Saturated Steam Pipe Drop Is Non-Isenthalpic, Not J-T

The number that matters is not the 2 bar of friction loss. It is the kilojoules per kilogram that leave through the insulation over 1 km of pipe. Those two quantities compete, and whichever wins decides whether steam arrives at the far end superheated, exactly saturated, or wet enough to hammer.

Four Models That Give the Wrong Exit State

Each of these appears routinely in steam distribution calculations, and each fails for a different reason.

Treating the line as a throttling device. A control valve or an orifice passing saturated steam is modelled as isenthalpic, and dry saturated vapour throttled to a lower pressure lands in the superheated region. Copying that model onto a pipeline predicts superheat at the outlet. It fails because the assumption behind isenthalpic throttling is negligible heat loss, and a kilometre of pipe is the opposite of that.

Running the simulator's adiabatic pipe model and believing the outlet temperature. An adiabatic pipe segment with 2 bar of friction loss will report superheated steam at the outlet, and it is internally consistent: no heat crossed the wall, so the enthalpy had to stay in the fluid. The model is right; the boundary condition is wrong. Adiabatic is a defensible input for the pressure-drop number and an indefensible input for the exit state.

Assuming friction destroys enthalpy. This is the tempting explanation for why the pipe cannot be isenthalpic, and it is the wrong physics. Friction in a duct is an internal irreversibility. The mechanical energy it consumes is dissipated into the fluid's own internal energy, not exported across the control surface. Entropy rises; stagnation enthalpy does not fall. Basing the exit state on "friction losses removed 2 bar worth of energy" double-counts nothing and subtracts something that never left.

Assuming isothermal flow and reading the outlet condition off the inlet temperature. Isothermal is a reasonable compressible-flow approximation for the pressure gradient in a long buried or heavily insulated line. It is not a statement about phase. Held at constant temperature while the pressure falls, the steam would have to be superheated at the outlet, which is again an artefact of an assumption chosen for a different purpose.

Stagnation Enthalpy in Adiabatic Pipe Flow

Take a steady, no-shaft-work control volume around the pipe. The steady-flow energy equation reduces to:

h_in + V_in^2/2 + g*z_in  =  h_out + V_out^2/2 + g*z_out + q_loss

h       = specific enthalpy, J/kg
V       = mean velocity, m/s
q_loss  = Q_wall / m_dot, J/kg (heat leaving through insulation)

Set q_loss = 0 and a horizontal run, and you get Fanno flow: stagnation enthalpy is constant along the pipe, exactly as it is across an orifice. Friction only moves the state along a Fanno line at constant h0, raising entropy and specific volume. So the adiabatic pipe and the adiabatic valve produce the same conclusion — superheat — because they are the same energy balance. Both are isenthalpic in the total sense.

The kinetic term is the only correction inside an adiabatic pipe, and in a header sized for low pressure drop it is small: at typical distribution velocities, V^2/2 is on the order of a few kJ/kg against a latent heat in the thousands. Static enthalpy is therefore very close to stagnation enthalpy at both ends, and the difference between a valve and a pipe cannot come from friction. It comes from q_loss.

Heat Transfer Area: Valve Versus a Kilometre of Pipe

This is heat, not logic. The valve body offers a surface area measured in fractions of a square metre and a residence time measured in milliseconds. The heat it sheds is negligible against the enthalpy flux through it, so q_loss ≈ 0 holds and isenthalpic is a good model.

The pipeline offers A = π·D·L. At 1 km, even a modest diameter gives hundreds of square metres of lagged surface sitting at a wall temperature well above ambient for the whole run. Heat loss per unit mass is:

q_loss = U * A * (T_steam - T_ambient) / m_dot        [J/kg]

U  = overall coefficient through insulation + air film, W/m^2K
     (read from the insulation manufacturer's thermal data,
      not from a generic table)

Against that, the enthalpy the pressure drop can convert to superheat is bounded by the flow work released, of order v̄ · ΔP, with v̄ the mean specific volume and ΔP = 2 bar = 2×10⁵ Pa. In a normally lagged distribution line, q_loss exceeds that term by a wide margin, which is why steam traps exist along the run and why the line stays on the saturation curve from end to end. If insulation losses could be neglected over kilometres, the trap and condensate-recovery industry would not exist.

The non-typical case is real but rare: a very well insulated line carrying a high pressure drop, where the saturation temperature falls faster than the wall losses can pull the steam down. Then superheat does develop. That situation has to be demonstrated by the energy balance, not assumed from the simulator default.

