Post Heat Treatment in Welding: The Complete Guide

The pipe passed every visual check, the radiographs were clean, and the hydrotest held. Eighteen months later it cracked at the weld toe, right where the hardness map showed a brittle band nobody had bothered to measure. That is the moment most fabricators finally learn the difference between a weld that looks finished and a weld that has actually been heat treated.

What Post Heat Treatment in Welding Actually Does

A 2-inch heavy-wall spool left a shop a few years ago with one line on the traveler: “PWHT not required — see exemption.” The exemption was real, but it belonged to a different P-number and an older code edition. Nobody checked. The spool went into wet H₂S service, ran about three years, and split along the fusion line. The postmortem found hardness in the coarse-grained heat-affected zone near 350 HV and a thin band of untempered martensite that a proper stress relief would have softened. The weld was sound. The thermal history after the weld was not.

That is the trap. Welding inspection looks at the joint. Post heat treatment looks at what the joint became. And in most fabrication shops, that step is the one nobody owns until something fails.

What “Post Heat Treatment” Covers

Taken broadly, post heat treatment in welding means any controlled thermal treatment applied after the arc is extinguished and before the part goes into service. That umbrella covers a lot: hydrogen bake-out, stress relief, normalizing, solution annealing, tempering, aging. In plain carbon and low-alloy steel fabrication, though, two treatments do almost all the work:

• Post heating for hydrogen control. A relatively low-temperature hold applied straight after welding, before the joint is allowed to reach ambient, to let diffusible hydrogen escape.
• Stress relief PWHT. A slower, hotter, separate operation performed below the steel’s lower critical temperature (Ac₁), designed to relax residual stress and temper the heat-affected zone.

They sound similar. They solve different problems, and confusing them causes a large share of field failures. The ESAB technical library draws the same distinction clearly, and it is worth reading before you write a procedure.

Where Residual Stress Comes From

A weld is a series of violent, local thermal cycles. Metal under the arc expands, but it is surrounded by cold, unyielding plate. When it cools, it tries to shrink and is held back again. Shrinkage plus restraint equals locked-in stress, and the numbers are not small. The James F. Lincoln Arc Welding Foundation notes that welding residual stresses reach magnitudes roughly equal to the yield strength of the base material. In a 50 ksi steel, that means 50 ksi of stress sitting in the part doing nothing useful.

Why that matters in practice:

• Brittle fracture. A crack-like flaw in a high-tensile residual field needs far less applied load to run.
• Stress corrosion cracking. Sour service, caustic, and amine environments attack stressed metal preferentially. Hard, stressed HAZ is prime real estate for SCC.
• Dimensional movement. Cut into a stressed weldment on a mill and the stress redistributes. Bores go oval, flanges go out of flat.
• Fatigue. Residual tension adds to cyclic service stress and eats fatigue life.

What the Heat Actually Changes

Hold steel below Ac₁ for long enough and two things happen. Yield strength falls with temperature, so the locked-in stress can relax through micro-yielding — the metal creeps a few microstrain and the stress drops with it. At the same time, the hard martensite and bainite formed by the fast weld cycle get tempered, which lowers hardness and raises toughness. In post heating, a third mechanism runs: hydrogen diffuses out while the joint is still warm. The Tetra Engineering overview is a good plain-language starting point on the mechanisms.

What PWHT Does Not Do

Set expectations early, because this list prevents more arguments than any table:

• It does not fix bad fit-up, excessive gap, or a joint that was never properly fused.
• It does not repair porosity, slag, or lack of fusion. Those are discontinuities; heat treatment is not a repair process.
• It does not restore strength lost to overheating or excessive heat input. Once the microstructure is coarsened, stress relief cannot undo it.
• It cannot fix a weld that was already cracked. It may even open the crack further through thermal stress.

Think of it as stress management, not weld repair.

The Question That Decides Everything

Before temperatures and soak times, one question governs the whole job: is PWHT required for this joint, and if so at what temperature and for how long? Every other decision — equipment, thermocouples, schedule, cost — flows from that answer.

Once you know it is required, four variables control the outcome: temperature, time, heating rate, and cooling rate. Get the temperature and time right but cool the part at 600°F per hour and you can put residual stress straight back in. Get the rates right but under-soak and the hardness map tells the story for you.

The rest of this guide works through each of those variables, the codes that set them, the metallurgy behind them, and the field mistakes that break them. It also covers filler metal behaviour: an alloy steel electrode classified on post-weld-heat-treated properties behaves very differently from one classified as-welded, and that difference decides which electrode you can actually use. First, though, we need to untangle the vocabulary — because in most shops, “post heating” and “PWHT” are used as if they mean the same thing, and they do not.

Post Heating vs. Stress Relieving vs. PWHT: Clearing Up the Confusion

Ask five people on the same job site what “PWHT” means and you will get at least three answers. One describes baking hydrogen out of a joint. Another points at a furnace chart running overnight at 1,150 F. A third shrugs and says it is whatever the drawing says. They are all describing real processes, but not the same process.

Here is the clean split. In North American practice, PWHT is shorthand for post-weld heat treatment, and nine times out of ten it means stress relief. Post heating is something else: a lower-temperature hydrogen bake-out performed immediately after welding, while the joint is still hot. Confuse the two and a bake-out gets written on a traveler as “PWHT 450 F,” signed off, then fails a week later at the toe of a restrained fillet.

Post Heating Is a Hydrogen Bake-Out

Hydrogen cracking is a delayed failure. Hydrogen picked up during deposition migrates toward the hard, highly stressed heat-affected zone and, given time and restraint, it pops. Post heating interrupts that clock by keeping the joint warm so hydrogen diffuses out through the steel before it concentrates anywhere dangerous.

A common baseline is 450 F (230 C) held for roughly one hour per inch (25 mm) of thickness, as described in JFLF’s key concepts guidance. The exact numbers vary by code, and the code on your drawing wins.

• ASME Section III, P-No. 1: 450–550 F (230–290 C), minimum 2 hours.
• NBIC: 500–550 F (260–290 C), minimum 2 hours.
• AASHTO/AWS D1.5 fracture control: 450–600 F (230–315 C) for one hour per inch of weld thickness or two hours, whichever is less.

Nothing metallurgical is being reset at 450 F. The steel is kept warm enough, long enough, for hydrogen to leave, which is why ESAB’s guidance on preheating and post-weld heat treatments treats them as two separate subjects with two separate purposes.

Stress Relief Is a Slow, Separate Operation

Stress relief PWHT is measured in the four figures. For P-No. 1 carbon steel above the thickness trigger, the window is around 1,100–1,200 F (593–649 C). At that heat the yield strength drops far enough that residual stress relaxes by micro-yielding, and hard transformation products in the HAZ get tempered.

That means stress relief happens later. The joint cools, gets inspected, maybe gets radiographed, then the component goes into a furnace or under resistance pads for a controlled heat-up, a soak, and a controlled cool-down. Hours of it, with ramps on both ends.

The Rest of the Family: Normalizing, Solution Annealing, Tempering, Aging

Stress relief is one member of a larger family of post-weld thermal treatments. Normalizing refines the coarse grain structure left by high-heat-input processes such as electro-slag welding and removes the hard HAZ zones they create. Solution annealing for austenitic stainless, typically 1,900–2,010 F, redissolves carbides that precipitated during welding. Tempering restores toughness in quenched-and-tempered and martensitic steels. Aging precipitates hardening phases in alloys designed for it, aluminum included.

Treatment	Purpose	Temperature	Timing	Materials	If Skipped
Post heating	Hydrogen bake-out	450–600 F (230–315 C)	Immediately, before ambient	Carbon steel, Cr-Mo, thick sections	Delayed hydrogen cracking
Stress relief PWHT	Relax stress, temper HAZ	1,100–1,425 F (593–775 C)	Hours later, after cooling and NDT	P-No. 1 above trigger, Cr-Mo, Q&T steels	Brittle fracture, SCC, distortion
Normalizing	Refine coarse grains, remove hard zones	~1,600–1,700 F (870–925 C)	After welding, full component	Electro-slag and high-heat-input welds	Low toughness, hard HAZ
Solution annealing	Redissolve carbides	~1,900–2,010 F (1,040–1,100 C)	After welding, with rapid quench	Austenitic stainless	Sensitization, corrosion attack
Aging	Precipitation hardening	Alloy-specific	After solution treatment or welding	PH stainless, aluminum, nickel alloys	Strength left on the table

Why Timing Matters More Than Temperature

Post heating must begin from interpass temperature. If the joint cools to ambient, the window is closed. Hydrogen is already sitting where it wants to sit, and warming the part afterward does not undo the damage. The procedure should say “maintain 450 F minimum until post heating begins,” not “post heat after welding.”

Filler metal matters too. A hydrogen-controlled electrode such as an E7018 is doing half the job before the heaters are switched on, and the low-hydrogen practices in this guide to alloy steel welding electrodes are the other half.

One Job, Both Treatments, In Sequence

These processes coexist. A 2.25Cr-1Mo reactor nozzle might get a hydrogen bake-out at 450 F straight off the cap pass, cool down, pass inspection, and be stress relieved at 704–760 C days later. Two operations, two procedures, two charts, two sign-offs. The bake-out protects the joint while it waits; the stress relief finishes the job. Treat them as one line item and you will get one wrong.

Why Codes Require It: ASME, AWS, and API Rules in Plain Language

Codes do not require post-weld heat treatment because heat is automatically good for a weld. They require it because the failure modes are predictable. A hard heat-affected zone, restrained shrinkage, residual hydrogen, and time in service can add up to cracking, and a code table is the industry’s agreed way of interrupting that sequence before it starts.

The logic is risk management, not ceremony. A code table compresses a metallurgical argument into a rule an inspector can enforce. If the steel is in this family, the joint is this thick, and the service is this severe, the weld must be held within a temperature range for a defined time. ESAB’s guidance on preheating and post-weld heat treatment makes the same point: the treatment is chosen to control hardness, hydrogen, and residual stress, not to improve appearance.

P-Numbers: The Shorthand That Makes Code Tables Work

ASME Section IX groups base metals into P-numbers by composition and metallurgical response to welding and heat treatment. P-No. 1 covers carbon steels. P-No. 3 and P-No. 4 cover low-alloy steels with different chromium and molybdenum levels. P-No. 5A is the familiar 2.25Cr-1Mo family, and Grade 91 sits in its own demanding group. The purpose is practical: one table can cover dozens of similar steels without listing every trade name. Filler metal selection is a separate conversation, which is why matching alloy steel welding electrodes to the base metal matters.

A P-number is not a strength grade. Two steels in the same group can behave differently in the HAZ, so the project specification and the material certificate still have to be read.

The ASME Family: Vessels, Boilers, and Piping

ASME Section VIII Division 1 is the pressure vessel benchmark. UW-40 sets the procedure: thermocouple placement, loading temperature, heating and cooling rates, and records. UCS-56 gives the when and how hot for carbon and low-alloy steel. UHA-32 does the same for high-alloy steel, usually with narrower triggers and different cooling rules.

ASME Section I PW-39 covers boiler parts, with its own thickness triggers and rules for attachments. ASME B31.1 covers power piping, while ASME B31.3 covers process piping. API RP 582 is not a code, but this recommended practice is often invoked alongside B31.3 for refining service and fills gaps the code leaves open.

