Nuclear reactor core power prediction method and device, nuclear reactor core monitoring method and system

Abstract

This application discloses a core power prediction method and apparatus, a core monitoring method and system, relating to the field of nuclear industry technology. The core power prediction method includes: acquiring target power detector data for a target core; the target power detector data includes values ​​from multiple power detectors; processing the target power detector data using a target power prediction model to obtain a predicted power value for the target core; wherein the target power prediction model employs a data grouping processing method, trained using historical power detector training data and their corresponding first historical power values. According to embodiments of this application, this core prediction method can monitor core power using a small number of detectors while ensuring the accuracy of core power prediction.

Description

Technical Field

[0001] This application belongs to the field of nuclear industry technology, specifically relating to a core power prediction method and device, and a core monitoring method and system. Background Technology

[0002] During the operation of a nuclear power plant, in order to ensure the safety of the nuclear power plant, it is necessary to understand the core state parameters such as the core power distribution in real time, so as to assess the core state based on the changes in the core state parameters and then carry out subsequent operations.

[0003] In related technologies, mapping and fitting methods are mainly used for core power prediction. However, this method requires many in-core detectors for fitting to ensure the accuracy of core power prediction. Therefore, how to monitor core power with a small number of detectors while ensuring the accuracy of core power prediction is a research hotspot in this field. Summary of the Invention

[0004] The technical problem to be solved by this application is to address the above-mentioned shortcomings of the prior art by providing a core power prediction method and apparatus, a core monitoring method and system. Using the core power prediction method, it is possible to monitor the core power with a small number of detectors while ensuring the prediction accuracy of the core power.

[0005] In a first aspect, embodiments of this application provide a core power prediction method, including: Acquire target power detector data for the target reactor core; the target power detector data includes values ​​from multiple power detectors. The target power detector data is processed using the target power prediction model to obtain the predicted power value of the target core. The target power prediction model is trained using a data grouping method, utilizing historical power detector training data and their corresponding first historical power values.

[0006] In some embodiments of the first aspect, before processing the target power detector data using a target power prediction model to obtain the predicted power value of the target core, the method further includes: S1. Obtain the training set and the test set; the training set includes multiple training samples; the training samples include historical power detector training data and their corresponding first historical power values; the test set includes multiple test samples; the test samples include historical power detector test data and their corresponding second historical power values. S2. Based on the training samples corresponding to any two historical power detector training data, construct the first polynomial of any two historical power detector training data and the first historical power value. S3. Solve for the first polynomial to obtain the first set of coefficients corresponding to the first polynomial; S4. Based on the first coefficient set and its corresponding two historical power detector training data, determine the neuron set of the e-th layer; the neuron set of the e-th layer includes multiple first neurons; e is the iteration number; S5. Calculate the first residual between each first neuron and the second historical power value using the test samples to obtain multiple first residuals; S6. Based on multiple first residuals, select at least one second neuron from the set of neurons in layer e to determine the set of neurons in layer e+1. S7. Use each second neuron in the (e+1)th layer neuron set as training data for the historical power detector, and repeat steps S2 to S7 until the preset iteration stop condition is met. Use the first neuron with the smallest first residual in the eth layer neuron set after iteration as the target power prediction model.

[0007] In some implementations of the first aspect, solving the first polynomial to obtain the first set of coefficients corresponding to the first polynomial includes: Solve the first polynomial using the least squares method to obtain the first set of coefficients corresponding to the first polynomial.

[0008] In some embodiments of the first aspect, the second neuron is a first neuron whose first residual is less than or equal to a preset residual.

[0009] In some embodiments of the first aspect, the preset iteration stopping condition includes: the minimum value of each first residual of the neuron set of layer e+1 after iteration is greater than the minimum value of each first residual of the neuron set of layer e after iteration.

[0010] In some embodiments of the first aspect, after processing the target power detector data using a target power prediction model to obtain the predicted power value of the target core, the method further includes: Piecewise polynomial fitting is performed on the predicted power values ​​to obtain the predicted power distribution.

[0011] Based on the same inventive concept, in a second aspect, embodiments of this application also provide a core monitoring method, comprising: For example, the core power prediction method of any of the first aspects determines the predicted power value of the target core; The target reactor core is monitored based on the predicted power value.

[0012] Based on the same inventive concept, in a third aspect, embodiments of this application provide a core power prediction device, comprising: The first acquisition module is used to acquire target power detector data of the target reactor core; the target power detector data includes multiple power detector values. The first processing module is used to process the target power detector data using the target power prediction model to obtain the predicted power value of the target core. The target power prediction model is trained using a data grouping method, utilizing historical power detector training data and their corresponding first historical power values.

