Abstract
In today's world of modern technology and rapid increasing demand of electronics, electricity has become an essential and a vital part of our daily life. The under-developed or developing countries face several different challenges in order to manage demand and supply of electricity. The gap between demand and supply of electricity has a very strong effect on the economic growth. The forecasting of energy will play an important role for the policy makers to timely identify the sudden change in demand of electricity under given conditions. To this end, we developed a model for forecasting of electricity generation from hydro-power plants. Besides the parameters from hydropower plant, the forecasting model incorporates the temperature and rainfall in the catchment area of the dam. In this research work we analyzed the data for the energy generation trends on live data of Tarbela Dam, which consist of daily electricity generation for the last five years augmented with the temperature and rainfall data of the dam catchment area. Moreover, we have applied different supervised machine learning and time-series based models to forecast the energy production. The proposed solution is based on future forecasting of energy generation for hydro power plant, which can assist the policy makers in better decision making. The proposed research work will help in minimizing the increasing gap between energy demand and production considering weather conditions of the area. It can also help the power plant management to detect any anomaly or a failure in the electricity production of electricity by studying the deviation from the predicted trend. Our proposed method can forecast the production of electricity generation with Mean Absolute Error of 2.47 only.