Method for rapidly calculating leakage rate of containment during operation of nuclear power unit

Abstract

The invention discloses a method for rapidly calculating the leakage rate of a containment during operation of a nuclear power unit, and relates to the technical field of nuclear power plant containment sealing performance evaluation.Data preprocessing, sliding window traversal, error elimination, data expansion, linear recognition, aggregation classification and rapid calculation are sequentially carried out on original data collected by a system; therefore, the linearization trend in the effective data is identified, and rapid calculation is realized through aggregation and classification. According to the method, original data collected during operation of a nuclear power plant unit is processed, rapid calculation of the leakage rate is finally realized, the leakage rate can be calculated according to the time of two days from shutdown to double shutdown of an EBA system, and power plant operation personnel are assisted in evaluating the sealing state of the containment; the leakage rate calculation precision can be improved, and the false alarm of the leakage rate on-line monitoring value during the unit operation period is eliminated.

Description

Technical Field

[0001] The present invention relates to the technical field of the evaluation of the containment seal performance of nuclear power plants, and more specifically to a method for quickly calculating the leakage rate of the containment during the operation of nuclear power units. Background Art

[0002] The containment of a nuclear power unit is a key barrier to prevent the leakage of radioactive substances, and its seal performance is directly related to the safe operation of the nuclear power plant. During the entire operation cycle of the nuclear power unit, accurately and quickly evaluating the leakage rate of the containment is crucial for ensuring the safe operation of the nuclear power unit. However, there are many problems in the existing on-line monitoring system for the containment leakage in practical applications, and it is difficult to meet the requirements of real-time and accurate monitoring of the containment seal performance during the operation of the nuclear power unit.

[0003] First of all, the calculation time of the existing monitoring system is too long. Generally, it takes at least 5 days to complete a calculation of the leakage rate, which makes it impossible to evaluate the seal performance of the containment in a timely manner during the critical stages of the operation of the nuclear power unit (such as during the period from the EBA shutdown to the double closing of the air gates, usually 2 - 3 days). This time delay may cause potential safety hazards not to be discovered and processed in a timely manner.

[0004] Secondly, the calculation accuracy of the existing monitoring system is relatively low. Due to the limitations of the system's own software and algorithms, the calculation results of the leakage rate are easily interfered with, resulting in the calculation accuracy not meeting the actual requirements. More seriously, the existing system is prone to false alarms during the normal operation of the unit, and it is difficult to eliminate the false alarms. This will not only interfere with the normal operation of the nuclear power plant, but also may cause additional attention from the regulatory authorities to nuclear safety, bringing unnecessary safety hazards.

[0005] In addition, the existing monitoring system also has deficiencies in data processing and analysis. Traditional monitoring methods usually rely on simple data collection and analysis, lacking the ability of in-depth data mining and processing. This makes it difficult to accurately identify and eliminate abnormal data under complex working conditions, thus affecting the accuracy and reliability of the leakage rate calculation.

[0006] In summary, the existing on-line monitoring system for the containment leakage cannot meet the requirements of real-time and accurate monitoring of the containment seal performance during the operation of the nuclear power unit in terms of calculation speed, accuracy, and reliability. Therefore, it is urgent to develop a new method that can quickly and accurately calculate the leakage rate of the containment during the operation of the nuclear power unit, so as to improve the safety and operation efficiency of the nuclear power unit and reduce the nuclear safety risk. Summary of the Invention

[0007] In view of this, the present invention provides a method for quickly calculating the containment leakage rate during the operation of a nuclear power unit, aiming to accurately and quickly evaluate the leakage of the containment of a nuclear power unit through real-time monitoring and data processing, thereby ensuring the safe operation of the nuclear power unit. Based on the operation data of the nuclear power unit, combined with the leakage characteristics and related physical parameters, the present invention adopts a combination of mathematical modeling and data fitting to calculate the leakage rate of the containment in a short time.

