DATA-DRIVEN STRATEGIC DECISION-MAKING IN THE ENERGY INDUSTRY: MULTI-LEVEL MULTI-PERIOD NATURAL GAS NETWORK AS A CASE STUDY IN QATAR

Abstract

Natural gas is the fastest-growing fossil-based energy source due to its economic and environmental characteristics. However, the natural gas trade demonstrated multiple shocks in the last five years, including the emergence of new producers and disruptions caused by the COVID-19 pandemic and the Russian sanctions. The latter contributed to changing the global natural gas trade and restructuring the energy mix in the European continent. As Qatar is a leading LNG exporter, new market opportunities have emerged for LNG sales in Asia and Europe. Simultaneously, the latest shocks signified interest in utilising renewable and cleaner energy resources, such as hydrogen and ammonia, to diversify the energy mix and support decarbonisation efforts. Hence, methanol and ammonia have grown as cleaner energy fuels, also known as hydrogen carriers, to replace conventional fossil fuels. In this regard, this work investigates future market opportunities to optimise a flexible multi-period and multi-level natural gas production network, which is responsive to market uncertainties. The novelty lies in using real market data and technically verified operational limits as primary inputs for optimal decision-making. First, the analysis primarily focuses on evaluating the impact of the latest market restructures on the Qatari LNG exports under deterministic production capacities and varied selling strategies. Second, the forecasted prices were then used for strategic natural gas network planning based on a single resource input (i.e., natural gas) and multi-product direct outputs (i.e., LNG, methanol, synthetic fuels, ammonia) and indirect outputs (i.e., urea and MTBE). Two emerging quantitative decision-making tools, agent-based modelling and mathematical programming, were used to optimise the allocation under fixed input data and various investigated case studies. Within this, operational flexibility represented by flexible production capacity was evaluated technically. The parameters were used as inputs to the decision-making tool to allow flexibility in allocating varied natural gas feed streams to the different processes throughout the planning horizon with the objective of maximising the net annual profit of the natural gas production network. Finally, the outputs of the optimisation tools were compared together. The analysis proved that agent-based modelling is more convenient for dynamic environments where new data emerge with time. Mathematical programming can aid decision-makers in making robust decisions at varied deterministic market conditions.

Date of Award	2024
Original language	American English
Awarding Institution	HBKU College of Science and Engineering