In this paper, the performance of the moving average and moving median load estimation techniques is investigated using aggregated measurements of domestic smart meters. The load estimation techniques were tested using forward and backward walk approaches. Forward walk aims to estimate future load measurements using past measurements while backward walk estimates missing past measurements of the load using more recent measurements. Simulation results show that the moving average combined with forward walk produces load estimates with higher accuracies than the moving median and backward walk.
Keywords load estimation, moving average, smart meter measurements