This is the lesson the whole faculty was building towards — the one-page referee for the question that should open every heating project and usually never gets asked: which carrier? Steam, pressurised hot water, thermic fluid, or an industrial heat pump each own part of the map. Choose right and every downstream lesson gets easier; choose by habit and the plant pays a small tax forever. Here is the matrix, the method, and the honest tie-breakers.

First cut: temperature decides most of it

Carriers are eliminated, not chosen — and temperature does the first round. Below ~90 °C: hot water — and increasingly the heat pump making that hot water (lesson F6): below this line a heat pump can deliver 2.5–4 kcal of heat per kcal of electricity, and at those COPs even expensive power beats cheap fuel for part of the map. 90–130 °C: the genuine contest — PHW (lesson E1) versus reduced-pressure steam; product sensitivity and condensate reality referee. 130–200 °C: steam's heartland — latent-heat punch, moderate pressures, the whole toolkit of Faculties C and D. 200–300 °C: thermic fluid (lesson E3) — steam pressures go exotic here while hot oil ambles along near atmospheric. Above ~300 °C: specialist territory (fired heaters, molten salts) beyond this curriculum's edge.

The matrix

SteamPressurised hot waterThermic fluidHeat pump (to hot water)
Sweet spot130–200 °C90–130 °C200–300 °C≤90 °C (creeping up)
Delivery per kg~500 kcal (latent)~25 kcal (ΔT 25)~35 kcal (ΔT 60)
Control qualitysteps (pressure)fine (mixing)fine (flow)fine
Distribution lossestraps + flash + radiationlowlowlow
Running cost basisfuel ÷ η (B2)fuel ÷ η + pumpingfuel ÷ η + pumping + fluid lifepower ÷ COP
Direct contact (DSI etc.)yesnonono
Cold-start punchexcellentmodestmodestpoor
Regulatory weightIBR (B10)lightlight (fire discipline E4)electrical + refrigerant
Quiet failure modetrap losses, wet steambalancing, cushionfluid degradation (E3/E5)COP collapse at high lift
1 · process temperature eliminates most carriers 2 · steam as a substance? DSI, sterilise, humidify → steam 3 · load shape & distance punchy/batch → steam · steady → loop 4 · ₹ per 1000 kcal delivered B2's method + pumping + COP 5 · the survivor wins state the runner-up + why five questions, asked in this order — temperature first, money last, so economics referees only among carriers that can actually do the job
The decision path. Money comes fifth on purpose: a cheap carrier that cannot hold the product's temperature uniformity was never cheap.

The tie-breakers, from the field

Mixed sites want a spine and limbs. Most real plants over ~1 TPH equivalent end up hybrid: a steam spine for the hot, punchy and direct-contact duties, feeding PHW sub-loops for the gentle band (lesson E1's closing pattern) — one boiler house, each duty on its best carrier. Count the whole crew. Steam brings the largest supporting cast (traps, PRSs, condensate, water treatment — Faculties B–D); TF brings fluid management (E5); PHW brings pumps and balancing (E2); heat pumps bring electrical capacity and refrigerant rules. The carrier's system, not its flange, is what you operate for twenty years. Existing assets bias honestly. A sound boiler house argues for steam solutions at the margin; a plant electrifying under a decarbonisation push (lesson F10's horizon) leans the other way. Name the bias in the study instead of letting it hide in the assumptions. And write down the loser. The best carrier studies end with one line — "PHW chosen; steam was second because condensate return could not be rebuilt in the old block" — because in five years, when someone asks why, that sentence is the whole answer.

At site
  • Before any new heating project: run the five questions in order, on one page, with the runner-up recorded. Twenty minutes; decades of consequences.
  • Audit existing duties against the map: every duty below 130 °C running on 10.5 kg/cm²g steam through a deep PRS is paying avoidable tax somewhere — E1's candidate register finds them.
  • Comparing carriers economically: deliver B2's ₹/1000 kcal at the product — include pumping power, trap/flash losses, fluid amortisation, COP. Boiler-house numbers flatter steam; delivered numbers tell the truth.
  • Distrust any proposal that starts from the equipment ("we supply hot-oil heaters") rather than the duty. The matrix runs forwards only.
Pin this
  • Temperature eliminates first: ≤90 heat-pump/hot water · 90–130 the PHW-vs-steam contest · 130–200 steam · 200–300 thermic fluid.
  • Steam as a substance (DSI, sterilising, humidifying) ends the debate — only steam applies.
  • Punchy batch loads favour latent heat; steady gentle loads favour loops.
  • Decide on ₹/1000 kcal delivered, among qualified survivors only.
  • Hybrid spine-and-limbs is the adult answer for mixed sites; record the runner-up always.
Steam stories

Industry has changed its favourite carrier before. A century ago, factories debated steam versus the new-fangled electric drive for power transmission — line shafts against motors — and the answer took thirty years to become obvious. Heat is slower-moving, but the same pattern is visible: electricity (through heat pumps) has already captured the low-temperature end in Europe's dairies and is probing upward each year. The matrix in this lesson is therefore dated the day it is printed — which is exactly why the method (temperature first, delivered cost last, runner-up recorded) matters more than today's answers.