A model for estimating the coolant temperature in the main pipes of a nuclear reactor, taking into account measurement errors.

Abstract

This invention discloses a model for estimating the temperature of coolant in the main pipeline of a nuclear reactor, considering measurement errors. Belonging to the field of nuclear power plant temperature measurement technology, the model includes the following steps: acquiring actual temperature measurement data from the power plant and establishing a dataset corresponding to the measured data and agreed-upon true values; analyzing the actual temperature measurement process and classifying measurement errors; considering different error distribution characteristics, combining a data-driven algorithm with a stochastic model to establish a temperature estimation model; designing a solution method for estimating prior information and calculating model coefficients in the temperature estimation model; optimizing and solving the temperature estimation model using the created dataset, and verifying its accuracy and reliability. This invention solves the problems of difficulty in detailed study of existing measurement errors and difficulty in solving for the true distribution of coolant temperature in nuclear pipelines.

Description

Technical Field

[0001] This invention belongs to the field of nuclear power plant temperature measurement technology, and in particular relates to the establishment of a model for estimating the temperature of the coolant in the main pipeline of a nuclear reactor, taking into account measurement errors. Background Technology

[0002] A nuclear power reactor is a high-power-density, high-operation-parameter, and high-safety-requirement device that rapidly and efficiently converts nuclear, thermal, and kinetic energy. Therefore, timely and accurate monitoring of the nuclear energy release and heat transfer status of the reactor core is crucial for ensuring the safe and reliable operation of the reactor. Based on the structure and operating principle of pressurized water reactors, the temperature of the coolant in the hot section of the main pipeline of the nuclear reactor system directly reflects the nuclear power and core heat transfer status, making it a core parameter for nuclear reactor power control and safety protection. If the primary loop temperature is too low, the power generation demand of the nuclear power plant cannot be met; if the primary loop temperature is too high, it may lead to fuel cladding damage or even fuel pellet meltdown, endangering the safety of the power plant. Therefore, accurately estimating the true temperature of the coolant in the main pipeline of the nuclear reactor is of significant practical importance and urgently requires relevant research.

[0003] To ensure the safe and reliable operation of nuclear power plants, measurement uncertainty analysis is used to measure the closeness of measured values ​​to the true values. This method traces the sources of uncertainty in a measurement parameter, synthesizes all uncertainties to obtain the overall uncertainty of the parameter, and then simply adds the calculated overall uncertainty to the nuclear power plant's design value to obtain a numerical range. It is assumed that adding the measured value to this range yields the true temperature distribution range. Traditional measurement uncertainty analysis methods assume that the true value of the measured parameter is unavailable and can only compensate for the difference between the measured and true values ​​by tracing the source of error terms. This approach suffers from problems such as difficulty in tracing the source, over-reliance on engineering experience, and a lack of accurate assessment of the measured parameter.

[0004] This invention aims to more accurately estimate the true temperature of the coolant in the main pipes of a nuclear reactor by using a data-driven model to directly invert and construct the actual temperature field based on actual measurement data. However, existing research has not carefully considered and analyzed the measurement errors carried in the measured values ​​when processing actual measurement data. Summary of the Invention

[0005] To address the aforementioned shortcomings in existing technologies, this invention provides a nuclear reactor main pipe coolant temperature estimation model that considers measurement errors. Based on acquired actual temperature measurement data from power plants, a dataset corresponding to the measurement data and agreed-upon true values ​​is established. By analyzing the actual temperature measurement process and classifying measurement errors, and considering different error distribution characteristics, a data-driven algorithm is combined with a stochastic model to establish a temperature estimation model and design a solution method. This enables the inversion of the true temperature value, solving the problems of difficulty in estimating the true value of the nuclear reactor main pipe coolant temperature and the relatively conservative assessment of the accuracy of measurement values.

[0006] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0007] This invention provides a model for estimating the coolant temperature in the main pipes of a nuclear reactor, taking into account measurement errors, comprising the following steps:

[0008] S1. Obtain actual temperature measurement data from the power plant and establish a dataset that corresponds to the measurement data and the agreed true values;

[0009] S2. By analyzing the actual temperature measurement process, the measurement errors are classified and processed.

[0010] S3. Considering different error distribution characteristics, combine the data-driven algorithm with the stochastic model to establish a temperature estimation model;

[0011] S4. Design a solution method for estimating prior information and calculating model coefficients in the temperature estimation model;

[0012] S5. Optimize and solve the temperature estimation model using the dataset in S1, and verify its accuracy and reliability.

