Supervisor:
Mohsen Soltani

Co-Supervisor:
Ahmad Arab-Kooshar

Assessment Committee:
Associate professor John-Josef Leth, AAU(Chair)
Gorm B. Andersen, Added Values, Denmark
Associate professor Anna Volkova, Tallinn Universoty of Technology, Estonia

Moderator:
Amin Hajizadeh

Abstract:

The global imperative to achieve Sustainable Development Goal 7 (SDG 7: affordable and clean energy) faces significant challenges, particularly in meeting its energy efficiency target. Current data indicates that the average annual improvement rate in energy efficiency programs (EEPs) falls substantially below the rate required to meet the 2030 milestone. This gap presents a critical research opportunity, especially within the urban building sector, which is emerging as a pivotal arena for developing, deploying, and evaluating EEPs due to its significant share of global energy consumption and greenhouse gas emissions. In traditional urban settings, buildings functioned solely as consumers (passive end-users), limiting their capacity to implement EEPs. 

However, electricity market liberalization and technological advances in small-scale distributed renewable energy systems (RESs) and storage solutions, to name a few, are transforming buildings into active prosumers. This shift allows buildings to be modeled as micro multi-carrier integrated energy hubs (MIEHs) capable of consuming, producing, storing, and supplying various energy carriers, offering more significant potential and flexibility for implementing EEPs. Optimizing the operational plan for the energy components within micro MIEHs can unlock a substantial portion of this potential.

It is widely recognized that micro MIEHs outperform single-carrier energy systems in several key areas, including technical efficiency, economic feasibility, environmental sustainability, and system reliability. However, micro MIEHs present significantly greater operational complexity due to multiple factors: their integration with diverse energy infrastructures, the intricate interdependencies between energy carriers, the distinct physical properties of each carrier (including response times and storage capabilities), and more stringent technical, economic, and environmental requirements. 

Furthermore, micro MIEHs face high uncertainty across the supply side, demand side, and system components due to the intricate energy coupling between multiple carriers—such as electricity, gas, and heat—and their complex interactions and interdependencies, the intermittency of RESs like wind and solar, dynamic regulatory frameworks aimed at promoting green transition, and evolving market mechanisms for emerging technologies such as storage solutions and multi-carrier demand-side management programs (DSMPs). Together, these factors make the operation of micro MIEHs increasingly complex and challenging. Therefore, a holistic optimization framework is essential for operating micro MIEHs, which simultaneously addresses technical efficiency, economic viability, and environmental sustainability while managing inherent system uncertainties through stochastic modeling.

To answer these challenges, this project develops a holistic bi-level stochastic optimization framework whose fundamental elements are techno-economic and environmental assessments at the lower level and risk management at the upper level to scrutinize many deep uncertainties in operating micro MIEHs for urban building applications. Driven by the imperative of SDG7, this framework aims to provide sustainable, efficient, and resilient energy solutions for diverse building types within urban settings. The optimal operation of the micro MIEHs is implemented at the lower-level optimization problem, serving as a computational core that drives and informs the upper-level optimization process.

In the lower level, the energy hub (EH) tool models the cluster of energy components across various building types as a micro MIEH. This model incorporates a range of energy carriers, including electricity, natural gas, solar radiation, heating, cooling, and oxygen, on both the input and output sides of the cluster. Multi-carrier energy converters and storage solutions are also integrated at the cluster core for efficient energy management. The operator (decision-maker) aims to determine the optimal configuration of this cluster over each time resolution of the operational horizon to generate, convert, transfer, and distribute multiple energy carriers to fulfill the multi-carrier energy demands.

At the lower-level optimization, the decision-maker looks at various problem objective functions across different case studies, including minimizing energy costs, emission costs, the cost of energy not supplied, and customer inconvenience costs, all while satisfying technical and logical constraints. Furthermore, multi-carrier demand-side management programs are implemented at this level to evaluate the micro MIEH’s operational flexibility comprehensively. This level incorporates various components’ nonlinear characteristics, enhancing the operational problem’s applicability and practicality for real-world applications. 

Uncertainty handling is implemented at the upper-level optimization problem that feeds the lower-level optimization process. At the upper level, the decision-maker simultaneously optimizes the horizon of many uncertain input parameters (UIPs), considering their interactions while applying information-gap decision theory (IGDT) to evaluate their adverse and beneficial impacts on the operation of the micro MIEH using robustness and opportunity functions, respectively. Depending on the building type, UIPs are electric, heating, and cooling power demands, oxygen demand, the price of electricity and gas, and the production capacity of the photovoltaic system. The boundaries set by minimum and maximum limits for the horizon of all UIPs, along with the entirety of the lower-level optimization problem, serve as constraints for the upper-level optimization problem.

The developed framework is a complex, practical, and wide-ranging optimization problem characterized by its non-convexity, nonlinearity, and mixed-integer structure. This problem is solved through an integrated solving procedure combining MATLAB and GAMS, with Excel as an interface between the platforms. The CPLEX and DICOPT solvers handle the linear and nonlinear versions of the lower-level optimization problems, respectively. Non-dominated sorting genetic algorithms II and III solve the upper-level multi- and many-objective optimization problems and generate a multi- and many-dimensional Pareto-efficient solution set, respectively.

A combination of the fuzzy satisfying method and either a distance metric methodology or conservative methodology (min-max formulation) is utilized to select the optimal solution from this Pareto set, providing the decision-maker with a well-informed choice from a spectrum of optimized solutions. Different real-world building types, including industrial and hospital settings, are used as case studies to evaluate the effectiveness of the developed framework in managing deep UIPs in operating micro MIEHs.