Thermal imaging based on-line monitoring device for thermal stress of pipeline

Abstract

The present application relates to the technical fields of pipeline thermal stress online monitoring, and provides a pipeline thermal stress online monitoring device based on thermal imaging, which comprises a device and a control device connected by communication, wherein the device comprises an infrared thermal imaging assembly and a regulating valve; the control device acquires an infrared radiation image sequence and collects fluid operation state parameters, extracts fluctuation frequency and initial phase generated by internal leakage disturbance by frequency domain conversion; a stress deduction module maps the image into a three-dimensional temperature field distribution matrix and calculates an alternating thermal stress amplitude; when the alternating thermal stress amplitude reaches a threshold value, a vibration suppression control module drives the regulating valve to generate an anti-phase interference wave according to phase lead compensation amount, offsets mechanical oscillation of a target pipe section, and outputs a warning information; the present application realizes real-time monitoring and active intervention on pipeline thermal stress risk.

Description

Technical Field

[0001] This invention relates to the field of pipeline thermal stress online monitoring technology, and in particular to a pipeline thermal stress online monitoring device based on thermal imaging. Background Technology

[0002] The pipeline thermal stress online monitoring device based on thermal imaging integrates infrared thermal imaging, data analysis, and early warning functions. Its core principle is to obtain continuous temperature distribution images by non-contact scanning of the pipeline surface with an infrared thermal imager, and then calculate the thermal stress field distribution of the pipeline in real time based on the physical relationship model between temperature and stress, so as to realize continuous online monitoring and visualization of the thermal stress state of the entire pipeline section.

[0003] Existing technologies for monitoring the thermal stress state of high-pressure steam heating pipeline systems in gas turbine power plants suffer from the following technical challenges: Currently, they primarily rely on periodic manual inspections and point temperature measurements, lacking the means to perform online, real-time, and visual imaging and analysis of pipe wall temperature field distortions caused by internal leaks in steam traps and regulating valves. For example, when an internal leak occurs in the electric valve before the high-pressure heating desuperheater, the leaking steam continuously washes over a specific pipe section, creating a localized low-temperature zone with a significant temperature gradient between this zone and adjacent sections heated by normal high-temperature steam. The non-uniform temperature distribution caused by the internal leak creates cyclic thermal stress within the pipe metal. Under long-term action, micro-cracks easily develop and gradually propagate at stress concentration points. Existing methods cannot capture and locate such dynamic temperature distribution anomalies and thermal stress concentration areas in real time. Operators struggle to detect potential hazards until leaks or crack propagation lead to non-shutdown accidents, posing significant safety and economic risks. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an online monitoring device for pipeline thermal stress based on thermal imaging. This invention solves the technical problem that the lack of online, real-time, and visual monitoring methods for uneven pipe wall temperature distribution and thermal stress concentration caused by internal leakage of valves in high-pressure steam pipelines makes it impossible to provide timely early warning and location of the risk of pipeline thermal stress crack initiation and propagation.

[0005] To solve the above-mentioned technical problems, the specific contents of the present invention are as follows: The present invention provides a pipeline thermal stress online monitoring device based on thermal imaging, comprising an equipment device and a control device communicatively connected to the equipment device; the equipment device includes an infrared thermal imaging component and a regulating valve; The control device includes: The data acquisition module is used to acquire an infrared radiation image sequence of the outer surface of the target pipe section through the infrared thermal imaging component, and to acquire fluid operation status parameters of the fluid inside the target pipe section. The parameter extraction module is used to receive the infrared radiation image sequence and perform frequency domain conversion to obtain a frequency domain signal, and extract the fluctuation frequency and initial phase from the frequency domain signal; The stress estimation module is used to receive the infrared radiation image sequence and the fluctuation frequency, map the infrared radiation image sequence into a three-dimensional temperature field distribution matrix inside the target pipe section, and calculate the alternating thermal stress amplitude corresponding to the three-dimensional temperature field distribution matrix. The vibration suppression control module is used to read the fluid operating state parameters and the initial phase when the alternating thermal stress amplitude reaches the preset thermal stress amplitude threshold, calculate the phase advance compensation amount, generate a feedforward pulse signal corresponding to the phase advance compensation amount, send the feedforward pulse signal to the regulating valve, and output a warning message.

[0006] Furthermore, in the thermal imaging pipeline thermal stress online monitoring system of the present invention, the parameter extraction module includes: The registration unit is used to extract the pixel coordinates in the infrared radiation image sequence and perform spatial registration calculations, and output a time-series temperature data matrix. The frequency domain conversion unit is used to receive the time-series temperature data matrix, perform a fast Fourier transform on the pixel points corresponding to the pixel point coordinates in the time-series temperature data matrix along the time axis, and convert them into the frequency domain signal. The parametric stripping unit is used to receive the frequency domain signal, filter out the low-frequency background component in the frequency domain signal using a preset low-pass cutoff frequency to obtain a retained frequency band signal, extract the dominant excitation frequency as the fluctuation frequency from the retained frequency band signal, and extract the fluctuation amplitude and the initial phase corresponding to the fluctuation frequency.

[0007] Furthermore, in the thermal imaging pipeline thermal stress online monitoring system of the present invention, the stress estimation module includes: The reverse mapping unit is used to input the infrared radiation image sequence into the heat conduction inverse problem solving algorithm model to derive the three-dimensional temperature field distribution matrix inside the target pipe segment; Tensor analysis unit is used to receive the three-dimensional temperature field distribution matrix and integrate the input material properties to establish a thermoelastic fluid-structure interaction model, and calculate the thermal stress tensor inside the target pipe section. An amplitude extraction unit is used to receive the thermal stress tensor, extract the peak-to-valley difference in the thermal stress tensor as the alternating thermal stress amplitude, and send the alternating thermal stress amplitude to the vibration suppression control module.

[0008] Furthermore, in the thermal imaging pipeline thermal stress online monitoring system of the present invention, the vibration suppression control module includes: The delayed solution unit is used to extract the pressure parameters, temperature parameters, and transient mass flow rate parameters from the fluid operating state parameters, and calculate the fluid velocity through fluid dynamics equations. The time calculation unit is used to receive the fluid flow rate, extract the physical pipe length between the regulating valve and the target pipe section, and divide the physical pipe length by the fluid flow rate to obtain the transmission delay time of the thermoelastic wave.

[0009] Furthermore, in the thermal imaging pipeline thermal stress online monitoring system of the present invention, the vibration suppression control module further includes: A forward calculation unit is used to receive the transmission delay time and the fluctuation frequency, and multiply the fluctuation frequency by the transmission delay time to output the natural phase lag. The estimation compensation unit is used to receive the natural phase lag, add the natural phase lag to a preset anti-phase constant, and output the phase advance compensation amount corresponding to the regulating valve.

[0010] Furthermore, in the thermal imaging pipeline thermal stress online monitoring system of the present invention, the control device further includes: The feedback receiving module is used to collect residual surface thermal fluctuation data of the outer surface of the target pipe section through the infrared thermal imaging component after the regulating valve receives the feedforward pulse signal. An adaptive fine-tuning module is used to receive the residual surface thermal fluctuation data and input it into the extreme value optimization model, and output the phase correction amount; The instruction issuing module is used to receive the phase correction amount and adjust the waveform properties of the feedforward pulse signal using the phase correction amount.

[0011] Furthermore, in the thermal imaging pipeline thermal stress online monitoring system of the present invention, the control device further includes: The source comparison unit is used to input the fluctuation frequency into the internal leakage excitation feature library for feature value comparison and output the internal leakage source location information. The positioning and rendering unit is used to receive the internal leakage source location information, calibrate the coordinates corresponding to the internal leakage source location information on a preset three-dimensional thermal stress cloud map, and output the warning information including the internal leakage source location information.

[0012] Furthermore, in the thermal imaging pipeline thermal stress online monitoring system of the present invention, the control device includes a coordinate system transformation unit, used to extract the pixel coordinate system parameters of the infrared thermal imaging component and the physical three-dimensional coordinate system parameters of the pipeline, establish a data mapping matrix from the pixel coordinate system parameters of the infrared thermal imaging component to the physical three-dimensional coordinate system parameters of the pipeline, and use the data mapping matrix to map the infrared radiation image sequence to the physical three-dimensional coordinate system of the pipeline.

