0033] The dynamic sensitive point identification module applies an advanced signal processing algorithm to automatically identify the FBG data points most sensitive to deformation response from the output signal to form a dynamic sensitive point set;

[0034] The signal preprocessing module pre-processes the output signal and the stress-wavelength data set;

[0035] The data analysis and processing module analyzes the pre-processed data, determines the sensitive point set that needs to be continuously monitored, selects a specific point in the sensitive point set to execute an adaptive segmented fitting algorithm, outputs the segmented points and the fitting polynomial coefficients of each sub-interval, and constructs a final fitting model;

[0036] The deformation measurement and calculation module calls the fitting model corresponding to the sub-interval where the wavelength value measured by the FBG sensor is located, calculates the corresponding stress value, and then obtains the deformation amount of the measured object.

[0037] Preferably, the FBG sensor network is composed of a plurality of fiber Bragg gratings (FBGs) distributed on the substrate.

[0038] Preferably, the dynamic sensitive point set reflects the core area of deformation characteristics of the structure under different stress states.

[0039] The preprocessing comprises smoothing filter processing, normalization processing, and gross error rejection;

[0040] The smoothing filter processing adopts an adaptive filter to suppress environmental noise;

[0041] The coarse error rejection adopts a stress-wavelength fitting system to reject abnormal data points with residuals greater than a set threshold, wherein an adaptive threshold determination algorithm dynamically adjusts the fitting residual threshold and error threshold according to the characteristics of real-time data to adapt to different measurement conditions and environmental changes.

[0042] Preferably, the analysis includes multi-scale analysis, model fusion, system identification and state estimation, physical model fusion, multi-sensor data fusion;

[0043] The multi-scale analysis includes analyzing FBG data at different scales, separating different frequency signal components, and fitting at different scales to improve accuracy, including wavelet transform or Fourier transform;

[0044] The model fusion includes combining various fitting models such as polynomials, exponents, logarithms, etc., dynamically selecting the optimal model for fitting according to data characteristics through a weighting or switching mechanism;

[0045] The system identification and state estimation include considering the stress-wavelength relationship of the FBG sensor as a nonlinear dynamic system, applying state estimation techniques including Kalman filtering to track and predict system state, and realizing real-time fitting in dynamic environment;

[0046] The physical model fusion includes establishing a physical model of the FBG sensor based on material mechanics and optical principles, combining experimental data, and implementing fitting through parameter estimation and optimization algorithms;

[0047] The multi-sensor data fusion includes fusing temperature, vibration and other sensor data, and performing fitting with the help of multi-sensor information fusion technology.

[0048] Preferably, the final results are visually displayed through a host computer interface;

[0049] The adaptive piecewise fitting algorithm includes piecewise determination, model selection, parameter initialization, adaptive fitting, and model evaluation;

[0050] The piecewise determination includes determining the segmentation points based on the distribution characteristics of the stress-wavelength data set, and dividing the entire stress range into several subintervals;

[0051] The model selection includes selecting or combining appropriate nonlinear fitting models for each subinterval data characteristics;

[0052] The parameter initialization includes initializing the parameters of the selected model, and the parameters of the fitting model in each subinterval are initialized separately;

[0053] The adaptive fitting includes using optimization algorithms to optimize and adjust the model parameters of each subinterval, minimizing the fitting error;

[0054] The model evaluation includes using mean square error, mean absolute error and other indicators to evaluate whether the performance of each sub-interval model meets the preset accuracy requirement, and the interval that does not meet the requirement needs to return to the parameter initialization module for re-optimization, and all of them meet the requirement to determine the final segmented fitting model.

[0055] Compared with the prior art, the present application has the following beneficial effects:

[0056] 1. The present application solves the contradiction between monitoring efficiency and accuracy of traditional uniform layout in complex stress scenarios through the system-level innovation of adaptive layout FBG network dynamic focusing high strain area, modular data processing architecture and three-dimensional real-time visualization interface, and the technical path and application target have significant industry uniqueness.

