ZHOU Wei, PAN Haipeng, WU Ping, et al. Tiny fault detection method based on DPCA and KL divergence[J]. 2020, (3): 135-138. DOI: 10.13873/J.1000-9787(2020)03-0135-04.
Aiming at the characteristics of small faults and inconspicuous features in industrial production process
a tiny fault detection method based on dynamic principal component analysis( DPCA) and KullbackLeibler( KL) divergence is proposed. The method first uses DPCA to construct the PCA model for dynamic process
and then uses KL divergence to quantify the similarity between the probability distributions of the score vectors
and establish the corresponding statistical limits according to the 3σ rule. In order to verify effectiveness of the proposed method
the method is applied to the Tennessee Eastman process( TEP). The results show that the proposed method can effectively detect tiny faults.