H41G-1447
Improving the global efficiency in small hydropower practice

Thursday, 17 December 2015
Poster Hall (Moscone South)
Lorenzo Gorla, EPFL Swiss Federal Institute of Technology Lausanne, Lausanne, Switzerland
Abstract:
The global increase in energy production from renewable sources has seen river exploitation for small hydropower plants to also grow considerably in the last decade. River intakes used to divert water from the main course to the power plant are at the base of such practice. A key issue concern with finding innovative concepts to both design and manage such structures in order to improve classic operational rules. Among these, the Minimal Flow Release (MFR) concept has long been used in spite of its environmental inconsistency.

In this work, we show that the economical and ecological efficiency of diverting water for energy production in small hydropower plants can be improved towards sustainability by engineering a novel class of flow-redistribution policies. We use the mathematical form of the Fermi-Dirac statistical distribution to define non-proportional dynamic flow-redistribution rules, which broadens the spectrum of dynamic flow releases based on proportional redistribution. The theoretical background as well as the economic interpretation is presented and applied to three case studies in order to systematically test the global performance of such policies. Out of numerical simulations, a Pareto frontier emerges in the economic vs environmental efficiency plot, which show that non-proportional distribution policies improve both efficiencies with respect to those obtained from some traditional MFR and proportional policies. This picture is shown also for long term climatic scenarios affecting water availability and the natural flow regime.

In a time of intense and increasing exploitation close to resource saturation, preserving natural river reaches requires to abandon inappropriate static release policies in favor of non-proportional ones towards a sustainable use of the water resource.