Intelligent Demand Forecasting Using Deep Learning

Authors

Lithicka Anandavel
Department of Computing Technologies, SRM Institute of Science and Technology, Kattankulathur, Chennai, India
Ansh Sharma
Department of Computing Technologies, SRM Institute of Science and Technology, Kattankulathur, Chennai, India
Naveenkumar S.
Department of Networking and Communications, Srm Institute of Science and Technology, Kattankulathur, Chennai, India
Suresh Sankaranarayanan
Department of Networking and Communications, Srm Institute of Science and Technology, Kattankulathur, Chennai, India
Anis Salwa Binti Mohd Khairuddin
Department of Electrical Engineering, Faculty of Engineering, Universiti Malaya, 50603, Kuala Lumpur, Malaysia

Synopsis

One type of energy demand is the electricity demand, which measures the electricity consumption Wh (watt-hour). Forecasting this electricity demand is very crucial and plays a fundamental role in the electrical industry, as it provides the basis for making decisions in the operation and planning procedures of power systems. Forecasting is important for development experts and are of great interest to energy authorities, power utilities, and private investors. Inaccurate projections can have disastrous social and economic implications, whether they over-or under-predict demand. Supply shortages and forced power outages occur from underestimating demand, wreaking havoc on productivity and economic growth. Overestimating demand can result in overinvestment in generation capacity, financial hardship, and, eventually, higher power costs. This paper has validated several methodologies such as ARIMA, XGBOOST, LSTM and Bi-LSTM towards forecasting the energy demand for different regions of Australia during different season. The models were validated towards energy demand forecasting in terms of error and accuracy resulting in LSTM with 2 layers outperforming the other models.

TechPost2022
Published
December 28, 2022
Online ISSN
2582-3922