NEURAL NETWORK APPLICATION TO SHORT TERM LOAD FORECAST

Power system planning and operation is an important part for power systems
industry. By having a good planning and operation, the quality of power supplied will
be improved ensuring both consumer and power provider getting their share equally.
In this case, the most challenging part is the prediction of how much the load that will
be used by the consumer for a short period of time. This prediction is called load
forecasting. This will be very useful to every power system company as the trending
for the load demand is different for each geographical location. There are different
methods to do the load forecasting. One of the project involved MATLAB program
for the short term load forecasting (STLF) using Artificial Neural Network (ANN)
model. We are using Multilayer Perceptron (MLP) Neural Network architecture, it
will improve the forecast value significantly by obtain a very small mean absolute
percentage error (MAPE). By getting a smaller MAPE, it represents higher forecast
accuracy of the model itself. The elements in this report contain of an introduction,
problem statement, objectives, literature review and methodology which was used to
solve the forecasting problems. The discussion of the obtained results will be looked
further in this project.