[0057] 2. The present application can dynamically select sensitive points to preprocess the original data under the premise of ensuring monitoring accuracy, remove gross error points, and adaptively determine the segmentation point according to the set fitting residual threshold and polynomial function form, realize adaptive high-precision fitting of stress-wavelength, and greatly improve the accuracy and stability compared with the traditional fitting method.

[0058] 3. The present application is not only suitable for structural health monitoring of satellite optical load, but also can be widely applied to other fields requiring high-precision deformation measurement, such as aerospace, civil engineering, mechanical manufacturing, etc. BRIEF DESCRIPTION OF DRAWINGS

[0059] Other features, objects and advantages of the present application will become more apparent from the following detailed description of non-limiting embodiments, made with reference to the accompanying drawings:

[0060] Figure 1 It is a working method flowchart of the present application;

[0061] Figure 2 It is a system composition schematic diagram of the present application;

[0062] Figure 3 It is a fiber grating sensor layout schematic diagram in the present application;

[0063] Figure 4 It is a filtering schematic diagram in the present application;

[0064] Figure 5 It is a stress-wavelength original data distribution diagram collected in the embodiment;

[0065] Figure 6 It is a stress-wavelength relationship curve after fitting by the method of the present application and the original data comparison diagram. DETAILED DESCRIPTION

[0066] The application will be described in detail below with specific examples. The following examples will help those skilled in the art to further understand the application, but do not limit the application in any form. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the application. These are within the scope of protection of the application.

[0067] The application does not change the position of the sensor, but selectively uses part of the FBG according to the existing sensor, focuses on monitoring some sensitive key points, which change with the change of external strain, which is the embodiment of self-adaptation. While ensuring accuracy, it improves the operation speed. Focus on the wavelength change of FBG sensor under stress for high-precision, dynamic adaptive fitting to realize accurate measurement of deformation.

[0068] Example 1

[0069] According to the substrate deformation measurement system based on fiber grating provided by the application, as shown in Figure 1 and 2 , comprising:

[0070] Data acquisition step: reading the output signal of the FBG sensor network, and using the fiber grating demodulator and the stress measurement system to collect the wavelength data of the grating sensor and the stress data of the stress sensor pasted at the same position as the grating sensor, and constructing a stress-wavelength data set. The FBG sensor network is composed of a plurality of fiber Bragg gratings (FBG) distributed on the substrate. Specifically, the demodulation device collects the spectrum data of the FBG sensor, and after noise filtering and standardization processing, the accuracy and reliability of the data are improved.

[0071] Dynamic sensitive point identification step: applying advanced signal processing algorithm to automatically identify the FBG data points most sensitive to deformation from the output signal to form a dynamic sensitive point set. The set reflects the core area of deformation characteristics of the structure under different stress states.

[0072] Signal preprocessing step: preprocessing the output signal and stress-wavelength data set. The preprocessing includes smoothing filter processing, normalization processing and gross error rejection. The smoothing filter processing adopts an adaptive filter to suppress environmental noise and improve data signal-to-noise ratio and fitting accuracy. The gross error rejection adopts a stress-wavelength fitting method to reject abnormal data points with residual error greater than a set threshold, wherein an adaptive threshold determination algorithm dynamically adjusts the fitting residual threshold and error threshold according to the characteristics of real-time data to adapt to different measurement conditions and environmental changes.

[0073] Data analysis and processing step: analyze the preprocessed data and determine the sensitive point set that needs to be continuously monitored. Select a specific point in the sensitive point set to perform the adaptive piecewise fitting algorithm, output the segmentation points and the fitting polynomial coefficients of each sub-interval, and construct the final fitting model. It should be noted that the monitoring process updates the sensitive point set in real time, and when the external conditions change, the sensitive points are re-identified and the monitoring strategy is adjusted. Only the selected sensitive points are subjected to data acquisition and processing.

[0074] The analysis includes multi-scale analysis, model fusion, system identification and state estimation, physical model fusion, multi-sensor data fusion, etc. Multi-scale analysis includes the use of tools such as wavelet transform or Fourier transform to analyze FBG data at different scales, separate different frequency signal components, and scale fitting to improve accuracy. Model fusion includes combining polynomial, exponential, logarithmic and other fitting models, using weighting or switching mechanisms to dynamically select the optimal model for fitting based on data characteristics to adapt to different application scenarios and data characteristics. System identification and state estimation include treating the stress-wavelength relationship of FBG sensors as a nonlinear dynamic system, applying state estimation techniques such as Kalman filtering to track and predict system state, and achieving real-time fitting in dynamic environments. Physical model fusion includes establishing a physical model of the FBG sensor based on material mechanics and optical principles, combining experimental data, and achieving high-precision fitting through parameter estimation and optimization algorithms. Multi-sensor data fusion includes fusing temperature, vibration and other sensor data to improve fitting accuracy and system robustness.

[0075] Deformation measurement and calculation step: according to the sub-interval where the wavelength value measured by the FBG sensor is located, the corresponding fitting model of the sub-interval is called to calculate the corresponding stress value, and then the deformation variable of the measured object is obtained. The final result is visualized through the host computer interface.

[0076] The adaptive segmented fitting algorithm includes segmentation determination, model selection, parameter initialization, adaptive fitting and model evaluation. The segmentation determination includes determining segmentation points based on stress-wavelength dataset distribution characteristics (such as data curvature variation, cluster analysis), dividing the entire stress range into several subintervals, and ensuring that the nonlinearity characteristics of the data in each subinterval are relatively consistent. The model selection includes selecting or combining appropriate nonlinear fitting models (such as polynomials, exponentials, Gaussian functions, etc.) for the data characteristics of each subinterval. Different models can be selected for different subintervals to better adapt to the characteristics of the data in each segment. The parameter initialization includes initializing the parameters of the selected model (randomly or based on experience to set the initial range), and initializing the parameters of the fitting model in each subinterval. The adaptive fitting includes using optimization algorithms (such as genetic algorithms, particle swarm optimization, simulated annealing) to optimize and adjust the model parameters in each subinterval, minimizing the fitting error. The subinterval fitting process is relatively independent, and collectively describes the overall stress-wavelength relationship. The model evaluation includes using mean square error (MSE), mean absolute error (MAE) and other indicators to evaluate whether the performance of each subinterval model meets the predetermined accuracy requirements. Intervals that do not meet the requirements need to return to the parameter initialization step for re-optimization; all meet the requirements to determine the final segmented fitting model.

[0077] The adaptive segmented fitting algorithm can dynamically adapt to the stress changes of the satellite optical load in complex space environment, and realize high-precision stress-wavelength relationship fitting. Adaptation ensures the accuracy and reliability of deformation monitoring under extreme conditions. Through high-precision data acquisition, preprocessing, and dynamic sensitive point identification and updating, the present application realizes real-time high-precision measurement of satellite optical load deformation, provides key technical support for satellite structural health monitoring, and helps to discover and handle potential structural problems in time. Because the present application has the characteristics of high precision, real-time, adaptability, etc., it is not only suitable for satellite optical load structural health monitoring, but also can be widely applied to other fields requiring high-precision deformation measurement, such as aerospace, civil engineering, mechanical manufacturing, etc.

[0078] Further, the adaptive fitting method of the substrate deformation measurement system based on fiber Bragg grating of the present application is described in detail as follows in combination with the drawings:

[0079] As Figure 1As shown, this figure is the flow chart of the adaptive fitting method of the deformation measurement system based on fiber Bragg grating (FBG) of the present application. First, data acquisition, using FBG sensor to measure the deformation of the measured object, to obtain the wavelength change data and stress value under stress, to form the stress-wavelength data set; then dynamic sensitive point identification, applying advanced signal processing algorithm, automatically identifying the FBG data points most sensitive to deformation response, forming a dynamic sensitive point set; then data preprocessing, removing outliers, smoothing filtering; then model selection, selecting or combining appropriate nonlinear fitting model according to the data characteristics of each subinterval of the data set; then parameter initialization, determining the initial value range of the parameters; then adaptive fitting, using optimization algorithm (such as genetic, particle swarm, simulated annealing algorithm) to adjust the parameters, minimizing the fitting error; then model evaluation, using mean square error, mean absolute error and other indicators to judge whether the model meets the accuracy requirements, if not, return to the adaptive fitting step, if yes, determine the final model; finally, in the deformation measurement and calculation, the wavelength value measured by the FBG sensor is substituted into the final model to calculate the stress value, realizing the accurate measurement of the deformation of the object. Specifically, each step is described as follows:

[0080] Data acquisition: using demodulation equipment to collect spectrum data from FBG sensor, and after noise filtering and standardization processing, using fiber grating demodulator and stress measurement system, wavelength data of grating sensor and stress data of stress sensor pasted at the same position are collected respectively, to construct stress-wavelength data set.

