Heavy-duty gas turbine anti-surge valve flexibility test method, device, equipment and storage medium

Abstract

The present application relates to the technical field of gas turbine testing, in particular to a heavy-duty gas turbine anti-surge bleed valve flexibility testing method, device, equipment and storage medium. The present application synchronously collects inlet and outlet pressure data, valve core angular displacement data and valve body vibration data during the action process of the anti-surge bleed valve under different operating conditions; time reference alignment and filtering and noise reduction are performed on the multi-source data to obtain standardized multi-physical field coupling signals; a valve digital twin model is constructed, the following error of the measured and ideal motion trajectories is calculated, and abnormal energy features in the vibration signals are extracted; the flexibility score is calculated and the risk trend of sticking is determined by comprehensively considering the following error, abnormal energy features and action response time, and the evaluation results containing fault root cause positioning are generated. The present application realizes synchronous acquisition and coupling analysis of multi-dimensional signals under high temperature and high pressure conditions, and significantly improves the flexibility evaluation accuracy and early sticking early warning capability of the valve.

Description

Technical Field

[0001] This invention relates to the field of gas turbine testing technology, and in particular to a method, apparatus, equipment, and computer storage medium for testing the flexibility of anti-surge venting valves in heavy-duty gas turbines. Background Technology

[0002] The anti-surge venting valve of a heavy-duty gas turbine is a core safety component of the compressor system, widely used in turbine startup and variable operating condition scenarios, directly determining the unit's anti-surge capability and operational reliability. As gas turbines develop towards higher parameters and higher power, related technologies have constructed a valve action characteristic testing system through the collaborative operation of pressure monitoring, displacement sensing, and vibration analysis. This system covers the entire process from sensor deployment to data acquisition and analysis, including key aspects such as inlet and outlet pressure difference capture, valve core stroke recording, and abnormal vibration identification, aiming to evaluate the valve's dynamic response performance under high temperature and high pressure environments.

[0003] However, existing testing methods mostly employ a single-parameter independent monitoring mode, failing to achieve spatiotemporal synchronization and deep coupling analysis of multi-physics field signals. This may result in the inability to accurately capture minute defects such as microsecond-level valve core start-up delays or mid-process jamming, or significant deviations between test data and actual operating conditions due to the lack of extreme condition simulation. Furthermore, the lack of correlation modeling between vibration spectrum and motion trajectory makes it difficult to quantify the causal relationship between mechanical wear and flexibility degradation, thus affecting the early warning capability for potential valve failures and the accuracy of overall system test evaluation. Summary of the Invention

[0004] Therefore, the technical problem to be solved by the present invention is to overcome the problems in the prior art that, due to the use of a single parameter for independent monitoring and the lack of spatiotemporal synchronization and coupling analysis of multi-physics field signals, it is impossible to accurately capture the subtle jamming defects of the valve core, difficult to quantify the causal relationship between mechanical wear and flexibility degradation, and affect the early warning capability.

[0005] To solve the above-mentioned technical problems, the present invention provides a method for testing the flexibility of a surge relief valve in a heavy-duty gas turbine, comprising:

[0006] Under different operating conditions on the gas turbine test bench, the inlet and outlet pressure data, valve core angular displacement data, and valve body vibration data were collected simultaneously during the operation of the anti-surge venting valve. The acquired multi-source data is time-base aligned and filtered to eliminate environmental interference, resulting in a standardized multi-physics coupled signal. A digital twin model of the valve is constructed based on standardized multi-physics field coupled signals. The following error between the measured motion trajectory and the ideal motion trajectory is calculated, and the abnormal energy characteristics in the vibration signal are extracted. Based on the following error, abnormal energy characteristics, and action response time, the valve flexibility score is calculated comprehensively, the jamming risk trend is determined, and an evaluation result including fault root cause location is generated.

[0007] Preferably, the synchronous acquisition of inlet and outlet pressure data, valve core angular displacement data, and valve body vibration data during the operation of the anti-surge and deflation valve includes: By integrating a cooling structure and a high-temperature resistant encapsulation, the pressure sensor can collect real-time data on transient pressure changes at the valve inlet and outlet, under the condition that a cooling medium is introduced to form a protective gas film to isolate high temperatures. By using a non-contact laser angle sensor with adaptive wavelength and power adjustment mechanism, the laser parameters are adjusted according to the reflectivity of the valve body surface, and the full stroke angular displacement data of the valve core from start to stop are continuously recorded. The original signal of valve body vibration acceleration is captured by a piezoelectric accelerometer rigidly fixed to a key position on the valve body.

[0008] Preferably, the real-time acquisition of transient pressure change data at the valve inlet and outlet includes: Using a temperature drift dynamic compensation algorithm, the pressure reading is corrected online based on the real-time monitored sensor body temperature to obtain the corrected pressure data. At the same time, two sets of redundant pressure sensors are used to monitor the same measuring point. By calculating the difference between the readings of the two sets of sensors and comparing it with a preset threshold, a single-point fault alarm is triggered when the difference exceeds the threshold.

[0009] Preferably, the continuous recording of the valve core's full-stroke angular displacement data from start to stop includes: When a change in the reflectivity of the valve body surface is detected, the laser power is automatically adjusted based on the reference power and the reference reflectivity. In an environment where the lens is integrated with a purge curtain to prevent dust adhesion, multi-dimensional angle data including angular displacement, angular velocity and angular acceleration are output.

[0010] Preferably, the step of performing time-base alignment processing on the collected multi-source data and performing filtering and noise reduction to eliminate environmental interference includes: By using the same hardware trigger pulse from the central clock to receive all sensors, the acquisition start time is ensured to be strictly consistent. Then, the cross-correlation analysis algorithm is used to correct the time delay of signals from different physical lines. The time delay corresponding to the maximum value of the cross-correlation function is found and phase compensation is performed to achieve strict alignment of multi-source data in the time domain. Next, an adaptive filtering algorithm is used to process the aligned original signal to eliminate environmental interference and electromagnetic noise, resulting in a standardized multi-physics coupled signal.

