MS Thesis Defense: Ryan Proulx

ZOOM LINK
Meeting ID: 948 8378 4703
Passcode: 793387

"A Multi-objective Consensus-Building Approach for Energy Systems Planning in Fragile States"

Abstract

Countries such as the Central African Republic, South Sudan, and Somalia face cycles of conflict, political instability, low electrification rates, and the impacts of climate change. Renewable energy-driven electrification is often considered an important way to mitigate these challenges. However, designing and planning green electrification initiatives in fragile states requires making complex planning decisions. We present a consensus-building framework and methodology to guide hybrid renewable energy systems planning in fragile states. Our study develops an iterative approach to Majority Judgment (MJ) for planning energy systems for United Nations (UN) compounds and the surrounding local community. The case study structure reflects a typical mini-grid demand profile involving an anchor client and a neighboring community.

We gather institutional data from 17 experts in African energy transition initiatives from groups including the African Development Bank, the UN Department of Operational Support, and the International Renewable Energy Agency. We elicit institutional knowledge via three institutional data collection channels: (1) organizations’ priorities when evaluating mini-grid projects, (2) organizations’ ranking of individual objectives, and (3) organizations’ evaluations of a set of Pareto-optimal mini-grid designs.

Leveraging these data, we implement three methods to process the institutional knowledge: (1) fitting utility functions to LLM (Large Language Model)-produced grades based on channel 1 institutional data, (2) generating utility functions based on channel 2 institutional data, and (3) fitting utility functions to stakeholder grades based on channel 3 institutional data. We compute the optimality gaps across all input methods in a 2-dimensional case (examining upfront cost and diesel consumption) and a 4-dimensional case (additionally examining community coverage, and system reliability). We find that low-diesel, high-community-coverage mini-grid designs, with reliability marginally below 100%, provide the best compromise in most cases across all input classes. Reliability proves to be a key lever that can significantly facilitate or hinder consensus-building. Overall, our results demonstrate that the proposed mechanism provides high-quality, near-optimal solutions under a limited-information setting, enabling deep insights into the practical nature of consensus solutions. Ultimately, our approach provides a blueprint backed by real stakeholder inputs and site-specific energy transition datasets with potential to accelerate electricity access in fragile states.

Thesis Committee

• (Chair) Vikrant Vaze
• Steven Peterson
• Geoffrey Parker
• Erin Mayfield
• Victoria Holt (Dickey Center for International Understanding)

Contact

For more information, contact Thayer Registrar at thayer.registrar@dartmouth.edu .