Practical Probabilistic Design Procedures for Medium Voltage Distribution Systems

Power system variables such as load demand, power generation from variable renewable energy plants and the placement of future distributed energy resources are characterized by uncertainty. Probabilistic network design based on probabilistic load flow techniques (PLF) is increasingly useful. However, most reported work on this topic focuses on developing faster and more robust statistical formulations of PLF algorithms. Such a unilateral approach diminishes the importance of understanding how the commonly assumed probabilistic load models are derived, and how the PLF analysis leads to planning decisions. This paper details a holistic approach to probabilistic network design for medium voltage (MV) distribution systems based on developing experience in South Africa. The framework includes the derivation of load models from measurements, their application in network analysis, and the interpretation of statistical outputs for planning decisions. The findings have implications for data preparation and the application of the results in MV network planning.