Towards the Establishment of a Robust Load Forecasting Mechanism in Tanzania Grid: A Study to Investigate Effect of Air Temperature and Daytime on Electricity Consumption in Residential Buildings

Abstract

- The current phenomenon of electric power management is towards the adoption of smart grid technologies to achieve efficient utility management processes such as transmission and distribution. Electric load forecasting has become an important aspect of smart grid technologies due to its capability of anticipating the power demand of a particular domain. The effective design of any load forecasting mechanism requires a critical investigation of electricity consumption determinants, following the diversification of social, economic, meteorological, and demographic grounds. Many research works have attempted to investigate the effect of temperature and daytime on electricity usage, targeting a specific country. The works have reported the existence of differing degrees of causality based on a particular country under investigation. The variation in findings from different research works has been a motivation to establish the study to examine the impact of air temperature and day time on load consumption in Tanzania. Four-years load and weather data have been collected from Tanzania Electric Supply Company (TANESCO) and Tanzania Meteorological Agency (TMA) respectively. The k-mean algorithm is used to detect outliers and missing values in the load dataset before further processing. Furthermore, the Shapiro-Wilk normality test method is applied to identify data distribution patterns which in turn leads us to the correct choice for Spearman’s rank correlation method. Results indicate that there is no linear relationship between electricity consumption and air temperature in residential buildings. Finally, findings indicate the existence of a strong causality degree between electricity consumption and day time in Tanzania.