PWHT Meaning in Welding: Post Weld Heat Treatment Guide

A pipefitter finishes a 12-inch carbon steel spool, bolts it into a line, and the hydrostatic test cracks a girth weld three days later. The radiography was clean, the weld procedure was qualified, and the inspector signed the paperwork — but nobody ever asked whether the spool needed post weld heat treatment. Months later the same shop gets a repair order that says “stress relieve to 1,200 °F before reinstallation,” and half the crew squints at the abbreviation like it is a foreign language. PWHT is one of those pieces of shop-floor shorthand that quietly decides whether a job is legal, on time, or a future failure report.

PWHT Meaning in Welding: The Definition Every Welder Should Know

PWHT stands for Post Weld Heat Treatment. In code language, it is the uniform reheating of a welded structure, or a defined portion of it, to a temperature high enough to relieve the major portion of the welding residual stresses, followed by uniform cooling. That is the AWS A3.0 wording, and it is the wording that matters when an inspector asks what you did to the weld.

Here is where the confusion starts. As Wikipedia puts it, PWHT is “often referred to as being any heat treatment performed after welding; however, within the oil, gas, petrochemical and nuclear industries, it has a specific meaning.” On a fabrication shop floor the phrase gets stretched to cover everything from a stress relief bake to a full solution anneal to re-ageing a welded aluminium bracket. Inside ASME and AWS codes, the meaning is narrow, and the narrowness decides whether a job is code compliant or a rework order waiting to happen.

The three heat treatments a welder actually meets

• Preheat happens before the arc is struck. It raises the base metal to a minimum temperature so the weld metal can cool slowly enough to avoid a hard, crack-prone HAZ.
• Interpass temperature is the maximum temperature the weld metal and HAZ are allowed to reach between passes. It keeps the heat input under control on thick, restrained joints.
• PWHT happens after welding is complete and the component has been inspected and often NDE’d. It is the deliberate, controlled heat cycle that relieves residual stress and does whatever metallurgical work the material needs.

Stress relief, or something more?

The single most important distinction is whether the treatment crosses a transformation temperature.

Thermal stress relief is performed below the A1 critical temperature. Nothing transforms. The microstructure stays exactly as the weld left it, and the only changes are stress relaxation and a little carbide precipitation. This is the classic PWHT for carbon steel piping and vessels.

A solution anneal is performed above A3, so the microstructure does transform. Austenitic stainless gets heated into the austenite field so chromium carbides dissolve, then it is quenched fast enough to keep that carbon in solution. Normalising sits in the same family: above A3, then air cooled to refine the grain structure. These are not interchangeable with stress relief, and running a stainless weld up to a carbon steel stress-relief temperature is a well-known way to destroy corrosion resistance. For aluminium, the equivalent after welding is solution treatment plus artificial ageing rather than PWHT, which is why aluminium MIG welding practice looks so different from carbon steel work.

Post-weld heat treatments compared

Treatment	What it actually does	Typical use
Thermal stress relief	Sub-critical heating below A1, relaxes residual stress, no phase change	Carbon steel pipe, vessels, structural weldments
Tempering	Reheats hardened steel to restore toughness and lower hardness	Quenched and tempered steels, alloy steel weld metal and HAZ
Normalizing	Heats above A3, air cools, refines the grain structure	Heavy carbon steel plate and forgings
Solution annealing	Heats above A3 and quenches, dissolving carbides	Austenitic stainless, some nickel alloys
Solution treat and age	Solution anneal then precipitation hardening	Aluminium (T6), age-hardenable stainless and nickel alloys
Dehydrogenation bake	Low-temperature soak to drive hydrogen out of the weld	Pre-shipment baking of high-strength steel
Intermediate stress relief	Partial relief between weld stages, below the final PWHT temperature	Thick weldments built up in several passes

Note the last four rows. TWI is explicit that PWHT is a specific term covering stress relief and tempering, and should not be confused with other post-weld heat treatments such as ageing of aluminium alloys, solution treatment of austenitic stainless, or hydrogen release. Shops that blur those terms end up with a purchase order, a rework order, or a certificate that says something the welding procedure never said.

Two practical consequences

First, not every weld needs PWHT. As ESAB’s university guidance points out, the requirement depends on material, thickness, service conditions and the governing industry code. A 3 mm mild steel bracket in a shop frame does not need it. A 40 mm P-No 5 pipe spool in sour service probably does.

Second, PWHT is not a free-standing decision. It is a qualification variable. Adding PWHT, removing it, or heat treating outside the qualified time and temperature range changes the procedure, and under ASME Section IX and ISO 15614-1 that change forces requalification of the welding procedure. This is the part that bites hardest: the weld was sound, the radiograph was clean, and the paperwork still says the spool is non-compliant because the final heat cycle did not match the PQR.

Terminology confusion is not a pedantic problem. It is a real and expensive source of field rework, and it starts with a crew hearing “PWHT” and not being sure whether it means stress relief, an anneal, or a bake. Get the definition straight, then read the code table for the actual job.

Why PWHT Exists: Residual Stress, Hydrogen and Distortion

Welding is a heat treatment in miniature. A molten pool, a hardened bead and a heat-affected zone (HAZ) that changes metallurgically in the space of a few minutes, all held in place by cold, rigid parent metal. Everything PWHT does is a direct response to that. If you understand what the weld is doing to the metal while it cools, the reason for post weld heat treatment becomes obvious.

How welding locks in residual stress

As a weld cools, the molten metal tries to contract. It is restrained by the surrounding cold steel, which has already stopped shrinking. The result is a self-balancing set of internal forces that nobody applied and nobody can remove without heating the part again. This is residual stress, and it is the primary reason PWHT exists.

The critical detail is that yield strength climbs as steel cools. While the metal is hot it can relieve stress by creeping or plastically deforming. As the temperature falls, yield strength rises and that relief mechanism shuts down. The stress that was manageable at 600 °C can be locked in place by the time the weld reaches ambient. The upshot is that residual stress in the weld region can approach the material’s proof strength at ambient temperature, and in a restrained joint it frequently does.

