In many developing countries, access to electricity remains a significant challenge. Electrification planners in these countries often have to make important decisions on the mode of electrification and the planning of electrical networks for those without access, while under resource constraints. An integrated approach to electrification planning in which traditional grid electrification is complemented off-the-grid technologies such as off-grid microgrids and stand-alone systems can enable the economic provision of electricity access in these regions. This integrated planning approach can be facilitated by determining the least-cost mode of electrification - i.e by electric grid extension or off-grid systems - for non-electrified consumers in a region under analysis, while considering technical, economic and environmental constraints. Computational clustering methods the identification of consumer clusters (either as clusters of off-grid microgrids, stand-alone systems or grid-extension projects) can be undertaken using computational clustering methods. This paper presents a novel computational approach to achieve this purpose. This methodology involves exploiting the grid network that connects all consumers, by greedily partitioning the network to identify clusters of consumers to be electrified by grid-extension and off-grid microgrid systems. Using test cases and sensitivity analyses, we implement and benchmark this top-down approach with those obtained from a bottom-up clustering methodology used by the Reference Electrification Model, a model obtainable in literature. Results presented show that the alternative top-down methodology proposed can compare favorably, in terms of global electrification costs, with a bottom-up approach to rural electrification planning.

翻译：在许多发展中国家，电力接入仍是一个重大挑战。这些国家的电气化规划者在资源约束下，常常需要就无电地区的电气化模式及电网规划做出重要决策。一种综合性的电气化规划方法——将传统电网延伸与离网技术（如离网微电网和独立系统）相结合——能够在这些区域实现经济可行的电力供应。这种综合规划方法可通过确定特定区域内无电用户的最低成本电气化模式（即电网延伸或离网系统）来实现，同时需考虑技术、经济和环境约束。利用计算聚类方法，可识别用户聚类（无论是离网微电网集群、独立系统集群还是电网延伸项目集群）。本文提出一种新颖的计算方法以实现此目标。该方法通过贪心式分割连接所有用户的电网网络，识别出适合电网延伸和离网微电网系统的用户集群。基于测试案例和敏感性分析，我们实施并比较了这一自上而下的方法与文献中可获取的参考电气化模型（Reference Electrification Model）所采用的基于自下而上聚类方法的基准方案。结果表明，就全局电气化成本而言，所提出的自上而下替代方法能够与农村电气化规划中的自下而上方法相媲美。