Wall thickness prediction method for uranium hexafluoride container

Abstract

According to the wall thickness prediction method for the uranium hexafluoride container, waste of resources such as maintenance personnel and maintenance equipment of the uranium hexafluoride container is reduced, the blank of artificial intelligence prediction of the state of the uranium hexafluoride container is filled up, professional inspection suggestions and maintenance suggestions are given according to the prediction result of the prediction model, and the prediction efficiency is improved. A direct prediction model and an indirect prediction model are constructed according to the time-varying rule of the wall thickness of the uranium hexafluoride container and used for calculating the residual wall thickness of the uranium hexafluoride container, and after the prediction precision of the prediction models is verified, the use conditions of the two prediction models are formulated in combination with the actual situation. The direct prediction model is M = thetaS * M0 + betaG * log (age) + delta, thetaS and betaG are parameters of the direct prediction model, M is the residual minimum wall thickness, M0 is the initial wall thickness, log (age) is the natural logarithm of the age of the uranium hexafluoride container, and delta is the random error of the model. The indirect prediction model is M (t) = C0-D (t), C0 is the initial wall thickness, and D (t) is the maximum corrosion depth obtained through power law relation calculation.

Description

Technical Field

[0001] The invention relates to the field of radioactive container maintenance, and in particular to a wall thickness prediction method for a uranium hexafluoride container. Background Art

[0002] With the development of uranium conversion production, people's awareness of production safety has gradually increased. Uranium hexafluoride is an important intermediate product in the production of nuclear fuel. The operation safety of uranium hexafluoride is very important and is closely related to the health of workers and the surrounding environment.

[0003] Uranium hexafluoride containers are an important barrier between uranium hexafluoride and the environment and humans, playing an important role in protecting, containing and isolating uranium hexafluoride, protecting humans and preventing the environment from radioactive contamination. The life of a uranium hexafluoride container is generally about several decades. During the long use of uranium hexafluoride containers, the wall thickness of the uranium hexafluoride containers becomes thinner due to corrosion, transportation wear and other factors, affecting their safety performance.

[0004] At present, for the safety inspection of uranium hexafluoride containers, precise instruments are manually carried to strengthen the inspection of uranium hexafluoride containers and increase the inspection frequency to prevent the failure of the uranium hexafluoride barrier and the direct release of uranium hexafluoride into the environment, causing harm to workers and the public. However, during the inspection process, since uranium hexafluoride itself contains radioactive nuclides, the transport container for transporting uranium hexafluoride is equivalent to a radiation source, which continuously emits radiation to its surroundings, thus causing radiation protection problems from external exposure. Therefore, even if the workers wear protective clothing, the frequent inspections and a large amount of inspection time will result in longer exposure time for the workers and there is still a radiation risk. Summary of the invention

[0005] The purpose of the present invention is to provide a wall thickness prediction method for uranium hexafluoride containers. By constructing a direct prediction model and an indirect prediction model, accurate wall thickness prediction values ​​are output according to different situations, and professional inspection opinions and maintenance suggestions are given according to the prediction results of the prediction model, thereby reducing the waste of resources such as maintenance personnel and maintenance equipment of uranium hexafluoride containers, reducing the inspection frequency and time, and filling the gap that artificial intelligence cannot predict the status of uranium hexafluoride containers.

[0006] The embodiment of the present invention is achieved as follows:

[0007] A wall thickness prediction method for uranium hexafluoride containers is proposed. A direct prediction model and an indirect prediction model are constructed according to the time-varying law of the wall thickness of the uranium hexafluoride container. The prediction model is used to calculate the remaining wall thickness of the uranium hexafluoride container. After the prediction accuracy of the prediction model is verified, the use conditions of the two prediction models are formulated in combination with the actual situation.

[0008] The direct prediction model is: M = θ S×M0+β G ×log(age)+δ, where θ S and β G To directly predict the model parameters, M is the remaining minimum wall thickness, M0 is the initial wall thickness, log(age) is the natural logarithm of the age of the UF6 container, and δ is the random error of the model;

[0009] The indirect prediction model is: M(t)=C0-D(t), where C0 is the initial wall thickness and D(t) is the maximum corrosion depth calculated by the power law relationship.

