A method for measuring thermal stress of a boiler structure based on fiber grating technology

Abstract

This application relates to the field of fiber optic sensing measurement and engineering structure stress analysis technology, and particularly to a method for measuring the thermal stress of a boiler structure based on fiber Bragg grating technology. The method includes: S1 sensor deployment; S2 spectral signal transmission; S3 spectral line preprocessing; S4 center wavelength demodulation; S5 thermal stress inversion; S6 stress map reconstruction; and S7 anomaly identification. This application achieves synchronous sensing and precise decoupling of temperature and strain in high-temperature and complex environments by deploying decoupled fiber Bragg grating sensors in key stress areas of the boiler structure and introducing a Bragg reflection center wavelength response model. This response mechanism, based on the nonlinear coupling relationship between wavelength drift and temperature or strain, effectively avoids the problems of traditional thermocouples and resistance strain gauges being susceptible to temperature drift interference and having limited deployment numbers, significantly improving the accuracy and distribution integrity of thermal stress identification.

Description

Technical Field

[0001] This application relates to the field of fiber optic sensing measurement and engineering structure stress analysis technology, and in particular to a method for measuring the thermal stress of boiler structures based on fiber Bragg grating technology. Background Technology

[0002] Boilers are widely used in high-temperature and high-pressure industries such as thermal power generation and chemical industry. Their structural parts are subjected to complex thermal loads for a long time, which can easily lead to thermal stress concentration and fatigue failure, resulting in serious accidents such as cracks, bulges, and even furnace tube rupture. To ensure the safety of boiler structure, it is crucial to obtain real-time information on the thermal stress status of key parts and provide early warnings.

[0003] Traditional thermal stress measurement typically employs contact sensors such as strain gauges and thermocouples. However, these methods have limitations such as poor anti-interference capabilities, limited deployment numbers, and difficulty in long-term stable operation. Especially in high-temperature, confined spaces, signal drift and limitations in measurement point placement become increasingly apparent. To address these issues, researchers have begun to introduce fiber Bragg grating (FBG) technology, which offers advantages such as resistance to electromagnetic interference, high-temperature resistance, and ease of distributed deployment. It is considered an alternative approach for thermal stress monitoring.

[0004] However, in practical applications, FBG monitoring still faces many technical obstacles. On the one hand, the Bragg wavelength is simultaneously affected by temperature and strain coupling, making it difficult to directly distinguish the causes of thermal stress. On the other hand, wavelength division multiplexing brings signal superposition and crosstalk problems, which can easily lead to the accumulation of demodulation errors. In addition, traditional center wavelength identification methods are difficult to deal with complex background noise and cannot achieve high-precision, high-resolution demodulation. Furthermore, in the process of deriving the thermal stress field from the demodulated value, there is a lack of a complete and sustainable closed-loop calculation and early warning mechanism. Summary of the Invention

[0005] The purpose of this application is to provide a method for measuring the thermal stress of boiler structures based on fiber Bragg grating technology, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, this application provides the following technical solution: a method for measuring the thermal stress of a boiler structure based on fiber Bragg grating technology, the specific steps of which are as follows: S1 Sensor Deployment: Decoupled Bragg grating sensors are deployed at key stress points of the boiler structure, and a sensing response model is established based on the Bragg reflection center wavelength formula to monitor temperature and strain changes at the structural location in real time. S2 Spectral Signal Transmission: A multi-channel reflection wavelength distribution is designed using wavelength division multiplexing (WDM) channel partitioning formulas, and a spectral multiplexing transmission link with fixed wavelength intervals is constructed to achieve interference-free synchronous transmission of multiple grating signals in a single optical fiber; S3 Spectral Line Preprocessing: S3 spectral preprocessing: Feature enhancement processing is performed on the received reflectance spectral data, including extracting the envelope curve and reflectance peak position information using Hilbert transform, in order to improve the signal-to-noise ratio of subsequent wavelength demodulation; S4 Center Wavelength Demodulation: The preprocessed reflectance spectrum sequence is input into a Transformer neural network model constructed based on a multi-head attention mechanism formula to achieve high-resolution center wavelength demodulation and reflectance spectrum peak localization; S5 Thermal Stress Inversion: Input the strain and temperature values ​​obtained by demodulation into the thermoelastic stress equation for calculation to obtain the thermal stress value corresponding to the sensing point; S6 Stress Map Reconstruction: Spatial interpolation fitting of thermal stress data from multiple sensing points is performed using a weighted reconstruction interpolation formula to generate a two-dimensional continuous thermal stress distribution map of key areas of the boiler structure. S7 Abnormal Change Identification: Based on the thermal stress change rate criterion formula, analyze whether the rate of change of thermal stress over time exceeds the set warning threshold. If the rate of change exceeds the warning threshold, an alarm signal is triggered.

[0007] Preferably, in step S1, the deployment of decoupled Bragg grating sensors at key stress-bearing locations of the boiler structure, and the establishment of a sensing response model based on the Bragg reflection center wavelength formula for real-time monitoring of temperature and strain changes at these structural locations, includes: selecting at least one high thermal stress location in the boiler structure, such as near the weld of the lower header of the steam drum, the outlet of the downcomer, or the water-cooled wall nozzle area, and deploying decoupled Bragg grating sensors thereon. The decoupled Bragg grating sensors are thermally isolated from the structure via a metal sheath and a high-temperature resistant ceramic gasket to distinguish between thermal strain and structural strain signals, generating a distinction result; based on the distinction result and the Bragg reflection center wavelength formula, an analytical mapping model between spectral drift and structural state is established; based on the analytical mapping model, effective reflection information is extracted using spectral line tracking and denoising processing, and the temperature and strain dual-channel responses at different deployment points are matched to perform dynamic inversion of the stress state and real-time monitoring of temperature and strain changes at the structural locations.

[0008] Preferably, in step S2, the step of designing a multi-channel reflection wavelength distribution using a wavelength division multiplexing (WDM) channel partitioning formula and constructing a spectral multiplexing transmission link with fixed-interval wavelengths to achieve interference-free synchronous transmission of multiple grating signals in a single optical fiber includes: designing an independent reflection center wavelength for each Bragg grating based on the WDM channel partitioning principle, generating a multi-channel wavelength group with fixed intervals; constructing a single-fiber multi-channel spectral multiplexing link based on the multi-channel wavelength group with fixed intervals using a multi-channel reflection wavelength scheme, using a unified incident and reflection path design with a grating array, and reducing channel crosstalk through transmission path temperature control and encapsulation isolation technology to achieve interference-free synchronous transmission of the multiple grating signals in the single optical fiber.

[0009] Preferably, in step S3, the feature enhancement processing of the received reflectance spectral data includes extracting the envelope curve and reflection peak position information using Hilbert transform. This includes: after the original reflectance spectral signal enters the feature enhancement processing module, analyzing the original reflectance spectral signal using the Hilbert transform envelope extraction formula to extract the instantaneous envelope curve of the original reflectance spectral signal; identifying and locating the reflection center wavelength based on the abrupt change points and gain characteristics of the instantaneous envelope curve; and extracting the reflection peak position information in the spectrum using a dynamic thresholding method and an envelope peak fitting method.

[0010] Preferably, in step S4, inputting the preprocessed reflectance spectrum sequence into a Transformer neural network model constructed based on a multi-head attention mechanism to achieve high-resolution center wavelength demodulation and reflectance spectrum peak localization includes: inputting the preprocessed reflectance spectrum sequence into the Transformer neural network model based on a multi-head attention mechanism; extracting structural information of the reflectance spectrum in different frequency band feature regions through multiple parallel attention heads to perform global perception and local enhancement of the full-band signal, and constructing a nonlinear mapping feature space between reflectance intensity and center wavelength; simultaneously setting a wavelength regression branch and a reflectance peak classification branch at the output of the Transformer neural network model to perform center wavelength localization and main peak identification of the reflectance spectrum; and achieving high-resolution center wavelength demodulation and reflectance spectrum peak localization by fusing and analyzing the attention weight results.

[0011] Preferably, in step S5, the step of inputting the demodulated strain and temperature values ​​into the thermoelastic stress equation for calculation to obtain the thermal stress value corresponding to the sensing point includes: using the demodulated strain and temperature values ​​as basic physical input parameters, and combining them with the elastic modulus, thermal expansion coefficient, and Poisson's ratio of the boiler structural material to construct a linear thermoelastic stress estimation model; based on the linear thermoelastic stress estimation model, coupling the thermal expansion effect caused by temperature change with the measured strain to quantitatively calculate the thermal stress value at each monitoring point.

