Sensitivity Analyses of Synthetic Power Grid Modeling Techniques in a Global Context: A Case Study in Ghana

Abstract

Synthetic power grid models play a pivotal role in algorithmic benchmarking and techno-economic policy analysis. Existing synthetic grid research has primarily focused on regions in the United States and Europe using methods based on these regions' characteristics. This paper examines these methods' suitability for creating representative models for other regions. Differing data availability, power consumption behaviors, and applicability of various modeling assumptions challenge the suitability of existing synthetic grid methods for non-Western countries. Our analysis focuses on the West African country of Ghana. We evaluate methods for estimating electric demand and transmission network topologies by benchmarking them against a representation of Ghana developed in our previous work that is based on an accurate network topology and public reports. Our results indicate that existing population-based demand assumptions may be inapplicable. Transmission topology methods can yield reasonable results when aggregate characteristics match those of the real system, but they do not capture the centralization of Ghana's grid.