Exit-State Symptoms and Their Causes

Observed at the outlet	Governing condition	Physical cause	Where to read it
T > T_sat(P_out), dry	q_loss < h_g(P_in) − h_g(P_out)	Pressure falls faster than heat is lost; near-adiabatic line with high friction drop	Outlet TE/PT pair vs. steam tables
T = T_sat(P_out), continuous trap discharge	q_loss > h_g(P_in) − h_g(P_out)	Normal insulated distribution line; wall losses condense part of the flow	Trap cycling rate, condensate meter
T = T_sat(P_out), heavy carryover, hammer	Large q_loss, or traps undersized/failed closed	Damaged or wet insulation, missing drip legs, waterlogged trap pockets	Thermal survey of lagging; trap survey
Simulator says superheated, field says saturated	Model run adiabatic	Zero heat-loss boundary condition, not real behaviour	Pipe segment heat-transfer settings in the model
ΔP matches prediction, T does not	Density insensitive to both effects	Cooling and depressurisation shift density in opposite directions	Compare v at inlet and outlet from tables

Segment-by-Segment Energy Balance Procedure

The pressure drop and the heat loss are coupled through the saturation curve, so march the line rather than solving it in one shot.

1. Divide the run into segments short enough that the saturation temperature changes by only a degree or two across each one. For 1 km at 2 bar total drop, 10 to 20 segments is enough.
2. Fix the inlet state: P_in and, for dry saturated feed, h_in = h_g(P_in) from the steam tables. If the supply is already wet, use h = h_f + x·h_fg with the measured or assumed quality.
3. For each segment, compute the friction drop with Darcy–Weisbach using the specific volume at the segment inlet. Verify the Mach number stays low; if the exit-end velocity climbs near choking, switch that segment to a compressible (Fanno or isothermal) solution.
4. For the same segment, compute Q = U·A·(T_sat − T_amb) using the insulation build actually installed, then q_loss = Q / m_dot. Include valve bodies, flanges, supports and any bare sections — uninsulated fittings are disproportionate heat sinks.
5. Update the enthalpy: h_next = h − q_loss. Update the pressure: P_next = P − ΔP_friction.
6. Compare h_next against h_g(P_next). If h_next < h_g, the steam is wet; get the quality from x = (h_next − h_f)/h_fg and log the condensate rate m_dot·(1−x) for that segment as trap load. If h_next > h_g, the steam is superheated; get the temperature from the superheat tables at h_next and P_next.
7. Carry the state forward and repeat. Sum the segment condensate rates to size drip legs and traps.

Note the size of the discriminator in step 6: h_g is nearly flat with pressure across normal distribution ranges, so the superheat allowance created by 2 bar of depressurisation is small — tens of kJ/kg at most. Wall losses over 1 km will usually exceed it, and the calculation lands on the saturation line with a condensate rate attached.

Density, and Why the Incompressible Assumption Survives

Two effects act on density along the run and they oppose each other. Cooling through the insulation drops the temperature and raises density. Friction drops the pressure and lowers density. In a line designed for low pressure drop — which is every line where the user wants usable pressure at the far end — the net change is small, and treating the flow as incompressible for the hydraulic calculation is defensible.

That tolerance does not transfer. An assumption accurate enough for a density that moves by a few percent is not accurate enough to place a state point relative to the saturation curve, where the whole question is which side of a line the enthalpy falls on. Use incompressible or isothermal compressible flow to get ΔP; use the enthalpy balance, and only the enthalpy balance, to get the exit condition.

Field Verification of the Exit State

The check is two instruments and a steam table.

1. Install or read a pressure transmitter and a temperature element at the same point near the outlet, in the same thermowell region, and let the line reach steady flow.
2. Look up T_sat at the measured pressure. If the measured temperature sits at T_sat within instrument error, the steam is saturated and any superheat prediction from the model is a boundary-condition artefact.
3. Confirm with the traps. Continuous or frequent trap discharge along the run is direct evidence that condensation is occurring, which fixes the steam on the saturation curve at that point. A genuinely superheated line downstream of a well-lagged high-drop section will show dry traps and a measurable temperature margin above T_sat.
4. Compare the measured condensate rate against the q_loss/h_fg figure from the calculation. A large discrepancy points at the insulation model, usually wet lagging, missing sections, or bare flanges and valve bodies.
5. Re-run the model with the heat-transfer coefficient that reproduces the measured condensate rate, then trust its exit state.

Instrument placement matters more than instrument class here: a temperature element reading a poorly insulated thermowell in a saturated line will read low and can be mistaken for subcooling that does not exist.

Limits of the Desktop Calculation

Stop calculating when the answer hinges on an overall heat-transfer coefficient you cannot bound — degraded or water-logged insulation, buried lines with unknown soil contact, or lagging whose installed thickness is not documented. At that point the measured condensate rate is worth more than any correlation. Take the trap sizing, drip-leg spacing and condensate-load numbers to the trap manufacturer's application engineering group, and take insulation performance data to the insulation supplier rather than deriving it.

Frequently Asked Questions