[0081] Dynamic sensitive point identification: based on advanced signal processing algorithm, automatically identifying the FBG data points most sensitive to deformation, forming a dynamic sensitive point set, reflecting the core area of deformation characteristics of the structure under different stress states.

[0082] Data preprocessing: preprocessing the collected data, using 3σ criterion to remove obvious abnormal data points, and using moving average method to smooth filter the data, to obtain the processed data.

[0083] Segmentation determination: curvature analysis is performed on the data to determine segmentation fitting in the stress value intervals [0, 5N], [5N, 10N], [10N, 15N]. The nonlinear change trends of the data in the three intervals are relatively obvious and different, and after segmentation, it is convenient to select appropriate fitting model.

[0084] Model selection: for the [0, 5N] interval, a quadratic polynomial function is selected as the fitting model λ=a0+a1σ+a2σ 2

[0085] ; for the [5N, 10N] interval, the combination of exponential function and linear polynomial function is selected λ=b0(e ^b1σ)+c0+c1σ;for [10N,15N] interval, select cubic polynomial function λ=d0+d1σ+d2σ 2 +d3σ 3 . Wherein a0, a1, a2, b0, b1, c0, c1, d0, d1, d2, d3 are parameters to be optimized.

[0086] Parameter initialization: randomly initialize the fitting model parameters of each interval, and set reasonable value range according to experience. For example, for the quadratic polynomial function parameters of the [0, 5N] interval, the value range of a0 is [1500, 1600], the value range of a1 is [-10, 10], and the value range of a2 is [0.5, 1.5]; other interval parameters are also similarly set to initialize the range.

[0087] Adaptive fitting: the particle swarm optimization algorithm is used to optimize and adjust the parameters of the fitting model of each interval. The population size of the particle swarm is set to 50, the maximum iteration number is set to 200, the learning factor c1=c2=2, and the inertia weight ω is linearly decreased from 0.9 to 0.4. The fitting model of the three intervals is iteratively calculated respectively to find the parameter value of each interval with the smallest fitting error.

[0088] Model evaluation: the mean square error (MSE) is used as the model evaluation index to calculate the MSE of the fitting model of each interval. After multiple iterations of optimization, the MSE of the [0, 5N] interval is 0.015nm 2 , the MSE of the [5N, 10N] interval is 0.02nm 2 , and the MSE of the [10N, 15N] interval is 0.018nm 2 , all of which meet the preset accuracy requirement (the preset accuracy is MSE<0.05nm 2 ), and the final segmented fitting model is determined.

[0089] Deformation measurement and calculation: during actual structure deformation monitoring, according to the sub-interval where the wavelength value measured by the FBG sensor is located, the corresponding fitting model of the interval is called to calculate the corresponding stress value, so as to realize accurate measurement of the structure deformation. As shown in Figure 6 , the comparison of the stress-wavelength relationship curve after segmented fitting by the method of the present application and the original data verifies the effectiveness and accuracy of the method of the present application.

[0090] As shown in Figure 2As shown in the diagram, the system composition diagram clearly illustrates the core components and their workflow. The FBG sensor network section displays multiple fiber Bragg grating (FBG) sensors distributed on the substrate, used to sense changes in external physical quantities and generate response signals; the data acquisition module reads the FBG sensor signals and converts the optical signals into processable electrical signals; the signal preprocessing module improves the quality and stability of the raw data through filtering (denoising) and normalization; the data analysis and processing module executes core algorithms, including in-depth data analysis, fitting algorithms (establishing physical quantity models), and adaptive layout algorithms; the host computer interface serves as the output, providing data visualization results and alarm information to assist operators in status monitoring and decision-making.

