Measurement-based load parameter modelling for technical and tariff studies of medium voltage feeders

Transformation in the South African power sector and new load patterns necessitated a review of models used for financial, technical tariff analysis. This pilot study took advantage available data customer measurements on medium Voltage (MV) feeders Eskom's database. Load with distinct profiles typical days were developed non-overlapping classes using set coherent parameters derived from MV chronological k-means algorithm. The results suggested that two can be to summer winter instead 365 hourly simulations. also reveal classification improved when are directed towards specific objectives, algorithm is supervised exogenous (external) loads. A comparison economic activity class suggests there sub-clusters identifiable within classes. proposed process practical, implementable suitable various studies networks.