Control test method and system for high-temperature gas cooled reactor burn-up measurement system

Abstract

The invention discloses a control test method and system for a high-temperature gas cooled reactor burn-up measurement system, and the method comprises the steps: adjusting an energy scale coefficient in a current environment through employing a self-adaptive energy scale algorithm after the measurement system is started; an edge cloud architecture is established to perform data acquisition and processing, measurement points independently acquire burnup data, key parameters are extracted and then uploaded to a central system, and pattern recognition and trend prediction are performed; and carrying out online calibration on the measurement system, recording all measurement and calibration data, and generating a detailed report. According to the method, the energy scale coefficient is adjusted through the self-adaptive energy scale algorithm, data acquisition and processing of the edge cloud architecture and online calibration and report generation are performed, so that comprehensive control and test of the high-temperature gas cooled reactor burnup measurement system are realized, the accuracy and reliability of the system are improved, real-time monitoring and analysis of burnup data are realized, and the real-time monitoring and analysis of the burnup data are realized. And important data support is provided for safe operation and maintenance of the nuclear power station.

Description

Technical Field

[0001] The invention relates to the field of nuclear energy science and engineering technology, and in particular to a control test method and system for a high temperature gas-cooled reactor burnup measurement system. Background Art

[0002] With the rapid development of nuclear energy technology, high temperature gas-cooled reactors, as an advanced form of nuclear energy utilization, have been widely used in my country's nuclear energy field. However, during the operation of high temperature gas-cooled reactors, the stability and accuracy of the burnup measurement system are crucial to the safe operation of the entire nuclear power plant. Therefore, the research on the control test method of the high temperature gas-cooled reactor burnup measurement system has important practical significance.

[0003] In the existing high-temperature gas-cooled reactor burnup measurement system, the traditional energy calibration method is generally used for burnup measurement. However, the traditional energy calibration method has certain limitations. First, the traditional energy calibration method cannot adapt to the changes in the energy calibration coefficient caused by environmental changes, which affects the accuracy of the burnup measurement. Secondly, the efficiency of data processing and transmission is low and cannot meet the needs of real-time monitoring. In addition, the existing burnup measurement system lacks effective fault diagnosis and prediction functions, and cannot warn of potential faults, which affects the stability and reliability of the system.

[0004] In response to the above problems, we have invented a control and test method for a high-temperature gas-cooled reactor fuel consumption measurement system. This method uses an adaptive energy calibration algorithm to automatically adjust the energy calibration coefficient and improve the accuracy of fuel consumption measurement. At the same time, through distributed computing and edge computing architecture, efficient data processing and transmission are achieved to meet the needs of real-time monitoring. In addition, the use of an intelligent fault diagnosis and prediction system enables real-time monitoring of measurement data, improving the stability and reliability of the system. Therefore, our invention belongs to the field of nuclear energy technology and solves the control and test problems of a high-temperature gas-cooled reactor fuel consumption measurement system. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is that the traditional burnup measurement method cannot adapt to the changes in the energy calibration coefficient caused by environmental changes, has poor real-time performance, and is not suitable for the burnup measurement of high-temperature gas-cooled reactors using a non-stop stacking and material replacement method.

[0007] To solve the above technical problems, the present invention provides the following technical solutions: a control and testing method for a high-temperature gas-cooled reactor fuel consumption measurement system, comprising: after the measurement system is turned on, an adaptive energy calibration algorithm is used to adjust the energy calibration coefficient under the current environment; an edge-cloud architecture is established for data collection and processing, and the measurement points independently collect fuel consumption data, extract key parameters and upload them to the central system for pattern recognition and trend prediction; the measurement system is calibrated online, all measurement and calibration data are recorded, and a detailed report is generated.

[0008] As a preferred solution of the control and test method for the high temperature gas-cooled reactor burnup measurement system of the present invention, the adaptive energy calibration algorithm includes: the measurement system automatically loads the energy calibration coefficient and calibration parameters of the previous time, selects n standard radioactive sources with known energy peak values, collects the γ energy spectrum data of the radioactive sources, identifies the photoelectric peak position in the γ energy spectrum through a fitting algorithm, and simultaneously fits the n photoelectric peaks to obtain the center position E0 of each photoelectric peak, which is expressed as:

[0009]

[0010] Where I(E) represents the intensity of the energy spectrum at energy E; I0 represents the peak intensity; and σ represents the peak width.

