CAJ | 학술논문

针对乏燃料剪切机剪切声音信号特征提取的难题，利用小波包分析方法，对不同磨损状况刀具的剪切声音信号进行小波包变换，提取变换信号的各频段归一化能量特征向量，根据声音信号的能量特征向量可辨识不同状况的乏燃料剪切机剪切声音，从而实现乏燃料剪切机故障诊断。实验表明，该特征向量能有效识别刀具的正常磨损、一级磨损、二级磨损三种状况，有效解决了基于隐马尔可夫模型的故障模式识别中特征提取的问题。
침대핍연료전절궤전절성음신호특정제취적난제，이용소파포분석방법，대불동마손상황도구적전절성음신호진행소파포변환，제취변환신호적각빈단귀일화능량특정향량，근거성음신호적능량특정향량가변식불동상황적핍연료전절궤전절성음，종이실현핍연료전절궤고장진단。실험표명，해특정향량능유효식별도구적정상마손、일급마손、이급마손삼충상황，유효해결료기우은마이가부모형적고장모식식별중특정제취적문제。
Aiming at the voice signal feature extraction problem of spent nuclear fuel shears,the wavelet packet analysis method was proposed. The wavelet packet transform was used to the cutting sound signal of the different status tool,extracting the normalized energy eigenvector of each frequency band signal,the energy eigenvector can distinguish different status cutting sound signal,and then diagnose the fault of spent nuclear fuel shears. Experi-mental results show that it can effectively distinguish three conditions of the shears:the lev-el of normal,the first or the second wear and tear,which effectively solved the feature ex-traction problem based on the Hidden Markov Model in shearing machine fault pattern rec-ognition.