[0011] Preferably, the process of constructing a digital twin model of the valve based on standardized multi-physics coupled signals, calculating the following error between the measured motion trajectory and the ideal motion trajectory, and extracting abnormal energy features from the vibration signal includes: A digital twin simulation model of a valve based on state-space equations is established, defining a state vector containing angular displacement, angular velocity, and angular acceleration. The ideal motion trajectory is solved using the system matrix, input matrix, and output matrix. The motion following error is calculated by comparing the measured trajectory with the ideal trajectory, and the root mean square error is used as the comprehensive error index. Simultaneously, the theoretical excitation vibration energy is obtained from the measured angular acceleration, and compared with the measured vibration signal energy to calculate the abnormal vibration energy residual to characterize the abnormal energy features.

[0012] Preferably, the step of obtaining the theoretical excitation vibration energy from the measured angular acceleration, comparing it with the measured vibration signal energy, and calculating the abnormal vibration energy residual to characterize the abnormal energy features includes: The vibration energy is excited based on the angular acceleration calculation theory, and the measured vibration signal energy is calculated at the same time. Then, the abnormal vibration energy residual is obtained through integration. This residual value directly reflects the additional vibration energy caused by valve core jamming or component wear.

[0013] Preferably, the step of comprehensively calculating the valve flexibility score and determining the jamming risk trend based on the following error, abnormal energy characteristics, and action response time, and generating an evaluation result including fault root cause localization, includes: A comprehensive scoring model for valve flexibility, including a normalized evaluation function, is constructed, and a comprehensive score is calculated. The comprehensive score is obtained by multiplying the normalized evaluation function based on the root mean square of the following error, the residual of abnormal vibration energy, and the response time by the adaptive weights of the operating conditions and then summing them. Simultaneously, the spectral kurtosis algorithm is applied to calculate the spectral kurtosis value of the vibration signal to identify early frictional impact characteristics. When the spectral kurtosis value continues to rise, it is determined that the early jamming trend is strengthening.

[0014] The present invention also provides a testing device for the flexibility of a surge relief valve in a heavy-duty gas turbine, comprising: The multi-source synchronous acquisition module is used to synchronously acquire inlet and outlet pressure data, valve core angular displacement data, and valve body vibration data during the operation of the anti-surge venting valve under different operating conditions on the gas turbine test bench. The signal processing module is used to perform time base alignment processing on the acquired multi-source data and perform filtering and noise reduction to eliminate environmental interference, so as to obtain a standardized multi-physics field coupled signal. The twin analysis and feature extraction module is used to construct a digital twin model of the valve based on standardized multi-physics field coupled signals, calculate the following error between the measured motion trajectory and the ideal motion trajectory, and extract abnormal energy features from the vibration signal. The assessment and early warning module is used to comprehensively calculate the valve flexibility score and determine the jamming risk trend based on the following error, abnormal energy characteristics and action response time, and generate assessment results including fault root cause location.

[0015] This invention also provides a testing device for the flexibility of a heavy-duty gas turbine anti-surge bleed valve, comprising: Memory, used to store computer programs; A processor is used to execute the computer program to implement the steps of the above-described method for testing the flexibility of a heavy-duty gas turbine anti-surge venting valve.

[0016] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for testing the flexibility of a heavy-duty gas turbine anti-surge venting valve.

[0017] The technical solution of the present invention has the following advantages compared with the prior art: The method for testing the flexibility of the anti-surge venting valve in heavy-duty gas turbines described in this invention simultaneously collects multi-source data on pressure, angular displacement, and vibration under different operating conditions. It then sequentially performs time reference alignment and filtering / noise reduction, constructs a digital twin model to calculate following errors and extract abnormal energy characteristics, comprehensively evaluates the flexibility score, and determines the trend of jamming risk. This method enables multi-dimensional synchronous perception and coupled analysis of the anti-surge venting valve's operating characteristics under high-temperature and high-pressure conditions, significantly improving the accuracy of flexibility assessment. It effectively captures minute defects such as microsecond-level valve core start-up delays and mid-course jamming, while simultaneously achieving early intelligent warning and root cause location of jamming risks. This overcomes the problem of inaccurate assessments caused by single-parameter monitoring and data asynchrony in existing technologies. Attached Figure Description

[0018] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart illustrating the implementation of a method for testing the flexibility of a surge relief valve in a heavy-duty gas turbine, as provided by the present invention. Figure 2 This is a structural block diagram of a heavy-duty gas turbine anti-surge venting valve flexibility testing device provided in an embodiment of the present invention. Detailed Implementation

[0019] The core of this invention is to provide a method, device, equipment, and computer storage medium for testing the flexibility of a surge relief valve in a heavy-duty gas turbine. This method can effectively achieve synchronous acquisition and deep coupling analysis of multi-physics field signals, accurately assess valve flexibility, and provide early warning of jamming risks.

[0020] To enable those skilled in the art to better understand the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. 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.

[0021] Please refer to Figure 1. Figure 1 The flowchart illustrates the implementation of a method for testing the flexibility of a surge prevention valve in a heavy-duty gas turbine, as provided by this invention. The specific operation steps are as follows: S101: Under different operating conditions on the gas turbine test bench, synchronously collect inlet and outlet pressure data, valve core angular displacement data, and valve body vibration data during the operation of the anti-surge venting valve; S102: Perform time base alignment processing on the acquired multi-source data and perform filtering and noise reduction to eliminate environmental interference, and obtain a standardized multi-physics coupling signal. S103: Construct a digital twin model of the valve based on standardized multi-physics field coupled signals, calculate the following error between the measured motion trajectory and the ideal motion trajectory, and extract abnormal energy features from the vibration signal; S104: Based on the following error, abnormal energy characteristics, and action response time, calculate the valve flexibility score and determine the jamming risk trend, and generate an evaluation result that includes fault root cause location.

[0022] In some embodiments, this embodiment will describe in detail the complete implementation of the method for testing the flexibility of the anti-surge venting valve of a heavy-duty gas turbine. This method relies on the whole machine test bench environment and, through the collaborative work of a multi-dimensional perception module and an AI intelligent analysis system, achieves a precise quantitative assessment of the operational flexibility, response characteristics, and potential faults of the anti-surge venting valve under high temperature, high pressure, and variable operating conditions.