[0011] As a preferred solution of the control test method for the high temperature gas-cooled reactor burnup measurement system of the present invention, the adaptive energy calibration algorithm further includes: for each standard radiation source gamma spectrum data f i (x), select the calibration interval [a i , b i ], calculate the energy spectrum data and reference energy spectrum data g within the calibration interval i The normalization factor of (x) is expressed as:

[0012]

[0013] Assume first that the relationship between energy and channel address is linear, calculate the energy calibration coefficient of each calibration point, and take the weighted average of the energy calibration coefficients of multiple calibration points to obtain the initial energy calibration coefficient, which is expressed as:

[0014]

[0015] In actual measurement, nonlinearity is introduced to compensate for the deviation of the system and adjust the measurement results. The sine function is used as part of the nonlinear correction, which is expressed as:

[0016] E′=E k (1+β1sin( β 2E k ))

[0017] Among them, N i represents the integral of the energy spectrum data actually collected within the calibration interval; M i represents the integral of the reference spectrum data within the calibration interval; α i Represents the energy calibration factor of each calibration point; E k represents the initial energy calibration coefficient; E′ represents the energy after correction; β represents the coefficient for adjusting the correction amplitude and frequency.

[0018] As a preferred solution of the control test method for the high temperature gas-cooled reactor burnup measurement system described in the present invention, wherein: the edge cloud architecture includes multiple edge nodes and a central processing system, the edge node is responsible for processing the data collected by the measurement point, each edge node has a fixed measurement point for collecting standard radioactive sources, when the measurement system is turned on, all the fixed measurement points for collecting standard radioactive sources collect data of the standard radioactive sources, upload them to the central processing system to adjust the initial energy calibration coefficient, and periodically collect data of the standard radioactive sources during the operation of the measurement system for online calibration;

[0019] The burnup data is collected independently at each measurement point. The background data is subtracted from the collected γ spectrum data. The shape of each peak is fitted by a Gaussian function to determine the peak position of the γ spectrum, which is expressed as:

[0020]

[0021] After the peak position is identified, the net area of ​​the identified peak position is calculated, and the key parameter vectors of peak position, net area and energy are extracted and uploaded to the central control system. The data of the standard radioactive source is marked, and the collected data and the extracted key parameter vector are uploaded accordingly, which is expressed as:

[0022]

[0023] Among them, d j represents the count of energy channel j in the γ spectrum data; A represents the peak height; C represents the number of channels; μ represents the peak position; σ represents the peak width; d bg represents background data; NA represents the net area after peak identification; a represents the peak width coefficient; P i represents the key parameter vector; E j Represents the energy value of the jth energy channel.

[0024] As a preferred solution of the control test method for the high temperature gas-cooled reactor burnup measurement system of the present invention, the central processing system includes: i For multidimensional data analysis, use P i Construct the data matrix X, centralize the data matrix X to obtain X centered, kernel principal component analysis is introduced to process nonlinear data, Gaussian kernel is selected as kernel function, and kernel matrix is ​​calculated and centered, which is expressed as:

[0025]

[0026] K ij = k(P i , P j )

[0027] K centered =K-1 m K-K1 m +1 m K1 m

[0028] Calculate K centered The eigenvalues ​​and eigenvectors of the first d largest eigenvalues ​​are selected to form the transformation matrix W kpca , use the transformation matrix to transform X centered Transform to the new feature space and get the reduced dimension data matrix X kpca , expressed as:

[0029] X kpca =K centered W kpca

[0030] Where k(x,y) represents the kernel function; x represents X centered The eigenvector in ; y represents X centered The eigenvector in ; σ 2 represents the width of the Gaussian kernel; K represents the kernel matrix; K centered represents the centralized kernel matrix; 1 m Represents an m×m matrix.