[0012] Preferably, in step S6, the step of using a weighted reconstruction interpolation formula to spatially interpolate and fit the thermal stress data of multiple sensing points to generate a two-dimensional continuous thermal stress distribution map of the key area of ​​the boiler structure includes: collecting thermal stress measurement data of multiple fiber Bragg sensing points in the key area of ​​the boiler structure; constructing a thermal stress field model based on the stress measurement data and physical location coordinates using a weighted interpolation method; setting weights according to the spacing between sensing points and data stability; fitting the thermal stress distribution trend in the two-dimensional area; applying the interpolation model to the two-dimensional structural mesh of the boiler based on the thermal stress distribution trend in the two-dimensional area; and calculating the stress value point by point using the inverse distance weighting principle to generate a two-dimensional continuous thermal stress distribution map of the key area of ​​the boiler structure.

[0013] Preferably, in step S7, the step of analyzing whether the rate of change of thermal stress over time exceeds a set warning threshold based on the thermal stress change rate criterion formula, and triggering an alarm signal if the rate of change exceeds the warning threshold, includes: constructing a monitoring mechanism based on the thermal stress change rate criterion formula to continuously collect thermal stress time-series data of key parts of the boiler, and performing time differentiation processing on the thermal stress time-series data to extract the rate of change of thermal stress; establishing a correlation mapping between strain change and actual operating conditions through the warning threshold to dynamically identify local abnormal stress fluctuations; comparing the rate of change of thermal stress with the warning threshold to trigger the alarm signal when the rate of change at the monitoring point exceeds a preset critical value.

[0014] A boiler structure thermal stress measurement device based on fiber Bragg grating technology, the specific steps of which are as follows: The sensor deployment module is used to deploy decoupled Bragg grating sensors at key stress-bearing parts of the boiler structure and establish a sensing response model based on the Bragg reflection center wavelength formula to monitor temperature and strain changes at the structural location in real time. The spectral signal transmission module is used to design multi-channel reflection wavelength distributions using wavelength division multiplexing channel partitioning formulas and to construct a spectral multiplexing transmission link with fixed wavelength intervals, enabling interference-free synchronous transmission of multiple grating signals in a single optical fiber. The spectral preprocessing module is used to perform feature enhancement processing on the received reflectance spectral data, including extracting the envelope curve and reflectance peak position information using Hilbert transform, so as to improve the signal-to-noise ratio of subsequent wavelength demodulation. The center wavelength demodulation module is used to input the preprocessed reflectance spectrum sequence into the Transformer neural network model constructed based on the multi-head attention mechanism formula to achieve high-resolution center wavelength demodulation and reflectance spectrum peak localization. The thermal stress inversion module is used to input the demodulated strain and temperature values ​​into the thermoelastic stress equation for calculation, so as to obtain the thermal stress value corresponding to the sensing point. The stress map reconstruction module is used to perform spatial interpolation fitting on the thermal stress data of multiple sensing points using a weighted reconstruction interpolation formula to generate a two-dimensional continuous thermal stress distribution map of the key areas of the boiler structure. The abnormal change identification module is used to analyze whether the rate of change of thermal stress over time exceeds a set warning threshold based on the thermal stress change rate criterion formula. If the rate of change exceeds the warning threshold, an alarm signal is triggered.

[0015] Preferably, the sensor deployment module includes: a sensor deployment unit, used to select at least one high thermal stress location in the boiler structure, such as near the weld of the lower header of the steam drum, the outlet of the downcomer, or the water-cooled wall nozzle area, and deploy a decoupled Bragg grating sensor. The decoupled Bragg grating sensor is thermally isolated from the structure through a metal sheath and a high-temperature resistant ceramic gasket to distinguish between thermal strain and structural strain signals and generate a distinction result; and a real-time monitoring unit, used to establish an analytical mapping model between spectral drift and structural state based on the distinction result and the Bragg reflection center wavelength formula. Based on the analytical mapping model, effective reflection information is extracted using spectral line tracking and denoising processing, and the temperature and strain dual-channel responses of different deployment points are matched to perform dynamic inversion of stress state and monitor temperature and strain changes at the structural location in real time.

[0016] Compared with the prior art, the beneficial effects of this application are as follows: 1. This application achieves synchronous sensing and precise decoupling of temperature and strain in high-temperature complex environments by deploying decoupled fiber Bragg grating sensors in key stress areas of boiler structures and introducing a Bragg reflection center wavelength response model. This response mechanism, based on the nonlinear coupling relationship between wavelength drift and temperature or strain, effectively avoids the problems of traditional thermocouples and resistance strain gauges being susceptible to temperature drift interference and having limited deployment numbers. It significantly improves the accuracy and distribution integrity of thermal stress identification, and is especially suitable for state sensing of multiple stress concentration parts such as boiler shell, tube sheet, and superheater.

[0017] 2. This application introduces a wavelength division multiplexing channel partitioning mechanism to construct a multi-channel narrowband reflection wavelength spectrum distribution scheme, enabling multiple grating sensing signals to be stably transmitted in a single optical fiber without interference. Compared with the traditional parallel deployment method, the spectrum multiplexing link not only significantly reduces the number of optical fibers and deployment space, but also improves the stability and anti-interference capability of the transmitted signal. Combined with constant interval wavelength matching design and channel interval control strategy, it ensures that different deployment points have stable identifiability in the spectrum space, thereby adapting to the needs of fine monitoring of thermal stress under high-density deployment.

[0018] 3. This application achieves precise demodulation of the center wavelength and spectral peaks by inputting the preprocessed reflectance spectrum sequence into a Transformer network model with a multi-head attention mechanism. Then, a high-resolution thermal stress field is constructed using the thermoelastic stress analytical formula. Finally, combined with the rate of change criterion, the stress change rate is dynamically judged to determine whether it exceeds the safety threshold, and an early warning mechanism is triggered. This closed-loop mechanism from perception to demodulation to modeling to early warning effectively overcomes the drawbacks of traditional manual reading or delayed response based on preset thresholds, and achieves highly timely monitoring of boiler operating status and proactive risk intervention, significantly enhancing the system's engineering deployability and operational safety assurance capabilities.

[0019] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0020] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating a method for measuring thermal stress in a boiler structure based on fiber Bragg grating technology, according to an embodiment of this application. Figure 2 This is a schematic diagram of a boiler structure thermal stress measurement device based on fiber Bragg grating technology provided in an embodiment of this application. Detailed Implementation

[0021] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0022] The following describes a method for measuring thermal stress in a boiler structure based on fiber Bragg grating (FBG) technology, according to an embodiment of this application, with reference to the accompanying drawings. Regarding the related technologies mentioned in the background section, FBG monitoring still faces many technical obstacles. On the one hand, the Bragg wavelength is simultaneously affected by temperature and strain coupling, making it difficult to directly distinguish the causes of thermal stress. On the other hand, wavelength division multiplexing (WDM) introduces signal superposition and crosstalk problems, easily leading to the accumulation of demodulation errors. Furthermore, traditional center wavelength identification methods struggle to cope with complex background noise and achieve high-precision, high-resolution demodulation. The process of deriving the thermal stress field from the demodulated value also lacks a complete and sustainable closed-loop calculation and early warning mechanism. This application provides a method for measuring thermal stress in a boiler structure based on FBG technology. In this method, decoupled FBG sensors can be deployed in key stress areas of the boiler structure, and a Bragg reflection center wavelength response model can be introduced to achieve synchronous sensing and accurate decoupling of temperature and strain in complex high-temperature environments. This response mechanism, based on the nonlinear coupling relationship between wavelength drift and temperature or strain, effectively avoids the problems of traditional thermocouples and resistance strain gauges being susceptible to temperature drift interference and having limited deployment numbers, significantly improving the accuracy and distribution integrity of thermal stress identification. This solves many technical obstacles that still exist in FBG monitoring in related technologies. On the one hand, the Bragg wavelength is simultaneously affected by temperature and strain coupling, making it difficult to directly distinguish the cause of thermal stress. On the other hand, wavelength division multiplexing brings signal superposition and crosstalk problems, which can easily lead to the accumulation of demodulation errors. In addition, traditional center wavelength identification methods are difficult to deal with complex background noise and cannot achieve high-precision, high-resolution demodulation. Furthermore, in the process of deriving the thermal stress field from the demodulated value, there is a lack of a complete and sustainable closed-loop calculation and early warning mechanism.

