CAJ | 학술논문

针对机械设备关键基础部件早期故障信号提取困难这一问题，提出了一种基于独立分量分析(ICA)的盲源分离去噪方法。采用 Fixed-point ICA算法和基于负熵的判据，对不同信噪比下金属裂纹信号进行提取。研究结果表明，此方法受噪声强度及信号频段的影响比较小，可有效提取出所需信号；且获得的信号波形失真很小，是一种较好的微弱信号提取方法。
침대궤계설비관건기출부건조기고장신호제취곤난저일문제，제출료일충기우독립분량분석(ICA)적맹원분리거조방법。채용 Fixed-point ICA산법화기우부적적판거，대불동신조비하금속렬문신호진행제취。연구결과표명，차방법수조성강도급신호빈단적영향비교소，가유효제취출소수신호；차획득적신호파형실진흔소，시일충교호적미약신호제취방법。
Aimed at the problem of hardly extraction early crack in critical infrastructure components of major equipments,a blind source separation for weak signals based on inde-pendent component analysis(ICA)is proposed.The Fixed-point ICA algorithm and the neg-ative entropy criterion are used.The metal crack signal with different SNRs is extracted and the results show that this method can extract the weak signal.The method is hardly affected by the SNR and frequency band of signals.And the waveform distortion of the obtained sig-nals is negligible,indicating the method is very suitable for weak signal extraction.