摘要
提出将联想记忆神经网络应用于电力变流系统的故障诊断。介绍了基于联想记忆模型的诊断算法 ,总结出交 直 交变频器的故障模式及故障分析表 ,然后将该样本向量进行HADAMARD预处理、存储记忆后 ,可据此模型进行样本和非样本的并行联想回忆 ,实现诊断功能。最后进行了数字仿真研究 。
An associative memory model of artificial neural network (ANN) is introduced for the fault diagnosis of inverter.Firstly,the diagnosis algorithm is derived and the fault table is concluded.Secondly,the fault sample vectors are preprocessed by HADAMARD transform,and learned by the model.Finally,the samples and non\|samples are associatively recalled in parallel based on this network.Furthermore,the digital simulation results demonstrate that the novel strategy is reliable.Additionally,a hardware implementation idea is presented,which uses large-scaled field programmable gate array (FPGA) circuits.
出处
《电工电能新技术》
CSCD
2004年第2期17-21,共5页
Advanced Technology of Electrical Engineering and Energy
基金
江苏省高校自然科学研究基金资助项目 (0 3KJD5 1 0 1 0 3)
