Nuclear power abnormal event intelligent identification and early warning method

Abstract

The invention particularly relates to a nuclear power abnormal event intelligent identification and early warning method. The method comprises the following steps: preprocessing collected nuclear power historical operation data; performing big data processing and analysis on the preprocessed nuclear power historical operation data, extracting key features and rule information of nuclear power abnormal events, and forming training set data; training a machine learning model through the training set data, and constructing a nuclear power abnormal event recognition model; monitoring nuclear power operation data in real time through a nuclear power abnormal event identification model, and identifying nuclear power abnormal events in the nuclear power operation data; when the nuclear power abnormal event recognition model recognizes a nuclear power abnormal event, early warning information is sent to related personnel; and starting emergency measures according to a preset emergency response plan. The invention further relates to a nuclear power abnormal event intelligent identification and early warning device, computer equipment and a storage medium. According to the invention, based on big data analysis and machine learning technologies, accurate capture and real-time early warning of nuclear power abnormal events are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of nuclear power abnormal event identification and early warning, and in particular to a method, device, computer equipment and storage medium for intelligent identification and early warning of nuclear power abnormal events. Background Art

[0002] In the field of traditional nuclear power operations, although experience feedback systems have been established to record and analyze abnormal events, these systems are often limited to simple data storage, data retrieval, and post-event review, and fail to fully tap their potential in accurate identification of abnormal events and real-time warning. Although the abnormal event data manually recorded by users is rich, due to the lack of efficient data mining and intelligent analysis methods, the value of the data has not been fully mined, resulting in low system utilization and failure to capture and warn of potential risks in a timely manner. Traditional anomaly identification methods mostly rely on manually set thresholds and comprehensive judgments of multiple statistical indicators. The process is cumbersome and easily affected by subjective factors, making it difficult to accurately capture the complex and changeable nuclear power operation status. Summary of the invention

[0003] Based on this, it is necessary to provide a method, device, computer equipment and storage medium for intelligent identification and early warning of nuclear power abnormal events in order to address the problem that the existing nuclear power experience feedback system is difficult to achieve accurate capture and real-time early warning of nuclear power abnormal events. Based on big data analysis and machine learning technology, it can achieve accurate capture and real-time early warning of nuclear power abnormal events.

[0004] In order to solve the above technical problems, the present invention provides a method for intelligent identification and early warning of abnormal nuclear power events, comprising the following steps:

[0005] S101, collecting historical operation data of nuclear power; preprocessing the collected historical operation data of nuclear power;

[0006] S102, performing big data processing and analysis on the pre-processed nuclear power historical operation data, extracting key features and regular information of nuclear power abnormal events, and forming training set data;

[0007] S103, training the machine learning model through the training set data to build a nuclear power abnormal event recognition model;

[0008] S104, monitoring nuclear power operation data in real time through a nuclear power abnormal event recognition model, and identifying nuclear power abnormal events in the nuclear power operation data;

[0009] S105. When the nuclear power abnormal event recognition model recognizes a nuclear power abnormal event, a warning message is sent to relevant personnel; and emergency measures are initiated according to a preset emergency response plan.

[0010] In the present invention, in S101, the historical operation data of nuclear power plants include user reported data, equipment operation data, environmental monitoring data and safety system status data.

[0011] In the present invention, in S101, the collected nuclear power historical operation data is preprocessed, including the following steps:

[0012] Clean the historical operation data of nuclear power plants to remove duplicate values, missing values ​​or abnormal values;

[0013] Integrate the cleaned nuclear power historical operation data and integrate the nuclear power historical operation data from different data sources into a unified format;

[0014] Transform the integrated nuclear power historical operation data to meet the requirements of the machine learning model;

[0015] The transformed historical nuclear power operation data is regulated.

[0016] In the present invention, in S102, the pre-processed nuclear power historical operation data is subjected to big data processing and analysis to extract key features and regular information of nuclear power abnormal events, including the following steps:

[0017] The pre-processed nuclear power historical operation data is stored in a database;

[0018] Perform big data processing and analysis on the historical nuclear power operation data stored in the database to extract the key characteristics and regular information of nuclear power abnormal events.

