Preheat and post-weld heat treatment — hydrogen, hardness, and reading the chart

Two thermal cycles bracket a weld: one before it and one after. They are specified separately, for different reasons, and both are routinely treated as formalities by people who have not had to explain a cracked joint.

Preheat

What it is for

Not to make the metal easier to melt. To slow the cooling rate.

A thick cold section acts as a heat sink and pulls the heat out of a weld very quickly. Fast cooling does two bad things at once:

Hydrogen is trapped. Hydrogen dissolves into the molten pool from moisture, oil, rust or damp electrodes. Given time it diffuses out. Cooled fast, it stays — and hydrogen in a hard structure under residual stress is the recipe for delayed cracking, which can appear days after the weld looked perfect.
A hard, brittle structure forms in the heat-affected zone.

Preheat attacks both by slowing the cooling. Which is why it is specified by material and thickness — the two things that set the cooling rate — and not by how cold the morning feels.

Interpass temperature

Preheat has a minimum, maintained throughout. Interpass also has a maximum.

Measuring it

A pipe joint on a run at grade wrapped in ceramic heating mats and insulation blankets bound with wire, with a bundle of thermocouple wires emerging from under the wrapping and running across the ground to a portable circular chart recorder standing on a frame nearby. The recorder door is open, showing the paper chart and its inked trace. A unit and workers are out of focus behind at dusk. — The wrapping, the thermocouple bundle and the chart, which are the three parts of the evidence. What the chart is worth depends entirely on where the other end of those wires is attached.

Post-weld heat treatment

When it is required

Driven principally by thickness, with material group setting where the threshold falls. B31.3 states requirements and exemptions below stated thicknesses; carbon steel is exempt well above where the chrome-moly alloys have long since required it.

Service can add its own requirement on top — sour service being the obvious case.

What it does

Holding the joint at temperature lets residual stresses relax, tempers the hard structure left by rapid cooling, and drives off remaining hydrogen.

Reading the chart

The record is a trace, and there are four things on it, not one.

What to read	Why
Heating rate	Too fast and the outside expands against a cold core, creating its own stresses. The spec controls it
Soak temperature	Within the specified band — and the band has a top as well as a bottom
Soak time	By thickness, and it starts when the coolest monitored point reaches temperature
Cooling rate	Controlled down to a stated temperature, then free

Reading only the peak and the hold time misses half of what the cycle was meant to control.

A temperature against time trace. The curve rises along a controlled heating ramp, holds through a soak band bounded by a minimum and a maximum, then falls along a controlled cooling ramp before being allowed to cool freely. Each of the four regions is annotated with what the specification controls there, and the thermocouple positions are marked on the soak section. — Read only the peak and the hold time and you have checked two of the four things the specification controls.

Thermocouples

The trace is evidence about the metal only if the sensors were on the metal, at the positions the procedure specified — including the coolest part of the assembly, which is what the soak time is counted from.

A beautiful chart from badly placed thermocouples records a cycle that part of the joint never experienced. Placement is part of the record, not a detail of the setup.

Hardness, and sour service

Where the service limits hardness — sour service above all, under NACE MR0175 and ISO 15156 — a hardness survey is the verification that the thermal cycles did their job.

Hardness is the practical proxy for susceptibility to sulphide stress cracking, and the heat-affected zone is where an uncontrolled cycle leaves a hard band. On those lines, preheat and PWHT stop being paperwork and become integrity items with a measurement attached.

See material degradation.

Welding to a spool that has already been treated

The auditable record

For PWHT: the chart trace, the thermocouple positions, the equipment and operator, the procedure, and — the one most often missing — the joints it covered, by weld number.

A chart that cannot be tied to specific welds proves a cycle happened somewhere. That is not the same as proving it happened to this joint, and at handover it is the difference between a closed line and an open question.

What to take away

Preheat slows cooling — against hydrogen cracking and a hard HAZ. Not to help melting.
Interpass has a maximum too, protecting toughness.
Measure preheat through the section, not beside the torch.
PWHT is driven by thickness and material group, with service on top.
It relaxes stress and tempers hardness — and repairs nothing.
Read the chart for ramp, soak band, soak time and cooling, not just the peak.
Thermocouple placement is part of the evidence; soak counts from the coolest point.
Hardness survey is the verification in sour service.
Welding to a treated spool undoes it locally — clamp instead.
Tie every chart to weld numbers.