[0091] like Figure 3 As shown in the diagram, the fiber optic grating sensor layout diagram details the distribution of FBG sensors on the substrate. The diagram clearly identifies the exact mounting point of each FBG sensor, crucial for accurately sensing local deformation. Each sensor is assigned a unique number for easy data traceability and management. Simultaneously, the diagram clearly displays the morphological characteristics of the supporting structure, aiding in understanding the relationship between the sensor layout and the substrate structure. Furthermore, key areas susceptible to load and exhibiting significant strain are identified and highlighted in the diagram through finite element analysis (FEA). Focusing on monitoring these points effectively captures crucial structural stress information, providing core support for structural health assessment.

[0092] like Figure 4 As shown in the figure, this graph presents the results of processing the original data using moving average filtering, Butterworth filtering, and median filtering, respectively. The horizontal axis typically represents stress variations, while the vertical axis represents the range of filtered data values. Each of the three curves represents a different filtering result. The moving average filtering curve smooths the data sequence by using the mean of the data within a window, effectively reducing short-term fluctuations and noise interference in the data, making the data trend clearer. The Butterworth filtering curve selectively retains useful frequency components based on specific frequency response characteristics, improving data purity and stability. The median filtering curve uses the median of the data points as the output and performs exceptionally well when processing data containing impulse noise. By comparing and analyzing the advantages and applicable scenarios of each filtering method, a data foundation is laid for subsequent analysis, modeling, and precise measurement to select the optimal solution.

[0093] like Figure 5 As shown in the figure, this is a distribution map of the raw stress-wavelength data collected in the embodiment. The horizontal axis represents the applied stress value, and the vertical axis corresponds to the wavelength value measured by the FBG sensor. Each data point represents a high-precision measurement result, and its density and shape intuitively reveal the preliminary relationship and trend of stress and wavelength, providing core data basis for subsequent fitting analysis and model construction.

[0094] likeFigure 6 As shown in the figure, the figure is a comparison chart of stress-wavelength relationship curve after fitting by the method of the application and original data. The discrete points in the figure represent the measured stress-wavelength original data, which are recorded by accurately measuring instruments under different stress conditions, reflecting the real (although with discreteness) corresponding relationship. The smooth curve is the stress-wavelength relationship fitting curve obtained based on the adaptive fitting method of the application. The curve is generated based on original data analysis, preprocessing (including denoising and removing outliers), model optimization and parameter optimization, aiming to accurately capture the inherent change law. The fitting curve closely follows the trend of the original data distribution, effectively reduces the uncertainty caused by data discreteness, proves the superiority of the method in dealing with the nonlinear relationship between stress and wavelength, overcomes the limitations of traditional methods, provides a reliable basis for high-precision deformation measurement, and significantly improves the system accuracy and stability.

[0095] The application aims to provide an adaptive fitting method for a fiber Bragg grating-based deformation measurement system, in particular, an adaptive fitting of stress-wavelength, which can dynamically select sensitive points to preprocess the original data and remove gross error points under the premise of ensuring monitoring accuracy, and can adaptively determine the segmentation points according to the set fitting residual threshold and polynomial function form, to realize adaptive high-precision fitting of stress-wavelength.

[0096] Embodiment 2

[0097] The application further provides an adaptive fitting system for a fiber grating-based substrate deformation measurement system, which can be realized by executing the process steps of the adaptive fitting method for the fiber grating-based substrate deformation measurement system, that is, the adaptive fitting method for the fiber grating-based substrate deformation measurement system can be understood by those skilled in the art as a preferred embodiment of the adaptive fitting system for the fiber grating-based substrate deformation measurement system.

[0098] According to the adaptive fitting system for the fiber grating-based substrate deformation measurement system provided by the application, the adaptive fitting system for the fiber grating-based substrate deformation measurement system comprises:

[0099] The data acquisition module reads the output signal of the FBG sensor network, and uses a fiber grating demodulator and a stress measurement method to respectively collect wavelength data of the grating sensor and stress data of the stress sensor pasted at the same position as the grating sensor, and constructs a stress-wavelength data set. The FBG sensor network is composed of a plurality of fiber Bragg gratings (FBGs) distributed on the substrate.

