Hybrid demand model for load estimation and short term loadforecasting in distribution electric systems

摘要

A new hybrid demand model to enhance load modeling in distributionnapplications is proposed in this paper. This model is specially wellnsuited for the applications emerging from the new structure of the powernsector worldwide. The modeling is performed in two steps. The first onenis a state space model for load estimation at the selected points in thennetwork. It uses information already available in the utility and alsonsome measurements, and it suggests measurement planning for meternlocation and bad data detection. The second step is an artificial neuralnnetwork (ANN) model for short-term load forecasting which is able toncope with the nonlinear behavior of the load. The model has beennvalidated in simulation studies and using historical data from thendistribution level