A method for evaluating micro fatigue failure of a nuclear fuel rack

Abstract

The application provides a nuclear fuel grid micro-oscillation fatigue failure evaluation method, and relates to the technical field of nuclear fuel grids, and comprises the following steps: obtaining an elastic wave signal and a normal clamping force of the nuclear fuel grid, then removing noise information in the signal to extract a pure high-frequency micro-oscillation impact signal, combining the normal clamping force to construct a time-domain load spectrum, calculating fatigue damage according to the load spectrum, and judging whether the grid fails or not. Through effective noise reduction and feature extraction on the elastic wave signal in a strong noise background, the real time-domain characteristics of the micro-oscillation impact can be accurately restored, so that the key time node of the transition of the contact surface from adhesion to sliding can be accurately positioned, the accuracy and reliability of the nuclear fuel grid micro-oscillation fatigue failure evaluation are significantly improved, the traditional offline estimation is realized to the full life cycle online evaluation based on the real dynamic load, and the structural integrity of the nuclear fuel assembly under the long-term service condition is effectively ensured.

Description

Technical Field

[0001] This invention generally relates to the field of nuclear fuel grid technology, and specifically to a method for assessing fretting fatigue failure of nuclear fuel grids. Background Technology

[0002] Nuclear fuel assemblies are the core of a reactor, and the nuclear fuel grid used for positioning is one of the most important structural components of the fuel assembly. The springs and rigid protrusions inside the grid support and position the fuel rods using frictional clamping forces.

[0003] During the long-term operation of a pressurized water reactor, the high-speed flow of the high-temperature, high-pressure coolant causes flow-induced vibrations in the fuel rods. This high-frequency, micro-amplitude vibration leads to continuous fretting friction at the contact points between the fuel rod cladding and the grid springs. Over time, the contact surface condition gradually deteriorates from initial "complete adhesion" to "local slippage" and even "macro-slippage," resulting in severe fretting wear and fatigue damage. Once the spring clamping force fails or the cladding is worn through, it will lead to a very serious radioactive material leakage accident.

[0004] To evaluate the fretting wear characteristics of grid springs, the industry currently typically uses mechanical fretting friction testing machines or offline calculations combined with Archard wear theory. However, existing testing and evaluation methods have shortcomings: In service environments or on simulated test benches, there is significant mechanical and fluid background noise. Existing testing methods mostly rely on macroscopic displacement or contact force sensors, whose sensitivity is simply insufficient to capture the "microscopic high-frequency abrupt changes" that occur on the contact surface in the early stages of wear. This makes it impossible to accurately define the critical time points for fatigue crack initiation and wear acceleration, and consequently, it is difficult to accurately determine whether the nuclear fuel grid has failed. Summary of the Invention

[0005] In view of the above-mentioned defects or deficiencies in the prior art, it is desirable to provide a method for assessing the fretting fatigue failure of nuclear fuel grids, which can improve the accuracy of nuclear fuel grid assessment.

[0006] This invention provides a method for assessing the fretting fatigue failure of nuclear fuel grids, comprising: S1: Acquire the elastic wave signal of the nuclear fuel grid and the normal clamping force of the nuclear fuel grid on the fuel rods; S2: Remove noise information from the elastic wave signal to obtain a high-frequency micro-motion impact signal that contains only the vibration state of the nuclear fuel grid; S3: Construct the time-domain load spectrum of the nuclear fuel grid based on the high-frequency micro-motion impact signal and the normal clamping force; S4: Calculate the fatigue damage of the nuclear fuel grid based on the time-domain load spectrum; S5: Use the fatigue damage to determine whether the nuclear fuel grid has failed.

[0007] According to the technical solution provided by the present invention, noise information in the elastic wave signal is removed to obtain a high-frequency micro-motion impact signal containing only the vibration state of the nuclear fuel grid, including: S2-1: Extract multiple modal function components with initial center frequencies based on the elastic wave signal; S2-2: Optimize the multiple modal function components to ensure reconstruction accuracy and obtain multiple first modal function components; S2-3: Calculate the Pearson correlation coefficient between each optimal modal function component and the elastic wave signal; S2-4: Eliminate the optimal mode function components whose Pearson correlation coefficient is less than the correlation coefficient threshold to remove noise information and obtain multiple remaining optimal mode function components; S2-5: The components of multiple residual optimal mode functions are linearly superimposed to obtain a high-frequency micro-motion impact signal.

[0008] According to the technical solution provided by the present invention, multiple modal function components are optimized to ensure reconstruction accuracy, resulting in multiple optimal modal function components with optimal center frequencies, including: S2-2-1: Construct an augmented Lagrange function with a quadratic penalty factor and Lagrange multipliers; S2-2-2: Using the augmented Lagrangian function as a constraint, the modal function components and center frequency are optimized to obtain the optimal center frequency and the corresponding optimal modal function components.