[0023] Specifically, Figure 1 This is a flowchart illustrating a method for measuring thermal stress in a boiler structure based on fiber Bragg grating technology, provided in an embodiment of this application.

[0024] like Figure 1 As shown, the method for measuring the thermal stress of a boiler structure based on fiber Bragg grating technology includes the following steps: S1 Sensor Deployment: Decoupled Bragg grating sensors are deployed at key stress points of the boiler structure, and a sensing response model is established based on the Bragg reflection center wavelength formula to monitor temperature and strain changes at these structural locations in real time.

[0025] In actual implementation, the embodiments of this application can deploy S1 sensors: decoupled Bragg grating sensors are deployed in key stress-bearing parts of the boiler structure, such as high-temperature steam pipe elbows and the weld area connecting the steam drum and downcomer. A sensing response model is established based on the Bragg reflection center wavelength formula to monitor the temperature and strain changes at the structural location in real time. This enables synchronous sensing and precise decoupling of temperature and strain in high-temperature complex environments. This response mechanism, based on the nonlinear coupling relationship between wavelength drift and temperature or strain, effectively avoids the problems of traditional thermocouples and resistance strain gauges being susceptible to temperature drift interference and having limited deployment numbers. It significantly improves the accuracy and distribution integrity of thermal stress identification, and is especially suitable for state sensing of multiple stress concentration parts such as boiler shells, tube sheets, and superheaters.

[0026] Optionally, in one embodiment of this application, in S1, decoupled Bragg grating sensors are deployed at key stress-bearing parts of the boiler structure, and a sensing response model is established based on the Bragg reflection center wavelength formula to monitor temperature and strain changes at the structural location in real time. This includes: selecting at least one high thermal stress location in the boiler structure, such as near the weld of the lower header of the steam drum, the outlet of the downcomer, and the water-cooled wall nozzle area, and deploying decoupled Bragg grating sensors. The decoupled Bragg grating sensors are thermally isolated from the structure through a metal sheath and a high-temperature resistant ceramic gasket to distinguish between thermal strain and structural strain signals, generating a distinction result; based on the distinction result and the Bragg reflection center wavelength formula, an analytical mapping model between spectral drift and structural state is established; based on the analytical mapping model, effective reflection information is extracted using spectral line tracking and denoising processing, and the temperature and strain dual-channel responses of different deployment points are matched to perform dynamic inversion of stress state and monitor temperature and strain changes at the structural location in real time.

[0027] In this embodiment, decoupled Bragg grating sensors can be installed in high thermal stress areas of the boiler structure, such as near the weld of the lower header of the steam drum, the outlet of the downcomer, and the water-cooled wall nozzle area. The sensors are thermally isolated from the structure by a metal sheath and a high-temperature resistant ceramic gasket (the decoupled Bragg grating sensor is thermally insulated by a high-temperature resistant ceramic gasket located between the structural substrate and the sensing element. The ceramic gasket is covered with a metal sheath, and the metal sheath is fixed to the structural surface by mechanical fastening or welding pre-drilled holes, thereby providing a good mechanical coupling channel while avoiding thermal interference). This can effectively distinguish between thermal strain and structural strain signals, ensuring measurement accuracy. At the same time, the placement density is optimized based on the structural simulation results to improve the response sensitivity to local stress changes (the placement density is determined based on the finite element thermo-mechanical coupling simulation analysis results, and is preferentially placed in areas with the maximum stress value and thermal strain gradient to improve the response sensitivity and spatial coverage accuracy to local dynamic stress changes).

[0028] Furthermore, in this embodiment, an analytical mapping model between spectral drift and structural state can be established by combining the Bragg reflection center wavelength formula. This model uses spectral line tracking and denoising to extract effective reflection information and matches the temperature and strain dual-channel responses of different deployment points to achieve dynamic inversion of stress state. The response mechanism considers thermal hysteresis and nonlinear factors in boiler operation to ensure the stability and accuracy of the demodulation model under high temperature fluctuations.

[0029] The formula for expressing the center wavelength of Bragg reflection is:

[0030] in, The wavelength of the Bragg reflection center (unit: nm). The effective refractive index in a fiber optic grating. The grating period is in nm.

[0031] The formula in this application embodiment is used to establish the response relationship between the reflection wavelength of FBG (fiber Bragg grating) and structural strain and temperature. It is the core theoretical basis of the S1 deployment stage. The formula in this application embodiment provides a sensing basis for the joint detection of temperature and strain, and can realize real-time dual-parameter monitoring of key parts of the boiler.

[0032] S2 spectral signal transmission: By designing a multi-channel reflection wavelength distribution using wavelength division multiplexing channel partitioning formula, and constructing a spectral multiplexing transmission link with fixed wavelength intervals, multiple grating signals can be transmitted synchronously in a single optical fiber without interference.

[0033] In this embodiment, S2 spectral signal transmission can be performed: a multi-channel reflection wavelength distribution is designed by using a wavelength division multiplexing channel partitioning formula, and a spectral multiplexing transmission link with fixed wavelength intervals is constructed to achieve interference-free synchronous transmission of multiple grating signals in a single optical fiber. Thus, by introducing a wavelength division multiplexing channel partitioning mechanism, a multi-channel narrowband reflection wavelength spectral distribution scheme is constructed, enabling multiple grating sensing signals to be stably transmitted in a single optical fiber without interference.

[0034] The spectral multiplexing link in this embodiment not only significantly reduces the number of optical fibers and deployment space, but also improves the stability and anti-interference capability of the transmitted signal. Combined with constant interval wavelength matching design and channel interval control strategy, it ensures that different deployment points have stable identifiability in the spectrum space, thereby adapting to the needs of fine monitoring of thermal stress under high-density deployment.

[0035] Optionally, in one embodiment of this application, in S2, a multi-channel reflection wavelength distribution is designed using a wavelength division multiplexing (WDM) channel partitioning formula, and a spectral multiplexing transmission link with fixed-interval wavelengths is constructed to achieve interference-free synchronous transmission of multiple grating signals in a single optical fiber. This includes: designing an independent reflection center wavelength for each Bragg grating based on the WDM channel partitioning principle, generating a multi-channel wavelength group with fixed intervals; constructing a single-fiber multi-channel spectral multiplexing link based on the multi-channel wavelength group with fixed intervals using a multi-channel reflection wavelength scheme, using a grating array to unify the incident and reflection path design, and reducing channel crosstalk through transmission path temperature control and encapsulation isolation technology to achieve interference-free synchronous transmission of multiple grating signals in a single optical fiber.

[0036] In this embodiment, based on the wavelength division multiplexing channel partitioning principle, an independent reflection center wavelength is designed for each Bragg grating to form a multi-channel wavelength group with a fixed interval. This design fully considers the spectral width and light source stability, so that each sensing unit does not interfere with each other when transmitting on the same optical fiber, thus meeting the requirements of high-density deployment and high-resolution demodulation.

[0037] The formula for wavelength division multiplexing (WDM) channel partitioning is as follows:

[0038] in, The wavelength interval (nm) between adjacent gratings is the reflection wavelength. For refractive index modulation depth, For the grating period, This represents the number of wavelength division multiplexing (WDM) channels.

[0039] The formulas in this application embodiment are used to design the wavelength spacing of each sensor in the FBG array, ensuring that the multi-channel signals do not overlap, thereby enabling multiple FBG sensors to transmit data synchronously on the same optical fiber without crosstalk, improving deployment efficiency and system response capability.

[0040] Furthermore, in this embodiment, a multi-channel spectral multiplexing link for a single optical fiber can be constructed based on a predefined multi-channel reflection wavelength group W. The incident end uses a bandwidth-controllable ASE light source for unified excitation, and the output end combines a spectral demodulator to separate the reflected wavelengths. The sensor array is arranged in a cascaded structure, with the incident and reflection paths collinear. Signal selection is achieved through FBG reflection characteristics. To reduce crosstalk, the system introduces dual suppression measures: first, a sheath-type encapsulation structure with thermal compensation is used to encapsulate each optical fiber sensing point in a temperature-controlled cavity, reducing spectral drift spread caused by thermal conduction interference; second, a layout model based on spacing management is constructed to ensure that the physical spacing meets the minimum crosstalk spacing, avoiding near-field interference. The overall structure supports synchronous demodulation of N≥16 points, with a demodulation system response delay of less than 50ms, meeting the requirements for rapid feedback of boiler operating thermal stress.

