A method for monitoring fatigue damage and evaluating residual life of a nuclear power steam turbine blade, an electronic device and a computer readable storage medium

Abstract

The application provides a nuclear power steam turbine blade fatigue damage monitoring and residual life evaluation method, an electronic device and a computer readable storage medium, the evaluation method comprising: obtaining multiple-source non-contact monitoring signals based on adaptive sampling frequency and time synchronization technology, and preprocessing the multiple-source non-contact monitoring signals; constructing a blade full-damage cycle characteristic database based on simulation data and test data; inputting the preprocessed multiple-source non-contact monitoring signals into the full-damage cycle characteristic database, and extracting a fatigue damage feature vector; constructing a multi-parameter coupled damage quantization equation based on the fatigue damage feature vector, and calculating a blade current cumulative damage degree; establishing a load spectrum dynamic correction triggering mechanism for iterative calibration to generate an actual load spectrum; introducing a crack propagation rate coefficient based on a crack propagation theory, and combining the cumulative damage degree and the actual load spectrum to calculate a residual life, and triggering a warning mechanism. The application can monitor and evaluate the residual life of a nuclear power steam turbine blade.

Description

Technical Field

[0001] This invention relates to the field of nuclear power turbine blade life assessment technology, and in particular to a method, electronic device and computer-readable storage medium for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades. Background Technology

[0002] As the core power component of a nuclear power unit, the nuclear power turbine plays a crucial role in converting nuclear energy into mechanical energy. Its operational safety and reliability directly determine the stable output and nuclear safety boundaries of the nuclear power unit. The blades, as the core load-bearing components of the nuclear power turbine, operate under extreme conditions for extended periods, including temperatures exceeding 350°C, pressures exceeding 10MPa, and high-speed alternating loads within the cylinder. They also endure multiple effects such as steam flow erosion, electromagnetic radiation interference, and material fatigue aging, making them susceptible to fatigue damage issues such as microcrack initiation, stiffness reduction, and deformation propagation. Blade failure can lead to unplanned unit shutdowns and even nuclear safety risks, directly threatening the operational safety of the nuclear power plant. Therefore, it is necessary to provide a method for monitoring fatigue damage and assessing the remaining life of nuclear power turbine blades to meet the requirements of high-precision, full-cycle damage control. Summary of the Invention

[0003] This invention provides a method, electronic device, and computer-readable storage medium for monitoring fatigue damage and assessing the remaining life of nuclear power turbine blades, which can monitor and assess the remaining life of nuclear power turbine blades.

[0004] This invention provides a method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades, comprising:

[0005] After the nuclear power turbine enters normal operation, it acquires multi-source non-contact monitoring signals based on adaptive sampling frequency and time synchronization technology, and preprocesses the multi-source non-contact monitoring signals. A database of blade damage cycle characteristics was constructed based on simulation and experimental data; The preprocessed multi-source non-contact monitoring signals are input into the full damage cycle feature database, and fatigue damage feature vectors are extracted. A multi-parameter coupled damage quantification equation is constructed based on fatigue damage feature vectors, and the current cumulative damage degree of the blade is calculated. A dynamic correction triggering mechanism for the load spectrum is established for iterative calibration to generate the actual load spectrum. A crack propagation rate coefficient is introduced based on crack propagation theory, and the remaining life is calculated by combining the cumulative damage degree with the actual load spectrum, and an early warning mechanism is triggered.

[0006] In one embodiment of the present invention, multi-source non-contact monitoring signals are acquired based on adaptive sampling frequency and time synchronization technology, and the multi-source non-contact monitoring signals are preprocessed, specifically including: The sampling frequency is dynamically adjusted based on the dominant frequency of blade vibration, and is set to 5 to 10 times the dominant frequency of blade vibration, with a range of 20kHz to 50kHz. GPS timing synchronization technology and linear interpolation correction method are used to control the sampling time difference within 1ms. An adaptive electromagnetic interference suppression algorithm is used to remove 50Hz power frequency interference and high-frequency electromagnetic radiation noise, while a threshold filtering method is used to remove invalid noise. The 3σ criterion is used to detect and remove outliers, eliminate abnormal data, and smooth the missing data using adjacent time data.