[0009] According to the technical solution provided by the present invention, the time-domain load spectrum of the nuclear fuel grid is constructed based on the high-frequency micro-motion impact signal and the normal clamping force, including: S3-1: Extract the voltage peak value based on the high-frequency micro-motion impact signal; S3-2: Calculate the maximum dynamic contact compressive stress amplitude between the nuclear fuel grid and the fuel rods based on the peak voltage; S3-3: Construct a time-domain load spectrum based on the normal clamping force and the maximum dynamic contact compressive stress amplitude.

[0010] According to the technical solution provided by the present invention, the maximum dynamic contact compressive stress amplitude between the nuclear fuel grid and the fuel rods is calculated based on the voltage peak value, including: S3-2-1: Calculate the equivalent dynamic impact force at the contact interface between the nuclear fuel grid and the fuel rods based on the voltage peak value; S3-2-2: Based on the elastic modulus and Poisson's ratio of the nuclear fuel grid and the elastic modulus and Poisson's ratio of the fuel rods, a nonlinear mapping relationship between the maximum dynamic contact compressive stress amplitude and the equivalent dynamic impact force is constructed. S3-2-3: The maximum dynamic contact compressive stress amplitude is calculated using the nonlinear mapping relationship and the equivalent dynamic impact force.

[0011] According to the technical solution provided by the present invention, a time-domain load spectrum is constructed based on the normal clamping force and the maximum dynamic contact compressive stress amplitude, including: S3-3-1: Extract the slip transition time point between the nuclear fuel grid and fuel rods from adhesion to slip based on high-frequency micro-motion impact signals; S3-3-2: Construct a time-domain load spectrum based on the slip change time point and the normal clamping force, such that when the time point is earlier than the slip change time point, the time-domain load spectrum is equal to the normal clamping force; when the time point is later than or equal to the slip change time point, the time-domain load spectrum is equal to the sum of the normal clamping force and the maximum dynamic contact compressive stress amplitude, or the difference between the normal clamping force and the maximum dynamic contact compressive stress amplitude.

[0012] According to the technical solution provided by the present invention, the slip change time point is extracted based on the high-frequency micro-motion impact signal, including: S3-3-1-1: Set a sliding window for the high-frequency micro-motion impact signal and calculate multiple real-time kurtosis; S3-3-1-2: Calculate the mean and standard deviation of the real-time kurtosis of the first set number of the earliest time points among the multiple real-time kurtosis; S3-3-1-3: Set the kurtosis abrupt change threshold based on the real-time kurtosis mean and standard deviation; S3-3-1-4: Among the multiple real-time kurtosis values ​​that are greater than or equal to the kurtosis mutation threshold and are consecutively set to a second set number, the earliest corresponding time point is taken as the slip mutation time point.

[0013] According to the technical solution provided by the present invention, calculating the fatigue damage of the nuclear fuel grid based on the time-domain load spectrum includes: S4-1: Input the time-domain load spectrum into the digital twin model of the nuclear fuel grid and fuel rods to calculate the stress-life curve of the nuclear fuel grid; S4-2: Calculate the number of cycles for each stress amplitude based on the time-domain load spectrum; S4-3: The fatigue damage is calculated based on the number of cycles and the stress-life curve.

[0014] According to the technical solution provided by the present invention, the method further includes: if the nuclear fuel grid has not failed, predicting the failure time point of the nuclear fuel grid using the time-domain load spectrum.

[0015] According to the technical solution provided by the present invention, predicting the failure time point of a nuclear fuel grid using the time-domain load spectrum includes: The fatigue damage growth rate is calculated based on the time-domain load spectrum. Based on the growth rate, the remaining number of cycles or remaining time required for fatigue damage to reach the failure threshold is calculated using the load spectrum extrapolation method. The failure time point can be calculated using the remaining number of cycles or the remaining duration.

[0016] The beneficial effects of this invention are as follows: To address the technical problem in existing technologies where strong background noise within the reactor makes it difficult for traditional sensors to capture the "microscopic high-frequency abrupt changes" at the contact surface between the nuclear fuel grid and fuel rods during the early stages of wear, thus hindering the accurate determination of the slip transition and fatigue damage accumulation process, ultimately resulting in low accuracy in failure assessment and difficulty in timely early warning, this invention adopts the following approach: First, the elastic wave signal and normal clamping force of the nuclear fuel grid are acquired. Then, noise information in the signal is removed to extract a pure high-frequency fretting impact signal. Next, a time-domain load spectrum is constructed by combining the normal clamping force. Finally, fatigue damage is calculated based on this load spectrum to determine whether the grid has failed. By effectively reducing noise and extracting features from the elastic wave signal against a strong noise background, the true time-domain characteristics of the fretting impact can be accurately restored, thereby accurately locating the critical time node of the contact surface transitioning from adhesion to slip. The time-domain load spectrum constructed on this basis provides a high-fidelity input for fatigue damage calculation, significantly improving the accuracy and reliability of nuclear fuel grid fretting fatigue failure assessment. This achieves a shift from traditional offline estimation to full-lifecycle online assessment based on real dynamic loads, effectively ensuring the structural integrity of nuclear fuel assemblies under long-term service conditions. Attached Figure Description

[0017] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a flowchart illustrating a method for assessing the fretting fatigue failure of a nuclear fuel grid. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0019] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0020] refer to Figure 1The present invention provides a method for assessing the fretting fatigue failure of nuclear fuel grids. This method is loaded into a computer system, and the computer system can execute the following steps according to the corresponding program to complete the assessment of the fretting fatigue failure of nuclear fuel grids.

