Simulation outputs and result data for: Boosters or Building Renovations? An evaluation of 4th Generation District Heating Strategies

This deposit contains the simulation outputs and intermediate result data accompanying the paper: Casamassima, L. et al. (2026). Boosters or Building Renovations? An evaluation of 4th Generation District Heating Strategies. Published on: Energy Conversion and Management: X (under review). The companion source code is available on GitHub: https://github.com/Tyler-lc/4th_gen_dh — release casamassima2026-ecmx. https://github.com/Tyler-lc/4th_gen_dh/releases/tag/casamassima2026-ecmx Together this deposit and that release constitute the full paper-citable artifact. Contents A 5.7 GB compressed archive (~20 GB extracted) containing the gitignored output trees of the analysis pipeline that cannot ship with the source repository due to size: - `building_analysis/results/` — per-building hourly energy demand outputs, scenario aggregate parquets (high-temperature DH, low-temperature DH with renovation, booster heat pump), and the booster sensitivity sub-tree across ten grid temperatures (25–70 °C). - `building_analysis/dhw_profiles/` — per-building domestic hot water demand profiles. - `sensitivity_analysis/` — sensitivity-analysis outputs across electricity/gas price, investment-cost multipliers, discount rate, and renovation-cost reduction factors. - `plots/` — publication figure assets and per-figure data exports. - `grid_calculation/cache/` — Pyomo/Gurobi district-heating network optimisation cache. 54,802 entries total. Compressed: 5.7 GB. Extracted: ~20 GB. ## How to use ```bash git clone https://github.com/Tyler-lc/4th_gen_dh cd 4th_gen_dh git checkout casamassima2026-ecmx conda env create -f environment.yml conda activate dh_sim pip install -e . python scripts/restore_paper_data.py # downloads this archive, verifies SHA256, extracts pytest -m regression # 29 regression tests pass against paper-aligned baseline Full step-by-step instructions in docs/PAPER_DATA_ARCHIVE.md of the release. Reproducibility note The data in this archive was produced by the codebase as it existed at git tag paper-submission-v1. The post-submission paper-ecmx-2026 snapshot contains a hardened deterministic-seeding refactor that draws a different sample of the same stochastic process. Re-running the pipeline from the snapshot code reproduces the methodology but not the bit-for-bit numbers. Directional drift between the two regimes is documented in docs/baseline_comparison_2026-04-20.md on the snapshot branch. To reproduce the paper's exact numerical results, use this archive together with the snapshot's git-tracked outputs. Integrity verification SHA256 of 4th_gen_dh_paper_data_2026-04-29.tar.gz: 18d00e3431460e08ba068ef47d9266abe6d8ec051915d519a18bdf2f0a78f01e Verify after download: shasum -a 256 4th_gen_dh_paper_data_2026-04-29.tar.gz本数据集包含伴随以下论文的仿真输出与中间结果数据： Casamassima, L. 等人(2026)《助推器还是建筑改造？第四代区域供热策略评估》 发表于《Energy Conversion and Management: X》（审稿中）。 附属配套源代码已发布于GitHub：https://github.com/Tyler-lc/4th_gen_dh —— 版本标签为casamassima2026-ecmx，对应发布链接：https://github.com/Tyler-lc/4th_gen_dh/releases/tag/casamassima2026-ecmx。 本数据集与该代码发布包共同构成可被论文引用的完整附属科研成果。 ### 数据集内容 该压缩归档文件大小为5.7 GB（解压后约20 GB），包含因体积过大无法随源代码仓库一同发布、被git忽略的分析流水线输出目录树： - `building_analysis/results/`：单建筑逐小时能源需求输出、场景聚合Parquet文件（高温区域供热、带改造的低温区域供热、助推式热泵），以及覆盖10种电网温度（25–70 °C）的助推器敏感性分析子目录。 - `building_analysis/dhw_profiles/`：单建筑生活热水需求曲线文件。 - `sensitivity_analysis/`：覆盖电价/气价、投资成本乘数、贴现率及改造成本降低系数的敏感性分析输出文件。 - `plots/`：期刊用图表素材及单图数据导出文件。 - `grid_calculation/cache/`：Pyomo/Gurobi区域供热网络优化缓存文件。 本数据集总计包含54,802条数据项。压缩后大小：5.7 GB，解压后约20 GB。 ### 使用方法 bash git clone https://github.com/Tyler-lc/4th_gen_dh cd 4th_gen_dh git checkout casamassima2026-ecmx conda env create -f environment.yml conda activate dh_sim pip install -e . python scripts/restore_paper_data.py # 下载本归档文件，验证SHA256哈希并完成解压 pytest -m regression # 针对与论文对齐的基线开展29项回归测试，所有测试均通过 详细分步操作指南请参见发布包内docs/PAPER_DATA_ARCHIVE.md文件。 ### 可复现性说明 本归档文件内的数据由git标签paper-submission-v1对应的代码库生成。投稿后发布的paper-ecmx-2026快照版本包含经过强化的确定性种子重构，其采用的随机过程采样集与原版本不同。基于该快照代码重新运行分析流水线，可复现研究方法，但无法得到完全一致的逐位数值结果。两种版本间的方向偏差已在快照分支的docs/baseline_comparison_2026-04-20.md文件中记录。 若需复现论文中的精确数值结果，请将本数据集与该快照版本的git跟踪输出文件配合使用。 ### 完整性验证 4th_gen_dh_paper_data_2026-04-29.tar.gz的SHA256哈希值： 18d00e3431460e08ba068ef47d9266abe6d8ec051915d519a18bdf2f0a78f01e 下载完成后可通过以下命令验证哈希完整性： shasum -a 256 4th_gen_dh_paper_data_2026-04-29.tar.gz