[0007] In one embodiment of the present invention, a full-damage cycle characteristic database of a blade is constructed based on simulation data and experimental data, specifically including: By integrating blade dynamics simulation data and fatigue testing machine test data, a full damage cycle characteristic database covering the undamaged state, microcrack initiation state, and crack propagation state is constructed. The microcrack initiation state corresponds to a crack length of less than 0.5 mm, and the crack propagation state corresponds to a crack length between 0.5 mm and 5 mm.

[0008] In one embodiment of the present invention, the preprocessed multi-source non-contact monitoring signal is input into a full damage cycle feature database, and fatigue damage feature vectors are extracted, specifically including: An improved dictionary sparse inversion model is adopted, and the atomic parameters are dynamically adjusted by an adaptive atomic update mechanism to separate fatigue damage-specific feature parameters in vibration, gap and acoustic dimensions from multi-source non-contact monitoring signals. The separated fatigue damage-specific feature parameters are weighted and fused by the entropy weight method to construct a fatigue damage feature vector.

[0009] In one embodiment of the present invention, a multi-parameter coupled damage quantification equation is constructed based on fatigue damage feature vectors, and the current cumulative damage degree of the blade is calculated, specifically including: Using fatigue damage feature vectors as the core input, and introducing the material fatigue accumulation coefficient calculated based on Miner's linear cumulative damage theory, a multi-parameter coupled damage quantification equation is constructed by combining blade material fatigue test data and material fatigue life curves, and the current cumulative damage degree of the blade is calculated.

[0010] In one embodiment of the present invention, a dynamic correction triggering mechanism for the load spectrum is established for iterative calibration to generate an actual load spectrum, specifically including: By comparing the actual operating parameters with the initial design load spectrum, the peak load, load cycle number, and load fluctuation amplitude are analyzed to establish a deviation distribution matrix. Deviation judgment criteria are set. When the peak deviation between the actual load and the design load is greater than 15%, or the cycle number deviation is greater than 20%, or the fluctuation amplitude is greater than 10% for 30 consecutive minutes, the load spectrum iterative correction process is triggered. The actual load spectrum is generated by reverse calibration in combination with the current cumulative damage degree.

[0011] In one embodiment of the present invention, a crack propagation rate coefficient is introduced based on crack propagation theory, and the remaining lifetime is calculated by combining the cumulative damage degree and the actual load spectrum, and an early warning mechanism is triggered, specifically including: Starting from the current damage state of the blade, and combining the actual load spectrum variation law, the crack propagation path and rate are simulated to calculate the remaining service life of the blade from the current state to the critical damage state, and the life prediction range is output. If the current state of the blade is the critical state, the remaining life calculation is immediately terminated, the corresponding blade position and cumulative damage degree are locked, and the early warning mechanism is triggered.

[0012] In one embodiment of the present invention, the early warning mechanism is a three-level early warning mechanism constructed based on the cumulative damage level. The first-level early warning issues a notice of concern and increases the monitoring frequency to once every 10 minutes to continuously track damage changes. The second-level early warning triggers a planned maintenance recommendation and, in conjunction with the unit operation schedule, outputs a specific maintenance window period and testing plan. The third-level early warning immediately activates an emergency early warning and simultaneously outputs the specific maintenance location, fault level, handling priority, and emergency response plan.

[0013] The present invention also provides an electronic device, including a memory and a processor. The memory is used to store a computer program, and the processor, when running the computer program, implements the above-mentioned method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades.

[0014] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-described method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades.