Material group	Typical B31.3 PWHT	Hardness ceiling
P-No. 1 carbon steel over 19 mm	593–649 C (1100–1200 F), 1 h/in, 1 h minimum	Not the usual basis
P-No. 4 (1.25Cr-0.5Mo)	704–746 C (1300–1375 F)	225 HB
P-No. 5A (2.25Cr-1Mo)	704–760 C (1300–1400 F)	241 HB
Grade 91	Roughly 1350–1425 F	Project specific

B31.3 also caps hardness at 225 HB for P-No. 3 and P-No. 4, and 241 HB for P-No. 5A and 5B. Those are acceptance criteria, not suggestions.

AWS and Bridge Work

AWS D1.1 addresses PWHT in Clause 5.8 of the 2015 edition and Clause 7.8 of the 2020 edition. That renumbering trips up people who assume every edition is organized the same way. AWS D1.5 takes a fracture-control view for bridges and adds post-heating requirements for thick or highly restrained joints. Where a bridge project also invokes AASHTO, compare both documents before choosing a cycle.

Thickness, Soak Band, and the Numbers That Decide the Job

“Nominal thickness” and “control thickness” are where many disputes begin. The control thickness is the dimension that actually drives hold time and heating rates, and it is not always the number printed on the drawing.

• Butt joints: use the thickness at the welded joint, including the full penetration depth.
• Fillet welds: use the throat dimension, not the leg size.
• Partial-penetration joints: use the depth of the joint preparation.
• Repair welds: use the repair depth and the original joint thickness as the code requires.
• Unequal thickness: the code definition decides whether the heavier section controls; some exemptions disappear when the thicker part crosses a threshold.

Section VIII Division 1 sets the minimum soak band as the weld plus 1t or 2 in. (50 mm), whichever is less, on each side. Soak band is the minimum zone that must reach temperature, not the whole heated band.

Exemptions, Notes, and Conflicts

The table gets you close. The notes decide. An exemption may hinge on carbon equivalent, thickness, preheat, service, or hardness, and a footnote that waives PWHT for one P-number can vanish when the joint is highly restrained or the service is sour. When more than one code applies, the most restrictive requirement governs. A client specification can also be stricter than the code, and it usually is.

Good PWHT planning guidance starts with a written procedure, not a chart borrowed from the last job.

The Metallurgy: What Happens Inside the Steel During a PWHT Cycle

A PWHT chart tells you what the furnace or heating pads did. Metallurgy tells you what the steel did in response. Those are not the same story, and confusing them is how a weld can pass a hardness check while still carrying a brittle, crack-prone microstructure.

Stress Relief Is Slow, Controlled Yielding

Welding leaves residual stress because the weld and its surroundings heat, expand, cool, and shrink at different rates while the joint is restrained. At PWHT temperature, the steel’s yield strength falls. Dislocations, the defects that carry plastic strain, gain enough mobility to climb and annihilate. The steel micro-yields under its own locked-in stress, and that stress relaxes. This is creep behavior, just at a temperature and time scale we can use in a shop.

The practical point is that temperature does most of the work. A long soak at a temperature too low to drop yield strength significantly will not produce the same result as a shorter soak in the correct window. Time matters, but it cannot fully compensate for an inadequate temperature.

Tempering the Heat-Affected Zone

The weld thermal cycle heats the HAZ rapidly and cools it quickly, producing hard martensite and bainite in susceptible steels. Holding below the lower transformation temperature, Ac1, tempers those structures. Carbon diffuses, carbides precipitate, hardness falls, and toughness generally improves. The aim is not to make the HAZ uniformly soft. It is to eliminate the hard, brittle band that can initiate cracking at the weld toe.

Carbides also change during the hold. Fine precipitates can grow and coalesce into larger particles. Moderate coalescence is part of tempering, but excessive coarsening can reduce strength and creep resistance. That is why time and temperature trade off only within limits. Doubling the hold time is not a safe substitute for raising temperature into the specified range, and raising temperature above the limit is never a shortcut.

The Ac1 Boundary: Where PWHT Can Backfire

Exceeding the lower transformation temperature, Ac1, forms fresh austenite. On cooling, that austenite can transform into new, untempered martensite or bainite. Instead of tempering the existing structure, you create a harder one. Codes set upper temperature limits with margin below Ac1 for exactly this reason. A chart that looks acceptable at the thermocouple can still hide a hotter core or a locally overheated band.

When PWHT Does Harm: Temper Embrittlement and Reheat Cracking

Temper embrittlement is a toughness problem that hardness testing will not reveal. Certain Cr-Mo steels, particularly 2.25Cr-1Mo, can lose notch toughness when held or cooled slowly through an embrittling temperature range. Impurity elements such as phosphorus, tin, antimony, and arsenic segregate to grain boundaries and weaken them. The step cooling test is used to simulate long-term embrittlement and assess toughness loss, a serious concern in hydroprocessing equipment.

Reheat cracking is another failure that appears after the cycle rather than during it. It occurs almost exclusively in coarse-grained HAZ regions of Cr-Mo and Cr-Mo-V steels, typically in the 350–550 C range during PWHT or elevated-temperature service. Impurity elements including antimony, arsenic, tin, sulfur, and phosphorus segregate to grain boundaries and increase susceptibility, as TWI’s guidance on reheat cracking explains. Once discovered, re-heating does not repair it.

Weld Metal and Base Metal Respond Differently

Filler metals have their own temper response. An E7018 deposit may have about 75 ksi tensile strength as-welded but only about 65 ksi after stress relieving, according to JFLF’s key concepts. E8018-B2 is classified to a minimum of 80 ksi after stress relief at 1275 F (690 C) for one hour, even though its as-welded strength may reach 120 ksi. The drop is not a defect. It is the intended metallurgical response, and the procedure must be qualified with that response in mind. The differences between alloy steel welding electrodes matter here, because assuming a deposit will behave like the base plate is a common planning error.