[0023] In this embodiment, the hardware architecture of the test system is first deployed and calibrated. The anti-surge valve, as the test object, is assembled according to the design drawings, actuator, and piping system. This ensures that the valve shaft coaxiality, connection rigidity, and actuator output force matching meet technical specifications. Flanges, joints, and hoses are tightened to eliminate assembly gaps. A multi-dimensional sensing module is integrated into the valve body and surrounding key locations, specifically including a pressure sensing submodule, an angle sensing submodule, and a vibration sensing submodule. The pressure sensing submodule consists of two sets of high-precision pressure sensors, installed at the valve inlet and outlet respectively, to capture transient changes in flow field pressure. To adapt to the high-temperature environment of 300-600℃ in heavy-duty gas turbine whole-machine testing, a microchannel air film cooling structure is integrated at the sensor front end, introducing approximately 50% low-temperature compressed air to form a stable protective air film in front of the sensing diaphragm, effectively isolating the mainstream high temperature. The sensor body is encapsulated in high-temperature resistant ceramic, capable of withstanding short-term temperatures up to 650℃. The angle sensing submodule employs a non-contact laser angle sensor, installed on the valve shaft extension or outside the valve body, avoiding mechanical interference from contact measurements on the valve core movement. Its lens integrates a purge air curtain, continuously spraying dry compressed air to prevent dust and oil mist adhesion. The vibration sensing submodule uses a piezoelectric accelerometer, rigidly bolted to key locations such as the valve cover or drive bracket, ensuring a clear vibration transmission path. Its frequency response range covers 0.5-10kHz, with a sensitivity of no less than 100mV / g. All sensors are connected to a data acquisition card and receive the same hardware trigger pulse from a central clock, with rise edge jitter controlled within 10μs to ensure strict consistency in the acquisition start-up timing.

[0024] After the test preparation phase is completed, the multi-condition signal acquisition process begins. The gas turbine test bench is subjected to acceleration and loading according to a typical start-up curve, covering different operating conditions such as ignition start-up, acceleration, grid connection and load increase, and normal shutdown. At each operating condition point, the anti-surge venting valve is triggered to open or close, and this process is repeated multiple times to ensure the validity of the data statistics.

[0025] In other embodiments, as an alternative implementation, the whole-machine test bench can also cover extreme transient conditions such as load shedding and rapid load changes to comprehensively evaluate the dynamic response characteristics of the anti-surge valve under extreme conditions. The number of operation cycles for each operating point can be set to 5 to 10 times according to statistical significance requirements to eliminate the influence of random errors on the test results.

[0026] In one specific embodiment, the gas turbine test bench accelerates and loads according to a typical start-up curve, specifically including: ignition start-up phase (0-1000 rpm), acceleration phase (1000-3000 rpm), grid connection and load increase phase (30% rated load, 60% rated load, 100% rated load), and normal shutdown phase. At each operating point, the anti-surge venting valve is triggered by the control system to perform three complete opening-closing cycles, and the data acquisition system continuously records the entire process at a sampling rate of 10 kHz.

[0027] Specifically, in this embodiment, the pressure sensor is a high-precision sensor with a range of 0-10 bar and an accuracy class of 0.1. Its microfluidic air film cooling structure introduces low-temperature compressed air at 20°C at a flow rate of 0.5 L / min. The laser angle sensor has a measurement range of 0-90° and a resolution of 0.01°. The piezoelectric accelerometer has a sensitivity of 100 mV / g and a frequency response range of 0.5-10 kHz.

[0028] It should be noted that in this embodiment, the "anti-surge release valve" refers to a safety valve in the compressor system of a heavy-duty gas turbine used to prevent surge; the "whole machine test bench" refers to a comprehensive test platform used to simulate the real operating environment of a gas turbine; and the "multi-dimensional sensing module" refers to a collaborative acquisition unit integrating multiple types of sensors. The beneficial effects of this embodiment are: by using multi-parameter synchronous sensing and high-precision time alignment technology, it overcomes the limitations of single-parameter monitoring and provides a comprehensive data foundation; and through special sensor protection design and compensation algorithms, it ensures measurement accuracy under extreme environments.

[0029] Based on the above embodiments, in some embodiments, the synchronous acquisition of inlet and outlet pressure data, valve core angular displacement data, and valve body vibration data during the operation of the anti-surge valve includes: By integrating a cooling structure and a high-temperature resistant encapsulation, the pressure sensor can collect real-time data on transient pressure changes at the valve inlet and outlet, under the condition that a cooling medium is introduced to form a protective gas film to isolate high temperatures. By using a non-contact laser angle sensor with adaptive wavelength and power adjustment mechanism, the laser parameters are adjusted according to the reflectivity of the valve body surface, and the full stroke angular displacement data of the valve core from start to stop are continuously recorded. The original signal of valve body vibration acceleration is captured by a piezoelectric accelerometer rigidly fixed to a key position on the valve body.

[0030] In some embodiments, the integrated microfluidic film cooling structure of the pressure sensor employs an annular distributed pore design to ensure uniform coverage of the cooling film on the sensing diaphragm surface; the high-temperature resistant ceramic encapsulation uses zirconia-based ceramic material with a thermal conductivity of less than 2 W / (m·K). The non-contact laser angle sensor operates at a wavelength of 650-850 nm and can automatically switch between visible and infrared bands based on the reflectivity of the valve body surface. The piezoelectric accelerometer is installed at various locations, including the upper surface of the valve cover, the side wall of the drive bracket, and the valve body flange connection, with at least one measuring point at each location.

[0031] In other embodiments, as an alternative implementation, the pressure sensor can also be a sapphire fiber optic grating sensor, which can achieve pressure measurement in high-temperature environments through wavelength demodulation without the need for additional cooling structures; the angle sensor can be a magnetoelectric non-contact encoder, which can obtain angular displacement signals by detecting changes in the magnetic field; and the vibration sensor can be a MEMS accelerometer to reduce cost and installation complexity.

[0032] In one specific embodiment, the pressure sensor integrates a microfluidic film cooling structure, forming a protective film approximately 0.1 mm thick in front of the sensing diaphragm, reducing the sensor body temperature from 600°C to below 120°C. When the laser angle sensor detects a decrease in the reflectivity of the valve body surface, it automatically increases the laser power from 5 mW to 20 mW and switches to the 850 nm infrared band. The piezoelectric accelerometer is rigidly fixed to the center of the valve cover with an M5 bolt and a torque of 10 N·m.

