The effects of global warming are severely recognizable and, according to the OECD, 47% of the world’s population will soon live in regions with insufficient drinking water. Already, many countries depend on desalination for fresh water supply, but such facilities are often powered by fossil fuels. This paper presents an energy self-sufficient desalination system that runs entirely on solar power. Sunlight is harvested using parabolic trough collectors with an effective aperture area of 1.5 m × 0.98 m and a theoretical concentration ratio of 150 suns, in which a concentrator photovoltaic thermal (CPV-T) hybrid-absorber converts the radiation to electricity and heat. This co-generated energy runs a multi-effect distillation (MED) plant, whereby the waste heat of multi-junction concentrator solar cells is used in the desalination process. This concept also takes advantage of synergy effects of optical elements (i.e., mirrors), resulting in a cost reduction of solar co-generation compared to the state of the art, while at the same time increasing the overall efficiency to ~75% (consisting of an electrical efficiency of 26.8% with a concurrent thermal efficiency of 48.8%). Key components such as the parabolic trough hybrid absorber were built and characterized by real-world tests. Finally, results of system simulations, including fresh water output depending on different weather conditions, degree of autonomy, required energy storage for off-grid operation etc. are presented. Simulation results revealed that it is possible to desalinate around 2,000,000 L of seawater per year with a 260 m² plant and 75 m³ of thermal storage.

Efficient and accurate grey-box building models, including water-based heating and cooling systems, are crucial for simulating and optimizing the energy demand of building, neighborhood, and network scenarios. However, the numerical effort and the amount of input data required for existing models are still high, and the parameterization of these systems is very labor-intensive. This paper presents a grey-box model that addresses these limitations by requiring minimal input data and offering a highly efficient parameterization method. Using physical principles, the model was validated against a detailed physical building model and measurement data. Our results show that the grey-box model accurately predicts return temperatures (σ = 0.37 K, µ = 0.05 K) and room air temperatures (σ = 0.62 K, µ = 0.28 K). Compared to 8229 s for the detailed physical model, the model requires only 18 s for a one-year simulation. The model also shows robust behavior with alternative weather data and control strategies. The key contribution of this work is the development of a grey-box model that combines high accuracy and numerical efficiency with significantly reduced data and parameterization requirements, with possible applications in large-scale building simulations, demand-side management, short-term energy storage strategies, and model predictive control.