Date of Award

2015

Degree Type

Thesis

Degree Name

Doctor of Philosophy

School

School of Engineering

Faculty

Faculty of Health, Engineering and Science

First Advisor

Dr Yasir Al-Abdeli

Second Advisor

Dr Ganesh Kothapalli

Abstract

The world’s increasing energy demand means the rate at which fossil fuels are consumed has increased resulting in greater carbon dioxide emissions. For many small (marginalised) or coastal communities, access to potable water is limited alongside good availability of renewable energy sources (solar or wind). One solution is to utilise small-scale renewably powered stand-alone energy systems to help supply power for everyday utilities and to operate desalination systems serving potable water (drinking) needs reducing diesel generator dependence. In such systems, on-site water production is essential so as to service electrolysis for hydrogen generation for Proton Exchange Membrane (PEM) fuel cells. Whilst small Reverse Osmosis (RO) units may function as a (useful) dump load, it also directly impacts the power management of stand-alone energy systems and affects operational characteristics. However, renewable energy sources are intermittent in nature, thus power generation from renewables may not be adequate to satisfy load demands. Therefore, energy storage and an effective Power Management Strategy (PMS) are vital to ensure system reliability.

This thesis utilises a combination of experiments and modelling to analyse the performance of renewably powered stand-alone energy systems consisting of photovoltaic panels, PEM electrolysers, PEM fuel cells, batteries, metal hydrides and Reverse Osmosis (RO) under various scenarios. Laboratory experiments have been done to resolve time-resolved characteristics for these system components and ascertain their impact on system performance. However, the main objective of the study is to ascertain the differences between applying (simplistic) predictive/optimisation techniques compared to intelligent tools in renewable energy systems. This is achieved through applying intelligent tools such as Neural Networks and Particle Swarm Optimisation for different aspects that govern system design and operation as well as solar irradiance prediction.

Results indicate the importance of device level transients, temporal resolution of available solar irradiance and type of external load profile (static or time-varying) as system performance is affected differently. In this regard, minute resolved simulations are utilised to account for all component transients including predicting the key input to the system, namely available solar resource which can be affected by various climatic conditions such as rainfall. System behaviour is (generally) more accurately predicted utilising Neural Network solar irradiance prediction compared to the ASHRAE clear sky model when benchmarked against measured irradiance data. Allowing Particle Swarm Optimisation (PSO) to further adjust specific control set-points within the systems PMS results in improvements in system operational characteristics compared to using simplistic rule-based design methods. In such systems, increasing energy storage capacities generally allow for more renewable energy penetration yet only affect the operational characteristics up to a threshold capacity. Additionally, simultaneously optimising system size and PMS to satisfy a multi-objective function, consisting of total Net Present Cost and CO2 emissions, yielded lower costs and carbon emissions compared to HOMER, a widely adopted sizing software tool. Further development of this thesis will allow further improvements in the development of renewably powered energy systems providing clean, reliable, cost-effective energy. All simulations are performed on a desktop PC having an Intel i3 processor using either MATLAB/Simulink or HOMER.

Share

 
COinS