Robust Multi-Objective Sizing and Operation of an Autonomous Bifacial Photovoltaic Water-Supply Pumping System

Photovoltaic (PV) water-supply pumping is an established and cost-effective solution for autonomous water supply in rural areas. However, recent studies typically treat the borehole water level as a fixed boundary condition, evaluate designs against a typical meteorological year, and rarely account for component degradation or extreme-year weather stress over the project lifetime. This study proposes a multi-objective sizing and operation framework for an autonomous bifacial photovoltaic (bPV) water-supply pumping system, coupling dynamic borehole hydraulics, battery and water storage, and component degradation within a consistent evaluation chain from optimization through life-cycle post-verification. The non-dominated sorting genetic algorithm III (NSGA-III) drives the optimization loop, typical and extreme meteorological-year scenarios screen water-service reliability, and, following life-cycle post-verification, the VIseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) procedure with hybrid analytic hierarchy process (AHP)–entropy weights and Monte Carlo testing of the criterion weights selects the final compromise design. In the reference workflow, the NSGA-III run comprised 6480 objective-function evaluations; subsequent typical-year, extreme-year, and life-cycle screening retained 100 feasible designs; the selected configuration achieved a levelized cost of water of 0.0675 USD/m3, a specific energy consumption of 0.238 kWh/m3, and an aggregate life-cycle water-shortage percentage of 0.204%. The results show that dynamic-head representation and life-cycle screening materially affect the selected capacity–cost–reliability trade-off and the post-verified feasible set, providing a reproducible basis for autonomous rural water-supply design.