Off-shore Solar Energy Generation System in the Energy Community of Providencia (Colombia) - Case Study

Abstract

This study evaluates the technical, reliability, environmental-operational, and financial feasibility of an offshorefloating photovoltaic generation system for an energy community on Providencia Island, Colombia. The proposed frameworkintegrates Monte Carlo simulation, Latin Hypercube Sampling, satellite-derived hourly irradiance and temperature data,cycle-based battery degradation, marine operating losses, local-grid-supported dispatch, diesel-equivalent accounting, andfinancial-support sensitivity. The analysis distinguishes between fully isolated photovoltaic–battery operation and a more realistic local-grid-supported pathway in which the existing island electricity infrastructure is represented as a diesel-equivalent backup source. The off-grid sizing analysis identifies a 420 kWp photovoltaic system with 400 kWh of battery storage as the lowest-costconfiguration satisfying the selected mean Loss of Load Probability threshold of 5%. However, this configuration doesnot provide uninterrupted autonomous supply: approximately 2.67% of annual demand remains unserved on average, andthe 95th-percentile LOLP reaches 10.87%. It also presents a high energy cost and a negligible probability of positive NetPresent Value. A local-grid-supported configuration is therefore evaluated as a transitional pathway in which the photovoltaic–battery system remains the project asset, while the existing island grid supplies residual demand deficits. No electricity sales tothe local grid are assumed in the base case, and surplus photovoltaic energy that cannot be consumed directly or stored iscurtailed. The recommended configuration consists of 240 kWp of photovoltaic capacity and 100 kWh of battery storage.It reduces the battery requirement by 75% relative to the off-grid reference while achieving a mean LOLP of 1.51%, arenewable fraction of 61.11%, and a diesel-equivalent grid-dependency fraction of 38.89%. NASA POWER irradiance data were additionally compared with ERA5 reanalysis data for 35,064 common hourly records over 2020–2023. The long-term mean irradiance differed by only 0.85%, supporting the use of NASA POWER as a reasonable long-term solar-resource proxy in the absence of on-site measurements. Under the realistic no-export assumption, the recommended configuration reaches a positive mean NPV only when CAPEX co-financing is close to 40%, with a probability of positive NPV of approximately 55.6%. The results indicate that offshore photovoltaic generation in Providencia is technically relevant as a diesel-reduction strategy, but its implementation should be interpreted as a policysupported transitional pathway rather than as an immediately autonomous and financially robust system.