[0021] The specific steps include: S1: The system acquires the elastic wave signal of the nuclear fuel grid and the normal clamping force of the nuclear fuel grid on the fuel rods; Specifically, a broadband acoustic emission sensor with a resonant frequency covering 100kHz to 1MHz is attached to the non-contact surface (the surface that does not contact the fuel rods) of the nuclear fuel grid using a special coupling agent to collect elastic wave signals; a high-precision piezoelectric force sensor is connected in series behind the base used to fix the nuclear fuel grid to collect macroscopic normal clamping force.

[0022] The broadband acoustic emission sensor is connected to the multi-channel high-speed data acquisition system. To avoid high-frequency signal aliasing distortion, according to the Nyquist-Shannon sampling theorem, the system sampling frequency must satisfy the following:

[0023] in, f s Sampling frequency, f max This represents the highest frequency generated by microscopic impacts at the contact interface. In this embodiment, the sampling frequency... f s The speed is set from 2MS / s to 5MS / s, and the normal clamping force and the elastic wave signal are synchronized at the microsecond level using the same hardware crystal clock of the acquisition card.

[0024] S2: The system removes noise from the elastic wave signal to obtain a high-frequency micro-motion impact signal containing only the vibration state of the nuclear fuel grid, including: S2-1: Extract multiple modal function components with different initial center frequencies based on the elastic wave signal; To address strong background noise, this embodiment employs the Variational Mode Decomposition (VMD) algorithm for processing, aiming to improve the elastic wave signal... f(t) Decomposed into K A number with a specific initial center frequency modal function components Its constrained variational model is defined as follows:

[0025] The constraints are:

[0026] in, This indicates a minimization problem. The goal is to find multiple optimal modal function components. u k The set , and the set of its multiple optimal center frequencies This minimizes the sum of the bandwidths of each mode calculated within the curly braces.

[0027] This indicates the pre-defined number of decomposition modes; Indicates all Summation is performed on each mode; This represents the k-th modal function component obtained from the decomposition; This represents the center frequency corresponding to the k-th modal function component; This represents the Dirac function, also known as the unit impulse function; Represents the imaginary unit; This represents the convolution operator; Indicates time t Find the partial derivatives; This represents a complex exponential term. Multiplying it by the preceding analytic signal is equivalent to performing a spectrum shift. This indicates the elastic wave signal collected at the input terminal; in this embodiment, it is the elastic wave signal. Represents a time variable; Pi is a mathematical constant. This represents the square of the norm of the signal.

[0028] S2-2: Optimize multiple modal function components to ensure reconstruction accuracy and obtain multiple optimal modal function components with optimal center frequencies; multiple first modal function components have optimal center frequencies (i.e., the sum of the center frequencies of multiple modal function components is the minimum). Step S2-2 includes: S2-2-1: Construct an augmented Lagrange function with a quadratic penalty factor and Lagrange multipliers; S2-2-2: Using the augmented Lagrangian function as a constraint, the alternating direction multiplier method (ADMM) is used to optimize the modal function components and center frequency to obtain the optimal center frequency and the corresponding optimal modal function components.

[0029] Specifically, the secondary penalty factor α The number of modal decompositions, K, is set between 1500 and 3000, and the number of modal decompositions, K, is set between 4 and 8 based on the experimental noise complexity.

[0030] In this embodiment, when transforming the constrained variational model into an unconstrained optimization problem, an augmented Lagrangian function is constructed to introduce a quadratic penalty factor. α and Lagrange multipliers l ( t The augmented Lagrangian function combines the good convergence of the quadratic penalty factor with the strict enforcement of constraints by the Lagrange multiplier. The formula is as follows:

[0031] The first term on the right side of the equals sign in the formula contains a quadratic penalty factor. α This is used to ensure the accuracy of signal reconstruction in the presence of Gaussian noise. α The larger the value, the narrower the decomposed frequency band; the third term on the right side of the equals sign in the formula contains Lagrange multipliers. l ( t This is used to ensure the strictness of the constraints, that is, the sum of the modal function components approximates the elastic wave signal.

