Nuclear power station pump 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 pump body anomaly detection method and device. The method comprises the steps that firstly, a pump body characteristic curve is obtained, and influence factors of a pump body are determined based on the pump body characteristic curve, instrument arrangement and business knowledge; then acquiring a plurality of sections of historical working data of the pump body, and establishing a training set through a data analysis method; training of a regression model, a reconstruction model and a statistic model is completed based on the training set, and an early warning threshold value is obtained based on residual error or statistic distribution; and finally, inputting the current related data into the reconstruction algorithm and the statistical magnitude model or the regression model in the state to obtain the statistical magnitude or the predicted value of the pump in the current state, and comparing the statistical magnitude or the predicted value with an early warning threshold to judge whether abnormity exists or not. According to the method, the abnormality of the pump body is judged by selectively applying a proper model, and the method is simple and convenient.

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 pump. 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 pump is an important power device in nuclear power plants and a driving mechanism for the flow of liquids. If a pump fails, it may cause unexpected downtime of the entire system, resulting in serious economic losses. Therefore, timely detection and elimination of pump failures is of great practical significance.

[0004] Currently, the detection of abnormal conditions during pump 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 device for detecting abnormalities in nuclear power plant pump bodies, using regression algorithm and reconstruction algorithm to calculate the predicted operating parameters of the pump body, and compare them with the actual operating parameters, so as to timely and accurately identify abnormal parameters during pump operation; at the same time, a statistical model is combined to realize abnormal detection of the overall state of the pump.

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

[0007] Step S100: obtaining a pump characteristic curve, and preliminarily determining the influencing factors of the pump based on the pump characteristic curve, business knowledge, and instrument layout;

[0008] Step S200: obtaining several sections of historical working data of the pump body, classifying the historical working data according to the types of the historical working data, and analyzing each type of historical working data by using a data analysis method to determine the final influencing factor of the pump body, and then generating a steady-state training set and a changing-state training set;

[0009] Step S300: establishing a first model, and training the first model using the steady-state training set to obtain a first prediction model of pump-related parameters and a first warning threshold that are fully trained, wherein the first model includes at least one of a reconstruction model and a statistical model;

[0010] Step S400: establishing a second model, using the change state training set to train the regression model to obtain a fully trained second prediction model and a second warning threshold, wherein the second model is a regression model;

[0011] Step S500: Obtain the working data and influencing factor data of the current pump body, determine the state of the current pump body, and input the corresponding data of the current pump body into the first prediction model or the second prediction model under the state based on the current pump body state. After obtaining the prediction data of the current pump body under the current state, determine whether the pump body has an abnormality based on the first warning threshold or the second warning threshold.

[0012] In a specific implementation of the present invention, the reconstruction model uses principal component analysis PCA reconstruction, autoencoder AE, and variational autoencoder VAE.

[0013] In a specific embodiment of the present invention, the regression model is one of a polynomial regression model, 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.

[0014] In a specific implementation manner of the present invention, the statistical model adopts a principal component analysis algorithm.

[0015] In a specific embodiment of the present invention, it also includes:

[0016] When it is determined that the cause of the abnormal alarm of the pump body is a false alarm or an under-alarm, the latest historical data is obtained and the model in this state is updated using the latest historical data.

[0017] In a specific embodiment of the present invention, the working data includes at least temperature parameters, current parameters and vibration parameters, and the influencing factor data includes at least pump fluid temperature, volume flow, pump sealing fluid temperature, flow and cooling motor fluid temperature, flow.

[0018] In a specific implementation of the present invention, the temperature parameters include motor bearing temperature, coil temperature, and pump bearing temperature; the current parameters include parameters such as motor current; and the vibration parameters include bearing vibration.

