Modelling and forecasting hourly electricity demand in West African countries

Abstract

The Economic Community of West African States aims to achieve 100% electrification rates by 2030 in all member countries. To achieve this ambitious target, electricity generation capacities need to be increased significantly. Forecasting hourly electricity demand is imperative for capacity planners in optimizing investment options and ensuring reliable electricity supply. However, modelling hourly electricity demand in developing countries can be a challenge due to paucity of historical demand data and methodological frameworks that adequately capture technology transitions and urban-rural communities. In this study, we address this gap by developing an hourly electricity demand model for 14 West African countries in the year 2016 and 2030.The model takes into accounts electrification rates, available household appliances, occupancy patterns of household members, type of day, available daylight hours and hourly weather conditions. Annual electricity demand in nonresidential sectors and electricity access rates in urban and rural households are forecasted using multiple regression analysis. We validated the developed model using actual 2016 monthly and annual electricity demand data. The results show the seasonal variations of electricity demand, with hourly electricity demand in dry seasons relatively higher than demand in wet seasons. The results also indicate that in 2030, electricity demand in the West African region is estimated to be five times its 2016 level. The methodology presented in this study can be applicable for modelling hourly electricity demand in developing countries that have scarce historical hourly demand data, a significant electricity supply-demand gap, and varying electricity access rates in urban and rural areas.

M-LED: Multi-sectoral Latent Electricity Demand Assessment for Energy Access Planning

“…The baskets of appliances are obtained through a literature review (ref. (Adeoye and Spataru, 2019;Blodgett et al, 2017;Kotikot et al, 2018;Lee et al, 2016b;Monyei et al, 2019;Monyei and Adewumi, 2017;Sprei, 2002;Thom, 2000)) supported by the authors' personal experience. The compiled database is reported in Supplementary File F1, where every category of users is characterized by a corresponding usage pattern of the owned appliances, differentiating every month to account for seasonality of the uses.…”

“…We construct 5x2 = 10 archetypical types (five in urban areas, and five in rural settlements) of households by electrical appliance ownership and use patterns. These are designed starting from a systematic screening of the literature (Adeoye and Spataru, 2019;Blodgett et al, 2017;Kotikot et al, 2018;Lee et al, 2016b;Monyei et al, 2019;Monyei and Adewumi, 2017;Sprei, 2002;Thom, 2000) about electricity consumption in developing countries and parametrised based on data from recent field visits in Kenya by the authors and their team (2019). The empirical screening provides the rationale to compile tables of appliances and usage patterns (refer to SI-A2) for each household type.…”

A network analysis of electricity demand and the cryptocurrency markets

“…For the OECD countries, the average household residential elasticity for electricity demand is 72% in absolute terms (Krishnamurthy & Kristrom, 2015). Furthermore, similar studies have been carried out in west African countries (Adeoye & Spataru, 2019), Greece (Angelopoulos, Siskos, & Psarras, 2019), Europe (Cialani & Mortazavi, 2018), Pakistan (Mirijat et al, 2018), Italy (Alberini, Pretico, Shen, & Torriti, 2019), Australia (Roberts, Haghdadi, Bruce, & MacGill, 2019), Portuguese (Figueiredo, Nunes, & Panao, 2020), Jamaica (Campbell, 2018), Japan (Lliopoulos, Esteban, & Kudo, 2020), United States (Obringer, Kumar, & Nateghi, 2019), Sweden (Jelica, Taljegard, Thorson, & Johnsson, 2018), Spain (Pérez‐García & Moral‐Carcedo, 2016), etc.…”

Short-Term Electricity Load Forecasting with Machine Learning

“…The short-term electricity load forecasting is implemented to solve a wide range of needs, providing a wide range of applications. The most evident difference between research is the load scale, from a single transformer [9] to buildings [10], to cities [11], regions [12], and even countries [13]. The second most crucial distinction among the research field is the forecasting horizon.…”

“…For example, the authors of Reference [26] used it to forecast the hourly weekly load in Thailand, obtaining an average mean absolute percentage error (MAPE) of 7.71% for 250 testing weeks and pointing out that temperature is a primary factor to predict load. Similarly, the authors of Reference [13] utilized MLR to forecast electricity consumption 24 h ahead for 14 west-African countries, considering weather variables like temperature, humidity, and daylight hours. The researchers that have implemented MLR agreed on the fast training and interpretability this model offers, although it shows poor performance for irregular load profiles.…”