[0013] The beneficial effects of this invention are as follows: The coolant temperature estimation model for the main pipeline of a nuclear reactor that considers measurement errors provided by this invention differs from the traditional method of treating measurement errors as a normal distribution. This invention refines and classifies the measurement errors present in the measurement process and makes assumptions for the different statistical characteristics of each type of error, thus alleviating the problem of conservative estimation that is prone to occur in traditional uncertainty assessment methods. The input used in this invention is actual measurement data from a nuclear power plant, and the probabilistic model improves the physical interpretability of the data-driven model, provides a reasonable explanation for the "many-to-many" functional relationship between the measured value and the true value, and effectively combines it with practical engineering applications.

[0014] Furthermore, the creation of the dataset in S1 includes the following steps:

[0015] S11. Actual measurement error is the difference between the measured value and the design value (theoretical calculation value), obtained from the working condition. Design value of temperature below ;

[0016] S12. Serialize the measured values ​​within the power plant over a period of time. With design value Subtracting them yields the measured error sequence. The statistical characteristics of the error magnitude were analyzed to obtain the measured error distribution. ;

[0017] S13. Based on the fluctuations in the actual operating conditions of the nuclear reactor, determine the range of temperature design value fluctuations. ;

[0018] S14. Based on the temperature range in S13 and the continuity of temperature change, the generation time length is... Temperature Convention True Value Sequence ;

[0019] S15. Based on the measured error distribution obtained in S12 to conduct Second-rate Sampling of an error sequence, and then comparing it with the sequence described in S14 Adding them together, we can get A series of measurement data sequences, each of which can be represented as: .

[0020] The beneficial effects of adopting the above-mentioned further scheme are as follows: Based on the actual temperature measurement data of the power plant and the design temperature values ​​in the design manual, this invention establishes a dataset corresponding to the measured temperature values ​​and the true values. In the creation of the dataset, this invention uses actual measured data from the power plant and fully considers the randomness of errors, establishing a "many-to-many" functional relationship between the true temperature values ​​and the measured values, making the method highly reliable and accurate.

[0021] Furthermore, the error analysis and classification processing method of S2 includes the following:

[0022] S21. The measurement data itself contains two parts: the true value and the measurement error.

[0023] The magnitude of measurement error is not a constant and has a certain degree of randomness. The statistical characteristics of errors from different sources are different, making it difficult to describe them quantitatively in a unified way.

[0024] The overall measurement error is divided into two categories: steady-state error and transient error. Steady-state error possesses stable statistical characteristics, and its distribution is independent of time. According to the uncertainty calculation standard for the main pipeline of a nuclear reactor, the error characteristics of instruments, processes, etc., can be described by white noise, and therefore can be considered as steady-state error. Transient error is defined as an error whose distribution is related to time. Observing the entire measurement process reveals that the temperature change in the main pipeline of the nuclear reactor is also time-dependent; therefore, this invention argues that transient error originates from changes in the actual temperature. Thus, a preliminary estimate of the current temperature can be made, and the transient error at that moment can be extracted using this prior information.

[0025] The beneficial effects of adopting the above-mentioned further solution are as follows: In the modeling process using data-driven methods, errors carried by the actual measurement data as input can lead to ill-conditioned predictions in the model. Existing research on this issue is limited, with most studies simply treating the error as a normal distribution. This invention refines the classification of errors present in measurements and describes the nature of errors differently based on different statistical characteristics, which helps to better achieve accurate estimation of the coolant temperature in the main pipes of nuclear reactors.

[0026] Furthermore, the temperature estimation model of S3 has the following characteristics:

[0027] S31. Because measurement data itself is not only random but also highly time-dependent, the commonly used model for processing random time series is the Autoregressive Moving Average (ARMA) model. This model assumes that the label values ​​fluctuate around a major trend over time, where the trend is influenced by historical labels, and the fluctuations are influenced by random events within a certain period, and the major trend itself is not necessarily stable. The temperature estimation model (NARMA) proposed in this invention draws on the basic idea of ​​this model.

