According to the Agenda 2030 launched by the United Nations in 2015, to ensure access to affordable, reliable, sustainable, and modern energy for all is now recognised as a fundamental goal to reach by 2030. Focusing on electrification, to ensure universal access to electricity, it is estimated that 2.6 billion people will have to be electrified by 2030, highlighting the need and the urgency to develop sustainable and appropriate approaches to electricity planning. According to this, this thesis deals with methods, approaches, and models for formulating and designing sustainable long-term electrification plans for rural off-grid areas of the world. In particular, the scientific literature highlights the lack of appropriate modelling frameworks for assessing, projecting, and integrating the electricity demand within the rural energy planning endeavour. It also reveals a weak understanding of the dynamic and multifaceted complexities that involve electricity access and socio-economic development.
To fill these gaps, this thesis sets a novel starting point for the research work on energy demand models and their integration in electrification planning procedures, by setting the following three specific objectives: (1) To investigate and discuss the challenge of electricity demand assessment and modelling for rural electrification. This objective is pursued through the development and analysis of specific case-studies, an extensive synthesis and capitalisation of the related scientific literature, and the characterisation of the main modelling fundamentals of this research field. The relevance of electricity demand in rural electricity planning is introduced, by discussing and demonstrating that unreliable forecasts and projections of short- and long-term electricity demand can negatively impact the techno-economic sizing of off-grid power systems. This implies a raising awareness on the criticality of electricity load assessment in rural electrification planning and advocates more research on this topic. The current methodologies adopted for projecting long-term energy demand along the planning horizon are then evaluated, finding that most of the rural energy planning literature neglects the aspect of long-term evaluation of electricity demand. It is also found that modelling long-term projections of energy demand needs to consider the multifaceted aspects related to it, which have both a technical and a socio-economic nature. This leads to the development of the main important causal loop diagrams that characterise the technical and socio-economic dimensions of the electricity-development nexus, proving that the evolution of rural electricity demand can be explained by endogenous dynamics. This result advocates the promotion of modelling techniques able to frame, understand, discuss, and quantitatively formulate the behaviour of complex systems, such as System Dynamics.
The second specific objective is (2) To assess and model the fundamental dynamics, variables, and exogenous policies that characterise the electricity-development nexus and determine the evolution of electricity demand. The chosen method to achieve this objective is system dynamics. All the steps are based on a real case-study as reference, i.e. a hydroelectric-based electrification programme implemented in the rural community of Ikondo, Tanzania, in 2005 by the Italian NGO named CEFA Onlus. The conceptualisation of the model leads to the analysis of the dynamic problem to solve and the purpose to achieve, the model boundary and key variables, and their behaviour. The formulation phase results in the development of a novel simulation model which simulates the impact of electricity access and use on the socio-economic development experienced in Ikondo, and the related feedback on the community’s electricity consumption. This result provides the first important goal in the research and modelling work committed to develop more general, flexible, and customizable energy demand models. The calibration of the model and the analysis of the uncertainties through the Markov-chain Monte-Carlo (MCMC) contributes to build confidence in the model structure by verifying its ability to replicate the observed historical behaviour of the system, and by uncovering model flaws and hidden dynamics. The calibration also confirms the appropriateness of system dynamics in modelling the complexities behind the evolution of rural electricity demand, and it provides new modelling insights on some presumed dynamics and their impact on the electricity-development nexus. The testing of the model leads to a novel assessment of the most relevant dynamics and it provides a novel discussion on model results when its inputs take on different values, until the extreme ones, and as if the model were tested for different contexts than Ikondo. Policy testing is also performed for exploring model behaviour when subjected to different polices and exogenous decision-making processes. It provides a list of complementary activities to couple with electrification programmes for enhancing their positive impact on rural communities. These results can support the definition of useful guidelines and best practices for rural electrification, and they advocate an updating of the traditional monitoring and evaluation frameworks commonly used for assessing energy access projects.
The last objective is (3) To integrate demand, load, and energy optimisation models in a more comprehensive electricity planning procedure. This is pursued by developing a computational soft-link between the system dynamics model, a stochastic load profiles generator, and a heuristic energy optimisation tool. The result is a more comprehensive modelling framework for investigating electrification processes – if compared with the traditional approaches and hypotheses commonly adopted to assess and integrate electricity demand in rural electricity planning –, and it provides an important contribution towards the employment of the robust multi-year energy optimisation as the referring standard for off-grid electricity planning.