[0019] In the present invention, big data processing and analysis are performed on the historical operation data of nuclear power plants stored in a database to extract key features and regular information of abnormal nuclear power events, including the following steps:

[0020] In the historical operation data of nuclear power plants, the data reported by users are unstructured data, and time series data are further generated for the data reported by users; the equipment operation data, environmental monitoring data and safety system status data are structured data;

[0021] Analyze the historical operation data of nuclear power through time series analysis model, and deeply explore the key regular information of nuclear power abnormal events in the historical operation data of nuclear power;

[0022] Extract key feature information of nuclear power abnormal events from historical nuclear power operation data.

[0023] In the present invention, the time series analysis model includes ARIMA and LSTM; the key regularity information of nuclear power abnormal events in the nuclear power historical operation data includes the time variation trend and potential periodicity of the nuclear power abnormal events in the nuclear power historical operation data.

[0024] In the present invention, time series data is further generated for the user reported data, including the following steps: generating time series data by counting the number of abnormal events reported by the user.

[0025] In the present invention, S103, training a machine learning model with training set data to construct a nuclear power abnormal event recognition model includes the following steps:

[0026] S1031. According to the characteristics and identification requirements of nuclear power abnormal events, select a suitable machine learning model to build a nuclear power abnormal event identification model;

[0027] S1032, training the selected machine learning model with the training set data, setting the machine learning model parameters during the training process, and optimizing the machine learning model parameter settings through a cross-validation strategy; after the training is completed, a nuclear power abnormal event recognition model is obtained;

[0028] S1033. Evaluate the nuclear power abnormal event recognition model using the test set data; if the evaluation index does not meet the requirements, optimize the nuclear power abnormal event recognition model until the evaluation index meets the requirements.

[0029] In the present invention, one or more of the IForest model, CBLOF model, HBOS model, KNN model, LOF model, OCSVM model and AutoEncoder model are selected to construct a nuclear power abnormal event recognition model.

[0030] In the present invention, the optimization strategy of the nuclear power abnormal event identification model includes adjusting model parameters, improving feature engineering, introducing new machine learning models or adopting machine learning model fusion.

[0031] In the present invention, in S105, when the nuclear power abnormal event recognition model identifies a nuclear power abnormal event, the early warning mechanism is triggered, and early warning information is sent to relevant personnel through pop-up warnings, sound alarms and / or SMS notifications; according to the type and severity of the nuclear power abnormal event, a corresponding emergency response plan is preset; emergency measures include emergency shutdown, personnel evacuation and pollution control.

[0032] In order to solve the above technical problems, the present invention also provides an intelligent identification and early warning device for abnormal nuclear power events, comprising:

[0033] The data collection and preprocessing module is used to collect the historical operation data of nuclear power plants and preprocess the collected historical operation data of nuclear power plants;

[0034] The big data processing and analysis module is used to process and analyze the pre-processed nuclear power historical operation data, extract the key features and regular information for identifying nuclear power abnormal events, and form training set data;

[0035] The machine learning model building module is used to train the machine learning model through the training set data and build a nuclear power abnormal event recognition model;

[0036] An abnormal event recognition module is used to monitor nuclear power operation data in real time through a nuclear power abnormal event recognition model and to recognize nuclear power abnormal events in the nuclear power operation data;

[0037] The early warning and emergency response module is used to send early warning information to relevant personnel when the nuclear power abnormal event recognition model identifies a nuclear power abnormal event; and initiate emergency measures according to the preset emergency response plan.

[0038] In order to solve the above technical problems, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the above-mentioned method for intelligent identification and early warning of abnormal nuclear power events are implemented.

[0039] In order to solve the above technical problems, the present invention also provides a computer-readable storage medium, on which computer-readable instructions are stored. When the computer-readable instructions are executed, the steps of the above-mentioned method for intelligent identification and early warning of abnormal nuclear power events are implemented.

