Operational and Economic challenges due to Run-of-River (RoR) Hydro and ways to address the challenges

Abstract

Run-of-River (RoR) hydro generating station is a unique and eco-friendly method to harvest the energy of water. Even though this type of hydro generating station has multiple benefits in comparison to storage type, several challenges are faced by hydro power plant operators and power system operators. Major portion of these challenges are attributed to sudden variation in water inflow due to changes in weather conditions in the upstream or catchment area of these power plants. This paper describes such challenges encountered due to sudden variation of water level in the pondage of a RoR hydro power plant named Ranganadi Hydro Electric Power (RHEP) plant with installed capacity of 405 MW, situated in the state of Arunachal Pradesh of India. The RHEP power plant has a small pondage capacity of around 3 hours (while running at its installed capacity). Sudden fluctuation in water inflow often leads to unanticipated variation of water level in its pondage. The fundamental approach to get rid of challenges posed by this unanticipated variation of water level is to have an accurate forecast system to predict the water level in different time horizons. Two independent methodologies of forecasting water level using Artificial Neural Network (ANN) and Vector Auto regression (VAR) have been discussed in this paper. Both the methods have been employed to predict half hourly water level in RHEP pondage for a day by using historical data of water level and weather parameters like rainfall, humidity, and temperature. The results obtained from the two methods have been compared in this paper. Accurate forecast of water level shall ease off the challenges pertaining to variation in RoR hydro generation.