[0019] The present invention provides a nuclear power plant pump body abnormality detection device, comprising:

[0020] An influencing factor preliminary determination module obtains a pump characteristic curve and preliminarily determines the influencing factor of the pump based on the pump characteristic curve, business knowledge, and instrument layout;

[0021] A training set generation module is used to obtain several sections of historical working data of the pump body, classify the historical working data according to the types of the historical working data, and analyze each type of historical working data using a data analysis method to determine the final influencing factor of the pump body, and then generate a steady-state training set and a changing-state training set;

[0022] A first model building module, used to build a first model, and train the first model using the steady-state training set to obtain a first prediction model of pump-related parameters and a first warning threshold that are fully trained, wherein the first model includes at least one of a reconstruction model and a statistical model;

[0023] A second model building module, used to build a second model, using the change state training set to train the regression model to obtain a fully trained second prediction model and a second warning threshold, wherein the second model is a regression model;

[0024] An abnormality detection module is used to obtain the working data and influencing factor data of the current pump body, judge the state of the current pump body, and input the corresponding data of the current pump body into the first prediction model or the second prediction model under this state based on the state of the current pump body. After obtaining the prediction data of the current pump body under the current state, it is judged whether the pump body has an abnormality based on the first warning threshold or the second warning threshold.

[0025] The present invention provides an electronic device, comprising: a processor and a memory;

[0026] The memory stores a computer program executable by the processor;

[0027] When the processor executes the computer program, the steps in the method for detecting abnormality of a nuclear power plant pump body are implemented.

[0028] 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 pump body abnormality detection method.

[0029] Compared with the prior art, the nuclear power plant pump abnormality detection method and device of the present invention obtains the influencing factors of the pump abnormality, and respectively uses statistical algorithm, reconstruction model and regression algorithm to construct three abnormality detection models of the pump body, and then selectively uses the appropriate model to judge the abnormality of the pump body, which is simple and convenient. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A flow chart showing a method for detecting abnormalities in a nuclear power plant pump;

[0031] Figure 2A schematic diagram showing the functional modules of a nuclear power plant pump abnormality detection device;

[0032] Figure 3 A schematic diagram showing the hardware structure of an electronic device;

[0033] In the figure, 10 is an electronic device, 110 is a processor, and 120 is a memory. DETAILED DESCRIPTION

[0034] 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.

[0035] An embodiment of the present invention provides a method for detecting abnormality of a nuclear power plant pump body, such as Figure 1 As shown, the following steps are included:

[0036] S100: Acquire a pump characteristic curve, and preliminarily determine an influencing factor of the pump based on the pump characteristic curve, business knowledge, and instrument layout;

[0037] S200: Acquire several sections of historical working data of the pump body, classify the historical working data according to the types of the historical working data, and analyze each type of historical working data using a data analysis method to determine the final influencing factor of the pump body, and then generate a steady-state training set and a variable-state training set;

[0038] S300: Establishing a first model, and training the first model using the steady-state training set to obtain a first prediction model of pump-related parameters that is fully trained and a first warning threshold, wherein the first model includes at least one of a reconstruction model and a statistical model;

[0039] S400: Establishing a second model, using the change state training set to train the regression model to obtain a fully trained second prediction model and a second warning threshold, wherein the second model is a regression model;

[0040] S500: Obtain the working data and influencing factor data of the current pump body, determine the state of the current pump body, and input the corresponding data of the current pump body into the first prediction model or the second prediction model in this state based on the state of the current pump body. After obtaining the prediction data of the current pump body in the current state, determine whether the pump body has an abnormality based on the first warning threshold or the second warning threshold.

[0041] In this embodiment, the pump characteristic curve is first obtained, and the influencing factor of the pump is determined based on the pump characteristic curve, instrument layout and business knowledge; then several sections of historical working data of the pump are obtained, and a training set is established through data analysis methods; the training of the regression model, reconstruction model and statistical model is completed based on the training set, and the warning threshold is obtained based on the residual or statistical distribution; finally, the current relevant data is input into the reconstruction algorithm and statistical model or regression model in this state, and the statistics or predicted values ​​of the pump in the current state are obtained, and it is judged whether there is an abnormality by comparing with the warning threshold. The present invention obtains the influencing factors of the pump abnormality, respectively uses the statistical algorithm, reconstruction model and regression algorithm to construct three abnormality detection models for the pump body, and then selectively uses the appropriate model to judge the abnormality of the pump body, which is simple and convenient.

[0042] In some embodiments, the abnormality detection method described in the present invention will reflect: 1) whether the pump parameters are consistent with historical values ​​under the same conditions; 2) whether the change of the pump current is reasonable in the non-startup transient state; 3) whether the pump as a whole may have an abnormality.

