摘要
总结了铝合金焊接气孔缺陷及原因,应用小波神经网络技术对铝合金焊接气孔缺陷进行智能诊断。描述了小波神经网络的建模过程,编制了智能诊断的实现软件。具体实例的诊断结果表明,小波神经网络缺陷诊断专家系统稳定性好,诊断速度快,能以较高的置信度诊断出多故障,提高了诊断精度及置信度,这对提高小波神经网络缺陷诊断专家系统的实际应用效果具有现实意义。
The defect symptoms-reason of aluminum alloy welding air hole is concluded.By using wavelet neural network,the defect of aluminum alloy welding air hole is diagnosed.A model of wavelet neural network is constructed,the software of intelligence diagnosis is compiled.With the aid of the expert system of wavelet neural network,multi-faults are diagnosed with a good stability,a high speed and a high reliability.The results show the method is workable.
出处
《振动．测试与诊断》
EI
CSCD
2005年第3期219-221,共3页
Journal of Vibration,Measurement & Diagnosis
关键词
小波神经网络
缺陷分析
诊断误差
置信度
wavelet neural network defect analysis diagnosis error reliability