[0013] In some embodiments of the third aspect, the device further includes: The second acquisition module is used to execute step S1, which includes: acquiring a training set and a test set; the training set includes multiple training samples; the training samples include historical power detector training data and their corresponding first historical power values; the test set includes multiple test samples; the test samples include historical power detector test data and their corresponding second historical power values. A construction module is used to execute step S2, which includes: constructing a first polynomial between the training data of any two historical power detectors and the first historical power value based on the training samples corresponding to the training data of any two historical power detectors. The solver module is used to execute step S3, which includes: solving the first polynomial to obtain the first set of coefficients corresponding to the first polynomial; The first determining module is used to execute step S4, which includes: determining the set of neurons in the e-th layer based on the first coefficient set and the training data of the two corresponding historical power detectors; the set of neurons in the e-th layer includes multiple first neurons; e is the number of iterations; The calculation module is used to execute step S5, which includes: using test samples to calculate the first residual between each first neuron and the second historical power value to obtain multiple first residuals; The first screening module is used to perform step S6, which includes: screening at least one second neuron from the neuron set of layer e based on multiple first residuals to determine the neuron set of layer e+1. As a module, it is used to execute step S7, which includes: using each second neuron in the (e+1)th layer neuron set as training data for the historical power detector, repeating steps S2 to S7 until the preset iteration stop condition is met, and using the first neuron with the smallest first residual in the iterated eth layer neuron set as the target power prediction model.

[0014] In some implementations of the third aspect, the solving module is specifically used for: Solve the first polynomial using the least squares method to obtain the first set of coefficients corresponding to the first polynomial.

[0015] In some embodiments of the third aspect, the second neuron is a first neuron whose first residual is less than or equal to a preset residual.

[0016] In some implementations of the third aspect, the preset iteration stopping condition includes: the minimum value of each first residual of the neuron set of layer e+1 after iteration is greater than the minimum value of each first residual of the neuron set of layer e after iteration.

[0017] In some embodiments of the third aspect, the device further includes: The fitting module is used to perform piecewise polynomial fitting on the power prediction values ​​to obtain the predicted power distribution.

[0018] Based on the same inventive concept, in a fourth aspect, embodiments of this application provide a reactor core monitoring system, including: The third aspect is the core power prediction device, used to determine the predicted power value of the target core; The monitoring device is used to monitor the target reactor core based on the predicted power value.

[0019] Based on the same inventive concept, in a fifth aspect, embodiments of this application provide an electronic device, the electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores one or more computer programs executable by the at least one processor, the one or more computer programs being executed by the at least one processor to enable the at least one processor to perform the above-described core power prediction method.

[0020] Based on the same inventive concept, in a sixth aspect, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-described core power prediction method.

[0021] Based on the same inventive concept, in a seventh aspect, embodiments of this application provide a computer program product that includes computer-readable code or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code is run in a processor of an electronic device, the processor in the electronic device executes the above-described core power prediction method.

[0022] According to the core power prediction method and apparatus, and core monitoring method and system provided in the embodiments of this application, target power detector data of the target core is first acquired, and then the target power prediction model is used to process the target power detector data to obtain the predicted power value of the target core. In the embodiments of this application, since the target power prediction model adopts a data grouping processing method, it is trained using historical power detector training data and its corresponding first historical power value. That is, by considering the historical power detector training data and its corresponding first historical power value, the predicted power value of the target core is determined, which can ensure the prediction accuracy of the predicted power value of the target core. Furthermore, the target power prediction model can reconstruct the axial power distribution using target power detector data (i.e., new detector signals) as input, thereby enabling the monitoring of the core power of the target core with a small number of detectors. In summary, the embodiments of this application can achieve the monitoring of core power with a small number of detectors while ensuring the prediction accuracy of core power.

[0023] Furthermore, the embodiments of this application predict the target power detector data using a power prediction model, which can improve the speed of core power prediction.

[0024] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description

[0025] The accompanying drawings are provided to further illustrate the present application and form part of the specification. They are used together with the embodiments of the present application to explain the application and do not constitute a limitation thereof. The above and other features and advantages will become more apparent to those skilled in the art from the detailed example embodiments described with reference to the accompanying drawings, in which: Figure 1 This illustration shows a flowchart of a core power prediction method provided in an embodiment of this application. Figure 2 This diagram illustrates a training set and a test set for the target power prediction model provided in an embodiment of this application. Figure 3 This diagram illustrates the training process of the target power prediction model provided in an embodiment of this application. Figure 4 This illustration shows another flowchart of the core power prediction method provided in an embodiment of this application; Figure 5 This paper presents a comparison chart of the predicted power value and the reference value of the target power prediction model provided in the embodiments of this application; Figures 6a to 6f This diagram shows the axial power distribution prediction results provided by the target power prediction model in an embodiment of this application. Figure 7 This diagram shows the power distribution prediction results for axial 15-section blocks and 6-core blocks provided in an embodiment of this application. Figure 8 This illustration shows a structural schematic diagram of a core power device provided in an embodiment of this application; Figure 9 This illustration shows a structural diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0026] To enable those skilled in the art to better understand the technical solutions of this application, exemplary embodiments of this application are described below in conjunction with the accompanying drawings, including various details of the embodiments of this application to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0027] Where there is no conflict, the various embodiments of this application and the features thereof may be combined with each other.

