CN113962299A - Intelligent operation monitoring and fault diagnosis general model for nuclear power equipment - Google Patents
Intelligent operation monitoring and fault diagnosis general model for nuclear power equipment Download PDFInfo
- Publication number
- CN113962299A CN113962299A CN202111201967.2A CN202111201967A CN113962299A CN 113962299 A CN113962299 A CN 113962299A CN 202111201967 A CN202111201967 A CN 202111201967A CN 113962299 A CN113962299 A CN 113962299A
- Authority
- CN
- China
- Prior art keywords
- data
- model
- sample
- nuclear power
- abnormal
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
- 238000003745 diagnosis Methods 0.000 title claims abstract description 44
- 238000012544 monitoring process Methods 0.000 title claims abstract description 35
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 24
- 238000013473 artificial intelligence Methods 0.000 claims abstract description 5
- 230000002159 abnormal effect Effects 0.000 claims description 46
- 238000012549 training Methods 0.000 claims description 31
- 238000001514 detection method Methods 0.000 claims description 21
- 238000000034 method Methods 0.000 claims description 19
- 238000012360 testing method Methods 0.000 claims description 19
- 238000013528 artificial neural network Methods 0.000 claims description 17
- 230000006870 function Effects 0.000 claims description 15
- 230000005856 abnormality Effects 0.000 claims description 13
- 238000010801 machine learning Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 9
- 238000010219 correlation analysis Methods 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 238000011161 development Methods 0.000 claims description 5
- 238000005516 engineering process Methods 0.000 claims description 5
- 230000002045 lasting effect Effects 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 238000007635 classification algorithm Methods 0.000 claims description 3
- 238000010224 classification analysis Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000000611 regression analysis Methods 0.000 claims description 3
- 238000012552 review Methods 0.000 claims description 3
- 230000026676 system process Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 4
- 238000005728 strengthening Methods 0.000 description 3
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Monitoring And Testing Of Nuclear Reactors (AREA)
Abstract
The invention relates to the technical field of nuclear power equipment monitoring, and particularly discloses a nuclear power equipment intelligent operation monitoring and fault diagnosis universal model, which comprises the following steps: s1, data acquisition: the invention provides a general model for intelligent operation monitoring and fault diagnosis of nuclear power equipment based on Python language and artificial intelligence algorithm, which can adapt to different modeling systems and equipment by flexibly adjusting parameters and configuring, wherein a model developer only needs to provide state data of a monitored object, and performs related parameter configuration to train and generate the model, and the model is finally used for monitoring and fault diagnosis of on-line nuclear power equipment.
Description
Technical Field
The invention relates to the technical field of nuclear power equipment monitoring, in particular to a general model for intelligent operation monitoring and fault diagnosis of nuclear power equipment.
Background
Nuclear power becomes a novel green energy source due to the characteristics of cleanness, high efficiency, less consumption of resources for fuel transportation and storage and the like, four departments such as 2018 development, reform and commission and the like jointly issue guidance opinions about further strengthening nuclear power operation safety management, wherein an important basic requirement is the application of new technologies such as 'insisting on method innovation, promoting informatization, intellectualization, big data and the like in nuclear power operation safety management, strengthening state monitoring on equipment, improving the safety management level', strengthening the operation state monitoring of key equipment, improving the monitoring technical means, improving the comprehensiveness, timeliness and accuracy of the operation state detection of the key equipment, realizing early warning of equipment faults and optimizing the maintenance strategy of the equipment; a key equipment management platform is established, and the intelligent detection, fault diagnosis, health evaluation and service life prediction levels of equipment are improved; the real-time tracking and monitoring of the operation risk of the nuclear power unit are enhanced, and the nuclear power operation safety level is accurately improved.
When a system or equipment of the nuclear power unit is abnormal or fails, state data sensitive to faults slightly or greatly change relative to state data in a normal state, and the state data change values can be used for analyzing and judging whether the abnormality occurs or not and revealing different fault types.
