The Measurement Algorithms Research On Electrical Parameters For Auto Qusi-Synchronizing Devices

Synchronizing device is simply the device which ties the two AC power together. It must be done when the voltage difference, frequency difference and phase difference of the both sides all meet certain conditions. Exiting voltage difference can cause reactive power impact. Exiting frequency difference can cause active power impact. Exiting great phase difference can cause great damage to the generator rotor shaft, even trigger sub-synchronous resonance.This paper briefly introduced synchronizing conditions, and summed up commonly electrical parameters measurement methods for qusi-synchronizing devices. In numerous synchronizing algorithms, this paper analyzed the measurement algorithms which based on the discrete Fourier transform theory. The discrete Fourier transform (DFT) is widely used in the field of electrical parameters detection. The basic principle of the algorithm is that using the discrete Fourier transform extraction base wave component to calculate electric parameters.This paper analyzes the spectrum leakage and picket fence effect phenomenon and the basic principle of windowing is talked. Based on all above, the traditional discrete Fourier transform algorithm was improved: Method one, Hanning window added to reduce spectrum and interpolation method to reduce the effect of picket fence effect; Method two, bartlett window added to reduce spectrum and recalibrates the result by the phase difference. The simulation showed that both than traditional algorithm have been improved greatly in reliability and accuracy.The discrete Fourier transform algorithm is based on the ideal synchronous sampling, but strict synchronous sampling is difficult to realize. In order to further improve the measurement precision, reduce the influence of synchronization error, the article analyzed the principle of synchronous sampling. The reasons of synchronization errors have been found. The improvement methods to quantizing error caused by processor count cycle and the error caused by periodic measurement hysteresis and dispersed interrupt response time have been found. These improvement methods are simple. Application scope is unrestricted and the amount of calculation added is not large. Also the synchronous precision can satisfy high precision requirements of measurement.