[0032] in, To augment the Lagrangian function, used to transform constrained optimization problems into unconstrained optimization problems; k The index number represents the modal function component (in this embodiment, the component of the intrinsic modal function), and takes the value 1, 2, ..., K; This indicates the pre-defined number of modal decompositions; Represents the decomposed The set of eigenmode function components; The decomposition yields the first... One eigenmode function component; This represents the set of center frequencies corresponding to each modal function component; Indicates the first The center frequencies corresponding to the modal function components; l , l ( t () All of these represent Lagrange multipliers, used to ensure that the reconstructed signal strictly approximates the elastic wave signal; This represents the quadratic penalty factor, also known as the data fidelity parameter. It is used to control the bandwidth of mode decomposition. The larger the value, the narrower the modal bandwidth. This represents the elastic wave signal acquired at the input terminal; This represents the inner product operator in mathematics.

[0033] Specifically, frequency domain iteration is subject to physical and mathematical limitations. In practical discrete signal processing, according to the Nyquist-Shannon sampling theorem, the effective frequency domain range of a signal is strictly limited to [specific range not specified in the original text]. Between, among The sampling frequency for acquiring acoustic vibration signals by the system is used. Therefore, the center frequency is updated in the frequency domain. In this case, the range of its iterative search is limited to the Nyquist frequency bound. Any frequency updates outside this range are considered invalid aliasing and are truncated, which guarantees the computational feasibility of the algorithm in real physical systems.

[0034] In this embodiment, the augmented Lagrangian function is used as the objective optimization function. The iterative process of ADMM involves fixing other variables in turn and optimizing the function individually. , and By taking the partial derivatives and setting them to zero, we obtain their update formulas in the frequency domain. Therefore, the entire iterative process is essentially about finding the augmented Lagrangian function. The ADMM algorithm minimizes the "saddle point" in each iteration through alternating optimization. To update and Then gradient ascent is performed to update .

[0035] The iterative formula for ADMM is as follows: using Fourier isochronous transform, each variable is updated sequentially in the frequency domain; where the... The specific formula for the next iteration is as follows: Modal function components Frequency domain update:

[0036] Center frequency Frequency domain update:

[0037] Lagrange multipliers Frequency domain update:

[0038] in, Represents the frequency variable in the frequency domain; Indicates the first Calculations; Indicates the first Calculations; Indicates the first After the nth iteration calculation, the nth The expression of each modal function component in the frequency domain; Represents elastic wave signal The expression in the frequency domain after Fourier transform; Indicates except Index labels of other modal function components besides; Indicates the current iteration state, except for the first... The sum of all modal function components in the frequency domain except for the one mode; Represents Lagrange multipliers The expression in the frequency domain after Fourier transform; Indicates the first The frequency domain expression of the Lagrange multipliers at the next iteration; Indicates the first After the iteration calculation, the updated result is... The new center frequency of each modal function component; Indicates the frequency range in non-negative frequencies Definite integral operations on; Indicates the first After the nth iteration Power spectral density of each modal function component; t This represents the update step size of the Lagrange multipliers, used to control the convergence and fidelity update rate of the algorithm under strong background noise.

[0039] Specifically, the convergence condition of ADMM is determined based on the relative squared error between the intrinsic mode functions calculated in two consecutive iterations. The convergence occurs when the relative changes of all mode function components in the frequency domain are less than a preset tolerance. The iteration stops when the time is reached. The formula is as follows:

[0040] in This represents the result of two consecutive iterations. Error energy of each modal function component in the frequency domain; Indicates the first During the nth iteration, the 1st The total energy of each modal function component itself; This represents the system's preset convergence tolerance, or decision threshold, which is typically set to 10. -5 Up to 10 -7 When the sum of relative errors is less than this minimum value When the algorithm has converged, the iteration stops.

[0041] When the sum of the relative errors of two consecutive iterations decreases to 10 -6 When the value reaches a certain order of magnitude, it indicates that the center frequency and waveform of each modal function component have reached a stable state. Using this minimum value as the judgment threshold can ensure the reconstruction accuracy and noise fidelity of high-frequency micro-motion impact signals while avoiding unnecessary redundant iterations caused by setting the threshold too small, thus balancing the decoupling accuracy and computational efficiency of the algorithm.

[0042] S2-3: Calculate the Pearson correlation coefficient between each optimal modal function component and the elastic wave signal; the calculation method is as follows:

[0043] in, This represents the correlation coefficient between the k-th intrinsic mode function (IMF) and the elastic wave signal; Represents the components of the optimal mode function Covariance between the elastic wave signal and the elastic wave signal; Represents the k-th optimal mode function component Standard deviation; This represents the standard deviation of the elastic wave signal.

[0044] S2-4: Eliminate the optimal mode function components whose Pearson correlation coefficient is less than the correlation coefficient threshold to remove noise information and obtain multiple remaining optimal mode function components; Correlation coefficient threshold The value is typically set between 0.2 and 0.3. When When this component is identified as low-frequency mechanical noise from the environment, it is removed. The remaining highly correlated optimal mode function components are then linearly superimposed to reconstruct a pure, noise-free high-frequency micro-motion impact signal. .