Disclosure of Invention
The invention aims to provide a general model for intelligent operation monitoring and fault diagnosis of nuclear power equipment, so as to solve the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme: a nuclear power equipment intelligent operation monitoring and fault diagnosis general model comprises the following steps:
s1, data acquisition: firstly, acquiring state data of a nuclear power system in the operation process, wherein the state data comprises normal operation condition data under various power levels and modes and abnormal data under possible abnormal events, the data form is mainly streaming data, high-dimensional time sequence data of a plurality of sensors form a training sample and a data matrix, a universal model training algorithm has no limit on the number and the time of the sensors, only the sample form is required to be unified, the acquired sample data reflects the main characteristics of the class of the sample form, the data is required to be classified when being acquired, and sample labels, normal data and abnormal data under different conditions in supervised learning are classified and stored, so that a training and testing data set is provided for a state monitoring model;
s2, preprocessing data: for all sample data, whether formats of the sample data are completely unified or not is mainly checked, whether missing data or abnormal values exist in the sample data or not is determined, whether a compensation mean value is adopted for compensation or not can be determined according to process characteristics for a small amount of missing data, samples with serious missing data can be removed, sample files of different types are stored according to a certain rule, normalization processing is carried out on the data, the data are packaged into a function, a plurality of selectable values are provided as input of an API (application programming interface), a user can complete different data processing by using a small amount of codes by using the API interface and only inputting the sample formats, so that original data are cleaned into pre-training data, all the data are scaled to the same range by adopting a unified algorithm, original information which is expressed by the data is stored, and algorithm identification is facilitated;
s3, developing a data correlation analysis and anomaly detection module: when the monitoring and diagnosing model is actually deployed and operated in a nuclear power system, firstly, abnormality detection is carried out on a monitored object, namely whether the monitored object is in an abnormal state or not is judged, and when the abnormality occurs, a fault mode identification algorithm needs to be called to identify the fault type. A model developer firstly needs to set some model hyper-parameters, reads the data processed in the step S2 by using interfaces of various packages related to Python language and artificial intelligence technology, performs correlation analysis first, automatically screens and selects a sensitive sensor for feature extraction, thereby reducing the dimensionality of the data, inputs the data into an algorithm to establish a model, trains a multi-classification fault diagnosis model according to historical data by using a machine learning algorithm, compares the test precision of various classification algorithms during training, and selects the model with the highest prediction precision to store as an anomaly monitoring module prediction calling model for online prediction;
s4, developing a fault mode diagnosis module: when the abnormality detection module in the step S3 finds that an abnormality occurs, that is, the abnormal data is transmitted to the failure mode diagnosis module to diagnose the failure type, so that all the data processed by the failure mode diagnosis module are abnormal data of different failure types, and because the actual nuclear power system processes a large number of failure types and is a unified interface, a universality effect is achieved, a neural network method is mainly used for model training and development, characteristics can be automatically extracted through a neural network, only data normalization is needed to be performed on the data processed in the step S2, a neural network structure is built by classifying a training set and a test set, then training is performed, a generated model can be tested through the test set, parameters of the network are continuously fed back and adjusted according to a test result, and the model precision is improved;
s5, model testing and application: adopting a stored prediction script in a general algorithm, opening an interface of real-time data of a nuclear power system, reading the latest real-time data according to the sequence of abnormal detection fault mode diagnosis, calling an abnormal detection module which has finished training for diagnosis, releasing the data if no abnormal condition exists, reading the latest operation data again for abnormal detection, transmitting the data to a fault mode identification module if abnormal condition occurs, diagnosing abnormal types through a trained fault mode diagnosis model, triggering warning, sending the fault type diagnosis result to a display front end in time, prompting an operator to remove the fault in time, and ensuring the reliable operation of nuclear power;
the essence of the machine learning model is a function, which is to implement mapping from one sample X to a labeled value Y of the sample, i.e., f (X) → Y, by assuming that a known function form f' (X) ═ Y is fitted as closely as possible to an objectively existing mapping function given a sufficient sample set { X |1, 2, 3 … … } and corresponding labels { Y1, Y2, Y3 … … }, with respect to the method of model construction, machine learning requires sample labels in the training process through supervised learning and then is used to train the model, and the state monitoring model and the fault diagnosis model are mainly constructed based on supervised learning techniques;
deep learning refers to a neural network with a deep network structure, a complex optimization model is constructed through deep neurons, weights and an activation function, and classification and regression analysis of complex tasks are completed by gradually converting initial abstract low-level features into high-level features through multi-layer neural network processing.