[0040] Beneficial technical effects of the present invention:

[0041] The method, device, computer equipment and storage medium for intelligent identification and early warning of abnormal nuclear power events of the present invention can process and analyze nuclear power operation data in real time, and timely discover and warn potential abnormal nuclear power events. This real-time performance enables operators to respond quickly and take effective measures to prevent the situation from deteriorating, thereby significantly reducing the probability of safety accidents. Through machine learning algorithms, complex patterns and features in historical nuclear power operation data are automatically learned to accurately identify abnormal nuclear power events. This accuracy reduces false alarms and missed alarms, making safety monitoring more reliable and effective. Through large data analysis and machine learning, historical nuclear power operation data are fully utilized to dig out hidden laws and features, thereby improving the overall utilization rate and value of historical nuclear power operation data. The abnormal nuclear power event identification model can not only identify abnormal nuclear power events, but also predict future trends based on historical data and current conditions, and provide decision makers with a variety of possible solutions and suggestions. This intelligent decision-making assistance capability can significantly improve the scientificity and accuracy of decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of an embodiment of the method for intelligent identification and early warning of abnormal nuclear power events of the present invention;

[0043] Figure 2 It is the logic diagram of IForest model;

[0044] Figure 3 It is the logic diagram of the Autoencoder model;

[0045] Figure 4 It is a schematic structural view of an embodiment of the intelligent identification and early warning device for abnormal nuclear power events of the present invention.

[0046] Figure 5 The figure is a schematic structural view of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used in the specification of the application are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned figure descriptions are intended to cover non-exclusive inclusions.

[0048] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0049] The technical solution of the present invention is clearly and completely described below in conjunction with the accompanying drawings and specific embodiments.

[0050] The present embodiment provides a method, device, computer equipment and storage medium for intelligent identification and early warning of abnormal events in nuclear power plants. By integrating massive nuclear power operation data, using big data analysis technology to deeply explore the correlation and pattern between nuclear power operation data, and combining the powerful pattern recognition and prediction capabilities of machine learning algorithms, automatic, rapid and accurate identification of abnormal events in nuclear power plants is achieved. It not only overcomes the limitation of single indicator identification of abnormal events in nuclear power plants, but also greatly improves the timeliness and accuracy of early warning of abnormal events in nuclear power plants, effectively utilizes rich experience feedback data resources that have been neglected in the past, promotes the intelligent upgrade of nuclear power operation management, and safeguards the safe and stable operation of nuclear power plants.

[0051] refer to Figure 1 , the method for intelligent identification and early warning of abnormal nuclear power events described in this embodiment includes the following steps: S101, collecting historical nuclear power operation data; preprocessing the collected historical nuclear power operation data;

[0052] S102, performing big data processing and analysis on the pre-processed nuclear power historical operation data, extracting key features and regular information of nuclear power abnormal events, and forming training set data;

[0053] S103, training the machine learning model through the training set data to build a nuclear power abnormal event recognition model;

[0054] S104, monitoring nuclear power operation data in real time through a nuclear power abnormal event recognition model, and identifying nuclear power abnormal events in the nuclear power operation data;

[0055] S105. When the nuclear power abnormal event recognition model identifies a nuclear power abnormal event, it sends a warning message to relevant personnel; and quickly initiates emergency measures according to the preset emergency response plan to prevent the situation from escalating.

[0056] Rapid assessment of abnormal nuclear power events provides a scientific basis for emergency decision-making and ensures the timeliness and effectiveness of emergency response.

[0057] In this embodiment, in S101, the historical operation data of nuclear power plants include user-reported data, equipment operation data, environmental monitoring data and safety system status data; the user-reported data include event description data and fault reports, which are collected in real time or periodically through electronic forms or databases; the equipment operation data include parameters such as temperature, pressure and flow monitored in real time by sensors, equipment operation logs and equipment historical fault data, which are collected through the equipment's own data acquisition system or dedicated data acquisition equipment.

[0058] Preprocessing the collected nuclear power historical operation data improves the quality of the nuclear power historical operation data and the efficiency of subsequent nuclear power historical operation data analysis. In this embodiment, in S101, preprocessing the collected nuclear power historical operation data includes the following steps:

[0059] Clean the historical operation data of nuclear power plants to remove duplicate values, missing values ​​or abnormal values;

[0060] Integrate the cleaned nuclear power historical operation data, and integrate the nuclear power historical operation data from different data sources into a unified format to ensure the consistency and integrity of the nuclear power historical operation data;

[0061] The integrated nuclear power historical operation data is normalized and encoded to meet the requirements of the machine learning model;

[0062] The transformed nuclear power historical operation data is regulated, and the amount of nuclear power historical operation data is reduced through data dimensionality reduction technology, thereby improving the efficiency of subsequent nuclear power historical operation data analysis.

[0063] Through the preprocessing of nuclear power historical operation data, a solid foundation has been laid for subsequent big data analysis and machine learning model training.