[0033] Specifically, in this embodiment, the temperature drift compensation of the pressure sensor is achieved by real-time monitoring of the body temperature using a built-in platinum resistance temperature sensor, resulting in a measurement error of less than ±0.1%FS after compensation. The purge air curtain of the laser angle sensor continuously sprays dry compressed air with a dew point temperature of -40°C at a flow rate of 10 L / min. The frequency response range of the piezoelectric accelerometer, after calibration, is confirmed to have a flatness better than ±5% within the range of 0.5-10 kHz.

[0034] It should be noted that "cooling structure" refers to a protective device that reduces the sensor's operating temperature by introducing a cooling medium; "protective gas film" refers to a cooling layer formed on the surface of the sensing element that isolates the high-temperature mainstream gas; and "adaptive wavelength and power adjustment mechanism" refers to the sensor's function of automatically optimizing laser parameters based on the optical characteristics of the measured surface. The beneficial effect of this embodiment is that, through the collaborative work of multiple types of sensors, simultaneous acquisition of three physical quantities—pressure, displacement, and vibration—is achieved, providing high-fidelity raw data for subsequent multiphysics coupling analysis.

[0035] Based on the above embodiments, in some embodiments, the real-time acquisition of transient pressure change data at the valve inlet and outlet includes: Using a temperature drift dynamic compensation algorithm, the pressure reading is corrected online based on the real-time monitored sensor body temperature to obtain the corrected pressure data. At the same time, two sets of redundant pressure sensors are used to monitor the same measuring point. By calculating the difference between the readings of the two sets of sensors and comparing it with a preset threshold, a single-point fault alarm is triggered when the difference exceeds the threshold.

[0036] In some embodiments, the temperature drift dynamic compensation algorithm constructs a compensation model based on the temperature-pressure response curves obtained by the sensor within the calibrated temperature range. The installation positions of the two sets of redundant pressure sensors completely overlap to ensure consistent measurement conditions. The preset threshold can be dynamically adjusted according to the sensor's accuracy level and operating requirements; a typical value is 1% of full scale.

[0037] In other embodiments, as an alternative implementation, temperature drift compensation can employ an adaptive compensation algorithm based on neural networks, which learns the nonlinear mapping relationship between temperature and pressure readings by training on historical data; the differential alarm of redundant sensors can also adopt a three-to-two voting mechanism to further improve reliability.

[0038]

[0039] in, For the corrected pressure data, Original pressure reading and For the pre-calibrated temperature compensation coefficient, For real-time monitoring of the sensor body temperature, The reference temperature is 20°C. This formula achieves non-linear compensation for temperature drift by multiplying the original reading by a second-order polynomial correction term related to the temperature deviation. The absolute value of the difference between the two sets of redundant sensors is denoted as... , The readings are from the first set of pressure sensors. This represents the readings from the second set of pressure sensors. t is a time variable, when... Exceeding the preset threshold If an abnormal data is detected, the system will immediately trigger a single-point fault alarm and remove the abnormal data to ensure the reliability of the test results.

[0040] Specifically, in this specific embodiment, the calibration coefficient Values Values reference temperature Set the temperature to 20℃ and the preset threshold. Take 1% of the full-scale 10 bar, which is 0.1 bar.

[0041] It should be noted that the "temperature drift dynamic compensation algorithm" refers to a calculation method that corrects the pressure measurement value based on the real-time temperature of the sensor, aiming to eliminate the influence of high-temperature environment on sensor accuracy; "redundant pressure sensor" refers to a design that arranges two independent sensors at the same measuring point to improve data reliability. The beneficial effects of this embodiment are: the temperature compensation algorithm ensures a measurement accuracy better than ±0.1%FS across the entire operating range, and the redundancy design enables real-time detection and isolation of sensor faults, significantly improving the reliability of pressure data.

[0042] Based on the above embodiments, in some embodiments, continuously recording the full stroke angular displacement data of the valve core from start to stop includes: When a change in the reflectivity of the valve body surface is detected, the laser power is automatically adjusted based on the reference power and the reference reflectivity. In an environment where the lens is integrated with a purge curtain to prevent dust adhesion, multi-dimensional angle data including angular displacement, angular velocity and angular acceleration are output.

[0043] In some embodiments, the laser power adjustment is continuous and bidirectional, i.e., the power is increased when reflectivity decreases and decreased when reflectivity increases, to maintain a constant intensity of the returned light signal. Angular velocity is obtained by performing a first-order differential operation on the angular displacement, and angular acceleration is obtained by performing a second-order differential operation on the angular velocity. The flow rate and pressure of the purge air curtain can be adaptively adjusted according to the ambient dust concentration.

[0044] In other embodiments, as an alternative implementation, angular displacement data can also be obtained by contact via a high-precision potentiometer or rotary transformer, but the potential impact of contact friction on the valve core movement must be considered; angular velocity and angular acceleration can also be directly measured by a laser Doppler vibrometer without the need for differential calculations.

[0045] In one specific embodiment, the laser power adjustment logic follows the following formula:

[0046] in, The adjusted laser power, The reference power (the power value that optimizes the signal-to-noise ratio of the returned light at the reference reflectivity). The reference reflectance is (usually taken as the reflectance of a standard polished stainless steel surface). This refers to the real-time reflectivity of the valve body surface. This formula ensures that power is proportionally increased as reflectivity decreases, maintaining a constant returned light signal intensity. The sensor directly outputs angular displacement. And angular velocity is generated in real time through differential operations. and angular acceleration This provides a multi-dimensional kinematic benchmark for subsequent analysis.

[0047] Specifically, in this embodiment, the reference power Take 5mW, reference reflectivity A value of 0.8 (corresponding to a polished stainless steel surface) is used. When the reflectivity of the valve body surface drops to 0.2 due to carbon buildup, the laser power is automatically increased to 20mW. The sampling frequency for angular displacement is 2kHz, and angular velocity and angular acceleration are calculated using a five-point cubic smoothing differential algorithm to reduce noise amplification effects.

[0048] It should be noted that the "adaptive wavelength and power adjustment mechanism" refers to the sensor's function of automatically optimizing laser parameters based on the reflectivity of the measured surface, aiming to overcome measurement difficulties under harsh optical conditions; "multi-dimensional angle data" refers to the set of three kinematic parameters: angular displacement, angular velocity, and angular acceleration. The beneficial effects of this embodiment are: by ensuring an angular resolution of up to 0.01° under harsh optical conditions through adaptive power adjustment, and by providing complete boundary conditions for digital twin modeling through multi-dimensional kinematic output.

[0049] Based on the above embodiments, in some embodiments, time-base alignment processing is performed on the collected multi-source data, and filtering and noise reduction are performed to eliminate environmental interference, including: By using the same hardware trigger pulse from the central clock to receive all sensors, the acquisition start time is ensured to be strictly consistent. Then, the cross-correlation analysis algorithm is used to correct the time delay of signals from different physical lines. The time delay corresponding to the maximum value of the cross-correlation function is found and phase compensation is performed to achieve strict alignment of multi-source data in the time domain. Next, an adaptive filtering algorithm is used to process the aligned original signal to eliminate environmental interference and electromagnetic noise, resulting in a standardized multi-physics coupled signal.

[0050] In some embodiments, the hardware trigger pulse is generated by a high-precision clock source, with its rising edge jitter controlled in the microsecond range. The computation window length of the cross-correlation analysis algorithm can be adaptively adjusted according to signal characteristics to ensure the statistical significance of the delay estimate. The adaptive filtering algorithm can employ recursive least squares (RLS) filtering or adaptive wavelet thresholding for noise reduction, with its filtering parameters dynamically updated according to the noise level.

[0051] In other embodiments, as an alternative implementation, time reference alignment can also be achieved via GPS synchronized clock or IEEE 1588 precise time protocol without the need for a separate hardware trigger line; filtering and noise reduction can also employ empirical mode decomposition (EMD) or variational mode decomposition (VMD) methods to separate the effective signal from the noise components.

[0052] In one specific embodiment, the cross-correlation analysis algorithm achieves time delay correction by calculating the following cross-correlation function:

[0053] in, Signals originating from two different physical lines, For time delay variables, This is the cross-correlation function. Its physical meaning is to describe the similarity between two signals at different time offsets, finding... The maximum value corresponding to As the actual time delay, phase compensation is applied to this delay to achieve strict time-domain alignment of multi-source data. Subsequently, the RLS filtering algorithm is used to align the original signal. Processed and filtered signal The iterative formula is:

[0054] in, For discrete-time indexing, This is the estimated value of the filtered signal at the current moment. This is the estimated value from the previous moment. It is the adaptive gain matrix (which determines the filter's tracking speed and steady-state error). This is the observation matrix (which describes the linear relationship between the measured values ​​and the state). This is the original measurement value at the current moment. The formula applies the gain matrix to the innovation... Weighted corrections are applied to achieve recursive estimation of the true signal.

[0055] Specifically, in this embodiment, the rising edge jitter of the hardware trigger pulse is controlled within 10μs. The calculation window length for cross-correlation analysis is 1024 sampling points, and the delay estimation accuracy reaches one sampling period (i.e., 1μs). The forgetting factor of the RLS filtering algorithm is set to 0.98, the initial value of the gain matrix $K(k)$ is set to the identity matrix, and the observation matrix... Pre-calibrated based on the sensor transfer function.

[0056] It should be noted that "time reference alignment" refers to the process of adjusting signals from different acquisition channels to strict synchronization on the time axis; "cross-correlation analysis algorithm" is a signal processing method that estimates time delay by calculating the correlation between signals; and "adaptive filtering algorithm" refers to a filtering method in which the filtering parameters can be automatically adjusted according to the statistical characteristics of the signal and noise. The beneficial effects of this embodiment are: by combining hardware triggering with cross-correlation correction, microsecond-level time synchronization accuracy is achieved; and adaptive filtering effectively suppresses environmental interference and electromagnetic noise, significantly improving the signal-to-noise ratio.

[0057] Based on the above embodiments, in some embodiments, a digital twin model of the valve is constructed based on standardized multi-physics coupling signals, the following errors between the measured motion trajectory and the ideal motion trajectory are calculated, and abnormal energy features in the vibration signal are extracted, including: A digital twin simulation model of a valve based on state-space equations is established, defining a state vector containing angular displacement, angular velocity, and angular acceleration. The ideal motion trajectory is solved using the system matrix, input matrix, and output matrix. The motion following error is calculated by comparing the measured trajectory with the ideal trajectory, and the root mean square error is used as the comprehensive error index. Simultaneously, the theoretical excitation vibration energy is obtained from the measured angular acceleration, and compared with the measured vibration signal energy to calculate the abnormal vibration energy residual to characterize the abnormal energy features.

[0058] In some embodiments, the digital twin model is constructed based on the physical mechanism of the valve, including the dynamic equations of the valve core-valve sleeve friction pair, the driving force equations of the actuator, and the coupling relationship between the flow field pressure and the valve core position. System matrix This reflects the valve's inertia, damping, and stiffness characteristics; input matrix Describes the effect of actuator thrust or pressure difference on the state; output matrix Map the state vector to observable output variables.

[0059] In other embodiments, as an alternative implementation, the digital twin model may also employ a data-driven black-box modeling method, such as neural networks or Gaussian process regression, to learn the input-output mapping relationship of the valve through a large amount of historical data; the calculation of the following error may also use alternative indicators such as maximum absolute error or integral absolute error.

[0060] In one specific embodiment, a state vector is defined. ,in These are angular displacement, angular velocity, and angular acceleration, respectively. (This is derived from the state-space equations.) and output equation Solving for the ideal motion trajectory ,in, For state vectors, For system matrix, For the input matrix, For the actuator thrust or differential pressure input, This is the output matrix. The measured trajectory... With ideal trajectory In comparison, define the motion following error. The root mean square error was used as the comprehensive error index. The calculation formula is as follows:

[0061] in The total duration of the action process. This represents the instantaneous following error. This indicator comprehensively evaluates the average deviation of the entire action process by integrating the square of the error over time and taking the square root, thus intuitively reflecting the accuracy with which the actual movement of the valve core follows the ideal command.

[0062] Specifically, in this specific embodiment, the system matrix in the state-space equation Input matrix and output matrix The input signal is obtained through a system identification method. This refers to the pneumatic thrust corresponding to the valve control command. Integral time. The time taken for the valve core to complete its stroke from start to stop is typically 0.5-2 seconds.