[0013] Furthermore, in the thermal imaging pipeline thermal stress online monitoring system of the present invention, the internal connection relationship of the adaptive fine-tuning module includes: The extreme value optimization model receives the residual surface thermal fluctuation data collected by the feedback receiving module; The extreme value optimization model performs derivative calculations on the residual surface thermal fluctuation data of the continuous acquisition period to obtain the partial derivative of mechanical oscillation change, and calculates and outputs the phase correction amount based on the partial derivative of mechanical oscillation change. The instruction issuing module receives the phase correction amount output by the extreme value optimization model and sends the feedforward pulse signal superimposed with the phase correction amount to the regulating valve.

[0014] Furthermore, in the thermal imaging pipeline thermal stress online monitoring system of the present invention, the vibration suppression control module extracts the alternating thermal stress amplitude, multiplies the alternating thermal stress amplitude using a preset proportional coefficient to obtain the opening adjustment amount, and sends the feedforward pulse signal including the opening adjustment amount information to the regulating valve.

[0015] Beneficial effects of this invention; The pipeline thermal stress online monitoring device based on thermal imaging described in this invention acquires infrared radiation image sequences non-contactly through an infrared thermal imaging component. Combined with a parameter extraction module, it performs in-depth analysis of the frequency domain signal, enabling accurate identification and extraction of transient hot spot characteristics caused by valve leakage in high-pressure steam pipelines. The stress inference module uses a heat conduction inverse problem solving algorithm to map two-dimensional surface temperature into a three-dimensional temperature field distribution matrix. It also integrates material properties and quantifies the amplitude of internal alternating thermal stress through a thermoelastic fluid-structure interaction model. This overcomes the technical limitations of existing point temperature monitoring, which cannot capture the stress concentration state inside the pipe wall, significantly improving the early diagnosis accuracy of crack initiation risk. The vibration suppression control module drives the regulating valve to generate anti-phase interference waves based on real-time calculated phase advance compensation, actively canceling the mechanical oscillations excited by leakage at the physical source, effectively delaying the accumulation of thermal fatigue damage to the pipeline metal. The collaborative operation of the source tracing and comparison unit and the positioning rendering unit can transform abstract frequency characteristics into intuitive leakage source location information, guiding operators to implement precise maintenance, reducing the risk of unplanned downtime while ensuring the inherent safety and economical operation of the pipeline system. Attached Figure Description

[0016] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on the drawings without creative effort.

[0017] Figure 1This is a system architecture diagram of the pipeline thermal stress online monitoring device based on thermal imaging according to the present invention. Detailed Implementation

[0018] To make the technical solution of the present invention clearer, the present invention will be clearly and completely described below with reference to specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. The present invention provided by various embodiments will be described in detail below with reference to the accompanying drawings. To better understand the purpose of the present invention, the present invention will be described in further detail below.

[0019] The present invention provides a pipeline thermal stress online monitoring device based on thermal imaging, comprising an equipment device and a control device communicatively connected to the equipment device; the equipment device includes an infrared thermal imaging component and a regulating valve; The control device includes: The data acquisition module is used to acquire an infrared radiation image sequence of the outer surface of the target pipe section through the infrared thermal imaging component, and to acquire fluid operation status parameters of the fluid inside the target pipe section. The parameter extraction module is used to receive the infrared radiation image sequence and perform frequency domain conversion to obtain a frequency domain signal, and extract the fluctuation frequency and initial phase from the frequency domain signal; The stress estimation module is used to receive the infrared radiation image sequence and the fluctuation frequency, map the infrared radiation image sequence into a three-dimensional temperature field distribution matrix inside the target pipe section, and calculate the alternating thermal stress amplitude corresponding to the three-dimensional temperature field distribution matrix. The vibration suppression control module is used to read the fluid operating state parameters and the initial phase when the alternating thermal stress amplitude reaches the preset thermal stress amplitude threshold, calculate the phase advance compensation amount, generate a feedforward pulse signal corresponding to the phase advance compensation amount, send the feedforward pulse signal to the regulating valve, and output a warning message.

[0020] The thermal imaging-based online monitoring device for pipeline thermal stress provided by this invention is deployed in the high-pressure steam heating pipeline system of a gas turbine power plant. The overall system architecture is divided into equipment installed on the pipeline and control devices located in a remote control room. The equipment includes infrared thermal imaging components fixed to the outside of the pipeline and regulating valves arranged in the upstream pipeline of the target section. An industrial Ethernet communication channel is established between the control device and the equipment for real-time data exchange. The system architecture of the control device includes a data acquisition module, a parameter extraction module, a stress estimation module, and a vibration suppression control module.

[0021] The data acquisition module, acting as the system's sensing front end, utilizes an infrared thermal imaging component to perform non-contact continuous scanning of the target pipe section downstream of the high-pressure bypass valve, which is prone to internal leakage. The infrared thermal imaging component acquires the infrared radiation intensity signal from the outer surface of the target pipe section and converts it into an infrared radiation image sequence including timestamps and two-dimensional spatial pixel coordinates. This infrared radiation image sequence reflects the dynamic temperature fluctuations on the outer wall of the target pipe section caused by internal fluid scouring. The data acquisition module simultaneously connects to the power plant's distributed control system network to read the real-time fluid operating status parameters of the steam fluid inside the target pipe section. These fluid operating status parameters specifically include real-time pressure parameters, instantaneous temperature parameters, and transient mass flow rate parameters of the main steam pipeline network. These fluid operating status parameters provide the fundamental boundary thermodynamic boundary condition inputs for subsequent thermoelastic dynamics calculations.

[0022] The parameter extraction module receives the infrared radiation image sequence transmitted by the data acquisition module. In the time domain, the infrared radiation image sequence represents a continuously fluctuating stream of temperature data. The parameter extraction module transforms the infrared radiation image sequence from the time domain to the frequency domain, performing a Fast Fourier Transform (FFT) on the pixel temperature fluctuation data along the time axis to generate a frequency domain signal reflecting the periodic temperature change. The frequency domain signal exposes high-frequency, minute thermal pulse components hidden against a macroscopic, slowly heating background. The parameter extraction module performs peak search within the high-frequency band of the frequency domain signal, extracting the transient hotspot features caused by the high-pressure steam-water turbulence impacting the pipe wall due to the internal leakage valve. The parameter extraction module identifies the dominant excitation frequency corresponding to the transient hotspot features as the fluctuation frequency and extracts the initial phase corresponding to the dominant excitation frequency. The fluctuation frequency and initial phase quantify the alternating pattern of high-frequency thermal shock induced by fluid disturbance on the outer surface of the target pipe section.

[0023] The stress deduction module synchronously receives the infrared radiation image sequence and the fluctuation frequency output by the parameter extraction module. Using the surface temperature data from the infrared radiation image sequence as known external boundary input conditions, the module substitutes this data into the heat conduction inverse problem solving algorithm model for inverse spatial derivation, calculating the three-dimensional temperature field distribution matrix inside the metal pipe wall of the target pipe segment. This three-dimensional temperature field distribution matrix reflects the spatial temperature gradient distribution from the outer surface to the inner surface of the pipe wall. The stress deduction module establishes a thermoelastic-fluid-structure interaction model based on the material's elastic modulus and Poisson's ratio. Using this model, the module performs finite element nodal stress calculations on the three-dimensional temperature field distribution matrix, deriving the dynamic thermal stress tensor inside the target pipe segment. The module extracts the stress peak and valley values ​​at various time points from the dynamic thermal stress tensor, defining the difference between the peak and valley values ​​as the alternating thermal stress amplitude. This alternating thermal stress amplitude characterizes the severity of the material's internal high-frequency thermal fatigue alternating load.

