Nepal Energy Wiki
Map

Winter Deficit Model

Page status
Working page
Updated
2026-05-12

Open in the interactive explorer and map

Summary

This data page documents a two-phase winter deficit model that estimates Nepal's Dec-Feb monthly energy balance and the 18:30 winter evening peak dispatch gap under parameterised 2025, 2030, and 2035 scenarios.

A two-phase reproducible model quantifying Nepal's winter electricity gap at both monthly energy (GWh) and sub-daily capacity (MW) resolution. Phase 1 (scripts/build_winter_deficit_model.py) models Dec–Feb monthly energy balances under parameterised 2030/2035 scenarios. Phase 2 (scripts/build_diurnal_peak_model.py) models the evening peak hour (18:30) dispatch to answer whether the portfolio that closes the energy gap also delivers evening firm capacity — the 6pm–10pm winter window where solar is zero.

Phase 1 — Monthly energy balance (all scenarios)

Phase 1 reads NEA FY 2024/25 monthly energy data as baseline, applies scenario assumptions (solar deployment, storage hydro additions, demand growth, BESS, demand-shaping), and outputs monthly Dec–Feb balances. Dec–Feb = Poush + Magh (two BS months, 58 days).

Scenario Horizon Solar MWp BESS GWh Demand CAGR Dec–Feb Demand GWh Hydro GWh Solar GWh BESS GWh Shaping GWh Gap GWh Gap %
baseline_fy2025 2025 142 0 0% 2,129 1,551 28 0 0 551 25.9%
current_1gw 2025 1,000 0 0% 2,129 1,551 193 0 0 385 18.1%
current_3gw 2025 3,000 0 0% 2,129 1,551 580 0 0 76 3.6%
current_5gw 2025 5,000 0 0% 2,129 1,551 966 0 0 0 0%
current_5gw_bess 2025 5,000 2 0% 2,129 1,551 966 1 0 0 0%
2030_baseline 2030 142 0 8% 3,128 1,926 28 0 0 1,175 37.6%
2030_17gw 2030 1,700 0 8% 3,128 1,926 329 0 0 874 27.9%
2035_full 2035 5,000 2 7% 4,188 3,247 966 1 400 0 0%
2035_solar_only 2035 5,000 0 7% 4,188 2,412 966 0 0 810 19.3%
2035_no_solar 2035 142 0 7% 4,188 3,247 28 0 0 914 21.8%
2035_no_budhigandaki 2035 5,000 2 7% 4,188 3,247 966 1 400 0 0%
2035_with_budhigandaki 2035 5,000 2 7% 4,188 4,187 966 1 400 0 0%

Assumptions: WECS Low GDP demand CAGRs (8% 2025–30, 7% 2030–35). Terai solar CF 16.5% fixed-tilt. Storage pipeline: Tanahu 140 MW (COD 2026), Dudhkoshi 670 MW (COD 2035, scenario_risk), Budhigandaki 1,200 MW (COD 2036, post-2035). Hydro growth CAGR: 3% (2030), 4% (2035). Dry-season storage energy distributed across four winter months with triangular weights [1,2,2,1] peaking in Poush-Magh. Source: data/processed/tables/winter_deficit_model/model_parameters.csv.

Phase 2 — Diurnal evening-peak dispatch (2035 scenarios)

Phase 2 takes Magh (mid-Jan to mid-Feb, ≈February) as the canonical winter reference month and models the 18:30 evening peak dispatch. The diurnal profile is sourced from Solar–Hydro Complementarity Profile lines 71–85: February weekday demand shape normalised to peak=100 at 18:30, RoR hydro flat at 30, solar bell curve dropping to zero by 17:30. Interpolation to hourly resolution yields an implied peak-to-average demand ratio of 1.35.

Dispatch logic: RoR hydro flat at average monthly MW. Reservoir hydro dispatched at rated MW in the 17:00–21:00 window (dispatch factor 0.9). Solar = 0 MW at 18:30 (sunset). BESS = 500 MW flat in the 17:30–21:30 window (2 GWh / 4h). Demand shaping = an explicit 400 MW peak reduction at 18:30 for scenarios that include demand-shaping energy.

