Preheat, hydrogen and cooling

The word preheat appears throughout the 1919 Airco booklet, but it does not mean what a modern welding procedure means by it. In the booklet, preheating is local and purposeful in a mechanical sense: heat the plate between two holes so it expands with the weld, then let both cool together. In a modern procedure, preheat is usually a specified minimum temperature for the whole joint region, chosen mainly to prevent hydrogen cracking in the heat-affected zone.

Both uses are legitimate. Confusing them is not. This chapter separates the two purposes and introduces the carbon equivalent, the simple number most often used to judge how much care a steel needs.

Two different reasons to heat the plate

Heating to share shrinkage. As Chapter 2 explained, a weld shrinks and the cold plate resists. Heating a region beside the weld makes it expand too, so it contracts with the weld instead of opposing it. The booklet used this when closing cracks and setting patches: heat the area marked on the drawing, keep it hot while welding, let it cool with the weld. The temperatures were judged by colour, such as red heat, and were not controlled or recorded.

Heating to slow cooling. In arc welding, hydrogen from moisture, contamination or consumables can enter the weld pool and diffuse into the heat-affected zone. If that zone cools quickly and forms a hard microstructure while it is restrained, cracks can form, sometimes hours after welding. TWI’s guidance on hydrogen cracking explains that preheat slows the cooling rate, allows some hydrogen to diffuse away and generally reduces hardness, and with it the susceptibility to cracking.

The second purpose is governed by procedure: a minimum preheat temperature over a defined width either side of the joint, a maximum interpass temperature, controlled consumables and, for susceptible steels, holding heat for a period after welding.

Figure 1Two reasons to preheat

	Sharing shrinkage, as in 1919	Preventing hydrogen cracking, today
Where	A chosen region beside the weld	A defined band either side of the joint
How hot	Judged by colour, such as red heat	A specified minimum, measured
Main purpose	Let nearby plate expand and shrink with the weld	Slow cooling, let hydrogen escape, limit hardness
Recorded	Rarely	In the procedure and the weld record

Carbon equivalent: a first measure of weldability

Steel’s tendency to form hard, crack-susceptible structures when welded rises with carbon and with alloying elements that increase hardenability. The carbon equivalent combines them into one number. TWI’s summary of the formulae gives the widely used IIW expression, with all values in weight per cent:

CE = C + Mn/6 + (Cr + Mo + V)/5 + (Ni + Cu)/15

TWI notes that steels with a CE below about 0.4 are generally not susceptible to heat-affected-zone hydrogen cracking when low-hydrogen consumables or processes are used. As CE rises, procedures typically call for more care, including higher preheat. The guidance mentions preheats as high as 200 °C for thick sections of high-CE steel, and refers to standards such as EN 1011-2 for recommended levels. Other formulae, such as Pcm, are often preferred for low-carbon modern steels, so the applicable code decides which to use.

The tank case: what is the plate?

This is an illustration. The tank plate has no surviving mill certificate. The workshop sends a sample from an offcut, removed during an earlier alteration, for chemical analysis. The illustrative result is: carbon 0.18%, manganese 1.20%, chromium 0.05%, molybdenum 0.02%, vanadium 0.00%, nickel 0.05% and copper 0.10%.

CE = 0.18 + 1.20/6 + (0.05 + 0.02 + 0.00)/5 + (0.05 + 0.10)/15

CE = 0.18 + 0.200 + 0.014 + 0.010 = 0.404

The plate sits right at the boundary TWI describes. The workshop’s procedure therefore specifies low-hydrogen consumables stored and handled to the manufacturer’s instructions, a modest measured preheat over a band either side of the groove, a maximum interpass temperature, and dye penetrant inspection repeated after a delay rather than only immediately after welding, because hydrogen cracks can appear late.

