JIPR v16 — Markov degradation modelling for fleet-scale substation preservation

Abstract No existing framework combines fleet-scale Markov degradation modelling, multi-hazard environmental acceleration, and empirical failure validation for electricity substations. This paper extends the CIGRE TB 761 five-state Markov model from single-asset condition assessment to fleet-scale preservation planning. We formalise environmental acceleration factors as multiplicative transition-probability modifiers, compute Expected Time to Critical (ETTC) distributions for 142,267 substations across 18 OECD countries via nested Monte Carlo simulation (1.4 billion realisations), and validate against 1,220 independently documented substation failures (2018–2023) on a temporally held-out test set. ETTC is a model-derived risk-ranking and preservation-planning indicator validated against documented failure occurrence, not against ground-truth physical condition inspection. Compound environmental hazards produce non-linear effects: simultaneous heat and coastal corrosion reduce ETTC by 2–4× (λ_combined = 2.3–2.7). Country heterogeneity is pronounced (Kruskal-Wallis H(17) = 48.3, p < 0.001): Spain and Italy have 38–43% of assets with ETTC < 15 years, vs. 14% for Austria/Switzerland. The environmentally-adjusted model achieves Spearman ρ = 0.71 and ROC AUC = 0.78 on the held-out test set (N_test = 406 positives within full test population of 142,267 substations; base rate 0.285%), significantly outperforming the age-only baseline (ρ = 0.46, AUC = 0.60; p < 0.001). The PR-AUC of 0.039 represents a 13.5× lift over the random-classifier baseline (0.0029); bottom-decile lift is 6.6× over base rate. Leave-one-country-out cross-validation confirms generalisability (mean ρ = 0.68). Under the central cost-parameter scenario, ETTC-based condition-based maintenance reduces lifecycle costs by 21% (95% CI 16–27%) relative to age-based replacement; across the full plausible range of cost parameters the reduction is 9–32% (median 19%). The Markov transition matrix is implemented as a Bayesian framework anchored to CIGRE TB 761 [7] and IEEE Standard C57.91 [26] informative priors, updated by openly published national DSO/TSO inspection records. An extended fleet (23 OECD countries plus Greenland) with dual-scenario climate sensitivity (SSP2-4.5 + SSP5-8.5) is reported in a companion Environmental Research: Energy paper. All input data sources are documented per provenance category (public, audit-on-request, model-generated). Code and data are published under GPL-3.0 and CC BY-SA 4.0; Zenodo DOI for the v16-tagged release will be minted at acceptance.