Optimizing Energy Output at Canyon Hydro Power Station, a Case Study

Abstract

Due to the rainfed central hills in Sri Lanka, the country enjoyed low cost, renewable hydroelectricity generation since 1950's. With the completion of nearly all major hydro site developments and the growth in electricity demand hydro generation currently contributes only 40% to the gross electricity generation in the country. In view of increasing dependence on carbon-polluting fuels for electricity generation, development and utilization of hydropower potential to the maximum level and in an optimum manner would bring multiple benefits. This paper presents a case study to achieve optimum generation at the Canyon, new Laxapana cascade. The current practice of loading cascaded Laxapana complex plants is mainly based on meeting pond balancing constraints and the spinning reserve requirements. Efficient performance of the plants are not given due consideration in the dispatch process. However, such considerations would lead to loading patterns which generate same amount of electricity using less water due to the fact that especially in case of Francis turbiness the efficiency significantly varies with the load. As an example a single Canyon unit delivering 14.56 MW at full reservoir level uses 0.578 m3 of water per second per MW whereas the same unit delivering 25.57 MW needs only 0.553 m3/s per MW. Further, when both machines are running in parallel the resulting head loss is higher than that when a single machine operates to give the same output. An algorithm is developed to obtain all the requested energy output within a specific period using the minimum of water. Rescheduling the dispatch of plants by the new algorithm saved 64388 m3 of water while generating 722MWh of energy. The water saved could generate 29.5 MWh on a later date and amounting to a saving of 4.1% in energy terms. As similar savings can be achieved almost everyday implementation of this algorithm would lead to avoidance of cost through reductions in the thermal generation and associated GHG emissions