[0015] The beneficial effects of this invention are: The present invention relates to a method, electronic equipment, and computer-readable storage medium for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades. First, it acquires multi-source non-contact monitoring signals using adaptive sampling frequency and timing synchronization technology. Preprocessing these signals effectively solves the problems of high-frequency data redundancy and loss of key transient information in traditional fixed sampling modes, significantly improving the temporal consistency and data quality of the multi-source non-contact monitoring signals and laying a high-fidelity data foundation for subsequent feature extraction. Second, it constructs a full-damage-cycle feature database for blades based on simulation and experimental data, overcoming the limitations of a single data source in covering the entire damage evolution lifecycle. This achieves complete feature mapping from initial damage initiation to critical failure state, enhancing the completeness and generalization ability of the assessment. Then, the preprocessed multi-source non-contact monitoring signals are input into the full-damage-cycle feature database, and fatigue damage feature vectors are extracted. Finally, a multi-parameter coupled damage quantification equation is constructed based on these fatigue damage feature vectors. The current cumulative damage degree of the blades is calculated, making the calculation result closer to the actual physical damage process. A dynamic correction triggering mechanism for the load spectrum is established for iterative calibration to generate the actual load spectrum. The introduction of the dynamic correction triggering mechanism for the load spectrum effectively addresses the load uncertainty caused by the time-varying characteristics of nuclear power turbine operating conditions. Real-time feedback correction ensures the dynamic adaptability of the remaining life assessment results to the current operating state. Finally, based on crack propagation theory, a crack propagation rate coefficient is introduced, and the remaining life is calculated by combining the cumulative damage degree with the actual load spectrum, triggering an early warning mechanism. This remaining life prediction based on crack propagation theory and the introduction of the crack propagation rate coefficient combines macroscopic damage mechanics with microscopic fracture mechanics. While ensuring the accuracy of the assessment, it achieves accurate characterization of the nonlinear crack propagation stage. Combined with the dynamic triggering of the early warning mechanism, it provides reliable technical support for the preventive maintenance and safe operation of nuclear power turbines, significantly improving the safety and economy of equipment operation. The method, electronic equipment, and computer-readable storage medium for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades of the present invention can monitor and assess the remaining life of nuclear power turbine blades, meeting the requirements for high-precision, full-cycle damage control. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0017] In the attached diagram: Figure 1This is a schematic diagram of the process structure of a method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades according to an embodiment of the present invention. Detailed Implementation

[0018] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0019] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. The drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0020] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0021] Before conducting fatigue damage monitoring and remaining life assessment of nuclear power turbine blades, corresponding preparatory work needs to be carried out, such as design optimization of the nuclear power turbine and sensor deployment and calibration, specifically including: During the design phase of the turbine cylinder and end cap, the design of optical windows, sensor mounting bases, and anti-interference wiring channels is carried out simultaneously, taking into account the running trajectory obtained from blade dynamics simulation, the airflow characteristics determined by fluid dynamics analysis, and the detection range and signal transmission path requirements of various types of sensors. The reserved design must be fully adapted to the service environment of nuclear power turbines with temperatures above 350°C and pressures above 10MPa. The sensor mounting base can be integrally cast from high-strength heat-resistant alloy material to ensure that the structural strength does not decrease under long-term operation. The wiring channel can adopt a double-layer sleeve structure, with an inner layer of ceramic heat insulation tube and an outer layer of stainless steel shielding tube, equipped with sealing flanges at both ends. This achieves triple protection of heat insulation, anti-electromagnetic interference, and sealing, while accurately covering key monitoring areas such as blade tips, leading edges, and trailing edges, without changing the surface morphology of the flow channel inside the cylinder or disturbing the airflow pattern, thus ensuring the original operating efficiency and structural stability of the turbine to adapt to the high-temperature, high-pressure, and high-speed operating conditions of nuclear power turbines.

[0022] After the sensor installation is completed, the sensors are calibrated. First, the turbine is started at low speed and the blade static position is used as a reference to calibrate the installation reference point and detection angle of each sensor. The detection paths of the laser and microwave sensors are checked one by one. Potential obstructions such as residual debris and oxide scale on the inner wall of the cylinder are removed. The sensor angles are adjusted to ensure that the detection light and microwave signal can continuously and stably cover the target monitoring area without blind spots or signal attenuation, laying a precise foundation for subsequent reference parameter setting.

[0023] Subsequently, the turbine underwent no-load and low-load steady-state operation. Benchmark parameters were specifically set for both operating conditions. For the laser Doppler vibration sensor, vibration data was collected under undamaged blade conditions to determine the inherent vibration benchmark values ​​and fluctuation ranges of the blades at different speeds. For the microwave radar sensor, the blade tip clearance benchmark value and dynamic fluctuation threshold under undamaged blade conditions were calibrated. For the fiber optic acoustic sensor, by comparing the signal differences between the shutdown and no-load operating states, the mechanical background noise and fluid disturbance noise of the equipment were distinguished, the background noise threshold was accurately calibrated, and invalid noise interference was eliminated. Through repeated testing and data recording under various operating conditions, including no-load, low-load, medium-load, and high-load, data was collected after 30 minutes of stable operation under each condition, repeated three times, and the average value was taken. Various benchmark parameters were integrated to construct a benchmark library of undamaged state signals covering the entire operating range, providing a traceable and comparable reliable reference for subsequent fatigue damage identification, location, and analysis.

[0024] Please see Figure 1 This invention provides a method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades, comprising: Step S1: After the nuclear power turbine enters normal operation, multi-source non-contact monitoring signals are acquired based on adaptive sampling frequency and time synchronization technology, and the multi-source non-contact monitoring signals are preprocessed.

[0025] This step acquires multi-source non-contact monitoring signals through adaptive sampling frequency and time synchronization technology, and preprocesses the multi-source non-contact monitoring signals. This effectively solves the problems of high-frequency data redundancy and loss of key transient information in the traditional fixed sampling mode, significantly improving the temporal consistency and data quality of multi-source non-contact monitoring signals, and laying a high-fidelity data foundation for subsequent feature extraction. The acquisition of multi-source non-contact monitoring signals can be achieved by relying on a high-precision synchronous acquisition module (16-bit sampling accuracy). The acquisition module can specifically include a laser Doppler vibration sensor, a microwave radar sensor, and a fiber optic acoustic sensor.

[0026] In one embodiment of the present invention, step S1 may specifically include: Step S11: Dynamically adjust the sampling frequency according to the main frequency of blade vibration, and set the sampling frequency to 5 to 10 times the main frequency of blade vibration, with the range controlled between 20kHz and 50kHz.

[0027] This step sets the sampling frequency to 5 to 10 times the dominant frequency of blade vibration, within the range of 20kHz to 50kHz, to ensure complete capture of subtle changes in blade vibration; simultaneously acquires blade vibration timing data, blade tip gap dynamic data, and fluid coupling acoustic data; specific data transmission can be achieved using fiber optic transmission to avoid electromagnetic interference.

[0028] Step S12: Use GPS time synchronization technology and linear interpolation correction method to control the sampling time difference within 1ms.

[0029] This step uses GPS timing synchronization technology and linear interpolation correction method to control the sampling time difference within 1ms, so as to uniformly calibrate the sampling frequency of each sensor to the same reference, thereby compensating for the sampling time difference between different sensors, accurately aligning the timing dimension of multi-source signals, completely solving the timing misalignment problem caused by inconsistent sampling frequencies, and ensuring the time synchronization of multi-source data.

[0030] Step S13: Use an adaptive electromagnetic interference suppression algorithm to remove 50Hz power frequency interference and high-frequency electromagnetic radiation noise, and use a threshold filtering method to remove invalid noise.

[0031] This step addresses the interference of strong electromagnetic radiation (i.e., electromagnetic field strength ≥100V / m) on signal transmission in nuclear power scenarios. It employs an adaptive electromagnetic interference suppression algorithm to eliminate 50Hz power frequency interference and high-frequency electromagnetic radiation noise, while using a threshold filtering method to eliminate sudden invalid noise such as equipment start-up and shutdown impacts and external environmental vibrations. This maximizes the retention of effective signal components related to blade vibration and damage, ensuring a signal-to-noise ratio ≥30dB.

[0032] Step S14: Use the 3σ criterion to detect and remove outliers, exclude abnormal data, and use adjacent time-time data to smoothly complete missing data; This step uses the 3σ criterion to detect and remove outliers from the collected data, eliminating abnormal data points caused by factors such as sensor failure, signal transmission interruption, and transient electromagnetic interference. Missing data is smoothly filled in using data from adjacent time points, and finally a standardized time series dataset is generated to ensure that the data has consistency, integrity and reliability, with a data validity rate of no less than 99.5%.

[0033] Step S2: Construct a full-cycle damage characteristic database for blades based on simulation and experimental data; This step constructs a full-cycle damage characteristic database for blades based on simulation and experimental data, overcoming the limitations of a single data source in covering the entire life cycle of damage evolution. It achieves a complete feature mapping from initial damage initiation to critical failure state, enhancing the completeness and generalization ability of the assessment.

[0034] In one embodiment of the present invention, step S2 may specifically include: By integrating blade dynamics simulation data and fatigue testing machine test data, a full damage cycle characteristic database covering the undamaged state, microcrack initiation state, and crack propagation state is constructed. The microcrack initiation state corresponds to a crack length of less than 0.5 mm, and the crack propagation state corresponds to a crack length between 0.5 mm and 5 mm.