[0043] Specifically, the influencing factors of the pump body operating parameters are first determined, and then the pump body abnormality detection model can be modeled based on the influencing factors. Optionally, the embodiment of the present invention determines the influencing factors based on the characteristic curve of the pump body, and the characteristic curve of the pump body reflects the working characteristics of the pump body. Optionally, the working data at least includes the pump fluid temperature, volume flow rate, pump sealing fluid temperature, flow rate, and cooling motor fluid temperature, flow rate, etc.

[0044] In this embodiment, the temperature parameters of the pump body include parameters such as the motor bearing temperature, the coil temperature, and the pump bearing temperature; the current parameters of the pump body include parameters such as the motor current; the vibration parameters of the pump body include parameters such as the bearing vibration, and the bearing vibrations are respectively X-direction vibrations and Y-direction vibrations. The working fluid parameters include parameters such as the working fluid temperature, flow rate, and pressure; the sealing parameters include parameters such as the sealing flow rate and the sealing temperature. For example, by analyzing the characteristic curve, it is found that the pump current continues to rise significantly during the cooling process; the opposite process can also be seen during the heating process. Therefore, when modeling, the effect of temperature on the pump current needs to be considered.

[0045] In some embodiments, in step S200, the abnormality detection model is constructed based on knowledge and experience, and the historical operation data of the pump body is obtained according to the operation experience of the nuclear power plant. The historical operation data at least includes the historical working parameters of the pump body and the corresponding influencing factor data, and then the historical operation data is used to predict whether the working parameters of the pump body are abnormal. Among them, for the historical working data, it is necessary to classify according to the data type, such as temperature sensor, vibration sensor, flow sensor, etc.; if there are many pump-related parameters, it is necessary to use clustering algorithms for classification, and use data analysis methods to determine the final influencing factors for each type of data and complete the required feature processing, and finally form the required steady-state training set and variable state training set, such as pump startup and pump shutdown processes. In addition, for more complex pumps, it is also necessary to divide the steady-state training set into multiple data subsets, such as the continuous rise or fall of current under steady-state operation.

[0046] In some embodiments, in step S300, 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. It should be noted that the present invention is provided with two training data sets, which are used to establish different models respectively. The first detection model is applicable to non-variable frequency pumps (except APA pumps) because the pump parameters do not change much after the pump is started. Therefore, the abnormality of the pump body can be determined by the reconstruction algorithm and the statistical algorithm. The second detection model is applicable to pump bodies that are more obviously affected by the relevant parameters or are variable frequency pumps themselves. Therefore, the abnormality of the pump body is determined by the regression algorithm.

[0047] Optionally, the reconstruction model adopts reconstruction algorithms such as principal component analysis PCA reconstruction, autoencoder AE, variational autoencoder VAE, etc.

[0048] In this embodiment, three reconstruction algorithms can be selected according to actual needs to establish a reconstruction model, specifically:

[0049] 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. ;

[0050] b. PCA reconstruction algorithm: Principal Component Analysis (PCA) is a dimensionality reduction method widely used in statistics, machine learning and data analysis;

[0051] 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.

[0052] 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.

[0053] In some embodiments, the input factors of the regression model are influencing factor data, and the output results of the regression model are pump operating parameters.

[0054] In this embodiment, after obtaining a fully trained pump body abnormality detection model, the influencing factor data of the current pump body is obtained, and after the influencing factor data of the current pump body is input into the pump body abnormality detection model, the predicted pump body working data can be quickly obtained.

[0055] In some embodiments, the statistical model uses one of a cluster analysis algorithm, a correlation analysis algorithm, and a principal component analysis algorithm.

[0056] Among them, the cluster analysis algorithm is to classify sample individuals or indicator variables according to their characteristics, and find reasonable statistics to measure the similarity of things. Common algorithms include Q-type cluster analysis, R-type cluster analysis, etc. The correlation analysis algorithm is to study whether there is a certain dependency relationship between phenomena, and to explore the relevant direction and degree of correlation for specific dependent phenomena; the principal component analysis algorithm is to transform a group of related indicators into a group of new independent indicator variables, and use a few of the new indicator variables to comprehensively reflect the main information contained in the original multiple indicator variables.

[0057] In this embodiment, when the pump body is a non-variable frequency pump, a regression model and a statistical model are used to judge whether the pump body is abnormal. The output result of the statistical model, such as the variance, can be directly used to judge whether the output result is within the normal range to determine whether the pump body is abnormal.

