Wireless sensor node fault diagnosis method and device based on deep learning

A wireless sensor and deep learning technology, applied in the field of fault diagnosis of wireless sensor nodes based on deep learning, can solve problems such as fusion error faults and stuck faults, achieve enhanced robustness and applicability, and facilitate continuous iterative improvement and update , the effect of strong model generalization ability

Pending Publication Date: 2022-02-15
NUCLEAR POWER INSTITUTE OF CHINA
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] However, due to the complexity of the environment in which wireless sensors are located in nuclear power plants, their own structures (including radio frequency chips, battery modules, etc.) and the particularity of signal transmission methods, compared with conventional sensors, wireless sensor nodes are more vulnerable to various Various types of faults (offset faults, gain faults, stuck faults, spike faults, data loss faults, fusion error faults, battery faults, hardware faults, etc.) are affected, and it is necessary to ensure that wireless sensors can complete monitoring work with high quality in nuclear power plants , it is necessary to predict and detect the possible faults of wireless sensors as early as possible

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Wireless sensor node fault diagnosis method and device based on deep learning
  • Wireless sensor node fault diagnosis method and device based on deep learning
  • Wireless sensor node fault diagnosis method and device based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] This embodiment proposes a wireless sensor node fault diagnosis method based on deep learning for the complex environment of the nuclear power plant, which improves the reliability of the wireless sensor network nodes and the entire network and the safety of each operating equipment of the nuclear power plant.

[0056] In this embodiment, through in-depth research and analysis of the fault model of the wireless sensor node mechanism, combined with the experimental fault data, the source domain training set data is obtained. The present invention makes full use of the deep essential feature extraction ability of SSDAE and the robustness and adaptability of the model, and obtains a deep fault feature extractor through training on the source domain data set; a fully connected adaptive layer is added after the SSDAE deep feature extractor to achieve A small amount of actual fault sample data in the target domain is used to train and correct the migration model, which can mak...

Embodiment 2

[0134] This embodiment proposes a wireless sensor node fault diagnosis device based on deep learning, such as Figure 7 As shown, the device includes a data preprocessing module, a training module, a migration module, an optimization module, and a detection module.

[0135] Wherein, the data preprocessing module is used to obtain SSDAE model source domain training set data according to wireless sensor node mechanism failure model data and experimental data.

[0136] The training module obtains the SSDAE model by training on the source domain training set data.

[0137] The migration module migrates the trained SSDAE model to obtain a migration model, trains and corrects the migration model with fault sample data in the target domain, and obtains the final deep fault diagnosis model.

[0138] The optimization module identifies the fault categories that cannot be judged by the deep fault diagnosis model by adding the "Don't know" fault response label to the classification layer...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a wireless sensor node fault diagnosis method and device based on deep learning. The method comprises the steps of obtaining SSDAE model source domain training set data; training by adopting source domain training set data to obtain an SSDAE model; migrating the trained SSDAE model to obtain a migration model, and training and correcting the migration model by target domain fault sample data to obtain a final deep fault diagnosis model; and detecting the fault of the wireless sensor in the nuclear power device by adopting the deep fault diagnosis model. The feature extraction capability of the stacked sparse automatic encoder is fully utilized, the deep essential feature extractor is obtained through training on the training set, the internal features of the fault data are deeply extracted, and accurate fault detection and diagnosis are facilitated.

Description

technical field [0001] The invention belongs to the technical field of wireless sensor networks (WSN), and in particular relates to a deep learning-based wireless sensor node fault diagnosis method and device. Background technique [0002] The wireless sensor in the nuclear power plant monitors the process parameters (temperature, pressure, etc.) of the equipment operation, and sends the parameter information to the upper-layer equipment by wireless transmission, which is used for the relevant control, display, alarm and equipment operation status judgment of the nuclear power plant Function, function is very important. [0003] However, due to the complexity of the environment in which wireless sensors are located in nuclear power plants, their own structures (including radio frequency chips, battery modules, etc.) and the particularity of signal transmission methods, compared with conventional sensors, wireless sensor nodes are more vulnerable to various Various types of ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06N3/04G06N3/08
CPCG06N3/04G06N3/084G06F18/2414G06F18/214
Inventor 邓志光徐思捷吕鑫吴茜朱毖微青先国杨洪润何正熙王雪梅赵阳卢川朱加良何鹏徐涛陈静李小芬李红霞叶宇衡
Owner NUCLEAR POWER INSTITUTE OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products