Scenario Peak Demand MW RoR MW Reservoir MW Solar at Peak BESS MW Shaping MW Residual MW Residual %
2035_no_solar 4,066 1,504 729 0 0 0 1,833 45.1%
2035_solar_only 4,066 1,504 126 0 0 0 2,436 59.9%
2035_full 4,066 1,504 729 0 500 400 933 23.0%
2035_no_budhigandaki 4,066 1,504 729 0 500 400 933 23.0%
2035_with_budhigandaki 4,066 1,504 1,809 0 500 400 −147 −3.6%

Model parameters: Magh reference month, 29 days, profile-implied peak-to-average ratio 1.348, reservoir dispatch factor 0.9, solar_mw_at_peak always 0.0 at the default 18:30 target hour. Budhigandaki (1,200 MW × 0.9 = 1,080 MW effective) flips the 933 MW deficit to a 147 MW surplus. Source: scripts/build_diurnal_peak_model.py, 2035_with_budhigandaki scenario.

Scenario definitions

Scenario names are inherited from Phase 1 energy-balance design. Note that 2035_solar_only excludes risky storage (Dudhkoshi) and 2035_no_solar includes it — the labels describe the additional resource being tested, not what is excluded.

Scenario ID Solar MWp Storage hydro active BESS Demand shaping Budhigandaki
2035_full 5,000 Tanahu + Dudhkoshi (810 MW) Yes (2 GWh) Yes (400 GWh Dec–Feb) No
2035_solar_only 5,000 Tanahu only (140 MW) No No No
2035_no_solar 142 Tanahu + Dudhkoshi (810 MW) No No No
2035_no_budhigandaki 5,000 Tanahu + Dudhkoshi (810 MW) Yes (2 GWh) Yes (400 GWh Dec–Feb) No
2035_with_budhigandaki 5,000 Tanahu + Dudhkoshi + Budhi (2,010 MW) Yes (2 GWh) Yes (400 GWh Dec–Feb) Yes (1,200 MW)

The counterintuitive result — 2035_solar_only has a worse evening peak (2,436 MW residual) than 2035_no_solar (1,833 MW) — follows from these definitions. Both have zero solar at 18:30. The difference is 670 MW of Dudhkoshi reservoir dispatch, excluded from 2035_solar_only by the Phase 1 design. The 603 MW residual gap between them (2,436 − 1,833) equals Dudhkoshi's dispatch contribution (670 × 0.9 = 603 MW).

Coverage / Method

Phase 1 reads NEA FY 2024/25 monthly energy data, applies scenario assumptions for solar deployment, storage hydro additions, demand growth, BESS, and demand shaping, and outputs Dec-Feb balances. Phase 2 uses the February weekday diurnal profile from Solar–Hydro Complementarity Profile and the scenario definitions to estimate 18:30 dispatch capacity by resource class.

Caveats

  • Import overcount: Phase 1 overcounts annual imports by ~282 GWh/year because it does not model shoulder-month simultaneous import/export or NEA curtailment. Winter Dec–Feb figures (Poush-Magh) are more reliable than annual totals. Source: data/processed/tables/winter_deficit_model/model_parameters.csv import_accounting_note.
  • BESS sizing: Phase 1 models BESS as a placeholder ~1 GWh Dec–Feb seasonal contribution (2 GWh distributed evenly across four winter months). Phase 2 diurnal model uses the full 500 MW / 2 GWh BESS power rating for evening dispatch, which is more realistic for the 4-hour evening window. The Phase 1 seasonal GWh understates BESS energy contribution; Phase 2 MW capacity is the correct metric for evening firm capacity analysis.
  • Scenario naming: 2035_solar_only excludes scenario_risk storage (Dudhkoshi); 2035_no_solar includes it. Names reflect Phase 1 energy-balance design and are counterintuitive for Phase 2 capacity-credit comparison. Read the scenario definitions table above when interpreting evening-peak results.
  • Reservoir dispatch: All reservoir capacity is assumed dispatched at full rated MW in the 17:00–21:00 window at 0.9 availability. This is optimistic — it assumes reservoirs are full and operators choose evening-peak dispatch. Use --reservoir-dispatch-factor to test lower availability.
  • BESS state of charge: No degradation, no state-of-charge constraints within the daily cycle (assumes full charge from mid-day solar). Optimistic.
  • Demand shaping: Phase 1 models 400 GWh of Dec–Feb energy shifted out of high-stress periods. Phase 2 separately models an explicit 400 MW reduction at the 18:30 target hour for scenarios that include demand shaping; it is a peak-MW assumption, not a GWh-to-MW conversion.
  • Weekday vs weekend: February weekday profile used as canonical. Weekend industrial load is lower but the evening residential peak still dominates.

Model parameters

Full parameter table and scenario definitions are in data/processed/tables/winter_deficit_model/model_parameters.csv. Key inputs: solar CF profile, storage hydro pipeline, NEA monthly energy balance FY 2024/25. Model source code is in scripts/build_winter_deficit_model.py and scripts/build_diurnal_peak_model.py.

Related