Mechanism	What Changes	Practical Risk
Stress relaxation	Dislocations move; residual stress drops	Too cool or too short leaves stress locked in
HAZ tempering	Martensite and bainite soften; toughness rises	Under-soaking leaves a hard band
Carbide coarsening	Precipitates grow during the hold	Excess time or temperature reduces strength
Exceeding Ac1	Fresh austenite forms and retransforms	New untempered structures, harder than before

What This Means on the Job

For the welder or inspector, the takeaway is simple: PWHT is a metallurgical treatment with a window, not a ritual of reaching a number on a chart. Too cold or too short leaves residual stress and hardness in place. Too hot risks fresh transformation. The wrong cooling path can embrittle a steel that passed every strength check. Watch thermocouple placement, because a surface reading can hide a hotter core, and remember that weld metal and base metal may end up with different strengths after the same cycle.

The best mental model is that you are not “baking” the weld. You are giving the steel enough thermal energy to relax, temper, and stabilize without crossing the boundaries that would undo the benefit.

Temperature, Time, and Ramp Rates: Reading a PWHT Cycle Like an Inspector

A PWHT chart tells a story in four parts, and most people read only the middle. They check the hold temperature, confirm it stayed in band, and move on. An inspector reads the whole curve. A steep climb, a short soak, or a furnace door opened early can undo much of the benefit even when the hold number looks perfect.

The Four Legs of a PWHT Cycle

Every cycle has heat-up, hold, cool-down, and unmonitored tails at each end. The tails matter because steel does not wait for the recorder to cross a control threshold before it starts changing.

• Heat-up: The part moves from ambient to holding temperature at a controlled rate. Too fast, and the outside runs ahead of the inside, creating a fresh thermal gradient and new stress.
• Hold: Time at temperature does the metallurgical work. The soak band must reach the target, not just the thermocouple nearest the heater.
• Cool-down: This is still part of the treatment. A controlled decline keeps residual stress from rebuilding and avoids hard, brittle structures in sensitive steels.
• The tails: Below the control threshold the chart may look unimportant, but the part is still cooling. Insulation stays in place, and the recorder keeps running until the component is safely down in temperature.

The Ramp Rules That Actually Govern the Job

ASME Section VIII Division 1 sets a familiar rule for carbon and low-alloy steel: above 800 F, heating and cooling rates shall not exceed 400 F/hr divided by the maximum thickness in inches, and need not be less than 100 F/hr. On a 3-inch component, that is about 133 F/hr. On a 1-inch component, the calculated number is 400 F/hr, but the 100 F/hr floor keeps very thin parts from being treated with unnecessary caution. The rule sits inside ASME’s PWHT guidance and is repeated across construction codes with local variations.

AWS D1.1 uses a different trigger. Above 600 F, heating is limited to 400 F/hr divided by thickness, while cooling is limited to 500 F/hr divided by thickness. That earlier threshold catches structural work ASME might not control until 800 F, which is why copying a ramp rate from one code to another is risky.

ASME B31.3 controls cooling above 1200 F at roughly 500 F/hr divided by nominal thickness in inches, about 278 C/hr per inch, with an absolute ceiling of 500 F/hr. In practice, thick-walled process piping cools more slowly than the ceiling allows, and the calculated rate often governs.

Code	Control threshold	Heating limit	Cooling limit
ASME VIII Div. 1	Above 800 F	400 F/hr ÷ thickness, min. 100 F/hr	400 F/hr ÷ thickness, min. 100 F/hr
AWS D1.1	Above 600 F	400 F/hr ÷ thickness	500 F/hr ÷ thickness
ASME B31.3	Above 1200 F	Per procedure and code	500 F/hr ÷ thickness, max. 500 F/hr

A Worked Example: The 3-Inch Nozzle

Take a 3-inch-thick P-No. 1 carbon steel nozzle. The hold time is 1 hour per inch with a 1-hour minimum, so the soak is 3 hours. The soak band extends 3 inches each side of the weld, and the heated band must be wider still to control the gradient. Heat-up above 800 F is capped at 400 ÷ 3, or about 133 F/hr. Climbing from 800 F to a 1200 F hold takes roughly three hours. Add the hold, controlled cool-down, and tails, and the job is closer to ten or twelve hours than three. PWHT is a scheduling item, not a quick finish step.

Thermocouples: The Weak Link That Looks Fine

Type K and Type N thermocouples are common for PWHT because they tolerate the temperature range and are widely available. Attachment matters more than many crews expect. Capacitor-discharge welding fuses the junction to the surface and gives a direct metal-to-metal reading. Clamps, bands, or wire can leave an air gap or a conduction path that skews the number. PARs Technical’s recommended practices report mechanically attached couples reading 30–40 F above welded junctions in the 1100–1200 F range, with one case showing roughly a 130 F difference. At a 1200 F hold, that is the difference between passing and scrapping.

Control thermocouples drive the controller. Monitoring thermocouples independently record what the component is doing. The LANL welding standard calls for at least six thermocouples on pipe larger than NPS 8 and four on NPS 8 and smaller, split between control and recording, with spares available. That is not bureaucratic excess. It is how you catch a hot spot one control couple would never see.

Placement: Where the Numbers Come From

Put thermocouples at the weld centerline, the soak band edge, and the heated band edge. On one-sided heating, measure the opposite side too. The heated face can be at temperature while the far wall lags by 50 F or more, and a chart that watches only the hot side will certify a soak the whole section never achieved. Nozzles and attachments deserve extra couples because shell and nozzle neck gain and lose heat at different rates.

Furnace vs. Local Heating: Equipment, Methods, and When to Use Each

You can hit the right temperature and still fail the job. How you deliver heat around the joint decides whether the weld relaxes or just gets hot in a narrow stripe. Get it wrong and you create a stressed shell with a baked ring in the middle.

Furnace Treatment: The Reference Standard

Furnace heating treats the whole component at once, and that is the point. Convection and radiation from every side mean each thermocouple sees the same environment, thick-to-thin transitions flatten naturally, and the chart is easy to defend. It suits shop vessels and small spools that fit without fighting the door.

The limits are physical: furnace size, crane capacity, hearth loading, and handling risk. A vessel that barely fits ends up on inadequate saddles, and a long spool can sag during the hold. Support it properly before you close the door.

Local Heating with Resistance Pads: The Field Workhorse

Ceramic pad heaters wrapped in insulation do most of the world’s field PWHT. They are flexible, they work on pipe, nozzles, and structural nodes, and they re-lay in hours. But everything depends on layout and craft: heaters too narrow, insulation too short, or a thermocouple in the wrong place and you get a hot band surrounded by cold, restrained steel.

Local heating exists because furnace treatment is often impossible during field fabrication and repair, so it has to reproduce furnace results inside the treated zone. The standard is simple: the soak band (weld plus heat-affected zone plus adjacent base metal) must reach and hold temperature as uniformly as a furnace would.

Induction Heating

Induction coils heat the pipe wall itself by eddy currents and hysteresis. The method is fast, clean, and efficient on carbon and low-alloy pipe, and it scales well on large-diameter girth seams. It is also unforgiving: coil spacing, frequency, and power all shape the pattern, and thin or variable wall joints show hot spots fast. This needs design work, not guesswork.

Exothermic and Chemical Kits

Cheap, portable, and no power required. They are also the least controllable option on the market, with crude temperature regulation and no meaningful ramp control. For code-critical PWHT they are generally unsuitable. Save them for preheat on a noncritical repair, not a stress relief an inspector will review.

Soak Band, Heated Band, and Gradient Control Band

These three terms are not interchangeable. The soak band is the region that must reach holding temperature: weld, HAZ, and adjacent base metal. The heated band extends beyond it so temperature can be controlled there and induced stress limited at the boundary. The gradient control band adds insulation, and sometimes supplementary heat, outside the heated band to keep the gradient gentle. Insulation beyond the heaters is not decoration; it stops a steep gradient from creating the stress you paid to remove.

Layout Around Nozzles and Attachments

Nozzles are where local PWHT goes wrong most often. The thin neck heats quickly; the thick shell lags. Heat the neck alone and you over-temper it while the shell never reaches temperature. The answer is separate control zones, heaters on the shell following the six-times-thickness rule, and enough thermocouples to prove both regions held. Branch connections need the same logic.

One-sided heating also induces distortion. Brace the component, plan lifting points before the cycle starts, and never lift a hot part by a temporary lug welded to the soak band. And no quenching, ever. Water or forced air at 1,100 °F introduces exactly the stress and hardness you paid to remove. Cool under insulation, on the chart, per the procedure.

Method Comparison

Method	Control	Uniformity	Cost	Typical Use	Limitation
Furnace	Excellent	Excellent	High setup, low per unit	Shop vessels, spools	Size, handling, hearth
Resistance pads	Good, craft-dependent	Good if banded right	Moderate	Field piping, nozzles	Gradients, power supply
Induction	Good with design	Good on plain pipe	Moderate to high	Large girth seams	Coil design, hot spots
Exothermic	Poor	Poor	Low	Noncritical preheat	No ramp control

Pick the method from the geometry, not the calendar. If it fits in a furnace, furnace it. If not, build the band correctly and prove the cycle on paper.

Materials That Need It, Materials That Do Not, and Ones That Get Hurt

The code table tells you what is mandatory. It does not tell you what is wise. Two joints can share a P-number, a thickness, and a weld profile while needing completely different treatment because one base metal arrived normalized and the other arrived quenched and tempered. The material’s prior thermal history matters as much as the clause you are reading.

Carbon Steel: Thickness Is the Trigger

P-No. 1 carbon steel is the workhorse of fabrication, and its PWHT rules are largely thickness-driven. Under ASME B31.3, ordinary carbon steel piping generally crosses the mandatory line above 19 mm, or about 3/4 in. Below that, many joints are exempt—but exemption is not the same as immunity. Sour service, thick restrained connections, low-temperature toughness requirements, or a different governing code can still push a joint over the line.

Low-Alloy Cr-Mo: Almost Always, and Hotter as Chromium Rises

With chromium and molybdenum additions, the default flips from “probably not” to “almost certainly.” These steels harden deeply, tolerate hydrogen poorly, and often operate at elevated temperature where creep matters. B31.3 puts P-No. 4, such as 1.25Cr-0.5Mo, at 704–746 °C, and P-No. 5A, 2.25Cr-1Mo, at 704–760 °C with a 241 HB maximum. Matching alloy steel welding electrodes to the base metal is part of the same decision, because the weld metal must survive the cycle without losing its required strength.

Quenched and Tempered Steels: The Over-Tempering Trap

Quenched-and-tempered alloy steels can be damaged if the PWHT exceeds the tempering temperature originally used to produce the plate. Cross that ceiling and you soften the base metal, reduce strength, and possibly lose toughness you cannot recover. The mill certificate is not paperwork to file and forget; it carries the tempering temperature you need to respect. The James F. Lincoln Arc Welding Foundation makes this point plainly: PWHT is beneficial only when the cycle stays below the steel’s original tempering range.

Austenitic Stainless: Usually Exempt, Not Always Safe

Most austenitic stainless weldments do not require PWHT. ASME B31.3 Table 331.1.1 lists no mandatory treatment for P-No. 8 austenitic steels, and ASME Section VIII UHA-32 generally exempts them as well. But “not required” does not mean “never affected.” Holding these alloys in roughly 800–1500 °F can precipitate chromium carbides and sensitize the grain boundaries, leaving the weld vulnerable to intergranular corrosion. Where corrosion resistance must be restored, solution annealing near 1900–2010 °F followed by rapid cooling redissolves those carbides. Tetra Engineering notes that this is a distinct heat treatment, not a routine stress relief.