[0058] When the pump body is a variable frequency pump or a pump body that is greatly affected by relevant parameters, a working data abnormal alarm threshold is preset. The working data abnormal 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 pump body. Therefore, after obtaining the working data through the model, the predicted working data is compared with the actual working data to obtain the difference between the two, and then the difference is compared with the preset working data abnormal alarm threshold. If it exceeds the working data abnormal alarm threshold, it indicates that there is an abnormality in the pump body. Otherwise, it indicates that there is no abnormality in the pump body.

[0059] In some embodiments, the method further comprises:

[0060] When it is determined that the pump body is abnormal, a corresponding alarm prompt message is issued.

[0061] In this embodiment, when it is determined that the pump body is abnormal, an alarm prompt information is issued, and the alarm prompt information includes factors such as the specific situation of the pump body abnormality and the cause of the pump body abnormality, so that the staff can quickly respond to the alarm prompt information.

[0062] In some embodiments, the method further comprises:

[0063] When it is determined that the cause of the pump abnormality alarm is a false alarm or an under-alarm, the latest historical data is obtained, and the first prediction model or the second prediction model is updated using the latest historical data.

[0064] In this embodiment, the above models are the first prediction model and the second prediction model established based on historical operation data, and the accuracy of the model is constrained by the data. Since the historical operation data of the unit fails to cover all normal or abnormal states, when the abnormal 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.

[0065] Specifically, the alarm prompt information is first analyzed and judged manually. 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 are extracted, and an abnormal feature library is established. Subsequently, new training is performed based on the abnormal sample data in the abnormal feature library to obtain the updated first prediction model and the second prediction model.

[0066] Another embodiment of the present invention provides a nuclear power plant pump abnormality detection device, referring to Figure 2 , the nuclear power plant pump body abnormality detection device includes:

[0067] An influencing factor preliminary determination module obtains a pump characteristic curve and preliminarily determines the influencing factor of the pump based on the pump characteristic curve, business knowledge, and instrument layout;

[0068] A training set generation module is used to obtain several sections of historical working data of the pump body, classify the historical working data according to the types of the historical working data, and analyze each type of historical working data using a data analysis method to determine the final influencing factor of the pump body, and then generate a steady-state training set and a changing-state training set;

[0069] A first model building module, used to build a first model, and train the first model using the steady-state training set to obtain a first prediction model of pump-related parameters and a first warning threshold that are fully trained, wherein the first model includes at least one of a reconstruction model and a statistical model;

[0070] A second model building module, used to build a second model, using the change state training set to train the regression model to obtain a fully trained second prediction model and a second warning threshold, wherein the second model is a regression model;

[0071] An abnormality detection module is used to obtain the working data and influencing factor data of the current pump body, judge the state of the current pump body, and input the corresponding data of the current pump body into the first prediction model or the second prediction model under this state based on the state of the current pump body. After obtaining the prediction data of the current pump body under the current state, it is judged whether the pump body has an abnormality based on the first warning threshold or the second warning threshold.

[0072] 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 pump body 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.

[0073] In some embodiments, the working data includes at least temperature parameters, current parameters and vibration parameters, and the influencing factor data includes at least pump fluid temperature, volume flow, pump sealing fluid temperature and flow, and cooling motor fluid temperature and flow.

[0074] In some embodiments, the reconstruction model uses principal component analysis PCA reconstruction, autoencoder AE, and variational autoencoder VAE.

[0075] In some embodiments, the regression model is one of a polynomial regression model, 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.

[0076] In some embodiments, the statistical model mainly adopts a cluster analysis algorithm.

[0077] In some embodiments, the method further comprises:

[0078] When it is determined that the cause of the abnormal alarm of the pump body is a false alarm or an under-alarm, the latest historical data is obtained and the model in this state is updated using the latest historical data.

[0079] In some embodiments, the apparatus further comprises:

[0080] The online update module is used to obtain the latest historical data when it is determined that the cause of the pump abnormal alarm is a false alarm or a missed alarm, and use the latest historical data to update the model in this state.

[0081] Another embodiment of the present invention provides an electronic device, such as Figure 3 As shown, the electronic device 10 includes:

[0082] One or more processors 110 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.

[0083] 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.

[0084] 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 pump body 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 pump body abnormality detection method in the above method embodiment is implemented.

[0085] 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.

[0086] One or more units are stored in the memory 120, and when executed by one or more processors 110, the nuclear power plant pump body 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 S500.

