Research On High-speed Signal Processing Method Of Optical Current Transformer Based On FPGA

With the development of large digital, intelligent grid, the optical current transformer(OCT) based on the Faraday effect is becoming an ideal device for measuring current for the advantages of anti-electromagnetic interference capability, good dynamic response capability and the digital output. However, the PIN diodes in the optical current transformer will generate strong random noise in the measurement process, which is one of the main factors affecting the OCT measurement accuracy. In addition, as one of the main equipments of intelligent substation, the OCT’s real-time requirement is very high. Therefore, this article focus on how to filter out the noise signal in the optical current transforme with high-speed, to meet it’s requirements of accuracy and real-time.This paper introduces the basic principles of OCT, analyzes the measurement process and structure as well as the noise source and properties of OCT, collected the noise data of PIN diode through experiment, which is the main noise sources of OCT, then analyzes of the signal to noise characteristics of OCT.According to the signal to noise characteristics of OCT, this paper proposes an improved particle filter algorithm to improve the output SNR of OCT. In the verification on the simulation platform of MATLAB, a particle filtering dynamic model is established and a fission bootstrap particle filter (FBPF) algorithm with proper parameters is implemented, for direct current and alternative current respectively. The results indicate that the proposed method effectively improves SNR and measurement accuracy.For FPGA hardware, establishes the model for AC OCT signal processing based on particle filter, on the basis of this model, derives and simplifes the calculation procedure of threshold-based particle filter algorithm. The resampling part, the sampling part and the weight calculation part of the algorithm in FPGA implementation and order of execution are improved, greatly improved the execution speed of the algorithm on FPGA.Uses the designed real-time particle filter to process the signal of low SNR for experiment, and compares the experimental results with the results of the Bessel filter,the results indicate that the particle filter designed in this paper is superior to Bessel filter, and meet the OCT requirements of measurement accuracy and real-time.