[0064] In this embodiment, in S102, the pre-processed nuclear power historical operation data is processed and analyzed for big data to extract key features and regular information of nuclear power abnormal events, including the following steps:

[0065] The pre-processed nuclear power historical operation data is stored in the database to ensure efficient access to the pre-processed nuclear power historical operation data;

[0066] Big data processing and analysis are performed on the historical nuclear power operation data stored in the database to extract the key characteristics and regular information of nuclear power abnormal events, laying a solid foundation for the subsequent construction of an efficient and accurate nuclear power abnormal event identification module.

[0067] Big data processing and analysis of pre-processed nuclear power historical operation data ensures in-depth mining and value maximization of nuclear power historical operation data, and improves the accuracy and timeliness of nuclear power abnormal event warnings.

[0068] In this embodiment, the historical operation data of nuclear power plants stored in the database is processed and analyzed for big data, and key features and regular information of abnormal nuclear power events are extracted, including the following steps:

[0069] In the historical operation data of nuclear power plants, the data reported by users are unstructured data, and time series data are further generated for the data reported by users; the equipment operation data, environmental monitoring data and safety system status data are structured data;

[0070] Analyze the historical operation data of nuclear power through the time series analysis model, and deeply explore the key regularity information of nuclear power abnormal events in the historical operation data of nuclear power; the time series analysis model includes ARIMA and LSTM; the key regularity information of nuclear power abnormal events in the historical operation data of nuclear power includes the time change trend and potential periodicity of nuclear power abnormal events in the historical operation data of nuclear power;

[0071] Extract key feature information of nuclear power abnormal events from historical nuclear power operation data.

[0072] In this embodiment, further generating time series data for the user reported data includes the following steps: generating time series data by counting the number of abnormal events reported by the user.

[0073] Constructing a nuclear power abnormal event recognition model through a machine learning model is the core step of intelligent recognition of nuclear power abnormal events, which directly determines the ability to accurately and efficiently identify potential nuclear power abnormal events. Constructing a nuclear power abnormal event recognition model through a machine learning model is an iterative optimization process. In this embodiment, S103, training a machine learning model through training set data to construct a nuclear power abnormal event recognition model includes the following steps:

[0074] S1031. According to the characteristics and identification requirements of nuclear power abnormal events, select a suitable machine learning model to build a nuclear power abnormal event identification model;

[0075] S1032, training the selected machine learning model with the training set data, setting the machine learning model parameters during the training process, and optimizing the machine learning model parameter settings through a cross-validation strategy; after the training is completed, a nuclear power abnormal event recognition model is obtained;

[0076] S1033. Evaluate the nuclear power abnormal event recognition model through independent test set data; if the evaluation index does not meet the requirements, optimize the nuclear power abnormal event recognition model until the evaluation index meets the requirements.

[0077] During the training process, it is necessary to set appropriate machine learning model parameters. For example, when selecting the IForest model, LOF model and / or OCSVM model to build a nuclear power abnormal event recognition model, it is necessary to select the number of trees in the IForest model, the neighborhood size in the LOF model, the kernel function and penalty term in the OCSVM model, etc. The selection of these parameters has a significant impact on the performance of the machine learning model. Therefore, it is necessary to optimize the machine learning model parameter settings through a cross-validation strategy.

[0078] Evaluation indicators can fully reflect the characteristics of the model. By comparing the performance of different models or model combinations, the model with the best performance can be selected for subsequent deployment. The evaluation indicator calculation formula is as follows:

[0079]

[0080] Among them, TP represents the number of samples correctly predicted as positive examples by the model; TN represents the number of samples correctly predicted as negative examples by the model; FP represents the number of samples that the model incorrectly predicts as positive examples; FN represents the number of samples that the model incorrectly predicts as negative examples; Accuracy represents accuracy; Precision represents precision; Recall represents recall; F1-score represents the harmonic mean of precision and recall.

[0081] In this embodiment, one or more of the IForest model, CBLOF model, HBOS model, KNN model, LOF model, OCSVM model and AutoEncoder model are selected to construct a nuclear power abnormal event recognition model.

[0082] The IForest model is also called the isolation forest model, the CBLOF model is also called the clustering-based local outlier factor model, the HBOS model is also called the histogram-based outlier score model, the KNN model is also called the K nearest neighbor model, the LOF model is also called the local outlier factor model, the OCSVM model is also called a first-class support vector machine model, and the AutoEncoder model is also called the autoencoder model.