[0028] As used herein, the term “and / or” includes any and all combinations of one or more related enumerated entries.

[0029] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application. As used herein, the singular forms “a” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that when the terms “comprising” and / or “made of” are used in this specification, the presence of the stated feature, integral, step, operation, element, and / or component is specified, but the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or groups thereof is not excluded. Terms such as “connected” or “linked” are not limited to physical or mechanical connections but can include electrical connections, whether direct or indirect.

[0030] Unless otherwise specified, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and this application, and will not be interpreted as having an idealized or overly formal meaning, unless expressly so defined herein.

[0031] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0032] Before introducing the core power prediction method provided in the embodiments of this application, the technical terms and related technologies that may be involved in the embodiments of this application will be explained.

[0033] Group Method of Data Handling (GMDH) is a self-organizing neural network algorithm that does not assume any fixed functional dependencies between inputs and outputs. It models complex input-output relationships by constructing a hierarchical multinomial regression network. This algorithm has been applied in fields such as climate prediction, financial market forecasting, and energy load forecasting.

[0034] Related technologies also use online monitoring methods based on fast calculation models to predict core power. However, in scenarios with smaller reactor types, the solution accuracy of these fast calculation models is difficult to meet engineering requirements.

[0035] The core power prediction method provided in this application is applicable to the operation of nuclear power plants. This core power prediction method can be executed by a core power prediction device and electronic equipment. The following description uses the execution of this core power prediction method by electronic equipment as an example.

[0036] like Figure 1 As shown, the core power prediction method provided in this application includes steps S110 to S120.

[0037] S110. Obtain target power detector data for the target reactor core; the target power detector data includes values ​​from multiple power detectors.

[0038] S120. The target power detector data is processed using the target power prediction model to obtain the predicted power value of the target core.

[0039] According to the core power prediction method provided in this application, target power detector data of the target core is first acquired, and then the target power prediction model is used to process the target power detector data to obtain the predicted power value of the target core. In this application embodiment, since the target power prediction model adopts a data grouping processing method, it is trained using historical power detector training data and its corresponding first historical power value. That is, by considering the historical power detector training data and its corresponding first historical power value, the predicted power value of the target core is determined, which can ensure the prediction accuracy of the predicted power value of the target core. Furthermore, the target power prediction model can reconstruct the axial power distribution using target power detector data (i.e., new detector signals) as input, thereby enabling the monitoring of the core power of the target core with a small number of detectors. In summary, this application embodiment can achieve the monitoring of core power with a small number of detectors while ensuring the prediction accuracy of core power.

[0040] The specific implementation methods for each of the above steps are described below.

[0041] In step S110, for example, the target core can be any core that requires core power prediction.

[0042] For example, multiple power detectors may be installed inside and / or outside the target reactor core. These power detectors are all communicatively connected to electronic devices, which can then acquire target core detector data of the target reactor core through these power detectors.

[0043] It should be noted that the power detector value can be set according to the actual situation and is not limited here.

[0044] In step S120, the target power prediction model can be a model capable of processing target power detector data of the target core to obtain the predicted power value of the target core. In the embodiments of this application, the target power core prediction model can use target power detector data as input to reconstruct the axial power distribution, thereby realizing the monitoring of core power with a small number of detectors.

[0045] In some embodiments, before processing the target power detector data using a target power prediction model to obtain the predicted power value of the target core, the method further includes: S1. Obtain the training set and the test set; the training set includes multiple training samples; the training samples include historical power detector training data and their corresponding first historical power values; the test set includes multiple test samples; the test samples include historical power detector test data and their corresponding second historical power values. S2. Based on the training samples corresponding to any two historical power detector training data, construct the first polynomial of any two historical power detector training data and the first historical power value. S3. Solve for the first polynomial to obtain the first set of coefficients corresponding to the first polynomial; S4. Based on the first coefficient set and its corresponding two historical power detector training data, determine the neuron set of the e-th layer; the neuron set of the e-th layer includes multiple first neurons; e is the iteration number; S5. Calculate the first residual between each first neuron and the second historical power value using the test samples to obtain multiple first residuals; S6. Based on multiple first residuals, select at least one second neuron from the set of neurons in layer e to determine the set of neurons in layer e+1. S7. Use each second neuron in the (e+1)th layer neuron set as training data for the historical power detector, and repeat steps S2 to S7 until the preset iteration stop condition is met. Use the first neuron with the smallest first residual in the eth layer neuron set after iteration as the target power prediction model.

[0046] In this embodiment, the target power prediction model is trained through steps S1 to S7, which provides a basis for subsequently using the target power prediction model to predict the predicted power value of the target core.