Preferably, the status data in step S1 is usually multi-sensor real-time data with time sequence characteristics, or the sample can be processed into high-dimensional time sequence data lasting for a short period of time, and the judgment criteria of abnormal data is mainly manual judgment, mainly from the expert experience and automatic marking provided during the operation of the system, and is subject to manual review.
Preferably, the sample data collected in step S1 is sample data collected in the same environment as the nuclear power system in actual operation as far as possible, and if the data contains noise, the noise level should be the same as or similar to the actual nuclear power system as far as possible.
Preferably, the sample data collected in step S1 has the main features: normal operating mode data are mild, and continuous value in case the breach appears, then can appear great pressure value fluctuation, along with flow and temperature sudden change.
Preferably, the model hyper-parameters in step S3 include dimension information of each sample, sample length, number of samples, number of features, and selection of question categories.
Compared with the prior art, the invention has the beneficial effects that: the invention provides a general model for intelligent operation monitoring and fault diagnosis of nuclear power equipment based on Python language and artificial intelligence algorithm, which can adapt to different systems and equipment for modeling by flexibly adjusting parameters and configuration, a model developer only needs to provide state data of a monitored object and perform related parameter configuration to train and generate the model, and finally, the model is used for monitoring and fault diagnosis of the on-line nuclear power equipment.
Drawings
FIG. 1 is a general technical route scheme of the monitoring and diagnostic model algorithm of the present invention;
FIG. 2 is a schematic diagram of a technical route of an anomaly detection model according to the present invention;
FIG. 3 is a schematic diagram of a fault pattern recognition model according to the present invention;
FIG. 4 is a schematic diagram of different break locations of the main circuit according to an embodiment of the present invention;
FIG. 5 is a graph comparing normal data and abnormal data according to an embodiment of the present invention;
FIG. 6 is a graph illustrating a feature correlation analysis in data preprocessing according to an embodiment of the present invention;
fig. 7 is a diagram of a five-layer hidden layer neural network structure according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-3, the present invention provides a technical solution: a nuclear power equipment intelligent operation monitoring and fault diagnosis general model comprises the following steps:
s1, data acquisition: firstly, acquiring state data of a nuclear power system in the operation process, wherein the state data comprises normal operation condition data under various power levels and modes and abnormal data under possible abnormal events, the data form is mainly streaming data, high-dimensional time sequence data of a plurality of sensors form a training sample and a data matrix, a universal model training algorithm has no limit on the number and the time of the sensors, only the sample form is required to be unified, the acquired sample data reflects the main characteristics of the class of the sample, the data is required to be classified when the data is acquired, and the normal data and the abnormal data under different conditions are classified and stored, so that a training and testing data set is provided for a state monitoring model;
s2, preprocessing data: for all sample data, whether formats of the sample data are completely unified or not is mainly checked, whether missing data or abnormal values exist in the sample data or not is determined, whether a compensation mean value is adopted for compensation or not can be determined according to process characteristics for a small amount of missing data, samples with serious missing data can be removed, sample files of different types are stored according to a certain rule, normalization processing is carried out on the data, the data are packaged into a function, a plurality of selectable values are provided as input of an API (application programming interface), a user can complete different data processing by using a small amount of codes by using the API interface and only inputting the sample formats, so that original data are cleaned into pre-training data, all the data are scaled to the same range by adopting a unified algorithm, original information which is expressed by the data is stored, and algorithm identification is facilitated;
s3, developing a data correlation analysis and anomaly detection module: when the monitoring and diagnosing model is actually deployed and operated in a nuclear power system, firstly, abnormality detection is carried out on a monitored object, namely whether the monitored object is in an abnormal state or not is judged, and when the abnormality occurs, a fault mode identification algorithm needs to be called to identify the fault type. A model developer firstly needs to set some model hyper-parameters, reads the data processed in the step S2 by using interfaces of various packages related to Python language and artificial intelligence technology, performs correlation analysis first, automatically screens and selects a sensitive sensor for feature extraction, thereby reducing the dimensionality of the data, inputs the data into an algorithm to establish a model, trains a multi-classification fault diagnosis model according to historical data by using a machine learning algorithm, compares the test precision of various classification algorithms during training, and selects the model with the highest prediction precision to store as an anomaly monitoring module prediction calling model for online prediction;