[0041] S3 spectral preprocessing: Feature enhancement processing is performed on the received reflectance spectral data, including extracting the envelope curve and reflectance peak position information using Hilbert transform, in order to improve the signal-to-noise ratio of subsequent wavelength demodulation.

[0042] Specifically, the embodiments of this application can perform S3 spectral preprocessing: feature enhancement processing is performed on the received reflectance spectral data, including extracting the envelope curve and reflectance peak position information using Hilbert transform, so as to improve the signal-to-noise ratio of subsequent wavelength demodulation.

[0043] Optionally, in one embodiment of this application, in S3, feature enhancement processing is performed on the received reflectance spectral data, including extracting the envelope curve and reflection peak position information using Hilbert transform. This includes: after the original reflectance spectral signal enters the feature enhancement processing module, the original reflectance spectral signal is analyzed using the Hilbert transform envelope extraction formula to extract the instantaneous envelope curve of the original reflectance spectral signal; the reflection center wavelength is identified and located based on the abrupt change point and gain characteristics of the instantaneous envelope curve; and the reflection peak position information in the spectrum is extracted using the dynamic threshold method and the envelope peak fitting method.

[0044] In this embodiment, the received grating reflection signal can be first input into the feature enhancement processing module. The system constructs a complex analytical signal based on the Hilbert transform envelope extraction formula. By taking its instantaneous amplitude, the signal envelope is extracted. This envelope reveals the amplitude variation trend of the reflection spectrum at each wavelength point, which helps to enhance the energy concentration characteristics of the main reflection peak and suppress high-frequency background noise. Compared with traditional smoothing processing, the Hilbert transform envelope can better preserve the main peak structure and avoid the loss of spectral details, thus laying a high signal-to-noise ratio foundation for center wavelength identification. The Hilbert transform envelope extraction expression formula is as follows:

[0045] in, The signal envelope amplitude, This is the original reflection spectrum signal. for The Hilbert transform.

[0046] The formula in this application embodiment is used to enhance the positional characteristics of reflection peaks in spectral signals, improve wavelength demodulation accuracy, thereby enhancing the reflection peak boundary, enabling the demodulation algorithm to more accurately identify the center wavelength, and improving the demodulation signal-to-noise ratio.

[0047] Furthermore, in this embodiment, after envelope enhancement is completed, the system identifies the center wavelength position in the reflection spectrum by analyzing the local maximum position and the abrupt change point of the first derivative of the envelope curve. The system uses a dynamic threshold method to match the main peak region and combines the fitted envelope peak value for precise wavelength positioning. At the same time, the system establishes a peak selection strategy under signal-to-noise ratio constraints to avoid the influence of multi-peak interference and weak signal false peaks. This processing flow significantly reduces the interference of random errors and light source instability on the demodulation results and improves the repeatability and stability of center wavelength extraction.

[0048] S4 Center Wavelength Demodulation: The preprocessed reflectance spectrum sequence is input into a Transformer neural network model constructed based on a multi-head attention mechanism formula to achieve high-resolution center wavelength demodulation and reflectance spectrum peak localization.

[0049] In actual implementation, the embodiments of this application can perform S4 center wavelength demodulation: inputting the preprocessed reflection spectrum sequence into the Transformer neural network model constructed based on the multi-head attention mechanism formula to achieve high-resolution center wavelength demodulation and reflection spectrum peak localization.

[0050] The embodiments of this application can achieve accurate demodulation of the center wavelength and spectral peaks by inputting the preprocessed reflection spectrum sequence into a Transformer network model with a multi-head attention mechanism, and then construct a high-resolution thermal stress field through the thermoelastic stress analytical formula.

[0051] Optionally, in one embodiment of this application, in S4, the preprocessed reflectance spectrum sequence is input into a Transformer neural network model constructed based on a multi-head attention mechanism formula to achieve high-resolution center wavelength demodulation and reflectance spectrum peak localization. This includes: inputting the preprocessed reflectance spectrum sequence into a Transformer neural network model based on a multi-head attention mechanism; extracting structural information of the reflectance spectrum in different frequency band feature regions through multiple parallel attention heads to perform global perception and local enhancement of the full-band signal, and constructing a nonlinear mapping feature space between reflectance intensity and center wavelength; simultaneously setting a wavelength regression branch and a reflectance peak classification branch at the output of the Transformer neural network model to perform center wavelength localization and main peak identification of the reflectance spectrum; and performing fusion analysis on the attention weight results to achieve high-resolution center wavelength demodulation and reflectance spectrum peak localization.

[0052] In this embodiment, the Bragg grating reflectance spectrum sequence after feature enhancement processing can be input into a Transformer neural network model with a multi-head attention mechanism. This model uses multiple parallel attention channels to discriminate spectral features in different frequency bands, realizing full-band perception and local feature enhancement of the reflectance spectrum. Each attention head can extract reflectance spectrum structure information at different scales, construct a nonlinear correlation representation between reflectance intensity and center wavelength, provide multi-dimensional feature support for subsequent high-precision wavelength identification, and enhance the model's ability to model complex spectral line changes.

[0053] The multi-head attention mechanism is expressed as follows (for peak demodulation): MultiHead(Q,K,V)=Concat(head1,…,headh)WO in, headi=Attention(QWiQ,KWiK,VWiV) Where Q, K, and V are the query, key, and value matrices (generated from the reflection spectrum sequence), respectively; WiQ, WiK, and WiV are the learnable linear mapping weights; WO is the output mapping weight; and h is the number of attention heads.

[0054] The embodiments of this application are applicable to high-precision wavelength demodulation and reflection peak localization, especially spectral line identification under nonlinear perturbation. By using an attention mechanism to enhance attention to local spectral features, it can effectively identify spectral detail changes and achieve high-resolution demodulation.

[0055] Furthermore, the Transformer model output is equipped with a wavelength regression branch and a main peak classification branch to locate the center wavelength and identify the main reflection peak in the reflection spectrum. By fusing the weight results of different attention channels, this structure can effectively cope with multi-peak interference and main peak occlusion, and improve the accuracy of reflection spectrum peak resolution and response to small wavelength changes. At the same time, the model training stage adopts a joint optimization strategy of center wavelength mean square error and classification loss to ensure that the output results have good generalization ability and demodulation stability under complex working conditions, which is suitable for stress monitoring needs in high temperature fluctuation scenarios during boiler operation.

[0056] S5 Thermal Stress Inversion: The strain and temperature values ​​obtained by demodulation are input into the thermoelastic stress equation for calculation to obtain the thermal stress value corresponding to the sensing point.

[0057] In actual implementation, the embodiments of this application can perform S5 thermal stress inversion: input the strain value and temperature value obtained by demodulation into the thermoelastic stress equation for calculation to obtain the thermal stress value corresponding to the sensing point.

[0058] Optionally, in one embodiment of this application, in S5, the demodulated strain value and temperature value are input into the thermoelastic stress equation for calculation to obtain the thermal stress value corresponding to the sensing point. This includes: using the demodulated strain value and temperature value as basic physical input parameters, and combining the elastic modulus, thermal expansion coefficient and Poisson's ratio of the boiler structural material to construct a linear thermoelastic stress estimation model; based on the linear thermoelastic stress estimation model, coupling the thermal expansion effect caused by temperature change with the measured strain to quantitatively calculate the thermal stress value of each monitoring point.

[0059] In this embodiment, the temperature and strain values ​​obtained by demodulating the Bragg grating can be used as input variables. Combined with the basic physical parameters such as the thermal expansion coefficient, elastic modulus and Poisson's ratio of the boiler structural materials, a linear thermoelastic stress calculation model under the background of thermo-mechanical coupling is established. The model sets constant boundary and quasi-static thermal field conditions at each monitoring point, and considers the coordinated response of material thermal deformation and structural strain caused by temperature rise. It can accurately reflect the real stress state of boiler components under high temperature load, and provide a generalizable theoretical basis and parameter framework for thermal stress inversion.

[0060] The expression for thermoelastic stress is:

[0061] in, Thermal stress (Pa), The elastic modulus of the material (Pa) To measure the total strain, The coefficient of linear expansion of the material (1 / °C) This represents the change in temperature (°C).