Duplex Stainless: Never a Conventional Stress Relief

Duplex stainless is a special case because the usual stress-relief temperature range is metallurgically hostile. Holding near 600–900 °C can form sigma phase and other unwanted precipitates that wreck toughness and corrosion resistance. Where heat treatment is applied, ASME B31.3 points to solution treatment at 1020–1100 °C (1870–2010 °F) for S31803 and S32205, followed by rapid cooling. A conventional stress relief is not a substitute.

Ferritic, Martensitic, and 9Cr Steels: Narrow Windows

Ferritic and martensitic stainless grades have their own rules, and the differences between grades are too large for a single recipe. Follow the code and the filler metal manufacturer’s procedure. Grade 91 is the most demanding common example: its PWHT window is tight, roughly 1350–1425 °F, and over-tempering can sacrifice creep strength. It is also prone to reheat cracking, so the cycle must be controlled rather than improvised.

Nickel Alloys and Aluminum: By Agreement, Not by Habit

Nickel alloys often receive PWHT only when specified by agreement, particularly for corrosion service or dissimilar joints. Aluminum behaves differently again: stress relief offers little benefit, while aging—or full solution treatment and aging—is used to develop properties. Neither family rewards a “carbon steel mindset.”

Material	PWHT normally required?	Typical treatment	Main risk
Carbon steel P-No. 1	Thickness-triggered	Stress relief above 3/4 in. in B31.3	Missed exemption or wrong thickness
1.25Cr-0.5Mo P-No. 4	Almost always	704–746 °C	Hydrogen cracking, hardness
2.25Cr-1Mo P-No. 5A	Almost always	704–760 °C, 241 HB max	Temper embrittlement
Quenched and tempered steel	Often, by code or specification	Below original tempering temperature	Over-tempering and strength loss
Austenitic stainless P-No. 8	Generally no	Solution anneal if corrosion resistance is impaired	Sensitization
Duplex stainless	Special case only	1020–1100 °C plus rapid cooling	Sigma phase, lost toughness
Grade 91	Yes, tightly controlled	Approximately 1350–1425 °F	Over-tempering, reheat cracking
Nickel alloys and aluminum	By specification or agreement	Alloy-specific aging or annealing	Applying the wrong thermal model

The practical rule is simple: identify the material family, check the governing code’s mandatory trigger, then check the base metal’s prior condition. A treatment that helps one P-number can ruin another. The next section deals with what happens inside the steel when the cycle is actually applied—and why those changes are not always beneficial.

Common Mistakes, Failures, and How to Troubleshoot Them

Most post weld heat treatment failures are not exotic. They come from about ten repeatable mistakes, and nearly every one leaves a signature you can read backwards from the chart, the hardness report, or the insulation. Once you know the signature, troubleshooting stops being guesswork and becomes evidence work.

The Ten Failures, Symptom by Symptom

1. The cycle that never happened. Symptom: no chart, or a chart dated after the hydrotest. Root cause: the exemption note read halfway, or control thickness measured at the wrong location on an unequal-thickness joint. Fix: engineering disposition and a documented cycle. Prevention: highlight the exemption clause on the drawing and measure thickness where the code defines it, not at the convenient spot.

2. Heating too fast. Symptom: a flange that bowed, or new cracking at attachments during heat-up. Root cause: ramp rate exceeded, often because the thickest section governed and nobody did the math. Fix: re-support and correct before final machining. Prevention: calculate ramp from maximum thickness and write the number on the layout sketch.

3. Cooling too fast. Symptom: hardness passes but the part still moves in machining. Root cause: furnace door cracked open early or insulation stripped while the component was still hot. Opening a furnace early or removing insulation too soon accelerates cooling above the controlled threshold and can re-introduce the residual stress you just paid to remove. Fix: controlled cool to the code threshold. Prevention: a written cooling instruction and one person authorized to remove insulation.

4. Over-tempering quenched-and-tempered steel. Symptom: tensile or hardness results below specification. Root cause: hold temperature above the mill’s original tempering temperature. This is not recoverable. Prevention: read the mill certificates before writing the procedure, and hold safely below that tempering temperature.

5. Under-soaking. Symptom: a beautiful chart with one thermocouple sitting 40 °F hotter than the rest. Root cause: the recorder averaged a hot zone and a cold zone. Extending soak time does not compensate for holding below the required minimum temperature; time and temperature only trade within narrow, code-defined limits. Prevention: control on the coldest thermocouple and add couples at heavy-to-light transitions.

6. Thermocouple errors. Symptom: scattered readings that disagree with a contact pyrometer. Root cause: mechanically attached couples, damaged extension wire, wrong thermocouple type, or reversed polarity. Mechanically attached couples are documented to read higher than attached-by-capacitance-discharge couples. Fix: verify with an independent instrument and disposition through engineering. Prevention: proper attachment per PARs recommended practices, spares, and continuity checks before power-up.

7. Uneven temperature on nozzles and transitions. Symptom: the shell overheats while the nozzle lags. Root cause: one control zone for two very different masses. Prevention: separate zones, heaters wrapped around the nozzle neck, and insulation carried well past the heated band.

8. Reheat cracking found after the cycle. Symptom: indications on UT or MT at the weld toe, or a failure during hydrotest. Reheat cracking is intergranular along prior austenite grain boundaries and usually sits in the coarse-grained HAZ beneath the weld or cladding, as TWI explains. Large weld beads make it worse: coarse columnar grains and a coarse HAZ that later passes cannot refine. It cannot be fixed by re-heating; the cracked metal must be excavated and repaired.