[0031] As a preferred solution of the control and testing method for the high temperature gas-cooled reactor burnup measurement system of the present invention, the central processing system also includes using a support vector machine for pattern recognition, and the input of the support vector machine is X obtained after multi-dimensional data analysis. kpca , use historical data and corresponding labels to train the support vector machine model. The labels of the support vector machine are defined as four categories: normal mode, equipment failure, data anomaly, and system overload. The training process uses polynomial kernel functions and penalty parameters. By optimizing the classification boundary, the support vector machine can find the best segmentation plane in the high-dimensional space. The trained support vector machine is used for pattern recognition;

[0032] Trend identification is achieved through the long short-term memory network to predict the future trend of fuel consumption measurement. The architecture of the long short-term memory network consists of an output layer, a three-layer bidirectional LSTM layer, a fully connected layer and an output layer. The input data is Xkpca , a Dropout layer is added after each bidirectional LSTM layer and fully connected layer, L2 regularization is used to constrain the network weights to prevent overfitting, and layer normalization is added after each LSTM layer to accelerate convergence, X kpca After passing through three bidirectional LSTM layers in sequence, it is mapped to the prediction result through the fully connected layer

[0033] As a preferred solution of the control test method for the high temperature gas-cooled reactor burnup measurement system of the present invention, wherein: the online calibration is realized by collecting standard radioactive source data, and the data used for online calibration is P uploaded by the edge node i , without multi-dimensional data analysis, the central processing system processes the data P uploaded by each edge node i Aggregate with the scale position data P0 and calculate the normalized value t of the data difference between the measurement position and the scale position j0 and degrees of freedom v j0 , expressed as:

[0034]

[0035] v j0 =2(φ-1)

[0036] Compare the t-test critical value table to determine the corresponding degrees of freedom v j0 The confidence level is used to determine the consistency between each measurement position and the calibration position. If the consistency deviation exceeds the preset threshold, the correction factor and relative error are calculated and sent to the edge node for energy calibration coefficient calibration, which is expressed as:

[0037]

[0038] Among them, P0 represents the data of the scale position; φ represents the number of measurements; f j0 represents the correction factor; represents the mean count rate; P j0 Indicates relative error.

[0039] In the second aspect, the present invention also provides a control and testing system for a high-temperature gas-cooled reactor fuel consumption measurement system, including an edge module, an edge end measurement point collects data from standard radioactive sources and non-standard radioactive sources, the measurement point uploads the collected data to the edge node, the edge node extracts the peak position, net area, and energy composition key parameter vectors and uploads them to the central control system; a central processing module, after startup, uses an adaptive energy calibration algorithm to calculate the initial energy calibration coefficient based on the data of the standard radioactive source, periodically performs online calibration during the test, and sends the calibrated coefficient to the edge node for adjustment. For the data of non-standard radioactive sources, the dimensionality is reduced through multi-dimensional data analysis, and the data is analyzed through pattern recognition and trend prediction; a report generation module records all measurement and calibration data, and generates a detailed report based on the results of pattern recognition and trend prediction.

[0040] In a third aspect, the present invention further provides a computing device, comprising: a memory and a processor;

[0041] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the control test method for the high temperature gas-cooled reactor burnup measurement system are implemented.

[0042] In a fourth aspect, the present invention further provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the control test method for a high temperature gas-cooled reactor burnup measurement system.

[0043] Beneficial effects of the invention: The method of the invention adjusts the energy calibration coefficient through an adaptive energy calibration algorithm, data collection and processing of the edge-cloud architecture, and online calibration and report generation, thereby realizing comprehensive control and testing of the high-temperature gas-cooled reactor fuel consumption measurement system, improving the accuracy and reliability of the system, and realizing real-time monitoring and analysis of fuel consumption data, providing important data support for the safe operation and maintenance of nuclear power plants. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0045] Figure 1 An overall flow chart of a control test method for a high temperature gas-cooled reactor burnup measurement system provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0047] Example 1

[0048] Reference Figure 1 , as one embodiment of the present invention, provides a control test method for a high temperature gas-cooled reactor burnup measurement system, comprising:

[0049] S1: After the measurement system is turned on, the energy calibration coefficient under the current environment is adjusted using the adaptive energy calibration algorithm.

[0050] Furthermore, the measurement system performs a self-check after being powered on to ensure that all components are working properly, automatically loads the last energy calibration coefficient and calibration parameters, selects n standard radiation sources with known energy peaks through the measurement points of the standard radiation sources, collects the gamma energy spectrum data of the radiation sources, and uploads them to the central processing system.

[0051] The adaptive energy calibration algorithm is implemented in the central processing system. The photoelectric peak position in the γ energy spectrum is identified by the fitting algorithm. The n photoelectric peaks are fitted simultaneously to obtain the center position E0 of each photoelectric peak, which is expressed as:

[0052]

[0053] Where I(E) represents the intensity of the energy spectrum at energy E; I0 represents the peak intensity; and σ represents the peak width.