[0100] Dynamic sensitive point identification module: applying advanced signal processing algorithms to automatically identify the FBG data points most sensitive to deformation response from the output signal, forming a dynamic sensitive point set; the dynamic sensitive point set reflects the deformation characteristic core area of the structure under different stress states.

[0101] Signal preprocessing module: preprocessing the output signal and stress-wavelength data set; the preprocessing includes smoothing filter processing, normalization processing and gross error rejection; the smoothing filter processing uses an adaptive filter to suppress environmental noise; the gross error rejection uses a stress-wavelength fitting system to reject abnormal data points with residual error greater than a set threshold, wherein an adaptive threshold determination algorithm dynamically adjusts the fitting residual threshold and error threshold according to the characteristics of real-time data to adapt to different measurement conditions and environmental changes.

[0102] Data analysis and processing module: analyze the preprocessed data, determine the sensitive point set that needs to be continuously monitored, select specific points in the sensitive point set to perform adaptive segmented fitting algorithm, output the segmented points and the fitting polynomial coefficients of each sub-interval, and construct the final fitting model; the analysis includes multi-scale analysis, model fusion, system identification and state estimation, physical model fusion, and multi-sensor data fusion; the multi-scale analysis includes combining wavelet transform or Fourier transform to analyze FBG data at different scales, separate different frequency signal components, and scale fitting to improve accuracy; the model fusion includes combining various fitting models such as polynomials, exponents, and logarithms, dynamically selecting the optimal model for fitting according to data characteristics through weighting or switching mechanism; the system identification and state estimation includes considering the stress-wavelength relationship of the FBG sensor as a nonlinear dynamic system, applying state estimation techniques including Kalman filter to track and predict system state, and realizing real-time fitting under dynamic environment; the physical model fusion includes establishing a physical model of the FBG sensor based on material mechanics and optical principles, combining experimental data, and realizing fitting through parameter estimation and optimization algorithm; the multi-sensor data fusion includes fusing temperature and vibration sensor data, and performing fitting with the help of multi-sensor information fusion technology.

[0103] The deformation measurement and calculation module: according to the sub-interval where the wavelength value measured by the FBG sensor is located, the corresponding fitting model of the sub-interval is called to calculate the corresponding stress value, and then the deformation of the measured object is obtained. The final result is visually displayed through the host computer interface; the adaptive segmented fitting algorithm includes segmentation determination, model selection, parameter initialization, adaptive fitting and model evaluation; the segmentation determination includes determining the segmentation point based on the stress-wavelength data set distribution characteristics, and dividing the entire stress range into several sub-intervals; the model selection includes selecting or combining appropriate nonlinear fitting models for the data characteristics of each sub-interval; the parameter initialization includes initializing the selected model parameters, and the parameters of the fitting model in each sub-interval are initialized respectively; the adaptive fitting includes using an optimization algorithm to optimize and adjust the model parameters of each sub-interval, and minimizing the fitting error; the model evaluation includes using the mean square error and the mean absolute error indicators to evaluate whether the performance of each sub-interval model meets the preset accuracy requirement, the interval that does not meet the requirement needs to return to the parameter initialization module for re-optimization, and all meet the requirement to determine the final segmented fitting model.

[0104] Those skilled in the art know that, in addition to implementing the system provided by the present application and each device, module and unit thereof in the form of pure computer readable program code, the system provided by the present application and each device, module and unit thereof can also be implemented in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers and embedded microcontrollers by logically programming the method steps to achieve the same functions. Therefore, the system provided by the present application and each device, module and unit thereof can be considered as a hardware component, and the devices, modules and units included therein for achieving various functions can also be considered as structures within the hardware component; the devices, modules and units for achieving various functions can also be considered as both software modules for implementing methods and structures within hardware components.

[0105] The specific embodiments of the present application are described above. It should be understood that the present application is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essential content of the present application. The embodiments of the present application and the features in the embodiments can be arbitrarily combined with each other without conflict.

