Abstract:
The main purposes of this study are to model annual electricity demand in both domestic and
nondomestic (industrial and commercial) sectors, and thereby forecast the electricity demand by
2020 in each sector separately in Sri Lanka. The annual electricity demand data of both domestic
and non-domestic sectors from 1969 to 2018 in Sri Lanka were collected for accurate modelling of
demand. Four statistical models including Exponential, Polynomial, Autoregressive Integrated
Moving Average (ARIMA) and Harvey Logistic were used to model the electricity demand in each
sector. The accuracy of the forecasted values from each model was tested using the mean absolute
percentage error (MAPE). The model with minimum MAPE was selected as the best model to
predict the electricity demand by 2020. It is revealed that, the Exponential model had the highest
predictive performance compared to other three models. Therefore, the Exponential model was
selected as the best model to predict future demand. According to the demand predictions, the
annual electricity demand in both domestic and non-domestic sectors in Sri Lanka will increase
exponentially with expected demands of 6,410GWh and 12,566GWh respectively by year 2020.
Indeed, electricity plays an important role in providing required energy for day to day life. Unlike
past, the demand in Sri Lanka is growing rapidly. Due to non-storable nature of electricity it is
important to assure that supply of electricity must be always available to satisfy the demand. The
results of this study help to construct a well design power development plan which assures
continuous supply of electricity to satisfy the growing demand in future. Also, results enable actions
to reduce wastage of overproducing of electricity and highlight the necessity of new avenues of
producing electricity to cater demands in future.
