Evaluating Power Usage Patterns

Precise electricity consumption predictions are essential for efficient energy management, resource allocation, and power system stability, particularly in expanding urban areas like Erbil. Time series models are crucial tools for capturing trends, seasonal variations, and structural shifts in energy use patterns. This study aims to forecast monthly electricity consumption in Erbil for 2025 using the seasonal autoregressive integrated moving average (SARIMA) approach. Historical monthly electricity consumption data from 2015 to 2024 (120 observations) were analyzed. Preprocessing involved logarithmic transformation to stabilize variance and appropriate differencing to achieve stationarity. Model selection prioritized evaluation criteria such as the Akaike information criterion and mean squared error (MSE). The SARIMA(1,1,1)×(0,1,1)12 model yielded the lowest MSE (0.0487347) and was identified as the optimal model, with statistically significant parameters. The resulting forecasts for 2025 indicate notable seasonality, with predicted monthly averages ranging from 955 MW in January to 916 MW in December. This study provides a validated time series model tailored to Erbil’s consumption dynamics, offering a robust foundation for improved energy demand forecasting and resource planning.