[0045] S2-5: The components of multiple residual optimal mode functions are linearly superimposed to obtain a high-frequency micro-motion impact signal.

[0046] High-fidelity extraction of high-frequency fretting impact signals under strong background noise was achieved through variational mode decomposition and Pearson correlation coefficient screening. The center frequencies of each component were optimized using the augmented Lagrangian function and the alternating direction multiplier method, ensuring signal reconstruction accuracy. By calculating the Pearson correlation coefficients between each optimal mode component and the original signal, spurious components related to environmental mechanical noise and fluid disturbances were accurately eliminated. The remaining highly correlated components were linearly superimposed to obtain a pure high-frequency fretting impact signal.

[0047] This method improves the sensitivity of identifying micro-impact characteristics of the contact surface, thereby enhancing the accuracy and robustness of fretting fatigue assessment of nuclear fuel grids.

[0048] S3: The system constructs the time-domain load spectrum of the nuclear fuel grid based on the high-frequency micro-motion impact signal and the normal clamping force, including: S3-1: Extract the voltage peak value based on the high-frequency micro-motion impact signal. U k ; S3-2: Calculate the maximum dynamic contact compressive stress amplitude between the nuclear fuel grid and fuel rods based on the peak voltage, including: S3-2-1: Calculate the equivalent dynamic impact force at the contact interface between the nuclear fuel grid and the fuel rods based on the voltage peak value; Based on the conversion relationship between acoustic emission energy and mechanical impact kinetic energy, a system calibration coefficient β (unit: N / V) is introduced to calculate the equivalent dynamic impact force at the contact interface. :

[0049] S3-2-2: Based on the elastic modulus and Poisson's ratio of the nuclear fuel grid and the elastic modulus and Poisson's ratio of the fuel rods, a nonlinear mapping relationship between the maximum dynamic contact compressive stress amplitude and the equivalent dynamic impact force is constructed. By introducing Hertzian contact mechanics, the elastic modulus of the fuel rod is obtained. E 1 and Poisson's ratio v 1. Elastic modulus of nuclear fuel grid E 2 and Poisson's ratio v 2. In order to calculate the equivalent elastic modulus E * and equivalent radius of curvature R * :

[0050]

[0051] in, Represented as the radius of curvature of the fuel rod; It represents the radius of curvature of the springs on the nuclear fuel grid that come into contact with the fuel rods.

[0052] According to Hertz's contact equation, the maximum dynamic contact compressive stress amplitude generated by this impact force in the contact area With impact force F There is a non-linear mapping relationship:

[0053] in, This is a geometric constant defined based on the actual topological morphology of the contact surface between the nuclear fuel lattice springs and the fuel rods. In actual fuel assemblies, the contact morphology between the positioning lattice springs and the cylindrical fuel rod cladding typically exhibits intersecting cylindrical contact or elliptical contact. This geometric constant is derived from the first and second kind of complete elliptic integrals in Hertzian contact theory. The value range is set to 0.3 to 0.8.

[0054] In this example, considering the spring morphology and the contact condition with the fuel rod, this geometric constant... It is set to 0.578.

[0055] S3-2-3: The maximum dynamic contact compressive stress amplitude is calculated using the nonlinear mapping relationship and the equivalent dynamic impact force.

[0056] S3-3: Constructing a time-domain load spectrum based on the normal clamping force and the maximum dynamic contact compressive stress amplitude, including: S3-3-1: Extract the slip transition time point between the nuclear fuel grid and fuel rods from adhesion to slippage based on high-frequency micro-motion impact signals, including: S3-3-1-1: Set a sliding window for the high-frequency micro-motion impact signal and calculate multiple real-time kurtosis; Assume the reconstructed high-frequency micro-motion impact signal is a discrete-time series. Set a sliding window of length N, and let the sliding window slide along the time axis with a step size S.

[0057] Within each sliding window, the kurtosis of the high-frequency micro-motion impact signal is calculated in real time. Kurtosis, as a fourth-order statistic, is extremely sensitive to high-frequency sudden shocks caused by early micro-motions. Its calculation formula is as follows:

[0058] in, Indicates the kurtosis value; This represents the total number of data points contained in the high-frequency micro-motion impact signal within the sliding window; This represents the specific amplitude of the i-th sampling point in the high-frequency micro-motion impact signal within the sliding window; This represents the average value of the high-frequency micro-motion impact signal within the sliding window; This indicates a summation.

[0059] S3-3-1-2: Calculate the average of the first set number of real-time kurtosis points among the multiple real-time kurtosis points. and standard deviation ; To ensure statistical reliability and avoid introducing subsequent early wear data, the first set number is set to a range of 50 to 200, and is set to 100 in this example.

[0060] S3-3-1-3: Set the kurtosis abrupt change threshold based on the real-time kurtosis mean and standard deviation; use Criteria for calculating kurtosis threshold : .

[0061] S3-3-1-4: Among the multiple real-time kurtosis values ​​that are greater than or equal to the kurtosis mutation threshold and are consecutively set to a second set number, the earliest corresponding time point is taken as the slip mutation time point.