s4, developing a fault mode diagnosis module: when the abnormality detection module in the step S3 finds that an abnormality occurs, that is, the abnormal data is transmitted to the failure mode diagnosis module to diagnose the failure type, so that all the data processed by the failure mode diagnosis module are abnormal data of different failure types, and because the actual nuclear power system processes a large number of failure types and is a unified interface, a universality effect is achieved, a neural network method is mainly used for model training and development, characteristics can be automatically extracted through a neural network, only data normalization is needed to be performed on the data processed in the step S2, a neural network structure is built by classifying a training set and a test set, then training is performed, a generated model can be tested through the test set, parameters of the network are continuously fed back and adjusted according to a test result, and the model precision is improved;
s5, model testing and application: adopting a stored prediction script in a general algorithm, opening an interface of real-time data of a nuclear power system, reading the latest real-time data according to the sequence of abnormal detection fault mode diagnosis, calling an abnormal detection module which has finished training for diagnosis, releasing the data if no abnormal condition exists, reading the latest operation data again for abnormal detection, transmitting the data to a fault mode identification module if abnormal condition occurs, diagnosing abnormal types through a trained fault mode diagnosis model, triggering warning, sending the fault type diagnosis result to a display front end in time, prompting an operator to remove the fault in time, and ensuring the reliable operation of nuclear power;
the essence of the machine learning model is a function, which is to implement mapping from one sample X to a labeled value Y of the sample, i.e., f (X) → Y, by assuming that a known function form f' (X) ═ Y is fitted as closely as possible to an objectively existing mapping function given a sufficient sample set { X |1, 2, 3 … … } and corresponding labels { Y1, Y2, Y3 … … }, with respect to the method of model construction, machine learning requires sample labels in the training process through supervised learning and then is used to train the model, and the state monitoring model and the fault diagnosis model are mainly constructed based on supervised learning techniques;
deep learning refers to a neural network with a deep network structure, a complex optimization model is constructed through deep neurons, weights and an activation function, and classification and regression analysis of complex tasks are completed by gradually converting initial abstract low-level features into high-level features through multi-layer neural network processing.
Further, the status data in step S1 is usually multi-sensor real-time data with time sequence characteristics, or the sample may be processed into high-dimensional time sequence data lasting for a short period of time, and the judgment criteria of the abnormal data is mainly manual judgment, mainly from the expert experience and automatic flag provided during the operation of the system, and is subject to manual review.
Further, the sample data collected in step S1 is sample data collected in the same environment as the nuclear power system in actual operation as far as possible, and if the data contains noise, the noise level should be the same as or similar to that of the actual nuclear power system as far as possible.
Further, the sample data collected in step S1 is mainly characterized by: normal operating mode data are mild, and continuous value in case the breach appears, then can appear great pressure value fluctuation, along with flow and temperature sudden change.
Further, the model hyper-parameters in step S3 include dimension information of each sample, sample length, number of samples, number of features, and selection of question categories.
In fig. 4 are shown: in the development stage, whether the main loop is possible to break is researched, and a large amount of data of normal operation of the main loop under different powers and abnormal power data when breaks occur in 10 different positions are collected on a simulator capable of reflecting real power plant indexes.
In fig. 5 are shown: each sample is high-dimensional time sequence data of 63 sensors lasting 150 seconds, and the data is mainly characterized in that: normal operating mode data are mild, and continuous value, concrete expression is the straight line form, in case appear the breach, then can appear great pressure value fluctuation, concrete expression is the curve form, accompanies flow and temperature sudden change.