[0008] In order to achieve the above object, the present invention adopts the following technical solution:

[0009] A method for quickly calculating the containment leakage rate during operation of a nuclear power unit comprises the following steps:

[0010] Data preprocessing: Preprocess the raw data collected during the operation of the nuclear power unit, construct the data into one-dimensional time series data, and eliminate the gross error data points through standard deviation;

[0011] Sliding window traversal: Perform sliding window traversal on the preprocessed data and calculate the intermediate data by combining different window sizes and sliding amounts;

[0012] Error elimination: Error identification of intermediate data, eliminating data points that deviate from the linear trend;

[0013] Data expansion: Through the results of sliding window traversal, the data is expanded to increase the frequency of use of data points;

[0014] Linearization identification: perform linearization processing on the expanded data to identify the linearization trend in the data;

[0015] Aggregation and classification: Aggregate the linearized data and classify them according to the pressure parameters;

[0016] Fast calculation: Fit a linear model to the classified data and use the slope to characterize the leakage rate to achieve fast calculation.

[0017] Through the above technical scheme, the present invention sequentially performs data preprocessing, sliding window traversal, error elimination, data expansion, linearization identification, aggregation classification and fast calculation on the raw data collected by the system, thereby identifying the linearization trend in the valid data and realizing fast calculation through aggregation and classification. The present invention processes the raw data collected during the operation of the nuclear power plant unit, and finally realizes the fast calculation of the leakage rate. It can calculate the leakage rate by using the two-day period from the shutdown of the EBA system to the shutdown, and assist the power plant operators in evaluating the containment sealing state; it can improve the calculation accuracy of the leakage rate and eliminate the false alarm of the online monitoring value of the leakage rate during the operation of the unit.

[0018] Preferably, in the above method for quickly calculating the containment leakage rate during the operation of a nuclear power unit, the specific calculation of the data preprocessing includes: calculating the mean and standard deviation of the data, setting the gross error threshold as the mean plus or minus a constant multiple of the standard deviation, and removing the data points outside the threshold range to obtain the cleaned time-series data.

[0019] Specifically, the original data collected by the system is preprocessed. The purpose of the preprocessing is to construct the original data into one-dimensional time-series data , where is the original measurement value of the i-th data point, ensuring that the data conforms to a consistent time interval.

[0020] By removing the gross error data points through the standard deviation, calculate the mean and the standard deviation , and then set the gross error threshold as , where is a constant, which is user-defined according to experience and usually takes 3. If , then remove the data point x i . After processing, the gross error data points outside the threshold range are removed to obtain the cleaned time-series data.

[0021] Preferably, in the above method for quickly calculating the containment leakage rate during the operation of a nuclear power unit, the sliding window traversal includes: performing a sliding window traversal on the cleaned time-series data by exhausting all possible window sizes and sliding amounts. For each set of window sizes and sliding amounts, a sliding window is formed within the data, and the data within each window is processed through a user-defined function and the result is output. By traversing all possible combinations of window sizes and sliding amounts, each data point is calculated multiple times, thereby expanding the data set.

[0022] Specifically, the cleaned time-series data is subjected to a sliding window traversal by exhausting all possible window sizes and sliding amounts . The value ranges of the window size and the sliding amount are:

[0023] ;

[0024] For each set of window sizes and sliding amounts , a sliding window is formed within the data, and the data within each window is . The processing result of each window is Calculate. This function processes the data within the window and outputs the result. The specific calculation formula is as follows:

[0025] ;

[0026] Among them, can be weighted average, linear fitting.

[0027] By traversing all possible combinations of window sizes and sliding amounts, each data point is calculated multiple times, thereby expanding the data set, increasing the usage frequency of data points, and ensuring that each data point participates in different sliding window calculations. The number of data points after expansion is:

[0028] ;

[0029] Among them, is the total amount of original data, is the window size, is the sliding amount. By exhausting all possible window combinations, the number of data points after expansion increases, ensuring the multiple participation and comprehensive processing of data points.

[0030] Preferably, in the above method for quickly calculating the containment leakage rate during the operation of a nuclear power unit, the error elimination includes: for each window calculation result obtained during the traversal of the sliding window, a corresponding physical model is established, the model is linearized, the perturbations existing in the linearized data are identified, and the perturbation identification is achieved by calculating the residual between the linearized data and the fitting model. If the residual exceeds the user-set threshold, it is considered that the data point has a perturbation and is thus eliminated.

[0031] Specifically, for each window calculation result obtained during the traversal of the sliding window , a corresponding physical model is established as follows:

[0032] ;

[0033] Among them, and are constants, is a time-related variable.

[0034] The physical model is linearized by taking the logarithm of both sides of the model:

[0035] ;

[0036] Identify the perturbations existing in the linearized data. The perturbation is the part where the data point deviates from the linear trend. The perturbation identification is achieved by calculating the residual between the linearized data and the fitting model:

[0037] ;

[0038] If the residual exceeds the user-set threshold , it is considered that the data point is disturbed and should be excluded.

[0039] Preferably, in the above method for quickly calculating the containment leakage rate during the operation of a nuclear power unit, the aggregation classification includes: obtaining aggregation parameters through training with historical data. After obtaining the aggregation parameters, aggregating the effective data after removing the linear error, establishing a correlation model between the aggregated data and the pressure parameter. When the correlation between the aggregated data and the pressure parameter exceeds a certain threshold, it is considered that the two are linearly correlated; otherwise, they are not correlated. After classification, only establish the functional relationship of the linearly correlated part.

[0040] Specifically, obtaining aggregation parameters through training with historical data , this parameter describes the aggregation relationship between different data groups. The historical data training process can be carried out through regression analysis, correlation analysis, and machine learning methods.

[0041] When training the aggregation parameters with historical data, the historical data is constructed into a set of known data and the corresponding labels , and the training process is:

[0042] ;

[0043] Among them, is the aggregation function during the training process.

[0044] After obtaining the aggregation parameters, aggregating the effective data after removing the linear error , and the aggregation process is:

[0045] ;

[0046] Among them, is the aggregation function, is the aggregated data.

[0047] Establish the correlation model with the pressure parameter , is the original data collected by the system, and the model is as follows:

[0048] ;

[0049] Among them, is the covariance between the aggregated data and the pressure parameter, and are the standard deviations of the aggregated data and the pressure parameter respectively, A coefficient for quantifying the degree of correlation between aggregated data and pressure parameters.

[0050] When and the correlation exceeds a certain threshold , it is considered that and are linearly correlated; otherwise, they are uncorrelated.

[0051] Classification is to classify and into two categories: linearly correlated and uncorrelated. After classification, only the linearly correlated part and is used to establish a functional relationship, that is:

[0052] ;

[0053] Preferably, in the above method for quickly calculating the containment leakage rate during the operation of a nuclear power unit, the collected original data includes, but is not limited to, pressure, temperature, and flow rate data during the operation of the nuclear power unit.

[0054] Preferably, in the above method for quickly calculating the containment leakage rate during the operation of a nuclear power unit, during the sliding window traversal, the value ranges of both the window size and the sliding amount are from 1 to the total number of data points.

[0055] Preferably, in the above method for quickly calculating the containment leakage rate during the operation of a nuclear power unit, the quick calculation: is a linear model. Fit the linear model, and use the slope to characterize the leakage rate after quick calculation.

[0056] Through the above technical solutions, compared with the prior art, the present invention discloses a method for quickly calculating the containment leakage rate during the operation of a nuclear power unit, which has the following beneficial effects:

[0057] 1. Significantly improve the calculation efficiency: It can quickly complete the calculation of the leakage rate during the critical stage of the operation of the nuclear power unit, solve the problem of too long calculation time in the prior art, thereby timely evaluate the sealing performance of the containment, and avoid the delayed discovery of potential safety hazards. Through the sliding window traversal and data augmentation technology, the data is processed and extended multiple times, increasing the usage frequency of data points, and improving the efficiency and comprehensiveness of data processing.