[0087] 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 S500.

[0088] In summary, the present invention shows a method, device, electronic device and storage medium for detecting abnormalities in a nuclear power plant pump body. First, the pump body characteristic curve is obtained, and the influencing factor of the pump body is determined based on the pump body characteristic curve, instrument layout and business knowledge; then several sections of historical working data of the pump body are obtained, and a training set is established through a data analysis method; based on the training set, the training of the regression model, reconstruction model and statistical model is completed, and the warning threshold is obtained based on the residual or statistical distribution; finally, the current relevant data is input into the reconstruction algorithm and the statistical model or regression model in this state, and the statistics or predicted values ​​of the pump in the current state are obtained, and the presence of abnormalities is judged by comparing with the warning threshold. The present invention obtains the influencing factors of the pump body abnormality, respectively uses the statistical algorithm, reconstruction model and regression algorithm to construct three abnormality detection models for the pump body, and then selectively uses the appropriate model to judge the abnormality of the pump body, which is simple and convenient.

[0089] 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.

[0090] 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 pump, characterized in that: The following steps are involved: Step S100: obtaining a pump characteristic curve, and preliminarily determining the influencing factors of the pump based on the pump characteristic curve, business knowledge, and instrument layout; Step S200: obtaining several sections of historical working data of the pump body, classifying the historical working data according to the types of the historical working data, and analyzing each type of historical working data by using a data analysis method to determine the final influencing factor of the pump body, and then generating a steady-state training set and a changing-state training set; Step S300: establishing a first model, and training the first model using the steady-state training set to obtain a first prediction model of pump-related parameters and a first warning threshold that are fully trained, wherein the first model includes at least one of a reconstruction model and a statistical model; Step S400: establishing a second model, using the change state training set to train the regression model to obtain a fully trained second prediction model and a second warning threshold, wherein the second model is a regression model; Step S500: Obtain the working data and influencing factor data of the current pump body, determine the state of the current pump body, and input the corresponding data of the current pump body into the first prediction model or the second prediction model under the state based on the current pump body state. After obtaining the prediction data of the current pump body under the current state, determine whether the pump body has an abnormality based on the first warning threshold or the second warning threshold.

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

3. The method for detecting abnormality of a nuclear power plant pump according to claim 1, characterized in that: The regression model is one of a polynomial regression model, 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 ExtraTREE regression model.

4. The method for detecting abnormality of a nuclear power plant pump according to claim 1, characterized in that: The statistical model adopts a principal component analysis algorithm.

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

6. The method for detecting abnormality of a nuclear power plant pump according to claim 1, characterized in that: The working data at least include temperature parameters, current parameters and vibration parameters, and the influencing factor data at least include pump fluid temperature, volume flow, pump sealing fluid temperature, flow, and cooling motor fluid temperature, flow.

7. The method for detecting abnormality of a nuclear power plant pump according to claim 6, characterized in that: The temperature parameters include motor bearing temperature, coil temperature, and pump bearing temperature; the current parameters include parameters such as motor current, and the vibration parameters include bearing vibration.

8. A nuclear power plant pump body abnormality detection device, characterized in that: include: An influencing factor preliminary determination module obtains a pump characteristic curve and preliminarily determines the influencing factor of the pump based on the pump characteristic curve, business knowledge, and instrument layout; A training set generation module is used to obtain several sections of historical working data of the pump body, classify the historical working data according to the types of the historical working data, and analyze each type of historical working data using a data analysis method to determine the final influencing factor of the pump body, and then generate a steady-state training set and a changing-state training set; A first model building module, used to build a first model, and train the first model using the steady-state training set to obtain a first prediction model of pump-related parameters and a first warning threshold that are fully trained, wherein the first model includes at least one of a reconstruction model and a statistical model; A second model building module, used to build a second model, using the change state training set to train the regression model to obtain a fully trained second prediction model and a second warning threshold, wherein the second model is a regression model; An abnormality detection module is used to obtain the working data and influencing factor data of the current pump body, judge the state of the current pump body, and input the corresponding data of the current pump body into the first prediction model or the second prediction model under this state based on the state of the current pump body. After obtaining the prediction data of the current pump body under the current state, it is judged whether the pump body has an abnormality based on the first warning threshold or the second warning threshold.

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 pump 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 pump body abnormality detection method according to any one of claims 1 to 7.

Images