Data and code for stochastic planning of building electricity-hydrogen-thermal integrated energy systems across multiple climate zones
Description
This dataset supports the article “Stochastic planning of building electricity-hydrogen-thermal integrated energy systems across multiple climate zones using conditional diffusion scenarios and an enhanced moss growth optimizer”. The dataset contains processed building electricity and useful-heat loads, photovoltaic and meteorological inputs, grid cost and emission scenarios, outage inputs, conditional-diffusion training artifacts, generated and reduced uncertainty scenarios, stochastic planning results, held-out reliability results, optimizer benchmarks, technology and grid future results, source tables, quality-assurance records, analysis code, and editable conceptual diagrams. Five climate-zone cases are represented by Miami, Phoenix, San Francisco, Baltimore, and Minneapolis. Public input coverage includes the 2012 planning year, the independent 2018 held-out year, and Cambium grid futures for 2030, 2040, and 2050. The public data sources include NREL ComStock End-Use Load Profiles, NSRDB PSM v4, NREL Cambium 2024, OEDI/USECPO outage data, EIA Form 861, NLR reliability parameters, and the BLS CPI-U series. Raw third-party files are not redistributed. Source URLs, release identifiers, file sizes, and SHA-256 hashes are retained in the download manifests so that the public inputs can be reconstructed from their original providers. Fixed random seeds, environment specifications, intermediate model artifacts, machine-readable source tables, and deterministic quality-assurance checks support independent inspection and reproduction. The dataset contains no personal, confidential, or restricted information.