TN-022 — Duration-Capable Node: CAPEX Model

Technical Note · Open Working Draft v0.1 · June 2026
Domain D-1 · D-4 · Companion to TN-001 v1.1
Relates: WP-019 · SM-010 · TN-011

§ 01 — Reference Node

The reference configuration is a 60 MW thermal output node supplying a district heating network of 15,000–25,000 households, with co-located chemical storage and VPP coordination. Three scenarios are modelled:

§ 01A — Node Scaling Classes and Baseload Compatibility

TN-001 nodes are not a single size. Peak heat demand determines configuration class. The following classification applies regardless of jurisdiction — representative Finnish municipalities are noted as examples only.

Representative examples at N-1 scale: municipalities similar to Varkaus, Iisalmi, Lieksa. N-2: Joensuu, Kouvola, Seinäjoki. N-3: Oulu, Jyväskylä. N-4: Tampere, Turku, Helsinki metropolitan area. These are illustrative only — configuration class depends on heat demand profile, not administrative category.

Baseload Fraction (BF) and Baseload Compatibility Index (BCI)

Minimum load relative to peak load determines how much constant-output heat a node can absorb without surplus problems:

where D_min = annual minimum heat demand (typically summer night load) and D_peak = design peak demand.

The Baseload Compatibility Index (BCI) formalises this for LDR-50 specifically:

where P_baseload = LDR-50 rated output (50 MW). Interpretation:

Design rule (updated for PEM+HP integration): Where BCI < 0.6, PEM + industrial heat pump delivers comparable heat output with full operational flexibility and no surplus constraint. LDR-50 adds duration but at the cost of structural rigidity that PEM+HP avoids. The crossover point is approximately BCI = 0.6–0.7 depending on electricity price and seasonal demand profile.

LDR-50 becomes structurally attractive when annual minimum heat demand exceeds approximately 30–40 MW and remains above 30% of peak load for a significant portion of the year (BF > 0.3, BCI > 0.6). Below this threshold, the operational complexity of managing constant LDR-50 output against variable demand outweighs the duration benefit.

§ 02 — Component CAPEX Breakdown

Property A — Dispatchable Generation

Property B — Chemical Energy Storage

Property C — Heat System Anchor

Property D — Market Flexibility Interface

Civil and Project Costs

§ 03 — Scenario Totals

§ 04 — Comparative Reference Points

At comparable thermal output, TN-001 S-2 and S-3 are in the same investment order of magnitude as LDR-50. Duration capability costs roughly 3–5× the conventional wood-chip baseline (S-0). LDR-50 at FOAK is not cost-competitive with S-2/S-3 for sub-100 MW thermal applications on a time-adjusted basis. Effective CAPEX = stated CAPEX + (delay cost × additional years) + regulatory risk premium. At a 7% cost of capital, 4 additional years of delay on a 200 M€ project adds ~60 M€ in financing cost alone — pushing effective LDR-50 CAPEX to 210–360 M€ vs S-2/S-3 at 100–145 M€ stated. NOAK LDR-50 (2035+) may change this comparison if serial production reduces unit cost substantially., but buildable on current regulatory frameworks, faster, with dispatchable output and self-produced fuel. The duration advantage of LDR-50 (fuel cycle years) is real; the cost and regulatory risk are also substantially higher at FOAK stage.

§ 05 — Revenue Model and Payback

The four-revenue-stream model from SM-010 applies directly. The risk floor principle: node viability does not depend on speculative revenue streams.

Indicative payback at central CAPEX (110 M€) and central EBITDA (45 M€/year): 2.4–3.1 years simple payback. IRR 14–20% at base case; 8–12% under stress (lower reserve prices, higher build cost). These figures match the SGFA Master Package analysis from the Kuopio node evaluation (April 2026).

Stress case note. The base case EBITDA assumes persistent reserve market participation and favourable electricity price volatility conditions. A critical reader will note that 2.4–3.1 year payback is exceptional by energy infrastructure standards — the question is why nodes of this type are not already being built everywhere. The answer is threefold: (1) reserve market revenues require Fingrid participation agreements not yet standardised for distributed nodes; (2) synthetic fuel revenue requires RFNBO certification with regulatory lead time; (3) FOAK integration risk is real. Stress case payback at reduced reserve revenue and lower fossil displacement savings: 6–10 years — still competitive with conventional district heating investment but without the headline numbers.

Key base case assumptions: district heat contract 55 €/MWh (long-term HPA), reserve market average 65 €/MWh, displaced fossil fuel 70 €/MWh (internal saving — replaces market gas). Sensitivity: heat price −30% → payback extends to 3.5–4.5 years; fossil fuel price −30% → payback extends to 4.0–5.0 years; both simultaneously → payback 5.5–7 years at central CAPEX.

CAPEX alone does not determine competitiveness. The relevant comparison metric is Levelized Cost of Heat (LCOH) — total system cost per MWh of useful heat output over the asset lifetime, including OPEX. A node with high CAPEX but very low fuel cost (self-produced SNG or F-T) and long asset life may have lower LCOH than a lower-CAPEX node dependent on market-priced fuel. LCOH quantification for each scenario requires site-specific OPEX inputs (electrolysis power cost, maintenance, catalyst cycles) and is outside the scope of this note; it is a necessary next step before investment appraisal.

