F3_Abstract.tex

%      3.   Abstract
%---------------------------

\chapter{Abstract}

Penetration of distributed energy resources in distribution networks is predicted to increase dramatically in the next seven years, bringing with it the opportunity for utilities to have a greater presence at low levels of the network.
To achieve this effectively, utilities will require accurate short term load forecasts.
This thesis presents several neural network-based load forecasting models that are applicable to forecasting both low and high levels of aggregate load.
The models are simple to implement and require no tuning when used on different feeders.

One model is implemented as part of the Bruny Island CONSORT residential battery trial to produce a 24-hour horizon half-hourly online forecast every five minutes.
When forecasting during anomalous peak holiday periods on a feeder that has a typical load of less than 1000kVA the forecasting system is able to successfully forecast major peaks with sufficient lead time and accuracy to enable the fleet of batteries to charge ahead of time and provide network support.