[0054] It should be noted that the center position of the photoelectric peak is the basis of energy calibration. In the multi-point calibration process, the detector's response to different energies is determined by identifying the center position of the photoelectric peak of different standard radiation sources. The precise identification and fitting of multiple photoelectric peak center positions can make subsequent energy calibration more accurate. Through multi-point calibration, the information of known energy points is combined to reduce the error caused by single-point calibration.

[0055] Furthermore, the initial energy calibration coefficient is calculated through multi-point calibration. The purpose of multi-point calibration is to calibrate the measurement system through multiple sets of data from known reference points to ensure the accuracy of the energy calibration. Multiple known radiation sources are used to obtain their energy spectrum data, and the initial energy calibration coefficient is calculated based on these data.

[0056] For each standard radioactive source, the γ spectrum data f i(x), select the calibration interval [a i , b i ], the energy spectrum signal is more significant in the calibration interval, and the energy spectrum data and reference energy spectrum data g are calculated in the calibration interval. i The normalization factor of (x) is expressed as:

[0057]

[0058] Assume first that the relationship between energy and channel address is linear, calculate the energy calibration coefficient of each calibration point, and take the weighted average of the energy calibration coefficients of multiple calibration points to obtain the initial energy calibration coefficient, which is expressed as:

[0059]

[0060] In actual measurement, the system response will have nonlinear deviations. By introducing nonlinearity to compensate for the deviation of the system, the measurement results are adjusted and the sine function is used as part of the nonlinear correction, which is expressed as:

[0061] E′=E k (1+β1sin(β2E k ))

[0062] Among them, N i represents the integral of the energy spectrum data actually collected within the calibration interval; M i represents the integral of the reference spectrum data within the calibration interval; α i Represents the energy calibration factor of each calibration point; E k represents the initial energy calibration coefficient; E′ represents the energy after correction; β represents the coefficient for adjusting the correction amplitude and frequency.

[0063] S2: Establish an edge-cloud architecture for data collection and processing. The measurement points independently collect fuel consumption data, extract key parameters and upload them to the central system for pattern recognition and trend prediction.

[0064] Furthermore, the edge-cloud architecture includes multiple edge nodes and a central processing system. The edge nodes are responsible for processing the data collected by the measurement points. Each edge node has a fixed measurement point for collecting standard radioactive sources. When the measurement system is turned on, all the measurement points for collecting standard radioactive sources collect data from the standard radioactive sources and upload the data to the central processing system to adjust the initial energy calibration coefficient. During the operation of the measurement system, the data of the standard radioactive sources are periodically collected for online calibration.

[0065] It should be noted that the data collected at the measurement point mainly include the following: 1. Gamma-ray energy spectrum data directly measured using detectors such as high-purity germanium spectrometers, including the count rates of rays of different energies. 2. For specific radioactive isotopes, the full energy peak count rate of gamma rays is obtained by direct measurement. 3. Background count rate, that is, the count rate of background radiation recorded by the measurement system in the absence of any radioactive source.

[0066] The burnup data is collected independently at each measurement point. The background data is subtracted from the collected γ spectrum data. The shape of each peak is fitted by a Gaussian function to determine the peak position of the γ spectrum, which is expressed as:

[0067]

[0068] Peak position identification is the basis for extracting key parameters. By accurately locating the peak position of the radioactive isotope in the gamma energy spectrum, an accurate reference point is provided for subsequent net area calculation and key parameter extraction.

[0069] After the peak position is identified, the net area of ​​the identified peak position is calculated. Based on the result of the peak position identification, the relative content of the radioactive isotope is obtained by calculating the net count within a certain range around the peak position, which is expressed as:

[0070]

[0071] The key parameter vectors of peak position, net area and energy composition are extracted and uploaded to the central control system. The data of the standard radioactive source is marked and the collected data and the extracted key parameter vectors are uploaded accordingly, which is expressed as:

[0072]

[0073] Among them, d j represents the count of energy channel j in the γ spectrum data; A represents the peak height; C represents the number of channels; μ represents the peak position; σ represents the peak width; d bg represents background data; NA represents the net area after peak identification; a represents the peak width coefficient; P i represents the key parameter vector; E j Represents the energy value of the jth energy channel.

[0074] It should be noted that background subtraction is achieved by subtracting background data from the collected data, eliminating background interference caused by factors such as environmental noise and detector response characteristics, thereby improving the accuracy and reliability of subsequent analysis.

