Abstract
Water was the first renewable resource to be used for producing both mechanical and electrical energy. In the early 90s, large-scale hydropower was developed in order to collect water at high geodetic altitudes, constituting an upper reservoir, and to exploit its energy by the means of hydraulic turbines inserted in penstocks. Hydraulic turbines are installed between the upper and the lower reservoirs with the task of converting the energy content of the water into mechanical power and, subsequently, through an electric generator into electrical energy. Nowadays, almost all the geodetic altitudes are already exploited and, for this reason, the hydropower sector is moving from the large-sale to the small-scale one. Some new hydraulic turbines (Agnew, Cross-flow and Turgo) suitable for the small-scale hydropower sector were redesigned considering several technical aspects related to the ones used in the large-scale hydropower plants. In particular, Pump-as-Turbine (PaT) technology is taking the field for being installed in rural and remote zones where an electric grid cannot be built or for energy recovery purposes in both civil and industrial plants. In literature, PaT’s performance, related to the Best Efficiency Point (BEP) and to off-design operating conditions, are not easy to find because pumps’ manufacturers provide only performance’s curves in direct mode. For this reason, this PhD thesis focuses the attention on the forecast models able to predict their performance. Theoretical methods for evaluating both BEP values and the off-design operating conditions, respectively, were developed, also by using Artificial Neural Network (ANN), and validated with a tested PaT. In addition, laboratory tests on a centrifugal PaT were performed together with numerical simulations on the same machine and also on different models in order to better analyse the characteristics of this machine on the fluid dynamic point of view and to propose geometry modifications to improve its efficiency in turbine mode, not affecting the pump mode operation. Finally, the application of this technology on both civil and industrial cases was studied on both technical and economical points of views.