A closed-cycle OTEC plant boils a low-boiling-point working fluid (ammonia) in an evaporator heated by warm surface seawater, expands the vapour through a turbine-generator, then condenses it against cold water pumped up from the deep ocean through the cold-water pipe (CWP). The only fuel is the surface-to-deep temperature difference ΔT, so the theoretical ceiling is the Carnot limit:
η_Carnot = ΔT / T_warm(K) = (T_warm − T_cold) / (T_warm + 273.15)
T_cold(depth) ≈ 4°C + (T_warm − 4°C)·e^(−depth / 220m) (thermocline decay)
Real ammonia-Rankine OTEC cycles achieve roughly 25–35% of the Carnot limit once turbine, heat-exchanger and pinch-point losses are included (this model uses a fixed 30%):
Q_thermal = Q_warm · ρ_sw · c_p · ΔT_pinch (ΔT_pinch = 2°C evaporator approach)
P_gross = 0.30 · η_Carnot · Q_thermal
The CWP itself is the main parasitic load. Darcy–Weisbach friction loss over the pipe length is pumped against, so a narrow, deep pipe wastes power on pumping even though it reaches colder water — the classic OTEC engineering trade-off:
v_cw = Q_cw / (π·D² / 4)
ΔP = f · (L/D) · (ρ_sw·v_cw²) / 2 (f = 0.02, L = 1.15·depth)
P_pump = ΔP · Q_cw / η_pump (η_pump = 0.75)
P_net = P_gross − P_pump
- Twarm — tropical surface water; higher values raise both ΔT and the Carnot ceiling.
- CWP depth — deeper reaches colder, more stable water (bigger ΔT) but lengthens the pipe, raising friction loss.
- CWP diameter — a wider pipe carries the same flow at lower velocity, cutting friction loss roughly with 1/D⁵, which is why real CWPs are built as wide as engineering and cost allow.
- Warm-water flow — sets plant scale; if the CWP can't carry enough cold water to match it (velocity capped at 2.5 m/s for erosion control), the cold side throttles the whole cycle.
This is why full-scale OTEC (e.g. NELHA, Hawaii; Makai Ocean Engineering test platforms) needs cold-water pipes several metres wide and up to a kilometre long, and why net power — not gross power — is the number that decides whether a plant is economical.