基于独立成分分析与FFT的火焰检测算法

Research on Fire Detection Algorithm Based on Independent Component Analysis and FFT

针对锅炉炉膛火焰检测产品普遍存在的偷看现象,采用独立成分分析(ICA)和FFT相结合的算法,对燃烧火焰进行盲源分离和功率谱分析。结果表明:FastICA+FFT算法能够正确分离出本体火焰信号、旁路火焰信号和炉膛背景火焰信号,判断出锅炉燃烧状态,提高了系统检火的可靠性,降低了偷看漏看概率。

Aiming at the commonly-existed misjudging in boiler furnace flame detectors,the algorithm of combining independent component analysis(ICA)and FFT was employed to analyze both blind source separation and power spectrum analysis of combustion flame.The results show that,FastICA+FFT algorithm can correctly separate the body flame signal,bypass flame signal and furnace background flame signal,and can effectively judge the boiler combustion state to improve the reliability of the system fire detection,and reduce the probability of misjudging.