[0024] The vibration suppression control module has a preset thermal stress amplitude threshold that characterizes the fatigue limit warning boundary of metallic materials. The module compares the real-time input alternating thermal stress amplitude with the preset thermal stress amplitude threshold. When the alternating thermal stress amplitude reaches the preset threshold, the system determines that there is a risk of high-frequency thermal stress crack initiation within the target pipe section. The module then reads the fluid operating state parameters and initial phase, and calculates the current fluid velocity within the pipe based on these parameters. Combining the physical distance between the regulating valve and the target pipe section with the fluid velocity, the module deduces the thermodynamic wave propagation delay time. Based on the thermodynamic wave propagation delay time and wave frequency, the module calculates the natural phase lag state when the wave reaches the downstream end, and superimposes an absolute reverse cancellation condition on top of this natural phase lag state to determine the phase lead compensation amount that needs to be applied at the upstream regulating valve. The module then generates a feedforward pulse signal with a specific opening fluctuation pattern based on the phase lead compensation amount. The vibration suppression control module sends feedforward pulse signals to the regulating valve, driving it to perform micro-switching actions according to the set pulsation frequency and opening degree. The periodic action of the regulating valve excites dynamic thermal interference waves in the fluid medium. When these dynamic thermal interference waves travel downstream to the target pipe section, they interact with the thermal shock caused by internal leakage in an anti-phase interference canceling state, thus canceling the high-frequency mechanical oscillations of the target pipe section online. Simultaneously with sending the feedforward pulse signals, the vibration suppression control module packages the data on excessive alternating thermal stress amplitude and the location data of the target pipe section to generate early warning information, which is then pushed to the human-machine interface in the main control room.

[0025] The registration unit within the parameter extraction module receives the infrared radiation image sequence transmitted by the infrared thermal imaging component and extracts the two-dimensional pixel coordinates from the infrared radiation image sequence. Addressing the physical vibration phenomenon accompanying the desuperheater and pressure reducer pipeline in a gas turbine power plant under high-temperature and high-pressure steam scouring, the registration unit performs spatial registration calculations on the pixel coordinates of adjacent frames to eliminate field-of-view pixel offset, outputting a spatially aligned time-series temperature data matrix. The time-series temperature data matrix includes continuous temperature change values ​​in a three-dimensional tensor format. The frequency domain conversion unit receives the time-series temperature data matrix and performs a Fast Fourier Transform along the time axis on the pixels corresponding to the pixel coordinates in the time-series temperature data matrix, converting the temperature change based on the time-series distribution into a frequency domain signal based on the frequency distribution. The frequency domain signal reflects the intensity of heat pulsation at different frequency periods. The parameter stripping unit receives the frequency domain signal. To eliminate baseline drift caused by the slow rise in workshop ambient temperature, the parameter stripping unit uses a preset low-pass cutoff frequency to filter the frequency domain signal, removing low-frequency background components and obtaining the retained frequency band signal. The parametric stripping unit performs a peak search algorithm within the high-frequency range included in the retained frequency band signal, locks the dominant excitation frequency corresponding to the energy peak point as the fluctuation frequency that causes hot spots on the pipe wall, and extracts the fluctuation amplitude and initial phase corresponding to the fluctuation frequency from the frequency domain signal.

[0026] The stress estimation module calls the inverse mapping unit to receive the surface temperature dataset included in the infrared radiation image sequence. The inverse mapping unit inputs the infrared radiation image sequence into the heat conduction inverse problem solving algorithm model as the external heat transfer boundary condition. The heat conduction inverse problem solving algorithm model derives the temperature gradient at various depth levels inside the target pipe section based on partial differential equations, and derives the three-dimensional temperature field distribution matrix inside the target pipe section. The three-dimensional temperature field distribution matrix expands the original two-dimensional temperature monitoring limited to the surface to quantitative data of the spatial temperature inside the pipe wall. The tensor analysis unit receives the three-dimensional temperature field distribution matrix and integrates the pre-input material properties such as the elastic modulus and Poisson's ratio of the pipe metal material to establish a thermoelastic-fluid-structure interaction model. The tensor analysis unit uses the thermoelastic-fluid-structure interaction model to apply finite element mesh node operations to the three-dimensional temperature field distribution matrix, transforming the non-uniform expansion deformation of the spatial temperature gradient into mechanical parameters, and calculating the thermal stress tensor representing the three-dimensional stress state inside the target pipe section. Under the condition of continuous internal leakage of the valve causing alternating heating and cooling of the pipe wall, the amplitude extraction unit receives the thermal stress tensor and compares the continuously changing thermal stress tensor values ​​along the time axis. The amplitude extraction unit extracts the peak and valley values ​​of the thermal stress tensor in the periodic cycle, calculates the peak-valley difference between the peak and valley values ​​as the alternating thermal stress amplitude, and sends the alternating thermal stress amplitude to the vibration suppression control module for subsequent mechanical damage assessment.

[0027] The vibration suppression control module initiates a delayed solution unit to intervene in the data processing flow mechanism. The delayed solution unit connects to the power plant's industrial network to read fluid operating state parameters, extracting pressure, temperature, and transient mass flow rate parameters. In a steam pipeline network operation scenario, steam density changes drastically with pressure and temperature parameters. The delayed solution unit substitutes the pressure and temperature parameters into the state equation to obtain the real-time steam density, and combines this with the transient mass flow rate parameter and the pipe's internal cross-sectional area to calculate the real-time fluid velocity using fluid dynamics equations. The time estimation unit receives the fluid velocity value, retrieves pre-stored pipeline mapping information from the system database, and extracts the actual physical pipeline length for steam flow between the regulating valve and the target pipe section. The time estimation unit divides the physical pipeline length by the dynamically changing fluid velocity to calculate the transmission delay time consumed by the thermoelastic wave caused by the upstream valve's action to travel along the pipeline fluid to the downstream target pipe section.

[0028] The vibration suppression control module utilizes the advance calculation unit to receive the transmission delay time and the fluctuation frequency extracted from the front end. The advance calculation unit performs multiplication logic, multiplying the fluctuation frequency by the transmission delay time to calculate the natural phase lag of the thermoelastic wave during pipeline transmission due to time passage. The natural phase lag reflects the angular shift caused by the waveform's spatial position change. The prediction compensation unit receives the natural phase lag data. In active vibration suppression scenarios, it needs to generate a waveform completely opposite to the internal leakage heat shock to achieve interference cancellation. The prediction compensation unit adds the natural phase lag to a preset anti-phase constant characterizing the peak-to-trough reversal property, deriving the phase advance compensation amount required to trigger the control valve action command relative to the current moment. The prediction compensation unit outputs the corresponding phase advance compensation amount for the control valve to the command synthesis layer to guide the system in arranging the pulse control timing.

[0029] After the control device sends a feedforward pulse signal to the regulating valve via a hardware interface, the feedback receiving module continuously monitors the system. Once the regulating valve receives the feedforward pulse signal and executes the corresponding valve stem mechanical pulsation action, the feedback receiving module continuously acquires residual surface thermal fluctuation data of the target pipe section's outer surface after waveform interference via an infrared thermal imaging component. This residual surface thermal fluctuation data reflects the minute temperature fluctuations that the active interference action failed to completely cancel out. The adaptive fine-tuning module receives the residual surface thermal fluctuation data and sets it as the objective function to be minimized, inputting it into the extreme value optimization model. The extreme value optimization model performs iterative calculations in a complex pipe network environment with nonlinear fluid resistance, outputting a phase correction amount to correct errors. The command issuing module receives the phase correction amount output by the extreme value optimization model and uses it to reshape the triggering timing of the feedforward pulse signal, dynamically adjusting the waveform properties of the feedforward pulse signal.