Why that stress causes trouble

Three failure mechanisms follow directly from it:

• Brittle fracture. A weld in a thick or highly restrained section can carry very high tensile residual stress across the notch-like defects of a bead. Combined with the low temperatures that make some steels brittle, that combination is the classic route to a sudden, brittle-looking crack. PWHT reduces the risk by relieving residual stress and improving HAZ toughness, as noted in this overview of post weld heat treatment.
• Stress corrosion cracking. In caustic, chloride, sulphide or hydrogen-containing service, a tensile residual stress overlay turns a corrosive environment into a cracking environment. Removing that overlay by PWHT measurably reduces the risk of stress corrosion cracking.
• Dimensional instability. Residual stress is a stored energy. When the component is machined, the stress redistributes and the part moves. It can also move during start-up shake-down, thermal cycling or pressure cycles. Either way, the item ends up outside acceptable dimensions, and the fix is usually to let the stress go before the final machining pass, not after.

Hydrogen and untempered martensite

Hydrogen is the other half of the problem. Hydrogen picked up from moisture, coatings, contaminated filler metal or a damp environment can diffuse into the weld and HAZ, and where it combines with susceptible microstructures it produces delayed hydrogen cracking. Heating the part accelerates hydrogen diffusion so the gas can escape rather than sit in the hardened zone. This is also why consumable selection matters: choosing alloy and low-hydrogen electrodes with properly baked storage reduces diffusible hydrogen in the first place, which PWHT then finishes off.

For quenched and tempered steels, the HAZ next to the fusion line can contain untempered martensite, a very hard and brittle structure. PWHT at the correct temperature tempers that martensite, restoring ductility and toughness. A general overview of preheat and post weld heat treatment mechanics is available in this ESAB technical overview.

Where PWHT has limits

Two honest caveats keep expectations realistic. First, PWHT has little beneficial effect on fatigue performance unless the resulting stress state is mostly compressive; if the fatigue driver is residual tensile stress in a badly detailed joint, PWHT will not save it. Second, PWHT is not free of side effects. Run at too high a temperature, it can reduce strength, coarsen grain, over-temper the weld metal or, in stainless grades, cause sensitisation and loss of corrosion resistance. Run at the wrong temperature entirely, it can do more harm than the residual stress it was meant to remove.

That is why the temperature band is not negotiable. TWI’s research on post weld heat treatment effects on welding residual stresses indicates roughly these soak ranges:

• Carbon manganese steels: 550 to 620 °C
• Creep-resistant steels: 650 to 750 °C
• Stainless steels: 800 to 850 °C

Code Requirements: When ASME and AWS Actually Make PWHT Mandatory

Welding codes never say “stress relieve this spool because it looks stressed.” They work through a chain: identify the material, find its P-Number and Group Number, look up the PWHT temperature band for that material, then test whether an exemption applies. Once you understand that chain, the question “does this job need PWHT?” stops being a guess and becomes a lookup. Thebackground on post weld heat treatment is well covered elsewhere, so this section focuses on how the codes themselves reach the answer.

Step one: P-Numbers and Group Numbers

The whole system starts in ASME Section IX, in Table QW-420 for welding and Table QB-420 for brazing. A P-Number is the broad material family: P-No 1 is carbon steel, P-No 4 low-alloy steel, P-No 5A and 5B Cr-Mo and Cr-Mo-V grades, P-No 6 and P-No 7 higher Cr-Mo steels, P-No 8 austenitic stainless, P-No 11A martensitic stainless, P-No 15E nickel alloys, P-No 62 the 9Cr-1Mo-V family. Inside each P-Number, the Group Number splits materials by strength level, chemistry and weldability.

Both numbers matter, because a PWHT chart may give a different temperature band for different groups inside the same P-Number. P-Numbers also drive preheat limits and impact test exemptions, which is why the same grade of steel can carry different preheat notes on two different jobs. Background on how these requirements stack up in practice is covered in heat treatment of welded joints.

ASME B31.3: process piping

For process piping, ASME B31.3 is the most prescriptive code you will meet. Table 331.1.1 gives the minimum holding temperature range and the minimum holding time for each P-No and Group No. The bands look like this:

P-No	Material family	Holding temperature range (B31.3 Table 331.1.1)
P-No 1 and P-No 3	Carbon steel	595–650 °C (1,100–1,200 °F)
P-No 4	Cr-Mo low-alloy steel	650–705 °C
P-No 5A / 5B	Cr-Mo and Cr-Mo-V alloy steel	675–760 °C
P-No 6	1% Cr alloy steel	760–800 °C
P-No 7	2.25Cr-1Mo alloy steel	730–775 °C
P-No 11A	Martensitic stainless steel	550–585 °C
P-No 15E	Nickel alloy	705–775 °C
P-No 62	9Cr-1Mo-V alloy steel	540–595 °C

Table 331.1.3 then lists the exemptions from mandatory PWHT. The most commonly used one covers P-No 1, all Group Nos.: PWHT may be omitted when preheat of 95 °C (200 °F) was applied to nominal material thickness over 25 mm (1 in.), and when multiple-layer welds were used on nominal thickness over 5 mm (3/16 in.). That is a narrow path, and it is a genuine path: thin carbon steel line pipe welded with multi-layer fillet and butt welds at modest preheat often never sees a furnace. For P-No 8 austenitic stainless steels, PWHT is neither required nor prohibited, which is why stainless work is often heat treated only when the designer or the service condition demands it.

Pressure vessels, boilers and structural steel

ASME Section VIII-1 points to Table UCS-56 for the mandatory PWHT requirements, and UCS-56.1 offers an alternate route: a lower holding temperature paired with a longer hold time. Section VIII-2 (alternative rules) handles the same ground through Table 6.8. In1984, J.L. Carpenter published “A Guide for Determining Postweld Heat Treatment” in the ASME Journal of Pressure Vessel Technology, building decision trees for Section VIII vessels because “uncertainties occasionally arise” when deciding whether PWHT is required. It remains the classic reference when a code interpretation turns into an argument.

ASME B31.1 covers power boiler and external piping, with boiler PWHT requirements in PW-39. The important distinction: the PW-39 rules govern boiler parts only. External piping attached to a boiler is evaluated under the piping code, not under the boiler rules, and mixing the two is a classic paperwork error.

Structural steel has its own logic. AWS D1.1 addresses preheat and postweld heat treatment in Clause 5.2, with minimum holding times set by Table 5.2 based on material thickness. Here the purpose is usually dimensional stability and metallurgical property restoration in heavier sections, rather than pressure containment. Other industries layer their own consensus specifications on top, which our guide to aerospace welding standards and practices touches on, and project specifications frequently add requirements beyond the base code.