Claims

1. A method for adaptive fitting of a fiber grating based substrate deformation measurement system, the method comprising: obtaining a plurality of fiber grating sensor data; determining a plurality of fiber grating sensor data fit parameters; and updating the plurality of fiber grating sensor data fit parameters based on the plurality of fiber grating sensor data. The method comprises the following steps: Data acquisition: reading the output signal of the FBG sensor network, and collecting the wavelength data of the FBG sensor and the stress data of the stress sensor pasted at the same position as the FBG sensor by using the fiber grating demodulator and the stress measurement system, to construct a stress-wavelength data set; Dynamic sensitive point identification: applying advanced signal processing algorithms to automatically identify the FBG data points most sensitive to deformation from the output signal, to form a dynamic sensitive point set; Signal preprocessing: preprocessing the output signal and the stress-wavelength data set; Data analysis and processing: analyzing the preprocessed data to determine the sensitive point set to be continuously monitored, selecting a specific point in the sensitive point set to perform an adaptive segmented fitting algorithm, outputting the segmented points and the fitting polynomial coefficients of each subinterval, and constructing a final fitting model; Deformation measurement and calculation: according to the subinterval where the wavelength value measured by the FBG sensor is located, calling the fitting model corresponding to the subinterval, calculating the corresponding stress value, and then obtaining the deformation of the measured object.

2. The FBG-based substrate deformation measurement system adaptive fitting method according to claim 1, wherein, The FBG sensor network is composed of a plurality of fiber Bragg gratings (FBGs) distributed on a substrate.

3. The FBG-based substrate deformation measurement system adaptive fitting method of claim 1, wherein, The dynamic sensitive point set reflects the core area of deformation characteristics of the structure under different stress states; The preprocessing includes smoothing filtering, normalization processing, and gross error rejection; The smoothing filtering uses an adaptive filter to suppress environmental noise; The gross error rejection uses a stress-wavelength fitting method to reject abnormal data points with a residual error greater than a set threshold, wherein an adaptive threshold determination algorithm dynamically adjusts the fitting residual threshold and the error threshold according to the characteristics of real-time data to adapt to different measurement conditions and environmental changes.

4. The FBG-based substrate deformation measurement system adaptive fitting method of claim 1, wherein, The analysis includes multi-scale analysis, model fusion, system identification and state estimation, physical model fusion, and multi-sensor data fusion; The multi-scale analysis includes combining wavelet transform or Fourier transform to analyze the FBG data at different scales, separate different frequency signal components, and perform scale-by-scale fitting to improve accuracy; The model fusion includes combining various fitting models such as polynomials, exponents, and logarithms, dynamically selecting the optimal model for fitting according to data characteristics through a weighting or switching mechanism; The system identification and state estimation include treating the stress-wavelength relationship of the FBG sensor as a nonlinear dynamic system, applying state estimation techniques including Kalman filtering to track and predict the system state, and realizing real-time fitting in a dynamic environment; The physical model fusion includes establishing a physical model of the FBG sensor based on material mechanics and optical principles, combining experimental data, and realizing fitting through parameter estimation and optimization algorithms; The multi-sensor data fusion includes fusing temperature and vibration sensor data and performing fitting with the help of multi-sensor information fusion technology.

5. The FBG-based substrate deformation measurement system adaptive fitting method of claim 1, wherein, The final result is visually displayed through a host computer interface; The adaptive segmented fitting algorithm includes segmentation determination, model selection, parameter initialization, adaptive fitting, and model evaluation; The segmentation determination includes determining segmentation points based on the distribution characteristics of the stress-wavelength data set, and dividing the entire stress range into several subintervals. The model selection comprises selecting or combining suitable nonlinear fitting models for the data characteristics of each sub-interval; The parameter initialization comprises initializing the selected model parameters, and the parameters of the fitting model in each sub-interval are initialized respectively; The adaptive fitting comprises optimizing and adjusting the model parameters of each sub-interval by using an optimization algorithm to minimize the fitting error; The model evaluation comprises evaluating whether the performance of each sub-interval model meets the preset accuracy requirement by using mean square error and mean absolute error indicators, returning to the parameter initialization step for re-optimization for the sub-intervals that do not meet the requirement, and determining the final segmented fitting model when all the sub-intervals meet the requirement.