[0062] The embodiments of this application can deduce thermal stress from strain and temperature data obtained by demodulation from grating sensors, thereby achieving accurate stress inversion after temperature compensation, which is a key calculation model for boiler operation safety monitoring.

[0063] Furthermore, based on the constructed thermoelastic model, the system in this embodiment couples the temperature changes collected at each sensing point with the corresponding structural strain to obtain the corresponding instantaneous thermal stress values. This process can effectively identify the stress concentration trend and variation range in the high-stress areas of the downcomer outlet and header weld. By inputting the results into the subsequent interpolation modeling and risk assessment process, quantitative monitoring of the overall heat load distribution of the boiler can be achieved, providing data support for fatigue life assessment, structural anomaly early warning, and dynamic operation and maintenance strategy formulation.

[0064] S6 Stress Map Reconstruction: Spatial interpolation fitting of thermal stress data from multiple sensing points is performed using a weighted reconstruction interpolation formula to generate a two-dimensional continuous thermal stress distribution map of key areas of the boiler structure.

[0065] In actual implementation, the embodiments of this application can perform S6 stress map reconstruction: using a weighted reconstruction interpolation formula to perform spatial interpolation fitting on the thermal stress data of multiple sensing points to generate a two-dimensional continuous thermal stress distribution map of the key areas of the boiler structure.

[0066] Optionally, in one embodiment of this application, in S6, a weighted reconstruction interpolation formula is used to spatially interpolate and fit the thermal stress data of multiple sensing points to generate a two-dimensional continuous thermal stress distribution map of the key area of ​​the boiler structure. This includes: collecting thermal stress measurement data of multiple fiber Bragg sensing points in the key area of ​​the boiler structure, and constructing a thermal stress field model based on the stress measurement data and physical location coordinates using a weighted interpolation method, setting weights according to the spacing between sensing points and data stability, and fitting the thermal stress distribution trend in the two-dimensional area; based on the thermal stress distribution trend in the two-dimensional area, applying the interpolation model to the two-dimensional structural mesh of the boiler, and calculating the stress value point by point using the inverse distance weighting principle to generate a two-dimensional continuous thermal stress distribution map of the key area of ​​the boiler structure.

[0067] In this embodiment, multiple fiber Bragg sensor points can be deployed in high-stress-sensitive areas such as the weld seam of the lower header of the boiler drum and the water-cooled wall nozzle area, and their thermal stress response data can be collected in real time. Combining the physical coordinates of the sensor points and the spatial distribution characteristics of the measurement points, a two-dimensional thermal stress field model is constructed using a weighted reconstruction interpolation method. This interpolation method uses the spacing between sensor points as the main weight factor and sets dynamic weights based on the volatility of historical data to ensure improved fitting smoothness in sparse areas and enhanced local gradient response in dense areas, thereby achieving effective fitting of the stress spatial change trend.

[0068] The weighted spatial interpolation expression is:

[0069] in, Let be the thermal stress value at the interpolation point (x,y). Let i be the thermal stress value of the i-th known sensing point. The distance from the interpolation point to the i-th point is... This is the weighting index (usually 2).

[0070] The embodiments of this application can smoothly transition the thermal stress data of discrete sensing points into a continuous regional thermal stress map, thereby realizing two-dimensional continuous visualization of the thermal stress of key boiler structures, which helps to determine areas of concentrated heat load and areas of potential fatigue risk.

[0071] Furthermore, in this embodiment, the interpolation model can be applied to the two-dimensional structural mesh of the boiler. Based on the inverse distance weighted interpolation principle, the system dynamically adjusts the interpolation weight according to the inverse square of the distance from the monitoring point to the target grid point, calculates the thermal stress response value of each grid node point by point, and draws a thermal stress isopleth map. This map intuitively presents the region of local stress concentration and the trend of thermal gradient change in the boiler structure. It is particularly suitable for the qualitative identification and quantitative prediction of fatigue crack initiation areas, providing a decision-making basis for boiler maintenance cycle design and service life management.

[0072] S7 Abnormal Change Detection: Based on the thermal stress change rate criterion formula, it analyzes whether the rate of change of thermal stress over time exceeds the set warning threshold. If the rate of change exceeds the warning threshold, an alarm signal is triggered.

[0073] In actual implementation, the embodiments of this application can analyze whether the rate of change of thermal stress over time exceeds the set warning threshold based on the thermal stress change rate criterion formula. If the rate of change exceeds the warning threshold, an alarm signal is triggered. Combined with the change rate criterion, the stress change rate is dynamically judged to determine whether it exceeds the safety threshold, and the warning mechanism is triggered in conjunction. This closed-loop mechanism from perception to demodulation to modeling to warning effectively overcomes the drawbacks of traditional manual reading or delayed response to preset thresholds, realizes highly timely monitoring of boiler operating status and proactive risk intervention, and significantly enhances the system's engineering deployability and operational safety assurance capabilities.

[0074] Optionally, in one embodiment of this application, in S7, based on the thermal stress change rate criterion formula, it is analyzed whether the rate of change of thermal stress over time exceeds a set warning threshold. If the rate of change exceeds the warning threshold, an alarm signal is triggered. This includes: constructing a monitoring mechanism based on the thermal stress change rate criterion formula to continuously collect thermal stress time-series data of key parts of the boiler, and performing time differentiation processing on the thermal stress time-series data to extract the rate of change of thermal stress; establishing a correlation mapping between strain change and actual working conditions through the warning threshold to dynamically identify local abnormal stress fluctuations; comparing the rate of change of thermal stress with the warning threshold to trigger an alarm signal when the rate of change at the monitoring point exceeds a preset critical value.

[0075] In this embodiment of the application, for high-risk parts of the boiler structure, such as the heat-affected zone of the water-cooled wall welding and the connection between the downcomer and the header, thermal stress time-series data output by fiber Bragg sensors are continuously collected. The thermal stress change rate criterion formula is used to perform time differentiation processing on the thermal stress curve to obtain the rate curve of stress change over time. By performing sliding window filtering and feature extraction on the change rate, and combining historical operating data of different parts, differentiated early warning thresholds are established to form a dynamic response mapping between the thermal stress change rate and the boiler operating status, thereby improving the sensitivity detection capability of abnormal fluctuations.

[0076] The expression for the thermal stress change rate criterion is:

[0077] Alarm conditions: If If so, an alarm will be triggered; in, For a moment The rate of change of thermal stress, For a moment thermal stress, The set threshold for the rate of change.

[0078] The embodiments of this application can identify dangerous operating conditions caused by rapid changes in thermal stress (such as rapid overheating and structural fatigue). When abnormal changes occur in thermal stress in critical parts of the boiler, the system can issue an alarm in time, improving the timeliness of fault response and system safety.

[0079] Furthermore, in this embodiment, the rate of change of thermal stress calculated in real time at each monitoring point can be compared with the corresponding threshold. When the rate of change of a certain node exceeds the set critical criterion, the system immediately calls the alarm logic module to output a structural thermal risk warning signal. It supports graded response according to fluctuation amplitude and duration. The warning module can be connected to the boiler operation control system, supporting local interface prompts and remote monitoring linkage, ensuring that the structural thermal stress mutation signs are captured and fed back in the first time, effectively preventing fatigue damage and life loss caused by stress surge, and providing highly timely auxiliary criterion support for safe boiler operation.

[0080] Specifically, this application includes Embodiment 1: High-Temperature Weld Thermal Stress Monitoring Based on Decoupled Fiber Bragg Gratings In a 600MW pulverized coal boiler, a decoupled fiber Bragg grating sensor was installed at the weld connecting the lower header of the steam drum and the water-cooled wall nozzle. The sensor was fixedly installed with thermal isolation by a metal sheath and ceramic gasket. The distribution of the sensor was optimized by combining the results of thermal-structural simulation. The temperature-strain dual-channel data was demodulated using the spectral drift analytical model. The sensor successfully captured the change in local thermal stress peak caused by heat load fluctuations during 72 hours of continuous operation, providing raw data for weld fatigue analysis.

[0081] Example 2: Densely Distributed Fiber Thermal Stress Monitoring in Multi-Channel Wavelength Division Multiplexing Links By adopting a fixed-interval reflection wavelength design, 32 FBG sensors with different center wavelengths are connected in series through a single optical fiber to construct a wavelength division multiplexing link. All sensor signals are encapsulated with low crosstalk and path temperature control to ensure stable and synchronous transmission. The system is deployed in the intersection area of ​​the downcomer bend and the header, and successfully achieves high-resolution continuous thermal stress measurement. The wavelength demodulation stability error is less than ±2pm, which meets the requirements of high-density boiler monitoring scenarios.