In fig. 7 are shown: in the embodiment of the main loop breach problem, a 0.8: the proportion of the training set and the test set is 0.2, a five-layer hidden layer neural network structure is adopted, and a model capable of identifying 10 fault types, namely a classification model for judging 10 abnormal types, is trained.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (5)
1. A nuclear power equipment intelligent operation monitoring and fault diagnosis general model is characterized by comprising the following steps:
s1, data acquisition: firstly, acquiring state data of a nuclear power system in the operation process, wherein the state data comprises normal operation condition data under various power levels and modes and abnormal data under possible abnormal events, the data form is mainly streaming data, high-dimensional time sequence data of a plurality of sensors form a training sample and a data matrix, a universal model training algorithm has no limit on the number and the time of the sensors, only the sample form is required to be unified, the acquired sample data reflects the main characteristics of the class of the sample, the data is required to be classified when the data is acquired, and the normal data and the abnormal data under different conditions are classified and stored, so that a training and testing data set is provided for a state monitoring model;
s2, preprocessing data: for all sample data, whether formats of the sample data are completely unified or not is mainly checked, whether missing data or abnormal values exist in the sample data or not is determined, whether a compensation mean value is adopted for compensation or not can be determined according to process characteristics for a small amount of missing data, samples with serious missing data can be removed, sample files of different types are stored according to a certain rule, normalization processing is carried out on the data, the data are packaged into a function, a plurality of selectable values are provided as input of an API (application programming interface), a user can complete different data processing by using a small amount of codes by using the API interface and only inputting the sample formats, so that original data are cleaned into pre-training data, all the data are scaled to the same range by adopting a unified algorithm, original information which is expressed by the data is stored, and algorithm identification is facilitated;
s3, developing a data correlation analysis and anomaly detection module: when the monitoring and diagnosing model is actually deployed and operated in a nuclear power system, firstly, abnormality detection is carried out on a monitored object, namely whether the monitored object is in an abnormal state or not is judged, and when the abnormality occurs, a fault mode identification algorithm needs to be called to identify the fault type. A model developer firstly needs to set some model hyper-parameters, reads the data processed in the step S2 by using interfaces of various packages related to Python language and artificial intelligence technology, performs correlation analysis first, automatically screens and selects a sensitive sensor for feature extraction, thereby reducing the dimensionality of the data, inputs the data into an algorithm to establish a model, trains a multi-classification fault diagnosis model according to historical data by using a machine learning algorithm, compares the test precision of various classification algorithms during training, and selects the model with the highest prediction precision to store as an anomaly monitoring module prediction calling model for online prediction;
s4, developing a fault mode diagnosis module: when the abnormality detection module in the step S3 finds that an abnormality occurs, that is, the abnormal data is transmitted to the failure mode diagnosis module to diagnose the failure type, so that all the data processed by the failure mode diagnosis module are abnormal data of different failure types, and because the actual nuclear power system processes a large number of failure types and is a unified interface, a universality effect is achieved, a neural network method is mainly used for model training and development, characteristics can be automatically extracted through a neural network, only data normalization is needed to be performed on the data processed in the step S2, a neural network structure is built by classifying a training set and a test set, then training is performed, a generated model can be tested through the test set, parameters of the network are continuously fed back and adjusted according to a test result, and the model precision is improved;
s5, model testing and application: adopting a stored prediction script in a general algorithm, opening an interface of real-time data of a nuclear power system, reading the latest real-time data according to the sequence of abnormal detection fault mode diagnosis, calling an abnormal detection module which has finished training for diagnosis, releasing the data if no abnormal condition exists, reading the latest operation data again for abnormal detection, transmitting the data to a fault mode identification module if abnormal condition occurs, diagnosing abnormal types through a trained fault mode diagnosis model, triggering warning, sending the fault type diagnosis result to a display front end in time, prompting an operator to remove the fault in time, and ensuring the reliable operation of nuclear power;
the essence of the machine learning model is a function, which is to implement mapping from one sample X to a labeled value Y of the sample, i.e., f (X) → Y, by assuming that a known function form f' (X) ═ Y is fitted as closely as possible to an objectively existing mapping function given a sufficient sample set { X |1, 2, 3 … … } and corresponding labels { Y1, Y2, Y3 … … }, with respect to the method of model construction, machine learning requires sample labels in the training process through supervised learning and then is used to train the model, and the state monitoring model and the fault diagnosis model are mainly constructed based on supervised learning techniques;
deep learning refers to a neural network with a deep network structure, a complex optimization model is constructed through deep neurons, weights and an activation function, and classification and regression analysis of complex tasks are completed by gradually converting initial abstract low-level features into high-level features through multi-layer neural network processing.
