Forecasting domestic hourly load profiles using vector regressions

Electricity distribution to households and business is the final step in the electricity generation, transmission and distribution value chain. Redistributed energy by municipalities to the South African domestic market is key to planning energy demand and supply. In this study we use the concept of vector regression modelling in forecasting total domestic hourly load profiles. Vector regression is done by aligning the time series so that days and hours match across the time period. A linear model is then fitted across this vector [1]. MSE, MAD and the MAPE are used in the study to evaluate model accuracy. Results from this study show that domestic load profiles are seasonal with peaks in the morning and in the evening around 8pm. Also winter months show high usage compared to the other seasons and this domestic load is by far the most dominant profile shape of netenergy sent out from generation.