§ 06 — Key Uncertainties

Electrolysis cost trajectory. PEM electrolysis costs have fallen rapidly (from ~1500 €/kW in 2020 to ~700–900 €/kW in 2026) and are expected to continue declining. S-2 and S-3 costs will improve materially by 2028–2030 deployment timeframe.

SOFC pricing. Elcogen/Convion SOFC systems remain at ~4–6 M€/MW — reflecting relatively low production volumes. At higher volumes the cost structure is fundamentally different from combustion engines; learning curve effects could bring this to 2–3 M€/MW within the decade.

NOAK cost multiplier. Subsequent nodes of identical design (2nd, 3rd) are substantially cheaper. Civil works and EPC costs — the largest reducible components — fall significantly with replication. Equipment follows learning curves but is less sensitive. Indicative NOAK multiplier: 0.70–0.80 × FOAK total. A programme of 5 identical S-2 nodes would bring per-node cost from ~110 M€ to ~80–90 M€ by node 3–4.

CO₂ feedstock availability. S-2 and S-3 methanation and F-T synthesis require local CO₂ supply. The S3 signal in TN-001 §04 applies: availability window is narrowing as industrial CO₂ streams commit to geological storage. Nodes planned after 2028 may face higher CO₂ sourcing costs.

FOAK premium. The ±30% FOAK contingency is the largest single uncertainty. For the second and subsequent nodes of identical design, civil and EPC costs fall significantly — analogous to the LDR-50 FOAK vs NOAK gap.

§ 07 — Complementarity with LDR-50 and Long-term Architecture

Node and Node⁺ Architecture

LDR-50 does not replace the TN-001 node — it enhances one of its properties. The architectural relationship is:

LDR-50 is a Property C enhancer, not a standalone configuration. It extends the heat anchor's duration from weeks (Property B fuel store) to years (nuclear fuel cycle) — but only where BCI > 0.6. Where BCI is low, Node without LDR is the correct configuration. Where BCI is high, Node⁺ becomes viable when LDR-50 reaches commercial deployment (2030+).

Expected Network Optimisation Result

If applied across the Finnish district heating stock, global optimisation would likely produce differentiated deployment rather than uniform rollout:

LDR-50 would not spread uniformly — it would concentrate in large load centres with high baseload fractions, while smaller municipalities deploy PEM-CHP-synthetic fuel combinations. This is structurally efficient: the technology's duration advantage is most valuable where constant output can always be absorbed.

TN-001 nodes and the LDR-50 small heat reactor are not competing configurations. They operate in different time windows, address different system functions, and are architecturally complementary.

LDR-50 — What It Does and Does Not Do

The LDR-50 (Steady Energy) produces thermal energy only — approximately 50 MW heat output delivered directly to a district heating network. It produces no electricity. Its function is to replace fossil fuel baseload in district heating: the fuel cycle extends to years, providing heat supply independence from both market price volatility and short-term fuel logistics. This is an exceptional duration capability for Property C (heat system anchor) — but it is limited to heat.

LDR-50 does not provide: dispatchable electricity generation (Property A), chemical energy storage or synthetic fuel production (Property B), or market flexibility interface (Property D). These remain necessary regardless of whether LDR-50 is present.

TN-001 in the 2029 Window

The 2027–2032 convergence window — SE1 capacity tightening, CHP closures, PPA liquidity compression, Layer 3 exposure — arrives before LDR-50 serial deployment. TN-001 nodes can be built now, on current regulatory frameworks, with current technology. They address the immediate gap with Properties A, B, C and D as an integrated system. LDR-50 is realistically a 2030–2035 deployment at scale.

Architectural Division of Labour

When LDR-50 arrives, it can replace or supplement Property C in existing TN-001 nodes — providing the district heat baseload from a multi-year fuel cycle while releasing TN-001's thermal capacity for higher-value applications. Property B (electrolysis, synthetic fuel production) and Property D (VPP, electricity markets) remain fully active and necessary: LDR-50 does not address electricity system flexibility or chemical energy storage.

LDR-50 is the heat anchor. TN-001 is the engine room, the fuel store, and the control layer. The Satama metaphor holds: LDR-50 is the warm berth; TN-001 is the working port.

Continuity and Defence Value

A distributed network of TN-001 nodes — each with its own chemical fuel store, dispatchable generation, and heat output — aligns with WP-002's Distributed Resilience Doctrine: many nodes, difficult to suppress, individually valuable. LDR-50, as a high-capacity heat anchor, provides a different resilience property: near-permanent thermal continuity from a single installation that requires protection but offers years of autonomous operation.

Together they form a layered architecture. A single strike cannot simultaneously disable distributed TN-001 nodes and a protected LDR-50 installation. For Puolustusvoimat and Huoltovarmuuskeskus, the combination covers both the resilience requirement (distributed, hard to target) and the duration requirement (years of heat supply without resupply logistics).

