Practical statistical methods in distribution load estimation

The paper deals with load profiles of electrical consumers of distribution networks (low voltage busbars of distribution transformers). These profiles are the most important data not only for performing off-line network analyses (what-if analysis, operation and development planning, etc.), but also for providing satisfactory results of running of all real-time functions (State Estimation, Power Flow, Fault Management, etc.). The load profiles are treated as random variables. The following two proofs are the main objective of the paper: 1) The load profiles with the necessary accuracy can be derived on the basis of a relatively small number of experiments (in-field measurements) and 2) The normal distribution is the best guess for the statistical behaviour of the most number of loads.