[0058] 2. Improve calculation accuracy: Through steps such as data preprocessing (removing gross error data points by standard deviation) and error elimination (linearization processing and residual analysis based on physical models), abnormal data points can be effectively identified and removed, avoiding the interference of these data on the calculation result of the leakage rate, thereby improving the calculation accuracy. By linearizing the data and performing aggregation classification based on pressure parameters, the linearization trend in the data can be more accurately identified, and a functional relationship is established only for the linearly correlated part, further improving the accuracy of the leakage rate calculation.

[0059] 3. Eliminate false alarms: Existing monitoring systems are prone to false alarms during the normal operation of the unit. However, the present invention can effectively eliminate false alarms through precise data processing and analysis methods, avoiding interference with the normal operation of the nuclear power plant and reducing unnecessary safety hazards and regulatory concerns.

[0060] 4. Enhance data processing ability: By adopting a combination of mathematical modeling and data fitting to deeply mine and process the operation data of nuclear power units, it can better meet the data processing requirements under complex working conditions and improve the reliability and accuracy of data processing. The collected original data includes various data types such as pressure, temperature, and flow rate. By comprehensively processing these data, the sealing performance of the containment can be more comprehensively evaluated.

[0061] 5. Optimize the monitoring process: The present invention provides a complete set of data processing and leakage rate calculation processes. From data preprocessing to rapid calculation, each step is interconnected, forming a systematic solution, improving the overall efficiency and reliability of the monitoring work. By training with historical data to obtain aggregation parameters, the monitoring model can be optimized according to the actual operation data, further improving the adaptability and accuracy of the monitoring system.

[0062] 6. Improve the operation safety of nuclear power units: By quickly and accurately evaluating the leakage situation of the containment, potential safety hazards can be timely detected, assisting power plant operators to take effective countermeasures, thereby ensuring the safe operation of nuclear power units and reducing nuclear safety risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0064] Figure 1 The drawings are the flowcharts for quickly calculating the leakage rate of the containment during the operation of the nuclear power unit provided by the present invention;

[0065] Figure 2 The accompanying drawing is a schematic diagram for quickly calculating the containment leakage rate during the operation of a nuclear power unit provided by the present invention. Specific embodiments

[0066] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0067] See the attached Figure 1 and the attached Figure 2 The embodiments of the present invention disclose a method for quickly calculating the containment leakage rate during the operation of a nuclear power unit, including the following steps:

[0068] Data preprocessing: Preprocess the original data collected during the operation of the nuclear power unit, construct the data into one-dimensional time series data, and eliminate gross error data points through the standard deviation.

[0069] Sliding window traversal: Traverse the preprocessed data with a sliding window, and calculate the intermediate data through combinations of different window sizes and sliding amounts.

[0070] Error elimination: Identify errors in the intermediate data and eliminate data points that deviate from the linear trend.

[0071] Data augmentation: Through the results of the sliding window traversal, perform augmentation processing on the data to increase the usage frequency of data points.

[0072] Linearization identification: Perform linearization processing on the augmented data to identify the linearization trend in the data.

[0073] Aggregation classification: Aggregate the data after linearization processing and classify it according to the pressure parameter.

[0074] Quick calculation: Fit a linear model to the classified data, and use the slope to represent the leakage rate to achieve quick calculation.

[0075] To further optimize the above technical solution, the specific calculation of data preprocessing includes: calculating the mean and standard deviation of the data, setting the gross error threshold as the mean plus or minus a constant multiple of the standard deviation, and eliminating the data points outside the threshold range to obtain the cleaned time series data.