The WPS, the designer and the final call

PWHT is an essential variable in procedure qualification. If your PQR was qualified with a postweld heat treatment condition, that condition belongs in the WPS, and welds made to a WPS that does not cover the PWHT actually performed are not properly qualified. A designer may also impose PWHT where the code minimum says it is optional, and that requirement travels with the drawing and the specification through fabrication, erection and pressure testing.

Temperature and Hold Time: How to Read a PWHT Chart

A PWHT chart looks like a simple lookup, but two conditions must be met at once: the metal must actually reach the soak temperature range, and it must stay in that band for the full holding time. Miss the temperature and the weld was never heat treated; miss the time and you get the same result. Find the row by P-No and Group No., then read across to the control thickness. TWI’s guide to heat treatment of welded joints shows how those ranges shift from carbon steel to stainless and nickel alloys.

The Standard Holding Rule

ASME B31.3 Table 331.1.1 sets the baseline hold: one hour per 25 mm (1 in.) of control thickness, with a 15-minute minimum for material up to 50 mm (2 in.) thick and a 30-minute minimum above 50 mm. A 1 in. schedule 40 pipe gets one hour of soak. A 6 in. spool gets six hours, plus the ramp up and the controlled cool, and that cycle is what eats a week of schedule.

Some P-Numbers step up past 2 in. For P-No 1 (carbon steel), P-No 3 and P-No 4, the hold becomes 2 hours plus 15 minutes for each additional 25 mm over 50 mm. P-No 7 (high-chromium stainless) has its own curve: 1 hour per 25 mm up to 125 mm (5 in.), then 15 minutes for each additional 25 mm above that. Read the row for your own P-No and Group No., not the nearest-looking row above it. Consumables for those low-alloy rows are a separate problem, covered in our Alloy Steel Welding Electrodes guide.

Control Thickness Is Not Nominal Thickness

The thickness that drives the clock is the control thickness: the governing thickness of material the heat must work through at the joint, generally the greater of the component wall thickness and the thickness of the weld joint itself, including reinforcement. It is not the pipe schedule and not the nominal wall. Where the joint carries extra metal, the number moves. A nozzle reinforcing pad raises the effective thickness, and a socket weld adds the socket depth and weld buildup on top of the pipe wall, either one pushing you into a longer hold than the schedule suggests. AWS D10.10 defines the heated and control zones where those temperatures are verified, which is where thickness questions usually begin.

The Lower Temperature, Longer Hold Trade

Sometimes the minimum soak temperature is a real problem, because the drawing calls for a coating, a seal, a dissimilar joint or a tolerance that cannot take the full heat. ASME B31.3 Table 331.1.2 offers a documented way out: come in 30 °C (50 °F) below the minimum and hold 2 hours, 55 °C (100 °F) below for 4 hours, 85 °C (150 °F) below for 10 hours, or 110 °C (200 °F) below for 20 hours. The last two reductions are allowed only for P-No 1 Group Nos. 1 and 2 at thicknesses of 25 mm (1 in.) or less, because dropping that far below the chart temperature on thicker carbon steel or on alloy steel can leave the weld short of the stress relief the code assumes. Whichever route you take belongs in the heat treatment procedure and the inspection plan, not a decision made the night before the furnace slot.

Cooling Notes and Material Traps

• P-No 7, Group Nos. 1 and 2: the cooling rate shall not exceed 55 °C (100 °F) per hour above 650 °C (1,200 °F), after which cooling must be fast enough to prevent embrittlement. Leaving the part in the furnace to cool is not automatically compliant.
• P-No 11A: the cooling rate shall be greater than 165 °C (300 °F)/h and no more than 315 °C (600 °F)/h, a controlled cool with a floor and a ceiling. A flat overnight cool in a closed furnace will usually miss it.
• P-No 15E: the Ni + Mn content of the filler metal shall not exceed 1.2% unless the designer specifies otherwise, and the maximum temperature shall not exceed 800 °C (1,470 °F). Exceed 800 °C and the weld plus the entire area affected by the PWHT must be removed and redone.
• BS EN 13445: limits gradients rather than soaks, with a maximum temperature difference of 150 °C within 4,500 mm below 450 °C and 100 °C within 4,500 mm above 450 °C.

None of these charts apply to aluminium, where the answer is solution treatment and artificial ageing rather than stress relief, and over-ageing a T6 assembly costs you strength. Aluminum MIG Welding: The Complete Practical Guide covers how non-ferrous assemblies are heat treated.

Before the furnace fires, check who signed the numbers. This summary of heat treatment of welded joints is a useful sanity check on the ranges you are about to weld to. Verify the P-No and Group No., the control thickness, the hold time and the cooling rate against the governing code and the project specification, then get the heat treatment procedure approved before the first weld is made. Once you are reading a chart, the decision has already been made for you.

Furnace PWHT vs Local PWHT: Methods, Equipment and Trade-offs

Once the code, the drawing and the WPS agree a weld needs post weld heat treatment, the shop faces a question with nothing to do with metallurgy: how do you actually get heat into the joint? Either the component goes into a furnace, or the heat goes to the weld. Both are standards-compliant and both end with a chart of time-temperature data, but they are not equivalent.

Furnace PWHT treats the entire component, or a major portion of it, as a single thermal mass. The part goes into an enclosed gas-fired or electric furnace, ramps to soak temperature at a controlled rate, holds, and cools in the closed furnace. The weld gets no special treatment: every piece of metal sees the same thermal cycle, so residual stress redistributes evenly and the peak comes down. That is why it is the preferred method and the yardstick for local heating.

The price of that uniformity is logistics: furnace capacity, crane access, transport to and from the heat treater, and the calendar cost of a long cycle. A part that will not clear the furnace door cannot be furnace PWHT’d at all.

Local PWHT: heating a band around the weld

Local PWHT heats a controlled band around the weld and leaves the rest of the component alone. The heat comes from gas-fired radiant heaters clamped in a ring around the joint, electric resistance heating with current driven through the part, induction heating, or infrared and oven-style heating blankets. It is usually the only realistic option for a spool already in the loop, a field repair, or a joint in a plant that cannot be shut down. It is also the method most often done badly, because the heat is not where the material wants it to be.

The vocabulary that makes local heating make sense

These terms come from AWS D10.10 (2021), the recommended practice for local heating of welds in piping and tubing:

• Soak band — the zone of material that must reach soak temperature, measured either side of the weld centre line.
• Heated band — the wider zone the heaters cover, so temperature tapers off instead of stepping down.
• Gradient control band — the region between soak band and heated band edge, where the gradient is kept within allowable limits.
• Control zone — the region the control thermocouple is deemed to represent.