[0082] Example 3: Verification of High-Precision Demodulation of Center Wavelength Based on Transformer Neural Network In the hot-state test apparatus, a preset heating curve is used to simulate the actual working conditions of the boiler heating wall. The reflection spectrum signal is enhanced by Hilbert transform and then input into a Transformer model with a multi-head attention mechanism. The trend of the center wavelength change is demodulated and compared with the measured temperature control value. The results show that the demodulation error is less than 0.1nm, and the robustness of the main peak localization is better than that of traditional convolutional networks, effectively solving the problems of multi-peak interference and small drift.

[0083] Example 4: Stress Inversion and Thermal Field Analysis Based on a Thermoelastic Mechanical Model Ten Bragg grating sensing points were set at the heating surface of a boiler tail section to record temperature and strain values. Based on the thermoelastic model, the material's elastic modulus, coefficient of thermal expansion, and boundary conditions were input to calculate the thermal stress distribution at each point. Combined with the inverse distance weighted interpolation results, a two-dimensional thermal stress field diagram was generated. The error was compared with that of the finite element simulation, and the error was controlled within 8%, verifying the inversion accuracy.

[0084] Example 5: Application of Dynamic Early Warning Mechanism for Thermal Stress Change Rate in Operating Boilers The FBG data of the water-cooled wall of the fourth boiler is input into the rate of change early warning module in real time. The critical threshold of the rate of change is set to 0.2MPa / min. When the boiler load suddenly increases and the rate of change at a local point exceeds the set threshold, the system automatically outputs an alarm signal and links to the control center to prompt maintenance. In actual operation, the early warning response time is less than 2 seconds, successfully avoiding the risk of structural overload.

[0085] In S1, sensor deployment refers to selecting high-thermal-stress areas in the boiler structure, such as near the weld seam of the lower header of the steam drum, the outlet of the downcomer, and the water-cooled wall nozzle area, and installing decoupled Bragg grating sensors. The sensors are thermally isolated from the structure via a metal sheath and a high-temperature resistant ceramic gasket (the decoupled Bragg grating sensor uses a high-temperature resistant ceramic gasket placed between the structural substrate and the sensing element for thermal insulation; the ceramic gasket is covered with a metal sheath, which is fixed to the structural surface by mechanical fastening or welding through pre-drilled holes, thus providing a good mechanical coupling channel while avoiding thermal interference). This effectively distinguishes between thermal strain and structural strain signals, ensuring measurement accuracy; simultaneously, combined with the structural... The simulation results were used to optimize the point density and improve the response sensitivity to local stress changes (the point density was determined based on the finite element thermo-mechanical coupling simulation analysis results, and the points were preferentially placed in the region with the maximum stress value and thermal strain gradient to improve the response sensitivity and spatial coverage accuracy to local dynamic stress changes). An analytical mapping model between spectral drift and structural state was established by combining the Bragg reflection center wavelength formula. This model uses spectral line tracking and denoising to extract effective reflection information and matches the temperature and strain dual-channel responses of different points to achieve dynamic inversion of stress state. The response mechanism takes into account the thermal hysteresis and nonlinear factors in boiler operation to ensure the stability and accuracy of the demodulation model under high temperature fluctuations.

[0086] In S2, spectral signal transmission refers to designing an independent reflection center wavelength for each Bragg grating based on the wavelength division multiplexing (WDM) channel partitioning principle, forming a multi-channel wavelength group with fixed intervals. This design fully considers spectral width and light source stability, ensuring that each sensing unit does not interfere with each other when transmitting on the same optical fiber, meeting the requirements of high-density deployment and high-resolution demodulation. Based on the defined multi-channel reflection wavelength group W, a multi-channel spectral multiplexing link is constructed on a single optical fiber. The incident end uses a bandwidth-controllable ASE light source for unified excitation, and the output end combines a spectral demodulator to separate the reflection wavelengths. The sensing array adopts a cascaded structure with collinear incident and reflected paths. Signal selection is achieved through the reflection characteristics of FBG. To reduce crosstalk, the system introduces dual suppression measures: First, a sheath-type encapsulation structure with thermal compensation is used to encapsulate each fiber optic sensing point in a temperature control cavity, reducing spectral drift spread caused by thermal conduction interference. Second, a layout model based on spacing management is constructed to ensure that the physical spacing meets the minimum crosstalk spacing and avoids near-field interference. The overall structure supports synchronous demodulation of N≥16 points, and the demodulation system response delay is less than 50ms, meeting the requirements for rapid feedback of thermal stress during boiler operation.

[0087] In S3, spectral preprocessing refers to the process where the received grating reflection signal is first input into the feature enhancement module. The system constructs a complex analytic signal based on the Hilbert transform envelope extraction formula and extracts the signal envelope by taking its instantaneous amplitude. This envelope reveals the amplitude variation trend of the reflection spectrum at various wavelengths, which helps to enhance the energy concentration characteristics of the main reflection peak and suppress high-frequency background noise. Compared with traditional smoothing, the Hilbert transform envelope can better preserve the main peak structure and avoid the loss of spectral details, thus laying a high signal-to-noise ratio foundation for center wavelength identification. After completing the envelope enhancement, the system identifies the center wavelength position in the reflection spectrum by analyzing the local maximum position and the first derivative abrupt change point of the envelope curve. The system uses a dynamic threshold method to match the main peak region and combines the fitted envelope peak value for accurate wavelength positioning. At the same time, the system establishes a spectral peak selection strategy under signal-to-noise ratio constraints to avoid the influence of multi-peak interference and weak signal false peaks. This processing flow significantly reduces the interference of random errors and light source instability on the demodulation results and improves the repeatability and stability of center wavelength extraction.

[0088] S4 center wavelength demodulation refers to inputting the Bragg grating reflection spectrum sequence after feature enhancement into a Transformer neural network model with a multi-head attention mechanism. This model uses multiple parallel attention channels to discriminate spectral features in different frequency bands, achieving full-band perception and local feature enhancement of the reflection spectrum. Each attention head can extract reflection spectrum structure information at different scales, constructing a nonlinear correlation representation between reflection intensity and center wavelength, providing multi-dimensional feature support for subsequent high-precision wavelength identification, and enhancing the model's ability to model complex spectral line changes. The Transformer model output is equipped with a wavelength regression branch and a main peak classification branch to locate the center wavelength and identify the main reflection peak in the reflection spectrum. By fusing the weight results of different attention channels, this structure can effectively cope with multi-peak interference, main peak shading, and other situations, improving the accuracy of reflection spectrum peak resolution and response to small wavelength changes. At the same time, the model training stage adopts a joint optimization strategy of center wavelength mean square error and classification loss to ensure that the output results have good generalization ability and demodulation stability under complex working conditions, which is suitable for stress monitoring needs in high-temperature fluctuation scenarios during boiler structure operation.

[0089] In S5, thermal stress inversion refers to using the temperature and strain values ​​obtained from Bragg grating demodulation as input variables, combined with basic physical parameters such as the thermal expansion coefficient, elastic modulus, and Poisson's ratio of the boiler structural materials, to establish a linear thermoelastic stress calculation model under thermo-mechanical coupling. This model sets constant boundaries and quasi-static thermal field conditions at each monitoring point, considering the coordinated response of material thermal deformation and structural strain caused by temperature rise. It can accurately reflect the true stress state of boiler components under high-temperature loads, providing a generalizable theoretical basis and parameter framework for thermal stress inversion. Based on the constructed thermoelastic model, the system couples the temperature changes collected at each sensing point with the corresponding structural strain variables to obtain the corresponding instantaneous thermal stress values. This process can effectively identify the stress concentration trend and variation range in the high-stress areas of the downcomer outlet and header welds. By inputting this result into subsequent interpolation modeling and risk assessment processes, quantitative monitoring of the overall boiler heat load distribution is achieved, providing data support for fatigue life assessment, structural anomaly early warning, and dynamic operation and maintenance strategy formulation.

