Nuclear power station box anomaly detection method and device

Abstract

The invention relates to the technical field of nuclear power safety protection, in particular to a nuclear power station box anomaly detection method and device. The method comprises the following steps: determining ideal influence factors of the box body; preliminarily determining an influence factor of the box; determining a final influence factor, and generating a corresponding training data set; establishing a reconstruction model and a regression model, and performing training; obtaining influence factor data of the current box body, and inputting the current influence factor data into the reconstruction model and the regression model to obtain two groups of prediction data of the box body; and acquiring actual operation data of the current box body, and comparing the actual operation data with the prediction data to judge whether the box body is abnormal or not. The device comprises an influence factor preliminary determination module; a potential impact factor determination module; a training set processing module; a reconstruction model training module; a regression model training module; a prediction module; and an abnormity judgment module. The method is accurate in judgment and high in early warning capability.

Description

Technical Field

[0001] The present invention relates to the technical field of nuclear power safety protection, and in particular to a method and device for detecting abnormality of a nuclear power plant box. Background Art

[0002] During the normal operation of a nuclear power plant, unexpected protection actions such as load shedding, shutdown, and reactor shutdown may occur due to faulty operation or failure of process equipment, instrumentation and control system failure, power grid failure, fire, insufficient personnel skills, and human error.

[0003] The box or tank plays an important role in nuclear power plants. Most boxes only have the function of storage, such as the common storage tank; some tanks also have the function of heat exchange, such as the stabilizer. If the box fails, it may cause the entire system to stop or shut down unexpectedly, causing serious economic losses. Therefore, it is of great practical significance to detect and eliminate the box failure in time.

[0004] Currently, the detection of abnormal conditions of liquid level, pressure or temperature signals during box operation mainly relies on the alarm information of the DCS system. The corresponding alarm is triggered only when a single measurement value or the combination of multiple measurement values ​​reaches the corresponding threshold. There is a lack of trend analysis and the abnormal warning capability is poor. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a method and a device for detecting abnormality of a nuclear power plant box, which have strong early warning capability.

[0006] The present invention provides a method for detecting abnormality of a nuclear power plant box, comprising the following steps:

[0007] Step S100: establishing a mass conservation equation for a box, and determining an ideal influencing factor for the box based on the mass conservation equation;

[0008] Step S200: preliminarily determine the influencing factor of the box based on the actual instrument deployment situation and business knowledge;

[0009] Step S300: obtaining a number of historical working data of the cabinet, analyzing the correlation between the potential influencing factors and the historical working data by using a data analysis method to determine the final influencing factors of the cabinet, and generating a corresponding training data set based on the historical working data and the final determined influencing factors;

[0010] Step S400: establishing a reconstruction model, and using the training data set to train the reconstruction model to obtain a fully trained box reconstruction model and warning thresholds of various influencing factors;

[0011] Step S500: establishing a regression model, and using the training data set to train the regression model to obtain a fully trained box regression model and a warning threshold value of a regression output parameter;

[0012] Step S600: Obtain the influencing factor data of the current box, and input the current influencing factor data into the reconstruction model and the regression model to obtain two sets of prediction data for the box;

[0013] Step S700: Acquire actual operation data of the current cabinet, and compare the actual operation data with the predicted data to determine whether the cabinet is abnormal.

[0014] In a specific embodiment of the present invention, the historical operation data at least includes historical liquid level difference data and its corresponding influencing factor data, and the influencing factor data includes the inlet flow and temperature of the box, the outlet flow and temperature of the box, the top pressure of the box, the liquid temperature, and the gas phase pressure.

[0015] In a specific embodiment of the present invention, when establishing the reconstruction model and the regression model, the influence of temperature and pressure on density is taken into account, and the inlet mass of the box is obtained by calculating the box inlet volume flow rate, the box inlet liquid density and the box inlet liquid pressure, and the outlet mass of the box is obtained by calculating the box outlet volume flow rate, the box outlet liquid density and the box outlet liquid pressure, wherein the box inlet liquid density and the box inlet liquid pressure are related to the box inlet liquid temperature.

[0016] In a specific implementation of the present invention, the reconstruction model adopts the reconstruction algorithm of principal component analysis PCA reconstruction, autoencoder AE, and variational autoencoder VAE.

