Method for evaluating service state of nuclear fuel assembly

Abstract

The invention provides a nuclear fuel assembly service state evaluation method, and relates to the technical field of fuel rod abrasion testing, and the method comprises the steps: obtaining the amplitudes of a plurality of measurement positions of a to-be-monitored fuel rod in real time, and generating a measurement amplitude vector of the to-be-monitored fuel rod corresponding to each moment based on the amplitudes of the plurality of measurement positions; inputting the measurement amplitude vector of the to-be-monitored fuel rod corresponding to each moment into a preset amplitude transfer relation model corresponding to the to-be-monitored clamping point to obtain the equivalent amplitude of the to-be-monitored clamping point corresponding to each moment; and inputting the equivalent amplitude of the to-be-monitored clamping point corresponding to each moment into a preset wear fatigue mapping model to obtain a clamping spring equivalent fatigue parameter and a fuel rod equivalent wear depth of the to-be-monitored clamping point corresponding to each moment. The dynamic trend of the wear depth of the fuel rod and the fatigue of the spring can be obtained in real time under the vibration working condition, and the defect that only the final states of the fuel rod and the spring can be obtained through traditional off-line detection is overcome.

Description

A method for assessing the service status of nuclear fuel assemblies Technical Field

[0001] This invention relates to the field of fuel rod erosion testing, and more particularly to a method for assessing the service condition of nuclear fuel assemblies. Background Technology

[0002] Nuclear fuel assemblies are the core components of a nuclear reactor, and their integrity directly affects the safe operation of the reactor. The positioning grid, through its clamping springs and rigid protrusions, elastically holds the fuel rods to suppress fuel rod vibrations caused by coolant flow, preventing wear and damage to the fuel rod cladding due to excessive vibration amplitude. During long-term reactor operation, continuous fretting friction between the fuel rods and the positioning grid leads to wear and material fatigue of the clamping springs, resulting in a decrease in the clamping force. This alters the dynamic characteristics of the fuel rod system, exacerbating vibration and wear, creating a vicious cycle that ultimately threatens assembly safety. Therefore, assessing the wear depth of the fuel rods and the fatigue state of the clamping springs is a key technical challenge for predicting fuel assembly life and ensuring reactor safety. Because the positioning grid clamping area is small and complex, it is impossible to place sensors without damaging the original clamping state and dynamic characteristics. As a result, it is difficult to directly obtain the vibration data of the fuel rods in this area. Moreover, existing technologies lack a reverse path that can effectively correlate vibration data with wear and fatigue conditions. Therefore, existing technologies mostly rely on offline disassembly to detect the final wear condition of the fuel rods and the fatigue condition of the springs. It is difficult to assess the changing trends of the wear condition of the fuel rods and the spring condition under vibration conditions, and thus cannot reflect its true service evolution process.

[0003] Therefore, there is an urgent need for a method that can assess the wear depth of fuel rods and the fatigue state of clamping springs under vibration conditions without disassembling nuclear fuel assemblies. Summary of the Invention

[0004] The purpose of this invention is to address the above problems by providing a method for assessing the service status of nuclear fuel assemblies.

[0005] This invention provides a method for assessing the service status of nuclear fuel assemblies, comprising: S1: acquiring the amplitude of multiple measurement positions of the fuel rod to be monitored in real time, and generating a measurement amplitude vector of the fuel rod to be monitored at each moment based on the amplitude of the multiple measurement positions; S2: inputting the measurement amplitude vector of the fuel rod to be monitored at each moment into a preset amplitude transfer relationship model corresponding to the clamping point to be monitored, to obtain the equivalent amplitude of the clamping point to be monitored at each moment; the preset amplitude transfer relationship model uses the measurement amplitude vector of the sample fuel rod as the model input and the actual amplitude of the sample clamping point as the model output. The sample clamping point is obtained by training a neural network model; the sample clamping point is the clamping point on the sample fuel rod corresponding to the clamping point to be monitored; S3: the equivalent amplitude of the clamping point to be monitored at each time moment is input into the preset wear fatigue mapping model to obtain the equivalent fatigue parameters of the clamping spring and the equivalent wear depth of the fuel rod at each time moment; the preset wear fatigue mapping model is obtained by training a neural network model with the equivalent amplitude of the sample clamping point as the model input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point as the model output.

[0006] According to certain embodiments of the present invention, the method further includes: acquiring the amplitude of multiple measurement positions of the sample fuel rod; generating a measurement amplitude vector of the sample fuel rod based on the amplitude of the multiple measurement positions; acquiring the actual amplitude of the sample clamping point; the actual amplitude of the sample clamping point is obtained by collecting the measurement value of a sensor installed at the installation position corresponding to the sample clamping point on the positioning grid; using the measurement amplitude vector of the sample fuel rod as the model input and the actual amplitude of the sample clamping point as the model output, performing neural network model training to obtain the preset amplitude transmission relationship model.

[0007] According to certain embodiments of the present invention, the method of training a neural network model to obtain the preset amplitude transfer relationship model by using the measured amplitude vector of the sample fuel rod as the model input and the actual amplitude of the sample clamping point as the model output includes: dividing the measured amplitude vector of the sample fuel rod into a training set and a test set; using the measured amplitude vector in the training set as the input and the actual amplitude of the sample clamping point as the output to train a neural network model to obtain a first intermediate model; inputting the measured amplitude vector in the test set into the first intermediate model to obtain the predicted amplitude of the sample clamping point; calculating the amplitude error between the predicted amplitude and the actual amplitude of the sample clamping point; and adjusting the network weight parameters of the first intermediate model based on the amplitude error until the amplitude error is less than or equal to a preset amplitude error threshold to obtain the preset amplitude transfer relationship model.

