In 2000  and 2008 , surveys were performed in Canada to measure the harmonics caused by domestic loads. Each survey was performed over one year in single week intervals for each location; together confirming observations made by Professor Alexander Emanuel in 1995  which revealed that the growth rate of harmonics can be up to 0.35% per year.
Although harmonic orders 9 and 15 were the only orders identified as being critical in the 2008 survey , the safety margin allowing for additional domestic loads is continuously shrinking. However, because of the limited observation time interval, it is possible that the levels of harmonics observed at past survey locations were actually more significant than measured. As demonstrated by a similar survey performed in France , the harmonic level measured in one specific week may as much as double when measured during a separate weekly interval.
Canadian utilities began conducting power quality surveys in 1991. Results from the monitored periods prior to September 2015 are found in CEATI reports numbered CEA 220 D 711A , T984700-5103 , T084700-5142 , T104700-5142B-1 , T104700-5142B-3  and T144700-5142C-3 . Prior to 2010, power quality recorders monitored the quality of supply for one weekper location. Since July 2010 , new harmonic recorders were permanently installed to continuously record harmonic levels at 18 sites across Canada and the United States. These recorders monitored the voltage harmonic distortion in distribution systems that were predominantly supplying residential loads. Beginning September 1st, 2015, this report covers a one year period over which 13 recorders at various locations measured voltage harmonic levels for time durations of up to 52 weeks.
Where possible, efforts were made to have harmonic recorders connected to the same locations since the survey inception in 2010. At several locations, the survey ceased in September 2014 due to non-availability of the sites; nevertheless, 11 locations remained unchanged and have supplied data since September 2010. However, two new locations were added for continuity of the survey: one in 2015 and another in 2016. Thus, the data of 13 locations is analyzed herein for the fourth phase of the project.
The greatest challenge was to record the voltage distortion values at each location for the entire 52-week period. By continuously processing the monitored waveform up to 9 kHz, each recorder saved 2700 values per second and thus accumulated an enormous amount of data. In order to distill pertinent information from such a vast amount of data, the planning level assessment procedure from IEC61000-3-6  was utilized (the IEEE519-2014 standard  also describes a similar procedure). According to this procedure, data recorded and assessed for each consecutive 12 cyclesof the mains frequency serves to compute the weekly 95th percentile of 10-min rms values.
In principle, utility planning levels should be used as the benchmark limit for the harmonic’s voltage levels. However, IEC compatibility levels (CL) were used for comparison purposes for this project since they are higher than the utility planning levels and thus provide prudently conservative, but still fully relevant, results. Compatibility levels are also important for what they represent relative to the EMC model as a whole. These levels are used by utilities to determine planning levels and by the IEC to determine equipment immunity requirements; thus, a violation of compatibility levels will lead to decreased immunity performance for equipment and the possibility of subsequent equipment damage and other undesirable effects. The weekly-assessed data sets were processed using a Rayleigh distribution function to assess the probability of finding a number of locations in the USA and Canada with harmonic levels exceeding the IEC compatibility levels.
The 95th percentile assessment is derived from the statistical cumulative distribution of a computed root mean square of samples that are taken over 10-min consecutive intervals over a one week period. According to this method, 5% of the rms values exceed the level estimated at the 95th percentile. Since the method sorts 1008 rms values per week, the countdown of the 50 highest values gives the range of the 95th percentile.
An analytical method was developed for CEATI project T16400-5174  to assess the impact of the mass penetration of LED bulbs and small chargers. This method was used for this project to estimate the cause of the harmonic voltage measured at one of the locations with the highest third harmonic level. The analytical method allows for the estimation of the voltage level of each customer connected to the feeders supplied by the same HV/MV transformer at the mainsubstation. It also allows for the estimation of the harmonic current emission for each of these customers.
Compliance assessment, Harmonics, Interharmonics, 2-9 Khz components, Power Quality, PQ Measurement, PQ survey methodology, Root cause analysis