Search

In many developing economies, a high percentage of domestic energy demand is for cooking based on fossil and biomass fuels. Their use has serious health consequences affecting almost 3 billion people. Cleaner cooking systems have been promoted in these countries such as solar cooking and smokeless stoves with varying degrees of success. In parallel, solar electrolytic hydrogen systems have been developed and increasingly used during the last 25 years for electricity, heat and automobile fueling applications. This study has developed and tested experimentally in the laboratory a solar hydrogen plant numerical model suitable for small communities, to generate and store cooking fuel. The numerical model was developed in TRNSYS and consists of PV panels supplying a PEM electrolyser of 63.6% measured stack efficiency and hydrogen storage in metal hydride cylinders for household distribution. The model includes novel components for the operation of the PEM electrolyser, its controls and the metal hydride storage, developed based on data of hydrogen generation, stack temperature and energy use from a purpose constructed small-scale experimental rig. The model was validated by a second set of experiments that confirmed the accurate prediction of hydrogen generation and storage rates under direct power supply from PV panels. Based on the validated model, large-scale case studies for communities of 20 houses were developed. The system was sized to generate enough hydrogen to provide for typical domestic cooking demand for three case-studies; Jamaica, Ghana and Indonesia. The daily cooking demands were calculated to be 2.5kWh/day for Ghana, 1.98kWh/day for Jamaica and 2kWh/day for Indonesia using data mining and a specific quantitative survey for Ghana. The suitability of weather data used in the model was evaluated through Finkelstein Schafer statistics based on composite and recent weather data and by comparing simulation results. A difference of 0.9% indicated that the composite data can be confidently used. Simulations results indicate that a direct connection system to the PV plant rather than using a battery is the optimal design option based on increased efficiency and associated costs. They also show that on average 10tonnes of CO2/year/household can be saved by replacing biomass fuel with hydrogen. The potential of total savings in the three case-study countries is shown in the form of novel solar hydrogen potential maps. The results of this study are a contribution towards better understanding the use of hydrogen systems and enhancing their role in renewable energy policy.

The increased penetration of grid-connected roof-top photovoltaic (PV) systems distribution feeders is leading to technical issues to electric utilities, mainly voltage fluctuations and violations to the grid codes limits. Furthermore, the legacy voltage control devices such as transformer's on-load tap changers (OLTC) and switched capacitor banks need to be operated more frequently than usual with PV penetration and could only alleviate slow-moving voltage fluctuations. Ancillary services from Photovoltaic (PV) inverters can alleviate the impact of the rapidly growing PV penetration by increasing distribution system flexibility and addressing voltage regulation issues. However, the required communication infrastructure and smart grid integration challenges limit broad deployments of PV ancillary services. This thesis proposes a cost-effective volt/var (VVC) control scheme of multi-string PV inverters for active voltage regulation and reactive power ancillary services using the existing smart distribution infrastructure to avoid upfront costs of PV ancillary services. The proposed VVC model is developed in MATLAB/Simulink to adapt PV reactive power compensation according to X/R characteristics of the distribution feeder for effective voltage regulation. A volt/var optimisation model is proposed based on particle swarm optimisation to coordinate the reactive power dispatch of PV inverters with the legacy voltage regulation devices to resolve the operational issues of distribution feeders at high PV penetration levels. Transient voltage analysis and quasi-static time series (QSTS) simulation analysis are conducted using Matlab/Simulink and OpenDSS co-simulation environment. A porotype is designed based on the compiled Simulink PLC code and verified in CODESYS IEC 61131-3 standardization tool to address the practical considerations of controlling multi-string PV systems and smart grid integration challenges. The proposed VVC scheme is implemented using remote terminal units (RTU) in a real-world PV system with multi-string inverters for experimental validation. Quasi-static time-series simulation and experimental results demonstrate the validated effectiveness of the proposed control scheme in controlling fast PV fluctuations, resolving voltage issues, voltage flickers, and enabling higher PV penetration. Furthermore, the proposed VVC control algorithm of PV inverters led to operational benefits for utilities by alleviating voltage flickers, reducing feeder losses, and significantly reducing regulator switching operations (up to 37.1% saving in SCB operations and 34.7% saving in OLTC operations), thereby extending their lifespan and potentially deferring the investment in new voltage regulators. The techno-economic assessment showed that the proposed VVC ancillary services achieved significant improvements of the benefits-cost ratio (BCR) and net present value (NPV) economic metrics of the system. The proposed BCR value increase by 13.3% from 1.27 to 1.44 when the proposed VVC was implemented. The respective (NPV) increased from $19,919.93 to $62,117.86 due to the additional benefits achieved from the proposed VVC ancillary services. Additionally, the payback period was reduced to 8 years using the proposed VVC compared to 12 years with the existing P/Pn method.