[0090] In S6, stress map reconstruction involves deploying multiple fiber Bragg sensor points in high-stress-sensitive areas such as the boiler drum lower header weld and water-cooled wall nozzle area, and collecting their thermal stress response data in real time. Combining the physical coordinates of the sensor points and their spatial distribution characteristics, a weighted reconstruction interpolation method is used to construct a two-dimensional thermal stress field model. This interpolation method uses the sensor point spacing as the main weighting factor and dynamically adjusts the weights based on historical data fluctuations to ensure improved fitting smoothness in sparse areas and enhanced local gradient response in dense areas, achieving effective fitting of the stress spatial variation trend. The interpolation model is applied to the boiler's two-dimensional structural mesh. Based on the inverse distance weighted interpolation principle, the system dynamically adjusts the interpolation weights according to the reciprocal of the square of the distance from the monitoring point to the target grid point, calculating the thermal stress response value of each grid node point by point, and generating a thermal stress isopleth map. This map visually presents the areas of local stress concentration and the trend of thermal gradient changes in the boiler structure, making it particularly suitable for the qualitative identification and quantitative prediction of fatigue crack initiation areas, providing a decision-making basis for boiler maintenance cycle design and service life management.

[0091] In S7, the abnormal change identification refers to continuously collecting thermal stress time-series data output by fiber optic Bragg sensors for high-risk parts of the boiler structure, such as the heat-affected zone of water-cooled wall welding and the connection between downcomers and headers. The thermal stress change rate criterion formula is used to perform time differentiation processing on the thermal stress curve to obtain the rate curve of stress change over time. By applying sliding window filtering and feature extraction to the change rate and combining it with historical operating data from different parts, differentiated early warning thresholds are established, forming a dynamic response mapping between the thermal stress change rate and the boiler operating status, improving the sensitivity detection capability of abnormal fluctuations. The real-time calculated thermal stress change rate of each monitoring point is compared with the corresponding threshold. When the change rate of a certain node exceeds the set critical criterion, the system immediately calls the alarm logic module to output a structural thermal risk early warning signal. It supports graded responses based on fluctuation amplitude and duration. The early warning module can interface with the boiler operation control system, supporting local interface prompts and remote monitoring linkage, ensuring that signs of sudden structural thermal stress changes are captured and fed back in the first instance, effectively preventing fatigue damage and life loss caused by sudden stress increases, and providing highly timely auxiliary criterion support for safe boiler operation.

[0092] According to the embodiments of this application, a method for measuring thermal stress in boiler structures based on fiber Bragg grating (FBG) technology can be proposed. This method involves deploying decoupled FBG sensors in key stress-bearing areas of the boiler structure and introducing a Bragg reflection center wavelength response model. This enables simultaneous sensing and precise decoupling of temperature and strain under high-temperature and complex environments. This response mechanism, based on the nonlinear coupling relationship between wavelength drift and temperature or strain, effectively avoids the problems of traditional thermocouples and resistance strain gauges being susceptible to temperature drift interference and having limited deployment numbers. This significantly improves the accuracy and distribution integrity of thermal stress identification. Therefore, this method solves many technical obstacles still existing in related technologies, such as the fact that the Bragg wavelength is simultaneously affected by temperature and strain coupling, making it difficult to directly distinguish the causes of thermal stress; the signal superposition and crosstalk problems caused by wavelength division multiplexing (WDM) easily lead to demodulation error accumulation; and traditional center wavelength identification methods struggle to cope with complex background noise and achieve high-precision, high-resolution demodulation. Furthermore, the process of deriving the thermal stress field from the demodulated value lacks a complete and sustainable closed-loop calculation and early warning mechanism.

[0093] Furthermore, Figure 2 This is a schematic diagram of the structure of a boiler structure thermal stress measurement device 10 based on fiber Bragg grating technology, including: a sensor deployment module 100, a spectral signal transmission module 200, a spectral line preprocessing module 300, a center wavelength demodulation module 400, a thermal stress inversion module 500, a stress spectrum reconstruction module 600, and an abnormal change identification module 700.

[0094] Specifically, the sensor deployment module 100 is used to deploy decoupled Bragg grating sensors at key stress-bearing parts of the boiler structure, and to establish a sensing response model based on the Bragg reflection center wavelength formula, for real-time monitoring of temperature and strain changes at the structural location.

[0095] The spectral signal transmission module 200 is used to design multi-channel reflection wavelength distribution through wavelength division multiplexing channel partitioning formula, and to construct a spectral multiplexing transmission link with fixed wavelength intervals, so as to realize the interference-free synchronous transmission of multiple grating signals in a single optical fiber.

[0096] The spectral preprocessing module 300 is used to perform feature enhancement processing on the received reflectance spectral data, including extracting the envelope curve and reflectance peak position information using Hilbert transform, so as to improve the signal-to-noise ratio of subsequent wavelength demodulation.

[0097] The center wavelength demodulation module 400 is used to input the preprocessed reflectance spectrum sequence into the Transformer neural network model constructed based on the multi-head attention mechanism formula to achieve high-resolution center wavelength demodulation and reflectance spectrum peak localization.

[0098] The thermal stress inversion module 500 is used to input the demodulated strain and temperature values ​​into the thermoelastic stress equation for calculation to obtain the thermal stress value corresponding to the sensing point.

[0099] The stress map reconstruction module 600 is used to perform spatial interpolation fitting on the thermal stress data of multiple sensing points using a weighted reconstruction interpolation formula to generate a two-dimensional continuous thermal stress distribution map of the key areas of the boiler structure.

[0100] The abnormal change identification module 700 is used to analyze whether the rate of change of thermal stress over time exceeds a set warning threshold based on the thermal stress change rate criterion formula. If the rate of change exceeds the warning threshold, an alarm signal is triggered.

[0101] Preferably, the sensor deployment module 100 includes: a sensor deployment unit and a real-time monitoring unit.

[0102] The sensor deployment unit is used to select at least one high thermal stress location in the boiler structure, such as near the weld of the lower header of the steam drum, the outlet of the downcomer, or the water-cooled wall pipe connection area, and deploy a decoupled Bragg grating sensor. The decoupled Bragg grating sensor is installed in thermal isolation from the structure through a metal sheath and a high-temperature resistant ceramic gasket to distinguish between thermal strain and structural strain signals and generate a distinction result.

[0103] The real-time monitoring unit is used to establish an analytical mapping model between spectral drift and structural state based on the differentiation results and the Bragg reflection center wavelength formula. Based on the analytical mapping model, it extracts effective reflection information by using spectral line tracking and denoising processing, and matches the temperature and strain dual-channel responses of different deployment points to perform dynamic inversion of stress state and monitor the temperature and strain changes at the structural position in real time.

[0104] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0105] Although embodiments of this application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for measuring thermal stress of a boiler structure based on fiber grating technology, characterized in that, Includes the following steps: S1 Sensor Deployment: Decoupled Bragg grating sensors are deployed at key stress points of the boiler structure, and a sensing response model is established based on the Bragg reflection center wavelength formula to monitor temperature and strain changes at the structural location in real time. S2 spectral signal transmission: By designing a multi-channel reflection wavelength distribution using wavelength division multiplexing channel partitioning formula, and constructing a spectral multiplexing transmission link with fixed wavelength intervals, multiple grating signals can be transmitted synchronously in a single optical fiber without interference. S3 spectral preprocessing: Feature enhancement processing is performed on the received reflectance spectral data, including extracting the envelope curve and reflectance peak position information using Hilbert transform, in order to improve the signal-to-noise ratio of subsequent wavelength demodulation; S4 Center Wavelength Demodulation: The preprocessed reflectance spectrum sequence is input into the Transformer neural network model constructed based on the multi-head attention mechanism formula to achieve high-resolution center wavelength demodulation and reflectance spectrum peak localization; S5 Thermal Stress Inversion: Input the strain and temperature values ​​obtained by demodulation into the thermoelastic stress equation for calculation to obtain the thermal stress value corresponding to the sensing point; S6 Stress Map Reconstruction: Spatial interpolation fitting of thermal stress data from multiple sensing points is performed using a weighted reconstruction interpolation formula to generate a two-dimensional continuous thermal stress distribution map of key areas of the boiler structure. S7 Abnormal Change Identification: Based on the thermal stress change rate criterion formula, analyze whether the rate of change of thermal stress over time exceeds the set warning threshold. If the rate of change exceeds the warning threshold, an alarm signal is triggered.