6. A fiber grating based substrate deformation measurement system adaptive fitting system, characterized in that, It comprises: A data acquisition module: reading the output signal of the FBG sensor network, and collecting wavelength data of the grating sensor and stress data of the stress sensor pasted at the same position as the grating sensor by using a fiber grating demodulator and a stress measurement method, to construct a stress-wavelength data set; A dynamic sensitive point identification module: applying advanced signal processing algorithms to automatically identify the FBG data points most sensitive to deformation from the output signal to form a dynamic sensitive point set; A signal preprocessing module: preprocessing the output signal and the stress-wavelength data set; A data analysis and processing module: analyzing the preprocessed data to determine the sensitive point set to be continuously monitored, selecting a specific point in the sensitive point set to execute an adaptive segmented fitting algorithm, outputting the segmented points and the fitting polynomial coefficients of each sub-interval, and constructing a final fitting model; A deformation measurement and calculation module: calling the fitting model corresponding to the sub-interval where the wavelength value measured by the FBG sensor is located to calculate the corresponding stress value, and then obtaining the deformation amount of the measured object.

7. The FBG-based substrate deformation measurement system adaptive fitting system of claim 6, wherein, The FBG sensor network is composed of a plurality of fiber Bragg gratings (FBGs) distributed on a substrate.

8. The FBG-based substrate deformation measurement system adaptive fitting system of claim 6, wherein, The dynamic sensitive point set reflects the core area of deformation characteristics of the structure under different stress states; The preprocessing comprises smoothing filter processing, normalization processing, and gross error rejection; The smoothing filter processing uses an adaptive filter to suppress environmental noise; The gross error rejection uses the stress-wavelength fitting system to reject abnormal data points with a residual error greater than a set threshold, wherein an adaptive threshold determination algorithm dynamically adjusts the fitting residual error threshold and the error threshold according to the characteristics of real-time data to adapt to different measurement conditions and environmental changes.

9. The FBG-based substrate deformation measurement system adaptive fitting system of claim 6, wherein, The analysis comprises multiscale analysis, model fusion, system identification and state estimation, physical model fusion, and multi-sensor data fusion; The multiscale analysis comprises combining wavelet transform or Fourier transform to analyze the FBG data at different scales, separate different frequency signal components, and perform scale-by-scale fitting to improve accuracy; The model fusion comprises combining various fitting models such as polynomials, exponents, and logarithms, dynamically selecting the optimal model for fitting according to data characteristics through a weighting or switching mechanism; The system identification and state estimation comprise regarding the stress-wavelength relationship of the FBG sensor as a nonlinear dynamic system, applying state estimation techniques including Kalman filtering to track and predict the system state, and realizing real-time fitting under dynamic environment. The physical model fusion comprises establishing a physical model of the FBG sensor based on material mechanics and optical principles, combining experimental data, and achieving fitting through a parameter estimation and optimization algorithm; The multi-sensor data fusion comprises fusing temperature and vibration sensor data and performing fitting by means of a multi-sensor information fusion technology.

10. The FBG-based substrate deformation measurement system adaptive fitting system of claim 6, wherein, Final results are visually displayed through a host computer interface; The adaptive segmented fitting algorithm comprises segmentation determination, model selection, parameter initialization, adaptive fitting and model evaluation; The segmentation determination comprises determining segmentation points based on stress-wavelength data set distribution characteristics and dividing the entire stress range into a plurality of subintervals; The model selection comprises selecting or combining appropriate nonlinear fitting models for the data characteristics of each subinterval; The parameter initialization comprises initializing the parameters of the selected models, and the parameters of the fitting models in each subinterval are initialized respectively; The adaptive fitting comprises optimizing and adjusting the model parameters of each subinterval by using an optimization algorithm to minimize the fitting error; The model evaluation comprises evaluating whether the performance of each subinterval model meets the preset accuracy requirement by using mean square error and mean absolute error indicators, returning to the parameter initialization module for re-optimization for the subinterval that does not meet the requirement, and determining the final segmented fitting model when all the subintervals meet the requirement.

Images