Statistical Analyses of Power Quality Mainly on Voltage Distortion

Abstract

It is important to grasp the actual condition of power quality before attempting to maintain appropriate levels. Some data on power quality, such as on the harmonic voltage, are available but their actual conditions have not been clarified as a whole. This report analyzes power quality data in further detail and clarifies the following points:
The level of power quality varies due to two reasons: location and time. Overall, the distribution of voltage distortion is similar to a logarithmic normal distribution because the distribution due to location is similar to a logarithmic normal distribution. However, the distribution of the cube root of voltage unbalance is similar to a normal distribution. The distributions due to time are nearly symmetrical for both voltage distortion and voltage unbalance.
Analysis of measured data from 1999 to 2006 revealed that the recent annual trend of voltage distortion was certainly decreasing. The Japanese government issued limits on harmonic currents from home appliances in 1994. Thus, it seems to have taken five years for the effect of these limits to appear as the appliances that were in use in 1994 were slowly replaced with ones that conformed to the limits.
When power quality is measured, it is important to determine the reliability of the measured data. The number of measuring points necessary to estimate confidence intervals was clarified. To estimate them within a ± 10% error, 50 and 120 measuring points are necessary for voltage distortion and voltage unbalance, respectively.