[0017] In a specific embodiment of the present invention, the regression model is one of a linear regression model, a polynomial regression, a decision tree regression model, a support vector machine regression model, a K nearest neighbor regression model, a random forest regression model, an Adaboost regression model, a gradient enhanced random forest regression model, a bagging regression model, and an Extra TREE regression model.

[0018] In a specific implementation of the present invention, step S700 is specifically as follows:

[0019] Acquire actual working data of the current cabinet, and compare the actual working data with the predicted working data to obtain a residual between the actual working data and the predicted working data;

[0020] Obtaining a working data abnormality alarm threshold according to the training result, comparing the residual with the working data abnormality alarm threshold, and judging whether the box is abnormal based on the comparison result;

[0021] When it is determined that the box is abnormal, a corresponding alarm prompt message is issued, and the alarm prompt message includes the specific situation of the abnormality of the box and the cause of the abnormality of the box.

[0022] In a specific implementation of the present invention, it further includes: when it is determined that the cause of the cabinet abnormality alarm is a false alarm or an under-alarm, the latest historical data is obtained, and the cabinet abnormality detection model is updated using the latest historical data.

[0023] The present invention provides a nuclear power plant box abnormality detection device, comprising:

[0024] An influence factor preliminary determination module, used to establish a mass conservation equation for the box, and determine the influence factor of the box based on the mass conservation equation;

[0025] Potential impact factor determination module, used to preliminarily determine the impact factor of the box based on the actual instrument deployment situation and business knowledge;

[0026] A training set processing module is used to obtain a number of historical working data of the box, analyze the correlation between the potential influencing factors and the historical working data using a data analysis method to determine the final influencing factors of the heat exchanger, and generate a corresponding training data set based on the historical working data and the final determined influencing factors;

[0027] A reconstruction model training module is used to establish a reconstruction model and train the reconstruction model using the training data set to obtain a fully trained box reconstruction model and warning thresholds of various influencing factors;

[0028] A regression model training module is used to establish a regression model and train the regression model using the training data set to obtain a fully trained box regression model and a warning threshold value of a regression output parameter;

[0029] A prediction module is used to obtain the influencing factor data of the current box, and input the current influencing factor data into the reconstruction model and the regression model for reconstruction to obtain two sets of prediction data of the box;

[0030] The abnormality judgment module is used to obtain actual operation data of the current cabinet, and compare the actual operation data with the predicted data to judge whether the cabinet has an abnormality.

[0031] The present invention provides an electronic device, comprising: a processor and a memory;

[0032] The memory stores a computer program executable by the processor;

[0033] When the processor executes the computer program, the steps in the method for detecting abnormality of a nuclear power plant box are implemented.

[0034] The present invention provides a computer-readable storage medium, which stores one or more programs. The one or more programs can be executed by one or more processors to implement the steps in the nuclear power plant box abnormality detection method.

[0035] Compared with the prior art, a method and device for detecting abnormalities in a nuclear power plant tank body of the present invention obtains the influencing factors of the tank body abnormalities, adopts reconstruction and regression algorithms to construct an abnormality detection model for the tank body, and other technical means, thereby achieving the technical effect that the liquid level changes of the tank body can be directly predicted through the abnormality detection model, and then compared with the actual liquid level changes, it can quickly and accurately determine whether the tank body has an abnormality. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A flow chart showing a method for detecting abnormalities in a nuclear power plant tank;

[0037] Figure 2 A schematic diagram showing the functional modules of a nuclear power plant box abnormality detection device;

[0038] Figure 3 A schematic diagram showing the hardware structure of an electronic device;

[0039] In the figure, 10 is an electronic device, 110 is a processor, and 120 is a memory. DETAILED DESCRIPTION

[0040] In order to further understand the present invention, the embodiments of the present invention are described below in conjunction with examples, but it should be understood that these descriptions are only for further illustrating the features and advantages of the present invention, rather than limiting the present invention.

