ISO New England Capacity Expansion Planning

Ward-Reduced Network Model & Scenario Analysis

Second region of the multi-region CEP framework (after SPP)
Iowa State University

Press → or Space to advance

ISO-NE as the Second Region

Same Methodology as SPP

Substation-based network reduction → Ward equivalent → DC-OPF capacity expansion
Pipeline parameterized by region (a registry + --region flag); SPP results unchanged
ISO-NE = PSS/E network area 101, FERC-714 respondent 76

Why ISO-NE

Gas-dominant fleet with significant nuclear, pumped storage, biomass
Aggressive decarbonization + electrification policy (winter-peak shift)
Tests the framework's portability on a very different system

Data Sources (all national, cut to ISO-NE)

Network: PSS/E 2024 ITP Eastern-Interconnection case — area-101 carve-out (4,906 buses)
Generator technology: EIA bus-match (Generator_to_PSSE_Bus_Match) → PSS/E gen-tech table → heuristic
Load: FERC-714 respondent 76 (ISO New England Inc.), 8,760 hourly (2024)
Renewable potential: NREL reV supply curves — limited / reference / open access scenarios
Capacity factors: NOAA HRRR weather backcast (2024), per site → per bus
Bus locations: national OSM geocoding (lat/lon for siting + maps)

Network Reduction: 4,906 → 440

EI case
93,300 buses

→

Area 101
4,906

→

Substation grouping
(xfmr + low-Z + short line)

→

≥69 kV backbone
+ junctions

→

Ward equivalent
440 buses

Buses in one physical substation (linked by transformers, low-impedance ties <0.001 pu, or short lines <2 mi) collapse to their highest-voltage representative. Result: 440 buses, 1,278 equivalent lines — an 8.97% retention (11.15× reduction), single connected island.

Existing Fleet — 22,459 MW

Technology	MW	Technology	MW
Combined Cycle	9,304	Oil GT/ST	906
Nuclear (Seabrook, Millstone)	3,527	Coal ST	645
Gas CT	2,303	Solar (existing)	667
Pumped Storage (Northfield, Bear Swamp)	1,507	Wind (existing)	518
Biomass	1,125	Gas ST	666
Hydro	1,056	Battery	21

Gas-dominant (~55%); classified via EIA bus-match + PSS/E gen-tech table (0.9% unclassified).

Load & Capacity Factors

Load (FERC-714 respondent 76)

2024 peak 24,854 MW → scaled to PSS/E snapshot 23,899 MW
America/New_York tz, summer-peaking (7 pm EDT)
17 segments: 4 seasons × 4 blocks + peak
Growth scenarios: 0.9% / 1.7% / 3.5% (low / CELT / deep-decarb)

Renewable shapes (HRRR backcast)

CF varies by every Season × Block
Wind mean ≈ 0.38, Solar mean ≈ 0.26
Solar ≈ 0 at night; ≈ 0.006 at the 7 pm peak → little capacity credit
Wind ≈ 0.23 at peak

Scenario Results

Scenario	NPV cost	Wind	Solar	Gas CT	CC
Low load (0.9%)	$66.8B	13.9	12.3	7.1	0.4
Reference (1.7%)	$76.2B	16.2	17.8	8.4	4.1
High load (3.5%)	$104.6B	44.5	21.9	19.6	8.8
High × limited renewables	$118.6B	56.9	30.6	13.1	9.8
High × open renewables	$103.4B	43.4	20.0	20.6	8.7

New build, GW added 2024→2049. All cases solve OPTIMAL with ~zero load shedding.

Key Findings

$67–105B

NPV cost across the load-growth band

~0 shed

Resource-adequate under the carbon cap

+$14B

Cost of limited renewable access at high load

Load is the dominant driver of cost and build.
Renewable access is non-binding at low/reference load (candidate potential ≫ what is built) but binds at high load — limited access costs +$14B vs reference; reference ≈ open.
A screening design (vary the binding dimension, test the other) gave the full picture in 5 solves — no need for all 9 or formal scenario reduction.

Caveats & Next Steps

Lower-fidelity items

Candidate renewable potential is large (reV technical potential), but a 1 GW/bus cap scenario moves cost only +$0.1B (ref) / +$1.1B (high) — the plan is robust to realistic interconnection limits
Single-rate growth can't capture ISO-NE's winter-peak electrification / summer→winter flip
~31% of pre-reduction gen capacity rests on a name/sheet heuristic
Area-101 carve-out of the SPP-ITP case (no neighbor imports modeled)

Deliverables

Interactive dashboards (low / ref / high), 440-bus map
Master documentation (methodology + data + results)
Fully reproducible via --region isone pipeline

Questions?