Universal access to electricity remains a critical challenge in rural Sub-Saharan Africa, where mini-grids present a viable alternative to national grid expansion. Effective off-grid energy planning typically relies on either Geospatial Electrification Models (GEMs), such as OnSSET, which offer broad spatial analyses, or Energy System Optimization Models (ESOMs), like MicroGridsPy, which provide high-resolution techno-economic optimization. Despite their inherent complementarities, integrating these models poses significant challenges, particularly regarding the consistent characterization of high-resolution electrical demand. This thesis proposes a synergistic framework to soft-link OnSSET and MicroGridsPy, using Kenya as a comprehensive case study. To bridge the gap in temporal granularity, high-resolution stochastic demand archetypes that account for wealth, latitude and specific cooling needs are integrated into OnSSET to enhance its demand definition and hybrid system modeling. Subsequently, 39 optimal mini-grid sites initially identified by the geospatial tool are individually optimized using MicroGridsPy. A calibrated Gaussian distribution of households is applied within the ESOM to ensure load coherence between the two software environments without sacrificing its settlement-level resolution. The comparative analysis reveals that OnSSET systematically overestimates the Levelized Cost of Energy (LCOE) for mini-grids compared to the dedicated optimization tool. MicroGridsPy generally prioritizes capital-intensive renewable technologies, such as photovoltaics and battery storage, to achieve a lower long-term LCOE. However, specific operational circumstances, such as intermittent seasonal cooling demands, can shift the optimal configuration back toward diesel generation to prevent battery underutilization. Crucially, the implementation of demand archetypes into OnSSET successfully reduces the divergence in system sizing and cost estimations between the two models. Ultimately, the sequential application of these tools leverages the spatial scale of GEMs alongside the technological rigor of ESOMs, offering a more robust and coherent methodology for rural electrification planning.
L’accesso universale all’elettricità è una sfida cruciale nell'Africa Subsahariana rurale, dove i sistemi decentralizzati offrono una valida alternativa all'espansione della rete nazionale. La pianificazione energetica si affida tipicamente a modelli geospaziali (GEMs) come OnSSET, per l'ampia analisi spaziale, o a modelli di ottimizzazione (ESOMs) come MicroGridsPy, per l'ottimizzazione tecno-economica ad alta risoluzione. Nonostante le complementarità, l'uso integrato di questi strumenti pone sfide significative nella definizione coerente della domanda elettrica. Questo lavoro propone un framework sinergico applicato al caso studio del Kenya. Per colmare il divario nella risoluzione temporale, in OnSSET sono stati integrati archetipi di carico ad alta risoluzione basati su fasce di reddito, latitudine ed esigenze di raffrescamento. Successivamente, 39 siti ottimali per mini-grid individuati dal modello geospaziale sono stati ottimizzati singolarmente con MicroGridsPy. Nell'ESOM è stata applicata una distribuzione Gaussiana delle famiglie, calibrata per garantire la coerenza del carico tra i due strumenti senza perdere la risoluzione a livello di singolo insediamento. L’analisi comparativa rivela che OnSSET sovrastima sistematicamente l'LCOE delle mini-grid rispetto allo strumento dedicato. MicroGridsPy predilige generalmente tecnologie rinnovabili ad alto costo iniziale e sistemi di accumulo per ottenere un LCOE inferiore nel lungo termine. Tuttavia, carichi stagionali intermittenti possono spostare la configurazione ottimale verso la generazione diesel per prevenire il sottoutilizzo delle batterie. L’introduzione degli archetipi in OnSSET riduce con successo la divergenza di dimensionamento e di costo tra i modelli. L'approccio sequenziale sfrutta così la scala spaziale dei GEMs e il rigore tecnologico degli ESOMs, offrendo una metodologia robusta per la pianificazione elettrica rurale.
Exploring synergies and complementarities between geospatial electrification modeling and an energy system optimization model for off-grid energy planning in Kenya
TOMASSO, VALENTINA;CURIA, EDOARDO
2025/2026
Abstract
Universal access to electricity remains a critical challenge in rural Sub-Saharan Africa, where mini-grids present a viable alternative to national grid expansion. Effective off-grid energy planning typically relies on either Geospatial Electrification Models (GEMs), such as OnSSET, which offer broad spatial analyses, or Energy System Optimization Models (ESOMs), like MicroGridsPy, which provide high-resolution techno-economic optimization. Despite their inherent complementarities, integrating these models poses significant challenges, particularly regarding the consistent characterization of high-resolution electrical demand. This thesis proposes a synergistic framework to soft-link OnSSET and MicroGridsPy, using Kenya as a comprehensive case study. To bridge the gap in temporal granularity, high-resolution stochastic demand archetypes that account for wealth, latitude and specific cooling needs are integrated into OnSSET to enhance its demand definition and hybrid system modeling. Subsequently, 39 optimal mini-grid sites initially identified by the geospatial tool are individually optimized using MicroGridsPy. A calibrated Gaussian distribution of households is applied within the ESOM to ensure load coherence between the two software environments without sacrificing its settlement-level resolution. The comparative analysis reveals that OnSSET systematically overestimates the Levelized Cost of Energy (LCOE) for mini-grids compared to the dedicated optimization tool. MicroGridsPy generally prioritizes capital-intensive renewable technologies, such as photovoltaics and battery storage, to achieve a lower long-term LCOE. However, specific operational circumstances, such as intermittent seasonal cooling demands, can shift the optimal configuration back toward diesel generation to prevent battery underutilization. Crucially, the implementation of demand archetypes into OnSSET successfully reduces the divergence in system sizing and cost estimations between the two models. Ultimately, the sequential application of these tools leverages the spatial scale of GEMs alongside the technological rigor of ESOMs, offering a more robust and coherent methodology for rural electrification planning.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_07_Tomasso_Curia.pdf
accessibile in internet per tutti
Descrizione: Tesi
Dimensione
5.56 MB
Formato
Adobe PDF
|
5.56 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/260657