[0076] To further optimize the above technical solution, the sliding window traversal includes: performing a sliding window traversal on the cleaned time series data by exhausting all possible window sizes and sliding amounts. For each set of window size and sliding amount, a sliding window is formed within the data, and the data within each window is processed by a user-defined function and the results are output. By traversing all possible combinations of window sizes and sliding amounts, each data point is calculated multiple times, thereby expanding the data set.

[0077] To further optimize the above technical solution, the error elimination includes: for each window calculation result obtained during the sliding window traversal, establishing a corresponding physical model, linearizing the model, identifying the perturbations existing in the linearized data. The perturbation identification is achieved by calculating the residuals between the linearized data and the fitted model. If the residuals exceed the user-set threshold, it is considered that the data point has a perturbation and is thus eliminated.

[0078] To further optimize the above technical solution, the aggregation classification includes: obtaining aggregation parameters through training with historical data. After obtaining the aggregation parameters, aggregating the valid data after eliminating the linear errors, establishing a correlation model between the aggregated data and the pressure parameters. When the correlation between the aggregated data and the pressure parameters exceeds a certain threshold, it is considered that they are linearly correlated, otherwise they are not correlated. After classification, only the functional relationship of the linearly correlated part is established.

[0079] To further optimize the above technical solution, the training process is carried out through regression analysis, correlation analysis, and / or machine learning methods.

[0080] To further optimize the above technical solution, the collected original data includes, but is not limited to, the pressure, temperature, and flow data during the operation of nuclear power units.

[0081] To further optimize the above technical solution, in the sliding window traversal, the value ranges of both the window size and the sliding amount are from 1 to the total number of data points.

[0082] To further optimize the above technical solution, in the linearization identification, when linearizing the physical model, take the logarithm of both sides of the model.

[0083] To further optimize the above technical solution, in the aggregation classification, when calculating the correlation coefficient between the aggregated data and the pressure parameters, use the ratio of covariance to standard deviation to quantify the degree of correlation.

[0084] Example 1:

[0085] The original data collected by the system is as follows:

[0086] Time point , and the corresponding measured value is , the preprocessing objective is to process the original data into time series data with consistent intervals. If there are cases where the time intervals of the data are inconsistent (such as missing data or irregular sampling), interpolation or elimination is required.

[0087] Eliminate gross error data points through standard deviation and calculate the mean of the data and the standard deviation , set the gross error threshold to , all data is within this range, so no data is eliminated.

[0088] Use a window size of 3 and a shift of 1 to expand the data to obtain the following window data:

[0089] Window 1: [5.0, 5.2, 5.4]; Window 2: [5.2, 5.4, 5.6]; Window 3: [5.4, 5.6, 5.1]; Window 4: [5.6, 5.1, 6.3].

[0090] For each window, use weighted average or linear fitting to calculate the processing result,

[0091] A new data set is obtained: 5.2, 5.4, 5.37, 5.67;

[0092] For each calculation result, error elimination is performed based on the following physical model:

[0093] ;

[0094] From historical data, and are constants.

[0095] Take the logarithm of each processed data point for linearization:

[0096] ;

[0097] For each data point, calculate the residual. For example, if the data 5.4 corresponds to time point 2, then

[0098] ;

[0099] Set the threshold , and at this time no data is eliminated.

[0100] Obtain the aggregation parameter through historical data training ;

[0101] For the valid data set 5.2, 5.4, 5.37, 5.67, use the weighted average method for aggregation:

[0102] ;

[0103] Perform a correlation analysis on the aggregated data and the pressure data, calculate the correlation coefficient r to determine the linear correlation between the data.

[0104] The pressure data is 48, 50, 53, 56. Calculate the correlation coefficient between the aggregated data and the pressure data:

[0105] , threshold , linearly correlated, establish a linear model:

[0106] , through historical data analysis ;

[0107] When the EBA system is out of service for two days until it is double-closed, it is necessary to calculate the leakage rate at a pressure parameter of 60. Substituting it into the model, the leakage rate is -2.3.