2. The method for measuring thermal stress in a boiler structure based on fiber Bragg grating technology according to claim 1, characterized in that, In step S1, the installation of decoupled Bragg grating sensors at key stress-bearing locations of the boiler structure, and the establishment of a sensing response model based on the Bragg reflection center wavelength formula, for real-time monitoring of temperature and strain changes at these structural locations, includes: In the boiler structure, at least one high thermal stress location is selected from the area near the weld of the lower header of the steam drum, the outlet of the downcomer, and the water-cooled wall pipe connection area. Decoupled Bragg grating sensors are installed thereon. The decoupled Bragg grating sensors are thermally isolated from the structure by a metal sheath and a high-temperature resistant ceramic gasket, so as to distinguish between thermal strain and structural strain signals and generate a distinction result. Based on the differentiation results and the Bragg reflection center wavelength formula, an analytical mapping model between spectral drift and structural state is established. Based on the analytical mapping model, effective reflection information is extracted by spectral line tracking and denoising processing, and the temperature and strain dual-channel responses of different deployment points are matched to perform dynamic inversion of stress state and monitor the temperature and strain changes at the structural location in real time.

3. The method for measuring thermal stress in a boiler structure based on fiber Bragg grating technology according to claim 1, characterized in that, In S2, the step of designing a multi-channel reflection wavelength distribution using a wavelength division multiplexing channel partitioning formula and constructing a spectral multiplexing transmission link with fixed-interval wavelengths to achieve interference-free synchronous transmission of multiple grating signals in a single optical fiber includes: Based on the wavelength division multiplexing channel partitioning principle, an independent reflection center wavelength is designed for each Bragg grating to generate a multi-channel wavelength group with a fixed interval; Based on the multi-channel wavelength group with fixed intervals, a single-fiber multi-channel spectral multiplexing link is constructed using a multi-channel reflection wavelength scheme. The incident and reflection paths are designed using a grating array, and channel crosstalk is reduced through transmission path temperature control and encapsulation isolation technology, so as to achieve interference-free synchronous transmission of the multiple grating signals in the single optical fiber.

4. The method for measuring thermal stress in a boiler structure based on fiber Bragg grating technology according to claim 1, characterized in that, In step S3, the feature enhancement processing of the received reflectance spectral data includes extracting the envelope curve and reflectance peak position information using Hilbert transform, including: After the original reflectance spectral signal enters the feature enhancement processing module, the original reflectance spectral signal is analyzed using the Hilbert transform envelope extraction formula to extract the instantaneous envelope curve of the original reflectance spectral signal. The reflection center wavelength is identified and located based on the abrupt change point and gain characteristics of the instantaneous envelope curve. The position information of the reflection peak in the spectrum is extracted by using the dynamic threshold method and the envelope peak fitting method.

5. The method for measuring thermal stress in a boiler structure based on fiber Bragg grating technology according to claim 1, characterized in that, In step S4, inputting the preprocessed reflectance spectrum sequence into a Transformer neural network model constructed based on a multi-head attention mechanism to achieve high-resolution center wavelength demodulation and reflectance spectrum peak localization includes: The preprocessed reflection spectrum sequence is input into the Transformer neural network model based on the multi-head attention mechanism. Multiple parallel attention heads extract the structural information of the reflection spectrum in different frequency band feature regions to perform global perception and local enhancement of the full-band signal and construct a nonlinear mapping feature space between reflection intensity and center wavelength. The output of the Transformer neural network model is simultaneously set with a wavelength regression branch and a reflection peak classification branch to locate the center wavelength and identify the main peak of the reflection spectrum. The high-resolution center wavelength demodulation and reflection spectrum peak location are achieved by fusing and analyzing the attention weight results.

6. The method for measuring thermal stress in a boiler structure based on fiber Bragg grating technology according to claim 1, characterized in that, In step S5, the step of inputting the demodulated strain and temperature values ​​into the thermoelastic stress equation for calculation to obtain the thermal stress value corresponding to the sensing point includes: Using the demodulated strain and temperature values ​​as basic physical input parameters, and combining them with the elastic modulus, thermal expansion coefficient and Poisson's ratio of the boiler structural materials, a linear thermoelastic stress estimation model is constructed. Based on the linear thermoelastic stress estimation model, the thermal expansion effect caused by temperature change is coupled with the measured strain to quantitatively calculate the thermal stress value at each monitoring point.

7. The method of claim 1, wherein the method is characterized by: In step S6, the step of spatially interpolating and fitting the thermal stress data of multiple sensing points using a weighted reconstruction interpolation formula to generate a two-dimensional continuous thermal stress distribution map of the key area of ​​the boiler structure includes: Thermal stress measurement data of multiple fiber Bragg sensor points in the key area of ​​the boiler structure were collected. Based on the stress measurement data and physical location coordinates, a thermal stress field model was constructed using a weighted interpolation method. The weights were set according to the sensor point spacing and data stability to fit the thermal stress distribution trend in the two-dimensional area. Based on the thermal stress distribution trend within the two-dimensional region, an interpolation model is applied to the two-dimensional structural mesh of the boiler, and stress values ​​are calculated point by point using the inverse distance weighting principle to generate a two-dimensional continuous thermal stress distribution map of the key areas of the boiler structure.

8. The method for measuring thermal stress in a boiler structure based on fiber Bragg grating technology according to claim 1, characterized in that, In step S7, the step of analyzing whether the rate of change of thermal stress over time exceeds a set warning threshold based on the thermal stress change rate criterion formula, and triggering an alarm signal if the rate of change exceeds the warning threshold, includes: Based on the aforementioned thermal stress change rate criterion formula, a monitoring mechanism is constructed to continuously collect thermal stress time series data of key parts of the boiler, and to perform time differentiation processing on the thermal stress time series data to extract the rate of change of thermal stress. By establishing a correlation mapping between strain changes and actual working conditions through early warning thresholds, local abnormal stress fluctuations can be dynamically identified. The rate of change of thermal stress is compared with the warning threshold so that the alarm signal is triggered when the rate of change at the monitoring point exceeds a preset critical value.

9. A device for measuring thermal stress of a boiler structure based on fiber grating technology, characterized by, include: The sensor deployment module is used to deploy decoupled Bragg grating sensors at key stress-bearing parts of the boiler structure and establish a sensing response model based on the Bragg reflection center wavelength formula to monitor temperature and strain changes at the structural location in real time. The spectral signal transmission module is used to design multi-channel reflection wavelength distributions using wavelength division multiplexing channel partitioning formulas and to construct a spectral multiplexing transmission link with fixed wavelength intervals, enabling interference-free synchronous transmission of multiple grating signals in a single optical fiber. The spectral preprocessing module is used to perform feature enhancement processing on the received reflectance spectral data, including extracting the envelope curve and reflectance peak position information using Hilbert transform, so as to improve the signal-to-noise ratio of subsequent wavelength demodulation. The center wavelength demodulation module is used to input the preprocessed reflectance spectrum sequence into the Transformer neural network model constructed based on the multi-head attention mechanism formula to achieve high-resolution center wavelength demodulation and reflectance spectrum peak localization. The thermal stress inversion module is used to input the demodulated strain and temperature values ​​into the thermoelastic stress equation for calculation, so as to obtain the thermal stress value corresponding to the sensing point. The stress map reconstruction module is used to perform spatial interpolation fitting on the thermal stress data of multiple sensing points using a weighted reconstruction interpolation formula to generate a two-dimensional continuous thermal stress distribution map of the key areas of the boiler structure. The abnormal change identification module is used to analyze whether the rate of change of thermal stress over time exceeds a set warning threshold based on the thermal stress change rate criterion formula. If the rate of change exceeds the warning threshold, an alarm signal is triggered.

10. The apparatus for measuring thermal stress of a boiler structure based on fiber grating technology according to claim 9, wherein, The sensor deployment module includes: The sensor deployment unit is used to select at least one high thermal stress location in the boiler structure, such as near the weld of the lower header of the steam drum, the outlet of the downcomer, or the water-cooled wall pipe connection area, and deploy a decoupled Bragg grating sensor. The decoupled Bragg grating sensor is installed in thermal isolation from the structure through a metal sheath and a high-temperature resistant ceramic gasket to distinguish between thermal strain and structural strain signals and generate a distinction result. The real-time monitoring unit is used to establish an analytical mapping model between spectral drift and structural state based on the differentiation results and the Bragg reflection center wavelength formula. Based on the analytical mapping model, it extracts effective reflection information by using spectral line tracking and denoising processing, and matches the temperature and strain dual-channel responses of different deployment points to perform dynamic inversion of stress state and monitor the temperature and strain changes at the structural position in real time.

Images