[0041] The embodiment of the present invention discloses a method for detecting abnormality of a nuclear power plant box. Figure 1 As shown, the following steps are included:

[0042] S100: establishing a mass conservation equation of the box according to instrument layout and business knowledge, and determining an influencing factor of the box based on the mass conservation equation;

[0043] S200: Establishing a mass conservation equation for the box, and determining an ideal influencing factor for the box based on the mass conservation equation;

[0044] S300: Acquire several historical working data of the cabinet, use a data analysis method to analyze the correlation between the potential influencing factors and the historical working data to determine the final influencing factors of the cabinet, and generate a corresponding training data set based on the historical working data and the final determined influencing factors;

[0045] S400: Establishing a reconstruction model, and using the training data set to train the reconstruction model to obtain a fully trained box reconstruction model and warning thresholds of various influencing factors;

[0046] S500: Establishing a regression model, and using the training data set to train the regression model to obtain a fully trained box regression model and a warning threshold value of a regression output parameter;

[0047] S600: Obtain the influencing factor data of the current box, and input the current influencing factor data into the box reconstruction model and the regression model to obtain two sets of prediction data for the box;

[0048] S700: Acquire actual operation data of the current cabinet, and compare the actual operation data with the predicted data to determine whether the cabinet is abnormal.

[0049] In this embodiment, firstly, a mass conservation equation of the box is established, and the influencing factor of the box is determined based on the mass conservation equation; then, several historical working data of the box are obtained, and based on the data analysis method, its influencing shadow is finally determined and the processing of the required features is completed, and a training set is established; a reconstruction model is established, and the regression model is trained using the training data set to obtain a well-trained box reconstruction model and early warning thresholds of various influencing factors; a regression model is established, and the regression model is trained using the training data set to obtain a well-trained box regression model and early warning thresholds of regression output parameters; then, the influencing factor data of the current box is obtained, and the current influencing factor data is input into the reconstruction model and the regression model to obtain the predicted working data of the box; finally, the actual working data of the current box is obtained, and the actual working data is compared with the predicted working data to determine whether the box is abnormal. The present invention obtains the influencing factors of the tank abnormality, adopts regression algorithm and reconstruction algorithm to construct two types of tank abnormality detection models, predicts the liquid level and temperature changes of the tank through the abnormality detection model, and then compares them with the actual liquid level and temperature changes, so as to quickly and accurately determine whether the tank has an abnormality.

[0050] In some embodiments, the mass conservation equation is used to reflect the change in the liquid level of the tank. Under ideal conditions, the increase in the tank liquid level is related to the mass change on both sides of the tank. Therefore, in order to obtain the liquid level change of the tank, the present invention can establish a mass conservation equation, and then obtain the ideal influencing factor of the tank liquid level change by converting the formula. After that, combined with the actual instrument deployment situation and business knowledge, the influencing factor of the tank is preliminarily determined.

[0051] In some embodiments, in step S300, the anomaly detection model is constructed based on knowledge and experience, and the historical operation data of the box is obtained according to the operating experience of the nuclear power plant, and the historical operation data at least includes historical liquid level difference data and its corresponding influencing factor data. Optionally, it includes the inlet flow and temperature of the box, the outlet flow and temperature of the box, the top pressure of the box (need to be considered for closed boxes), liquid temperature, and gas phase pressure; some boxes also need to consider the flow, temperature, pressure, etc. on the cooling side or electric heating, etc. It should be noted that some boxes need to be divided into multiple operating conditions. The influencing factors of the box under each operating condition are finally determined through data analysis methods, and the features to be processed are processed, thereby forming a training data set under this condition.

[0052] In this embodiment, according to the law of conservation of mass, there is no leakage, and the mass conservation is satisfied in the entire box, that is, Therefore, in the embodiment of the present invention, the regression model is built based on the law of mass conservation, deltalevel=function(Min1, Min2, Mout1, level), that is, the change of the tank liquid level is obtained by the mass difference on both sides of the tank.

[0053] In some embodiments, under the premise that the density does not change much, the mass is only related to the flow rate. However, in actual situations, both temperature and pressure will oscillate, which will cause the density to change. Therefore, in the embodiment of the present invention, during the modeling process, it is necessary to consider the influence of temperature and pressure on density. Optionally, the inlet mass of the box is obtained by calculating the box inlet volume flow rate, the box inlet liquid density and the box inlet liquid pressure, and the outlet mass of the box is obtained by calculating the box outlet volume flow rate, the box outlet liquid density and the box outlet liquid pressure, wherein the box inlet liquid density and the box inlet liquid pressure are related to the box inlet liquid temperature.