[0075] Furthermore, the edge node transmits key parameter data to the central control system through a high-speed transmission channel. The central processing system first processes the P of the non-standard radioactive source uploaded by the edge node. iFor multidimensional data analysis, use P i Construct the data matrix X.

[0076]

[0077] Calculate the mean vector μ of the data matrix and transform each key parameter vector P i Center the data matrix X by subtracting the mean vector to obtain X centered , expressed as:

[0078]

[0079] X centered =X-μ

[0080] Kernel principal component analysis is introduced to process nonlinear data. Gaussian kernel is selected as the kernel function. The kernel function is a nonlinear mapping that maps the original data points x and y to a high-dimensional space. The similarity between x and y in the high-dimensional space is represented by the inner product. The kernel matrix is ​​calculated and centered, which is expressed as:

[0081]

[0082] K ij = k(P i , P j )

[0083] K centered =K-1 m K-K1 m +1 m K1 m

[0084] Calculate K centered The eigenvalues ​​and eigenvectors of the first d largest eigenvalues ​​are selected to form the transformation matrix W kpca , use the transformation matrix to transform X centered Transform to the new feature space and get the reduced dimension data matrix X kpca , expressed as:

[0085] X kpca =K centered W kpca

[0086] Where k(x, y) represents the kernel function; x represents X centered The eigenvector in ; y represents X centered The eigenvector in ; σ 2 represents the width of the Gaussian kernel; K represents the kernel matrix; K centered represents the centralized kernel matrix; 1 m Represents an m×m matrix, each element is 1 / m. The reduced data Xkpca It contains the main features of the original data and removes redundant information.

[0087] After completing the multidimensional data analysis, support vector machine is used for pattern recognition. The input of support vector machine is X obtained after multidimensional data analysis. kpca , the support vector machine model is trained using historical data and corresponding labels. The labels of the support vector machine are defined as four categories: normal mode, equipment failure, data anomaly, and system overload. The training process uses polynomial kernel functions and penalty parameters. By optimizing the classification boundary, the support vector machine can find the best segmentation plane in high-dimensional space. The optimization problem is expressed as:

[0088]

[0089] subject to you i (w·φ(x i )+b)≥1-ξi,ξi≥0

[0090] Use the trained support vector machine for pattern recognition, input the reduced-dimensional data into the support vector machine model, and obtain the predicted label through model calculation.

[0091]

[0092] Among them, φ(x i ) represents the polynomial kernel function mapping; ξ i represents the slack variable; C represents the penalty parameter; α i represents the Lagrange multiplier; y i Represents the predicted label.

[0093] Equipment failure is caused by sensor failure or connection problems, which is manifested in the long-term deviation of certain sensor readings from the normal range; data anomalies are problems in the transmission from edge nodes to the central processing system, signal loss or interference during data transmission; system overload is caused by insufficient resources of the central processing system or exceeding the design load; when an abnormal pattern is detected, an alarm is issued and staff are dispatched to conduct detection and repair based on the abnormal label.

[0094] It should be noted that the fuel consumption data has a strong nonlinear characteristic, so the kernel function of the support vector machine selects the polynomial kernel function instead of the Gaussian kernel function. In the experiment of the fuel consumption measurement system, the fuel consumption data is often affected by multiple factors and shows obvious nonlinear characteristics. The polynomial kernel function can flexibly model these nonlinear relationships by adjusting the order of the polynomial, thereby improving the classification and prediction capabilities of the support vector machine.

[0095] Furthermore, trend identification is achieved through the long short-term memory network to predict the future trend of fuel consumption measurement. The architecture of the long short-term memory network consists of an output layer, a three-layer bidirectional LSTM layer, a fully connected layer and an output layer. The input data is X kpca , a Dropout layer is added after each bidirectional LSTM layer and fully connected layer, and L2 regularization is used to constrain the network weights to prevent overfitting. A layer normalization is added after each LSTM layer to accelerate convergence. The input of the first bidirectional LSTM layer is X kpca The output value H1 is used as the input of the second bidirectional LSTM layer. After passing through three bidirectional LSTM layers in sequence, it is mapped to the prediction result through the fully connected layer.