Standard	Band requirement
ASME Section VIII	Soak band of twice the weld thickness, or 50.8 mm either side of the weld, whichever is the lesser
ASME B31.3	Weld width plus 25.4 mm either side of the weld
BS EN 13445	No soak band width specified; heated band of 5√Rt centred on the weld (R = inside radius, t = thickness)

Why local PWHT is harder to trust

TWI is blunt about it: local PWHT can leave residual stress higher than the levels recommended for furnace heat treatment, so expert advice should be sought or conservative assumptions used. Two details explain why. A heated band edge temperature of at least half the soak temperature is generally assumed to give an acceptable axial gradient, and anything less drags a steep gradient through the joint. Unbalanced or rapid heating and cooling creates thermally induced stresses that can exceed yield and generate fresh residual stress as the part cools.

Geometry matters even more on large pipe. LANL’s GWS 1-08 calls for heater elements giving four-quadrant control above NPS 12 (DN 300), because one ring of heaters will not heat the bottom of a large pipe evenly. TWI also offers a very approximate rule of thumb: for flat plate the heated band should be at least twice the length of the weld, though practical constraints often prevent it.

A shop scenario: 40-foot vessel versus in-line spool

Two jobs in the same shop, the same week:

Job A is a 40-foot carbon steel vessel with 32 mm walls, longitudinal and girth welds, and a design that calls for stress relief. It is a candidate for furnace PWHT: rigged, trucked to the heat treater, loaded, soaked, cooled, returned, inspected and back on the crane. It costs real money and real days, but one chart covers the whole soak, the gradient across the joint is trivial, and the inspector gets a clean, defensible record.

Job B is a 10-inch line spool with a single girth weld already wired into the loop. Shipping it out means breaking welds, draining the line and re-terminating. It gets local PWHT instead: a ring of radiant heaters, thermocouples on the weld centre line and at the soak band edges, a soak band of weld width plus 25.4 mm each side, and a controlled cool under blanket. At NPS 10 there is no four-quadrant requirement, and the job comes back in days rather than weeks. The caveat sits in the file: residual stress around that weld is higher than furnace PWHT would have left, and the record shows a narrow band reaching temperature rather than a whole body of steel doing it.

Both jobs can be perfectly legal. Furnace PWHT removes stress across the whole component; local PWHT manages stress in a zone and the gradient around it. Choose furnace when the part can move, and plan local heating with real care when it cannot.

For thin-wall aluminium work, local heating is usually the wrong tool entirely, since raising the temperature risks distortion and over-ageing of the joint. If you are new to those alloys, our guide to aluminium MIG welding covers the filler, gas and technique choices that matter before heat is ever considered.

Running a PWHT Job: Thermocouples, Heating Rates, Soaking and Cooling

A compliant PWHT is not a “put it in the oven and call it done” operation. It is a documented thermal cycle with a written procedure, a charted record and defined checkpoints. The post weld heat treatment definition only works if somebody measured the part while it heated, held it in the required band and controlled the way it came down.

1. Pre-job: paperwork, cleanliness and support

Before anything gets hot, three things get confirmed. First, the WPS and its supporting PQR must have been qualified with PWHT. Heat treatment is an essential variable, so a procedure qualified without it does not cover a PWHT joint, and one qualified with it cannot be swapped for one that was not. Second, all welding must be complete, since PWHT is not a repair allowance for unfinished work. Third, the component is cleaned of oil, grease, thread lubricants, paint and any lead- or sulfur-containing substance before it sees heat, because those vapors attack furnace elements and contaminate the part.

Then the component is blocked and supported, where jobs often go sideways. As metal temperature rises, yield strength falls, so at soak temperature the part may be physically unable to carry its own weight. A vessel free to sag in the furnace will come out with a distorted shell. Use shaped trestles sized for the geometry, and where internal supports are needed, use a material with a similar coefficient of thermal expansion so the supports move with the shell rather than fighting it. Heat treatment of welded joints covers the same reasoning in more depth.

2. Instrumentation: getting a reading you can defend

Type K and Type N thermocouples are the common choices for carbon and low alloy work, with Type J on lower temperature ranges. Attachment method matters when an inspector asks how you know the metal was actually at temperature.

• Capacitor discharge (CD) welding gives the most accurate attachment, with the two wires typically separated by 3-4 mm. The discharge welds the junction directly onto the metal, so it stays put and reads true through the whole cycle.
• Mechanically attached wires need heat-resistant putty or insulation so the heater cannot overheat the junction and give you a false reading.

Placement matters as much as attachment. Put a thermocouple on the weld centre line, one at each edge of the soak band and one at each edge of the heated band. Where geometry allows, locate them on the surface opposite the heater, since that surface sees true metal temperature rather than radiant heat. Double up on critical points: a single open thermocouple on night shift should not end the job. The AWS D10.10 heat treatment standard defines the soak band, heated band and gradient control terminology these requirements sit on.

Minimum pipe coverage is usually specified like this:

Pipe diameter	Control zone	Recording thermocouples
Under 100 mm	One	One
100 mm to 200 mm	One	One at 12 o’clock and one at 6 o’clock
Above 250 mm	One	One at 12, 3, 6 and 9 o’clock

Bigger pipe holds more thermal mass, so more of the circumference gets monitored.

3. Heating and cooling rates

Rate control matters as much as soak temperature, because thermal shock and uneven expansion cause distortion. Under ASME guidance on post weld heat treatment practice, heat-up and cool-down above 600 °F (315 °C) are capped at 400 °F/h (205 °C/h) for thicknesses of 1 in. (25 mm) or less, and 400 °F/h divided by the thickness of the thicker member above that, with a floor of 100 °F (38 °C) per hour.

Parameter	ASME Section VIII and III	AWS D1.1
Loading the furnace	Component at ambient	Furnace not above 600 °F (315 °C) when the assembly goes in
Rate control threshold	Above 600 °F (315 °C)	Above 600 °F (315 °C)
Maximum heat-up rate	400 °F/h (205 °C/h) up to 1 in. (25 mm); 400 °F/h divided by thickness of the thicker member above that	400 °F per hour divided by the maximum thickness of the thicker part, in inches
Minimum rate	Never less than 100 °F (38 °C) per hour	Same practical floor
Uniformity	Max 250 °F (140 °C) variation within any 15 ft (4.5 m) of weld length	Max 250 °F (140 °C) variation within any 15 ft (4.5 m) interval

4. Soak, then controlled cooling

Once the part reaches the required band, the soak begins and the chart proves it. Hold the specified time at the specified temperature, with control and recording thermocouples both in the zone. Cooling follows the same discipline, in a closed furnace or chamber where possible; a part that comes out to air with lagging around it has not had a controlled cool-down.