[0054] In some embodiments, in step S400, since the power plant may undergo maintenance, the data will change after the maintenance, which may lead to false alarms in model prediction. Therefore, in order to avoid this situation, the embodiment of the present invention reconstructs the influencing factor data. The reconstructed data can characterize the characteristics of the influencing factors to the greatest extent and avoid false alarms. Moreover, the influencing factor data is reconstructed by reconstructing the model. When the influencing factor data of the box deviates during actual work, it can be used as a basis for determining whether the box is abnormal.

[0055] Optionally, the reconstruction model adopts autoencoder AE, variational autoencoder VAE, PCA reconstruction, and TransFormer algorithm.

[0056] In this embodiment, three reconstruction algorithms can be selected according to actual needs to establish a reconstruction model, specifically:

[0057] a. Autoencoder AE: It is an unsupervised learning model. In machine learning and deep learning, its main goal is to learn how to represent data from the input space into a compact representation in the latent variable space and restore it back to the original input space as accurately as possible;

[0058] b. PCA reconstruction algorithm: Principal Component Analysis (PCA) is a dimensionality reduction method widely used in statistics, machine learning and data analysis;

[0059] c. Variational Autoencoder VAE: is an unsupervised learning method for generating and understanding the underlying structure of data. It combines the concepts of autoencoder (AE) and Bayesian reasoning, and can compress and reconstruct input data, and can also generate new data samples. VAE is particularly good at discovering low-dimensional representations from high-dimensional data, which makes them widely used in many fields, such as image generation, semantic embedding, and anomaly detection.

[0060] In some embodiments, in step S400, the regression model is a predictive modeling technique that studies the relationship between the dependent variable (target) and the independent variable (predictor) and is used for predictive analysis, time series models, and discovering causal relationships between variables. Optionally, the regression model can be one of a linear regression model, a decision tree regression model, a support vector machine regression model, a K nearest neighbor regression model, a random forest regression model, an Adaboost regression model, a gradient enhanced random forest regression model, a bagging regression model, and an Extra TREE regression model.

[0061] In some embodiments, the input factors of the regression model are influencing factor data, and the output result of the regression model is the liquid level difference.

[0062] In this embodiment, after a fully trained tank anomaly detection model is obtained, the influencing factor data of the current tank is obtained, and after the influencing factor data of the current tank is input into the tank anomaly detection model, the predicted liquid level difference can be quickly obtained.

[0063] In some embodiments, the step S600 specifically includes:

[0064] Acquire actual working data of the current cabinet, and compare the actual working data with the predicted working data to obtain a residual between the actual working data and the predicted working data;

[0065] The working data abnormality alarm threshold is obtained according to the training result, the residual is compared with the working data abnormality alarm threshold, and based on the comparison result, it is determined whether the box is abnormal.

[0066] In this embodiment, a liquid level difference normal alarm threshold is preset, and the liquid level difference normal alarm threshold is the maximum value of the error between the actual value and the predicted value, that is, within this error range, it indicates that there is no abnormality in the tank. Therefore, after the predicted liquid level difference is obtained through the model, the predicted liquid level difference is compared with the actual liquid level difference to obtain the difference between the two, and then the difference is compared with the preset liquid level difference normal alarm threshold. If the liquid level difference normal alarm threshold is exceeded, it indicates that there is an abnormality in the tank. Otherwise, it indicates that there is no abnormality in the tank.

[0067] In some embodiments, the method further comprises:

[0068] When it is determined that the box is abnormal, a corresponding alarm prompt message is issued.

[0069] In this embodiment, when it is determined that an abnormality occurs in the box, an alarm prompt message is issued, and the alarm prompt message includes factors such as the specific situation of the abnormality of the box and the cause of the abnormality of the box, so that the staff can quickly respond to the alarm prompt message.

[0070] In some embodiments, the method further comprises:

[0071] When it is determined that the cause of the cabinet abnormality alarm is a false alarm or an under-alarm, the latest historical data is obtained, and the cabinet abnormality detection model is updated using the latest historical data.

