Affiliation:
1. Software Centered University Project Group, Chonnam National University, 300 Yongbong-dong, Buk-gu, Gwangju 61186, Republic of Korea
2. JLG Corporation, 93 Hyou-ro, Nam-gu, Gwangju 61756, Republic of Korea
Abstract
Electricity consumption forecasting plays a crucial role in improving energy efficiency, ensuring stable power supply, reducing energy costs, optimizing facility management, and promoting environmental conservation. Accurate predictions help optimize energy system operations, reduce energy wastage, cut costs, and decrease carbon emissions. Consequently, the research on electricity consumption forecasting algorithms is thriving. However, to overcome challenges like data imbalances, data quality issues, seasonal variations, and event handling, recent forecasting models employ various approaches, including probability and statistics, machine learning, and deep learning. This study proposes a short- and medium-term electricity consumption prediction algorithm by combining the GRU model suitable for long-term forecasting and the Prophet model suitable for seasonality and event handling. (1) The preprocessed data propose the Prophet model in the first step for seasonality and event handling prediction. (2) In the second step, seven multivariate data are experimented with using GRU. Specifically, the seven multivariate data consist of six meteorological data and the residuals between the predicted data from the proposed Prophet model in Step 1 and the observed data. These are utilized to predict electricity consumption at 15 min intervals. (3) Electricity consumption is predicted for short-term (2 days and 7 days) and medium-term (15 days and 30 days) scenarios. The proposed approach outperforms both the Prophet and GRU models, reducing prediction errors and offering valuable insights into electricity consumption patterns.
Funder
MSIT (Ministry of Science and ICT), Korea
IITP
Subject
Management, Monitoring, Policy and Law,Renewable Energy, Sustainability and the Environment,Geography, Planning and Development,Building and Construction
Reference43 articles.
1. International Renewable Energy Agency (2020). Renewable Capacity Statistics, International Renewable Energy Agency.
2. (2023, August 01). Rethink Energy Site. Available online: https://rethinkresearch.biz/product/rethink-energy/.
3. Lv, L., Luo, L., and Yang, Y. (2022). Distribution Line Load Predicting and Heavy Overload Warning Model Based on Prophet Method. Sustainability, 14.
4. Grazioli, G., Chlela, G., Selosse, S., and Maïzi, N. (2022). The Multi-Facets of Increasing the Renewable Energy Integration in Power Systems. Energies, 15.
5. Xue, B., and Keng, J. (2012, January 12–14). Dynamic Transverse Correction Method of Middle and Long Term Energy Forecasting Based on Statistic of Forecasting Errors. Proceedings of the Conference on Power and Energy IPEC, Ho Chi Minh City, Vietnam.