|BERTONI FEDERICA||Cycle: XXXII |
Section: Systems and Control
Tutor: PIRODDI LUIGI
Advisor: CASTELLETTI ANDREA FRANCESCO Major Research topic
:Robust Planning of the Water, Energy and Food Nexus in large river systems of fast-developing countriesAbstract:
Water, energy and food are three key resources for human development that are strictly interconnected. Water is essential to both produce food and generate energy, as agriculture by itself is responsible for 70% of global water withdrawals and cooling thermal processes in electricity production account for almost another 22%. Energy is consumed to supply water (e.g., water extraction, treatment and delivery) and produce food (e.g., produce fertilizers, pump water for irrigation). Food is expected to be even more exploited in the near future to produce energy, e.g. sugarcane for biofuels production, according to estimates of land employed to grow biofuel feedstocks rising up to 37 million hectares by 2030. The demand for water, energy and food is forecasted to increase due to population and economic growth, while their availability is likely to be affected by changes in climate. Many of large international river basins worldwide and several fast-developing African countries will eventually suffer more severely from these growing pressures on their natural resources, whose sustainable use will thus be endangered. For example, the population of the Zambezi river basin in Africa is estimated to reach 51 million by 2025 (+27.5% of the population in 2008), leading to a 60% increase in the food needed to feed the region by 2050 and a subsequent growth in energy and water demand. Energy consumption and irrigation water withdrawals are foreseen to rise up to 50% by 2035 and 10% by 2050, respectively. As far as water availability is concerned, climate experts predict some riparian countries of the Zambezi Basin to be water stressed (e.g., Tanzania and Zimbabwe) or to suffer from water scarcity (e.g., Malawi) by 2025. It is thus of the utmost importance to gain a deeper understanding of the Water-Energy-Food (WEF) nexus in complex and transboundary water resources systems in order to develop adaptation strategies to the challenges brought on by population growth and climate change in these fast-growing economies. An integrated water resources management approach has to be employed to analyse the WEF nexus in terms of trade-off between conflicting objectives, e.g. hydropower production vs irrigation water supply, and quantify its impacts on the ecology of the system. This will allow a sustainable management and development of water and land resources employed to produce energy and food.The aim of my PhD research is to develop a Decision-Analytic-Framework (DAF) for large-scale, transboundary water resources systems in order to determine optimal planning and control decisions, producing alternative efficient pathways (i.e., temporal sequence of structural and management actions allowing to explore adaptation options to changing hydrological and socio-economic conditions) for advancing water management strategies. Within this framework, a strategic WEF model at the river basin scale will be implemented to perform a simulation-based optimization for the planning and the control of this integrated system with respect to different objectives, leading to the identification of robust efficient pathways under current and future scenarios. The Zambezi river basin will be used to illustrate the methodological procedure and test the innovate robust DA Framework, as it is an exemplary transboundary case study characterised by a fast growth and thus by an increasing energy and food demand, likely resulting in significant environmental and socio-economic impacts. First, a state-of-the-art review of all the works and projects focused on the Zambezi river basin will be performed, followed by a reconnaissance analysis of the current status of the basin, aimed at identifying the key stakeholders, their management objectives and the principal components of the WEF nexus to be modelled. Then, planned infrastructures (i.e., dams and irrigated agriculture) and future development plans at the basin scale will be investigated, in order to highlight the main planning and management actions that will be evaluated in the pathway design and optimization process. After the implementation and the coupling of the water allocation scheme conceptual model and the energy system model, the WEF nexus dynamics will be analysed at the Zambezi basin scale by carrying out a simulation-based multi-objective optimization of the resulting integrated strategic model under current and future scenarios. The WEF nexus will be quantified in terms of trade-off between conflicting optimization objectives, e.g. hydropower production vs environmental protection vs irrigation water supply, and alternative efficient pathways will be generated. Two novel algorithms will be developed and employed in the optimization process to jointly optimize the infrastructure planning (e.g., new reservoir size) and control (e.g., water reservoir operating policy) decisions: a Multi-Objective Direct Policy Search (DPS) and a Reinforcement Learning (RL) algorithm, which extends the traditional Fitted-Q Iteration (FQI). The classic FQI will also be extended by a third innovative batch-mode RL algorithm called Scenario-based Fitted-Q Iteration (sFQI), which will be used for designing optimal, adaptive policies for controlling water reservoir systems in the Zambezi basin under climate-related uncertainty of future scenarios. Finally, robust adaptation pathways will be designed via robust optimization as the ones performing satisfactorily over a wide range of uncertain future scenarios, generated by co-varying hydro-climatic and socio-economic projections.