2. The nuclear power equipment intelligent operation monitoring and fault diagnosis general model as claimed in claim 1, wherein: the status data in step S1 is usually multi-sensor real-time data with time sequence characteristics, or the sample may be processed into high-dimensional time sequence data lasting for a short period of time, and the judgment criteria of the abnormal data is mainly manual judgment, mainly from the expert experience and automatic flags provided during system operation, and is subjected to manual review.
3. The nuclear power equipment intelligent operation monitoring and fault diagnosis general model as claimed in claim 1, wherein: the sample data acquired in step S1 is sample data acquired as far as possible in the same environment as the actual operation of the nuclear power system, and if the data contains noise, the noise level should be the same as or similar to that of the actual nuclear power system as far as possible.
4. The nuclear power equipment intelligent operation monitoring and fault diagnosis general model as claimed in claim 1, wherein: the sample data acquired in step S1 is mainly characterized in that: normal operating mode data are mild, and continuous value in case the breach appears, then can appear great pressure value fluctuation, along with flow and temperature sudden change.
5. The nuclear power equipment intelligent operation monitoring and fault diagnosis general model as claimed in claim 1, wherein: the model hyper-parameters in step S3 include the dimension information of each sample, the sample length, the number of samples, the number of features, and the selection of the problem category.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111201967.2A CN113962299A (en) | 2021-10-15 | 2021-10-15 | Intelligent operation monitoring and fault diagnosis general model for nuclear power equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111201967.2A CN113962299A (en) | 2021-10-15 | 2021-10-15 | Intelligent operation monitoring and fault diagnosis general model for nuclear power equipment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113962299A true CN113962299A (en) | 2022-01-21 |
Family
ID=79464666
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111201967.2A Withdrawn CN113962299A (en) | 2021-10-15 | 2021-10-15 | Intelligent operation monitoring and fault diagnosis general model for nuclear power equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113962299A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114547970A (en) * | 2022-01-25 | 2022-05-27 | 中国长江三峡集团有限公司 | Intelligent diagnosis method for abnormity of top cover drainage system of hydraulic power plant |
CN114660399A (en) * | 2022-02-28 | 2022-06-24 | 湖北拓普电力有限公司 | Ring main unit equipment state monitoring and fault diagnosis method |
CN116647586A (en) * | 2023-07-24 | 2023-08-25 | 山东工泵电机有限公司 | Method for realizing remote control of water and fertilizer integrated intelligent pump house by cloud computing |
CN116702030A (en) * | 2023-05-31 | 2023-09-05 | 浙江大学 | Blast furnace state monitoring method and device based on sensor reliability analysis |
CN116723083A (en) * | 2023-08-10 | 2023-09-08 | 山东智拓大数据有限公司 | Cloud server online fault diagnosis method and device |
CN117171576A (en) * | 2023-11-01 | 2023-12-05 | 博纯材料股份有限公司 | Abnormality monitoring method and system applied to material purification system |
-
2021
- 2021-10-15 CN CN202111201967.2A patent/CN113962299A/en not_active Withdrawn
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114547970A (en) * | 2022-01-25 | 2022-05-27 | 中国长江三峡集团有限公司 | Intelligent diagnosis method for abnormity of top cover drainage system of hydraulic power plant |
CN114547970B (en) * | 2022-01-25 | 2024-02-20 | 中国长江三峡集团有限公司 | Intelligent diagnosis method for abnormality of top cover drainage system of hydropower plant |
CN114660399A (en) * | 2022-02-28 | 2022-06-24 | 湖北拓普电力有限公司 | Ring main unit equipment state monitoring and fault diagnosis method |
