Data and code for "Weather-year choice outweighs algorithm choice in sizing standalone hybrid renewable energy systems"

Input data, generated results and code for the revised paper "Weather-year choice outweighs algorithm choice in sizing standalone hybrid renewable energy systems" (Results in Engineering, manuscript RINENG-D-26-21837). Version 3 accompanies the revision. It holds every file of the version-2 archive, now split over twelve archives; extracting all twelve into one folder gives the version-2 tree. Ten files differ from version 2 only in reworded status messages and operating notes (REPRODUCE.md, seven scripts, and the matching lines of verify_ist.txt and tables_rev/REVISION_RESULTS.md); data, computations and results are unchanged. Version 3 adds tables_rev_utc/, the revision tables of the as-submitted (UTC-clock) study, which paper_numbers.py reads for the corrections quoted in the response letter. Version 2 replaced version 1: demand shapes are laid on Indian Standard Time, every three-variable reference design carries a branch-and-bound certificate, and the scenario, population and five-variable re-search studies of the revision are included. Archives:hres_v3_code.zip : the Python scripts (simulation, dispatch, economics, the seven metaheuristics and random search, reference construction, certification and analysis; the three synthetic demand shapes are defined in hres_bench.py), manifest.json with every parameter and library version, README_IST.md and REPRODUCE.md, the run logs, main_runs_ist.csv.gz (the run-level record of the 164,160 main-benchmark runs), and the derived tables, figures and extra exports (tables/, tables_rev/, tables_rev_utc/, figures/, figures_rev/, export_extra/).hres_v3_weather.zip : raw/, the 228 cached hourly PVGIS-ERA5 series (PVGIS 5.3, horizontal array, 14 per cent system loss) for 12 Indian sites and the years 2005 to 2023, with their request metadata.hres_v3_weather_tilt.zip : raw_tilt/, the 228 series for a south-facing array at the PVGIS optimum inclination of each site, with the tilt angles.hres_v3_results.zip : results/, the reference designs and run records of the main benchmark (684 site-year blocks), the budget-curve, convergence and five-variable studies; and the cert, bench, population and extcheck folders of results_rev/: branch-and-bound certificates, scenario benchmarks, population study and five-variable re-search.hres_v3_robust_*.zip, eight archives of two scenarios each : results_rev/robust/, the multi-year reliability matrices of every robust-sizing scenario. paper_scripts.zip : the scripts that produce every number, table and figure of the revised paper from the stored outputs, with the numbers ledger; all of their inputs are in hres_v3_code.zip. Seeds are derived deterministically from the block name, the algorithm name and a fixed base seed, so any run can be reproduced from its identifiers. All weather values are reanalysis, not site measurements, and all demand profiles are synthetic.