Figure 2Carbon equivalent of the tank plate

1. Carbon0.18 · 45%
2. Manganese ÷ 60.2 · 50%
3. Chromium, molybdenum, vanadium ÷ 50.014 · 3%
4. Nickel and copper ÷ 150.01 · 2%

Old plate deserves extra suspicion

Repair work often meets steel whose composition is unknown. Old plate may contain more carbon, sulphur or phosphorus than modern steel, may be laminated, and may have been heated, cold-worked or repaired before. Before welding a part whose integrity matters:

• find the original specification or a mill certificate if one exists;
• if not, test a sample for composition, and consider mechanical testing where strength or toughness matters;
• examine the plate near the repair for laminations or earlier repairs;
• be cautious with materials the welding procedure was not qualified on.

In the United States, the steam locomotive regulation discussed in Chapter 8 distinguishes carbon steel above and below 0.25% carbon for welding on unstayed boiler portions, requiring prior approval above that level. Whatever the jurisdiction, composition is not a detail to guess.

Cooling after welding

Cooling deserves as much thought as heating. A repair cooled quickly in a cold workshop, or quenched to save time, can harden and crack. Procedures may call for slow cooling under insulation, holding at temperature for a period to let hydrogen escape, or a formal post-weld heat treatment for stress relief where a standard or design requires it. The booklet’s practice of reheating areas after welding to relieve shrinkage was an early, unmeasured step in this direction. Modern heat treatment is controlled, recorded and justified by the applicable standard, because uncontrolled heating can damage as easily as it can help.

Where hydrogen comes from, and how it is kept out

Hydrogen cracking needs three things together: hydrogen in the weld, a susceptible microstructure in the heat-affected zone, and tensile stress. Preheat works on the second and helps the first. Controlling the hydrogen at its source is just as important.

The main sources are:

• Moisture in consumables. Low-hydrogen electrodes and fluxes absorb water from the air. Manufacturers specify storage, drying or baking conditions and the time an electrode may be exposed before it must be re-dried or discarded.
• Moisture and contamination on the joint. Condensation, rust, oil, grease, paint and cutting fluids all release hydrogen in the arc. A cold steel surface in a humid workshop can carry condensation that is invisible until the arc strikes it.
• The process. Some processes and consumables deposit less hydrogen than others. Consumables are commonly classified by their diffusible hydrogen level, with designations such as H5 or H10 indicating a maximum of about 5 or 10 millilitres of hydrogen per 100 grams of weld metal.

For the tank repair, the procedure therefore specifies a low-hydrogen consumable class, the storage and re-drying conditions from the manufacturer’s data sheet, and cleaning of the groove and a band either side to bright metal immediately before welding.

Thickness and restraint change the preheat

Carbon equivalent is only one input to a preheat decision. Methods such as those in EN 1011-2 also consider the combined thickness of the joint, the heat input and the hydrogen level of the consumable. Combined thickness reflects how many paths heat has to flow away from the weld: more paths mean faster cooling and a greater need for preheat.

For a butt weld between two plates of equal thickness, the combined thickness is the sum of the thicknesses measured over a short distance either side of the joint. The tank’s 10 mm butt repair therefore has a combined thickness of about 20 mm. A fillet weld attaching a 10 mm plate to the middle of another 10 mm plate has heat paths in three directions, two along the continuous plate and one into the attached plate, giving a combined thickness of about 30 mm. The same steel, consumable and heat input can need more preheat at the fillet than at the butt weld.

Figure 3What a preheat decision depends on

• Preheat temperature
  • Carbon equivalentcomposition of the steel
  • Combined thicknessheat paths away from the weld
  • Heat inputarc energy and process efficiency
  • Hydrogen levelconsumable class and condition
  • Restrainthow hard the joint is held

Restraint adds a further factor. A highly restrained repair, such as a crack excavation in the middle of a large plate, develops higher stresses during cooling than a free joint, so procedures for repairs sometimes specify more preheat than the same material would need in an unrestrained fabrication.

Measure preheat where it matters

A preheat requirement is only as good as its measurement. Practical points include:

• Measure at the right place. Procedures commonly specify measurement at a set distance from the joint edge, and where possible on the face opposite the heating, so that the reading represents the full thickness rather than a hot surface.
• Allow the heat to soak. After a gas torch or heating element is removed, the surface temperature falls as heat spreads into the plate. Measure after a short equalising time, as the procedure specifies.
• Use suitable instruments. Contact thermometers and temperature-indicating crayons are common. Infrared thermometers are quick but can misread on shiny or scaled surfaces.
• Heat the whole band. Preheat applied to a narrow strip leaves cold metal close by, which cools the weld faster than the reading suggests.
• Record the readings. The repair record should show the preheat and interpass temperatures actually measured, not only the values specified.