[0008] According to certain embodiments of the present invention, the method further includes: obtaining the equivalent amplitude of the sample clamping point; the equivalent amplitude of the sample clamping point is obtained by inputting the measured amplitude vector of the sample fuel rod into a preset amplitude transfer relationship model corresponding to the sample clamping point; obtaining the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point; the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point are obtained by offline calibration of the clamping spring and the sample fuel rod after each experiment; using the equivalent amplitude of the sample clamping point as the model input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod as the model output, performing neural network model training to obtain the preset wear fatigue mapping model.

[0009] According to certain embodiments of the present invention, the method of training a neural network model to obtain the preset wear fatigue mapping model includes: dividing the equivalent amplitude of the sample clamping point into a training set and a test set; using the equivalent amplitude in the training set as input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod as output, and obtaining a second intermediate model; inputting the equivalent amplitude in the test set into the second intermediate model to obtain the predicted fatigue parameters of the clamping spring and the predicted wear depth of the fuel rod at the sample clamping point; calculating the fatigue parameter error between the predicted fatigue parameters of the clamping spring and the actual fatigue parameters of the clamping spring, and the depth error between the predicted wear depth of the fuel rod and the actual wear depth of the fuel rod; adjusting the network weight parameters of the second intermediate model based on the fatigue parameter error and the depth error until the fatigue parameter error is less than or equal to a preset fatigue parameter error threshold and the depth error is less than or equal to a preset depth error threshold, thereby obtaining the preset wear fatigue mapping model.

[0010] According to the technical solutions provided by certain embodiments of the present invention, the method further includes: machining mounting holes at mounting positions corresponding to sample clamping points on the positioning grid of the nuclear fuel assembly fretting wear accelerated reliability test device; embedding a sensor into the mounting holes, such that the detection surface of the sensor probe and the sample clamping point of the sample fuel rod meet a preset distance; and acquiring the amplitude of the sample clamping point of the sample fuel rod through the sensor to obtain the actual amplitude of the sample clamping point.

[0011] According to the technical solutions provided by certain embodiments of the present invention, the method further includes: comparing the equivalent wear depth of the fuel rod at the clamping point to be monitored at each time moment with a safety threshold range of different levels; if the equivalent wear depth of the fuel rod at the clamping point to be monitored falls into the safety threshold range of the corresponding level, then outputting the prompt information corresponding to the safety threshold range of that level.

[0012] According to the technical solutions provided by certain embodiments of the present invention, the method further includes: calculating the absolute rate of change of the equivalent wear depth of the fuel rod at the clamping point to be monitored within a preset time; comparing the absolute rate of change with a preset rate of change threshold; and issuing a corresponding warning message if the absolute rate of change is greater than the preset rate of change threshold.

[0013] According to the technical solutions provided by certain embodiments of the present invention, the method further includes: obtaining the initial wall thickness of the clamping point to be monitored; calculating the wear ratio of the clamping point to be monitored at each moment based on the initial wall thickness of the clamping point to be monitored and the equivalent wear depth of the fuel rod of the clamping point to be monitored at each moment; comparing the wear ratio with a preset ratio threshold, and if the wear ratio is greater than or equal to the preset ratio threshold, outputting corresponding warning information.

[0014] According to the technical solutions provided by certain embodiments of the present invention, the method further includes: after the experiment, performing offline calibration on the fuel rod to be monitored and the clamping spring to obtain the actual wear depth of the clamping point to be monitored and the actual fatigue parameters of the clamping spring; comparing the equivalent wear depth of the fuel rod at the clamping point to be monitored with the actual wear depth of the clamping point to be monitored at the last moment, and comparing the equivalent fatigue parameters of the clamping spring at the clamping point to be monitored with the actual fatigue parameters of the clamping spring at the clamping point to be monitored at the last moment, and correcting the preset amplitude transmission relationship model, the preset wear fatigue mapping model, the safety threshold range, the preset rate of change threshold, and the preset ratio threshold according to the comparison results.

[0015] In summary, this invention provides a method for assessing the service status of nuclear fuel assemblies, comprising: S1: acquiring the amplitude of multiple measurement locations of the fuel rod to be monitored in real time, and generating a measurement amplitude vector of the fuel rod to be monitored at each moment based on the amplitude of the multiple measurement locations; S2: inputting the measurement amplitude vector of the fuel rod to be monitored at each moment into a preset amplitude transfer relationship model corresponding to the clamping point to be monitored, to obtain the equivalent amplitude of the clamping point to be monitored at each moment; the preset amplitude transfer relationship model uses the measurement amplitude vector of the sample fuel rod as the model input and the actual amplitude of the sample clamping point as the model. The model output is obtained through neural network model training; the sample clamping point is the clamping point on the sample fuel rod corresponding to the clamping point to be monitored; S3: the equivalent amplitude of the clamping point to be monitored at each moment is input into the preset wear fatigue mapping model to obtain the equivalent fatigue parameters of the clamping spring and the equivalent wear depth of the fuel rod at each moment; the preset wear fatigue mapping model is obtained by training a neural network model with the equivalent amplitude of the sample clamping point as the model input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod as the model output. Compared with the prior art, the present invention realizes the reverse calculation from the amplitude of the easily measurable position of the fuel rod to the equivalent amplitude of the clamping point, and then to the wear depth and fatigue state by constructing a preset amplitude transmission relationship model and a preset wear fatigue mapping model. It can complete the monitoring of the wear depth of the fuel rod and the fatigue state of the clamping spring under vibration conditions without disassembling the nuclear fuel assembly, effectively avoiding the defect that traditional offline disassembly and detection cannot reflect the real service evolution process.