5. After the heat

Parts must be realigned, warpage corrected and bolts retightened to the vessel drawing after PWHT. PWHT normally follows completion of the welds and generally precedes final NDE and pressure testing, though owner specifications often call for radiography or ultrasonic testing before PWHT with final UT afterward, and magnetic particle testing before PWHT repeated after it. Sectors with their own rules, such as aerospace welding, add their own sequences on top.

The Ten Most Common PWHT Mistakes and How to Avoid Them

PWHT rarely fails because of a spectacular error. It fails because a number was never checked, a footnote never read, a chart filed in the wrong binder. These ten findings keep surfacing in audits.

1. Treating PWHT as optional because the weld passed radiography. Radiography proves the joint is sound, not the residual stress, hydrogen or the condition of the heat-affected zone. A girth weld can be clean on film and still leave a code-required spool out of compliance. PWHT is a deliverable on the drawing, not a reward for good welding.
2. Missing the requirement because nobody read the code note. Exemption tables such as ASME B31.3 Table 331.1.3 and Section VIII UCS-56 carry footnotes that change the answer, and designer specifications often demand PWHT on a joint the code would exempt. The code is the floor, the specification the ceiling. Put the PWHT requirement on the drawing, the WPS and the ITP before the first arc.
3. Using the wrong soak temperature band for the P-No. A generic "stress relieve to 1,200 °F" callout is not a code. TWI’s guidance on stress versus temperature behaviour shows why: carbon manganese needs roughly 550 to 620 °C,
   

   The Ten Most Common PWHT Mistakes and How to Avoid Them

   PWHT rarely fails because of a spectacular error. It fails because a number was never checked, a footnote never read, a chart filed in the wrong binder. These ten findings keep surfacing in audits.

   1. Treating PWHT as optional because the weld passed radiography. Radiography proves the joint is sound, not the residual stress, hydrogen or the condition of the heat-affected zone. A girth weld can be clean on film and still leave a code-required spool out of compliance. PWHT is a deliverable on the drawing, not a reward for good welding.
   2. Missing the requirement because nobody read the code note. Exemption tables such as ASME B31.3 Table 331.1.3 and Section VIII UCS-56 carry footnotes that change the answer, and designer specifications often demand PWHT on a joint the code would exempt. The code is the floor, the specification the ceiling. Put the PWHT requirement on the drawing, the WPS and the ITP before the first arc.
   3. Using the wrong soak temperature band for the P-No. A generic "stress relieve to 1,200 °F" callout is not a code. TWI’s guidance on stress versus temperature behaviour shows why: carbon manganese needs roughly 550 to 620 °C, creep-resistant steels 650 to 750 °C, stainless 800 to 850 °C. Too low leaves stress locked in; too high coarsens grain and cuts toughness.
   4. Running the hold time against nominal thickness. The rule is one hour per 25 mm (1 inch) of control thickness with a 15-minute minimum, and control thickness is not always the nominal wall. Reinforcement, a socket weld, a thicker mating part or a nozzle set changes the number. "It’s a 12-inch spool" is not an answer to the inspector.
   5. Cooling too fast outside the furnace. Inside a closed furnace the mass of the component controls the cool; outside it, natural convection does, and a long spool’s ends cool faster than its middle. ASME B31.3 caps the cooling rate for P-No 7 materials at 55 °C/h above 650 °C. An uncontrolled cool can leave a hard, untempered HAZ and bow the part.
   6. Falling short on the soak band during local PWHT. Local heating relieves stress only where metal reaches soak temperature. ASME VIII wants a soak band of twice the weld thickness or 50.8 mm either side, whichever is lesser; B31.3 wants the weld width plus 25.4 mm either side. Miss that band and cold metal outside fights the hot zone. Definitions are in AWS D10.10.
   7. Bad thermocouple attachment. A junction taped over mill scale, welded to a rough spot or buried in a drilled hole reads the error, not the metal. Direct attachment, ideally capacitor discharge welded, on the weld centreline plus soak band and heated band edges is the practice. Doubling up thermocouples, with one on the opposite surface, guards against one failed element.
   8. Insufficient support and bracing. PWHT removes locked-in stress and the joint then moves. A long free-standing spool will bow. Block and brace it before it heats, align it, and record the alignment before and after. Warpage found at final NDE is expensive. Warpage found after machining is a rebuild.
   9. Heating stainless to carbon-steel temperatures, often irreversible. Carbon steel soak temperatures put austenitic stainless squarely in the sensitization range, where chromium carbides precipitate at grain boundaries and the HAZ loses corrosion resistance. Austenitic stainless weldments generally do not require PWHT, and corrosion there is often localised in the HAZ. Haynes International adds: stress relief at carbon steel temperatures is normally ineffective for nickel-base alloys and can precipitate secondary phases that harm corrosion resistance; for most alloys PWHT in the 1,000 to 1,500 °F (538 to 816 °C) range should be avoided, and the only acceptable treatment for solid-solution strengthened alloys is a full solution anneal. Where the service environment attacks sensitized areas, solution anneal at the annealing temperature to redissolve the carbides.
   10. Welding outside the qualified WPS, then heat treating to "fix" it. PWHT is an essential variable in procedure qualification. A WPS and PQR qualified without PWHT stay unqualified because the part went through the furnace. Your options are requalification, or removal of the weld and a redo: a gouging and rewelding job with its own qualified procedure, not a touch-up.

   Note: Thermal stress relief is not free. It can reduce strength and toughness in the HAZ of low-carbon, high-strength steel weldments, and can contribute to hydrogen embrittlement, which matters when picking consumables for an alloy steel joint with preheat and soak. Bimetallic and multimetallic assemblies create thermal stress problems because the two metals want different temperatures, and large wall-thickness variations develop thermal stresses as thick and thin sections heat and contract differently. Some components are too large for the available furnaces, a planning problem, not a heat treatment one. Aluminium is another special case: it needs re-aging or solution plus artificial ageing to restore properties, and over-ageing cuts strength, so our aluminium MIG welding guide does not transfer to a steel furnace cycle.

   Quick Note: Run this self-audit before the spool goes to the furnace.