[0108] In this specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple. For the relevant parts, refer to the description in the method section.

[0109] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for quickly calculating the containment leakage rate during operation of a nuclear power unit, characterized in that: The following steps are involved: Data preprocessing: Preprocess the raw data collected during the operation of the nuclear power unit, construct the data into one-dimensional time series data, and eliminate the gross error data points through standard deviation; Sliding window traversal: Perform sliding window traversal on the preprocessed data and calculate the intermediate data by combining different window sizes and sliding amounts; Error elimination: Error identification of intermediate data, eliminating data points that deviate from the linear trend; Data expansion: Through the results of sliding window traversal, the data is expanded to increase the frequency of use of data points; Linearization identification: perform linearization processing on the expanded data to identify the linearization trend in the data; Aggregation and classification: Aggregate the linearized data and classify them according to the pressure parameters; Fast calculation: Fit a linear model to the classified data and use the slope to characterize the leakage rate to achieve fast calculation.

2. A method for rapidly calculating the containment leakage rate during operation of a nuclear power unit according to claim 1, characterized in that: The specific calculation of the data preprocessing includes: calculating the mean and standard deviation of the data, setting the gross error threshold to the mean plus or minus a constant times the standard deviation, removing data points that exceed the threshold range, and obtaining the cleaned time series data.

3. A method for rapidly calculating the containment leakage rate during operation of a nuclear power unit according to claim 1, characterized in that: The sliding window traversal includes: performing sliding window traversal on the cleaned time series data by exhausting all possible window sizes and sliding amounts, forming a sliding window in the data for each group of window size and sliding amount, processing the data in each window by a user-defined function and outputting the result, and performing multiple calculations on each data point by traversing all possible combinations of window size and sliding amount, thereby expanding the data set.

4. A method for rapidly calculating the containment leakage rate during operation of a nuclear power unit according to claim 1, characterized in that: The error elimination includes: for each window calculation result obtained during the sliding window traversal process, a corresponding physical model is established, the model is linearized, and the disturbance in the linearized data is identified. The disturbance identification is achieved by calculating the residual between the linearized data and the fitting model. If the residual exceeds the user-set threshold, it is considered that the data point has a disturbance and is eliminated.

5. A method for rapidly calculating the containment leakage rate during operation of a nuclear power unit according to claim 1, characterized in that: The aggregation classification includes: obtaining aggregation parameters through historical data training, after obtaining the aggregation parameters, aggregating the valid data after eliminating linear errors, and establishing a correlation model between the aggregated data and the pressure parameters. When the correlation between the aggregated data and the pressure parameters exceeds a certain threshold, it is considered that the two are linearly correlated, otherwise they are unrelated. After classification, only the functional relationship of the linearly correlated part is established.

6. A method for rapidly calculating the containment leakage rate during operation of a nuclear power unit according to claim 5, characterized in that: The training process is performed by regression analysis, correlation analysis and / or machine learning methods.

7. A method for rapidly calculating the containment leakage rate during operation of a nuclear power unit according to claim 1, characterized in that: The collected raw data include but are not limited to pressure, temperature and flow data during the operation of the nuclear power unit.

8. A method for rapidly calculating the containment leakage rate during operation of a nuclear power unit according to claim 1, characterized in that: In the sliding window traversal, the window size and the sliding amount both range from 1 to the total number of data points.

9. A method for rapidly calculating the containment leakage rate during operation of a nuclear power unit according to claim 1, characterized in that: In the linearization identification, when the physical model is linearized, logarithms of both sides of the model are taken simultaneously.

10. A method for rapidly calculating the containment leakage rate of a nuclear power unit during operation according to claim 1, characterized in that: In the aggregate classification, when calculating the correlation coefficient between the aggregate data and the pressure parameter, the ratio of the covariance to the standard deviation is used to quantify the degree of correlation.

Images