Neural network model based on anti-error data fusion

Trend of error of the measured data is found by using wavelet transform, and the reliability of the measured data is tested according to the error trend, and the weights of the measured data are determined. Then anti-error data fusion method is proposed. After the data fusion, a model for three-phas...

Full description

Saved in:
Bibliographic Details
Main Authors: Mei Wang, Yuan-Bin Hou
Format: Conference Proceeding
Language:English
Subjects:
cable fault
Communication cables
Data fusion
Distortion measurement
Electric variables measurement
Error correction
Fault detection
neural network
Neural networks
Power system modeling
prediction
Predictive models
Sensor systems
System testing
wavelet
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Trend of error of the measured data is found by using wavelet transform, and the reliability of the measured data is tested according to the error trend, and the weights of the measured data are determined. Then anti-error data fusion method is proposed. After the data fusion, a model for three-phase cable fault system is constructed by choosing BP neural network with an improved BP algorithm, and the prediction and location of cable fault can be implemented based on neural network model. Simulation shows that the outputs of neural network model are nearly close to the outputs of the practical system, and the mean value of errors of cable fault distance predicted by the neural network model that is constructed by using the anti-error data is quite less than that by using the data before fusion. So the anti-error data fusion method is correct and the NN model of cable fault system is reliable.
ISSN:2160-133X
DOI:10.1109/ICMLC.2005.1527667