   • WPS and PQR qualified with PWHT; specified range and hold time match the code chart for the P-No and Group.
   • Hold time calculated from control thickness, with the calculation on the heat treatment procedure.
   • Soak band, heated band and gradient control band widths on the thermocouple sketch, checked against the actual weld.
   • Thermocouples direct attached, placement sketched, spare set and spare recorder on site.
   • Component cleaned of oil, grease, paint and thread lubricant, then blocked and braced.
   • Cooling method in writing, rate limit included, charts and calibration certificates ready.

   PWHT by Material and Industry: Piping, Vessels, Structural Steel, Stainless and Aluminium

   Post weld heat treatment is not one practice with one rulebook. Requirements shift with the material, the governing code, the service and the wall thickness. Line pipe in a refinery, a vessel shell, a bridge girder, a stainless skid spool and an aluminium trailer deck share the same four letters and little else except the worry about residual stress. Know your world before you read a chart.

   Quick Note: Esab groups post-weld heat treatments into four categories: stress relieving, obtaining a suitable metallurgical structure (tempering), normalizing after welding (for example, refining a coarse electroslag weld structure), and regaining original properties in precipitation-hardening aluminium alloys. Your category decides the code and the temperature chart that apply.

   Pressure piping: ASME B31.1 and B31.3

   Process piping is the most prescriptive corner of the industry. B31.3 gives each material a P-Number and a Group Number, then points to a holding temperature range and a minimum hold time. Two traps catch people out: the definition of control thickness, and exemption tables that can remove PWHT from a joint that would need it. Exemptions apply only when the designer did not call for PWHT and the qualified WPS excludes it, so read the specification and the procedure first.

   Materials with no listed P-Number still have a rule: ASME B31.3 Note (8) requires heat treatment within 14 days of welding, hold time increased by 1.2 hours for every 25 mm (1 in.) of thickness above 25 mm, and cooling to 425 °C (800 °F) at no more than 280 °C (500 °F) per hour. That window catches the common reality of a spool sitting in the yard for a month.

   Pressure vessels: ASME Section VIII Division 1 and Division 2

   Division 1 handles vessels through Table UCS-56 and UCS-56.1, the latter an alternate route for dissimilar joints whose two materials cannot tolerate the same soak temperature. Division 2 is stricter at higher design stresses, and Table 6.8 lists where PWHT is mandatory rather than discretionary. The hard cases are thick shells and heavy nozzle openings: a 150 mm shell needs a slow ramp and a long soak, and nozzle reinforcement welds may still need heat treatment after the vessel is otherwise complete, meaning a partial furnace load or a local band around an already machined opening. Plan that rigging early; it can rewrite the erection schedule.

   Structural steel: AWS D1.1

   AWS D1.1 splits heat treatment into two jobs. Stress relief for dimensional stability uses the minimum hold time from Table 5.2, based on the thicker part’s thickness, holding at 1,100 to 1,200 °F (600 to 650 °C) after the mean temperature is reached. The second is metallurgical property restoration, for quenched and tempered steel where welding locally untempered the heat-affected zone. Clause 5.2 then requires the assembly to be held within specified limits after the maximum temperature of 1,100 °F (600 °C) is reached, with no difference greater than 150 °F (85 °C) between the highest and lowest temperature during the hold. If a fabricator quotes stress relief on a Q&T girder, ask which problem they are solving. TWI guidance on heat treatment of welded joints is a reminder that target temperatures track the parent metal.

   Power and nuclear: ASME Section I and Section III

   Section III turns PWHT into a documented heating and cooling campaign with controlled rates, gradient limits and thermocouple coverage. On large pipe that means heating the full circumference in controlled zones so no quadrant lags. Section I applies its rules to boiler parts only; external piping falls under B31.1 or B31.3 instead. Shops in this niche, including Alamo Welding & Boiler Works, live with that distinction weekly.

   Stainless steel and nickel alloys

   For austenitic stainless, PWHT is often prohibited outright. Heating 304 or 316L into the 600 to 650 °C range carbon steel likes can precipitate chromium carbides at grain boundaries, which is sensitisation, and it costs the corrosion resistance that justified the alloy. The route is a solution anneal followed by rapid cooling, above the transformation range, not stress relief. Duplex solution anneal windows are narrow and grade specific: S31803 and S32205 at 1,020 to 1,100 °C (1,870 to 2,010 °F), S32550 at 1,040 to 1,120 °C, S32750 at 1,025 to 1,125 °C, other duplex alloys at 980 to 1,040 °C, each followed by rapid cooling.

   Nickel-base alloys have their own discipline. Haynes International recommends welding nickel alloys in the annealed condition, and solution annealing material with more than 7% cold work before welding to avoid cracking. That is crack prevention rather than PWHT, but the same condition-and-heat-treatment conversation. For dissimilar or clad joints, the Alloy 20 welding procedures guide shows how those joints are handled in practice.

   Aluminium

   Aluminium does not get stress relief in the carbon steel sense; its post-weld treatment restores properties the weld’s thermal cycle disturbed. For 6000- and 7000-series material in the T6 or T73 condition, the options are a re-age after welding or a full solution and artificial ageing cycle, recovering strength lost to over-ageing in the heat-affected zone. Over-ageing pushes the alloy past peak strength, so a cycle that runs too long or too hot can leave the joint weaker than the parent metal. Getting filler and preheat right from the start, as our aluminium MIG welding guide explains, often removes the need for a post-weld ageing cycle altogether.

   Industry	Governing code	What PWHT looks like
   Process piping	ASME B31.3 / B31.1	P-Number and Group tables, control thickness, exemptions, Note (8) for unlisted materials
   Pressure vessels	ASME Section VIII Div 1 / Div 2	UCS-56 and UCS-56.1, Table 6.8 mandatory cases, extra work at thick shells and nozzle openings
   Structural steel	AWS D1.1	Clause 5.2, Table 5.2 hold times, separate rules for Q&T steel
   Power and nuclear	ASME Section I / Section III	Documented rates and gradients, multi-zone heating on large pipe, boiler-only rules in Section I
   Stainless and nickel	Code plus material specification	Solution anneal or solution plus age; PWHT often prohibited to avoid sensitisation
   Aluminium	Code plus material specification	Re-age or solution plus artificial ageing; watch for over-ageing

   Note: Before a single weld is cut, confirm four things: which code governs the assembly, which P-Number or material condition applies, what the control thickness is, and whether the WPS was qualified with PWHT. Each one can change the answer, and each is far cheaper to resolve on paper than in a furnace.