[0072] In this embodiment, the above models are all box anomaly detection models established based on historical operation data, and the accuracy of the models is restricted by the data. Since the historical operation data of the unit fails to cover all normal or abnormal states, when the anomaly detection model developed based on historical operation data encounters a new normal or abnormal state, false alarms or missed alarms will occur. If the previous model is used for detection at this time, the false alarm rate or missed alarm rate will increase. Therefore, the present invention adds an online update process, uses historical data for verification, and makes adjustments based on the new data.

[0073] Specifically, the alarm prompt information is first manually analyzed and judged. If it is judged to be a false alarm, it means that there may be a new abnormal pattern, and the new pattern does not exist in the historical data. Correspondingly, it is necessary to add such samples to the training samples, and retrain and update the model based on such samples; if it is judged to be a missed alarm, it means that there is a situation that should be identified by the existing nuclear power important parameter abnormality identification system, but it is not actually identified. In this case, the features of the missed alarm will be extracted, and an abnormal feature library will be established. Subsequently, new training will be carried out based on the abnormal sample data in the abnormal feature library to obtain an updated cabinet abnormality detection model.

[0074] Another embodiment of the present invention provides a nuclear power plant box abnormality detection device, see Figure 2 , the nuclear power plant box abnormality detection device comprises:

[0075] The influence factor determination module is used to establish a mass conservation equation of the box, and determine the influence factor of the box based on the mass conservation equation;

[0076] The potential impact factor determination module is used to preliminarily determine the impact factor of the box based on the actual instrument deployment situation and business knowledge;

[0077] The training set processing module is used to obtain a number of historical working data of the box, analyze the correlation between the potential influencing factors and the historical working data using a data analysis method to determine the final influencing factors of the heat exchanger, and generate a corresponding training data set based on the historical working data and the final determined influencing factors;

[0078] A reconstruction model training module is used to establish a reconstruction model and train the reconstruction model using the training data set to obtain a fully trained box reconstruction model and early warning thresholds of various influencing factors;

[0079] The regression model training module is used to establish a regression model, and train the regression model using the training data set to obtain a fully trained box regression model and a warning threshold value of the regression output parameter;

[0080] The prediction module is used to obtain the influencing factor data of the current box, and input the current influencing factor data into the reconstruction model and the regression model for reconstruction to obtain two sets of prediction data of the box;

[0081] The abnormality judgment module is used to obtain actual operation data of the current cabinet, and compare the actual operation data with the predicted data to judge whether the cabinet has an abnormality.

[0082] It should be noted that the module referred to in the present invention refers to a series of computer program instruction segments that can perform specific functions, which is more suitable for describing the execution process of nuclear power plant box abnormality detection than a program. For the specific implementation method of each module, please refer to the corresponding method embodiment above, which will not be repeated here.

[0083] In some embodiments, the influencing factor data include the inlet flow rate and temperature of the box, the outlet flow rate and temperature of the box, the top pressure of the box, the liquid temperature, and the gas phase pressure.

[0084] In some embodiments, the inlet mass of the tank is obtained by calculating the tank inlet volume flow rate, the tank inlet liquid density and the tank inlet liquid pressure, and the outlet mass of the tank is obtained by calculating the tank outlet volume flow rate, the tank outlet liquid density and the tank outlet liquid pressure, wherein the tank inlet liquid density and the tank inlet liquid pressure are related to the tank inlet liquid temperature.

[0085] In some embodiments, the reconstruction model adopts reconstruction algorithms such as principal component analysis PCA reconstruction, autoencoder AE, variational autoencoder VAE, etc.

[0086] In some embodiments, the regression model is one of a linear regression model, a polynomial regression, a decision tree regression model, a support vector machine regression model, a K nearest neighbor regression model, a random forest regression model, an Adaboost regression model, a gradient enhanced random forest regression model, a bagging regression model, and an Extra TREE regression model.

[0087] In some embodiments, the abnormality determination module is specifically used to:

[0088] Acquire actual working data of the current cabinet, and compare the actual working data with the predicted working data to obtain a residual between the actual working data and the predicted working data;

[0089] The working data abnormality alarm threshold is obtained according to the training result, the residual is compared with the working data abnormality alarm threshold, and based on the comparison result, it is determined whether the box is abnormal.

