Nonlinear error recognition of FOCT based on CEEMDAN-ZCR

Abstract

Focusing on the problem that it is difficult to distinguish the nonlinear errors such as drift error and ratio error of fiber optic current transformer (FOCT), an error recognition method based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN)and zero-crossing rate (ZCR) is proposed. Firstly, the output signal of the FOCT is decomposed using CEEMDAN algorithm to obtain the intrinsic mode functions (IMF) containing nonlinear error characteristics, which constitutes the original error vector data set. Then, the number of components under different errors is compared. The ZCR index of each IMF component under different errors is calculated by ZCR algorithm. The results are used to classify the IMF. Finally, the IMF component signals are divided into three categories according to the ZCR index. IMF components are superimposed and reorganized into three components, and IMF component signals with stable number of decomposition results are constructed. Error identification is realized according to the expression forms of different components. Experiment results show that the CEEMDAN-ZCR based error recognition method can effectively identify the two kinds of errors. The drift error characteristics are mainly concentrated in the third layer of IMF (C3), and the variation ratio error is mainly concentrated in the second layer of IMF (C2).

Keywords

• fiber optic current transformer (foct)
• drift error
• ratio error
• complete ensemble empirical mode decomposition with adaptive noise (ceemdan)
• zero-crossing rate (zcr)
• error recognition