[0035] The dynamic simulation data in this step can be obtained from ANSYS blade dynamic simulation, covering blade vibration, clearance, and acoustic simulation data under different loads and speeds, for example, a total of 1000 sets; fatigue testing machine test data can be obtained by selecting 3 undamaged standard blades and 5 blades with pre-fabricated microcracks (e.g., crack length 0.1-0.5mm) to simulate damaged blades, for a total of 300 sets, and conducting multi-condition tests on the fatigue testing machine to obtain test data; based on the dynamic simulation data and fatigue testing machine test data, a database covering the characteristics of the entire damage cycle of the blade is constructed. The database clearly divides the state into three stages: undamaged state, microcrack initiation state (crack length ≤0.5mm), and crack propagation state (crack length 0.5-5mm). Each stage includes three types of characteristic parameters: vibration, clearance, and acoustics, covering 12 indicators such as amplitude, frequency, phase, and peak spectral value, providing a comprehensive and detailed reference for accurate identification of damage characteristics.

[0036] Step S3: Input the preprocessed multi-source non-contact monitoring signal into the full damage cycle feature database and extract the fatigue damage feature vector.

[0037] In one embodiment of the present invention, step S3 may specifically include: Step S31: Using an improved dictionary sparse inversion model and combining it with an adaptive atom update mechanism to dynamically adjust the atom parameters, fatigue damage-specific feature parameters of vibration dimension, gap dimension and acoustic dimension are separated from the multi-source non-contact monitoring signals.

[0038] This step employs an improved dictionary sparse inversion model combined with an adaptive atom update mechanism. It designs a dedicated atomic structure based on the blade's inherent vibration modes and material mechanical properties, and dynamically adjusts the atomic parameters to improve the matching degree between the dictionary and the damage signal by more than 40%. This significantly enhances the separation accuracy between the damage signal and the interference signal, effectively extracting weak damage features. The preprocessed multi-source non-contact monitoring signal is input into the improved dictionary sparse inversion model to separate fatigue damage-specific feature parameters in the vibration dimension (stiffness attenuation, frequency shift), gap dimension (radial deformation, gap fluctuation), and acoustic dimension (spectral distortion, coupling noise).

[0039] Step S32: Use the entropy weighting method to weight and fuse the separated fatigue damage-specific feature parameters to construct a fatigue damage feature vector.

[0040] This step uses the entropy weighting method to weight and fuse the three types of feature parameters separated—vibration dimension, gap dimension, and acoustic dimension—to construct a multidimensional fatigue damage feature vector. This reduces the interference of non-damage factors such as load fluctuations and airflow disturbances, ensuring that the feature vector truly and accurately reflects the blade damage state.

[0041] Step S4: Construct a multi-parameter coupled damage quantification equation based on fatigue damage feature vectors, and calculate the current cumulative damage degree of the blade.

[0042] This step constructs a multi-parameter coupled damage quantification equation based on fatigue damage feature vectors to calculate the current cumulative damage degree of the blade, making the calculation result of the cumulative damage degree closer to the actual physical damage process and improving the accuracy of subsequent life assessment.

[0043] In one embodiment of the present invention, step S4 may specifically include: Using fatigue damage feature vectors as the core input, and introducing the material fatigue accumulation coefficient calculated based on Miner's linear cumulative damage theory, a multi-parameter coupled damage quantification equation is constructed by combining blade material fatigue test data and material fatigue life curves, and the current cumulative damage degree of the blade is calculated.

[0044] This step uses the damage feature vector generated in step S32 as the core input, and combines the fatigue test data of the blade material (e.g., martensitic stainless steel) with the material's SN fatigue life curve to construct a multi-parameter coupled damage quantification equation. The equation incorporates basic parameters such as material tensile strength, fatigue limit, and elastic modulus, while fully considering operating conditions (speed, load, steam temperature) and damage accumulation law. Based on Miner's linear cumulative damage theory, a material fatigue accumulation coefficient is introduced to accurately calculate the current cumulative fatigue damage degree of the blade. The quantification result error is controlled within 5% to accurately reflect the blade damage state.