CN116702030A (en) * | 2023-05-31 | 2023-09-05 | 浙江大学 | Blast furnace state monitoring method and device based on sensor reliability analysis |
CN116702030B (en) * | 2023-05-31 | 2024-01-23 | 浙江大学 | Blast furnace state monitoring method and device based on sensor reliability analysis |
CN116647586A (en) * | 2023-07-24 | 2023-08-25 | 山东工泵电机有限公司 | Method for realizing remote control of water and fertilizer integrated intelligent pump house by cloud computing |
CN116723083A (en) * | 2023-08-10 | 2023-09-08 | 山东智拓大数据有限公司 | Cloud server online fault diagnosis method and device |
CN116723083B (en) * | 2023-08-10 | 2024-01-26 | 山东智拓大数据有限公司 | Cloud server online fault diagnosis method and device |
CN117171576A (en) * | 2023-11-01 | 2023-12-05 | 博纯材料股份有限公司 | Abnormality monitoring method and system applied to material purification system |
CN117171576B (en) * | 2023-11-01 | 2024-02-27 | 博纯材料股份有限公司 | Abnormality monitoring method and system applied to material purification system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113962299A (en) | Intelligent operation monitoring and fault diagnosis general model for nuclear power equipment | |
EP1393141B1 (en) | Provision of process related information | |
CN109800127A (en) | A kind of system fault diagnosis intelligence O&M method and system based on machine learning | |
CN108027611B (en) | Decision assistance system and method for machine maintenance using expert opinion supervised decision mode learning | |
US20050015217A1 (en) | Analyzing events | |
US20050043922A1 (en) | Analysing events | |
Tiddens et al. | The adoption of prognostic technologies in maintenance decision making: a multiple case study | |
JP2004523843A (en) | System, apparatus and method for diagnosing a flow system | |
Kontogiannis | A framework for the analysis of cognitive reliability in complex systems: a recovery centred approach | |
CN115358155A (en) | Power big data abnormity early warning method, device, equipment and readable storage medium | |
CN116228186A (en) | Ship cabin intelligent operation and maintenance system based on human engineering | |
CN116629627A (en) | Intelligent detection system of power transmission on-line monitoring device | |
KR20230021272A (en) | Bridge health monitoring system using artificial intelligence | |
CN102929241B (en) | Safe operation guide system of purified terephthalic acid device and application of safe operation guide system | |
KR102618023B1 (en) | Failure prediction diagnosis system and method through pattern analysis according to failure type | |
AU2020202134A1 (en) | Data driven machine learning for modeling aircraft sensors | |
Merkt | Predictive models for maintenance optimization: an analytical literature survey of industrial maintenance strategies | |
Knight et al. | Intelligent management of helicopter health and usage management systems data | |
Song et al. | A Shape-constrained Transfer Temporal Transformer Network for remaining useful life prediction of rotating machines | |
CN116223031A (en) | Health diagnosis method and system based on vibration signals of radar driving mechanism | |
CN118096131A (en) | Operation and maintenance inspection method based on electric power scene model | |
Wen et al. | An evidential similarity-based regression method for the prediction of equipment remaining useful life under complex conditions | |
Frank | AI-based fault diagnosis and troubleshooting in automotive systems | |
Frank | The role of AI in predicting automotive failures | |
CN117891684A (en) | Hardware fault prediction method and system for embedded computer |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: No. 29 Hong Cao Road, Xuhui District, Shanghai Applicant after: Shanghai Nuclear Engineering Research and Design Institute Co.,Ltd. Address before: No. 29 Hong Cao Road, Xuhui District, Shanghai Applicant before: SHANGHAI NUCLEAR ENGINEERING RESEARCH & DESIGN INSTITUTE Co.,Ltd. |
|
WW01 | Invention patent application withdrawn after publication |
Application publication date: 20220121 |