Electrical Load Forecasting using ARIMA, Prophet and LSTM Networks

Abstract

Forecasting electrical load plays a vital role in power system planning. However, it is quite difficult to forecast electrical load, as the load on the system varies continuously concerning time and seasons. In this paper, we are proposing an advanced artificial neural network model to forecast short-term electrical load. The proposed method tested on historical data collected from Karnataka power corporation, India, and test results compared with other data-driven models viz. ARIMA, RNN, LSTM, and Prophet. The accuracy and RMSE values were calculated and observed that the proposed model was superior in a day and weekly ahead electrical load forecasting.

Energy Demand Forecasting and Optimizing Electric Systems for Developing Countries

“…Compared to other techniques under different study periods and forecasting horizons, the NARX technique proved superior to other econometric approaches in all studied cases. Another comparison between LSTM and ARMA, ARIMA, and SARIMA shows that LSTM has the slightest deviation in prediction, which is only 3.2%, opting for its applicability in generation and distribution planning [70]. Phyo et al [29] observed in their experiment that LSTM outperformed the Deep belief network (DBN) with a yearly average MAPE of 3.79% only through smoothing out raw data.…”

“…Dayahead is most relevant in the case of optimal generator unit commitment, whereas annual forecasting helps in the system and economic planning. [13], [16], [19], [24], [25], [31], [38], [42], [43], [45], [53]- [58], [60]- [67], [70], [81], [95], [98] the limitations of single models, though they sometimes introduce complex structures. Time series techniques are computationally inexpensive compared to other models due to having a non-complex structure.…”