[0062] The second set number is used for the algorithm's "anti-false triggering" determination. Considering the possibility of accidental electromagnetic interference or single random mechanical impacts in the test environment, the kurtosis must continuously exceed the threshold to confirm a true degradation of the micro-motion slip state. The value of this second set number is set to a range of 3 to 10, and is set to 5 in this example.

[0063] Specifically, when multiple consecutive real-time kurtosis values ​​are greater than or equal to the kurtosis abrupt change threshold, it is determined that the stiffness of the contact surface between the nuclear fuel grid and the fuel rods has degraded, and the state between them changes abruptly from adhesion to fretting slip. The physical timestamp at this point is recorded. t k The slip change time point is used as the reference point, and the number of peak values ​​exceeding the baseline within the slip window is counted as the impact frequency.

[0064] The baseline refers to the amplitude threshold used to identify the effective peak value of a micro-impact in a time-domain signal.

[0065] The specific calculation method for this baseline is as follows: High-frequency micro-motion impact signals are extracted when the nuclear fuel grid and fuel rods are in an initial fully adhered state (i.e., before the slip change occurs, within the aforementioned first set number of sliding windows), and the standard deviation of the initial signal amplitude in this segment is calculated; subsequently, based on... The criterion is to set the absolute value of the baseline to a set multiple of the standard deviation (usually 3 to 5 times, and in this example it is set to 3 times, that is, the baseline threshold is plus or minus three times the standard deviation).

[0066] S3-3-2: Based on the aforementioned slip change time point t k The normal clamping force is used to construct a time-domain load spectrum. When the time point is earlier than the slip change time point, the time-domain load spectrum is equal to the normal clamping force. When the time point is later than or equal to the slip change time point, the time-domain load spectrum is equal to the sum of the normal clamping force and the maximum dynamic contact compressive stress amplitude, or the difference between the normal clamping force and the maximum dynamic contact compressive stress amplitude.

[0067] Specifically, with t k Construct the time-domain load spectrum using the time boundary points. : when t < t k At time (adhesive state): ; when t ≥ t k Time (slipping state): .

[0068] in, The measured normal clamping force.

[0069] By fusing the denoised high-frequency micro-motion impact signal with the normal clamping force, a time-domain load spectrum that accurately reflects the evolution of the contact interface state is constructed. Dynamically correlating the micro-slip state switching with the macro-load spectrum overcomes the limitations of the traditional constant load assumption, providing a high-fidelity, time-varying input boundary for fatigue damage calculation, significantly improving the realism and sensitivity of the assessment.

[0070] S4: The system calculates the fatigue damage of the nuclear fuel grid based on the time-domain load spectrum, including: S4-1: Input the time-domain load spectrum into the digital twin model of the nuclear fuel grid and fuel rods to calculate the stress-life curve of the nuclear fuel grid; S4-2: Calculate the number of cycles for each stress amplitude based on the time-domain load spectrum; S4-3: The fatigue damage is calculated based on the number of cycles and the stress-life curve.

[0071] Specifically, a three-dimensional geometric model containing nuclear fuel grids and fuel rods is imported into the digital twin model, a high-precision mesh is generated, frictional contact pairs are set, and structural displacement constraints are applied.

[0072] The transient equivalent stress field of the springs on the nuclear fuel lattice at the point of sudden slip change is extracted using the finite element solver; by introducing the Basquin equation for the material, the stress-life curve is obtained:

[0073] in, This represents the maximum dynamic contact compressive stress amplitude. This represents the number of failure cycles.

[0074] 'm' represents the fatigue strength exponent of the material, which physically characterizes the sensitivity of the fatigue life of nuclear fuel lattice materials to changes in stress amplitude. Based on the properties of commonly used metallic materials in nuclear fuel assembly lattices, the empirically chosen value for this parameter 'm' is typically set between 3 and 15. In this example, the spring is made of a high-strength nickel-based alloy. Due to its extremely high yield strength and relatively flat high-cycle fatigue SN curve, the exponent 'm' is set between 8 and 14.

[0075] C is the fatigue strength constant of the material, which physically characterizes the basic fatigue resistance of the material under a specific stress state. Its value is highly correlated with the ultimate tensile strength of the material. When the stress amplitude... When the unit is MPa, the empirical range of values ​​for the constant C is usually set to 10. 10 Up to 10 25 between; In practical engineering applications, the precise values ​​of parameters m and C can be obtained by conducting standard high-frequency fatigue tensile and compressive tests on the same batch of nuclear fuel lattice materials, and by using the least squares method to perform double logarithmic linear fitting calibration on the experimental data.

[0076] Based on Miner's linear cumulative damage rule using measured impact frequencies, the current fatigue damage variables are calculated. :

[0077] in, This represents the actual number of cycles under a specific stress amplitude in the measured load spectrum. This represents the maximum number of cycles allowed under this stress.

