Locating Desuperheater Thermocouples After Water Spray

A desuperheater outlet thermocouple that sits inside the evaporation zone does not measure steam temperature. It measures a wetted metal surface, and the control loop built on it will drive the spray valve open until the downstream pipe is running water. Every check below names a reading, what each outcome means, and which check follows.

Check 1: Wet Probe or Bad Loop

Before anything else, confirm the indication is physically possible. Take line pressure at the measurement point and look up saturation temperature at that pressure.

1. Record steam pressure P2, indicated temperature T2_meas, upstream temperature T1, steam flow, and spray water flow and temperature.
2. Compute the heat-balance outlet temperature from m_s*(h1 - h2) = m_w*(h2 - h_w), solved for h2, then read T2_calc at h2 and P2.
3. Compare T2_meas with Tsat(P2) and with T2_calc.

If T2_meas pins within a few degrees of Tsat(P2) while the heat balance says you should have 40-100 degF of superheat, the probe is wetted. Go to the mechanism section, then Check 2. If T2_meas tracks T2_calc within instrument error at every load, the measurement location is sound and the problem is loop tuning, spray valve stroke, or water flow measurement — stop here.

A third outcome exists and is diagnostic: T2_meas matches the balance at high steam flow and collapses toward saturation at low flow. That is a turndown failure at the nozzle, not a wrong thermowell location. It takes you to Check 2 as well.

Symptom	Reading to take	Mechanism	Next check
Indication clamps at saturation, will not rise when spray closes	Tsat(P2) vs T2_meas	Liquid film on well; latent heat holds the tip at Tsat	Check 2
Fast drop on spray step, slow recovery	Open-loop step response, both directions	Asymmetric wetting/drying time constant	Check 4
Correct at high load, wet at low load	Nozzle differential pressure vs steam flow	Below minimum atomizing DP the nozzle dribbles	Check 2
Bottom-of-pipe skin TC reads well below top	Top/bottom skin thermocouple pair	Stratified liquid film running along the invert	Check 5
Downstream elbow or tube bank shows tube-to-tube spread	Header/tube outlet temperature scan	Slug carryover past the spray station	Check 5
Bang or shock on load change downstream	Pipe support and hanger inspection, bypass line	Water hammer from accumulated condensate	Check 5

Mechanism: What a Droplet Does to a Thermowell

The question of continuous versus discrete treatment resolves itself the moment you fix the objective. You are not looking for a mixed-mean temperature; you are looking for the axial station at which the last droplet disappears. That is a Lagrangian problem — track a droplet of the largest diameter the nozzle produces, integrate its energy balance along the pipe, and find the length. A continuous, homogeneous-mixture model gives you the correct thermodynamic endpoint and tells you nothing about where it occurs, because it assumes equilibrium at every station. Use the continuous model only downstream of the length the discrete model returns.

At the well itself, three effects stack. A droplet striking the sheath deposits its latent heat directly into the tip; while any liquid film persists the surface cannot exceed Tsat. Below the Leidenfrost temperature the film wets and spreads, which is the worst case; above it the droplet rides on a vapor cushion and rebounds, giving an intermittent, noisy error instead of a steady one. And the stem conducts: a wetted well in a superheated line pulls heat out of the tip through the sheath, so even partial wetting biases the reading low by more than the droplet's own energy content. The result is a measurement that is always biased in the direction that tells the controller to spray more.

Check 2: Water Subcooling and Nozzle Differential Pressure

Two water-side variables dominate evaporation time, and both are read at the nozzle, not at the pump.

1. Measure spray water temperature at the nozzle inlet and compute subcooling as dT_sub = Tsat(P_nozzle_discharge) - T_water. Sensible heating to saturation happens before any mass leaves the droplet, and that time is dead length in your pipe.
2. Measure the differential across the nozzle at minimum, normal, and maximum spray demand. Compare each against the supplier's minimum atomizing differential.

Vendor statements that evaporation completes before the mixture leaves the desuperheater body are written for a design point, not for the operating envelope. Read them as conditional: such a claim may hold only where available differential pressure across the nozzle exceeds roughly 200 psid and the spray water is no more than about 5 degF subcooled. Outside that window the same hardware produces coarse drops that leave the body as liquid. Source the water from a point that is only slightly cooler than saturated conditions, and specify a nozzle with wide turndown — modern spring-loaded designs hold differential pressure across the orifice as flow falls, which is exactly what fixes the low-load branch of Check 1.

If the differential collapses at low spray demand, stop here and fix the nozzle or add a minimum-flow interlock that blocks spray below the steam velocity the supplier specifies. If differential and subcooling are both in range, proceed to Check 3.