9. Hardness failures after PWHT. Symptom: a reading above the acceptance limit. Root cause is often the test itself: measuring on decarburized surface, grinding too shallow, or testing base metal instead of the agreed location. Genuine failures mean the soak was short or cold. ASME B31.3 sets Brinell limits of 225 HB for P-No. 3 and 4 and 241 HB for P-No. 5A/5B as a check on PWHT effectiveness. Prevention: an agreed hardness map and documented surface preparation.

10. Treating PWHT as a cure for hydrogen cracking. Symptom: delayed cracking days after welding despite a full stress relief. Root cause: hydrogen was never baked out. Post heating at roughly 450 °F for an hour per inch, started before the joint cools, is the hydrogen measure; ESAB’s guidance makes the distinction clear. Stress relief hours later does not substitute.

Troubleshooting at a Glance

Problem	What you see	Usual cause	Fix
Missed PWHT	No chart or late date	Exemption or thickness misread	Engineering disposition, run cycle
Fast heating	New distortion	Ramp above limit	Correct before machining
Fast cooling	Movement in service	Insulation removed early	Re-run with cooling control
Under-soak	Wide TC spread	Averaged zones	Control on coldest TC
Reheat cracking	Toe indications	Coarse HAZ, large beads	Excavate and repair

Two habits prevent most of this list. First, read the chart during the cycle, not the morning after. Second, when a filler metal choice drives hardness or toughness risk, get it right up front with a proper alloy steel welding electrode selection guide rather than solving it with heat treatment later.

The pattern across all ten mistakes is the same: the process was treated as a ritual rather than a controlled, measured operation. Fix the measurement, fix the documentation, and the failures mostly stop.

Planning a PWHT Job: Procedure, Qualification, Documentation, and Cost

Most post-weld heat treatment failures are not furnace failures. They are planning failures: a controller set to the wrong number, a thermocouple in the wrong place, a WPS never qualified with a heat treatment cycle attached. Good PWHT jobs are won on paper long before the heaters switch on.

Step 1: Read the Drawing and Pin Down the Code

Start with three questions: which code governs, which edition, and what the drawing actually calls for? A note reading “stress relieve per code” is a trap, not an instruction. The code, the material P-number, and the nominal or control thickness together decide whether PWHT is required, at what temperature, and for how long.

Record the P-number, the thickness basis (butt, fillet, partial penetration, repair, and unequal thickness all calculate differently), and the soak band width before writing a line of procedure. On Cr-Mo work, confirm which filler metal the WPS names — the classification matters, as covered in this guide to alloy steel welding electrodes.

Step 2: Build the Heat Treatment Procedure

A usable procedure states, in plain numbers: temperature range, hold time and its basis, maximum heating and cooling rates, soak and heated band dimensions, thermocouple count and attachment method, insulation specification, and recorder details. Anything vague gets interpreted differently by the night shift. Sketch the soak band on the component with a paint stick so the crew can see where the heaters belong.

Step 3: The WPS/PQR Trap Under QW-407

This is where projects quietly go wrong. Under ASME Section IX, PWHT is an essential variable (QW-407). A procedure qualified without post-weld heat treatment cannot simply be used on a joint that will be heat treated — the mechanical properties change and the qualification no longer represents the work. Changes to the time and temperature range, or to the method of performance, require requalification.

Practical consequence: when qualification coupons are welded for a job that includes PWHT, the coupons themselves must go through the cycle — same temperature range, hold time basis, and cooling approach — before machining and testing. Testing as-welded coupons and then heat treating production welds is a documentation failure waiting for an auditor.

Step 4: Documentation That Survives the Job

PFI guidance in ES-45 sets out what a proper package contains: the PWHT procedure, heater and thermocouple layout drawings, certificates for thermocouples and extension wire, recorder calibration records, hardness results, and the complete chart record. Recording instruments should be calibrated at least annually under a quality program.

• Chart recorder output covering heat-up, hold, and cool-down, not just the hold
• Calibration certificates for every recorder and thermocouple lot
• Layout sketches showing actual, not intended, thermocouple positions
• Hardness results with marked locations — hardness after PWHT is the evidence that tempering actually happened
• A signed closeout report listing deviations and how they were resolved

Step 5: Sequencing and Scheduling

PWHT normally comes before hydrotest and before final machining that depends on dimensional stability, since stress relaxation can move a machined face. Repairs made after PWHT generally need their own heat treatment, so the schedule needs a rework buffer rather than a hope that none occurs.

Step 6: What Actually Drives the Cost

Tonnage and geometry set the baseline. From there: furnace versus local heating, total hold time, thermocouple count, insulation quantity, mobilization distance, and rework risk. Local heating on a remote site with 12 thermocouples and two shifts of technicians is not the same purchase as a batch of small spools in a shop furnace. Boiler and vessel specialists such as Alamo Welding & Boiler Works build this into their pricing for a reason. A documented cycle costs a fraction of a rejected spool, a re-hydrotest, or an in-service cracking investigation.

Step 7: Safety Is Part of the Procedure

Insulated surfaces stay dangerous long after the chart shows cooldown. Plan barricades and signage, a burn-hazard briefing, and a no-touch rule until the surface is verified cool. Resistance heating draws serious current — check supply capacity, cable sizing, and grounding. Confined-space thermocouple work needs gas testing and an attendant, insulation releases dust, and every heater circuit needs lockout/tagout so nobody energizes a band while a colleague is attaching a couple.

The chart record is the only part of the job that survives. Plan the paperwork as carefully as the heaters, and heat treatment becomes a controlled, provable process instead of a ritual at the end of the shift.

Field Checklist and Final Takeaways

The unit is hooked up, insulation is on, and someone must decide whether the job is ready. Print this checklist and work through it in order. If a line fails, fix it before the burners come on; once the chart starts drawing, you are committed.

