2026 Theses Doctoral

Data-Driven Rural Communities Electrification Planning in Sub-Saharan Africa

Wu, Yuezi

Electricity access remains one of the central development challenges in Sub-Saharan Africa, where large rural populations still lack reliable electricity services. Solar mini-grids have emerged as an important electrification approach in remote rural communities. However, planning economically viable systems remains difficult due to low and uncertain electricity demand, heterogeneous settlement structures, and limited information during project development. These conditions introduce substantial uncertainty into engineering decisions related to generation sizing and distribution network design.

From the load profile perspective, the analysis identifies two key aspects that strongly influence solar mini-grid generation economics: the peak-day to average-day load ratio, where higher ratios increase generation capacity requirements, and the daytime load fraction, where greater alignment between demand and solar availability reduces costs. Based on these aspects, three strategies are examined to improve system performance. Load integration, by aggregating a larger number of households and productive loads, improves load diversity and reduces generation costs. Load shedding, where a small fraction of electricity demand is curtailed during rare peak events, can substantially reduce generation capacity requirements and system costs. Load flexibility, particularly from productive loads operating during daytime hours, improves alignment between demand and solar generation and further enhances system efficiency.

Spatial settlement structure is a critical determinant of electrification feasibility. Nucleated settlements reduce per-customer connection costs compared with dispersed settlement patterns. Network design decisions, including spatial outlier treatment and pole-based connection layouts, further influence distribution cost estimation. System boundary definition also plays a key role: as system scale expands to include more households, generation benefits from improved load integration, but distribution networks must extend farther, creating a trade-off between generation scale economies and distribution expansion.

Using spatial data from Northern Uganda, the analysis shows that productive loads are significantly more likely to occur in larger and more nucleated communities. Communities with similar settlement characteristics also tend to cluster spatially. Communities located farther from the existing grid are generally smaller and more dispersed, while a small subset of remote productive-load–hosting settlements remain large and nucleated and therefore represent strong candidates for mini-grid development.

Overall, this dissertation develops analytical tools for solar mini-grid sizing with integrated load management strategies and for distribution system estimation. It also provides empirical insights that support more robust engineering planning for rural electrification under conditions of uncertainty and resource constraints. The findings offer guidance for planners and developers seeking economically viable electrification strategies in underserved regions.