Real-time modeling and online filtering of the stochastic error in a fiber optic current transducer

The stochastic error characteristics of a fiber optic current transducer (FOCT) influence the relay protection, electric-energy metering, and other devices in the spacer layer. Real-time modeling and online filtering of the FOCT’s stochastic error tends to be an effective method for improving the measurement accuracy of the FOCT. This paper first pretreats and inspects the FOCT data, statistically. Then, the model order is set by the AIC principle to establish an ARMA (2,1) model and model’s applicability is tested. Finally, a Kalman filter is adopted to reduce the noise in the FOCT data. The results of the experiment and the simulation demonstrate that there is a notable decrease in the stochastic error after time series modeling and Kalman filtering. Besides, the mean-variance is decreased by two orders. All the stochastic error coefficients are decreased by the total variance method; the BI is decreased by 41.4%, the RRW is decreased by 67.5%, and the RR is decreased by 53.4%. Consequently, the method can reduce the stochastic error and improve the measurement accuracy of the FOCT, effectively.