[2210.14555] Modeling Hourly Power Consumption During a Power Crisis

[Submitted on 26 Oct 2022]

Download PDF

Overview: In this paper, two-state Markov-switching autoregressive (MS-AR) and autoregressive (AR) models are used to study the dynamic hourly power consumption during the period of the power crisis in Ghana (the “damsa” period). capture behavior. Hourly data from 01/01/2014 to 12/31/2014 was obtained from Ghana Grid and used for the study. Using various information criteria, MS(2)-AR(4) was selected as the best model to describe the dynamic behavior of power consumption during the power crisis in Ghana. The parameters of the MS(2)-AR(4) model are estimated using the expectation-maximization algorithm. Based on this result, we estimate an 87\% chance of staying in the low power regime. The expected duration of the low power consumption regime is 7.8 hours each day and the high power consumption regime is expected to last 2.3 hours each day. The proposed model is robust compared to autoregressive models. This is because it effectively captures the dynamics of power demand over time through peaks and large fluctuations in consumption patterns. Similarly, the model can identify distinct regime changes related to power consumption during periods of power crises.