[0083] Each machine learning model has its own unique advantages. The IForest model is good at quickly identifying anomalies, the LOF model can capture local density changes in data, the OCSVM model is suitable for situations where there are only normal samples, and the AutoEncoder model detects anomalies by reconstructing errors. Based on factors such as the performance of nuclear power historical operating data, computing resource limitations, and the interpretability of the model, one or more of these models can be selected and combined to build a nuclear power abnormal event recognition model.

[0084] The Forest model uses the idea of ​​random forest to isolate outliers by constructing a large number of binary trees. Outliers are easily isolated due to their low density, while normal points require multiple cuts to be isolated.

[0085] The LOF model is an unsupervised anomaly detection method that calculates the local density deviation of a given data point relative to its neighbors. It regards samples with much lower density than their neighbors as outliers.

[0086] The CBLOF model first uses a clustering algorithm to divide the data into multiple clusters, then distinguishes large clusters from small clusters by setting a threshold, and calculates the distance from each point to the nearest large cluster to identify outliers.

[0087] The HBOS model constructs a frequency histogram by dividing the data into multiple intervals according to features and calculating the number of samples in each interval, and then calculates an anomaly score for each sample. The higher the score, the more likely the sample is an anomaly.

[0088] The KNN model evaluates the degree of abnormality of a sample by calculating the distance between it and its K nearest neighbors. The larger the distance, the more likely the sample is an abnormal point.

[0089] The OCSVM model maps data samples to a high-dimensional feature space and solves an optimal hyperplane to achieve maximum separation between the target data and the coordinate origin, thereby identifying outliers.

[0090] The AutoEncoder model uses the learning ability of neural networks to identify abnormal data. It trains a neural network through unsupervised learning so that it can reconstruct input data. In anomaly detection, if the reconstruction error of a data point is much larger than the reconstruction error of a normal data point, the data point is considered an anomaly.

[0091] IForest model logic is as follows Figure 2 As shown in the figure, the IForest model constructs multiple isolated trees and uses the path length of the data point in the tree to detect outliers. The path of abnormal data points in the isolated tree is usually shorter, so it is easier to isolate. The isolation forest model calculates the anomaly score of each data point. For each sample x, the calculation formula of its anomaly score s(x,n) is as follows:

[0092]

[0093] Where E(h(x)) is the average path length of data point x in all trees, n is the sample size of training data, c(n) is the average path length of the tree, and the anomaly score ranges from [0,1]. The closer it is to 1, the easier it is for the sample to be isolated and the more likely it is to be anomaly data.

[0094] The Autoencoder model is a typical anomaly detection model based on neural networks. Figure 3 As shown, it includes an encoder and a decoder. The encoder is used to find the compressed representation of the given data, and the decoder is used to reconstruct the original input. If the input sample is: X = (X 1 , X 2 , …, X 10 ), the result after reconstruction by the Autoencoder model is: The restoration error MAE is: Then, when the restoration error is greater than the set threshold, it is marked as an outlier.

[0095] In this embodiment, the optimization strategy of the nuclear power abnormal event identification model includes adjusting model parameters, improving feature engineering, introducing new machine learning models, or adopting machine learning model fusion. For example, you can try to combine the advantages of the IForest model and the LOF model, and fuse the prediction results of the two by weighted voting to improve the overall recognition accuracy.

[0096] In this embodiment, in S105, when the nuclear power abnormal event recognition model identifies a nuclear power abnormal event, the early warning mechanism is triggered, and early warning information is sent to relevant personnel through pop-up warnings, sound alarms and / or SMS notifications; according to the type and severity of the nuclear power abnormal event, a corresponding emergency response plan is preset; emergency measures include but are not limited to emergency shutdown, personnel evacuation and pollution control to minimize accident losses.

[0097] refer to Figure 4 As an implementation of the above method, the present invention provides an embodiment of an intelligent identification and early warning device for abnormal nuclear power events. The embodiment of the device corresponds to the embodiment of the above method, and the device can be specifically applied to various electronic devices.

