Answers are grounded in this paper and cite their sources.
Reliable comparisons of battery storage requirements depend on an explicitly defined flexibility service. We applied a common hourly residual-load shifting rule to 25 European power systems during 2015–2025 (2,409,740 complete country–hours), using daily cyclic state-of-charge accounting, 90% round-trip efficiency, a 90% usable SOC range and independent P99 energy and power sizing. In 2025, nameplate energy spanned 1.28–142.17 GWh, power 0.28–23.36 GW and implied duration 4.43–7.38 h. Median energy relative to mean national load increased from 1.11 h in 2015 to 2.24 h in 2025. VRE penetration showed a positive country- and year-adjusted association with normalized energy (0.042 h per percentage point; country-clustered 95% confidence interval, 0.025–0.059), without supporting causal inference. In a matched temporal-service comparison, 2025 median energy was 2.240 h for the daily specification versus 5.027 h for the weekly specification; the contrast also reflects balancing and aggregation differences. With fixed P99 ratings, continuous SOC replay fully served a median 82.74% of days, compared with 98.08% marginal daily-envelope coverage. These are conditional technical benchmarks, not guaranteed service levels, observed curtailment recovery, economically optimal investments or total national storage requirements.
No takes yet. Share an insight, caveat, or question.
Mykhei et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: