IMPLEMENTATION OF SHORT-TERM LOAD FORECASTING USING ANN
Advisor:
Dr. Suhail Aftab Qureshi
Abstract:
In a deregulated, competitive power market, utilities tend to maintain their generation
reserve close to the minimum required by an independent system operator. This makes
a need for an accurate instantaneous-load forecast for the coming several dozen minutes.
This work is related to the Short-Term Load Forecasting (STLF) in power system
operations. A three layered Feed-Forward Neural Network (FFN), trained by Resilient BackPropagation(RBP),
is used. This algorithm considered more powerful as compared to the other
traditional approach when the actual loads are forecasted and used as input variables. It
provides more reliable forecasts, especially when the weather conditions are different
from those represented in the training data. The proposed model can provide 24 hour
ahead load forecasts. Separate ANNs are utilized for load forecasting of one hour to
twenty-four hours ahead. The inputs to the ANN are previous day load with other
factors related to it, and the output of the ANN is the predicted load for the next day
for a given hour. Artificial Neural Network (ANN) in MATLAB software was used in
solving the forecasting problem. The average error percentage was determined by using
the Mean Absolute Percentage Error (MAPE). The result of Multi-Layer Perceptron
(MLP) network model, used for one day ahead short term load forecast, shows that
MLP network has a good performance and accurate prediction was achieved for this
model. Its forecasting reliabilities were evaluated by computing the mean absolute error
between the actual and predicted values. We were able to obtain Mean Absolute
Percentage Error (MAPE) of 1.9214% which represents a high degree of accuracy.