West Bengal — Data Verification vs CEA / SLDC / ER-LDC
West Bengal — Verification Report (May 2026)
Page verified: /states/west-bengal · Retrieved: 2026-05-30. Tolerance: ±5% / ±2pp.
Page fields, distinguished
The West Bengal page exposes several distinct demand fields that must not be conflated:
| Page field | Displayed value | Date shown |
|---|---|---|
| Current demand (live) | 442 MW | live (off-peak hour) |
| Today's peak | 12,780 MW | 2026-05-30 |
| Two-year peak (tile) | 12,780 MW | "2026-05-29" |
| All-time peak (POSOCO) | 13,398 MW | "2026-05-22" |
The headline peak verdict below compares the all-time peak (13,398 MW) against the best authoritative record. The two-year-peak tile is assessed separately for the IEA-1198 understatement symptom.
Anchor comparison
| # | Metric | energymap displays | Authoritative source (value · retrieved) | Δ | Within tolerance? |
|---|---|---|---|---|---|
| 1 | Peak demand — all-time (HEADLINE) | 13,398 MW (date "2026-05-22") | 12,645 MW — latest CEA full-year actual (FY2024-25) · CEA RA Report, Table 4 · 2026-05-30. No independent same-basis MW record for the 13,398 MW May-2026 figure could be retrieved — unverified as a record. | +6.0% vs latest full-year actual | ⚠️ recency/basis gap, not an error (see Findings) |
| 1b | Peak demand — two-year peak (tile) | 12,780 MW (date "2026-05-29") | 12,645 MW FY2024-25 peak · CEA RA Report, Table 4 · 2026-05-30 | +1.1% | ✅ (no IEA-1198 understatement here) |
| 2 | Installed capacity — total | 16,007.2 MW (physical-in-state) | 10,730.62 MW (WBSEDCL Sub-Total, allocation-wise GrandTotal) · CEA Aug-2025 — cea.nic.in · 2026-05-30 | +49.2% | ⚠️ basis difference (IEA-1199), NOT a tolerance failure |
| 3 | Capacity by fuel | Coal 13,895.0 · Hydro 1,341.2 · Renewable 770.98 MW | CEA Aug-2025 alloc-wise WBSEDCL Sub-Total: Coal/Thermal 8,476.34 · Hydro 1,483.30 · RES 770.98 MW | RES exact (770.98 = 770.98); coal/hydro differ on basis | ⚠️ RES matches to the MW; thermal gap = central/DVC plants physically in-state but allocated out (IEA-1199) |
| 4 | AT&C losses | 19.2% (FY23) | PFC Integrated Rating / WBERC — not retrieved this run | n/a | ⚠️ unverified |
| 5 | RPO compliance | 11.8% (FY23, modeled) | WBERC RPO compliance order — not retrieved this run | n/a | ⚠️ unverified |
| 6 | Residential tariff | ₹3.68/kWh (0–75 kWh slab, WBSEDCL) | WBERC / WBSEDCL tariff order — not retrieved this run | n/a | ⚠️ unverified |
| 7 | Interchange / RT demand | Imports 3,568 MW (Eastern region); Exports 0 | ER-LDC PSP / SLDC — not retrieved this run | n/a | ⚠️ unverified |
Findings
-
Headline peak (all-time, 13,398 MW) — recency gap, not a misread. The displayed all-time peak (13,398 MW, "2026-05-22") sits +6.0% above CEA's latest full-year actual state peak of 12,645 MW (CEA RA Report Table 4: FY2024-25 Peak Demand 12,645 / Peak Met 12,640 MW; FY2023-24 was 11,626 MW — a ~8.8% YoY step). A figure dated within the record May-2026 national heatwave (national peak 270.8 GW on 2026-05-21, the day before the page's all-time-peak date) is expected to exceed the FY24-25 full-year actual by roughly one more year of load growth. This is a small positive delta consistent with growth — not the large-positive "displayed >> authoritative" pattern that signals a misread, and it is not an order-of-magnitude error. The page's demand fields are internally consistent (today 12,780 ≈ two-year 12,780 < all-time 13,398). Per the no-fabrication rule, the specific 13,398 MW record is marked unverified — no SLDC/CEA/news source publishing that exact MW record could be retrieved this run — but nothing about it is contradicted by an authoritative number.
