Flaw Identification Based on Layered Multi-subnet Neural Networks
Authors:
- Sun XiaoyunSXSun Xiaoyun
- Li AihuaLALi Aihua
- 1 November 2009
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 05,p. 118-121
Abstract
Pointed to the disadvantages such as low recognizing precision, long training time and limited recognizing range of single neural network in eddy current testing, layered multi-subnet neural network is presented. It is composed by a sumnet and several layered subnets, and can divide a complex task into a series of subtasks, so it could quickly identify whether the defect is existed, and also the defect location and dimension. Because of Fisher ratio method used to select the RBF centers, the network structure is simplified much. The result shows that layered multi-subnet neural network is suitable to online eddy current testing.Keywords
This publication has 3 references indexed in Scilit:
- Eddy-current testing by flexible microloop magnetic sensor arrayIEEE Transactions on Magnetics, 1998
- NDT identification of a crack using ANNs with stochastic gradient descentIEEE Transactions on Magnetics, 1995
- Neural network approach to holographic nondestructive testingApplied Optics, 1995