[0030] The control device's built-in source tracing and comparison unit connects to the power plant's historical defect database and inputs the fluctuation frequency output by the parameter extraction module into the internal leakage excitation feature library for feature value comparison. The internal leakage excitation feature library stores fixed excitation spectrum templates generated by minor leaks in different types of valves under specific pressure differentials. The source tracing and comparison unit calculates the similarity index between the fluctuation frequency and each spectrum template, selects the template with the highest similarity, and outputs the specific valve number that induced the mechanical oscillation as the internal leakage source location information. The location rendering unit receives the internal leakage source location information, calls the graphics rendering engine to load the structural model of the target pipe section and surrounding valves, and maps the coordinates corresponding to the internal leakage source location information onto a preset three-dimensional thermal stress cloud map. The location rendering unit highlights specific valve nodes in the three-dimensional thermal stress cloud map using different contrast colors, combines the three-dimensional thermal stress cloud map image with text numbers to generate and output warning information including the internal leakage source location information.

[0031] The control unit invokes the coordinate system transformation unit to process the spatial correspondence between the infrared equipment and the physical pipeline. The coordinate system transformation unit extracts the pixel coordinate system parameters of the infrared thermal imaging component and simultaneously reads the pipeline's physical three-dimensional coordinate system parameters established based on the power plant design drawings. For the scenario of significant thermal expansion displacement of gas turbine power plant pipelines under high-temperature steam heating, the coordinate system transformation unit calculates the spatial rotation matrix and translation vector in conjunction with the constraint boundary conditions of the pipeline supports, establishing a data mapping matrix that maps the infrared thermal imaging component pixel coordinate system parameters to the pipeline's physical three-dimensional coordinate system parameters. The coordinate system transformation unit uses the data mapping matrix to perform spatial coordinate transformation on the acquired infrared radiation image sequence, directly mapping and projecting the grayscale values ​​of the two-dimensional pixels in the infrared radiation image sequence onto the corresponding entity coordinate nodes in the pipeline's physical three-dimensional coordinate system.

[0032] The adaptive fine-tuning module internally constructs a closed-loop connection topology logic based on data flow drive. The extreme value optimization model within the adaptive fine-tuning module receives residual surface thermal fluctuation data continuously collected by the feedback receiving module. The extreme value optimization model incorporates gradient descent algorithm logic to extract the residual surface thermal fluctuation data record matrix for the continuous acquisition cycle. It then performs derivative operations along the time dimension on the residual surface thermal fluctuation data for the continuous acquisition cycle to obtain the partial derivative of mechanical oscillation change, reflecting the trend of oscillation decay or intensification. The extreme value optimization model determines the optimization step size based on the sign and absolute value of the partial derivative of mechanical oscillation change, and calculates and outputs a phase correction amount to compensate for the nonlinear resistance of the pipeline network based on the partial derivative of mechanical oscillation change. The command issuing module receives the phase correction amount output in real time from the extreme value optimization model, performs algebraic addition processing on the original pulse control timing and the phase correction amount, and sends the feedforward pulse signal with the superimposed phase correction amount to the regulating valve to execute the compensation pulsation action.

[0033] The vibration suppression control module performs quantitative adjustment of the energy intensity of the offset pulse signal. It extracts the amplitude of alternating thermal stress and analyzes the absolute magnitude of the fatigue load currently borne by the pipe wall. For high-pressure steam heating pipe sections with different diameters and wall thicknesses, the module retrieves a preset proportional coefficient calibrated experimentally for the specific pipe section material. Using this preset proportional coefficient, the module multiplies the alternating thermal stress amplitude, linearly amplifying the thermodynamic stress value and converting it into a mechanical opening control value to obtain the opening adjustment amount. The module encodes this opening adjustment amount information into the amplitude parameters of the pulse waveform, synthesizes a feedforward pulse signal including the opening adjustment amount information, and sends this feedforward pulse signal to the regulating valve.

[0034] To address the potential hazards encountered in the operation of high-pressure steam heating pipeline systems in gas turbine power plants, existing monitoring methods, which rely excessively on periodic manual inspections and point temperature measurements, cannot effectively identify pipe wall temperature field distortions caused by leaks in steam traps or regulating valves. This leads to a failure to provide early warnings of the risk of thermal stress cracks caused by localized temperature gradients. The present invention provides a pipeline thermal stress online monitoring device based on thermal imaging. This device integrates an infrared thermal imaging component, a regulating valve, and a control device with deep analysis capabilities to construct a closed-loop system capable of capturing dynamic temperature distribution anomalies in real time and implementing proactive vibration suppression intervention. The overall system architecture uses the field-installed infrared thermal imaging component and the regulating valve located upstream of the target pipe section as the execution front end. The logical flow from physical perception to risk control is achieved through data acquisition, parameter extraction, stress estimation, and vibration suppression control modules within the control device.

[0035] The data acquisition module, acting as the sensing entry point, drives the infrared thermal imaging component to perform non-contact scanning of key pipe sections such as the electric valve before the high-pressure heating desuperheater. It converts the acquired infrared radiation intensity signals into an infrared radiation image sequence including timestamps and two-dimensional coordinates, thus characterizing the dynamic temperature fluctuations on the outer surface of the pipe wall caused by internal steam scouring. To support subsequent mechanical modeling, the data acquisition module simultaneously connects to the power plant control network to read fluid operating state parameters, including pressure, temperature, and transient mass flow rate parameters, establishing boundary conditions for thermoelastic dynamics calculations. After receiving the infrared radiation image sequence, the parameter extraction module uses a registration unit to spatially register adjacent frames to eliminate pixel offsets caused by pipe physical vibrations, generating a spatially aligned time-series temperature data matrix. The frequency domain conversion unit further performs a fast Fourier transform along the time axis on the time-series temperature data matrix, mapping the complex time-domain temperature fluctuations into frequency-domain signals. The parameter stripping unit uses a preset low-pass cutoff frequency to filter out the ambient heating background and pinpoint the fluctuation frequency, initial phase, and fluctuation amplitude caused by valve leakage.

[0036] The stress derivation module receives infrared radiation image sequences through the inverse mapping unit, uses them as external boundary conditions and substitutes them into the heat conduction inverse problem solving algorithm model to deduce the temperature gradient distribution at various depth levels inside the pipe wall, forming a three-dimensional temperature field distribution matrix. The tensor analysis unit integrates material properties such as the thermal expansion coefficient, elastic modulus, and Poisson's ratio of the pipe material, and uses a thermoelastic fluid-structure interaction model to transform the non-uniform temperature drop in the three-dimensional temperature field distribution matrix into a mechanical response, calculating the thermal stress tensor characterizing the three-dimensional stress state. The amplitude extraction unit tracks the extreme value changes of the thermal stress tensor in periodic cycles, calculates the peak-to-valley difference between the peak and valley values ​​as the alternating thermal stress amplitude, and directly quantifies the severity of thermal fatigue loads borne by the material.

[0037] When the vibration suppression control module detects that the alternating thermal stress amplitude reaches the preset thermal stress amplitude threshold, it determines that there is a risk of crack initiation. It then activates the delay calculation unit to calculate the real-time steam density based on pressure and temperature parameters, and derives the fluid velocity using transient mass flow rate parameters. The time estimation unit divides the physical pipe length between the regulating valve and the target pipe section by the fluid velocity to obtain the transmission delay time of the thermoelastic wave to the downstream. The lead calculation unit multiplies the fluctuation frequency by the transmission delay time to obtain the natural phase lag, which is then superimposed by the preset anti-phase constant by the prediction compensation unit to calculate the phase lead compensation amount that needs to be applied at the upstream regulating valve. The command issuing module synthesizes a feedforward pulse signal based on the phase lead compensation amount and the opening adjustment amount calculated from the preset proportional coefficient, driving the regulating valve to perform a micro-pulsation action to generate an anti-phase interference wave, thus canceling the thermoelastic mechanical oscillation of the downstream pipe section.

[0038] The system continuously monitors residual surface thermal fluctuation data after interference through the feedback receiving module. It uses the extreme value optimization model in the adaptive fine-tuning module to perform derivative calculations on the data to obtain the partial derivatives of mechanical oscillation changes, thereby fine-tuning the phase correction amount in real time and reshaping the waveform properties of the feedforward pulse signal. The source tracing and comparison unit matches the real-time fluctuation frequency with the internal leakage excitation feature library, determining the specific valve number that induced the anomaly as the internal leakage source location information. The positioning and rendering unit maps this information onto a preset three-dimensional thermal stress cloud map for highlighting. The coordinate system transformation unit is responsible for establishing a mapping matrix between infrared pixel coordinates and the physical three-dimensional coordinates of the pipeline, compensating for the spatial displacement of the target pipe segment caused by thermal expansion, and ensuring that the infrared radiation image sequence can be accurately projected onto the physical coordinate nodes.