[0047] For example, the training set and the testing set constitute the dataset of the target power prediction model. That is, the dataset of the target power prediction model is divided into a training set and a testing set. The testing set can also be called the validation set.

[0048] For example, the set of neurons in the (e+1)th layer may include at least one second neuron. The value of e can be set according to the actual situation and is not limited here. For example, the value of e can be 1, 2, 3, etc.

[0049] For example, prepare A set of historical core power detector data as input variables , The corresponding historical core power detector data Historical power values ​​of reactor cores are used as output variables (Target value), and according to a ratio of 7:3 or 8:2, Group of historical core power detector data and their corresponding The historical power values ​​of the reactor cores were divided into training and test sets, and the constructed dataset is as follows: Figure 2 As shown. Among them, n t The number of training samples in the training set. n c = n - n t The number of test samples in the test set. The number of input variables, i.e., the number of power detectors included in the historical power detector training data.

[0050] For example, such as Figure 3 As shown, the values ​​for each pair of power detectors and ( , =1,2,…, , Using the training samples in the training set, we construct a polynomial with respect to the target value, and obtain formula (1).

[0051] Formula (1) includes: (1) in, , , , , , , , , , These are the coefficients of the polynomial that need to be calculated. The order of the polynomial; It can be the first The power detector input variable, that is, the first power detector input variable. The detected values ​​of each power detector; It can be the first The power detector input variable, that is, the first power detector input variable. The detected value of each power detector.

[0052] In some examples, solving the first polynomial yields the first set of coefficients corresponding to the first polynomial, including: Solve the first polynomial using the least squares method to obtain the first set of coefficients corresponding to the first polynomial.

[0053] In this example, the first polynomial can be solved quickly using the least squares method.

[0054] For example, the GMDH method typically uses polynomials of orders 1 to 3. Taking order 2 as an example, the first polynomial is solved using the least squares method to obtain the first set of coefficients. It satisfies formula (2).

[0055] Formula (2) includes: (2) in, For the first The first historical power detector value measured by the power detector; For the first The second historical power detector value measured by the first power detector; and so on. For the first The power detector measured the first Historical power detector values. For the first The first historical power detector value measured by the power detector; For the first The second historical power detector value measured by the first power detector; and so on. For the first The power detector measured the first Historical power detector values.

[0056] , , , , , The first set of coefficients The amount.

[0057] The first historical power value corresponding to the first historical power detector training data; This is the first historical power value corresponding to the training data of the second historical power detector; and so on. For the first The first historical power value corresponding to the training data of the historical power detector.

[0058] For example, the first layer of neurons is obtained using input variables and least squares coefficients. . , indicating from The number of combinations of selecting two elements from a set of distinct elements is calculated as follows: . This represents the first neuron; This represents the second first neuron; and so on. Indicates the first The first neuron.

[0059] For example, the residual (i.e., the first residual) between each first neuron in the first layer of neurons and the target value (i.e., the second historical power value) is calculated using test samples from the test set; the calculation process satisfies formula (3).

[0060] Formula (3) includes: (3) in, Indicates the first residual; Indicates the first The first historical power detector data corresponding to the first The first neuron.

[0061] In some examples, the second neuron is the first neuron whose first residual is less than or equal to a preset residual.

[0062] In other words, if the first residual of the first neuron is greater than a preset residual, the first neuron is discarded; if the first residual of the first neuron is less than or equal to the preset residual, the first neuron is used as the second neuron to obtain... A second neuron.

[0063] It should be noted that the preset residual can be set according to the actual situation, and there is no limitation here. Figure 3 In this context, "iteration" represents iteration.

[0064] In some examples, the preset iteration stopping condition includes: the minimum value of each first residual of the neuron set of layer e+1 after iteration is greater than the minimum value of each first residual of the neuron set of layer e after iteration.

[0065] For example, in each iteration, check whether the minimum value DMIN of the first residual in the set of neurons in the (e+1)th layer after the iteration is less than the minimum value of the first residual in the set of neurons in the eth layer after the iteration. If DMIN increases, stop the iteration and select the neuron with the smallest first residual in the set of neurons in the eth layer after the iteration as the final construction result, which satisfies the form of formula (4).

[0066] Formula (4) includes: (4) in, The target value; For the first Historical power detector training data; The number of historical power detector values ​​included in the historical power detector training data; , , , All are polynomial coefficients.

[0067] For example, by substituting the target power detector data, including the values ​​of multiple power detectors, into formula (4), the power output can be predicted (i.e., the predicted power value is output).

[0068] In some embodiments, after processing the target power detector data using a target power prediction model to obtain the predicted power value of the target core, the method further includes: Piecewise polynomial fitting is performed on the predicted power values ​​to obtain the predicted power distribution.

[0069] In this embodiment, piecewise polynomial fitting of the predicted power values ​​can yield a fine predicted power distribution, providing a basis for subsequent monitoring of the target reactor core.

[0070] In this embodiment, the GMDH method is applied to reactor core power prediction, training a surrogate model (i.e., a target power prediction model) that takes detector values ​​(i.e., values ​​from multiple power detectors) as input and core power values ​​(i.e., predicted power values) as output. This model can quickly predict core power with a small amount of detector data. The method provided in this embodiment can significantly improve the speed of core power prediction and achieve high prediction accuracy.

[0071] This application uses the GMDH algorithm to predict the core power distribution. This method is trained using several known sets of core power distributions and detector data. The trained model can reconstruct the axial power distribution with new detector signals as input, thereby achieving the purpose of monitoring core power with a small number of detectors.

[0072] To better understand the core power prediction method provided in the embodiments of this application, the following description is provided in conjunction with specific application scenarios.

[0073] The following explanation uses a small-scale reactor model as an example. The reactor core contains 90 fuel assemblies, arranged in three axial layers. Each layer contains 30 hexagonal prism fuel assemblies arranged radially. The six assemblies located in the third axial layer, at the innermost radial edge, have lower fuel enrichment than the other assemblies. The fuel is cylindrical coated particulate fuel. The core moderator and reflector are made of graphite, and the fuel channels and coolant channels are regularly arranged within the hexagonal graphite assemblies.

[0074] like Figure 4 As shown, the core power prediction method provided in this application includes: (1) Use a fine modeling program for calculation.

[0075] (2) Acquire detector data and core power data, and construct a dataset.

[0076] A detailed 3D core model was built using the Monte Carlo method (RMC), and the power distribution under different operating conditions was statistically analyzed. Based on different burnup, temperatures, and control rod positions, 39 operating conditions were constructed. During power statistics, each axial layer was divided into 5 segments, resulting in a total of 15 axial segments. Considering the radial symmetry of the core, only the power in the radial 1 / 6 region was statistically analyzed, that is, only the power in 5 out of the 30 regions in each core layer was analyzed. Therefore, for each operating condition, 5 sets of axial power distributions were obtained, each set containing 15 power data segments. , ..., Therefore, we can obtain 39 working conditions × 5 sets of axial power = 195 sets of axial power distribution data. = =[ (1) , (2) ,…, (15) The average power value of the five segments corresponding to each layer along the axis is taken as the detector value for that layer. Therefore, each set of data corresponds to three detector values. , , ), constituting 195 sets of detector data ( = =[ (1) , (2) , (3) The total dataset consists of 195 sets of detector data. and power data And randomly divided them into training sets in an 8:2 ratio ( train×3 = 156×3 , train×15 = 156×15 =[ train (1) , train (2) , …, train (15) ]) and test set ( check×3 = 39×3 , check×15 = 39×15 ).

[0077] (3) GMDH model training

[0078] detector data train×3 As input features, respectively train (1) , train (2) , …, train (15) Train the GMDH network using the target value to establish the output. =( (1) , (2) ,…, (15) ) and input parameters The functional mapping relationship between them is shown in formula (5).

[0079] Formula (5) includes: (5) in, Indicates the first output (1) With input parameters Functional mapping relationship between them; Indicates the second output (2) With input parameters The function mapping relationship between them; and so on, This indicates the 15th output. (15) With input parameters The functional mapping relationship between them.

[0080] Each layer of neurons is constructed using a second-order polynomial to construct the next layer of neurons.

[0081] The fitting performance of the 15 trained surrogate models (i.e., target power prediction models) on the test set is as follows: Figure 5 As shown. Figure 5 In the diagram, any blue dot corresponds to a segment in a certain operating condition of the test set. Its horizontal axis corresponds to the actual power value of the segment, and its vertical axis corresponds to the predicted power value of the GMDH proxy model. The closer the scatter point is to the red dashed line, the closer the predicted value is to the actual value. Figure 5 The mean squared error (MSE) and mean absolute error (MAE) are labeled in the upper left corner. The surrogate model's predictions on the test set show good agreement with the true values.

[0082] (4) Input the new detector values ​​into the trained model to obtain the core power prediction value.

[0083] Test set detector data check×3 Substitute the trained GMDH surrogate model (i.e., the target power prediction model) into the reconstructed axial 15-segment power distribution of the test set. Thus, we obtain formula (6).

[0084]

[0085] The reconstructed 15-segment power distribution is as follows Figures 6a to 6f As shown. Figures 6a to 6f This shows the comparison between the predicted and actual values ​​of the GMDH surrogate model for three examples in the test set. The horizontal axis represents the axial segment number. Figure 6a , Figure 6c , Figure 6e The vertical axis of the graph represents the power value, the red scatter dots represent the predicted values ​​of the GMDH surrogate model, and the blue bars represent the actual values. Figure 6b , Figure 6d , Figure 6f The vertical axis represents the relative deviation between the predicted and actual values ​​(predicted value - actual value) / actual value. The mean squared error is labeled for each example in the figure. and mean absolute error ,in =15 represents the number of axial segments. Number the nodes / blocks. The test set is numbered. The three examples in the figure represent three different shapes of axial power distribution. The GMDH surrogate model has good prediction accuracy for both the right-hand low-power distribution and the left-hand low-power distribution.

