Problem or Challenge
In district energy systems, economic risk, regulatory uncertainty, and technology lock-in all weigh on decision makers' choices. Accordingly, model parameters are coming under scrutiny for their inherent uncertainty. More recently, optimisation methods which account for uncertainty have been proposed. However, the quantification, propagation, and management of uncertainty in model parameters continues to pose challenges.
Furthermore, current models are fundamentally ill-suited to project future evolutions of energy demand. In the context of economic volatilities, organizational restructuring, and environmental vulnerabilities, we are likely to see shifts in historic trends of energy consumption. At the moment we have no mechanism to efficiently assimilate these shifts.
MRes/PhD project objectives
This project will investigate the role of digital technologies and new data sources to quantify uncertainties in district energy system optimization. It will explore hybrid-models that combine statistical and numerical modelling for improved management and reduction of uncertainties in model outcomes. Specific emphasis will be on communication of uncertainties for effective decision-making.
PhD project description
The PhD will extend a recent model of district energy optimization in the following aspects: (a) representation of additional components and technologies (for eg. EV charging within the district energy network), (b) develop a systematic methodology to quantify uncertainties in model inputs, especially those pertaining to resilience of the system. This will be carried out by exploiting digital technologies (c) propagation, management, and communication of uncertainties through novel combinations of statistical and numerical modelling.
MRes component
- A thorough literature review of the state-of-the-art in stochastic optimization
- The component (a) listed above, along with identification of relevant digital technologies.
PhD - Expected Outcomes, Contributions to Knowledge & Practice
- A tool for stochastic energy system optimization
- Data assimilation across the system lifetime to assess changes in system operation
- Efficient display and communication of model outputs through appropriate computational platforms.