[0096] It should be noted that the first layer of bidirectional LSTM processes the input multidimensional data matrix X kpca , extract short-term and long-term features from the sequence data through LSTM units in both the forward and backward directions; the second layer of bidirectional LSTM receives the feature matrix H1 output by the first layer, and processes it again through the bidirectional LSTM unit to further extract deeper features; the third layer of bidirectional LSTM receives the feature matrix H2 output by the second layer, performs the last layer of bidirectional LSTM processing, and finally extracts the deepest features, and outputs H3 to the fully connected layer to obtain the prediction result

[0097] The bidirectional LSTM can simultaneously consider the contextual information of the sequence data, allowing the model to capture more comprehensive features and enhance the accuracy of the prediction. The three-layer LSTM structure extracts features at different levels layer by layer to fully mine the information in the data. At the same time, the three-layer LSTM processes the nonlinear relationship in the data and adapts to the complex changing trends in the fuel consumption data.

[0098] For the prediction results when When , it means the predicted value is on an upward trend, increase the inspection frequency and monitor whether there are potential problems that lead to increased fuel consumption; when When , it means that the predicted value remains stable, the system runs stably and no additional adjustment is required; when , it means that the predicted value shows a continuous downward trend. Trend prediction is a relatively long-term process. Only when the trend continues to rise or fall over a period of time can it be judged as a trend change. Trend fluctuations within a short period of time are acceptable within the normal range.

[0099] S3: Perform online calibration of the measurement system, record all measurement and calibration data, and generate detailed reports.

[0100] Furthermore, the online calibration is realized by collecting standard radioactive source data, and the data used for online calibration is P uploaded by the edge node. iWithout multi-dimensional data analysis, the central processing system collects the total energy peak net count N, error σ N , measuring time t and dead time τ to calculate the full-energy peak net count rate, error, relative error and count rate mean, the full-energy peak net count rate n:

[0101]

[0102] Error σ n :

[0103]

[0104] Relative error η n :

[0105]

[0106] Count rate mean

[0107]

[0108] The central processing system processes the data P uploaded by each edge node i Aggregate with the scale position data P0 and calculate the normalized value t of the data difference between the measurement position and the scale position j0 and degrees of freedom v j0 , expressed as:

[0109]

[0110] v j0 =2(φ-1)

[0111] Compare the t-test critical value table to determine the corresponding degrees of freedom v j0 The confidence level is used to determine the consistency between each measurement position and the calibration position. If the consistency deviation exceeds the preset threshold, the correction factor and relative error are calculated and sent to the edge node for energy calibration coefficient calibration, which is expressed as:

[0112]

[0113] Among them, P0 represents the data of the scale position; φ represents the number of measurements; f j0 represents the correction factor; represents the mean count rate; P j0 Indicates relative error.

[0114] The central processing system calculates the calibration data fj0 and P j0 After receiving the calibration data, each edge node adjusts the system parameters and applies the correction factor fj0 and the relative error P j0Calibrate the energy scale factor.

[0115] The present embodiment also provides a control test system for a high-temperature gas-cooled reactor burnup measurement system, including an edge module, wherein the edge end measurement point collects data from standard radioactive sources and non-standard radioactive sources, and the measurement point uploads the collected data to the edge node, and the edge node extracts the peak position, net area, and energy composition key parameter vectors and uploads them to the central control system; the central processing module uses an adaptive energy calibration algorithm after startup to calculate the initial energy calibration coefficient based on the data of the standard radioactive source, periodically performs online calibration during the test, and sends the calibrated coefficient to the edge node for adjustment. For the data of the non-standard radioactive source, the dimensionality is reduced through multi-dimensional data analysis, and the data is analyzed through pattern recognition and trend prediction; a report generation module records all measurement and calibration data, and generates a detailed report based on the results of pattern recognition and trend prediction.

[0116] This embodiment also provides a computing device, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the control test method for the high temperature gas-cooled reactor burnup measurement system proposed in the above embodiment.

[0117] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the control test method for a high temperature gas-cooled reactor burnup measurement system as proposed in the above embodiment is implemented.

[0118] The storage medium proposed in this embodiment and the control test method for the high temperature gas-cooled reactor burnup measurement system proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0119] Through the above description of the implementation method, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccessMemory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0120] Example 2

[0121] The following is an embodiment of the present invention, which provides a control test method for a high temperature gas-cooled reactor burnup measurement system. In order to verify the beneficial effects of the present invention, a scientific demonstration is carried out through simulation experiments.

[0122] A simulation experiment was conducted using historical data from a high-temperature gas-cooled reactor burnup measurement system to simulate the processes of initial energy scale calibration, trend analysis, pattern recognition, and online calibration. The t-test critical value table used in online calibration is shown in the following table.