[0063] It should be noted that a "digital twin model" refers to a simulation model that reflects the dynamic characteristics of a valve, capable of calculating an ideal motion trajectory based on input conditions; "following error" refers to the deviation between the measured motion trajectory and the ideal motion trajectory; and "root mean square error" is a comprehensive error index that reflects the overall energy level of the error. The beneficial effect of this embodiment is that by constructing a digital twin model and calculating the following error, complex mechanical dynamic responses can be transformed into quantifiable accuracy indicators, effectively overcoming the limitations of traditional single-parameter monitoring in identifying subtle jamming and hysteresis defects.

[0064] Based on the above embodiments, in some embodiments, the theoretical excitation vibration energy is obtained from the measured angular acceleration, and compared with the measured vibration signal energy to calculate the abnormal vibration energy residual to characterize the abnormal energy features, including: The vibration energy is excited based on the angular acceleration calculation theory, and the measured vibration signal energy is calculated at the same time. Then, the abnormal vibration energy residual is obtained through integration. This residual value directly reflects the additional vibration energy caused by valve core jamming or component wear.

[0065] In some embodiments, the theoretical excitation vibration energy is physically defined as the vibration energy generated by the inertial force produced by angular acceleration under ideal motion conditions of the valve core. The measured vibration signal energy is the vibration energy directly measured by the accelerometer, including contributions from both ideal excitation and abnormal impact. The abnormal vibration energy residual is the difference between the measured energy and the theoretical energy.

[0066] In other embodiments, as an alternative implementation, abnormal energy characteristics can also be extracted through envelope spectrum analysis or wavelet packet energy spectrum decomposition of the vibration signal without comparison with theoretical excitation; the abnormal vibration energy residual can also be replaced by alternative indicators such as peak energy ratio or energy entropy.

[0067] In one specific embodiment, the excitation vibration energy is calculated based on the angular acceleration calculation theory. Its expression is:

[0068] in It is a proportionality coefficient (reflecting the gain characteristics of valve core mass, moment of inertia, and vibration transmission path). This represents the measured angular acceleration. Simultaneously, the energy of the measured vibration signal is calculated. ,in The raw signal is collected by a vibration acceleration sensor. The residual energy of the abnormal vibration is then obtained through integration. The calculation formula is as follows:

[0069] in For the duration of points, To measure the energy of the vibration signal, The theoretical excitation vibration energy is calculated using this formula. First, the square of the difference between the measured vibration energy and the theoretical excitation vibration energy is calculated. Then, the square root is taken after integrating over time to obtain the residual value that characterizes the abnormal energy. This residual value directly reflects the additional vibration energy caused by valve core jamming or component wear.

[0070] Specifically, in this specific embodiment, the scaling factor The typical value range was determined to be 0.01-0.1 through calibration tests. Vibration signal The sampling frequency is consistent with the angular acceleration, both being 2kHz. Integration time Take the total travel time of the valve core from start to stop.

[0071] It should be noted that "theoretical excitation vibration energy" refers to the vibration energy theoretically generated by the inertial force of the valve core motion, which is the benchmark for judging whether there is abnormal impact; "abnormal vibration energy residual" refers to the portion of the measured vibration energy that exceeds the theoretical excitation energy, which is a key indicator for quantifying the severity of mechanical abnormalities. The beneficial effect of this embodiment is that by comparing theoretical excitation and measured vibration, abnormal vibrations caused by valve core jamming or component wear can be quantified from the energy dimension, providing a quantitative basis for assessing the severity of faults.

[0072] Based on the above embodiments, in some embodiments, the valve flexibility score is comprehensively calculated and the jamming risk trend is determined based on the following error, abnormal energy characteristics, and action response time, generating an evaluation result including fault root cause localization, including: A comprehensive scoring model for valve flexibility, including a normalized evaluation function, is constructed, and a comprehensive score is calculated. The comprehensive score is obtained by multiplying the normalized evaluation function based on the root mean square of the following error, the residual of abnormal vibration energy, and the response time by the adaptive weights of the operating conditions and then summing them. Simultaneously, the spectral kurtosis algorithm is applied to calculate the spectral kurtosis value of the vibration signal to identify early frictional impact characteristics. When the spectral kurtosis value continues to rise, it is determined that the early jamming trend is strengthening.

[0073] In some embodiments, the normalized evaluation function maps indicators of different dimensions and magnitudes to a unified scoring range (e.g., 0-1), with a larger value indicating a better individual indicator. The operating condition adaptive weight can be dynamically adjusted according to the operating conditions of the gas turbine (e.g., startup, load increase, steady state, variable load) to reflect the differences in the importance of each evaluation indicator under different operating conditions.

[0074] In other embodiments, as another implementation method, the comprehensive score can also be constructed using fuzzy comprehensive evaluation or analytic hierarchy process (AHP), and the weights can be determined through expert experience or historical data optimization; the determination of stagnation trends can also be achieved using machine learning methods such as support vector machine (SVM) or long short-term memory network (LSTM).

[0075] In one specific embodiment, a valve flexibility comprehensive scoring model including a normalized evaluation function is constructed, and a comprehensive score is calculated. Its formula is:

[0076] in For working condition adaptive weights and satisfying To follow the root mean square error, For abnormal vibration energy residual, The action response time (the time interval from when the command is issued to when the valve core begins to move). Based on The normalized evaluation function can usually be selected as... Monotonically decreasing form, This is the normalization coefficient. (Score) A higher value indicates better valve flexibility. Simultaneously, a spectral kurtosis algorithm is applied to calculate the spectral kurtosis value of the vibration signal. The formula for identifying early frictional impact characteristics is as follows:

[0077] in, It is a vibration signal. For analysis duration. Spectral kurtosis. It is the ratio of the fourth moment to the square of the second moment of the vibration signal, characterizing the degree to which the signal deviates from a Gaussian distribution; for normal random vibration, Approaching 3; when impact or friction features are present, a sharp peak appears in the signal. Significantly greater than 3. When the spectral kurtosis value A sustained upward trend indicates an strengthening of the early stagnation trend.

[0078] Specifically, in this specific embodiment, the adaptive weight of the operating condition is taken under the starting operating condition. , Under steady-state conditions, take Normalization coefficient Based on the statistical distribution of historical data, typical values ​​were determined to be 0.5, 2.0, and 0.2. The analysis time for spectral kurtosis calculation was also considered. Set a time interval of 1 second, a sliding step size of 0.1 seconds, and continuously monitor. The trend of value changes.

