Characterization of power quality in low voltage networks based on modeling by mixture distributions

Abstract

The paper describes a new method for modeling power quality parameters (e.g. harmonics) particularly in public low voltage networks. The technique is based on the correlation between the trend of power quality parameters in a low voltage network and the characteristics of the consumer structure connected to the MV/LV-substation. A mixture of single Gaussian distributions is fitted to the data of each power quality parameter measured during a predefined time-interval (e.g. 1 day) at the low voltage busbar of the substation. The parameters of the mixture distribution represent the typical properties of the consumer structure very closely. In most cases the value of each parameter can be assigned to a characteristic of the electrical behavior of the consumer structure. This is demonstrated by several examples. The quantification of the correlation between the characteristic of consumer structure on the one hand side and the power quality on the other hand side enables efficient and easy-to-use methods for power quality evaluation. Such methods can e.g. support utilities in network and investment planning or regulatory authorities developing enhanced regulation schemes including power quality