[0123] Table 1 t test critical value table

[0124]

[0125] Experiments were conducted using data from two standard radioactive sources, cobalt-60 and cesium-137, and four non-standard radioactive sources. Data such as peak height and net area were calculated, and the calculated errors were analyzed. For the convenience of calculating the data, the experimental data are shown in the following table.

[0126] Table 2 Experimental data table

[0127]

[0128]

[0129] It can be seen from the experimental data that the measurement accuracy of the present invention is very high, the peak position deviations of standard radiation source 1 and standard radiation source 2 are both within a reasonable range, the correction factors are 1.02 and 1.01 respectively, and the relative errors are small, which shows the adaptive ability of the system under different standard radiation source conditions.

[0130] Under the edge-cloud architecture, edge nodes can independently collect the burnup data of non-standard radioactive sources and extract key parameters to upload to the central system, making data transmission and processing efficient and stable. The peak position of the energy spectrum is fitted by Gaussian function to determine the peak height and peak width, and the calculation of the net area also shows high accuracy, thus ensuring the accuracy of the energy spectrum data.

[0131] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A control test method for a high temperature gas-cooled reactor burnup measurement system, characterized in that: include: After the measurement system is turned on, the energy calibration coefficient is adjusted in the current environment using the adaptive energy calibration algorithm; Establish an edge-cloud architecture for data collection and processing. The measurement points independently collect fuel consumption data, extract key parameters and upload them to the central system for pattern recognition and trend prediction. Perform online calibration of the measuring system, record all measurement and calibration data and generate detailed reports.

2. The control and testing method for a high temperature gas-cooled reactor burnup measurement system according to claim 1, characterized in that: The adaptive energy calibration algorithm includes: the measurement system automatically loads the energy calibration coefficient and calibration parameters of the last time, selects n standard radioactive sources with known energy peaks, collects the γ energy spectrum data of the radioactive sources, identifies the photoelectric peak position in the γ energy spectrum through a fitting algorithm, and simultaneously fits the n photoelectric peaks to obtain the center position E0 of each photoelectric peak, which is expressed as: Where I(E) represents the intensity of the energy spectrum at energy E; I0 represents the peak intensity; and σ represents the peak width.

3. The control and testing method for a high temperature gas-cooled reactor burnup measurement system according to claim 2, characterized in that: The adaptive energy calibration algorithm also includes, for each standard radiation source gamma energy spectrum data f i (x), select the calibration interval [a i , b i ], calculate the energy spectrum data and reference energy spectrum data g within the calibration interval i The normalization factor of (x) is expressed as: Assume first that the relationship between energy and channel address is linear, calculate the energy calibration coefficient of each calibration point, and take the weighted average of the energy calibration coefficients of multiple calibration points to obtain the initial energy calibration coefficient, which is expressed as: In actual measurement, nonlinearity is introduced to compensate for the deviation of the system and adjust the measurement results. The sine function is used as part of the nonlinear correction, which is expressed as: E′=E k (1+β1sin(β2E k )) Among them, N i represents the integral of the energy spectrum data actually collected within the calibration interval; M i represents the integral of the reference spectrum data within the calibration interval; α i Indicates the energy calibration factor of each calibration point; E k represents the initial energy calibration coefficient; E′ represents the energy after correction; β represents the coefficient for adjusting the correction amplitude and frequency.

4. The control and testing method for a high temperature gas-cooled reactor burnup measurement system according to claim 3, characterized in that: The edge cloud architecture includes multiple edge nodes and a central processing system. The edge node is responsible for processing the data collected by the measurement point. Each edge node has a fixed measurement point for collecting standard radioactive sources. When the measurement system is turned on, all the fixed measurement points for collecting standard radioactive sources collect data from the standard radioactive sources and upload them to the central processing system to adjust the initial energy calibration coefficient. During the operation of the measurement system, the data of the standard radioactive source is periodically collected for online calibration. The burnup data is collected independently at each measurement point. The background data is subtracted from the collected γ spectrum data. The shape of each peak is fitted by a Gaussian function to determine the peak position of the γ spectrum, which is expressed as: After the peak position is identified, the net area of ​​the identified peak position is calculated, and the key parameter vectors of peak position, net area and energy are extracted and uploaded to the central control system. The data of the standard radioactive source is marked, and the collected data and the extracted key parameter vector are uploaded accordingly, which is expressed as: Among them, d j represents the count of energy channel j in the γ spectrum data; A represents the peak height; C represents the number of channels; μ represents the peak position; σ represents the peak width; d bg represents background data; NA represents the net area after peak identification; a represents the peak width coefficient; P i represents the key parameter vector; E j Represents the energy value of the jth energy channel.