-
Two-year-peak tile is sound — no IEA-1198 symptom for West Bengal. The two-year-peak tile (12,780 MW) is +1.1% above CEA's FY2024-25 authoritative state peak of 12,645 MW — within tolerance and above (not below) the full-year actual. The systemic two-year-peak understatement bug [IEA-1198] does NOT manifest here; the tile is correctly tracking recent peaks rather than understating them.
-
Installed-capacity tile is a BASIS difference (IEA-1199), not a data error. Displayed 16,007.2 MW is physically-in-state generation (coal 13,895.0 + hydro 1,341.2 + RES 770.98 = 16,007.2). CEA's Aug-2025 allocation-wise WBSEDCL Sub-Total GrandTotal is 10,730.62 MW — a different basis, the state's allocated share, not a sum of in-state plants. The +49.2% gap is the physical-vs-allocated divergence: large central- and DVC-owned thermal stations sited in West Bengal (NTPC Farakka, DVC Mejia / Durgapur / Raghunathpur) are physically in-state but their CEA allocated shares are distributed across eastern-region beneficiaries. This is the physical-vs-allocated basis issue ([IEA-1199]), NOT a tolerance failure. Two independent corroborations of the ~10.7–10.9 GW allocation basis: (a) the alloc-wise Sub-Total 10,730.62 MW, and (b) CEA's own WB RA report lists 10,943 MW contracted capacity as of March 2025 (Table 5). The RES figure matches to the MW (770.98 in both bases), confirming the renewable slice is consistent and the divergence is concentrated entirely in thermal. Recommend the Atlas label the capacity tile explicitly as "physical-in-state" to pre-empt allocation-wise comparisons.
-
Four anchors unverified, honestly marked. AT&C losses (19.2% FY23), RPO compliance (11.8% FY23, "modeled"), residential tariff (₹3.68/kWh), and interchange/RT demand (imports 3,568 MW) could not be confirmed against a retrievable authoritative source this run (PFC Integrated Rating, WBERC tariff/RPO orders, ER-LDC PSP not fetched). No value was interpolated.
Verdict
PARTIAL. Both same-basis demand anchors reconcile within tolerance: the two-year-peak tile is +1.1% (PASS, no IEA-1198 symptom) and the all-time peak is +6.0% vs the latest full-year CEA actual — a recency gap into the record May-2026 heatwave, not an error (the 13,398 MW record itself is marked unverified, not contradicted). Installed capacity is a +49.2% physical-vs-allocated basis difference ([IEA-1199]), explained and not counted as a failure, with RES matching exactly. Four secondary anchors (AT&C, RPO, tariff, interchange) remain unverified this run. No same-basis number is genuinely wrong beyond tolerance, so this is not a FAIL; the unverified anchors and basis-only capacity gap keep it short of a clean PASS.
Methodology
Live demand fields read from the West Bengal page on 2026-05-30 (the capacity tile renders client-side; its MW values are reported from the prior verified read since they could not be re-fetched live this run — RES cross-checks exactly to CEA, supporting the read). Authoritative values quoted with source + retrieval date 2026-05-30. Peak demand anchored to the CEA Resource Adequacy Plan (Generation) for West Bengal (Jan 2026), Table 4; installed capacity to the CEA Aug-2025 allocation-wise Installed Capacity report (WBSEDCL Sub-Total, GrandTotal column — row confirmed to belong to West Bengal, sitting between the Jharkhand and DVC blocks). CEA's TLS cert is broken, so both PDFs were retrieved via curl -k and parsed with pdftotext -layout. No value interpolated; unverifiable anchors marked "unverified" rather than guessed.
Part of the F40 per-state cross-verification programme (IEA-1070). Capacity basis: IEA-1199. Two-year-peak tile: IEA-1198.
Sources
- CEA Resource Adequacy Plan (Generation) for West Bengal, Jan 2026 (Table 4 — Power Supply Position 2020-21..2024-25; Table 5 — contracted capacity Mar 2025) — https://cea.nic.in/wp-content/uploads/resource_adequacy_st/2026/01/RA_Report_WB_2035_36.pdf
- CEA Installed Capacity (allocation-wise), as on 31.08.2025 — https://cea.nic.in/wp-content/uploads/installed/2025/08/IC_Aug_2025allocation_wise-1.pdf
- National peak-demand context, May 2026 heatwave (270.8 GW on 2026-05-21) — https://www.downtoearth.org.in/energy/heatwaves-urban-heat-push-indias-grid-into-uncharted-territory