   Planning, Costs, Documentation and Safety for a PWHT Job

   PWHT is not a checkbox at the end of fabrication. It is a scheduled operation with a furnace booking, a rigging plan, a thermocouple layout, a chart package and a safety plan, and arranging it after the welders finish is the biggest scheduling mistake. By then crane slots are booked and the furnace queue is three weeks deep. On heavy vessels PWHT is routinely the longest lead item, since a cycle means ramp-up, soak, controlled cooling and cool-down that cannot be compressed without risking non-compliance.

   Plan PWHT before the welding starts

   PWHT belongs on four documents before welding begins:

   • Fabrication drawing, with the heat treatment callout and the owner’s temperature range.
   • WPS and PQR, where PWHT is an essential variable the qualification covers.
   • Heat treatment procedure (HP), distinct from the welding procedure.
   • Inspection and test plan (ITP), with hold and witness points agreed before anyone lights a torch.

   Quick Note: A familiar owner requirement calls for the heat treatment procedure to include a vessel sketch showing thermocouple locations and types, attachment method, insulation details, support methods, bracing against distortion and the heating control system. Without that drawing, it is not really an HP.

   Where the money actually goes

   Quotes for PWHT look thin and balloon later. The real cost drivers:

   • Furnace time, charged hourly and often booked at a minimum.
   • Fuel or electricity, plus standby time waiting for a rigging slot.
   • Rigging and transport, maybe a larger crane and route surveys.
   • Thermocouples, extension leads, recorders, calibration, consumables.
   • Post-PWHT machining to remove weld caps and blend the surface.
   • Re-inspection if dimensional checks or NDE must be repeated.

   That last pair is where savings evaporate: a spool that misses tolerance after PWHT can cost more to rework than the furnace charge.

   Build the documentation package as you go

   The heat treatment record is what an auditor reads months later. Assemble it in one place:

   1. WPS and PQR showing the PWHT range qualified.
   2. Mill certificates and traceability for base metal and consumables.
   Heat treatment procedure, thermocouple sketch and insulation details.
   3. Heater and control system details, including the chart recorder setup.
   4. Calibration certificates for thermocouples, recorders and instruments.
   5. Signed time and temperature charts from the run.
   6. Support and bracing records for how the component was held.
   7. Post-PWHT dimensional report.

   Keep charts and calibration certificates with the fabrication record: when a client or authority reviewer asks two years later how stress relief was verified, that binder is the answer. The Los Alamos National Laboratory welding engineering standards show the detail inspectors expect.

   Safety: the fume is the hazard people forget

   The arc is a chemical hazard and the furnace needs guarding, interlocks and controlled access. But the exposure crews underestimate is welding fume, and stainless gives the most toxic fume of all, since its filler and base metal contain chromium that can convert to hexavalent chromium during the arc.

   OSHA sets the permissible exposure limit for hexavalent chromium at 5 µg/m³ as an 8-hour time-weighted average, action level 2.5 µg/m³ (29 CFR 1910.1026 and 1926.1126). Welding fume also falls under 29 CFR 1910 Subpart Q, 1926 Subpart J and 1915 Subpart D; the OSHA welding fact sheet is the readable version.

   NIOSH’s engineering controls study found total welding fume ranged from 2 to 60 mg/m³ without local exhaust ventilation, dropping to 3 to 13 mg/m³ with it. Field controls ranged from completely ineffective, in the case of a canopy hood, to an 83% cut with a fume extraction gun, and pulsed GMAW cut fume 24% against conventional GMAW. A peer-reviewed comparison of stainless steel arc welding processes found GMAW, in short-circuit and pulsed spray modes, had the lowest Cr(VI) emission rates and SMAW the highest, several GMAW variants falling below 2% of the SMAW rate; the NIOSH engineering controls database entry has the detail.

   Prolonged exposure to welding fume may cause lung damage and cancers of the lung, larynx and urinary tract, and manganese fume exposure can cause Parkinson’s-like symptoms.

   Note: A peer-reviewed review of arc welding fume and Cr(VI) emissions sits at PMC4270263 on PubMed Central, worth keeping in the shop library for fume control decisions.

   No furnace? Rent the equipment

   Not every shop owns a stress relief furnace. Rented induction heating equipment, local PWHT rigs and contract furnace capacity fill the gap, and for a single spool or repair job that is often the sensible commercial answer. The trade-offs: mobilisation costs, a shorter window for drawing approval and less control over record quality. Confirm the supplier’s control system meets the code’s heating rate, soak band and recording requirements, not just temperature.

   Affiliate disclosure: Some suppliers and services referenced in this article may pay us a commission if you buy through the links. That never changes our recommendations, which is why the equipment guidance stays general rather than pointing to one vendor.

   For other materials, our guides to aluminium MIG welding and alloy steel electrodes cover the filler and preheat decisions that sit alongside heat treatment planning. For the metallurgy behind it all, TWI’s work on heat treatment of welded joints is still the reference most engineers quote.

   Key Takeaways: PWHT Meaning in Welding, Summarised

   PWHT stands for post weld heat treatment. In code language it is the uniform reheating of a welded structure, or a defined part of it, to a sufficient temperature to relieve the major portion of welding residual stress, followed by uniform cooling. Outside piping, pressure vessels and structural work, the term gets used loosely to mean “any heat treatment after welding”, and that loose usage is a leading cause of rework on job sites. Inside ASME and AWS, the definition is narrow, and the difference decides whether a job is compliant or merely finished. Wikipedia’s post weld heat treatment summary draws the same line between thermal stress relief and a full solution anneal or normalising.

   Six things determine whether PWHT is required

   1. The material P-No and Group No. These drive which chart and temperature band you use. Carbon manganese, low alloy, stainless, nickel alloys and aluminium all behave differently.
   2. Thickness, and specifically control thickness. Hold time scales with it, and a nozzle reinforcement or a socket weld can change the number you calculate.
   3. The governing code. ASME B31.3, Section VIII Div 1 and2, B31.1, Section III, AWS D1.1 and national standards such as PD 5500 each set their own temperature, hold time, heating and cooling rates, and heating method. Because those rules differ, the level of stress relief achieved differs too.
   4. Service conditions. Caustic, chloride, sour or high-cycle service, and design codes for specific equipment, can push the requirement well beyond the generic table.
   5. The WPS and PQR qualification. PWHT is an essential variable in procedure qualification. TWI notes that it is an essential variable in specifications such as ISO 15614-1 and ASME IX, and that adding or removing PWHT outside the qualified time and temperature ranges forces requalification.
   6. The designer’s own specification. A drawing or piping class can call for PWHT even where the code allows an exemption, and the specification always wins.

