Abstract:
Accurate load forecasting is very beneficial for the efficient and economical operation of
a power supply utility. The power system load forecast can be categorized into long-term
forecast and short-term forecast. Long term forecast usually covers a time span of one
year to ten years and is needed for decision making regarding capacity expansion and
long term capital investment return studies. Short term forecast on the other hand
provides information about system load characteristics from one hour to twenty four
hours or upto a few days into the future. Short term forecast is necessary for the efficient
and reliable operation of an electric utility. It is needed for economic dispatch, unit
commitment, energy sale/purchase decision and load management.
Keeping in view the importance of accurately forecasting the system load an improved
algorithm has been developed for short term load forecasting of an electric utility. In this
perspective the approach adopted is based upon time series analysis. Emphasis is on
modeling the interconnected power system in state space form. Stochastic Approximation
algorithm and Kalman Predictor are applied for load forecasting upto one hour. A
software package is developed with the proposed implementation method for electric
supply utilities. The application of this software package to real load data obtained from
WAPDA (Water and Power Development Authority) has shown good results. The salient
feature of this software package is that it can provide a breakup of load demand at each
grid station, which is not currently available with WAPDA. This breakup can result in
reduced transmission losses with corresponding economic benefits.
