Inzinion/power-plant-climate-exposure-index

Global Power-Plant Climate Exposure Screening Index (CESI)

Per-plant outdoor-environment severity for the world's power fleet: a 0-100 index, a CES1-CESX screening class and the dominant stressor, derived from each plant's own coordinates (WorldClim normals, Koppen-Geiger class, distance to coast). The full model ships as severity_model.py.

Canonical record: doi.org/10.5281/zenodo.22172589 · Publisher: Inzonex

Load

from datasets import load_dataset

ds = load_dataset("Inzinion/power-plant-climate-exposure-index", split="train")
print(ds)

Or straight into pandas:

import pandas as pd

df = pd.read_csv("https://huggingface.co/datasets/Inzinion/power-plant-climate-exposure-index/resolve/main/power_plant_environment_severity.csv")

42,094 power plants · 196 countries · per-plant outdoor-environment severity: 0–100 index, CES1–CESX class, dominant stressor · CC BY 4.0

For every plant, climate stressors are derived from its own coordinates: WorldClim temperature normals (annual mean, summer max, seasonal swing), Köppen–Geiger class, and distance to coast computed from a 0.5° land grid (1,661 plants sit within 15 km of a coastline). These are combined into a composite 0–100 severity index, a screening class on Inzonex's own CESI scale (CES1 very low … CES5 very high, CESX extreme), and the dominant stressor for outdoor hardware: marine salt corrosion, thermal cycling, heat/UV, dust abrasion, or humidity/wetness.

Class distribution: CES1 6,060 · CES2 10,839 · CES3 13,669 · CES4 7,857 · CES5 3,337 · CESX 332.

What CESI is — and is not

CESI is a heuristic screening index, not an ISO 9223 implementation. ISO 9223:2012 classifies atmospheric corrosivity from measured time-of-wetness, SO₂ and airborne-salinity dose–response inputs; none of those are measured here. CESI approximates exposure from open climate data only: a wetness proxy from Köppen class and temperature, a marine-chloride proxy from coast distance (with a coastal-microclimate uplift: humid coasts remain severe even inside arid Köppen zones), plus heat, thermal-cycling, UV and dust terms with fixed weights and thresholds (the full formula is the shipped reference implementation severity_model.py, which build_dataset.py runs to produce every value in this dataset; fields are summarised in data_dictionary.csv). ISO 9223's qualitative category ladder served as a conceptual reference for the class structure only — CESI classes must not be quoted as ISO 9223 categories, and neither is a substitute for site measurement or materials selection.

The scale has had a single-point field plausibility check (a Persian-Gulf coastal site classifies CES5 "very high" with marine salt corrosion as the dominant stressor, consistent with the severe marine corrosivity observed there in Inzonex field work) — that is a sanity check, not a validation study.

Files

Intended use

Screening and ranking context for the durability of outdoor power-plant hardware — HRSGs, expansion joints, valves, flanges and their insulation: which parts of a fleet sit in severe marine or high-cycling environments, and where protection matters most. Useful for fleet-level prioritisation, asset-management context and climate-exposure research; not for site-specific engineering decisions.

Sources and attribution

WRI Global Power Plant Database (plants) · WorldClim (temperature normals) · Köppen–Geiger classification · ISO 9223:2012 (conceptual reference for the category ladder only). CC BY 4.0; see datapackage.json.

Versions

v1.1.0 (2026-08-30): the dataset is now computed directly by the shipped severity_model.py (deterministic and order-independent) — v1.0.0 packaged a precomputed file whose coast-distance values carried an order-dependent 0.5° cell-cache artifact, and its build script only repackaged that file rather than containing the formula. Coast distances are now cached per coordinate, which moves a small number of plants across class thresholds (e.g. CESX 354 → 332) and adds 3 plants (42,091 → 42,094). Negative numeric values (longitudes, latitudes, sub-zero temperatures) are now written as plain numbers — v1.0.0 escaped them with a leading apostrophe as an over-broad spreadsheet-formula guard, which broke numeric parsing. croissant.json upgraded to conformant MLCommons Croissant 1.0.

Cite

See CITATION.cff. Published by Inzonex — engineering data for high-temperature plant equipment.

Citation

Canonical DOI (always resolves to the newest version):

Aheiev, D. (Inzonex) (2026). Global Power-Plant Climate Exposure Screening Index (CESI) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.22172589

@dataset{inzonex_power_plant_climate_exposure_index,
  author    = {Aheiev, Dmytro},
  title     = {Global Power-Plant Climate Exposure Screening Index (CESI)},
  year      = {2026},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.22172589},
  url       = {https://doi.org/10.5281/zenodo.22172589}
}