[0086] (5) Piecewise polynomial fitting is used to fit the predicted power to obtain the fine power distribution prediction (i.e., the predicted power distribution).

[0087] The reconstructed 15 axial power distribution segments construct×15 The system is divided into three segments, and a fourth-order polynomial is used to fit each segment to reconstruct the axial power distribution of the 60-core chip. construct×60 . Figure 7 This chart compares the predicted and actual values ​​for a specific example in the test set, using 15-cell and 60-cell modules. The purple scatter plots represent the prediction results for the 60-cell module, while the blue line represents the statistical results from the RMC fine-tuning model. The relative deviations between the predicted and actual values ​​for the 60-cell module are all less than 5%, indicating high reliability of the prediction results.

[0088] Based on the same inventive concept, this application also provides a core monitoring method that can be applied to the operation of nuclear power plants.

[0089] The core monitoring method provided in this application embodiment may include steps S210 to S220.

[0090] S210. Determine the predicted power value of the target core according to the core power prediction method of the above embodiment.

[0091] S220. Monitor the target reactor core based on the predicted power value.

[0092] According to the core monitoring method provided in this application embodiment, target power detector data of the target core is first acquired, and then the target power prediction model is used to process the target power detector data to obtain the predicted power value of the target core. In this application embodiment, since the target power prediction model adopts a data grouping processing method, it is trained using historical power detector training data and its corresponding first historical power value. That is, by considering the historical power detector training data and its corresponding first historical power value, the predicted power value of the target core is determined, which can ensure the prediction accuracy of the predicted power value of the target core. Furthermore, the target power prediction model can reconstruct the axial power distribution using target power detector data (i.e., new detector signals) as input, thereby enabling the monitoring of the core power of the target core with a small number of detectors. In summary, this application embodiment can achieve the monitoring of core power with a small number of detectors while ensuring the prediction accuracy of core power.

[0093] For example, piecewise polynomial fitting can be performed on the predicted power values ​​to obtain the predicted power distribution. Then, the operating status of the target reactor core can be monitored by comparing the trend of the predicted power distribution or by analyzing the deviation from the baseline data.

[0094] The core monitoring method provided in this application includes the core power prediction method provided in the above embodiments, that is, it has the beneficial effects and implementation methods of the core power prediction method provided in the above embodiments. For details, please refer to the specific description of the core power prediction method in the above embodiments, which will not be repeated here.

[0095] Based on the same inventive concept, embodiments of this application provide a reactor core power prediction device. For example... Figure 8 As shown, the core power prediction device provided in this application embodiment includes a first acquisition module 310 and a first processing module 320.

[0096] The first acquisition module 310 is used to acquire target power detector data of the target reactor core; the target power detector data includes multiple power detector values. The first processing module 320 is used to process the target power detector data using the target power prediction model to obtain the predicted power value of the target core. The target power prediction model is trained using a data grouping method, utilizing historical power detector training data and their corresponding first historical power values.

[0097] According to the core power prediction device provided in this application embodiment, target power detector data of the target core is first acquired, and then the target power prediction model is used to process the target power detector data to obtain the predicted power value of the target core. In this application embodiment, since the target power prediction model adopts a data grouping processing method, it is trained using historical power detector training data and its corresponding first historical power value. That is, by considering the historical power detector training data and its corresponding first historical power value, the predicted power value of the target core is determined, which can ensure the prediction accuracy of the predicted power value of the target core. Furthermore, the target power prediction model can reconstruct the axial power distribution using target power detector data (i.e., new detector signals) as input, thereby enabling the monitoring of the core power of the target core with a small number of detectors. In summary, this application embodiment can achieve the monitoring of core power with a small number of detectors while ensuring the prediction accuracy of core power.

[0098] In some embodiments, the device further includes: The second acquisition module is used to execute step S1, which includes: acquiring a training set and a test set; the training set includes multiple training samples; the training samples include historical power detector training data and their corresponding first historical power values; the test set includes multiple test samples; the test samples include historical power detector test data and their corresponding second historical power values. A construction module is used to execute step S2, which includes: constructing a first polynomial between the training data of any two historical power detectors and the first historical power value based on the training samples corresponding to the training data of any two historical power detectors. The solver module is used to execute step S3, which includes: solving the first polynomial to obtain the first set of coefficients corresponding to the first polynomial; The first determining module is used to execute step S4, which includes: determining the set of neurons in the e-th layer based on the first coefficient set and the training data of the two corresponding historical power detectors; the set of neurons in the e-th layer includes multiple first neurons; e is the number of iterations; The calculation module is used to execute step S5, which includes: using test samples to calculate the first residual between each first neuron and the second historical power value to obtain multiple first residuals; The first screening module is used to perform step S6, which includes: screening at least one second neuron from the neuron set of layer e based on multiple first residuals to determine the neuron set of layer e+1. As a module, it is used to execute step S7, which includes: using each second neuron in the (e+1)th layer neuron set as training data for the historical power detector, repeating steps S2 to S7 until the preset iteration stop condition is met, and using the first neuron with the smallest first residual in the iterated eth layer neuron set as the target power prediction model.