[0079] It should be noted that the "normalized evaluation function" refers to a function that maps original indicators with different dimensions to a unified scoring interval; "spectral kurtosis" is a signal processing method based on higher-order statistics, which can effectively identify non-Gaussian impact components in vibration signals; and "condition-adaptive weighting" refers to a strategy that dynamically adjusts the weights of each evaluation indicator according to the current operating state of the gas turbine. The beneficial effects of this embodiment are: it achieves a leap from single-parameter monitoring to multi-parameter coupled intelligent evaluation through a comprehensive scoring model, providing intuitive and quantitative flexible indicators; and through spectral kurtosis analysis, it can accurately identify stalling trends in the early stages of fault occurrence, significantly improving the sensitivity of fault early warning.

[0080] Please refer to Figure 2 , Figure 2 This invention provides a structural block diagram of a heavy-duty gas turbine anti-surge bleed valve flexibility testing device; the specific device may include: The multi-source synchronous acquisition module 100 is used to synchronously acquire inlet and outlet pressure data, valve core angular displacement data, and valve body vibration data during the operation of the anti-surge venting valve under different operating conditions on the gas turbine test bench. The signal processing module 200 is used to perform time base alignment processing on the acquired multi-source data and perform filtering and noise reduction to eliminate environmental interference, so as to obtain a standardized multi-physics field coupled signal. The twin analysis and feature extraction module 300 is used to construct a digital twin model of the valve based on standardized multi-physics field coupled signals, calculate the following error between the measured motion trajectory and the ideal motion trajectory, and extract abnormal energy features from the vibration signal. The assessment and early warning module 400 is used to comprehensively calculate the valve flexibility score and determine the jamming risk trend based on the following error, abnormal energy characteristics and action response time, and generate an assessment result including the location of the root cause of the failure.

[0081] In some embodiments, the multi-source synchronous acquisition module 100 further includes a pressure sensing submodule, an angle sensing submodule, and a vibration sensing submodule. The pressure sensing submodule consists of two sets of high-precision pressure sensors, installed at the valve inlet and outlet respectively. The sensor front end integrates a microfluidic film cooling structure, and the body is encapsulated in high-temperature resistant ceramic. The angle sensing submodule uses a non-contact laser angle sensor, installed at the valve shaft extension end or outside the valve body, with a lens integrated with a purge air curtain. The vibration sensing submodule uses a piezoelectric accelerometer, fixed to key locations such as the valve cover or drive bracket by rigid bolts. All sensors are connected to a data acquisition card and receive the same hardware trigger pulse from a central clock.

[0082] In other embodiments, as an alternative implementation, the pressure sensor in the multi-source synchronous acquisition module 100 can be a sapphire fiber optic grating sensor, the angle sensor can be a magnetoelectric non-contact encoder, and the vibration sensor can be a MEMS accelerometer. Time reference alignment in the signal processing module 200 can be achieved using a GPS synchronized clock or the IEEE 1588 precise time protocol, and filtering and noise reduction can be achieved using empirical mode decomposition (EMD) or variational mode decomposition (VMD) methods.

[0083] In one specific embodiment, the temperature compensation of the pressure sensor in the multi-source synchronous acquisition module 100 follows the formula... The meanings of the symbols are as described above. The power adjustment of the angle sensor follows the formula... . As the reference power, As the reference reflectivity, The reflectivity of the valve body surface is detected in real time. The cross-correlation function in the signal processing module 200 is... ,in, Signals from a physical line Signals from another physical line, Let t be the time delay variable, and t be the time variable. The RLS filtering iteration formula is: The comprehensive error index in the twin analysis and feature extraction module 300. Abnormal vibration energy residual The assessment and early warning module has a comprehensive score of 400. spectral kurtosis .

[0084] Specifically, in this embodiment, the jitter of the rising edge of the hardware trigger pulse of the multi-source synchronous acquisition module 100 is controlled within 10μs, and the calibration coefficient of the pressure sensor is... reference temperature The RLS filter forgetting factor of signal processing module 200 is set to 0.98. The scaling factor of twin analysis and feature extraction module 300 is... The value range is 0.01-0.1. The adaptive weighting of the operating condition in the evaluation and early warning module 400 is set to [value missing] under the startup condition. .

[0085] It should be noted that the specific implementation of each module in this embodiment can be referred to the description of the corresponding steps in the previous method embodiment. The beneficial effects of the heavy-duty gas turbine anti-surge venting valve flexibility testing device in this embodiment are as follows: by integrating a multi-dimensional sensing module and an AI intelligent analysis system, it realizes comprehensive testing of the flexibility of the heavy-duty gas turbine anti-surge venting valve under extreme operating conditions such as high temperature and high pressure; by using multi-parameter synchronous sensing and high-precision time alignment technology, it overcomes the limitations of single-parameter monitoring and provides a panoramic data foundation; through the early warning mechanism of digital twin and multi-algorithm fusion, it can keenly identify early weak fault characteristics and accurately locate the root cause, significantly improving the intelligence level of testing and its engineering guidance value.

[0086] The heavy-duty gas turbine anti-surge vent valve flexibility testing device of this embodiment is used to implement the aforementioned heavy-duty gas turbine anti-surge vent valve flexibility testing method. Therefore, the specific implementation of the heavy-duty gas turbine anti-surge vent valve flexibility testing device can be found in the previous embodiment section of the heavy-duty gas turbine anti-surge vent valve flexibility testing method. For example, the multi-source synchronous acquisition module 100, signal processing module 200, twin analysis and feature extraction module 300, and evaluation and early warning module 400 are respectively used to implement steps S101, S102, S103, and S104 in the above-mentioned heavy-duty gas turbine anti-surge vent valve flexibility testing method. Therefore, its specific implementation can be referred to the description of the corresponding embodiment, and will not be repeated here.

[0087] A specific embodiment of the present invention also provides a heavy-duty gas turbine anti-surge venting valve flexibility testing device, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the above-described heavy-duty gas turbine anti-surge venting valve flexibility testing method.

[0088] A specific embodiment of the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for testing the flexibility of a heavy-duty gas turbine anti-surge venting valve.

[0089] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0090] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0091] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0092] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0093] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for testing the flexibility of a surge relief valve in a heavy-duty gas turbine, characterized in that, include: Under different operating conditions on the gas turbine test bench, the inlet and outlet pressure data, valve core angular displacement data, and valve body vibration data were collected simultaneously during the operation of the anti-surge venting valve. The acquired multi-source data is time-base aligned and filtered to eliminate environmental interference, resulting in a standardized multi-physics coupled signal. A digital twin model of the valve is constructed based on standardized multi-physics field coupled signals. The following error between the measured motion trajectory and the ideal motion trajectory is calculated, and the abnormal energy characteristics in the vibration signal are extracted. Based on the following error, abnormal energy characteristics, and action response time, the valve flexibility score is calculated comprehensively, the jamming risk trend is determined, and an assessment result including the location of the root cause of the failure is generated.

2. The method according to claim 1, characterized in that, The synchronous acquisition of inlet and outlet pressure data, valve core angular displacement data, and valve body vibration data during the operation of the anti-surge valve includes: By integrating a cooling structure and a high-temperature resistant encapsulation, the pressure sensor can collect real-time data on transient pressure changes at the valve inlet and outlet, under the condition that a cooling medium is introduced to form a protective gas film to isolate high temperatures. By using a non-contact laser angle sensor with adaptive wavelength and power adjustment mechanism, the laser parameters are adjusted according to the reflectivity of the valve body surface, and the full stroke angular displacement data of the valve core from start to stop are continuously recorded. The original signal of valve body vibration acceleration is captured by a piezoelectric accelerometer rigidly fixed to a key position on the valve body.

3. The method according to claim 2, characterized in that, The real-time acquisition of transient pressure change data at the valve inlet and outlet includes: Using a temperature drift dynamic compensation algorithm, the pressure reading is corrected online based on the real-time monitored sensor body temperature to obtain the corrected pressure data; Two sets of redundant pressure sensors are used to monitor the same measuring point. The difference between the readings of the two sets of sensors is calculated and compared with a preset threshold. When the difference exceeds the threshold, a single-point fault alarm is triggered.

4. The method according to claim 2, characterized in that, The continuously recorded angular displacement data of the valve core from start to stop includes: When a change in the reflectivity of the valve body surface is detected, the laser power is automatically adjusted based on the reference power and the reference reflectivity. In an environment where the lens is integrated with a purge curtain to prevent dust adhesion, multi-dimensional angle data including angular displacement, angular velocity and angular acceleration are output.

5. The method according to claim 1, characterized in that, The process of aligning the acquired multi-source data with a time reference and performing filtering and noise reduction to eliminate environmental interference includes: By using the same hardware trigger pulse from the central clock to receive all sensors, the acquisition start time is ensured to be strictly consistent. Then, the cross-correlation analysis algorithm is used to correct the time delay of signals from different physical lines. The time delay corresponding to the maximum value of the cross-correlation function is found and phase compensation is performed to achieve strict alignment of multi-source data in the time domain. Next, an adaptive filtering algorithm is used to process the aligned original signal to eliminate environmental interference and electromagnetic noise, resulting in a standardized multi-physics coupled signal.

6. The method according to claim 1, characterized in that, The valve digital twin model is constructed based on standardized multi-physics field coupled signals. The tracking error between the measured motion trajectory and the ideal motion trajectory is calculated, and abnormal energy features in the vibration signal are extracted, including: A digital twin simulation model of a valve based on state-space equations is established, defining a state vector containing angular displacement, angular velocity, and angular acceleration. The ideal motion trajectory is solved using the system matrix, input matrix, and output matrix. The motion following error is calculated by comparing the measured trajectory with the ideal trajectory, and the root mean square error is used as the comprehensive error index. Simultaneously, the theoretical excitation vibration energy is obtained from the measured angular acceleration, and compared with the measured vibration signal energy to calculate the abnormal vibration energy residual to characterize the abnormal energy features.

7. The method according to claim 6, characterized in that, The process of obtaining the theoretical excitation vibration energy from the measured angular acceleration and comparing it with the measured vibration signal energy to calculate the abnormal vibration energy residual to characterize the abnormal energy features includes: The vibration energy is excited based on the angular acceleration calculation theory, and the measured vibration signal energy is calculated at the same time. Then, the abnormal vibration energy residual is obtained through integration. This residual value directly reflects the additional vibration energy caused by valve core jamming or component wear.

8. The method according to claim 6, characterized in that, The process involves comprehensively calculating the valve flexibility score and determining the jamming risk trend based on following error, abnormal energy characteristics, and action response time, generating an assessment result that includes fault root cause localization. A comprehensive scoring model for valve flexibility, including a normalized evaluation function, is constructed, and a comprehensive score is calculated. The comprehensive score is obtained by multiplying the normalized evaluation function based on the root mean square of the following error, the residual of abnormal vibration energy, and the response time by the adaptive weights of the operating conditions and then summing them. Simultaneously, the spectral kurtosis algorithm is applied to calculate the spectral kurtosis value of the vibration signal to identify early frictional impact characteristics. When the spectral kurtosis value continues to rise, it is determined that the early jamming trend is strengthening.

9. A device for testing the flexibility of a surge relief valve in a heavy-duty gas turbine, characterized in that, include: The multi-source synchronous acquisition module is used to synchronously acquire inlet and outlet pressure data, valve core angular displacement data, and valve body vibration data during the operation of the anti-surge venting valve under different operating conditions on the gas turbine test bench. The signal processing module is used to perform time base alignment processing on the acquired multi-source data and perform filtering and noise reduction to eliminate environmental interference, so as to obtain a standardized multi-physics field coupled signal. The twin analysis and feature extraction module is used to construct a digital twin model of the valve based on standardized multi-physics field coupled signals, calculate the following error between the measured motion trajectory and the ideal motion trajectory, and extract abnormal energy features from the vibration signal. The assessment and early warning module is used to comprehensively calculate the valve flexibility score and determine the jamming risk trend based on the following error, abnormal energy characteristics and action response time, and generate assessment results including the location of the root cause of the failure.

10. A testing device for the flexibility of a heavy-duty gas turbine anti-surge venting valve, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the method for testing the flexibility of a heavy-duty gas turbine anti-surge venting valve as described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method for testing the flexibility of a heavy-duty gas turbine anti-surge venting valve as described in any one of claims 1 to 8.

Images