[0039] The infrared radiation image sequence involved in this invention is a set of two-dimensional data matrices generated by continuously scanning the surface of a target pipe segment using a high-frequency infrared thermal imaging component. Each pixel carries a radiation intensity value reflecting the thermal state of the pipe wall. To accurately reconstruct the dynamic trajectory of temperature fluctuations in the time domain, the infrared radiation image sequence includes millisecond-accurate timestamp information and corresponding two-dimensional spatial pixel coordinates, enabling the system to capture minute thermodynamic disturbances caused by internal steam leakage. By performing a linear mapping from grayscale to temperature on the infrared radiation image sequence, the abstract infrared radiation intensity can be transformed into an intuitive surface temperature field distribution.

[0040] Fluid operating state parameters are the core real-time dataset describing the physical properties of the medium inside the target pipe section, directly provided by the power plant's distributed control system network. Specifically, these parameters encompass the main steam network pressure parameters, instantaneous temperature parameters, and transient mass flow rate parameters. Pressure and instantaneous temperature parameters together determine the physical density and enthalpy of the high-pressure steam under current operating conditions, while the transient mass flow rate parameter reflects the macroscopic motion characteristics of the medium. These parameters provide essential thermodynamic boundary inputs for the subsequent thermoelastic dynamics model, directly impacting the accuracy of subsequent derivations of flow velocity and stress transmission delay.

[0041] The frequency domain signal is transformed domain data generated by the parameter extraction module through a Fast Fourier Transform (FFT) process on the time-series temperature data matrix. This process separates the originally chaotic, alternating temperature fluctuations over time into spectral density information based on frequency distribution. The frequency domain signal can effectively separate the slow thermal drift component caused by the background environment from the high-frequency excitation component induced by valve leakage. By analyzing the amplitude and phase spectra in the frequency domain signal, the dominant excitation frequency with concentrated energy can be identified, providing a digital basis for identifying the periodic thermal shocks generated by turbulent fluid flow on the pipe wall.

[0042] Fluctuation frequency is a technical indicator that measures the rate of thermal pulsation caused by internal leakage steam in a target pipe section. It is extracted from the energy peak point of the frequency domain signal. Fluctuation frequency represents the characteristic frequency of fluid disturbance to the pipe wall and is closely related to the orifice size of the internal leakage point, the steam pressure difference, and the turbulence intensity. As the basic input frequency for subsequent active vibration suppression algorithms, fluctuation frequency determines the triggering pace of the upstream control valve's pulsation interference action.

[0043] The fluctuation amplitude is defined as the absolute value of the difference between the peaks and troughs of temperature fluctuations at the fluctuation frequency, reflecting the intensity of thermal shock energy caused by internal leakage. The fluctuation amplitude is positively correlated with the amplitude of the thermal gradient inside the pipe wall, directly determining the quantitative value of alternating thermal stress. The fluctuation amplitude provides a quantitative reference for the amplitude of the pulse waveform adjusted by the vibration suppression control module, used to determine how much energy of interference wave needs to be excited to counteract mechanical oscillations.

[0044] The initial phase describes the initial angular position of the thermal pulse waveform at the start of data acquisition and serves as the time reference point for establishing the interference cancellation logic. Combining the initial phase with the transmission delay time, the instantaneous phase state of the thermoelastic wave upon reaching the downstream target pipe section can be derived. The accuracy of the initial phase directly affects whether the reverse interference wave can achieve absolute anti-phase superposition with the original thermal shock wave in spatial location.

[0045] The three-dimensional temperature field distribution matrix is ​​a set of temperature data along the entire thickness of the pipe wall, derived by the stress extrapolation module using an inverse heat conduction problem-solving algorithm. This extends the monitoring range from the outer surface of the pipe wall to the inner surface. The three-dimensional temperature field distribution matrix includes the transient temperature gradient of the target pipe segment on the three-dimensional spatial grid nodes. This spatialized temperature distribution data is the direct physical source for calculating the distribution of internal thermal stress in the metal, reflecting the impact of non-uniform cooling caused by internal leakage on the deep structure of the pipe wall.

[0046] Alternating thermal stress amplitude is a key mechanical indicator for evaluating the degree of thermal fatigue damage suffered by the metal material of a target pipe section. It is derived from the peak-to-valley difference of the thermal stress tensor during a periodic cycle. The alternating thermal stress amplitude quantifies the cumulative reciprocating plastic deformation energy generated by each thermal cycle on the metal lattice. The magnitude of the alternating thermal stress amplitude determines whether the system needs to issue vibration suppression commands and numerically guides the control valve to generate a corresponding proportion of interference energy.

[0047] The phase lead compensation is a time deviation control command generated by the vibration suppression control module to prematurely trigger the regulating valve's action, aiming to compensate for the phase lag caused by the transmission of thermoelastic waves in the pipeline. The phase lead compensation comprehensively considers the phase shift effects caused by the fluctuation frequency, physical pipeline length, and fluid flow velocity. The phase lead compensation ensures that the pulsating interference wave generated upstream arrives at the target pipe section with exactly the opposite phase.

[0048] The feedforward pulse signal is a control level sequence generated by the command issuing module, which directly drives the actuator of the regulating valve to produce high-frequency micro-amplitude switching action. The feedforward pulse signal includes timing information determined by the phase lead compensation amount and energy amplitude information determined by the opening adjustment amount. By exciting a dynamic thermal interference wave of a specific frequency in the fluid, the feedforward pulse signal suppresses the oscillations excited by internal leakage at the physical source.

[0049] Residual surface thermal fluctuation data refers to the residual temperature fluctuation signal collected by the feedback receiving module after the system performs an interference action. It represents the residual deviation after the main interference wave and the original thermal shock are superimposed. The residual surface thermal fluctuation data reflects the adaptability of the current control logic to the physical environment. The residual surface thermal fluctuation data is set as the objective function input of the extreme value optimization model for real-time closed-loop correction of phase offset.

[0050] The phase correction is a fine-tuning parameter calculated by the extreme value optimization model based on the changing trend of residual surface thermal fluctuation data. It is used to dynamically correct the calculation error of the estimated phase lead compensation. The phase correction can compensate for the decrease in control accuracy caused by pipeline resistance, flow velocity fluctuations, and nonlinear disturbances. By superimposing the phase correction with the initial phase of the feedforward pulse signal, the optimal interference reduction state is dynamically locked.

[0051] The internal leakage source location information is the specific valve node number and spatial coordinates that induce the risk, determined by the tracing and comparison unit through feature matching. This information concretizes the abstract oscillation alarm into a maintainable physical point. Combined with the location rendering of a 3D thermal stress cloud map, this information guides operators to accurately identify drain valves or electric gates with internal leakage defects without requiring a complete shutdown.

[0052] The data mapping matrix is ​​a transformation operator generated by the coordinate system transformation unit for converting between the pixel coordinate system and the physical 3D coordinate system. The data mapping matrix includes a rotation matrix and a translation vector, which can eliminate lens distortion and installation angle deviations during infrared imaging. The data mapping matrix can automatically compensate for large thermal displacements caused by pipeline heating, ensuring that positioning accuracy does not fail due to pipeline expansion.

[0053] The valve opening adjustment amount is the absolute stroke of the regulating valve stem, calculated by the vibration damping control module based on the alternating thermal stress amplitude. The opening adjustment amount determines the intensity level of the excited interference wave. The opening adjustment amount is linearly converted using a preset proportional coefficient, ensuring that the generated destructive interference energy reaches a dynamic balance with the internal thermal leakage impact energy.