[0045] Step S5: Establish a dynamic correction triggering mechanism for the load spectrum to perform iterative calibration and generate the actual load spectrum; This step introduces a dynamic load spectrum correction triggering mechanism for iterative calibration to generate an actual load spectrum. This effectively addresses the load uncertainty caused by the time-varying characteristics of nuclear power turbine operating conditions and ensures the dynamic adaptability of the remaining life assessment results to the current operating state through real-time feedback correction.

[0046] In one embodiment of the present invention, step S5 may specifically include: Step S51: Compare the actual operating parameters with the initial design load spectrum, analyze the peak load, load cycle number, and load fluctuation amplitude to establish a deviation distribution matrix.

[0047] This step can be based on real-time monitoring of operating data such as rotational speed, load, steam temperature and pressure by sensors, combined with the deviation between actual and design operating conditions fed back by damage feature vectors, to carry out preliminary analysis of dynamic correction of load spectrum. By comparing the differences between actual operating parameters and initial design load spectrum, the deviation of three core indicators—peak load, number of load cycles, and load fluctuation amplitude—is analyzed in detail, and a deviation analysis matrix is ​​established.

[0048] Step S52: Set deviation judgment criteria. When the peak deviation between the actual load and the design load is greater than 15%, or the deviation of the number of cycles exceeds 20%, or the fluctuation amplitude exceeds 10% for 30 consecutive minutes, the load spectrum iterative correction process is triggered, and the actual load spectrum is generated by reverse calibration in combination with the current cumulative damage degree.

[0049] This step immediately initiates the load spectrum iterative correction process when the peak deviation between the actual load and the design load exceeds 15%, or the deviation in the number of cycles exceeds 20%, or the load fluctuation exceeds 10% within 30 consecutive minutes. This avoids the accumulation of deviations that could lead to distortion in damage quantification. The iterative correction generates an actual load spectrum that fits the current service status. During the correction process, the load parameters are calibrated in reverse by combining the blade damage quantification results. For example, the process is iterated 3-5 times until the deviation between the calculated load spectrum value and the actual monitored value is ≤5%. This avoids damage quantification errors caused by the mismatch between the design load and the actual operating conditions, significantly improves the accuracy of damage calculation, and provides accurate load data support for subsequent remaining service life assessment.

[0050] Step S6: Introduce a crack propagation rate coefficient based on crack propagation theory, calculate the remaining life by combining the cumulative damage degree with the actual load spectrum, and trigger an early warning mechanism.

[0051] This step, based on crack propagation theory and incorporating a crack propagation rate coefficient for remaining service prediction, combines macroscopic damage mechanics with microscopic fracture mechanics. While ensuring assessment accuracy, it accurately characterizes the nonlinear crack propagation stage. Coupled with the dynamic triggering of an early warning mechanism, it provides reliable technical support for preventative maintenance and safe operation of nuclear power turbines, significantly improving the safety and economy of equipment operation. The crack propagation rate coefficient can be obtained through joint inversion of tip clearance fluctuation and vibration frequency offset. A binary regression analysis method is used to fit the coefficient value, leveraging the synergistic effect of the two parameters to improve the accuracy of coefficient calculation, reducing prediction error by 30% and enhancing reliability.

[0052] In one embodiment of the present invention, step S6 may specifically include: Step S61: Starting from the current damage state of the blade, and combining the actual load spectrum variation law, simulate the crack propagation path and rate, calculate the remaining service life of the blade from the current state to the critical damage state, and output the life prediction range. The critical damage state standard in this step can be: when the cumulative damage degree of the blade is ≥80%, or the dynamic change in the blade tip clearance exceeds 5mm, or the crack length is ≥5mm, it is judged as a critical state. Furthermore, this step can dynamically adjust the update cycle according to operating conditions, establishing an iterative update mechanism for remaining life assessment. Specifically, under harsh operating conditions such as high load (≥80% of rated load) and high speed (fluctuation ±50r / min), the damage accumulation rate accelerates, shortening the update cycle to 1 hour / time, improving the timeliness of the assessment. Under normal stable operating conditions, a regular update cycle of 4 hours / time is adopted, continuously iteratively correcting the cumulative damage degree and remaining life value using the latest data collected by sensors, ensuring that the assessment results match the actual blade state in real time. Step S61: If the current state of the blade is critical, immediately terminate the remaining life calculation, simultaneously lock the corresponding blade position and cumulative damage degree, and trigger the early warning mechanism.

