Increasing the reliability of indices for power quality assessment in distribution networks

Nowadays power quality, especially voltage quality becomes more and more important for utilities, manufacturers and customers as well as for the regulatory authorities. Therefore a growing number of voltage quality monitoring campaigns with continuously increasing measurement times, considerably longer than one week, are observed. The assessment of voltage quality according to common standards (e.g. IEC 61000-2-x or EN 50160) is usually based on an evaluation interval of one week while the handling of measurements lasting more than a week is not clearly specified. Furthermore the measurement data contain lots of information that is not used by the actual applied assessment methods in the most efficient way. Especially for internal planning purposes within the utilities the knowledge of medium- and long-term trends and variations as well as information about the typical time-dependent customer behavior can be very useful. The paper presents a method that includes medium- and long- term variations (e.g. seasonal variations) of continuous voltage quality parameters (e.g. harmonics) into the calculation of actual levels. The method is based on statistical tolerance bounds and produces more reliable values that can be compared with limits given by actual standards or used for further analyses. After a description of actual assessment methods the paper gives a systematical overview on the variations of voltage quality parameters and typical reasons for that. Next the statistical basics are explained on several examples. Based on the experiences of the authors finally some recommendations for practical application of the suggested method are given.