Work

Seasonal Autoregressive Integrated Moving Average Model for Prediction of Electrical Energy Demand of a Base Transceiver Station

Year: 2026

Type: article

Abstract: Efficient energy management is essential for the continuous operation of Base Transceiver Stations (BTS), which are critical to mobile communication infrastructure. Power shortages can lead to service disruptions, necessitating accurate energy demand forecasting for optimal energy planning. This study developed a dual SARIMA models using 36 months (May 2021–April 2024) of field-measured hourly and daily electrical energy data from a BTS in Abeokuta, Nigeria with average load of 4.2 kW. The long-term daily model SARIMA (0,1,0)(1,0,1)₇ captures weekly traffic seasonality, while the short-term hourly model SARIMA(1,0,0)(0,1,2)₂₄ addresses diurnal peaks. The performance of the model was evaluated using Mean Absolute Percentage Error (MAPE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE). Results showed that the long-term approach had a lower MAPE (7.67%) compared to the short-term approach (13.32%), indicating better accuracy. However, the short-term approach outperformed in terms of RMSE (0.63 kW) and MSE (0.40 kW), compared to 0.78 kW and 0.62 kW, respectively, for the long-term approach. Utilizing the forecasted energy demand, HOMER Pro software was employed to design an optimized hybrid energy system integrating solar energy and a biogas co-fired generator. Results yielded an economically optimal solar-biogas-battery configuration: 40 kW Canadian Solar PV, 16 kW CAT biogas generator, and 32.8 kWh LiFePO₄ storage delivering zero unmet loads. Key metrics include Levelized Cost of Energy (COE) of $0.01666/kWh, Net Present Cost (NPC) of $15,000, and 8-year payback, representing a 75% cost reduction versus diesel-only alternatives. SARIMA forecasting thus bridges precise BTS load prediction with cost-effective renewable sizing, enabling sustainable telecom expansion across off-grid regions.

Source: Nigerian Journal of Technology

DOI: 10.4314/njt.v45i2.2

Authors: TOLULOPE ERINOSHO, K. A. Amusa, Abdultaofeek, Fatai Kazeem, Imhade Okokpujie

Institutions: Federal University of Agriculture, Abeokuta, Summit University, Offa, Afe Babalola University

Language: English

Topic: Energy Load and Power Forecasting

Subfield: Electrical and Electronic Engineering

Field: Engineering

Domain: Physical Sciences

SDGs: Affordable and clean energy, Industry, innovation and infrastructure, Climate action

Open Access status: gold

Estimated APC paid: $185