[0010] As a preferred technical solution, the parameter θ of the direct prediction model S and β G The calculation methods include:

[0011] Acquire historical wall thickness data, the historical wall thickness data including a measured data group composed of historical measurement time and historical measured wall thickness at corresponding time points;

[0012] The historical wall thickness data were grouped and collated according to the container characteristics of the UF6 containers;

[0013] Perform model weighted optimization on the grouped data according to the time relationship;

[0014] Calculate θ for each set of data by directly predicting the model S and β G ;

[0015] According to the regression method, the parameter θ is obtained S and β G Initial threshold;

[0016] Determine the direct prediction model parameters θ through fitting optimization S and β G .

[0017] As a preferred technical solution, the container characteristics include geographical characteristics of the container storage location, weather characteristics, uranium hexafluoride state characteristics, usage condition characteristics, and container material characteristics.

[0018] As a preferred technical solution, the random error δ obtains the confidence interval or probability of the random error by means of statistical distribution.

[0019] As a preferred technical solution, the direct prediction model optimizes the parameters of the direct prediction model by fitting historical wall thickness data, and at the same time verifies or evaluates the prediction ability of the direct prediction model.

[0020] As a preferred technical solution, the power law relationship calculation method of the maximum corrosion depth is: D(t) = A×(age) b, where A and b are indirect prediction model parameters, and age is the service life of the UF6 container.

[0021] As a preferred technical solution, the calculation method of the indirect prediction model parameters A and b includes:

[0022] Obtain historical pit depth data, the historical pit depth data including a measured data set consisting of historical measurement time and historical corrosion depth at corresponding time points;

[0023] The maximum corrosion depth calculation model was logarithmically transformed, and the formula after logarithmic transformation was: log(D(t))=log(A)+b×log(age). The estimated values ​​of log(A) and b were obtained by linear regression analysis, and the estimated values ​​of A and b were calculated after exponential transformation.

[0024] The indirect prediction model parameters A and b were determined by fitting optimization.

[0025] As a preferred technical solution, after obtaining the estimated values ​​log(A) and b through linear regression analysis, a goodness of fit test is also performed.

[0026] As a preferred technical solution, in the process of calculating the maximum corrosion depth, statistical distribution is used to describe the variability of the corrosion depth in the basic model to calculate the predicted confidence interval or probability.

[0027] As a preferred technical solution, the direct prediction model and the indirect prediction model are selected or combined according to the service life of the uranium hexafluoride container to output the optimal remaining wall thickness prediction value.

[0028] The beneficial effects of the embodiments of the present invention are:

[0029] 1. The present invention predicts the wall thickness of uranium hexafluoride containers through direct prediction models and indirect prediction models, thereby providing targeted inspection opinions and maintenance suggestions, significantly improving the efficiency of analyzing the walls of uranium hexafluoride containers, and being more conducive to the operation units to quickly and efficiently overhaul or replace uranium hexafluoride containers, thereby further improving the safety of uranium hexafluoride containers during storage or transportation;

[0030] 2. The data sets required for constructing the direct prediction model and the indirect prediction model of the present invention are sorted and optimized, and the data sets can be used by the two prediction models at the same time, which not only provides sufficient support for establishing the prediction model, but also improves the prediction speed;

[0031] 3. With respect to the existing uranium hexafluoride containers that still use the original enhanced inspection frequency to determine the status of the uranium hexafluoride containers, the wall thickness prediction method for the uranium hexafluoride container in the present invention fills the gap in the intelligent evaluation of uranium hexafluoride containers and reduces the waste of human, material and financial resources in the maintenance process of uranium hexafluoride containers;

[0032] 4. The direct prediction model and the indirect prediction model can be selected or combined according to the different usage status or usage time of uranium hexafluoride containers to obtain the optimal prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0034] Figure 1 A flow chart of a method for constructing a direct prediction model for the wall thickness of a uranium hexafluoride container according to an embodiment of the present invention;

[0035] Figure 2 Flow chart of a method for calculating parameters θS and βG of a direct prediction model according to an embodiment of the present invention;

[0036] Figure 3 A flow chart of a method for constructing an indirect prediction model for the wall thickness of a uranium hexafluoride container according to an embodiment of the present invention;

[0037] Figure 4 This is a flow chart of a method for calculating indirect prediction model parameters A and b according to an embodiment of the present invention;

[0038] Figure 5 This is a functional structural block diagram of a wall thickness prediction system for a uranium hexafluoride container according to an embodiment of the present invention.

[0039] Icon: historical wall thickness data acquisition unit C1; direct prediction model and indirect prediction model construction unit C2; direct prediction model and indirect prediction model calling unit C3; result output unit C4. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0041] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0043] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more comprehensive and complete and fully convey the concept of the example embodiments to those skilled in the art.

[0044] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0045] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0046] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0047] This embodiment provides a wall thickness prediction method for uranium hexafluoride containers. According to the time-varying pattern of the wall thickness of uranium hexafluoride containers or the changing pattern of other uranium hexafluoride containers susceptible to corrosion / wear, a direct prediction model and an indirect prediction model are constructed to calculate the remaining wall thickness of the uranium hexafluoride container. After verifying the prediction accuracy of the prediction model, the prediction results of the direct prediction model and the indirect prediction model are selected or further integrated in combination with actual conditions, so as to provide highly accurate prediction values ​​for personnel in the field of uranium hexafluoride container use, and give targeted inspection opinions and maintenance suggestions.

[0048] First embodiment

[0049] See also Figure 1 and 2 In this embodiment, the remaining wall thickness of the uranium hexafluoride container is calculated by constructing a direct prediction model, and the method includes:

[0050] A1: Collect and organize historical wall thickness data; in the process of collecting historical wall thickness, these data usually come from regular measurement data of precision instruments, such as measured data obtained by using high-precision ultrasonic thickness gauges, which are reliable. For recording errors, data preprocessing can be performed through data cleaning and enhancement, data type conversion, etc. to ensure that the historical wall thickness data is smooth and continuous, and adapts to and matches the requirements of the model. When missing values ​​appear, it is necessary to select appropriate filling methods according to the attributes and distribution of the missing data in the data preprocessing stage, such as linear regression, mean filling, nearest neighbor filling, and mode filling. Select an appropriate method for data filling.

[0051] During the sorting process, the historical wall thickness data is a measured data group composed of the historical measurement time and the historical measured wall thickness at the corresponding time point; then, the historical wall thickness data is grouped and sorted according to the container characteristics of the uranium hexafluoride container.

[0052] Among them, container characteristics include characteristics that may easily affect the use of the container, such as the storage location or transportation location of the container, including external characteristics such as geographical characteristics (storage location, altitude, soil acidity, etc.), weather characteristics (air humidity, nitrogen oxides, sulfur dioxide content, etc.), as well as internal characteristics such as the container itself (coating quality protection measures, the wear rate of the container during driving, the material of the container, etc.) and the uranium hexafluoride medium itself (physical characteristics of uranium hexafluoride, temperature, air pressure, filling amount, etc.).

[0053] In order to quickly obtain the above container characteristics, this embodiment can also use K-means clustering algorithm to classify the data generated during the storage or transportation of uranium hexafluoride, so that the data can play its own value.

[0054] In terms of container feature processing, a correlation analysis model for factors affecting uranium hexafluoride container defects can also be constructed. The key factors that lead to uranium hexafluoride container corrosion and the importance of the key factors can be discovered from numerous risk factors. Data weighting can be performed according to the importance, which is conducive to screening more accurate data for uranium hexafluoride container integrity evaluation and obtaining more reasonable decision-making methods.

[0055] The numerical values ​​of container characteristics can be used to establish a simulation model through software to explore the influence of various characteristic data on the corrosion rate of uranium hexafluoride containers.

[0056] A2: Construct a direct prediction model, which is used as a basic model for predicting the change of container wall thickness over time. The direct prediction model is: M = θ S ×M0+β G ×log(age)+δ, where θ S and β G To directly predict the model parameters, M is the remaining minimum wall thickness, M0 is the initial wall thickness, log(age) is the natural logarithm of the age of the UF6 container, and δ is the random error of the model;

[0057] The data of M0 can be obtained through the nominal thickness on the design drawing or through the thickness measurement at the factory; directly predict the model parameter θ S and β G It can be calculated by directly predicting the model and the actual measured parameters after grouping; the random error δ uses statistical distribution to describe the variability of the basic model in order to calculate the confidence interval or probability of the prediction.

[0058] Specifically, directly predict the model parameters θ S and β G The calculation methods include:

[0059] A21: Retrieve historical wall thickness data and retrieve the sorted historical wall thickness data to facilitate model training and optimization;

[0060] A22: The historical wall thickness data after grouping is optimized by weighting according to the time relationship. Since the variance of the data changes with the increase of service life, the model needs to be weighted to ensure that the containers with different service life are given appropriate weights during the fitting process.

[0061] A23: Preliminary calculation of θ for each set of data using the direct prediction model and historical wall thickness data S and β G , and the parameter θ is obtained according to the regression method S and β G Initial threshold; at the same time, the goodness of fit of the regression model is checked to ensure that the model can reasonably describe the data;

[0062] A24: Refit and optimize to determine the direct prediction model parameters θ S and β G .

[0063] In this embodiment, the random error δ uses statistical distribution to describe the variability of the basic model. Various random errors in the use of uranium hexafluoride containers also have a great influence on the prediction. Therefore, the random error is considered to be optimized again, for example, by using sensitivity analysis, introduction of random concepts, etc. to make it obey the normal distribution, so as to ensure the accuracy of the random error.

[0064] A3: Use historical wall thickness data to perform classification training, iterative fitting and verification on the direct prediction model to obtain a mature direct prediction model, and verify or evaluate the prediction ability of the direct prediction model.

[0065] The direct prediction model in this embodiment has a simple mathematical model, a short calculation time, and accurate calculation results, which are easy to verify.

[0066] In order to verify the reliability of the classifier, the performance test of the classifier can be carried out by cross-validation. The data is divided into two groups: training set and test set. The classifier is trained by the training set, and the training model is evaluated by the test set.

[0067] In the verification process, the confidence interval method is used. The confidence interval method is an important statistical test method for quantitative analysis and evaluation of the effectiveness of the model. By determining the calculation interval, it is judged whether the model meets the acceptable accuracy range, and then the effectiveness of the model is quantitatively evaluated and the confidence of the model is determined.

[0068] Evaluation indicators include accuracy, precision, F1 score, and confusion matrix. Accuracy indicates the proportion of all samples that are accurately classified. Precision is used to measure the proportion of true positive samples in instances predicted to be positive. F1 score represents the weighted harmonic mean of recall and accuracy. Confusion matrix is ​​a comprehensive evaluation indicator system that describes the classification performance of the model. It describes the relationship between the diagnostic results of the diagnostic model and the true value.

[0069] Second embodiment

[0070] See also Figure 3 and 4 In this embodiment, the wall thickness of the uranium hexafluoride container is calculated by constructing an indirect prediction model. The principle is to indirectly obtain the minimum remaining wall thickness of the uranium hexafluoride container by predicting the maximum corrosion depth of the surface of the uranium hexafluoride container. The method includes:

[0071] B1: Obtain and organize historical pit depth data. The historical pit depth data includes a measured data set consisting of the historical measurement time and the historical corrosion depth at the corresponding time point. In the process of collecting historical pit depth data, it is still indirectly obtained by regular precision instrument measurements. It is indirectly obtained by assuming that the wall thickness of the original uranium hexafluoride container is uniform, and then subtracting the actual wall thickness measurement value obtained by the precision instrument. The advantage of this is that it can eliminate the prediction deviation caused by the uneven wall thickness of the uranium hexafluoride container. Similarly, after acquisition, the historical pit depth data needs to be cleaned and enhanced, and statistical distribution is used to describe the variability of corrosion depth in the basic model to calculate the confidence interval or probability of the prediction.

[0072] The historical pit depth data includes a measured historical data set formed by collation of the historical measurement time and the historical corrosion depth value at the corresponding measurement time point.

[0073] B2: Construct an indirect prediction model. The indirect prediction model is: M(t)=C0-D(t), where C0 is the initial wall thickness, which can also be obtained through the nominal thickness on the design drawing or through the thickness measurement at the factory. D(t) is the maximum corrosion depth predicted by the power law relationship.

[0074] Specifically, the power law relationship calculation method for the maximum corrosion depth is: D(t) = A×(age) b , where A and b are indirect prediction model parameters, age is the service life of the uranium hexafluoride container, and the calculation method of the indirect prediction model parameters A and b includes:

[0075] B21: The maximum corrosion depth calculation model is logarithmically transformed. The formula after logarithmic transformation is: log(D(t))=log(A)+b×log(age). The historical pit depth data is retrieved and linear regression analysis is used to obtain the estimated values ​​of log(A) and b. The estimated values ​​of A and b are calculated after exponential conversion.

[0076] B22: Determine the indirect prediction model parameters A and b through fitting optimization.

[0077] After obtaining the estimated values ​​of log(A) and b through linear regression analysis, a goodness of fit test is also performed to ensure that the indirect prediction model can reasonably describe the data. If the indirect prediction model does not fit well, it is necessary to reconsider the quality of the data, the choice of model, or whether there are outliers. After adjustment, refit and perform diagnosis and verification here until the parameters A and b are obtained.

[0078] B3: Use historical wall thickness data to train, fit and verify the indirect prediction model to obtain a mature indirect prediction model, and verify or evaluate the prediction ability of the indirect prediction model.

[0079] For example:

[0080] Assume we have the following simplified historical data:

[0081] Container No. Age (years) The largest pit depth in history (μm) 1 10 50 2 15 70

[0082] The linear regression analysis and optimization method in the second embodiment are used to estimate A and b for the above data, and A=7.3982 and b=0.8298 are obtained.

[0083] Now, for a 30-year-old uranium hexafluoride container, we predict that its maximum pit depth D(30) is: D(30) = 7.3982 × (30) 0.8298 =124μm.

[0084] The initial thickness and the predicted maximum pit depth are then used to estimate the minimum wall thickness of the 30-year-old container: M = initial thickness - D(30). If M is below a certain safety threshold, the cylinder may need to be inspected and confirmed on site and repaired or replaced.

[0085] Third embodiment

[0086] In order to obtain highly accurate prediction values, further optimizations have been made for the application of direct prediction models and indirect prediction models.

[0087] When the value of the remaining minimum wall thickness M calculated by the direct prediction model or the remaining minimum wall thickness M(t) calculated by the indirect prediction model is less than the safety threshold, different UF6 containers have different materials or sizes, and their safety thresholds are also different. For example, the internal volume of the 30B UF6 transport container is 740L, the wall thickness is 1.27cm, and the thinnest wall thickness allowed is 0.8cm. For example, when the safety threshold is 0.9cm, the UF6 container is inspected and confirmed on site and repaired. When the safety threshold is 0.8cm, the UF6 container is replaced.

[0088] Alternatively, all prediction models are weighted and optimized according to the accuracy of the two prediction models for uranium hexafluoride containers with different service lives, and the optimal weighted combination model of the remaining wall thickness is obtained, and then the most weighted combination model is used to predict the remaining wall thickness of the pressure-bearing equipment. For example, if the service life of a uranium hexafluoride container is one or two years, and its wear degree is less affected by the characteristics of the container, then the weight of the indirect prediction model is greater than that of the direct prediction model. The value of the weight can be obtained based on the accuracy and correlation analysis of the prediction curve and the standard curve of the indirect prediction model and the direct prediction model in the law of change over time.

[0089] Alternatively, based on the optimal weighted combination model and the research results of the correlation analysis model, the weights of key risk factors can be calculated in combination with the hierarchical analysis method, and a risk scoring system correction model can be established to obtain the safety factor of corrosion defect evaluation and further improve the prediction accuracy.

[0090] Fourth embodiment

[0091] See also Figure 5 This embodiment provides a wall thickness prediction system for a uranium hexafluoride container, which includes:

[0092] The historical wall thickness data acquisition unit C1 includes a measured data group composed of historical measurement time and historical measured wall thickness at corresponding time points; the historical wall thickness data are grouped and sorted according to the container characteristics of the uranium hexafluoride container;

[0093] The direct prediction model and the indirect prediction model construct unit C2, and use or indirectly use the historical wall thickness data to train and optimize the model. The direct prediction model is: M = θ S ×M0+β G ×log(age)+δ, where θ S and β G For direct prediction of model parameters, M is the remaining minimum wall thickness, M0 is the initial wall thickness, log(age) is the natural logarithm of the age of the uranium hexafluoride container, and δ is the random error of the model; the indirect prediction model is: M(t) = C0-D(t), where C0 is the initial wall thickness and D(t) is the maximum corrosion depth calculated by the power law relationship;

[0094] The direct prediction model and the indirect prediction model calling unit C3 selects or combines the direct prediction model and the indirect prediction model according to the service life of the uranium hexafluoride container, obtains the optimal prediction model through algorithm selection and verification, and is used to output an accurate prediction value;

[0095] The result output unit C4 outputs inspection opinions and maintenance suggestions to the staff based on the predicted values.

[0096] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0097] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, or it can be an electrical, mechanical or other form of connection.

[0098] The units described as separate components may or may not be physically separated. As a unit, it can be appreciated by those skilled in the art that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0099] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0100] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a grid device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0101] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for predicting the wall thickness of a uranium hexafluoride container, characterized in that: According to the time-varying law of the wall thickness of the uranium hexafluoride container, a direct prediction model and an indirect prediction model are constructed to calculate the remaining wall thickness of the uranium hexafluoride container; The direct prediction model is: M = θ S ×M0+β G ×log(age)+δ, where θ S and β G To directly predict the model parameters, M is the remaining minimum wall thickness, M0 is the initial wall thickness, log(age) is the natural logarithm of the age of the UF6 container, and δ is the random error of the model; The indirect prediction model is: M(t)=C0-D(t), wherein C0 is the initial wall thickness and D(t) is the maximum corrosion depth calculated by the power law relationship.

2. The wall thickness prediction method for uranium hexafluoride container according to claim 1, characterized in that: The parameters θ of the direct prediction model S and β G The calculation methods include: Acquire historical wall thickness data, wherein the historical wall thickness data includes a measured data group composed of historical measurement time and historical measured wall thickness at corresponding time points; Grouping and arranging the historical wall thickness data according to the container characteristics of the uranium hexafluoride containers; Perform model weighted optimization on the grouped data according to the time relationship; Calculate θ for each set of data using the direct prediction model S and β G ; According to the regression method, the parameter θ is obtained S and β G Initial threshold; Determine the direct prediction model parameters θ through fitting optimization S and β G .

3. The wall thickness prediction method for uranium hexafluoride container according to claim 2, characterized in that: The container characteristics include geographical characteristics of the container storage location, weather characteristics, uranium hexafluoride state characteristics, usage condition characteristics, and container material characteristics.

4. The wall thickness prediction method for uranium hexafluoride container according to claim 1, characterized in that: The random error δ obtains the confidence interval or probability of the random error by means of statistical distribution.

5. The wall thickness prediction method for uranium hexafluoride container according to claim 1, characterized in that: The direct prediction model optimizes the parameters of the direct prediction model by fitting the historical wall thickness data, and at the same time verifies or evaluates the prediction ability of the direct prediction model.

6. The wall thickness prediction method for uranium hexafluoride container according to claim 1, characterized in that: The power law relationship calculation method of the maximum corrosion depth is: D(t) = A×(age) b , where A and b are indirect prediction model parameters, and age is the service life of the UF6 container.

7. The wall thickness prediction method for uranium hexafluoride container according to claim 6, characterized in that: The calculation method of the indirect prediction model parameters A and b includes: Acquire historical pit depth data, wherein the historical pit depth data includes a measured data group consisting of historical measurement time and historical corrosion depth at corresponding time points; The maximum corrosion depth calculation model is logarithmically transformed, and the formula after logarithmic transformation is: log(D(t))=log(A)+b×log(age), and the estimated values ​​of log(A) and b are obtained by linear regression analysis, and the estimated values ​​of A and b are calculated after exponential conversion; The indirect prediction model parameters A and b were determined by fitting optimization.

8. The wall thickness prediction method for uranium hexafluoride container according to claim 7, characterized in that: After the estimated values ​​log(A) and b are obtained by linear regression analysis, a goodness of fit test is also performed.

9. The wall thickness prediction method for uranium hexafluoride container according to claim 7, characterized in that: During the calculation of the maximum corrosion depth, a statistical distribution is used to describe the variability of the corrosion depth in the basic model to calculate the confidence interval or probability of the prediction.

10. The wall thickness prediction method for uranium hexafluoride container according to any one of claims 1 to 9, characterized in that: According to the service life of the uranium hexafluoride container, the direct prediction model and the indirect prediction model are selected or combined to output the optimal remaining wall thickness prediction value.

Images