   Three questions to ask before you strike an arc

   • What code governs this job, and which edition?
   • Is PWHT required here, and if so, at what temperature and hold time?
   • Is my WPS qualified with PWHT, at the temperature and time I intend to use?

   Ask those three before the first weld and the heat treatment becomes a planned step with time on the schedule. Ask them afterwards and you are negotiating with a furnace, a repair procedure and a re-inspection.

   Quick Note: For controlled furnace PWHT performed to procedures such as PD 5500, residual stress remaining in carbon manganese and 2.25Cr-1Mo welds after treatment in the 580 to 620 °C range is advised at roughly 30 percent of yield. That is good relief, not complete relief, and it is why the TWI review of post weld heat treatment effects on residual stress matters as much to designers as the weld map does.

   The biggest practical lesson

   Furnace PWHT gives the most complete and most uniform stress relief, because the whole component comes up to temperature together and soaks together. Local PWHT, heated with radiant burners, resistance blankets or induction around the girth weld, can be faster and far cheaper, but it leaves higher residual stress and demands expert planning of the band width, the gradient and the heat input. The AWS D10.10 standard for local heat treatment spells out the soak band, heated band and gradient control band for exactly that reason.

   PWHT is part of a system, not an oven cycle

   Welder qualification, preheat, PWHT, non-destructive examination and the fabrication record all have to line up. A clean radiograph does not cancel a PWHT requirement, and a compliant furnace cycle does not rescue a weld that was made outside a qualified procedure. Teams working to aerospace or high-integrity specifications should also read how heat treatment limits are handled in practice, such as in aerospace TIG welding standards and techniques.

   For the exact numbers on your own job, go to the source: the ASME Journal of Pressure Vessel Technology guide to determining PWHT for the decision logic, a qualified welding procedure specification for the procedure itself, and the governing code plus the project specification for the numbers that apply to your job. Always check both. They rarely say the same thing twice.

   Frequently Asked Questions

   What does PWHT mean in welding?

   PWHT stands for Post Weld Heat Treatment. AWS A3.0 defines it as stress-relief heat treatment: the uniform heating of a structure, or a portion of it, to a sufficient temperature to relieve the major portion of the residual stresses, followed by uniform cooling. In code work it always means a controlled thermal cycle with specified soak temperature, holding time, and heating and cooling rates.

   Is PWHT the same thing as stress relieving?

   In practice, yes — PWHT is the code term for the stress-relief heat treatment applied to a welded assembly. Some specifications use “stress relief” and “stress relieving” interchangeably with PWHT. PWHT is narrower than the general phrase “heat treatment after welding,” which also covers solution annealing, normalizing, tempering, and ageing treatments that serve a different metallurgical purpose.

   Does every weld need post weld heat treatment?

   No. The requirement depends on the material’s P-No and Group No., the control thickness, the governing code, the service conditions, and whether the welding procedure specification requires it. ASME B31.3 Table 331.1.3 lists specific exemptions, such as carbon steel P-No 1 welds with mandatory preheat and multiple-layer welding. Austenitic stainless steels generally do not require PWHT, and P-No 8 is recorded as neither required nor prohibited.

   What temperature is used for PWHT on carbon steel?

   For carbon steel P-No 1 and P-No 3 materials, ASME B31.3 Table 331.1.1 specifies a holding range of 595-650 °C (1,100-1,200 °F). Alloy steels step up from there — Cr-Mo steels at 650-705 °C, 1% Cr steels at 760-800 °C, and 2.25Cr-1Mo steels at 730-775 °C. Always work from the chart for your actual P-No rather than from a general rule of thumb.

   How long must a weld be held at the PWHT temperature?

   The common rule is 1 hour per 25 mm (1 inch) of control thickness with a 15-minute minimum, and longer holds beyond 50 mm or 125 mm depending on the material group. Codes also allow an alternate route: dropping the soak temperature by 30 °C (50 °F) doubles the hold to 2 hours, and dropping by 110 °C (200 °F) means 20 hours — though the larger reductions are limited to P-No 1 Group Nos. 1 and 2 at 25 mm (1 in.) thickness or less.

   What is the difference between furnace PWHT and local PWHT?

   Furnace PWHT heats the whole component in an enclosed oven, giving the most uniform temperature and the most complete stress relief, and it is the preferred method where the component will fit. Local PWHT heats a controlled band around the weld using gas-fired radiant heaters, electric resistance or induction heating. Local treatment leaves higher residual stress, needs a wider heated and insulated band, and requires careful planning to avoid damaging thermal gradients.

   Should NDE be done before or after PWHT?

   Common owner specifications require the initial radiography or ultrasonic test before PWHT, with magnetic particle testing performed before and repeated after the heat treatment. A final UT after final PWHT is frequently specified. The governing point is that PWHT must be performed within the qualified WPS range, so any welding done after PWHT has to be covered by the procedure and re-examined.

   Does PWHT remove all the residual stress in a weld?

   No. British Standard PD 5500:2012 estimates that a component still retains residual stress of about 30% of its yield strength after a compliant furnace PWHT, and BS 7910 puts the transverse component nearer 20% of yield. TWI’s research also found that holding temperature matters more than holding time, which is why the soak band and temperature control are the parts of the job to get right.

   Can you PWHT stainless steel the same way as carbon steel?

   Generally no. Austenitic stainless steels usually do not require PWHT, and stress relief at typical carbon steel temperatures can cause sensitization and a permanent loss of corrosion resistance. Where heat treatment is needed after welding stainless, it is normally a full solution anneal — around 1,050-1,150 °C for austenitic grades and 980-1,125 °C for duplex grades depending on the alloy — followed by rapid cooling.

   What is an intermediate stress relief and a dehydrogenation heat treatment?

   They are partial treatments used on high-restraint welds before final PWHT. An intermediate stress relief (ISR) is essentially a partial stress relief, typically above 600 °C for a couple of hours, used on shell butt welds and nozzle attachment welds in high-restraint areas. A dehydrogenation heat treatment (DHT) works at a much lower temperature, typically 300-425 °C for at least one hour, to drive off hydrogen and prevent delayed cracking, and normally requires written approval from the purchaser to be used in place of an ISR.