Conventional and future prediction methods for turbine design parameters in drinking water and irrigation pipelines

This thesis aims to pioneer the utilization of untapped energy in water supply and irrigation canal lines by installing hydroelectric power plants (HEPP) with pressure reducing valves. The study presents work conducted on HEPP installation using inline pipe turbine systems on the P11 Pumping Water Conveyance Line in the Izmir region and the Akbaş HES Irrigation Canal in the Denizli region. The efforts include feasibility studies based on existing datasets and future time predictions to determine turbine characteristic values. Flow rate-time series for both pipelines were organized on a daily and hourly basis. Traditional methods were employed to transition from the time domain to the frequency domain, resulting in the creation of flow-duration curves. Design flow rate values were determined from these curves, leading to the selection of a Pelton turbine for Denizli and a Francis turbine for Izmir. The financial return to the investor and the annual energy production capacity of the system were calculated based on the design flow rate and head values determined through traditional methods. LSTM and GRU models in artificial intelligence were utilized to forecast future time series by analyzing historical data, providing insights into the historical impact on turbine design and production values. Predictions indicate an increase in flow rate and profit for the P11 Line over the next year, while the Akbaş Line exhibits a decrease in flow rate and profit in the feasibility study conducted with hourly data, and a decrease in flow with an increase in profit conducted with daily data.