[0099] In some implementations, the solver module is specifically used for: Solve the first polynomial using the least squares method to obtain the first set of coefficients corresponding to the first polynomial.

[0100] In some implementations, the second neuron is a first neuron whose first residual is less than or equal to a preset residual.

[0101] In some implementations, the preset iteration stopping condition includes: the minimum value of each first residual of the neuron set of layer e+1 after iteration is greater than the minimum value of each first residual of the neuron set of layer e after iteration.

[0102] In some embodiments, the device further includes: The fitting module is used to perform piecewise polynomial fitting on the power prediction values ​​to obtain the predicted power distribution.

[0103] The core power prediction device provided in this application embodiment can be used to execute the core power prediction method, that is, it has the beneficial effects and implementation methods of the core power prediction method provided in the above embodiments. For details, please refer to the specific description of the core power prediction method in the above embodiments. This embodiment will not repeat the description here.

[0104] Based on the same inventive concept, in a fourth aspect, embodiments of this application provide a core monitoring system, which includes the core power prediction device and monitoring device described in the above embodiments.

[0105] Among them, the core power prediction device is used to determine the predicted power value of the target core; The monitoring device is used to monitor the target reactor core based on the predicted power value.

[0106] The core monitoring system provided in this application embodiment can be used to execute the core monitoring method, that is, it has the beneficial effects and implementation methods of the core monitoring method provided in this application embodiment. For details, please refer to the specific description of the core monitoring method in the above embodiment, which will not be repeated here.

[0107] It is understood that the various method embodiments mentioned above in this application can be combined with each other to form combined embodiments without violating the principle and logic. Due to space limitations, this application will not elaborate further. Those skilled in the art will understand that in the above methods of specific implementation, the specific execution order of each step should be determined by its function and possible internal logic.

[0108] Figure 9 This is a block diagram of an electronic device provided in an embodiment of this application.

[0109] Reference Figure 9 This application provides an electronic device, which includes: at least one processor 701; at least one memory 702; and one or more I / O interfaces 703 connected between the processor 701 and the memory 702; wherein the memory 702 stores one or more computer programs that can be executed by at least one processor 701, and the one or more computer programs are executed by at least one processor 701 to enable at least one processor 701 to perform the above-described core power prediction method or core monitoring method.

[0110] This application also provides a computer-readable storage medium storing a computer program thereon, wherein the computer program, when executed by a processor / processor core, implements the aforementioned core power prediction method or core monitoring method. The computer-readable storage medium can be volatile or non-volatile.

[0111] This application also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code is run in the processor of an electronic device, the processor in the electronic device executes the above-described core power prediction method or core monitoring method.

[0112] Those skilled in the art will understand that all or some of the steps, systems, and apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software can be distributed on a computer-readable storage medium, which may include computer storage media (or non-transitory media) and communication media (or transient media).

[0113] As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable program instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technologies, portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is known to those skilled in the art that communication media typically contain computer-readable program instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0114] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0115] The computer program instructions used to perform the operations of this application may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuits, such as programmable logic circuits, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), are personalized by utilizing the status information of the computer-readable program instructions. These electronic circuits can execute the computer-readable program instructions to implement various aspects of this application.

[0116] The computer program product described herein can be implemented specifically through hardware, software, or a combination thereof. In one alternative embodiment, the computer program product is specifically embodied in a computer storage medium; in another alternative embodiment, the computer program product is specifically embodied in a software product, such as a software development kit (SDK), etc.

[0117] Various aspects of this application are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer-readable program instructions.

[0118] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that, when executed by the processor of the computer or other programmable data processing apparatus, they create means for implementing the functions / actions specified in one or more blocks of the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium that causes a computer, programmable data processing apparatus, and / or other device to operate in a particular manner; thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing aspects of the functions / actions specified in one or more blocks of the flowchart and / or block diagram.

[0119] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to perform the functions / actions specified in one or more boxes of a flowchart and / or block diagram.

[0120] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of an instruction containing one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0121] Example embodiments have been disclosed herein, and while specific terminology has been used, it is for general illustrative purposes only and should not be construed as limiting. In some instances, it will be apparent to those skilled in the art that features, characteristics, and / or elements described in conjunction with particular embodiments may be used alone, or in combination with features, characteristics, and / or elements described in conjunction with other embodiments, unless otherwise expressly indicated. Therefore, those skilled in the art will understand that various changes in form and detail may be made without departing from the scope of this application as set forth by the appended claims.