The booklet’s judgement by colour, red heat, was a measurement of a kind, but it described temperatures far above the preheats a modern procedure uses for carbon steel, and it measured the surface only. A modern record replaces the eye with an instrument and a number.

The upper limit: interpass temperature

Procedures also set a maximum interpass temperature: the hottest the joint may be when the next pass begins. Letting the joint accumulate heat pass after pass widens the hot zone, increases distortion and, in some steels, reduces strength and toughness in the heat-affected zone. Quenched and tempered steels are particularly sensitive, because reheating above their tempering conditions can soften them.

For the tank repair, the welder checks the joint temperature before each pass and waits if it is above the procedure’s limit. The waiting time is part of the job, and the sequence plan from Chapter 4 helps by giving the welder somewhere else to work while one area cools.

Delay the final inspection

Because hydrogen cracks can form hours after welding, an inspection carried out while the joint is still warm can miss them. Standards for susceptible steels and critical joints may therefore specify a hold time before final inspection, sometimes of a day or more. Even where no delay is mandated, inspecting a repair again after it has fully cooled, as the tank procedure requires, is cheap insurance.

Post-weld heat treatment, briefly

Some repairs go one step further than controlled cooling. Post-weld heat treatment heats the welded region, or the whole part, to a specified temperature, holds it there for a time related to the thickness, and cools it at a controlled rate. Its usual purposes are to relax residual stresses towards a lower level, to temper hard microstructures in the heat-affected zone and, in some steels, to restore toughness.

Whether it is required is not a workshop preference. Pressure equipment standards and design codes specify when post-weld heat treatment is mandatory, typically depending on the material, the thickness and the service, and they specify the temperature, hold time, heating and cooling rates and the width of the heated band for local treatment. A repair to a part that was heat treated after its original welding may need the same treatment again, or an approved alternative, to keep the part’s condition consistent.

Local heat treatment of a repair in a large structure brings its own risks. Heating a band across a plate creates thermal gradients that can distort the part or induce new stresses if the band is too narrow or the heating uneven. For that reason it is planned with thermocouples placed to the procedure, controlled heating equipment and a record of the temperature cycle actually achieved.

The booklet’s practice of reheating areas after welding to relieve shrinkage strains was a distant ancestor of this process, done by eye with a torch. The modern version keeps the intention and replaces judgement with specified temperatures, measured cycles and a written record.

A preheat and hydrogen checklist

Before welding a repair whose cracking would matter, confirm that:

1. the steel’s composition is known, or tested, and its carbon equivalent calculated;
2. the combined thickness and restraint of the joint have been considered;
3. the consumable’s hydrogen class and storage condition match the procedure;
4. the joint and a band either side are clean and dry;
5. preheat is applied across the full band and measured at the specified point;
6. interpass temperature limits are set and will be checked;
7. post-weld cooling and inspection timing are planned.

None of these steps appears in the 1919 booklet, and nearly all of them address mechanisms its authors could not yet see. That gap is the clearest measure of how much welding repair has changed.

Exercise: compare two steels

Calculate the IIW carbon equivalent for a plate with 0.12% carbon, 0.90% manganese, 0.03% chromium, 0.01% molybdenum, no vanadium, 0.03% nickel and 0.06% copper. Compare it with the tank plate.

CE = 0.12 + 0.90/6 + (0.03 + 0.01 + 0)/5 + (0.03 + 0.06)/15 = 0.12 + 0.150 + 0.008 + 0.006 = 0.284

This steel is well below 0.4 and would normally be easier to weld than the tank plate at 0.404. Explain why a low CE does not remove the need for low-hydrogen practice and a qualified procedure, and why thickness and restraint still affect the preheat a procedure specifies.