[0028] The temperature estimation model described in this invention comprises three terms. The first term assumes that the agreed-upon true value at the current moment is influenced by measurements taken over a period of time (including past, present, and future times), thus compensating for the lack of consideration for the influence of future data on the current value in traditional ARMA. The second term is transient error information, which can be obtained through prior information about the temperature field. The third term is steady-state error information, consistent with the white noise information described in ARMA. The mathematical expression of this model can be written as:

[0029] in, Is The conventional true value of the time-time model estimation may take different values ​​under the same input due to input uncertainty. In the time period The sequence of measured values ​​within; It is a measurement coefficient matrix, which represents the influence of the measured values ​​on the agreed true value; In the time period Prior information within is... A preliminary estimate of the temperature value at each moment within the time frame, which is: The function; It is the transient error coefficient matrix, used to measure the impact of transient errors on the conventional truth value; It is the steady-state error during the measurement process, and its statistical characteristics are consistent with those of white noise; It is a random error coefficient matrix, used to measure the impact of steady-state error on the agreed-upon true value; since the error is random, and The elements in the array are not fixed values, but random variables that follow a certain distribution.

[0030] The beneficial effects of adopting the above-mentioned further scheme are as follows: In order to realize the inversion of the real temperature field based on actual temperature measurement data, combined with the ARMA concept, the components of the true temperature value are divided into the contribution of the measured value and the contribution of the error term, and the basic framework of the temperature estimation model is built, laying the foundation for the subsequent model establishment and solution.

[0031] Furthermore, the solution process for the temperature estimation model of S4 consists of the following steps:

[0032] S41. Use the Attention-LSTM model to estimate prior information;

[0033] S42. Solve the model coefficients using the FCNN-BN model.

[0034] Furthermore, the Attention-LSTM model in S41 estimates prior information by including the following steps:

[0035] S411, Collect n resistance temperature measurement data at time t. The data is fed into the attention mechanism module to obtain data carrying the location information of other measurement points. ;

[0036] S412, update n inputs By feeding the data into an LSTM neural network structure, prior estimates of the temperature at n measurement points at time t can be obtained.

[0037] S413, the time length is By repeating the above steps with the measured data, prior temperature estimates can be obtained. .

[0038] The beneficial effects of adopting the above-mentioned further scheme are as follows: Experiments have shown that spatial correlations in the dataset carry a large amount of information. In the Attention-LSTM model designed in this invention, the spatial correlations between several measurement points are first calculated through a self-attention mechanism to replace the original input, and then the new input is passed to the LSTM unit. When the LSTM module calculates its temporal correlations, since the input itself carries spatial information about the locations of other measurement points, a secondary extraction of spatial characteristics is achieved, ensuring that the model fully mines the spatiotemporal data features.

[0039] Furthermore, the FCNN-BN model in step S42 solves for the model coefficients by including the following steps:

[0040] S421. Calculate the n resistance temperature measurement data at time t. The data is fed into a fully connected neural network (FCNN) and processed through weights. Calculations yielded ;

[0041] Will Feed it into a Bayesian neural network (BN) and process it. The characteristics are assumed (usually that they follow a normal distribution with a mean of 0 and a variance of 1), and weights are used to determine their properties. , Calculations yielded ;

[0042] Will and Adding them together gives the final result. The final result is fed into the loss function to train and optimize the model parameters. , , .

[0043] The beneficial effects of adopting the above-mentioned further scheme are as follows: for the constant coefficient W1 and the variable coefficients W2 and W3, fully connected neural networks and Bayesian neural networks are used to solve them according to their different characteristics, which ensures the accuracy of the final result; at the same time, the introduction of the probability model (Bayesian network) also provides a more reasonable explanation for the randomness of the error.

[0044] Furthermore, the accuracy and reliability verification in S5 includes the following steps:

[0045] S51. Due to the introduction of the probability model, for the same input, the output is different each time. This invention uses commonly used absolute mean error, mean square error, and root mean square error to analyze the accuracy of each experimental result;

[0046] S52. Considering that the results of the probability model are not constant, the evaluation index in S51 cannot assess the accuracy of the distribution estimation. It is also necessary to use continuous hierarchical probability scores and interpretable variance to evaluate the accuracy of the model's estimated distribution. The beneficial effect of adopting the above further solution is that this invention provides a method for calculating the expectation and variance of the measured temperature field, and uses probabilistic methods to statistically analyze the randomness inherent in the measured temperature field, which can more accurately represent its inherent laws.

[0047] The beneficial effects of adopting the above-mentioned further scheme are as follows: Based on the characteristics of the temperature estimation model itself, the present invention verifies the accuracy of the model from two aspects: single-point prediction and distribution prediction, thus ensuring the accuracy of the final solution model.

[0048] Other advantages of the present invention will be analyzed in more detail in the following embodiments. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart illustrating the steps of a nuclear reactor main pipe coolant temperature estimation model that takes into account measurement errors, as described in an embodiment of the present invention.

