Rural electrification in remote areas of developing countries has several challenges which hinder energy access to the population. For instance, the extension of the national grid to provide electricity in these areas is largely not viable. The Kenyan Government has put a target to achieve universal energy access by the year 2020. To realize this objective, the focus of the program is being shifted to establishing off-grid power stations in rural areas. Among rural areas to be electrified is Habaswein, which is a settlement in Kenya’s northeastern region without connection to the national power grid, and where Kenya Power installed a stand-alone hybrid mini-grid. Based on field observations, power generation data analysis, evaluation of the potential energy resources and simulations, this research intends to evaluate the performance of the Habaswein mini-grid and optimize the existing hybrid generation system to enhance its reliability and reduce the operation costs. The result will be a suggestion of how Kenyan rural areas could be sustainably electrified by using renewable energy based off-grid power stations. It will contribute to bridge the current research gap in this area, and it will be a vital tool to researchers, implementers and the policy makers in energy sector.

Microgrids play a pivotal role in the attempts to achieve universal access to electricity; hence, their accurate planning, which also includes the evolution of the system over the project lifetime, is essential for maximizing the profitability of the investments. This paper proposes an iterative multi-year mixed-integer linear programming (MILP) optimization, in which power-dependent efficiency and capacity degradation of storage as functions of the operative strategy are considered. In this study, the computational burden of such a large problem is significantly reduced by means of an external loop that accounts for the hourly storage efficiency and residual capacity and serves to update the parameters for the MILP optimization of the following iteration. This approach is tested on a case study of a community in Soroti, in central-eastern Uganda. Comparison with the output of a one-shot MILP and with a procedure that overlooks the effects of the dispatching strategy on storage characteristics is shown, highlighting the benefits gained with the proposed algorithm in terms of both simulation time and accuracy of the design.