Despite Russia’s significant climatic potential and growing demand for sustainable energy, solar water heating (SWH) development faces climatic variability, economic limitations, and regional prioritization challenges. This study proposes a multi-criteria decision-making (MCDM) framework based on the TOPSIS method with nine weighting strategies, including objective, subjective, and hybrid approaches. Eight Russian regions were evaluated using TSOL software considering technical, economic, environmental, and social criteria. Simulated and real data were normalized and analyzed. Vladivostok ranked first in most methods due to high CO2 reduction, low heating cost, favorable payback period, and strong solar contribution, while Nizhny Novgorod ranked last. Variations in Moscow and Yekaterinburg rankings indicated sensitivity to weighting schemes. Cross-method comparison showed relative stability for the highest- and lowest-ranked alternatives, whereas the middle-ranked regions were more sensitive to the selected weighting scheme. Multiple weighting approaches were employed as a robustness mechanism rather than as the sole methodological novelty, allowing the stability and interpretability of regional priorities to be explicitly assessed. Beyond the comparison of weighting methods, the proposed framework provides a robust decision-support tool that integrates technical, environmental, economic, and socio-economic indicators to improve the reliability of regional prioritization for the deployment of SWH systems.
Numerical errors in convective flows are often diffusive and can make computational fluid dynamics (CFD) simulations depart markedly from physical reality. Savonius wind turbine (SWT) simulations are especially affected by transient flow separation and inconsistent numerical methodologies. This paper introduces two sensitivity metrics, the grid sensitivity indicator (GSI), which aggregates the power-coefficient spread over several mesh levels and tip speed ratios, and the combined sensitivity indicator (CSI), which adds the transient torque fluctuation. Both measure numerical sensitivity rather than discretization error, and they are used as objectives to systematically optimize 2D and 3D unsteady Reynolds-averaged Navier–Stokes simulation settings. Mesh, domain, timestep, and initialization settings were optimized on three rotors of very different performance: conventional, three-blade, and B-spline scooplet. Within the tested ranges, the selected settings reduce the GSI to about 2% in 2D and 1%–2% in 3D CFD; key settings include a blade-edge element size of about one tenth of the blade thickness and a rotor zone diameter of 1.8 rotor diameters. A multi-stage initialization with adaptive blending between first- and second-order schemes eliminates the instability associated with constant-timestep initialization at high tip speed ratios and requires about 60% fewer initialization timesteps in 3D. The GSI varies smoothly where the grid convergence index behaves erratically, and a mesh adequate for one rotor proves inadequate for another. The settings, determined for a 1m rotor at a Reynolds number of about 8×105 and compared with experiments for the conventional and three-blade rotors, are recommended starting guidelines for comparable SWT simulations rather than universal rules.
