外文文献.pdf

Load Forecasting for Substations in a Distribution
System ABSTRACT
Accurate load forecasting is essential for distribution planning. This paper
presents an efficient method to predict the load growth load data for substations only.
The method not only eliminates errors resulting from load transfer between
substations, but also predicts the load growth accurately for substations with
insufficient historical load data.
The method combines the load transfer coupling between substations, improved
vacant area inference and disaggregation with the conventional time trend method to
improve the accuracy of load forecasting. Four district areas in Taiwan Power
Company (Tai power) were selected for demonstration of the proposed load
forecasting model. It is proved that more accurate load forecasts can be achieved than
those obtained from the conventional trend models.
INTRODUCTION
Load forecasting for a distribution system is required to predict the load growth
of all substations in a district for the next 5 to 10 years; this is medium-term
forecasting. The bulk of distribution system planning involves the siting and capacity
of substations, the routing of feeders and many other decisions relating to both the
locations and amounts of future load growth. Load forecasting is the fundamental
basis of optimal dist