[0016] It should be understood that the descriptions of technical features, technical solutions, beneficial effects, or similar language in this invention do not imply that all features and advantages can be achieved in any single embodiment. Rather, it is understood that the description of a feature or beneficial effect means that a specific technical feature, technical solution, or beneficial effect is included in at least one embodiment. Therefore, the descriptions of technical features, technical solutions, or beneficial effects in this specification do not necessarily refer to the same embodiment. Furthermore, the technical features, technical solutions, and beneficial effects described in this embodiment can be combined in any suitable manner. Those skilled in the art will understand that embodiments can be implemented without one or more specific technical features, technical solutions, or beneficial effects of a particular embodiment. In other embodiments, additional technical features and beneficial effects may be identified in specific embodiments that do not embody all embodiments. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 is a flowchart illustrating a method for assessing the service status of nuclear fuel assemblies according to an embodiment of the present invention; Figure 2 is a structural schematic diagram illustrating a reliability test device for accelerating fretting wear of nuclear fuel assemblies according to an embodiment of the present invention.

[0019] The text labels in the figure represent: 1. Sample fuel rod; 2. Positioning grid. Detailed Implementation

[0020] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. This description is merely illustrative and explanatory, and should not be construed as limiting the scope of protection of the present invention in any way. Specifically, the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort should fall within the scope of protection of the present invention.

[0021] It should be noted that similar reference numerals and letters in the following figures denote similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or apparatus.

[0022] Nuclear fuel assemblies are core components of nuclear reactors, and their integrity directly affects the safe operation of the reactor. During long-term reactor operation, the continuous fretting friction between the fuel rods and the positioning grid causes wear and fatigue of the clamping springs, resulting in a decrease in their clamping force. This, in turn, alters the dynamic characteristics of the fuel rod system, exacerbating a vicious cycle of vibration and wear, and threatening assembly safety. Therefore, assessing the wear depth of the fuel rods and the fatigue state of the clamping springs is a key challenge in predicting assembly life and ensuring reactor safety.

[0023] In actual testing, to avoid damaging the original clamping state and dynamic characteristics of the testing device, sensors cannot be placed at the fuel rod clamping position, making it difficult to directly obtain vibration data in this area. Furthermore, existing technologies lack an effective correlation path between vibration data and wear and fatigue conditions. Therefore, the detection of the state of fuel rods and clamping springs largely relies on offline disassembly and calibration, which can only obtain the final state of the two and cannot track the changing trends under vibration conditions, making it difficult to reflect the actual service evolution process.

[0024] It should be clarified that this method only aims to capture the aforementioned dynamic change trends through the model. The final state of the fuel rod and clamping spring is still based on the offline disassembly and inspection results, which will not affect the accuracy of the final state detection of the two, ensuring the reliability of the core results. During the model training phase, although the actual amplitude is collected by adding sensors at the sample clamping position, it will cause slight damage to the local clamping mechanical properties of the sample. However, this effect can be gradually mitigated through sufficient sample data training and iterative optimization of model parameters, thereby establishing the correlation trend between the clamping point amplitude and wear and fatigue state.

[0025] As mentioned above, in response to the problems in the prior art, this embodiment provides a method for assessing the service status of nuclear fuel assemblies, as shown in Figures 1 and 2, including: S1: Real-time acquisition of the amplitude of multiple measurement positions of the fuel rod to be monitored, and generation of the measurement amplitude vector of the fuel rod to be monitored at each moment based on the amplitude of the multiple measurement positions; Specifically, the selection of multiple measurement positions needs to be combined with the structural characteristics and vibration transmission law of the fuel rod, and select areas that are easy to arrange sensors and can effectively reflect the overall vibration response of the fuel rod to be monitored (such as the free section, end, etc. of the fuel rod to be monitored). The amplitude data of these positions are collected in real time by sensors, and the measurement amplitude vector of the fuel rod to be monitored at each moment is generated based on the amplitude of different measurement positions at the same moment, laying the foundation for subsequent data input. The relationship between the measurement amplitude vector and the amplitude of different measurement positions at the same moment can be expressed by formula (1): Formula (1) where, To measure the amplitude vector; This represents the amplitude at different measurement locations at the same moment.

[0026] S2: Input the measured amplitude vector of the fuel rod to be monitored at each moment into the preset amplitude transfer relationship model corresponding to the clamping point to be monitored, and obtain the equivalent amplitude of the clamping point to be monitored at each moment; the preset amplitude transfer relationship model is obtained by training a neural network model with the measured amplitude vector of the sample fuel rod 1 as the model input and the actual amplitude of the sample clamping point as the model output; the sample clamping point is the clamping point on the sample fuel rod 1 corresponding to the clamping point to be monitored; specifically, input the measured amplitude vector of the fuel rod to be monitored at each moment into the preset amplitude transfer relationship model corresponding to the clamping point to be monitored, and obtain the equivalent amplitude of the clamping point to be monitored at each moment (since the vibration is more obvious when the fuel rod is closer to the vibration source, the preset amplitude transfer relationship model only needs to output the equivalent amplitude of the uppermost clamping point to meet the requirements), realize the conversion of the easily measurable position amplitude of the fuel rod to be monitored into the amplitude of the unmeasurable clamping point, and the process is shown in the following formula (2): Formula (2) where, The equivalent amplitude of the clamping point to be monitored; To measure the amplitude vector; The amplitude at different measurement locations at the same moment; This is a preset amplitude transfer function.