[0051] Figure 2 This is a flowchart illustrating the steps involved in creating the dataset in an embodiment of the present invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0053] All other embodiments are within the scope of protection of this invention.

[0054] like Figure 1As shown, in one embodiment of the present invention, the present invention provides a model for estimating the coolant temperature of the main pipe of a nuclear reactor that takes into account measurement errors, comprising the following steps:

[0055] S1. Obtain actual temperature measurement data from the power plant and establish a dataset that corresponds to the measurement data and the agreed true values;

[0056] Creating the dataset in S1 includes the following steps:

[0057] S11. Actual measurement error is the difference between the measured value and the design value (theoretical calculation value), obtained from the working condition. Design value of temperature below ;

[0058] S12. Serialize the measured values ​​within the power plant over a period of time. With design value Subtracting them yields the measured error sequence. The statistical characteristics of the error magnitude were analyzed to obtain the measured error distribution. ;

[0059] S13. Based on the fluctuations in the actual operating conditions of the nuclear reactor, determine the range of temperature design value fluctuations. ;

[0060] S14. Based on the temperature range in S13 and the continuity of temperature change, the generation time length is... Temperature Convention True Value Sequence ;

[0061] S15. Based on the measured error distribution obtained in S12 to conduct Second-rate Sampling of an error sequence, and then comparing it with the sequence described in S14 Adding them together, we can get A series of measurement data sequences, each of which can be represented as: .

[0062] The flowchart of the dataset creation method is as follows: Figure 2 As shown.

[0063] By analyzing the actual temperature measurement process, measurement errors are classified and processed.

[0064] S2 includes the following steps:

[0065] S21. The measurement data itself contains two parts: the true value and the measurement error.

[0066] S22. There are many reasons for measurement errors, including instrument factors, environmental factors, and human factors.

[0067] S23. The magnitude of measurement error is not a constant value and has a certain degree of randomness. The statistical characteristics of non-homogeneous errors are different, making it difficult to describe them quantitatively in a unified manner.

[0068] S24. The overall measurement error is divided into two categories: steady-state error and transient error. Steady-state error has stable statistical characteristics, and its distribution is independent of time. According to the uncertainty calculation standard for the main pipeline of a nuclear reactor, the error characteristics of instrument errors, process errors, etc., can be described by white noise, so they can be regarded as steady-state errors. Transient error is defined as an error whose distribution is related to time. Observing the entire measurement process, it can be found that the temperature change of the main pipeline of the nuclear reactor is also related to time, that is, the transient error originates from the change in the actual temperature. Therefore, a preliminary estimate of the current temperature can be made, and then the transient error at the current moment can be extracted using this prior information.

[0069] S3. Considering different error distribution characteristics, combine the data-driven algorithm with the stochastic model to establish a temperature estimation model;

[0070] S3 includes the following steps:

[0071] S31. Because measurement data itself is not only random but also highly time-dependent, the commonly used model for processing random time series is the Autoregressive Moving Average (ARMA) model. This model assumes that the label values ​​fluctuate around a major trend over time, where the trend is influenced by historical labels, and the fluctuations are influenced by random events within a certain period, and the major trend itself is not necessarily stable. The temperature estimation model (NARMA) proposed in this invention draws on the basic idea of ​​this model.

[0072] S32. The temperature estimation model described in this invention comprises three terms. The first term assumes that the agreed-upon true value at the current moment is influenced by measured values ​​over a period of time (including past, present, and future times), thus compensating for the lack of consideration for the influence of future data on the current value in traditional ARMA. The second term is transient error information, which can be obtained through prior information about the temperature field. The third term is steady-state error information, consistent with the white noise information described in ARMA. The mathematical expression of this model can be written as:

[0073] in, Is The conventional true value of the time-time model estimation may take different values ​​under the same input due to input uncertainty. In the time period The sequence of measured values ​​within; It is a measurement coefficient matrix, which represents the influence of the measured values ​​on the agreed true value; In the time period Prior information within is... A preliminary estimate of the temperature value at each moment within the time frame, which is: The function; It is the transient error coefficient matrix, used to measure the impact of transient errors on the conventional truth value; It is the steady-state error during the measurement process, and its statistical characteristics are consistent with those of white noise; It is a random error coefficient matrix, used to measure the impact of steady-state error on the agreed-upon true value; since the error is random, and The elements in the array are not fixed values, but random variables that follow a certain distribution.

