A Model Based Approach to Simulate Excess Water of Reservoirs in Sri Lanka

Abstract

Water reservoirs have been a part and parcel of the life style of community living around it. Its operation and maintenance directly influence the dependents as well as community living around and suburbs of the reservoir. While reservoir mechanism is an excellent solution to preserve water, its operations can become precarious during a time of a heavy rainy season where the amount of water accumulated to the reservoir increases significantly during this period resulting in reservoir being unable to handle the subsequent water inflow and ultimately releases this water at a daunting rate making the downstream flood. So our research focuses on addressing this scenario through Information and Communication Technology based approach where decision making related to reservoir operations is supported as well as enhanced through ICT. We have identified that monitoring reservoir water level and smooth water release are two critical factors of a flood scenario and the research is centered on how ICT can improve on these two critical factors. Mathematical and Artificial Neural Network approaches have been taken in the research. Neural Network model showed some satisfactory results and it can be enhanced with availability of accurate and critical data. We implemented a prototype system which supports decision making and monitor, how reservoir water level fluctuate and how excess water can be released. Through this we were able to achieve our target of conceptualizing the reservoir operations though an ICT solution. Evaluation of the domain experts validated the success of our concept.