[0027] The preset amplitude transmission relationship model is obtained by training a BP neural network model using the measured amplitude vector of sample fuel rod 1 as the model input and the actual amplitude of the sample clamping point as the model output. The sample clamping point is the clamping point on sample fuel rod 1 corresponding to the clamping point to be monitored. Its position, structural constraints, and assembly relationship with the positioning grid 2 are completely consistent with the clamping point to be monitored. This is used to control the variables of the BP neural network, eliminate interference from non-essential factors such as different clamping positions on the vibration transmission law, and ensure that the model training focuses on the transmission relationship of the fuel rod amplitude between the easily measurable position and the clamping point.

[0028] S3: Input the equivalent amplitude of the clamping point to be monitored at each moment into the preset wear fatigue mapping model to obtain the equivalent fatigue parameters of the clamping spring and the equivalent wear depth of the fuel rod at each moment. The preset wear fatigue mapping model is obtained by training a neural network model with the equivalent amplitude of the sample clamping point as the model input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point as the model output.

[0029] Specifically, the equivalent amplitude of the clamping point to be monitored at each moment is input into the preset wear fatigue mapping model to obtain the equivalent fatigue parameters of the clamping spring and the equivalent wear depth of the fuel rod at each moment. The equivalent fatigue parameters of the clamping spring are used to characterize the fatigue state of the clamping spring, and the equivalent wear depth of the fuel rod is used to characterize the wear depth of the fuel rod. This process is shown in the following formula (3): Formula (3) where, The equivalent fatigue parameters of the clamping spring at the clamping point to be monitored; The equivalent wear depth of the fuel rod at the clamping point to be monitored; Preset wear fatigue mapping function; The equivalent amplitude of the clamping point to be monitored.

[0030] The preset wear fatigue mapping model is obtained by training a BP neural network model with the equivalent amplitude of the sample clamping point as the model input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point as the model output.

[0031] This invention, by constructing a preset amplitude transmission relationship model and a preset wear fatigue mapping model, realizes the reverse calculation from the amplitude of the fuel rod at an easily measurable location to the equivalent amplitude at the clamping point, and then to the wear depth and fatigue state. It can monitor and evaluate the changing trends of fuel rod wear depth and clamping spring fatigue state under vibration conditions without disassembling the nuclear fuel assembly. This solves the defect of traditional offline disassembly and testing that cannot reflect the actual service evolution process, and avoids the problem that traditional offline disassembly and testing can only obtain the final wear condition of the fuel rod and the fatigue state of the spring, but cannot obtain the changing trends of fuel rod wear depth and spring state under vibration conditions.

[0032] In a preferred embodiment, the method further includes: acquiring the amplitude of multiple measurement positions of the sample fuel rod 1; generating a measurement amplitude vector of the sample fuel rod 1 based on the amplitude of the multiple measurement positions; acquiring the actual amplitude of the sample clamping point; the actual amplitude of the sample clamping point is obtained by collecting the measurement value of the sensor installed at the installation position corresponding to the sample clamping point on the positioning grid 2; using the measurement amplitude vector of the sample fuel rod 1 as the model input and the actual amplitude of the sample clamping point as the model output, performing neural network model training to obtain a preset amplitude transmission relationship model.

[0033] Specifically, when acquiring the amplitudes at multiple measurement locations of the sample fuel rod and generating the measurement amplitude vector of sample fuel rod 1 based on the amplitudes at these multiple measurement locations, the multiple measurement locations of the sample fuel rod are exactly the same as the multiple measurement locations of the fuel rod to be monitored. This ensures that the preset amplitude transfer relationship model trained based on the amplitudes at multiple measurement locations of the sample fuel rod can reflect the common laws of fuel rod vibration transfer, eliminate errors caused by differences in the distribution of measurement locations, and guarantee that the mapping relationship obtained based on the sample fuel rod can be directly transferred to the fuel rod to be monitored. The measured values ​​of the sensors installed at the installation positions corresponding to the sample clamping points on the positioning grid 2 are collected as the actual amplitude of the sample clamping points. It should be clarified here that the installation of sensors is limited to the training and calibration stage of the preset amplitude transfer relationship model. In actual experiments, to avoid any impact on the original clamping state of the positioning grid 2 and the dynamic characteristics of the fuel rod, sensors should not be installed on the positioning grid 2. Based on the preset amplitude transfer relationship model obtained from the above steps and the actual amplitude of the sample clamping point, the measured amplitude vector of sample fuel rod 1 is used as the model input, and the actual amplitude of the sample clamping point is used as the model output. The preset amplitude transfer relationship model can be obtained by using a BP neural network to train the model.

[0034] In a preferred embodiment, training a neural network model using the measured amplitude vector of the sample fuel rod 1 as model input and the actual amplitude of the sample clamping point as model output to obtain a preset amplitude transfer relationship model includes: dividing the measured amplitude vector of the sample fuel rod 1 into a training set and a test set; using the measured amplitude vector in the training set as input and the actual amplitude of the sample clamping point as output to train a neural network model to obtain a first intermediate model; inputting the measured amplitude vector in the test set into the first intermediate model to obtain the predicted amplitude of the sample clamping point; calculating the amplitude error between the predicted amplitude and the actual amplitude of the sample clamping point; and adjusting the network weight parameters of the first intermediate model based on the amplitude error until the amplitude error is less than or equal to a preset amplitude error threshold to obtain the preset amplitude transfer relationship model.

[0035] Specifically, the weight parameters of the BP neural network are randomly initialized. The measured amplitude vectors from the training set are used as input, and the actual amplitude of the sample clamping point is used as output to train the neural network model and obtain a first intermediate model. The measured amplitude vectors from the test set are input into the first intermediate model, and forward propagation is completed through hidden layer transformation, outputting the predicted amplitude value of the clamping point. The amplitude error between the predicted amplitude value of the clamping point and the actual amplitude of the sample clamping point is calculated, and the weight parameters of the first intermediate model are adjusted backward using an error backpropagation algorithm until the amplitude error is less than or equal to a preset amplitude error threshold, thus obtaining a preset amplitude transfer relationship model. By using the training set to dominate the model training process and the test set to dominate the model validation process, the accuracy of the preset amplitude transfer relationship model is improved.

[0036] In a preferred embodiment, the method further includes: obtaining the equivalent amplitude of the sample clamping point; the equivalent amplitude of the sample clamping point is obtained by inputting the measured amplitude vector of the sample fuel rod 1 into the preset amplitude transfer relationship model corresponding to the sample clamping point; obtaining the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point; the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point are obtained by offline calibration of the clamping spring and the sample fuel rod after each experiment; using the equivalent amplitude of the sample clamping point as the model input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod as the model output, performing neural network model training to obtain a preset wear fatigue mapping model.

[0037] Specifically, the measured amplitude vector of sample fuel rod 1 is input into the preset amplitude transfer relationship model corresponding to the sample clamping point to obtain the equivalent amplitude of the sample clamping point. The purpose is to completely unify the acquisition path and physical meaning with the equivalent amplitude of the clamping point to be monitored. This ensures that the preset wear fatigue mapping model trained based on the equivalent amplitude of the sample clamping point can be directly transferred and applied to the fuel rod to be monitored, realizing the inverse calculation of the equivalent fatigue parameters of the clamping spring and the equivalent wear depth of the fuel rod. After each experiment, the clamping spring and sample fuel rod are calibrated offline to obtain the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point. Offline calibration is the process of removing the clamping spring and sample fuel rod from the experimental device and directly measuring them using professional testing equipment without changing their fatigue and wear state after the experiment. For the actual fatigue parameters of the clamping spring, a controllable load is applied using a stiffness testing device, and the force-displacement response curve of the spring is recorded to calculate the equivalent stiffness reflecting its fatigue degradation degree. For the actual wear depth of the fuel rod, a surface profilometer or microscopic measuring device is used to scan and detect the sample clamping point of sample fuel rod 1 to obtain the actual wear depth of the fuel rod. Based on the equivalent amplitude of the sample clamping point, the actual fatigue parameters of the clamping spring at the sample clamping point, and the actual wear depth of the fuel rod obtained from the above steps, the equivalent amplitude of the sample clamping point is used as the model input, and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod are used as the model output to train the BP neural network model to obtain the preset wear fatigue mapping model.

[0038] In a preferred embodiment, training a neural network model to obtain a preset wear fatigue mapping model, using the equivalent amplitude of the sample clamping point as model input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod as model output, includes: dividing the equivalent amplitude of the sample clamping point into a training set and a test set; using the equivalent amplitude in the training set as input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod as output, training a neural network model to obtain a second intermediate model; inputting the equivalent amplitude in the test set into the second intermediate model to obtain the predicted fatigue parameters of the clamping spring and the predicted wear depth of the fuel rod at the sample clamping point; calculating the fatigue parameter error between the predicted fatigue parameters of the clamping spring and the actual fatigue parameters of the clamping spring, and the depth error between the predicted wear depth of the fuel rod and the actual wear depth of the fuel rod; adjusting the network weight parameters of the second intermediate model based on the fatigue parameter error and the depth error until the fatigue parameter error is less than or equal to a preset fatigue parameter error threshold and the depth error is less than or equal to a preset depth error threshold, thereby obtaining the preset wear fatigue mapping model.

[0039] Specifically, the weight parameters of the BP neural network are randomly initialized. The equivalent amplitude of the sample clamping point is divided into independent training and test sets. The equivalent amplitude in the training set is used as input, and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point are used as output to train a second intermediate model. The equivalent amplitude of the test set is input into the second intermediate model to obtain the predicted fatigue parameters of the clamping spring and the predicted wear depth of the fuel rod at the sample clamping point. The fatigue parameter error and depth error between the predicted and actual values ​​are calculated respectively. Based on these two errors, the model network weight parameters are adjusted in reverse. The process is iterated until both types of errors meet the corresponding preset threshold requirements. The final preset wear fatigue mapping model is obtained. The accuracy of the preset wear fatigue mapping model is ensured through the verification process.

[0040] In a preferred embodiment, the method further includes: machining mounting holes at mounting positions corresponding to the sample clamping points on the positioning grid 2 of the nuclear fuel assembly fretting wear accelerated reliability test device; embedding the sensor into the mounting holes, and ensuring that the detection surface of the sensor probe and the sample clamping point of the sample fuel rod 1 meet a preset distance; and collecting the amplitude of the sample clamping point of the sample fuel rod 1 through the sensor to obtain the actual amplitude of the sample clamping point.

[0041] Specifically, mounting holes are machined on the positioning grid 2 of the nuclear fuel assembly fretting wear accelerated reliability test device at the mounting positions corresponding to the sample clamping points. A sensor is embedded in the mounting hole, ensuring a preset distance (typically 0.8-2.8 mm) between the sensor probe's detection surface and the sample clamping point of the sample fuel rod 1. The sensor collects the amplitude of the sample clamping point of the sample fuel rod 1 to obtain the actual amplitude of the sample clamping point. The sensor is an eddy current sensor, but laser displacement sensors, miniature capacitive displacement sensors, or other vibration detection devices with non-contact measurement capabilities, sensitive signal response, and size adaptable to the mounting hole can also be used; no specific limitation is made here. It should be noted that during actual monitoring, to avoid any impact on the original clamping state of the positioning grid 2 and the dynamic characteristics of the fuel rod, sensors should not be installed on the positioning grid 2. However, drilling holes in the positioning grid 2 to install sensors during model training is to calibrate the actual amplitude of the sample clamping point, thereby training and generating a preset amplitude transfer relationship model and a preset wear fatigue mapping model based on the actual amplitude of the sample clamping point.

[0042] In a preferred embodiment, the method further includes: comparing the equivalent wear depth of the fuel rod at the clamping point to be monitored at each moment with the safety threshold range of different levels; if the equivalent wear depth of the fuel rod at the clamping point to be monitored falls into the safety threshold range of the corresponding level, then outputting the prompt information corresponding to the safety threshold range of that level.

[0043] Specifically, safety threshold ranges are divided into four levels based on the depth of wear, with each level increasing sequentially and without overlap. The equivalent wear depth of the fuel rod at each monitored clamping point is compared to the safety threshold range of each level. If the equivalent wear depth falls within the first level's safety threshold range, a normal warning is issued; if it falls within the second level's, a slight wear warning is issued; if it falls within the third level's, a moderate wear warning is issued; and if it falls within the fourth level's, a severe wear warning is issued. This indicates that the monitored fuel rod is at risk of failure, its remaining wall thickness is insufficient to guarantee the structural integrity of the reactor for long-term operation, and continued service could easily lead to insufficient strength and rupture due to excessive cladding thinning, resulting in serious safety accidents such as pellet leakage and fuel rod fallout.

[0044] In a preferred embodiment, the method further includes: calculating the absolute rate of change of the equivalent wear depth of the fuel rod at the clamping point to be monitored within a preset time; comparing the absolute rate of change with a preset rate of change threshold; and issuing a corresponding warning message if the absolute rate of change is greater than the preset rate of change threshold.

[0045] Specifically, the absolute rate of change of the equivalent wear depth of the fuel rod at the clamping point to be monitored within a preset time is calculated. The absolute rate of change is calculated according to the following formula (4): Formula (4) where, The absolute rate of change of the equivalent wear depth of the fuel rod at the clamping point to be monitored within a preset time period; The equivalent wear depth of the fuel rod at the start of a preset time period; The equivalent wear depth of the fuel rod at the end of the preset time period; This is a preset time period.

[0046] The obtained absolute rate of change is compared with the preset rate of change threshold. If the absolute rate of change is greater than the preset rate of change threshold, it indicates that the wear of the fuel rods is accelerating and there is a certain safety risk in the future. In this case, the corresponding early warning information should be issued in a timely manner.

[0047] In a preferred embodiment, the method further includes: obtaining the initial wall thickness of the clamping point to be monitored; calculating the wear ratio of the clamping point to be monitored at each moment based on the initial wall thickness of the clamping point to be monitored and the equivalent wear depth of the fuel rod at the clamping point to be monitored at each moment; comparing the wear ratio with a preset ratio threshold, and if the wear ratio is greater than or equal to the preset ratio threshold, outputting the corresponding warning information.

[0048] Specifically, the initial wall thickness of the clamping point to be monitored is obtained; based on the initial wall thickness of the clamping point to be monitored and the equivalent wear depth of the fuel rod at each moment, the wear ratio of the clamping point to be monitored at each moment is calculated, and the wear ratio is calculated according to the following formula (5): Formula (5) where, This represents the wear ratio; The initial wall thickness of the clamping point to be monitored; The equivalent wear depth of the fuel rod at the clamping point to be monitored.

[0049] For fuel rods of the same type under monitoring, if the calculated wear ratio exceeds a preset threshold, it indicates that the wear level is insufficient to guarantee the structural integrity of the reactor for long-term operation, and continued service could easily lead to safety accidents. For different fuel rods, due to differences in initial wall thickness, the wear ratio can be used as a basis for judging the rationality of experimental conditions. For example, if the wear ratio of fuel rods with different wall thicknesses is too low, it may be due to problems such as abnormal vibration source output or insufficient experimental time. In this case, it is necessary to shut down the reactor to check the vibration source condition or appropriately extend the experimental time to ensure that the experiment covers the entire life cycle of fuel rod wear and avoid data distortion due to insufficient experimental conditions.