[0098] The intelligent recognition and early warning device for abnormal nuclear power events described in this embodiment includes:

[0099] The data collection and preprocessing module is used to collect the historical operation data of nuclear power plants and preprocess the collected historical operation data of nuclear power plants;

[0100] The big data processing and analysis module is used to process and analyze the pre-processed nuclear power historical operation data, extract the key features and regular information for identifying nuclear power abnormal events, and form training set data;

[0101] The machine learning model building module is used to train the machine learning model through the training set data and build a nuclear power abnormal event recognition model;

[0102] An abnormal event recognition module is used to monitor nuclear power operation data in real time through a nuclear power abnormal event recognition model and to recognize nuclear power abnormal events in the nuclear power operation data;

[0103] The early warning and emergency response module is used to send early warning information to relevant personnel when the nuclear power abnormal event recognition model identifies a nuclear power abnormal event; and quickly initiate emergency measures according to the preset emergency response plan to prevent the situation from escalating.

[0104] As an implementation of the above method, the present invention provides an embodiment of a computer device, and the embodiment of the computer device corresponds to the embodiment of the above method for intelligent identification and early warning of abnormal nuclear power events.

[0105] See also Figure 5The computer device described in this embodiment includes a memory, a processor and a network interface that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device with a memory, a processor and a network interface, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application-specific integrated circuits, programmable gate arrays, digital processors, embedded devices, etc.

[0106] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0107] The memory includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory, random access memory, static random access memory, read-only memory, electrically erasable programmable read-only memory, programmable read-only memory, magnetic memory, disk, optical disk, etc. In some embodiments, the memory may be an internal storage unit of the computer device, such as a hard disk or memory of the computer device. In other embodiments, the memory may also be an external storage device of the computer device, such as a plug-in hard disk, smart memory card, secure digital card, flash memory card, etc. equipped on the computer device. Of course, the memory may also include both the internal storage unit of the computer device and its external storage device. In this embodiment, the memory is generally used to store the operating system and various application software installed on the computer device, such as the computer-readable instructions of the above-mentioned nuclear power abnormal event intelligent identification and early warning method. In addition, the memory may also be used to temporarily store various types of data that have been output or are to be output.

[0108] In some embodiments, the processor may be a central processing unit, a controller, a microcontroller, a microprocessor, or other data processing chip. The processor is generally used to control the overall operation of the computer device. In this embodiment, the processor is used to run computer-readable instructions stored in the memory or process data, such as computer-readable instructions for running the above-mentioned nuclear power abnormal event intelligent identification and early warning method.

[0109] The network interface may include a wireless network interface or a wired network interface, which is generally used to establish a communication connection between the computer device and other electronic devices.

[0110] As an implementation of the above method, the present invention provides an embodiment of a computer-readable storage medium, and the embodiment of the computer-readable storage medium corresponds to the embodiment of the above-mentioned method for intelligent identification and early warning of abnormal nuclear power events.

[0111] The computer-readable storage medium described in this embodiment stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the above-mentioned method for intelligent identification and early warning of abnormal nuclear power events.

[0112] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for intelligent identification and early warning of abnormal nuclear power events, characterized in that: The following steps are involved: S101, collecting historical operation data of nuclear power; preprocessing the collected historical operation data of nuclear power; S102, performing big data processing and analysis on the pre-processed nuclear power historical operation data, extracting key features and regular information of nuclear power abnormal events, and forming training set data; S103, training the machine learning model through the training set data to build a nuclear power abnormal event recognition model; S104, monitoring nuclear power operation data in real time through a nuclear power abnormal event recognition model, and identifying nuclear power abnormal events in the nuclear power operation data; S105. When the nuclear power abnormal event recognition model recognizes a nuclear power abnormal event, a warning message is sent to relevant personnel; Initiate emergency measures according to the preset emergency response plan.

2. The method for intelligent identification and early warning of abnormal nuclear power events according to claim 1, characterized in that: Historical nuclear power operation data includes user-reported data, equipment operation data, environmental monitoring data and safety system status data.

3. The method for intelligent identification and early warning of abnormal nuclear power events according to claim 2 is characterized in that: In S101, the collected nuclear power historical operation data is preprocessed, including the following steps: Clean the historical operation data of nuclear power plants to remove duplicate values, missing values ​​or abnormal values; Integrate the cleaned nuclear power historical operation data and integrate the nuclear power historical operation data from different data sources into a unified format; Transform the integrated nuclear power historical operation data to meet the requirements of the machine learning model; The transformed nuclear power historical operation data is regulated.