Claims

1. A method for predicting reactor core power, characterized in that, include: Acquire target power detector data for the target reactor core; the target power detector data includes multiple power detector values. The target power detector data is processed using a target power prediction model to obtain the predicted power value of the target core. The target power prediction model is trained using a data grouping method, utilizing historical power detector training data and their corresponding first historical power values.

2. The method according to claim 1, characterized in that, Before processing the target power detector data using the target power prediction model to obtain the predicted power value of the target reactor core, the method further includes: S1. Obtain a training set and a test set; the training set includes multiple training samples; the training samples include historical power detector training data and their corresponding first historical power values; the test set includes multiple test samples; the test samples include historical power detector test data and their corresponding second historical power values. S2. Based on the training samples corresponding to any two of the historical power detector training data, construct a first polynomial between any two of the historical power detector training data and the first historical power value. S3. Solve the first polynomial to obtain the first set of coefficients corresponding to the first polynomial; S4. Determine the e-th layer neuron set based on the first coefficient set and the corresponding training data of the two historical power detectors; the e-th layer neuron set includes multiple first neurons; e is the iteration number; S5. Calculate the first residual between each of the first neurons and the second historical power value using the test samples to obtain multiple first residuals; S6. Based on the multiple first residuals, at least one second neuron is selected from the set of neurons in the e-th layer to determine the set of neurons in the (e+1)-th layer. S7. Use each of the second neurons in the (e+1)th layer neuron set as training data for the historical power detector, and repeat steps S2 to S7 until the preset iteration stop condition is met. Use the first neuron with the smallest first residual in the eth layer neuron set after iteration as the target power prediction model.

3. The method according to claim 2, characterized in that, Solving the first polynomial to obtain the first set of coefficients corresponding to the first polynomial includes: The first polynomial is solved using the least squares method to obtain the first set of coefficients corresponding to the first polynomial.

4. The method according to claim 2, characterized in that, The second neuron is the first neuron whose first residual is less than or equal to a preset residual.

5. The method according to claim 2, characterized in that, The preset iteration stopping condition includes: the minimum value of each of the first residuals of the (e+1)th layer neuron set after iteration is greater than the minimum value of each of the first residuals of the eth layer neuron set after iteration.

6. The method according to claim 1, characterized in that, After processing the target power detector data using the target power prediction model to obtain the predicted power value of the target reactor core, the method further includes: Piecewise polynomial fitting is performed on the predicted power values ​​to obtain the predicted power distribution.

7. A core monitoring method, characterized in that, include: The core power prediction method according to any one of claims 1 to 6 determines the predicted power value of the target core; The target reactor core is monitored based on the predicted power value.

8. A reactor core power prediction device, characterized in that, include: The first acquisition module is used to acquire target power detector data of the target reactor core; the target power detector data includes multiple power detector values. The first processing module is used to process the target power detector data using the target power prediction model to obtain the predicted power value of the target core. The target power prediction model is trained using a data grouping method, utilizing historical power detector training data and their corresponding first historical power values.

9. The apparatus according to claim 8, characterized in that, The device further includes: The second acquisition module is used to execute step S1, which includes acquiring a training set and a test set; the training set includes multiple training samples; the training samples include historical power detector training data and their corresponding first historical power values; the test set includes multiple test samples; the test samples include historical power detector test data and their corresponding second historical power values. A construction module is used to execute step S2, which includes: constructing a first polynomial between the training data of any two historical power detectors and the first historical power value based on the training samples corresponding to any two historical power detector training data; The solution module is used to execute step S3, which includes: solving the first polynomial to obtain the first coefficient set corresponding to the first polynomial; The solution module is used to execute step S3, which includes: determining the set of neurons in the e-th layer based on the first coefficient set and the corresponding training data of the two historical power detectors; the set of neurons in the e-th layer includes multiple first neurons; e is the number of iterations; The calculation module is used to execute step S5, which includes: using the test sample to calculate the first residual between each of the first neurons and the second historical power value to obtain a plurality of first residuals; The first screening module is used to perform step S6, which includes: screening at least one second neuron from the set of neurons in the e-th layer based on a plurality of first residuals, so as to determine the set of neurons in the (e+1)-th layer. The first screening module is used to execute step S6, which includes: using each of the second neurons in the (e+1)th layer neuron set as training data for the historical power detector, repeating steps S2 to S7 until a preset iteration stop condition is met, and using the first neuron with the smallest first residual in the iterated eth layer neuron set as the target power prediction model.

10. A reactor core monitoring system, characterized in that, include: The core power prediction device according to claim 8 or 9 is used to determine the predicted power value of a target core. A monitoring device is used to monitor the target reactor core based on the predicted power value.

Images