5. The control and testing method for a high temperature gas-cooled reactor burnup measurement system according to claim 4, characterized in that: The central processing system includes: i For multidimensional data analysis, use P i Construct the data matrix X, centralize the data matrix X to obtain X centered , kernel principal component analysis is introduced to process nonlinear data, Gaussian kernel is selected as kernel function, and kernel matrix is ​​calculated and centered, which is expressed as: K ij =k(P i ,P j ) K centered =K-1 m K-K1 m +1 m K1 m Calculate K centered The eigenvalues ​​and eigenvectors of the first d largest eigenvalues ​​are selected to form the transformation matrix W kpca , use the transformation matrix to transform X centered Transform to the new feature space and get the reduced dimension data matrix X kpca , expressed as: X kpca =K centered W kpca Where k(x, y) represents the kernel function; x represents X centered The eigenvector in ; y represents X centered The eigenvector in ; σ 2 represents the width of the Gaussian kernel; K represents the kernel matrix; K centered represents the centralized kernel matrix; 1 m Represents an m×m matrix.

6. The control and testing method for a high temperature gas-cooled reactor burnup measurement system according to claim 5, characterized in that: The central processing system also includes using a support vector machine for pattern recognition, the input of the support vector machine is X obtained after multi-dimensional data analysis. kpca , use historical data and corresponding labels to train the support vector machine model. The labels of the support vector machine are defined as four categories: normal mode, equipment failure, data anomaly, and system overload. The training process uses polynomial kernel functions and penalty parameters. By optimizing the classification boundary, the support vector machine can find the best segmentation plane in the high-dimensional space. The trained support vector machine is used for pattern recognition; Trend identification is achieved through the long short-term memory network to predict the future trend of fuel consumption measurement. The architecture of the long short-term memory network consists of an output layer, a three-layer bidirectional LSTM layer, a fully connected layer and an output layer. The input data is X kpca , a Dropout layer is added after each bidirectional LSTM layer and fully connected layer, L2 regularization is used to constrain the network weights to prevent overfitting, and layer normalization is added after each LSTM layer to accelerate convergence, X kpca After passing through three bidirectional LSTM layers in sequence, it is mapped to the prediction result through the fully connected layer 7. The control and testing method for a high temperature gas-cooled reactor burnup measurement system according to claim 6, characterized in that: The online calibration is achieved through the collected standard radioactive source data. The data used for online calibration is the P uploaded by the edge node. i , without multi-dimensional data analysis, the central processing system processes the data P uploaded by each edge node i Aggregate with the scale position data P0 and calculate the normalized value t of the data difference between the measurement position and the scale position j0 and degrees of freedom v j0 , expressed as: v j0 =2(φ-1) Compare the t-test critical value table to determine the corresponding degrees of freedom v j0 The confidence level is used to determine the consistency between each measurement position and the calibration position. If the consistency deviation exceeds the preset threshold, the correction factor and relative error are calculated and sent to the edge node for energy calibration coefficient calibration, which is expressed as: Among them, P0 represents the data of the scale position; φ represents the number of measurements; f j0 represents the correction factor; represents the mean count rate; P j0 Indicates relative error.

8. A control and testing system for a high temperature gas-cooled reactor burnup measurement system using the method according to any one of claims 1 to 7, characterized in that: include, Edge module: The edge measurement point collects data from standard radioactive sources and non-standard radioactive sources. The measurement point uploads the collected data to the edge node. The edge node extracts the key parameter vectors of peak position, net area, and energy and uploads them to the central control system. The central processing module uses an adaptive energy calibration algorithm after startup to calculate the initial energy calibration coefficient based on the data of the standard radioactive source. It periodically performs online calibration during the test process and sends the calibrated coefficient to the edge node for adjustment. For the data of non-standard radioactive sources, it reduces the dimension through multi-dimensional data analysis and analyzes the data through pattern recognition and trend prediction. The report generation module records all measurement and calibration data and generates detailed reports based on the results of pattern recognition and trend prediction.

9. A computing device comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.

Images