[0074] S4. Design a solution method for estimating prior information and calculating model coefficients in the temperature estimation model;

[0075] S4 includes the following steps:

[0076] S41. Use the Attention-LSTM model to estimate prior information;

[0077] S42. Solve the model coefficients using the FCNN-BN model.

[0078] S5. Optimize and solve the temperature estimation model using the dataset in S1, and verify its accuracy and reliability.

[0079] S5 includes the following steps:

[0080] S51. Due to the introduction of the probability model, for the same input, the output is different each time. This invention uses commonly used mean absolute error (MAE), mean square error (MSE), and root mean square error (RMSE) to analyze the accuracy of each experimental result. The calculation method of the evaluation index is as follows:

[0081]

[0082]

[0083]

[0084] in, For the true value, These are the model's predicted values. This represents the total number of results;

[0085] S52. Considering that the probability model results are not constant, the evaluation metrics in S51 cannot assess the accuracy of the distribution estimation. It is also necessary to use the Continuous Graded Probability Score (CRPS) and Explainable Variance (EVS) to evaluate the accuracy of the model's estimated distribution. The calculation methods for these evaluation metrics are as follows:

[0086]

[0087] in, This is the cumulative distribution function.

[0088] The beneficial effects of adopting the above-mentioned further solutions are as follows: the present invention provides a method for calculating the expectation and variance of the measured temperature field, and uses a probabilistic method to statistically analyze the randomness inherent in the measured temperature field itself, which can more accurately represent its inherent laws.

[0089] This invention addresses the problem of assessing the accuracy of temperature measurement data for nuclear reactor main ducts by proposing a temperature estimation model for nuclear reactor main ducts that considers measurement errors. This model can be input with actual measurement data for a specific time period. The Attention-LSTM module is used to first obtain a preliminary estimate of the temperature during this time period. Then, both are input into the FCNN-BNN model to solve for the coefficients of the temperature estimation model. The model ultimately outputs the predicted temperature values ​​at the corresponding measuring points within that time period. These predicted values ​​can then be used to determine the actual temperature distribution in the main piping of the nuclear reactor, thereby assessing the reliability and accuracy of the measurements.

[0090] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A model for estimating the coolant temperature in the main pipes of a nuclear reactor, considering measurement errors, characterized in that, Includes the following steps: S1. Obtain actual temperature measurement data from the power plant and establish a dataset that corresponds to the measurement data and the agreed true values; S2. By analyzing the actual temperature measurement process, the measurement errors are classified and processed. S3. Considering different error distribution characteristics, a temperature estimation model is established by combining data-driven algorithms with stochastic models. The model contains three terms: the first term assumes that the agreed-upon true value at the current moment is affected by the measured values ​​over a period of time (including past, present, and future time), thus compensating for the fact that traditional ARMA does not consider the influence of future data on the current value; the second term is transient error information, which can be obtained through prior information of the temperature field; the third term is steady-state error information, which is a white noise term, mathematically expressed as Y. t =W1X T +W2Y T +W3ε T ; where Y t X is the conventional true value estimated by the model at time t. Due to input uncertainty, it may take different values ​​for the same input; T It is the sequence of measured values ​​within the time period T; W1 is the measurement coefficient matrix, representing the influence of the measured values ​​on the agreed true value; Y T It is prior information within the time period T, a preliminary estimate of the temperature value at each moment within T, denoted as X. T The function; W2 is the transient error coefficient matrix, used to measure the impact of transient errors on the conventional truth value; ε T W2 represents the steady-state error during the measurement process, which is consistent with the statistical characteristics of white noise. W3 is the random error coefficient matrix, used to measure the impact of steady-state error on the agreed true value. Since the error is random, the elements in W2 and W3 are not constant values, but random variables that follow a certain distribution. S4. Design a solution method for estimating prior information and calculating model coefficients in the temperature estimation model; S5. Optimize and solve the temperature estimation model using the dataset in S1, and verify its accuracy and reliability.

2. The nuclear reactor main pipe coolant temperature estimation model considering measurement errors according to claim 1, characterized in that, Creating the dataset in S1 includes the following steps: S11. Define the actual measurement error as the difference between the measured value and the design value (theoretical calculation value), and obtain the design value A of the temperature under operating condition P; S12. The sequence of measured values ​​within the power plant over a period of time is B = {B1,…,B...}. n Subtracting the design value A from the measured error result yields the sequence e = {e1, ..., e}. n The statistical characteristics of the error magnitude are analyzed to obtain the measured error distribution F(e); S13. Based on the fluctuations during actual operation of the nuclear reactor, determine the range of temperature design value fluctuations [T]. min ,T max ]; S14. Based on the temperature range in S13 and the continuity of temperature change, generate a temperature-defined truth sequence C = {C1, ..., C...} with a time length of t. t }; S15. Based on the measured error distribution F(e) obtained in S12, sample t error sequences m times, and then add them to the sequence C described in S14 to obtain m sets of measurement data sequences. Each set of measurement data sequences can be represented as D = {D1, ..., D...} t }; S16. Repeat S14-S15 to create a large dataset.