[0078] S5: The system uses the fatigue damage to determine whether the nuclear fuel grid has failed.

[0079] Specifically, when D≥1, the nuclear fuel grid is determined to have experienced fatigue failure; when D<1, the nuclear fuel grid is determined not to have experienced fatigue failure.

[0080] Furthermore, it also includes: if the nuclear fuel grid has not failed, predicting the failure time point of the nuclear fuel grid using the time-domain load spectrum, including: The fatigue damage growth rate is calculated based on the time-domain load spectrum. Based on the growth rate, the remaining number of cycles or remaining time required for fatigue damage to reach the failure threshold is calculated using the load spectrum extrapolation method. The failure time point can be calculated using the remaining number of cycles or the remaining duration.

[0081] Specifically, the current time is calculated using Miner's linear cumulative damage rule. fatigue damage variables Then, the system will execute the following two parallel evaluation logics: I. Real-time status monitoring and cloud map output Regardless of damage variables Whether the value is greater than 1, the system will output a 3D evolution cloud map of fatigue damage at the contact point of the finite element model in real time (i.e., numerical simulation results). The 3D evolution cloud map displays the spatial distribution of fatigue damage on the entire spring structure through color gradients, which is used to intuitively locate the most likely dangerous nodes to fracture.

[0082] II. Failure Determination and Remaining Life Prediction Extracting fatigue damage variables from the most dangerous nodes in the 3D evolution cloud map : If the current is detected If the spring has failed due to fatigue, the system will issue a failure alarm. If the current is detected The system then determines that the spring has not yet failed. At this point, the system calculates the fatigue damage growth rate per unit time based on the measured historical load spectrum data. .

[0083] Based on the load spectrum extrapolation algorithm, assuming that the statistical characteristics of subsequent fretting loads remain stable, the calculation is performed to determine when the cumulative damage reaches the failure threshold (i.e., The remaining number of cycles or remaining time required is added to the current cycle number or current time point to obtain the predicted failure time point.

[0084] Load spectrum extrapolation algorithms include: (1) Rainflow counting: Perform rainflow counting on the currently constructed time-domain load spectrum to extract the stress amplitude and corresponding occurrence frequency of each cycle in the load spectrum; (2) Probability distribution fitting: The extracted stress amplitude and frequency data are fitted with the probability density using the Weibull distribution model to establish a load probability distribution function that reflects the current service status; (3) Lifetime Limit Extrapolation: Set the target lifetime cycle length and calculate the magnification factor between the target duration and the measured duration. Apply the magnification factor to the above probability distribution function and extrapolate the frequency to the high-stress tail region to predict the extreme high load stress amplitude and frequency that occur with a very low probability within the lifetime cycle; (4) Load spectrum reconstruction: The extrapolated cumulative frequency curve is reconstructed into a full-lifetime true micro-motion load spectrum containing the predicted extreme load through the Markov transfer matrix.

[0085] This embodiment directly compares the calculated fatigue damage variables with the failure threshold, realizing the quantitative discrimination of the fretting fatigue state of the nuclear fuel grid and establishing an online failure judgment mechanism based on the real dynamic load spectrum, which can reflect the cumulative damage degree of the contact interface in real time.

[0086] Meanwhile, the determination result provides a clear trigger condition for subsequent remaining lifetime prediction—an extrapolation warning is activated when D < 1, avoiding invalid calculations and false alarms, significantly improving the timeliness and accuracy of failure judgment, and providing a reliable decision-making basis for the structural integrity assessment of nuclear fuel assemblies.

[0087] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention is not limited to the specific combination of the above-described technical features, but also includes other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in this invention.

Claims

1. A method for assessing fretting fatigue failure of nuclear fuel grids, characterized in that, include: S1: Acquire the elastic wave signal of the nuclear fuel grid and the normal clamping force of the nuclear fuel grid on the fuel rods; S2: Remove noise information from the elastic wave signal to obtain a high-frequency micro-motion impact signal that contains only the vibration state of the nuclear fuel grid; S3: Construct the time-domain load spectrum of the nuclear fuel grid based on the high-frequency micro-motion impact signal and the normal clamping force; S4: Calculate the fatigue damage of the nuclear fuel grid based on the time-domain load spectrum; S5: Use the fatigue damage to determine whether the nuclear fuel grid has failed.

2. The method for assessing fretting fatigue failure of nuclear fuel grids according to claim 1, characterized in that, After removing noise from the elastic wave signal, a high-frequency micro-motion impact signal containing only the vibration state of the nuclear fuel grid is obtained, including: S2-1: Extract multiple modal function components with initial center frequencies based on the elastic wave signal; S2-2: Optimize the multiple modal function components to ensure reconstruction accuracy and obtain multiple first modal function components; S2-3: Calculate the Pearson correlation coefficient between each optimal modal function component and the elastic wave signal; S2-4: Eliminate the optimal mode function components whose Pearson correlation coefficient is less than the correlation coefficient threshold to remove noise information and obtain multiple remaining optimal mode function components; S2-5: The components of multiple residual optimal mode functions are linearly superimposed to obtain a high-frequency micro-motion impact signal.