§ 08 — Thermal Balance Optimisation

A node possessing Properties A–D and configured with a baseload heat source (LDR-50) faces a non-trivial optimisation problem: three heat sources with different temporal profiles must supply a district heating load that varies seasonally by a factor of 5–10.

Thermal State Equation

The hard constraint is LDR-50 rigidity: Q_LDR(t) = 50 MW ∀t. This is the system's stiff mode condition — it forces summer surplus and makes winter shortfall structurally dependent on CHP. VTT's containerised PEM system (2024) adds an industrial heat pump to the electrolysis unit, changing the thermal coupling fundamentally. Q_recover(t) now includes both electrolysis waste heat and heat pump output: at COP = 3–4, the heat pump can deliver 3–4 MWh of heat per MWh of electrical input, recovering low-grade waste heat to district heating temperatures. PtX is therefore not only a market-driven heat valve — it is simultaneously a heat producer. This has a counterintuitive consequence: in summer, PEM + heat pump may worsen the LDR-50 surplus problem rather than solve it. LDR-50 and PEM are not automatically synergistic heat sources; under certain conditions they compete. CHP functions as the residual controller — activated only when LDR-50 + PtX + storage do not meet demand:

CHP is a stress buffer actuator, not a production source. This distinction matters for fuel store sizing: winter CHP dispatch consumes Property B inventory at a rate determined by demand residual, not by generation economics.

Three Heat Sources, Three Logics

Seasonal Load Asymmetry

Typical Finnish district heating load (Kaukolämpötilasto 2025) for a 60 MW node serving 15,000–25,000 households: winter peak 150–200 MW, summer minimum 15–25 MW, winter/summer ratio 6–10×.

Winter (150 MW load): LDR-50 provides 50 MW. PtX waste heat 10–30 MW. CHP must cover remaining 70–90 MW — consuming Property B stored fuel.

Summer (20 MW load): LDR-50 alone produces 50 MW — a 30 MW surplus. If PtX also runs to produce winter fuel stock, its waste heat adds to the surplus. This surplus must be stored or dumped.

Critical Constraint

LDR-50 cannot ramp below minimum stable level (25–35 MW for a 50 MW unit). Summer load (20 MW) may fall below this threshold even at minimum. Three solutions exist, singly or in combination:

1. Thermal storage (Property C enlargement). Store excess summer heat in hot water accumulators. Requires 200–400 MWh thermal capacity (2,000–8,000 m³ water tanks). CAPEX addition: 5–15 M€.

PEM system value beyond hydrogen. The key insight from VTT's PEM system is that its value in a TN-001 node is not only the hydrogen it produces but its role as a thermal flexibility element. PEM operating at minimum load in summer avoids heat surplus; at maximum load in winter it simultaneously produces hydrogen (refilling Property B store), recovers heat (supplementing Property C), and participates in reserve markets (Property D). CHP fuel consumption is reduced because PEM covers part of the thermal load. This expands the S-2 economic frontier: the same investment now delivers hydrogen, SNG precursor, waste heat recovery, heat pump integration, and reserve market flexibility — without separate components. The CAPEX efficiency of S-2 improves relative to TN-022 §02 assumptions.

2. High-temperature electrolysis (SOEC). Route LDR-50 surplus heat into electrolysis at 700–850°C, improving efficiency from ~85% to ~95% and transforming summer surplus into higher hydrogen yield. TRL note: PEM (VTT containerised system, commercial 2024) is significantly further along than SOEC for this application. SOEC carries a higher FOAK penalty in the 2027–2032 deployment window — realistic commercial availability is 2030+. In the near term, PEM with industrial heat pump is the more reliable thermal recovery path. CAPEX premium over PEM when commercially available: 2–5 M€.

3. Seasonal CHP dispatch shift. Reduce CHP to near zero in summer; run LDR-50 at minimum; dump surplus heat to ambient if necessary. Operationally feasible but wasteful — acceptable only if thermal storage is constrained by site.

Property D Role

Real-time coordination across LDR-50 output (fixed but ramped to minimum), PtX power consumption (price-driven, heat output must be forecast), CHP dispatch (fuel cost), and thermal storage state of charge involves multiple timescales — hourly for PtX, seasonal for storage — and non-linear efficiencies (SOEC heat coupling). Property D must include a thermal balance layer in its optimisation engine, not only electrical market signals.

Optimisation objective extension. Thermal balance adds a fourth objective to the CAPEX/Duration/Resilience triplet — a thermal infeasibility penalty that forces the system to account for LDR-50 rigidity:

This penalises both surplus (summer dump) and deficit (winter shortfall). It makes SOEC and thermal storage CAPEX decisions dynamic rather than static — their value depends on the full annual load profile, not on peak sizing alone.

Without thermal balance optimisation, a node with LDR-50 baseload will either dump summer heat (waste) or over-build CHP (inefficient). With SOEC integration, the summer surplus becomes higher hydrogen yield — transforming a thermal liability into a chemical asset.