[0053] In one embodiment of the present invention, the early warning mechanism can be a three-level early warning mechanism constructed according to the cumulative damage degree. The first-level early warning (minor damage, cumulative damage degree 20%-30%) issues a notice of concern and increases the monitoring frequency to once every 10 minutes to continuously track damage changes. The second-level early warning (moderate damage, cumulative damage degree 30%-60%) triggers a planned maintenance recommendation and outputs specific maintenance windows and testing plans in conjunction with the unit operation schedule. The third-level early warning (severe damage, cumulative damage degree 60%-80%) immediately activates an emergency early warning and simultaneously outputs specific maintenance locations, fault levels, handling priorities, and emergency response plans to guide maintenance personnel to complete emergency response within 2 hours, prevent the fault from escalating, and realize a closed-loop management of the entire chain from minor damage tracking, planned maintenance of moderate damage to emergency handling of severe damage, adapting to the maintenance needs under different damage states.

[0054] In summary, the method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades of this invention, through synchronous optimization of turbine design and sensor deployment, combined with structural design and sealing protection measures adapted to high-temperature and high-pressure conditions, achieves accurate monitoring of critical blade areas under all operating conditions without disturbing the in-cylinder flow field or attenuating the original operating efficiency and structural strength of the unit. This completely solves the technical pain points of poor adaptability of traditional sensor deployment to the harsh operating conditions of nuclear power and its tendency to interfere with unit operation. Furthermore, through adaptive anti-interference algorithms, time-series synchronous calibration, and abnormal data screening, it effectively eliminates complex noise interference such as electromagnetic radiation and fluid disturbances, ensuring a stable signal-to-noise ratio. A high-precision full-operation-range benchmark signal library is constructed, providing reliable data support for damage feature extraction, significantly improving the sensitivity and accuracy of weak damage signal identification, and achieving accurate monitoring of the blade microcrack initiation stage. Early and accurate damage detection; simultaneously, based on an improved sparse inversion algorithm for damage feature extraction and a multi-parameter coupled damage quantification model, combined with a dynamic iterative correction mechanism for the load spectrum, it fully adapts to the operating characteristics of nuclear power turbines under varying operating conditions, effectively compensating for the deviation between design loads and actual service conditions, controlling the damage quantification error to within 5%, significantly improving the accuracy of fatigue damage calculation, and achieving dynamic and accurate assessment of remaining life. This breaks through the limitations of traditional static calculations, which are difficult to adapt to the damage accumulation patterns under varying operating conditions. Finally, it accurately outputs differentiated operation and maintenance suggestions and emergency plans based on the degree of damage, ensuring rapid emergency response in the event of severe damage, avoiding major safety risks such as blade fracture and unit shutdown, significantly reducing operation and maintenance costs and resource consumption, and possessing engineering adaptability and scalability. It can be compatible with the structural characteristics and operating conditions of different types of nuclear power turbines, significantly broadening the application scenarios of the technology.

[0055] This invention also provides an electronic device, including a memory and a processor. The memory stores a computer program, and the processor, when running the computer program, implements the aforementioned method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades. The assessment method is the same as above and will not be repeated here.

[0056] This invention also provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the aforementioned method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades. The assessment method is the same as described above and will not be repeated here.

[0057] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0058] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades, characterized in that, include: After the nuclear power turbine enters normal operation, it acquires multi-source non-contact monitoring signals based on adaptive sampling frequency and time synchronization technology, and preprocesses the multi-source non-contact monitoring signals. A database of blade damage cycle characteristics was constructed based on simulation and experimental data; The preprocessed multi-source non-contact monitoring signal is input into the full damage cycle feature database, and fatigue damage feature vector is extracted. A multi-parameter coupled damage quantification equation is constructed based on the fatigue damage feature vector, and the current cumulative damage degree of the blade is calculated. A dynamic correction triggering mechanism for the load spectrum is established for iterative calibration to generate the actual load spectrum. A crack propagation rate coefficient is introduced based on crack propagation theory, and the remaining lifetime is calculated by combining the cumulative damage degree with the actual load spectrum, and an early warning mechanism is triggered.