3. The dataset corresponding to the measurement data and the agreed true value according to claim 2, characterized in that, The measured error used in S15 is not a constant value, but follows a certain probability distribution. Therefore, a set of agreed true value sequences corresponds to m sets of measurement data sequences.

4. The nuclear reactor main duct coolant temperature estimation model considering measurement errors according to claim 1, characterized in that, The error characteristics and classification methods of S2 include the following: S21. The measurement data itself contains two parts: the true value and the measurement error. S22. There are many reasons for measurement errors, including instrument factors, environmental factors, and human factors. S23. The magnitude of measurement error is not a constant value and has a certain degree of randomness. The statistical characteristics of non-homogeneous errors are different, making it difficult to describe them quantitatively in a unified manner. S24. This invention divides the overall measurement error into two categories: steady-state error and transient error. Steady-state error has stable statistical characteristics, and its distribution is independent of time. According to the uncertainty calculation standard for the main pipeline of a nuclear reactor, the error characteristics of instruments, processes, etc., can be described by white noise, and therefore can be considered as steady-state error. Transient error is defined as an error whose distribution is related to time. Observing the entire measurement process reveals that the temperature change of the main pipeline of the nuclear reactor is also related to time. This invention believes that transient error originates from the change in actual temperature. Therefore, a preliminary estimate of the current temperature can be made, and then the transient error at the current moment can be extracted using this prior information.

5. The nuclear reactor main duct coolant temperature estimation model considering measurement errors according to claim 1, characterized in that, The temperature estimation model of S3 has the following characteristics: S31. Since the measurement data itself has not only randomness but also strong time correlation, the commonly used model for processing random time series is the Autoregressive Moving Average (ARMA) model. This model assumes that the label value fluctuates around the general trend of time, where the trend is formed by the influence of historical labels, the fluctuation is formed by the influence of random events within a period of time, and the general trend itself is not necessarily stable. The temperature estimation model (NARMA) proposed in this invention draws on the basic idea of ​​this model.

6. The nuclear reactor main duct coolant temperature estimation model considering measurement errors according to claim 1, characterized in that, The solution process for the temperature estimation model of S4 consists of the following steps: S41. Use the Attention-LSTM model to estimate prior information; S42. Solve the model coefficients using the FCNN-BN model.

7. The nuclear pipeline coolant temperature estimation model considering measurement errors according to claim 6, characterized in that, The Attention-LSTM model in S41 estimates prior information through the following steps: S411, Collect n resistance temperature measurement data at time t. The data is fed into the attention mechanism module to obtain data carrying the location information of other measurement points. S412, update n inputs By feeding the data into an LSTM neural network structure, prior estimates of the temperature at n measurement points at time t can be obtained. S413. Repeat the above steps with the measurement data of time length T to obtain the prior estimate information Y of temperature. T .

8. The coolant temperature estimation model according to claim 6, characterized in that, The FCNN-BN model in S42 solves for the model coefficients through the following steps: S421. Calculate the n resistance temperature measurement data at time t. The data is fed into a fully connected neural network (FCNN) and, through weight W1, calculated to obtain... S422, Y T The data is fed into a Bayesian neural network (BN) and processed for ε. T The characteristics are assumed (usually that they follow a normal distribution with a mean of 0 and a variance of 1), and the following is calculated using weights W2 and W3: S423, will and Adding them together gives the final result. The final result is fed into the loss function to train and optimize the model parameters W1, W2, and W3.

9. A nuclear reactor main duct coolant temperature estimation model considering measurement errors according to claim 1, characterized in that, The accuracy and reliability verification in S5 includes the following steps: S51. Due to the introduction of the probability model, the output is different for the same input each time. This invention uses the commonly used absolute average error, mean square error, and root mean square error to analyze the accuracy of the experimental results for each time. S52. Considering that the results of the probability model are not constant, the evaluation index in S51 cannot assess the accuracy of the distribution estimation. It is also necessary to use continuous hierarchical probability scores and interpretable variance to assess the accuracy of the model's estimated distribution.

Images