Water Flow Patterns Generation Model for Grid Connected Run-of-River Hydro Power Plant

Integration of renewable energy in the electric grid is an important aspect of grid operation. Hydropower plants, especially the run-of-river is increasing continuously, and it is the significant renewable energy source. High uncertainty and unpredictability is associated with hydropower generation due to change in weather conditions and water flows. For reliable and continuous operation of power plants modelling of water flow patterns is considered as a critical step to meet operational and maintenance challenges. In this paper, a probabilistic model consists of Weibull distribution and Neural Networks is developed. The proposed probabilistic model estimates the uncertainties of high-water flow and low water flow season patterns. Different scenarios of water flow patterns are generated based on the estimated Weibull distribution parameters. Proposed probabilistic model is validated using Goodness-of-Fit (GOF) technique such as Average Mean Absolute Percentage Error (AMAPE). Generated scenarios are helpful for operational planning, flood frequency analysis, reservoir management, offline testing and distributed generation studies.