论文部分内容阅读
本文首先用BP网络完成了涡流无损检测中缺陷的快速识别。针对BP算法存在着收敛速度慢、易陷入局部极小的缺点 ,作者将RBF(RadialBasisFunction)网络应用于涡流无损检测。结果表明RBF算法的收敛速度较BP算法快得多。RBF网络适于在线检测。
In this paper, BP network is used to complete the rapid identification of defects in eddy current non-destructive testing. For BP algorithm, it has the disadvantage of slow convergence rate and easy falling into local minimum. The author applies RBF (Radial Basis Function) network to eddy current non-destructive testing. The results show that RBF algorithm converges much faster than BP algorithm. RBF network suitable for online testing.