[0090] In some embodiments, the apparatus further comprises:

[0091] The online update module is used to obtain the latest historical data when it is determined that the cause of the cabinet abnormality alarm is a false alarm or an under-alarm, and to update the cabinet abnormality detection model using the latest historical data.

[0092] Another embodiment of the present invention provides an electronic device, such as Figure 3 As shown, the electronic device 10 includes:

[0093] One or more processors and memory 120, Figure 3 A processor 110 is used as an example for description. The processor 110 and the memory 120 may be connected via a bus or other means. Figure 3 The example of connecting through bus is taken in the following.

[0094] The processor 110 is used to complete various control logics of the electronic device 10. It can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a single-chip microcomputer, an ARM (Acorn RISC Machine) or other programmable logic device, a discrete gate or transistor logic, a discrete hardware component, or any combination of these components. In addition, the processor 110 can also be any traditional processor, microprocessor or state machine. The processor 110 can also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors combined with a DSP and / or any other such configuration.

[0095] The memory 120 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions corresponding to the nuclear power plant box abnormality detection method in the embodiment of the present invention. The processor 110 executes various functional applications and data processing of the electronic device 10 by running the non-volatile software programs, instructions and units stored in the memory 120, that is, the nuclear power plant box abnormality detection method in the above method embodiment is implemented.

[0096] The memory 120 may include a program storage area and a data storage area, wherein the program storage area may store an operating platform, an application required for at least one function; the data storage area may store data created according to the use of the electronic device 10, etc. In addition, the memory 120 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 120 may optionally include a memory remotely arranged relative to the processor 110, and these remote memories may be connected to the electronic device 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0097] One or more units are stored in the memory 120, and when executed by one or more processors 110, the nuclear power plant box abnormality detection method in any of the above method embodiments is executed, for example, the above described Figure 1 The method comprises steps S100 to S600.

[0098] Another embodiment of the present invention provides a computer-readable storage medium, which stores computer-executable instructions. The computer-executable instructions are executed by one or more processors, for example, to execute the above-described Figure 1 The method comprises steps S100 to S600.

[0099] In summary, a method, device, electronic device and storage medium for detecting abnormality of a nuclear power plant box shown in the present invention first establishes a mass conservation equation for the box, and determines the influencing factor of the box based on the mass conservation equation; then obtains several sections of historical working data of the box, and finally determines its influencing shadow based on the data analysis method and completes the processing of the required features, and establishes a training set; establishes a reconstruction model, and uses the training data set to train the regression model to obtain a well-trained box reconstruction model and early warning thresholds for each influencing factor; establishes a regression model, and uses the training data set to train the regression model to obtain a well-trained box regression model and early warning thresholds for regression output parameters; then obtains the influencing factor data of the current box, and inputs the current influencing factor data into the reconstruction model and the regression model to obtain the predicted working data of the box; finally obtains the actual working data of the current box, and compares the actual working data with the predicted working data to determine whether the box is abnormal. The present invention obtains the influencing factors of the tank abnormality and adopts reconstruction and regression algorithms to build an abnormality detection model for the tank. Then, the liquid level change of the tank can be directly predicted by the abnormality detection model. After comparing with the actual liquid level change, it can quickly and accurately determine whether the tank has an abnormality.

[0100] The above embodiments are only used to help understand the method and core idea of ​​the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made to the present invention without departing from the principles of the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

[0101] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may 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 the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting abnormality of a nuclear power plant box, characterized in that: The following steps are involved: Step S100: establishing a mass conservation equation for a box, and determining an ideal influencing factor for the box based on the mass conservation equation; Step S200: preliminarily determine the influencing factor of the box based on the actual instrument deployment situation and business knowledge; Step S300: obtaining a number of historical working data of the cabinet, using a data analysis method to analyze the correlation between the potential influencing factors and the historical working data to determine the final influencing factors of the cabinet, and generating a corresponding training data set based on the historical working data and the final determined influencing factors; Step S400: establishing a reconstruction model, and using the training data set to train the reconstruction model to obtain a fully trained box reconstruction model and warning thresholds of various influencing factors; Step S500: establishing a regression model, and using the training data set to train the regression model to obtain a fully trained box regression model and a warning threshold value of a regression output parameter; Step S600: Obtain the influencing factor data of the current box, and input the current influencing factor data into the reconstruction model and the regression model to obtain two sets of prediction data for the box; Step S700: Acquire actual operation data of the current cabinet, and compare the actual operation data with the predicted data to determine whether the cabinet is abnormal.