[0054] The model constructed in this invention is based on the physical structure and thermodynamic characteristics of high-pressure steam pipelines in gas turbine power plants. The thermoelastic-fluid-structure interaction model establishes a mechanical mapping relationship between the temperature field and the stress field by integrating basic material properties such as the elastic modulus, Poisson's ratio, and coefficient of thermal expansion of the target pipe section's metallic materials, combined with spatial temperature gradient data provided by a three-dimensional temperature field distribution matrix. The data processing path begins with the infrared thermal imaging component capturing a sequence of infrared radiation images. The parameter extraction module converts the image data into a frequency domain signal using a fast Fourier transform, extracting the fluctuation frequency, initial phase, and fluctuation amplitude characterizing internal leakage. Subsequently, the inverse mapping unit uses a heat conduction inverse problem solving algorithm to inversely extrapolate the surface temperature data into a three-dimensional temperature field distribution matrix along the entire thickness of the pipe wall. The thermoelastic-fluid-structure interaction model receives the three-dimensional temperature field distribution matrix and performs finite element nodal stress calculations to obtain the thermal stress tensor characterizing the internal stress state, which is then quantified into alternating thermal stress amplitude. Finally, the vibration suppression control module derives the phase advance compensation amount based on the comparison between the alternating thermal stress amplitude and the preset thermal stress amplitude threshold, combined with the fluid operating state parameters, generates and sends a feedforward pulse signal to the regulating valve, and achieves closed-loop suppression of mechanical oscillation by exciting anti-phase interference waves.

[0055] The core logic of the registration unit's spatial registration calculation relies on the spatial mapping equation constructed by the affine transformation matrix. The registration unit extracts the coordinates of feature pixels from two adjacent frames in the infrared radiation image sequence, establishing a two-dimensional coordinate mapping model that includes translation and rotation feature parameters. For any pixel coordinate in the previous frame, the registration unit performs matrix multiplication using the affine transformation matrix to calculate and output the aligned target pixel coordinates in the current frame. The specific affine transformation spatial registration formula is expressed as follows: In the affine transformation space registration formula Represents the coordinate vector of the target pixel after alignment of the current frame image; M represents the coordinate vector of the feature pixels extracted from the previous frame; M represents the two-dimensional affine rotation matrix including rotation and scaling variables; T represents the translation vector representing the horizontal and vertical displacements. The data processing path of the frequency domain transformation unit performing a Fast Fourier Transform on the time-series temperature data matrix is ​​to convert the discrete time-series temperature signal into a frequency distribution map. The frequency domain transformation unit extracts the set of discrete temperature values ​​within a preset time window along the time axis and substitutes them into the standard equation of the Discrete Fourier Transform for series expansion. The mathematical expression of the Discrete Fourier Transform is designed as follows: In the mathematical expression of the discrete Fourier transform The first element in the frequency domain signal obtained after the transformation is represented by the first element. The complex amplitude of each frequency component; This represents an accumulation operator that performs a summation operation on discrete-time index numbers from 0 to the total number of samples minus 1; This represents the total number of discrete temperature values ​​sampled within a preset time window. Represents the discrete-time index number in the time series; Representing the The specific temperature value corresponding to each discrete-time index number; The base constant representing the natural logarithm; It represents the imaginary unit in complex number operations; The index number representing the discrete frequency component; It represents the constant value of pi.

[0056] The inverse mapping unit derives the three-dimensional temperature field distribution matrix by inversely solving the partial differential equation of the unsteady-state heat conduction inverse problem. The inverse mapping unit uses the surface temperature distribution analyzed from the infrared radiation image sequence as the convective heat transfer boundary input term, and combines it with the thermal conductivity and specific heat capacity parameters of the target pipe section material, performing inward differential iterative calculations along the pipe wall thickness direction. The specific expression of the partial differential equation of the unsteady-state heat conduction inverse problem is as follows: In the partial differential equation of the inverse problem of unsteady heat conduction Density parameter representing the metallic material of the target pipe section; The specific heat capacity constant of the target pipe section's metallic material; The term represents the partial derivative of the three-dimensional temperature field distribution matrix variables with respect to the time variable; The three-dimensional temperature field distribution matrix variable represents each spatial node s inside the target pipe segment; The time variable representing the occurrence of the heat transfer process; Thermal conductivity parameter representing the target pipe section's metallic material; Represents the Laplacian operator, used to measure the divergence distribution characteristics of the temperature gradient in three-dimensional space; The volumetric heat source generation rate variable represents the internal volume of the target pipe section. The data flow for calculating the thermal stress tensor in the tensor analysis unit is based on the generalized Hooke's law combined with the theory of thermal expansion and strain.

[0057] The tensor analysis unit receives local temperature difference data generated from the calculation of the three-dimensional temperature field distribution matrix, and multiplies the local temperature difference data by the coefficient of thermal expansion to obtain the thermal strain matrix. The tensor analysis unit integrates the elastic modulus and Poisson's ratio parameters of the pipe material to construct an elastic stiffness matrix, and multiplies the elastic stiffness matrix by the thermal strain matrix to obtain the thermal stress tensor induced by the temperature gradient. The formula for calculating the thermal stress tensor is as follows: In the formula for calculating the thermal stress tensor, This represents the thermal stress tensor components calculated on the nodes of the three-dimensional spatial mesh. The elastic modulus parameter representing the material of the target pipe section; Poisson's ratio parameter representing the material of the target pipe section; Represents the total strain tensor components at the mesh nodes; The volumetric strain trace number summation term representing the total strain tensor; Represents the matrix variables of the Kronecker function; The linear thermal expansion coefficient parameter representing the material of the target pipe section; This represents the local temperature difference variable of a grid node in a three-dimensional temperature field distribution matrix relative to a reference state.

[0058] The data stream used by the delay calculation unit to deduce the propagation delay time of the thermoelastic wave relies on the continuous equation of state and the fundamental laws of kinematics. The delay calculation unit divides the acquired transient mass flow rate parameters by the flow cross-sectional area and fluid density inside the pipe to obtain the real-time fluid velocity. The time estimation unit then uses the physical pipe length as the numerator and the real-time fluid velocity as the denominator, performing a division operation to obtain the propagation time of the thermodynamic wave. The propagation delay time estimation formula is expressed as: In the formula for calculating transmission delay time This represents the propagation delay time of thermoelastic waves inside the pipe. This parameter represents the physical pipe length between the control valve and the target pipe section. This represents the fluid density parameter calculated by interpolation from tables based on pressure and temperature parameters; The cross-sectional area parameter representing the actual flow within the target pipe section; This represents the transient mass flow rate parameter read from the power plant's distributed control system.

[0059] The lead calculation unit and the prediction compensation unit jointly and serially execute the task of calculating the phase lead compensation. The lead calculation unit multiplies the fluctuation frequency extracted from the front end with the transmission delay time to obtain the natural phase lag angle. To achieve the reverse cancellation effect of peaks and troughs, the prediction compensation unit superimposes an absolutely out-of-phase radian value onto the natural phase lag angle, synthesizing the final timing command deviation sent to the control valve. The mathematical formula for calculating the phase lead compensation is expressed as follows: In the mathematical formula for calculating phase lead compensation... This represents the phase advance compensation value that needs to be triggered in advance to activate the regulating valve and perform the pulsating action. Represents the constant value of pi; This represents the fluctuation frequency parameter locked in the output of the parameter extraction module; This represents the propagation delay time of the thermoelastic wave output by the time estimation unit. The extreme value optimization model internally encapsulates an optimization iterative equation based on gradient descent theory to correct execution errors. The feedback receiving module continuously collects residual surface thermal fluctuation data and constructs the target cost function matrix. The extreme value optimization model performs differentiation on the target cost function matrix within continuous time steps to obtain the partial derivatives of the mechanical oscillation change. The extreme value optimization model updates the phase attribute of the next feedforward pulse signal along the negative gradient direction of the partial derivatives of the mechanical oscillation change to approximate the minimum amplitude state.