Check 3: Droplet Diameter Without CFD

Ask the nozzle supplier for Sauter mean diameter at your differential pressure and water properties. If that data does not exist, bound the problem from above with the aerodynamic breakup limit, which is the diameter that survives in the steam stream:

d_max = We_crit * sigma / (rho_steam * u_rel^2)

We_crit ~ 12   (bag breakup onset)
sigma       = surface tension of water at Tsat(P_nozzle discharge)
rho_steam   = steam density at P2, T1
u_rel       = |u_droplet - u_steam| at injection

Surface tension is evaluated at saturation conditions corresponding to the pressure at the nozzle discharge, not at the pump. It falls steeply as pressure rises, which is why the same nozzle atomizes better in a high-pressure line than a low-pressure one. Droplet diameter, degree of subcooling, and that surface tension are the three variables that set evaporation time; everything else is second order.

Do not chase the smallest possible droplet by raising water pressure without checking momentum — see Check 5. Carry two diameters forward: the supplier SMD for the expected case, and d_max for the bounding case.

Check 4: Evaporation Time and Straight-Length

Integrate the classical d-squared law with a convection correction, using an effective latent heat that absorbs the subcooling:

Evaluate steam properties at the film condition between Tsat and the local steam temperature, and recompute Re_d as the drop decelerates — a two- or three-step hand integration is enough. Run it for both diameters from Check 3.

Now compare against field experience, because the analytical number is only as good as d0. The older attemperator rule of thumb calls for enough straight pipe downstream of the nozzle to give 0.35 s of steam transit time before any obstruction or measurement. Compute L_rule = u_steam * 0.35 s. Measured plant behavior brackets this widely: complete vaporization has been observed within 5 pipe diameters of the injection point in favorable installations, and no detectable loss of superheat at 15 diameters in unfavorable ones. Take the largest of L_evap(d_max), L_rule, and 15 pipe diameters as the thermowell station. For the underlying theory, the non-equilibrium droplet evaporation treatments published in the International Journal of Multiphase Flow between 1981 and 1994 for droplets inside boiling furnace waterwall tubes apply directly to this geometry.

Check 5: Momentum, Impingement and Downstream Geometry

Finer is not automatically better. Raising water pressure raises droplet injection velocity, and past a point the spray crosses the duct before the steam can turn it. In one oxidizer quench service, changing from 150 psig spray heads to 1200 psig fine-droplet nozzles produced water impingement on the opposite wall 6 ft away; backing the supply down to roughly 500 psig restored gas/water contact. Check the same way here: look for wall thinning, thermal fatigue cracking, or a cold streak on the pipe opposite the nozzle.

Then walk the downstream piping. Unevaporated water collects on the first elbow, runs as a film along the invert, and releases as slugs. In HRSG attemperator service this path produces tube-to-tube temperature imbalance in the next bank, thermal stress, tube overheating, tube-to-header weld failures, and water hammer in turbine bypass piping. If the spray station discharges into a short run followed by an elbow, the thermocouple location is not your only problem — the station itself needs relocation or a longer straight run, ideally with the flow horizontal or upward so gravity does not build a film on the invert.

Procedure: Relocating the Measurement

1. Fix L_min = max(L_evap(d_max), u_steam * 0.35 s, 15 pipe diameters), evaluated at minimum steam flow, since that is the longest transit time but also the coarsest spray and lowest velocity — run the calculation at minimum, normal, and maximum load and take the worst case in length.
2. Confirm the new station has at least L_min of unobstructed straight pipe upstream of it and no elbow between it and the nozzle. If the run is too short, move the desuperheater, not the thermowell.
3. Orient the well horizontally in a horizontal run, or in the top half of the pipe, never at the invert. Set insertion depth so the tip sits between one-third and one-half of the internal diameter and away from the wall boundary layer.
4. Use the smallest sheath diameter the velocity and wake-frequency calculation allows, and verify the wake calculation at maximum steam velocity before ordering.
5. Install a second element 5-10 diameters further downstream, or a spare well at that station, as a permanent evaporation-completion check.
6. Add top-and-bottom skin thermocouples on the pipe just upstream of the first downstream elbow. A persistent bottom-minus-top delta is a running film and needs no interpretation.
7. Re-range the controller and set a minimum-superheat trip that closes spray when T2_meas - Tsat(P2) falls below the design margin.

Do not move on to loop tuning until every one of these is in place; a well-tuned loop on a wetted probe drives the failure faster.

Verification

1. Compare the relocated element with the downstream element. Both should read the same value within a few degrees. A downstream reading that is consistently higher means evaporation is still finishing between them — extend the length.
2. Step the spray valve open 10 percent in manual and watch the response. A dry probe gives a smooth first-order fall and a symmetric recovery. A step down followed by a hang at Tsat(P2), or a fast fall with a much slower rise, means liquid is still reaching the tip.
3. Repeat the step test at minimum steam flow with spray at its lowest controllable demand. This is where nozzle turndown fails and where a marginal location shows itself.
4. Confirm the top/bottom skin thermocouple delta upstream of the elbow stays below a few degrees through a full load ramp.
5. With the unit at minimum load and spray active, verify the superheat margin T2_meas - Tsat(P2) never falls below the design value for one full hour before returning the loop to automatic.

FAQ