2. The method for detecting abnormality of a nuclear power plant box according to claim 1, characterized in that: The historical operation data at least includes historical liquid level difference data and its corresponding influencing factor data, and the influencing factor data includes the inlet flow and temperature of the box, the outlet flow and temperature of the box, the top pressure of the box, the liquid temperature, and the gas phase pressure.

3. The method for detecting abnormality of a nuclear power plant box according to claim 1, characterized in that: In establishing the reconstruction model and the regression model, Guo Chen considered the influence of temperature and pressure on density, and obtained the inlet mass of the box by calculating the box inlet volume flow rate, the box inlet liquid density and the box inlet liquid pressure; the outlet mass of the box was obtained by calculating the box outlet volume flow rate, the box outlet liquid density and the box outlet liquid pressure, wherein the box inlet liquid density and the box inlet liquid pressure are related to the box inlet liquid temperature.

4. The method for detecting abnormality of a nuclear power plant box according to claim 1, characterized in that: The reconstruction model adopts principal component analysis PCA reconstruction, autoencoder AE, and variational autoencoder VAE reconstruction algorithms.

5. The method for detecting abnormality of a nuclear power plant box according to claim 1, characterized in that: The regression model is one of a linear regression model, a polynomial regression, a decision tree regression model, a support vector machine regression model, a K nearest neighbor regression model, a random forest regression model, an Adaboost regression model, a gradient enhanced random forest regression model, a bagging regression model, and an Extra TREE regression model.

6. The method for detecting abnormality of a nuclear power plant box according to claim 1, characterized in that: The step S700 is specifically as follows: Acquire actual working data of the current cabinet, and compare the actual working data with the predicted working data to obtain a residual between the actual working data and the predicted working data; Obtaining a working data abnormality alarm threshold according to the training result, comparing the residual with the working data abnormality alarm threshold, and judging whether the box is abnormal based on the comparison result; When it is determined that the box is abnormal, a corresponding alarm prompt message is issued, and the alarm prompt message includes the specific situation of the abnormality of the box and the cause of the abnormality of the box.

7. The method for detecting abnormality of a nuclear power plant box according to claim 1, characterized in that: Also includes: When it is determined that the cause of the cabinet abnormality alarm is a false alarm or an under-alarm, the latest historical data is obtained, and the cabinet abnormality detection model is updated using the latest historical data.

8. A nuclear power plant box abnormality detection device, characterized in that: include: An influence factor preliminary determination module, used to establish a mass conservation equation for the box, and determine the influence factor of the box based on the mass conservation equation; Potential impact factor determination module, used to preliminarily determine the impact factor of the box based on the actual instrument deployment situation and business knowledge; A training set processing module is used to obtain a number of historical working data of the box, analyze the correlation between the potential influencing factors and the historical working data using a data analysis method to determine the final influencing factors of the heat exchanger, and generate a corresponding training data set based on the historical working data and the final determined influencing factors; A reconstruction model training module is used to establish a reconstruction model and train the reconstruction model using the training data set to obtain a fully trained box reconstruction model and warning thresholds of various influencing factors; A regression model training module is used to establish a regression model and train the regression model using the training data set to obtain a fully trained box regression model and a warning threshold value of a regression output parameter; A prediction module is used to obtain the influencing factor data of the current box, and input the current influencing factor data into the reconstruction model and the regression model for reconstruction to obtain two sets of prediction data of the box; The abnormality judgment module is used to obtain actual operation data of the current cabinet, and compare the actual operation data with the predicted data to judge whether the cabinet has an abnormality.

9. An electronic device, characterized in that: include: Processor and memory; The memory stores a computer program executable by the processor; When the processor executes the computer program, the steps of the nuclear power plant box abnormality detection method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the nuclear power plant box abnormality detection method according to any one of claims 1 to 7.

Images