4. The method for intelligent identification and early warning of abnormal nuclear power events according to claim 2 is characterized in that: In S102, the pre-processed nuclear power historical operation data is processed and analyzed for big data to extract key features and regular information of nuclear power abnormal events, including the following steps: The pre-processed nuclear power historical operation data is stored in a database; Perform big data processing and analysis on the historical operation data of nuclear power plants stored in the database to extract key features and regular information of abnormal nuclear power events; The big data processing and analysis of the nuclear power historical operation data stored in the database is carried out to extract the key characteristics and regular information of nuclear power abnormal events, including the following steps: In the historical operation data of nuclear power plants, the data reported by users are unstructured data, and time series data are further generated for the data reported by users; the equipment operation data, environmental monitoring data and safety system status data are structured data; Analyze the historical operation data of nuclear power through the time series analysis model, and deeply explore the key regularity information of nuclear power abnormal events in the historical operation data of nuclear power; the key regularity information of nuclear power abnormal events in the historical operation data of nuclear power includes the time change trend and potential periodicity of nuclear power abnormal events in the historical operation data of nuclear power; Extract key characteristic information of nuclear power abnormal events from historical nuclear power operation data.

5. The method for intelligent identification and early warning of abnormal nuclear power events according to claim 2, characterized in that: S103, training a machine learning model using training set data to build a nuclear power abnormal event recognition model, including the following steps: S1031. According to the characteristics and identification requirements of nuclear power abnormal events, select a suitable machine learning model to build a nuclear power abnormal event identification model; S1032, training the selected machine learning model with the training set data, setting the machine learning model parameters during the training process, and optimizing the machine learning model parameter settings through a cross-validation strategy; after the training is completed, a nuclear power abnormal event recognition model is obtained; S1033. Evaluate the nuclear power abnormal event recognition model using the test set data; if the evaluation index does not meet the requirements, optimize the nuclear power abnormal event recognition model until the evaluation index meets the requirements.

6. The method for intelligent identification and early warning of abnormal nuclear power events according to claim 5 is characterized in that: Select one or more of the IForest model, CBLOF model, HBOS model, KNN model, LOF model, OCSVM model and AutoEncoder model to build a nuclear power abnormal event recognition model; the optimization strategy of the nuclear power abnormal event recognition model includes adjusting model parameters, improving feature engineering, introducing new machine learning models or adopting machine learning model fusion.

7. The method for intelligent identification and early warning of abnormal nuclear power events according to claim 1, characterized in that: When the nuclear power abnormal event recognition model identifies a nuclear power abnormal event, the early warning mechanism is triggered and early warning information is sent to relevant personnel through pop-up warnings, sound alarms and / or SMS notifications; according to the type and severity of the nuclear power abnormal event, the corresponding emergency response plan is preset; emergency measures include emergency shutdown, personnel evacuation and pollution control.

8. An intelligent identification and early warning device for abnormal nuclear power events, characterized in that: include: Data acquisition and preprocessing module, used to collect historical nuclear power operation data; Preprocess the collected nuclear power historical operation data; The big data processing and analysis module is used to process and analyze the pre-processed nuclear power historical operation data, extract the key features and regular information for identifying nuclear power abnormal events, and form training set data; The machine learning model building module is used to train the machine learning model through the training set data and build a nuclear power abnormal event recognition model; An abnormal event recognition module is used to monitor nuclear power operation data in real time through a nuclear power abnormal event recognition model and to recognize nuclear power abnormal events in the nuclear power operation data; The early warning and emergency response module is used to send early warning information to relevant personnel when the nuclear power abnormal event recognition model identifies a nuclear power abnormal event; Initiate emergency measures according to the preset emergency response plan.

9. A computer device comprising a memory and a processor, wherein the memory stores computer-readable instructions, characterized in that: When the processor executes the computer-readable instructions, the steps of the method for intelligent identification and early warning of nuclear power abnormal events as described in any one of claims 1-7 are implemented.

10. A computer-readable storage medium having computer-readable instructions stored thereon, characterized in that: When the computer-readable instructions are executed, the steps of the method for intelligent identification and early warning of nuclear power abnormal events as described in any one of claims 1 to 7 are implemented.

Images