[0060] The phase update formula for the extreme value optimization model is expressed as: In the phase update formula of the extreme value optimization model This represents the phase correction amount for the next control cycle, which is prepared to be sent to the instruction issuing module after the iterative update. This represents the actual phase variable being executed in the current control cycle; The iteration step size ratio coefficient represents the convergence speed of the extreme value search algorithm; The cost function, which includes residual surface thermal fluctuation data, is obtained from feedback under the current actual phase variable driving the action. The partial derivative of the mechanical oscillation change is obtained by taking the derivative of the cost function with respect to the actual phase variable.

[0061] The specific implementation plan addresses the complete data flow numerical scenario of the aforementioned calculation logic. The infrared thermal imaging component scans the pipe section behind the electric valve before the high-pressure desuperheater, and the parameter stripping unit locks the fluctuation frequency parameter induced by internal leakage, setting it to 5 Hz. The time estimation unit reads the physical pipe length parameter from the regulating valve to the target pipe section, setting it to 15 meters. The delay solution unit reads the transient mass flow rate parameter, setting it to 30 kg / s, the internal flow cross-sectional area parameter of the target pipe section, setting it to 0.05 square meters, and the fluid density parameter obtained from the fluid dynamics equations, setting it to 20 kg / m³. The delay solution unit multiplies the physical pipe length parameter of 15 meters by the fluid density parameter of 20 kg / m³ and the flow cross-sectional area parameter of 0.05 square meters to calculate 15 kg. Dividing 15 kg by the transient mass flow rate parameter of 30 kg / s yields a transmission delay time of 0.5 seconds for the thermoelastic wave. The advance calculation unit retrieves the fluctuation frequency parameter of 5 Hz and multiplies it by the transmission delay time of 0.5 seconds to obtain 2.5. Multiplying 2.5 by twice the constant of pi yields five times the constant of pi as the natural phase lag. The prediction compensation unit adds five times the constant of pi to the absolute antiphase constant of 1 times the constant of pi, resulting in a phase advance compensation of six times the constant of pi. The vibration suppression control module extracts six times the constant of pi to adjust the waveform properties of the feedforward pulse signal, driving the regulating valve to perform the opening change action three full cycles in advance. The tensor analysis unit integrates the elastic modulus of carbon steel material set to 200 gigapascals and the linear thermal expansion coefficient set to 1.2 x 10⁻⁵ degrees Celsius. For a grid node with a temperature difference of 50 degrees Celsius in the three-dimensional temperature field distribution matrix, it calculates that the alternating thermal stress amplitude caused by internal leakage reaches 120 megapascals.

[0062] To verify the actual vibration suppression and monitoring effect of the pipeline thermal stress online monitoring device based on thermal imaging described in this invention, the project team conducted a comparative test on a high-pressure steam heating system (rated pressure 5.6MPa, temperature 450℃) in a gas turbine power plant.

[0063] The experiment selected the outlet pipe section of the desuperheater and pressure reducer, which had a slight risk of internal leakage, as the monitoring target. The experimental group activated the closed-loop vibration suppression control system of this invention, while the control group only performed conventional infrared monitoring without intervening in the pulsation of the regulating valve. During the experiment, the data acquisition module continuously acquired the main steam pressure, temperature, and mass flow rate parameters, and the dynamic range of the fluid velocity calculated in real time was between 25 m / s and 42 m / s.

[0064] Accuracy Verification: The accuracy of the "three-dimensional temperature field distribution matrix" generated by the stress extrapolation module was verified by pre-embedding a high-precision thermocouple sensor inside the pipe wall as a true reference. Experimental results show that the spatial root mean square error of the inner wall temperature distribution extrapolated by the inverse mapping unit using the IHCP algorithm remains within a very small range. This confirms the high physical reliability of the technique for inverting the internal thermal state of the pipe wall based on the sequence of infrared radiation images from the outer surface.

[0065] Vibration suppression control effect test: When the initial value of the alternating thermal stress amplitude caused by internal leakage scouring is detected to be 135MPa (which has exceeded the preset thermal stress amplitude threshold), the vibration suppression control module drives the upstream regulating valve to perform 5Hz micro-pulsation interference according to the phase advance compensation amount.

[0066] Initial interference period: After the feedforward pulse signal is sent, the residual surface thermal fluctuation data collected by the feedback receiving module drops to 82MPa, and the initial suppression effect is evident.

[0067] Closed-loop optimization period: The extreme value optimization model outputs phase correction in real time based on the partial derivative of mechanical oscillation changes, and fine-tunes the feedforward control phase. After 3 iteration cycles, the alternating thermal stress amplitude of the target pipe section finally stabilizes below 28 MPa, and the stress alternating amplitude reduction rate is extremely significant.

[0068] Experimental Data Flow Analysis: During the experiment, the parameter stripping unit successfully extracted feature data with a fluctuation frequency of 5.2Hz and an initial phase of 0.85rad. The propagation delay time calculated by the delay solving unit was dynamically adjusted from 0.48s to 0.52s. The phase lead compensation was dynamically corrected accordingly, and the waveform properties of the feedforward pulse signal and the internal leakage excitation characteristics remained in a phase-locked state. Experimental data demonstrate that this system, through the thermoelastic wave phase interference principle, can effectively eliminate mechanical oscillations caused by non-uniform thermal stress and avoid the continuous deterioration of local thermal fatigue damage caused by internal leakage.

[0069] Embodiment 1 of this invention focuses on real-time monitoring of thermal stress in the thinned area of ​​the outlet pipe wall of a high-pressure bypass valve in a gas turbine power plant, analyzing the localized low-temperature chilling effect formed when high-pressure steam leaks into the outlet pipe section through the valve. During the operation of a certain type of gas turbine unit, an infrared thermal imaging component is deployed in a ring array near the critical weld seam at the outlet of the high-pressure bypass valve, continuously capturing infrared radiation image sequences including spatial pixel coordinates. The parameter extraction module performs spatial registration calculations on the acquired continuous images through the registration unit, outputting a time-series temperature data matrix, effectively filtering out false signals generated by pipeline thermal displacement during unit operation. The frequency domain conversion unit maps the extracted time-domain fluctuation data to the frequency domain, and the parameter stripping unit uses a preset low-pass cutoff frequency to remove baseline interference from the power plant's ambient temperature rise, locking the fluctuation frequency, fluctuation amplitude, and initial phase generated by the leaking jet. The stress derivation module uses a reverse mapping unit combined with a heat conduction inverse problem solving algorithm to derive the three-dimensional temperature field distribution matrix along the entire thickness of the pipe wall, and calculates the alternating thermal stress amplitude inside the metal using a thermoelastic fluid-structure interaction model. When the monitored alternating thermal stress amplitude exceeds the preset thermal stress amplitude threshold, the vibration suppression control module immediately extracts the fluid operating state parameters from the DCS, calculates the phase advance compensation amount, and sends a feedforward pulse signal to the regulating valve. The regulating valve generates a dynamic thermal interference wave that is inversely phase to the thermal shock, thereby canceling the high-frequency mechanical oscillations in the outlet pipe section and mitigating the risk of fatigue cracks initiating at the weld due to internal leakage erosion.

[0070] Embodiment 2 of this invention focuses on the system's dynamic compensation capability for the transmission delay of thermoelastic waves under complex operating conditions, specifically addressing the non-uniform temperature distribution caused by internal leakage of the electric valve before the high-pressure heating desuperheater. In this application scenario, a leaking valve before the desuperheater leads to continuous leakage of high-temperature steam. The infrared thermal imaging component uses a coordinate system transformation unit to map the captured infrared radiation image sequence into the physical three-dimensional coordinate system of the pipeline in real time, accurately locating the scoured area. The tensor analysis unit of the stress extrapolation module integrates the elastic modulus and Poisson's ratio of the pipeline material, transforming the three-dimensional temperature field distribution matrix into a thermal stress tensor. The amplitude extraction unit calculates the current alternating thermal stress amplitude. The delay solving unit in the vibration suppression control module extracts pressure parameters, temperature parameters, and transient mass flow rate parameters in real time, dynamically calculating the fluid velocity under the current operating conditions using fluid dynamics equations. The time extrapolation unit calculates the transmission delay time based on the physical pipeline length from the regulating valve to the internal leakage area. The advance calculation unit then derives the phase advance compensation amount, which, in conjunction with the opening adjustment amount, controls the regulating valve to generate precise energy compensation. The adaptive fine-tuning module continuously receives residual surface thermal fluctuation data collected by the feedback receiving module through an extreme value optimization model, and iteratively optimizes the phase correction amount, thus solving the problem of thermal wave distortion caused by nonlinear fluid friction during remote control.