[0050] In addition, the changing trend of the equivalent fatigue parameter (K) of the clamping spring at the clamping point to be monitored can be used as an auxiliary criterion for wear evolution. If K is continuously decreasing (indicating that the spring is becoming looser) and the growth rate of d is also accelerating, it can be clearly determined that the spring is not clamping tightly, which leads to more intense fretting friction between the fuel rod and the spring, thereby accelerating wear. This provides a quantitative basis for the decision-making on preventive maintenance and replacement of fuel assemblies by showing the impact of clamping performance degradation on the accelerated wear effect of fuel rods.

[0051] In a preferred embodiment, the method further includes: after the experiment, performing offline calibration on the fuel rod and clamping spring to obtain the actual wear depth of the clamping point and the actual fatigue parameters of the clamping spring; comparing the equivalent wear depth of the fuel rod at the clamping point corresponding to the last moment with the actual wear depth of the clamping point, and comparing the equivalent fatigue parameters of the clamping spring at the clamping point corresponding to the last moment with the actual fatigue parameters of the clamping spring; and correcting the preset amplitude transmission relationship model, the preset wear fatigue mapping model, the safety threshold range, the preset rate of change threshold, and the preset ratio threshold based on the comparison results.

[0052] Specifically, after the experiment, the fuel rod to be monitored was calibrated offline using equipment such as a thickness gauge and a 3D topology analyzer to obtain the actual wear depth of the clamping point. The clamping spring was also calibrated offline using static mechanical tests to obtain its actual fatigue parameters. The equivalent wear depth of the fuel rod at the last clamping point was compared with the actual wear depth of the clamping point, and the equivalent fatigue parameters of the clamping spring at the last clamping point were compared with the actual fatigue parameters of the clamping spring. Based on the comparison results, the preset amplitude transmission relationship model, the preset wear fatigue mapping model, the safety threshold range, the preset rate of change threshold, and the preset proportional threshold were corrected. For example, if the deviation stems from insufficient model prediction accuracy (such as a large gap between the equivalent value and the true value), the network weights and structural parameters of the preset amplitude transmission relationship model and the preset wear fatigue mapping model are adjusted to optimize the fitting accuracy of the mapping relationship. If the deviation stems from a mismatch between the warning criteria and the actual safety boundary (such as warning too early or too late according to the original threshold), the safety threshold range, the preset rate of change threshold, and the preset ratio threshold are corrected to make the warning logic more consistent with the actual degradation law of fuel rods and springs.

[0053] The method for monitoring fuel rod wear depth and clamping spring fatigue state provided by this invention constructs a complete inverse chain from the amplitude of easily measurable points on the fuel rod to the equivalent amplitude at the clamping point, and then to the wear depth and spring fatigue state, by building a preset amplitude transmission relationship model and a preset wear fatigue mapping model. This solution can obtain the dynamic change trend of fuel rod wear depth and clamping spring fatigue state in real time under vibration conditions without disassembling the nuclear fuel assembly, effectively making up for the shortcomings of traditional offline disassembly and testing, and reconstructing the dynamic evolution process of fuel rods and springs under vibration conditions.

[0054] To facilitate understanding by those skilled in the art, the workflow of the fuel rod wear depth and clamping spring fatigue state monitoring method provided by this invention is as follows: Vibration amplitudes at multiple preset measurement positions of the fuel rod to be monitored are acquired in real time, forming a measurement amplitude vector of the fuel rod to be monitored at the corresponding time. The measurement amplitude vector at each time is input into a preset amplitude transfer relationship model matching the clamping point to be monitored, and the equivalent amplitude of the clamping point to be monitored at the corresponding time is output through model calculation. The equivalent amplitude of the clamping point to be monitored at each time is input into a preset wear fatigue mapping model, and the equivalent fatigue parameters of the clamping spring and the equivalent wear depth of the fuel rod at the corresponding time are obtained through model calculation. This invention can obtain the dynamic change trend of fuel rod wear depth and clamping spring fatigue state in real time under vibration conditions, making up for the shortcomings of traditional offline disassembly and testing, and restoring the dynamic evolution process of the fuel rod and spring under vibration conditions.

[0055] This article uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. The above descriptions are only preferred embodiments of the present invention. It should be noted that due to the limitations of textual expression, and the objective existence of infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of the present invention, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes, or combinations, or the direct application of the inventive concept and technical solution to other situations without modification, should all be considered within the scope of protection of the present invention.

Claims

1. A method for assessing the service status of nuclear fuel assemblies, characterized in that, include: S1: Real-time acquisition of amplitude at multiple measurement locations of the fuel rod to be monitored, and generation of the measurement amplitude vector of the fuel rod to be monitored at each moment based on the amplitude at the multiple measurement locations; S2: Input the measured amplitude vector of the fuel rod to be monitored at each moment into the preset amplitude transfer relationship model corresponding to the clamping point to be monitored, and obtain the equivalent amplitude of the clamping point to be monitored at each moment; The preset amplitude transmission relationship model is obtained by training a neural network model with the measured amplitude vector of the sample fuel rod (1) as the model input and the actual amplitude of the sample clamping point as the model output; the sample clamping point is the clamping point on the sample fuel rod (1) that corresponds to the clamping point to be monitored. S3: Input the equivalent amplitude of the clamping point to be monitored at each moment into the preset wear fatigue mapping model to obtain the equivalent fatigue parameters of the clamping spring and the equivalent wear depth of the fuel rod at each moment; the preset wear fatigue mapping model is obtained by training a neural network model with the equivalent amplitude of the sample clamping point as the model input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point as the model output.