3. The method for assessing fretting fatigue failure of a nuclear fuel grid according to claim 2, characterized in that, Optimize multiple modal function components to ensure reconstruction accuracy, resulting in multiple first modal function components, including: S2-2-1: Construct an augmented Lagrange function with a quadratic penalty factor and Lagrange multipliers; S2-2-2: Using the augmented Lagrangian function as a constraint, the modal function components and center frequency are optimized to obtain the optimal center frequency and the corresponding optimal modal function components.

4. The method for assessing fretting fatigue failure of nuclear fuel grids according to claim 1, characterized in that, The time-domain load spectrum of the nuclear fuel grid is constructed based on the high-frequency micro-motion impact signal and the normal clamping force, including: S3-1: Extract the voltage peak value based on the high-frequency micro-motion impact signal; S3-2: Calculate the maximum dynamic contact compressive stress amplitude between the nuclear fuel grid and the fuel rods based on the peak voltage; S3-3: Construct a time-domain load spectrum based on the normal clamping force and the maximum dynamic contact compressive stress amplitude.

5. The method for assessing fretting fatigue failure of a nuclear fuel grid according to claim 4, characterized in that, The maximum dynamic contact compressive stress amplitude between the nuclear fuel grid and fuel rods is calculated based on the peak voltage, including: S3-2-1: Calculate the equivalent dynamic impact force at the contact interface between the nuclear fuel grid and the fuel rods based on the voltage peak value; S3-2-2: Based on the elastic modulus and Poisson's ratio of the nuclear fuel grid and the elastic modulus and Poisson's ratio of the fuel rods, a nonlinear mapping relationship between the maximum dynamic contact compressive stress amplitude and the equivalent dynamic impact force is constructed. S3-2-3: The maximum dynamic contact compressive stress amplitude is calculated using the nonlinear mapping relationship and the equivalent dynamic impact force.

6. The method for assessing fretting fatigue failure of a nuclear fuel grid according to claim 4, characterized in that, A time-domain load spectrum is constructed based on the normal clamping force and the maximum dynamic contact compressive stress amplitude, including: S3-3-1: Extract the slip transition time point between the nuclear fuel grid and fuel rods from adhesion to slip based on high-frequency micro-motion impact signals; S3-3-2: Construct a time-domain load spectrum based on the slip change time point and the normal clamping force, such that when the time point is earlier than the slip change time point, the time-domain load spectrum is equal to the normal clamping force; when the time point is later than or equal to the slip change time point, the time-domain load spectrum is equal to the sum of the normal clamping force and the maximum dynamic contact compressive stress amplitude, or the difference between the normal clamping force and the maximum dynamic contact compressive stress amplitude.

7. The method for assessing fretting fatigue failure of a nuclear fuel grid according to claim 6, characterized in that, Extract slip change time points based on high-frequency micro-motion impact signals, including: S3-3-1-1: Set a sliding window for the high-frequency micro-motion impact signal and calculate multiple real-time kurtosis; S3-3-1-2: Calculate the mean and standard deviation of the real-time kurtosis of the first set number of the earliest time points among the multiple real-time kurtosis; S3-3-1-3: Set the kurtosis abrupt change threshold based on the real-time kurtosis mean and standard deviation; S3-3-1-4: Among the multiple real-time kurtosis values ​​that are greater than or equal to the kurtosis mutation threshold and are consecutively set to a second set number, the earliest corresponding time point is taken as the slip mutation time point.

8. The method for assessing fretting fatigue failure of a nuclear fuel grid according to claim 1, characterized in that, The fatigue damage of the nuclear fuel lattice is calculated based on the time-domain load spectrum, including: S4-1: Input the time-domain load spectrum into the digital twin model of the nuclear fuel grid and fuel rods to calculate the stress-life curve of the nuclear fuel grid; S4-2: Calculate the number of cycles for each stress amplitude based on the time-domain load spectrum; S4-3: The fatigue damage is calculated based on the number of cycles and the stress-life curve.

9. The method for assessing fretting fatigue failure of a nuclear fuel grid according to claim 1, characterized in that, Also includes: If the nuclear fuel grid has not failed, the failure time point of the nuclear fuel grid can be predicted using the time-domain load spectrum.

10. The method for assessing fretting fatigue failure of a nuclear fuel grid according to claim 9, characterized in that, Predicting the failure time point of the nuclear fuel grid using the aforementioned time-domain load spectrum includes: The fatigue damage growth rate is calculated based on the time-domain load spectrum. Based on the growth rate, the remaining number of cycles or the remaining time required for fatigue damage to reach the failure threshold is calculated using the load spectrum extrapolation method. The failure time point can be calculated using the remaining number of cycles or the remaining duration.

Images