With the increase in technological
advancement, the need for an energy conservation
system is increasing. It is necessary to build an effective
and predictive system now that existing energy
consumption models emphasize on and forecast
accuracy. Consumption of energy has increased many
folds and now has become a concerning issue. The
utilization of energy has increased with human
development and growth. The main reasons for these
problems are uncontrolled power usage, including
excessive consumption, lack of optimal design, and
energy wastage.
The purpose of this work is to predict the future
trend in power usage for any system that monitors and
requires this information in real-time. The best model to
accomplish this goal was evaluated using recurrent
neural networks (RNN) and long short-term memory
(LSTM). Results from experiments demonstrate great
accuracy and require fewer computer resources during
model training than competing models.
Keywords :
Power Consumption, Short-term load forecasting, electricity markets, spot prices, Recurrent Neural Networks (RNN).