2. The evaluation method according to claim 1, characterized in that, The method further includes: acquiring the amplitude of multiple measurement positions of the sample fuel rod (1); generating a measurement amplitude vector of the sample fuel rod (1) based on the amplitude of the multiple measurement positions; acquiring the actual amplitude of the sample clamping point; the actual amplitude of the sample clamping point is obtained by collecting the measurement value of the sensor installed at the installation position corresponding to the sample clamping point on the positioning grid (2); using the measurement amplitude vector of the sample fuel rod (1) as the model input and the actual amplitude of the sample clamping point as the model output, performing neural network model training to obtain the preset amplitude transmission relationship model.

3. The evaluation method according to claim 2, characterized in that, Using the measured amplitude vector of the sample fuel rod (1) as the model input and the actual amplitude of the sample clamping point as the model output, a neural network model is trained to obtain the preset amplitude transfer relationship model. This includes: dividing the measured amplitude vector of the sample fuel rod (1) into a training set and a test set; using the measured amplitude vector in the training set as the input and the actual amplitude of the sample clamping point as the output, a neural network model is trained to obtain a first intermediate model; inputting the measured amplitude vector in the test set into the first intermediate model to obtain the predicted amplitude of the sample clamping point; calculating the amplitude error between the predicted amplitude and the actual amplitude of the sample clamping point; adjusting the network weight parameters of the first intermediate model based on the amplitude error until the amplitude error is less than or equal to a preset amplitude error threshold to obtain the preset amplitude transfer relationship model.

4. The evaluation method according to claim 1, characterized in that, The method further includes: obtaining the equivalent amplitude of the sample clamping point; the equivalent amplitude of the sample clamping point is obtained by inputting the measured amplitude vector of the sample fuel rod (1) into the preset amplitude transfer relationship model corresponding to the sample clamping point; obtaining the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point; the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point are obtained by offline calibration of the clamping spring and the sample fuel rod (1) after each experiment; using the equivalent amplitude of the sample clamping point as the model input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point as the model output, performing neural network model training to obtain the preset wear fatigue mapping model.

5. The evaluation method according to claim 4, characterized in that, Using the equivalent amplitude of the sample clamping point as model input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point as model output, a neural network model is trained to obtain the preset wear fatigue mapping model. This includes: dividing the equivalent amplitude of the sample clamping point into a training set and a test set; using the equivalent amplitude in the training set as input and the actual fatigue parameters of the clamping spring and the actual wear depth of the fuel rod at the sample clamping point as output, a neural network model is trained to obtain a second intermediate model; inputting the equivalent amplitude in the test set into the second intermediate model to obtain the predicted fatigue parameters of the clamping spring and the predicted wear depth of the fuel rod at the sample clamping point; calculating the fatigue parameter error between the predicted fatigue parameters and the actual fatigue parameters of the clamping spring, and the depth error between the predicted wear depth and the actual wear depth of the fuel rod; adjusting the network weight parameters of the second intermediate model based on the fatigue parameter error and the depth error until the fatigue parameter error is less than or equal to a preset fatigue parameter error threshold and the depth error is less than or equal to a preset depth error threshold, thus obtaining the preset wear fatigue mapping model.

6. The evaluation method according to claim 2, characterized in that, The method further includes: machining mounting holes on the positioning grid (2) of the nuclear fuel assembly micro-motion wear accelerated reliability test device at mounting positions corresponding to the sample clamping points; embedding the sensor into the mounting holes, and ensuring that the detection surface of the sensor probe and the sample clamping point of the sample fuel rod (1) meet a preset distance; and collecting the amplitude of the sample clamping point of the sample fuel rod (1) through the sensor to obtain the actual amplitude of the sample clamping point.

7. The evaluation method according to claim 1, characterized in that, The method further includes: comparing the equivalent wear depth of the fuel rod at the clamping point to be monitored at each moment with the safety threshold range of different levels; if the equivalent wear depth of the fuel rod at the clamping point to be monitored falls into the safety threshold range of the corresponding level, then outputting the prompt information corresponding to the safety threshold range of that level.

8. The evaluation method according to claim 7, characterized in that, The method further includes: calculating the absolute rate of change of the equivalent wear depth of the fuel rod at the clamping point to be monitored within a preset time; comparing the absolute rate of change with a preset rate of change threshold; and issuing a corresponding warning message if the absolute rate of change is greater than the preset rate of change threshold.

9. The evaluation method according to claim 8, characterized in that, The method further includes: obtaining the initial wall thickness of the clamping point to be monitored; calculating the wear ratio of the clamping point to be monitored at each moment based on the initial wall thickness of the clamping point to be monitored and the equivalent wear depth of the fuel rod at the clamping point to be monitored at each moment; comparing the wear ratio with a preset ratio threshold, and if the wear ratio is greater than or equal to the preset ratio threshold, outputting the corresponding warning information.

10. The evaluation method according to claim 9, characterized in that, The method further includes: after the experiment, performing offline calibration on the fuel rod to be monitored and the clamping spring to obtain the actual wear depth of the clamping point to be monitored and the actual fatigue parameters of the clamping spring; comparing the equivalent wear depth of the fuel rod at the clamping point to be monitored with the actual wear depth of the clamping point to be monitored at the last moment, and comparing the equivalent fatigue parameters of the clamping spring at the clamping point to be monitored with the actual fatigue parameters of the clamping spring at the clamping point to be monitored at the last moment, and correcting the preset amplitude transmission relationship model, preset wear fatigue mapping model, safety threshold range, preset rate of change threshold, and preset ratio threshold according to the comparison results.

Images