[0071] Embodiment 3 of this invention: Applied to group monitoring and fault tracing scenarios of power plant condensate drain valve groups, achieving closed-loop management of potential hazards through visualized early warning and multi-dimensional positioning. The data acquisition module of the control device simultaneously monitors target pipe sections downstream of multiple condensate drain valves and converts infrared radiation image sequences into dynamic stress field views. When the stress estimation module determines that the alternating thermal stress amplitude of a certain condensate drain pipeline reaches the warning level, the tracing comparison unit inputs the extracted fluctuation frequency into the internal leakage excitation feature library for fingerprint matching, determining the specific valve number that caused the fault as the internal leakage source location information. After receiving the internal leakage source location information, the positioning rendering unit marks the abnormal stress distribution in red highlight form on the three-dimensional thermal stress cloud map and simultaneously pushes it to the DCS interface in the main control room for display. The vibration suppression control module assesses the current thermal damage accumulation state based on the fatigue life SN curve, adjusts the waveform attributes of the feedforward pulse signal to suppress local oscillations, and generates an "Online Vibration Suppression Intervention Report" for archiving. Based on the precise positioning results displayed on the visualization platform, operators can arrange planned troubleshooting without shutting down the system, avoiding sudden pipe burst accidents caused by the inability to detect internal leaks in time.

Claims

1. A pipeline thermal stress online monitoring device based on thermal imaging, characterized in that, It includes a device and a control device communicatively connected to the device; the device includes an infrared thermal imaging assembly and a regulating valve. The control device includes: The data acquisition module is used to acquire an infrared radiation image sequence of the outer surface of the target pipe section through the infrared thermal imaging component, and to acquire fluid operation status parameters of the fluid inside the target pipe section. The parameter extraction module is used to receive the infrared radiation image sequence and perform frequency domain conversion to obtain a frequency domain signal, and extract the fluctuation frequency and initial phase from the frequency domain signal; The stress estimation module is used to receive the infrared radiation image sequence and the fluctuation frequency, map the infrared radiation image sequence into a three-dimensional temperature field distribution matrix inside the target pipe section, and calculate the alternating thermal stress amplitude corresponding to the three-dimensional temperature field distribution matrix. The vibration suppression control module is used to read the fluid operating state parameters and the initial phase when the alternating thermal stress amplitude reaches the preset thermal stress amplitude threshold, calculate the phase advance compensation amount, generate a feedforward pulse signal corresponding to the phase advance compensation amount, send the feedforward pulse signal to the regulating valve, and output a warning message.

2. The online monitoring system for thermal stress in pipelines using thermal imaging according to claim 1, characterized in that, The parameter extraction module includes: The registration unit is used to extract the pixel coordinates in the infrared radiation image sequence and perform spatial registration calculations, and output a time-series temperature data matrix. The frequency domain conversion unit is used to receive the time-series temperature data matrix, perform a fast Fourier transform on the pixel points corresponding to the pixel point coordinates in the time-series temperature data matrix along the time axis, and convert them into the frequency domain signal. The parametric stripping unit is used to receive the frequency domain signal, filter out the low-frequency background component in the frequency domain signal using a preset low-pass cutoff frequency to obtain a retained frequency band signal, extract the dominant excitation frequency as the fluctuation frequency from the retained frequency band signal, and extract the fluctuation amplitude and the initial phase corresponding to the fluctuation frequency.

3. The online monitoring system for thermal stress in pipelines using thermal imaging according to claim 2, characterized in that, The stress estimation module includes: The reverse mapping unit is used to input the infrared radiation image sequence into the heat conduction inverse problem solving algorithm model to derive the three-dimensional temperature field distribution matrix inside the target pipe segment; Tensor analysis unit is used to receive the three-dimensional temperature field distribution matrix and integrate the input material properties to establish a thermoelastic fluid-structure interaction model, and calculate the thermal stress tensor inside the target pipe section. An amplitude extraction unit is used to receive the thermal stress tensor, extract the peak-to-valley difference in the thermal stress tensor as the alternating thermal stress amplitude, and send the alternating thermal stress amplitude to the vibration suppression control module.

4. The online monitoring system for thermal stress in pipelines using thermal imaging according to claim 3, characterized in that, The vibration suppression control module includes: The delayed solution unit is used to extract the pressure parameters, temperature parameters, and transient mass flow rate parameters from the fluid operating state parameters, and calculate the fluid velocity through fluid dynamics equations. The time calculation unit is used to receive the fluid flow rate, extract the physical pipe length between the regulating valve and the target pipe section, and divide the physical pipe length by the fluid flow rate to obtain the transmission delay time of the thermoelastic wave.

5. The online monitoring system for thermal stress in pipelines using thermal imaging according to claim 4, characterized in that, The vibration suppression control module also includes: A forward calculation unit is used to receive the transmission delay time and the fluctuation frequency, and multiply the fluctuation frequency by the transmission delay time to output the natural phase lag. The estimation compensation unit is used to receive the natural phase lag, add the natural phase lag to a preset anti-phase constant, and output the phase advance compensation amount corresponding to the regulating valve.

6. The online monitoring system for thermal stress in pipelines using thermal imaging according to claim 5, characterized in that, The control device further includes: The feedback receiving module is used to collect residual surface thermal fluctuation data of the outer surface of the target pipe section through the infrared thermal imaging component after the regulating valve receives the feedforward pulse signal. An adaptive fine-tuning module is used to receive the residual surface thermal fluctuation data and input it into the extreme value optimization model, and output the phase correction amount; The instruction issuing module is used to receive the phase correction amount and adjust the waveform properties of the feedforward pulse signal using the phase correction amount.

7. The online monitoring system for thermal stress in pipelines using thermal imaging according to claim 6, characterized in that, The control device further includes: The source comparison unit is used to input the fluctuation frequency into the internal leakage excitation feature library for feature value comparison and output the internal leakage source location information. The positioning and rendering unit is used to receive the internal leakage source location information, calibrate the coordinates corresponding to the internal leakage source location information on a preset three-dimensional thermal stress cloud map, and output the warning information including the internal leakage source location information.

8. The online monitoring system for thermal stress in pipelines using thermal imaging according to claim 7, characterized in that, The control device includes a coordinate system transformation unit, used to extract the pixel coordinate system parameters of the infrared thermal imaging component and the physical three-dimensional coordinate system parameters of the pipeline, establish a data mapping matrix that maps the pixel coordinate system parameters of the infrared thermal imaging component to the physical three-dimensional coordinate system parameters of the pipeline, and use the data mapping matrix to map the infrared radiation image sequence to the physical three-dimensional coordinate system of the pipeline.

9. The online monitoring system for thermal stress in pipelines using thermal imaging according to claim 8, characterized in that, The internal connection relationships of the adaptive fine-tuning module include: The extreme value optimization model receives the residual surface thermal fluctuation data collected by the feedback receiving module; The extreme value optimization model performs derivative calculations on the residual surface thermal fluctuation data of the continuous acquisition period to obtain the partial derivative of mechanical oscillation change, and calculates and outputs the phase correction amount based on the partial derivative of mechanical oscillation change. The instruction issuing module receives the phase correction amount output by the extreme value optimization model and sends the feedforward pulse signal superimposed with the phase correction amount to the regulating valve.

10. The online monitoring system for thermal stress in pipelines using thermal imaging according to claim 9, characterized in that, The vibration suppression control module extracts the alternating thermal stress amplitude, multiplies the alternating thermal stress amplitude using a preset proportional coefficient to obtain the opening adjustment amount, and sends the feedforward pulse signal including the opening adjustment amount information to the regulating valve.

Images