Identification and Filtering of Random Error in Fiber Optic Current Transducer

Abstract

With the advantages of excellent insulation performance, wide dynamic measurement range, good frequency response characteristics and high sensitivity, fiber optic current transducer (FOCT) has become the mainstream of current measurement technology. The structure of fiber-optic current transformer is complex, and random error becomes one of the major problems that affect the application of fiber-optic current transformers in power engineering. It is urgent to study the random error characteristics of fiber current transformers and propose effective measures to suppress random errors. The characteristics of random error in FOCT are analyzed, and time series model is established by preprocessing and statistical inspection of FOCT output data. Based on the random error characteristics of the FOCT output signal, it makes prerequisite preparation for the signal filtering optimization modelling and the construction of the state function equation. With the derived random error state function of FOCT, the Wavelet filtering algorithm combining with Kalman filtering algorithm is proposed, then, the time domain and frequency domain local area transforms are used to detect the random error characteristics and identify various types of random error information accurately. Real-time multi-scale decomposition and optimal estimation of random error in FOCT are realized. Experimental results demonstrate that the time series modelling and Wavelet filtering algorithm can accurately identify the random error characteristics of the FOCT and automatically match the parameter filtering to improve current measurement accuracy.