2. The method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades according to claim 1, characterized in that, Multi-source non-contact monitoring signals are acquired based on adaptive sampling frequency and time synchronization technology, and the multi-source non-contact monitoring signals are preprocessed, specifically including: The sampling frequency is dynamically adjusted based on the dominant frequency of blade vibration, and is set to 5 to 10 times the dominant frequency of blade vibration, with a range of 20kHz to 50kHz. GPS timing synchronization technology and linear interpolation correction method are used to control the sampling time difference within 1ms. An adaptive electromagnetic interference suppression algorithm is used to remove 50Hz power frequency interference and high-frequency electromagnetic radiation noise, while a threshold filtering method is used to remove invalid noise. The 3σ criterion is used to detect and remove outliers, exclude abnormal data, and smooth the missing data using adjacent time data.

3. The method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades according to claim 1, characterized in that, A database of blade damage cycle characteristics was constructed based on simulation and experimental data, specifically including: By integrating blade dynamics simulation data and fatigue testing machine test data, a full damage cycle feature database covering the undamaged state, microcrack initiation state, and crack propagation state is constructed. The microcrack initiation state corresponds to a crack length of less than 0.5 mm, and the crack propagation state corresponds to a crack length between 0.5 mm and 5 mm.

4. The method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades according to claim 1, characterized in that, The preprocessed multi-source non-contact monitoring signals are input into the full damage cycle feature database, and fatigue damage feature vectors are extracted, specifically including: An improved dictionary sparse inversion model is adopted, and the atomic parameters are dynamically adjusted by an adaptive atomic update mechanism to separate fatigue damage-specific feature parameters of vibration dimension, gap dimension and acoustic dimension from the multi-source non-contact monitoring signal. The separated fatigue damage-specific feature parameters are weighted and fused by the entropy weight method to construct a fatigue damage feature vector.

5. The method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades according to claim 1, characterized in that, Based on the fatigue damage feature vector, a multi-parameter coupled damage quantification equation is constructed, and the current cumulative damage degree of the blade is calculated, specifically including: Using the fatigue damage feature vector as the core input, and introducing the material fatigue accumulation coefficient calculated based on Miner's linear cumulative damage theory, combined with the blade material fatigue test data and material fatigue life curve, a multi-parameter coupled damage quantification equation is constructed, and the current cumulative damage degree of the blade is calculated.

6. The method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades according to claim 1, characterized in that, Establish a dynamic load spectrum correction triggering mechanism for iterative calibration to generate the actual load spectrum, specifically including: By comparing the actual operating parameters with the initial design load spectrum, the peak load, load cycle number, and load fluctuation amplitude are analyzed to establish a deviation distribution matrix. A deviation judgment standard is set. When the peak deviation between the actual load and the design load is greater than 15%, or the cycle number deviation is greater than 20%, or the fluctuation amplitude is greater than 10% for 30 consecutive minutes, the load spectrum iterative correction process is triggered. The actual load spectrum is generated by reverse calibration in combination with the current cumulative damage degree.

7. The method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades according to claim 1, characterized in that, Based on crack propagation theory, a crack propagation rate coefficient is introduced, and the remaining lifetime is calculated by combining the cumulative damage and the actual load spectrum, triggering an early warning mechanism, specifically including: Starting from the current damage state of the blade, and combining the variation law of the actual load spectrum, the crack propagation path and rate are simulated, the remaining service life of the blade from the current state to the critical damage state is calculated, and the life prediction range is output. If the current state of the blade is the critical state, the remaining life calculation is immediately terminated, the corresponding blade position and the cumulative damage degree are locked, and the early warning mechanism is triggered.

8. The method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades according to claim 7, characterized in that, The early warning mechanism is a three-level early warning mechanism built according to the cumulative damage level. The first-level early warning issues a notice of concern and increases the monitoring frequency to once every 10 minutes to continuously track damage changes. The second-level early warning triggers planned maintenance recommendations and outputs specific maintenance windows and testing plans in conjunction with the unit operation schedule. The third-level early warning immediately activates the emergency warning and simultaneously outputs specific maintenance locations, fault levels, handling priorities, and emergency response plans.

9. An electronic device comprising a memory and a processor, characterized in that, The memory is used to store computer programs, and when the processor runs the computer programs, it implements the method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for monitoring fatigue damage and assessing remaining life of nuclear power turbine blades as described in any one of claims 1 to 8.

Images