Pre-Job Checklist

1. Confirm the code and edition. Write the governing standard and revision on the job packet. ASME B31.3, Section VIII Div. 1, AWS D1.1, and API 1104 have all changed their PWHT provisions over the years, and a procedure written to a superseded edition is a finding waiting to happen.
2. Verify P-number and control thickness. Re-check material certs against the drawing. Measure actual thickness at the joint, not nominal plate callout. Control thickness includes weld reinforcement and the combined thickness of dissimilar joints.
3. Confirm PWHT is required and no exemption applies. ESAB notes that not all welds require PWHT; the need depends on material, thickness, service conditions, and the applicable code. Read the exemptions, not just the table.
4. Confirm the WPS was qualified with PWHT. PWHT is an essential variable under ASME Section IX QW-407. A procedure qualified without it cannot simply be applied with it.
5. Verify temperature range, hold-time basis, and ramp rates. The soak temperature must match the P-number, the hold time must be calculated from control thickness, and heating and cooling rates must respect code limits.
6. Mark the soak band and heated band on the component, not just in the procedure. Field crews need a chalk line or soapstone band, not a paragraph.
7. Verify thermocouple count, type, attachment, placement, and spares. Confirm control versus monitoring designation, couples at the weld centerline and band edges, and at least two spares on site.
8. Verify recorder calibration and chart scaling. Calibration certificates must be current, the chart range sensible for the target temperature, and the pen or digital log advancing correctly before the cycle starts.
9. Confirm insulation and weather protection. Wet insulation, wind, and rain wreck uniformity. Cover the assembly and keep a dry margin around the heated band.
10. Confirm supports and bracing for distortion control. One-sided local heating moves steel. Brace it, allow it to float where it needs to, and never quench.

During the Cycle

Monitor heat-up and hold in real time. Watch the ramp against the planned curve, a thermocouple that flatlines or spikes, and soak-band uniformity across top and bottom. If a couple fails mid-cycle, replace it and document the swap rather than accept a blind spot. On cooling, resist pulling insulation early; fast cooling reintroduces the residual stress you just spent hours removing.

After the Cycle

• Perform hardness checks at agreed locations. Record the instrument, location, surface preparation, and values. Wrong locations and decarburized surfaces produce false failures.
• Compile the documentation package. Chart records, calibration and thermocouple certificates, layout sketches, hardness results, and a signed closeout report. The PFI guidance is a solid reference for what belongs in that package.
• Disposition deviations through engineering, not field judgment. A missed hold, failed couple, or short soak goes to the engineer in writing, with a documented, qualified disposition attached.

The Broader Takeaway

PWHT is cheap insurance relative to a brittle fracture, a sour-service failure, or a rejected spool sitting on a dock. It is a documented, controlled process, not a ritual performed because a drawing said so.

So the principal answer, one last time: not every weld needs post weld heat treatment, but every weld that needs it must be done correctly and proved on paper. The purpose of the cycle is metallurgical, yet the evidence is administrative.

Years from now, the welder will have moved on, the inspector will have retired, and the drawings may be archived or lost. What survives is the chart: a strip of paper or data file showing what temperature the steel reached, for how long, and how gently it cooled. That chart is the only witness left. Make sure it tells the truth. If you are still building your foundation in the codes, start with our welding codes primer before your next procedure lands on the desk.

Frequently Asked Questions

What is the difference between post heating and post weld heat treatment?

Post heating is a relatively low-temperature hydrogen bake-out, typically around 450 F (230 C), performed immediately after welding before the joint cools to room temperature. Post weld heat treatment is a separate, higher-temperature stress relief cycle, usually 1100-1400 F depending on the material, performed to relax residual stress and temper the heat-affected zone.

Is post heat treatment mandatory for all welds?

No. Most routine welds in carbon steel below the code thickness threshold do not require it. Requirements are driven by material type, thickness, restraint, service environment, and the governing code such as ASME Section VIII, ASME B31.3, or AWS D1.1.

At what temperature is PWHT performed for carbon steel?

For P-No. 1 carbon steel, ASME B31.3 specifies roughly 593-649 C (1100-1200 F) for thicknesses above 19 mm, with a holding time of one hour per inch and a one-hour minimum. ASME Section VIII uses a similar range but the exact values come from the applicable table and notes.

How long must a weld be held at PWHT temperature?

Holding time is based on nominal or control thickness, commonly one hour per inch (25 mm) of thickness with a minimum of one hour for carbon steel. Thicker or more highly alloyed sections require longer holds, and some codes allow the time to be accumulated across multiple cycles.

Can PWHT be done locally instead of in a furnace?

Yes. Local heating with resistance pads, induction coils, or other controlled heat sources is common for field fabrication and repairs, provided the soak band, heated band, and gradient control band are correctly sized and the cycle is properly recorded.

Does post weld heat treatment increase or decrease strength?

It usually decreases strength slightly while improving toughness and reducing residual stress. For example, an E7018 deposit may fall from about 75 ksi as-welded to about 65 ksi after stress relief, which is why procedures must be qualified in the heat-treated condition.

What is reheat cracking and can PWHT cause it?

Reheat cracking is a failure that occurs during PWHT or elevated-temperature service in susceptible Cr-Mo and Cr-Mo-V steels, typically in the coarse-grained heat-affected zone. It is driven by stress relaxation combined with grain boundary weakening, and once it occurs the component generally cannot be salvaged by simply repeating the heat treatment.

Do stainless steel welds need post weld heat treatment?

Austenitic stainless steels generally do not require PWHT and are often exempt in ASME B31.3. However, if the service is aggressive and sensitization is a concern, a full solution anneal at high temperature followed by rapid cooling may be specified to restore corrosion resistance.

How do I prove a PWHT cycle was performed correctly?

The documented package should include the approved procedure, heater and thermocouple layout sketches, calibrated recorder charts showing the full heating, holding, and cooling cycle, thermocouple and instrument calibration certificates, and hardness test results where required.