Multi-stage optimization of rural electrification planning at regional level considering multiple criteria. Case study in Nigeria

•A MILP model is presented to design a rural electrification plan in a region.•Electrification actions are organized in temporal stages to increase satisfaction.•On-grid and off-grid microgrids with different generation sources are considered.•The model is validated and applied to a case study in Nigeria.•Solutions obtained show an increase in the total satisfaction of the region. Rural electrification should plan the multi-stage implementation of electrification systems considering different distribution options (national grid extension and decentralized off-grid systems) with different energy sources (photovoltaic, wind, diesel, etc.). Most of the existing planning assessment tools search for a static final design and only consider economical and technical criteria, although social and environmental issues have proven essential in ensuring long-term sustainability. In this context, the objective of this paper is to investigate and define how different actions of electricity extension in a region can be gradually organized to maximize the region’s overall satisfaction, and which technology options for electricity generation and distribution should be promoted on a regional scale according to economic, technical, socio-institutional and environmental criteria. The obtained solution defines, at each planning stage, to which communities should electricity access be extended and which configuration and sources are used in each community in order to maximize the total satisfaction of the region. This total satisfaction is calculated by means of multicriteria indices that evaluate the energy provided and the suitability of the chosen system designs according to stakeholders’ opinions (experts, promoters, users, among others). A mixed-integer linear programming (MILP) model is developed to solve the regional multi-stage planning. The model is validated with a computational experiment of 540 randomly generated instances; the results highlight its efficiency. Finally, it is applied to solve a case study of 544 real non-electrified communities in Plateau State (Nigeria) calculating two electrification plans (with a more and less budget-constrained scenario) which are compared with the currently available one. The results show that the proposed new plans significantly increase the total satisfaction obtained in the electrification of the region. The consideration of a less-constrained budget scenario allows the promotion of on-grid MGs based on only photovoltaic wit