Hydropower Energy Recovery from Water Supply Networks Using Pump-As-Turbines: Optimal Selection, Prediction of Global Efficiency, Economic Viability and Large-Scale Resource Assessment
Citation:
Mitrovic, Dorde, Hydropower Energy Recovery from Water Supply Networks Using Pump-As-Turbines: Optimal Selection, Prediction of Global Efficiency, Economic Viability and Large-Scale Resource Assessment, Trinity College Dublin.School of Engineering, 2021Download Item:
Thesis_manuscript_Dorde_Mitrovic_Final_For_Printing.pdf (Dorde Mitrovic - PhD thesis - Final document for printing) 20.86Mb
Abstract:
It is well documented in the literature that the drinking water sector is among the most energy intensive sectors in developed countries. This is because all processes involved in their operation such as extraction, treatment and distribution of water to consumers, require significant amounts of energy. Moreover, all predictions suggest that the energy consumption of the sector is only going to increase in the future, with the increase in population, urbanisation and wealth.
Consequently, researchers and practitioners from the field are constantly exploring new technical solutions to reduce the energy dependency of water supply networks (WSNs). Since the beginning of the last decade, a concept that gained particular interest among the researchers is the concept of hydropower energy recovery (HPER). Namely, this concept assumes installation of different kinds of hydroelectric converters at sites within WSNs with excess pressure, aiming to exploit this excess energy for electricity generation. Nevertheless, majority of the sites from WSNs have small installed powers (usually less than 100kW), which usually makes them not economically viable for installation of the traditional custom-made hydro turbines.
An unconventional type of hydro converters that has been extensively investigated by researchers to be used as a low cost HPER technology at sites within WSNs is so called pump-as-turbine (PAT) technology. PATs are conventional water pumps utilized in reverse as turbines. Their most important advantage and the one that makes PAT technology relevant in comparison to the traditional turbines is their several times lower cost, which originates from their mass production. As these are not intended to work as turbines these of course have their disadvantages as well. Some of the most pronounced are lower part load efficiency and particularly absence of flow regulation device. These disadvantages are particularly relevant for their application in WSNs whose sites are characterized with large flow and head variability
The present PhD dissertation focused on addressing several gaps that have been identified in the literature in the area of application of PATs as HPER devices in WSNs, aiming to foster the HPER concept as a means to reduce the energy dependency of the sector.
After studying the problem of PAT selection to be used as HPER device at locations of existing PRVs, characterized by large variation of flow and head operating conditions, the first part of the thesis describes development of a novel optimisation-based methodology for addressing this problem. The novelties of the presented methodology includes application of a derivative free heuristic optimisation algorithm, a new formulation of the constraints imposed to solution space defined with the boundaries of PATs available on the market, and implementation of the PAT?s operation limits using the values of the mechanical power. This part of the thesis also applies the developed methodology to several real-world PRV case studies to investigate two literature gaps: 1) Can the selection of different objectives lead to the selection of different commercially available PAT families? and 2) Can different operational limits lead to the selection of different commercially available PAT families?
As the selection of an optimal PAT for a particular PRV site and assessment of the HPER plant?s global efficiency is not very straightforward, the second part of this thesis investigates if the global efficiency of the plant could be predicted using only the statistical parameters of PRV operating conditions. For this purpose, a large database with high resolution recordings at 38 PRVs from Dublin and Seville was compiled. Statistical metrics representing centrality and dispersion of these samples were than quantified and used as predictor variables in subsequent regression analysis to predict the global efficiency of the HPER plants equipped with PATs. The results of the regression analysis showed the global efficiency of these plants could be predicted with an accuracy of around 0.89 (expressed with R2 adjusted), using only the statistical parameters of the recorded PRV operating conditions. Another aim of this part of the thesis was to define the minimal average operating conditions at PRV sites so their upgrade to PAT based HPER plants is expected to have a positive NPV after 10 years.
The high resolution recordings from 38 PRV sites mentioned previously were also used in the third part of this thesis to investigate the ratio between their average operating flow and head conditions and the BEP flow and head of their theoretically optimal PATs. The aim of this part of the thesis was to investigate if this ratio could be generalised regardless of dispersion of the operating conditions of a considered site, its size and optimal operating limits of the selected PAT.
The previously described parts of this thesis present different solutions on how to assess HPER potential of PRV site within a WSN when the information about its operating conditions are available. However, it was found that in many cases these are not available. Thus, the forth part of this thesis conducts a spatial regression analysis to investigate if the potential of PRV sites could be estimated with reasonable accuracy using only population and topography data in their proximity.
The fifth and the final part of this thesis addresses the problem of assessment of HPER potential within WSNs on a multi-country scale. For this purpose, a large database of several thousands potential sites across Ireland and the UK was complied. The methodology developed in this part of the thesis utilizes the findings from the second part of the thesis to determine the economically viable sites and the prediction model derived in the same part to assess their global efficiency and thus their potential. To extrapolate the potential to areas without the data the methodology defines a coefficient, which designates the average power potential per 1000 people. Finally, the extrapolated values of the total potential in the investigated countries are compared to the total energy consumption of their drinking water sectors, thus evaluating the percentage of the possible energy savings using HPER concept.
Sponsor
Grant Number
European Regional Development Fund (ERDF) through the Interreg Atlantic Area Programme 2014-2020
Author's Homepage:
https://tcdlocalportal.tcd.ie/pls/EnterApex/f?p=800:71:0::::P71_USERNAME:MITROVIDDescription:
APPROVED
Author: Mitrovic, Dorde
Advisor:
McNabola, AonghusPublisher:
Trinity College Dublin. School of Engineering. Disc of Civil Structural & Environmental EngType of material:
ThesisCollections:
Availability:
Full text availableLicences: