Stainless and hygienic lines

Stainless steel piping in a food, dairy or beverage plant is welded for a different purpose from a steam line. Strength is rarely the issue; the walls are thin and the pressures modest. What matters is the inside surface. Every weld must be smooth, flush and fully fused inside, with no crevice, no oxidised scale, no pits and no step where milk, juice or cleaning chemicals can lodge and bacteria grow. The line must drain completely and be cleaned in place, by circulating hot cleaning solutions through it, without being taken apart. A weld that would pass on a water main can make a dairy line unusable.

This chapter covers how thin stainless tube is welded, compares orbital and manual welding, describes the test coupons that keep orbital welding honest, explains heat tint and how it is removed, and sets out the layout rules, slopes and dead legs, that let a hygienic line drain and clean.

Tube, not pipe

Hygienic stainless lines are usually built from tube sized by its outside diameter, such as 25.4, 38.1, 50.8, 63.5 and 76.2 mm, with walls of around a millimetre and a half, made to food-industry standards with a controlled inside finish. The thin wall changes the welding. There is no bevel and often no filler: the ends are cut square, faced so that they meet with no gap, and fused together by the tungsten arc, an autogenous weld, with the inside bead forming from the melted wall itself. The weld bead inside should be just visible, slightly convex or flush, the same width all round, and fully penetrated.

Because the wall is thin and there is no gap to fill, the squareness and cleanliness of the cut ends decide the weld. Tube is cut with saws or tube cutters that leave a square end, faced with a facing tool so that both ends meet completely, and deburred inside and out. Any gap, mismatch or out-of-square end makes the arc melt one side more than the other, and the inside bead wanders or fails to penetrate.

Orbital and manual welding

Thin tube can be welded by hand, rotating the tube or moving round it, but the best results and the most consistent ones come from orbital welding: a weld head that clamps round the tube, with the tungsten electrode rotating round the joint at a programmed speed while the power source changes the current in steps as it goes, to suit each position.

Figure 1Manual and orbital welding of 38.1 mm hygienic tube

	Manual gas tungsten arc	Orbital gas tungsten arc
Arc time per joint	about 4–6 minutes	about 1–2 minutes
Consistency	depends on the welder and their fatigue	repeats the programme exactly
Inside bead	varies in width and penetration	even width and penetration all round
Access	needs room for the welder's hand and view	needs room for the head round the tube
Records	welder's own notes	programme, current and travel logged for each weld
Skill	highly skilled tube welder	trained operator; programme developed by a specialist

Orbital welding does not replace skill so much as move it. A specialist develops and qualifies the weld programme for each tube size, wall and material; operators then make every weld with that programme, checking the fit-up, purge and electrode before each weld and the coupon at the start of each shift. The head is fitted at the joint in seconds, and the weld is made in a minute or two. On a job with hundreds of joints in a few tube sizes, orbital welding is faster and far more consistent. On a job with a handful of joints in many sizes, or where the head will not fit, skilled manual welding is still used.

Weld coupons

Orbital welding repeats whatever it is set to do, which is its strength and its weakness. If the tungsten has worn, the purge is poor, or the tube in today’s batch has a slightly different composition, every weld will come out the same, and the same wrong. The answer is the coupon: a test weld made on two short lengths of the same tube, with the same programme and settings, at the start of each shift, after any change of tube batch or electrode, and whenever the programme is changed. The coupon is cut open and the inside bead examined for penetration, width and colour, and the result is recorded before production welding begins.

Figure 2The contractor's start-of-shift orbital weld check

• Avoid: Electrodecorrect type and diameter, ground to the programme's tip angle, set to the arc gap
• Avoid: Programmecorrect for the tube size, wall and material; locked against changes
• Avoid: Purgeargon flow and oxygen reading at the vent within the procedure's limit
• Avoid: Coupon weldmade on tube from the batch in use; cut open and inspected inside
• Avoid: Inside beadfull penetration, even width, colour within the client's reference
• Avoid: Recordcoupon tagged, photographed and filed with the shift's weld log
• Do: ? Second checkafter any change of tube heat, electrode or programme

Heat tint

When stainless steel is heated in the presence of even a little oxygen, a thin oxide film forms on its surface, coloured from pale straw through gold, brown and blue to dark grey as the temperature and the oxygen rise. That film is heat tint. On the outside of a weld it is a cosmetic matter on many jobs; on the inside of a product line, the darker tints mean a surface depleted of chromium beneath the oxide, less resistant to corrosion, and rougher, so that it holds product and bacteria. The client’s specification sets the acceptable colour, often by a reference chart of photographed weld colours; a good purge gives an inside bead with little or no colour.

Heat tint that exceeds the limit is removed by pickling, with acid pastes or solutions that dissolve the oxide and the depleted layer beneath it, and the surface is then passivated so that its protective chromium oxide film reforms. On the inside of an installed line, pickling and passivation are done by circulating the chemicals through the system. The chemicals are hazardous and their waste must be handled under environmental rules, so a purge good enough not to need pickling is cheaper as well as better.

Keeping carbon steel out

Stainless steel resists corrosion only if its surface is clean. Particles of carbon steel embedded in it, from a grinding disc used on carbon steel, a steel wire brush, a carbon steel clamp or a bench where carbon steel was cut, rust and leave brown stains and pits that look like corrosion of the stainless itself. On a hygienic line they can be traps for bacteria too. The contractor’s stainless work is done in a separate area of the shop, with tools, brushes, discs, clamps and benches reserved for stainless and marked as such, and tube is stored off the floor away from carbon steel work.

Slopes and drainage

A hygienic line must drain completely when it is emptied, so that no product or cleaning solution is left standing in it. The lines are laid to a slope, falling towards the drain points, and the slope must be kept along every length and through every fitting.

Figure 3Fall along a sloped product line

fall = run length × slope

Run

30 m

Slope

1 in 100

Supports

every 2.5 m

1. fall = 30,000 × 0.01 = 300 mm from the high end to the drain
2. each support is 2,500 × 0.01 = 25 mm lower than the one before
3. a weld that kinks the tube by 1° adds or removes about 17 mm of height over a metre

The welds matter to the slope. A weld that shrinks unevenly can kink the tube by a degree or so, enough to create a low spot where liquid pools between two supports, or a high spot that traps air. Orbital welds, with their even heat, are less prone to this than manual ones, but supports are set to the slope with a level or laser before the tube is welded into them, and the line is checked after welding, with a level on the tube at each support, so that any low spot is found while the scaffold is still up and the fix is easy.

Dead legs

A dead leg is a branch, closed by a valve or cap, in which product sits without flowing when the main line is cleaned. Cleaning solutions circulating in the main barely reach into a long dead leg, and product left there spoils and contaminates the next batch. Hygienic designs keep branches short, measured in tube diameters from the main’s wall to the valve seat, with limits set by the client’s hygienic standard, and orient them so they drain. The welder’s part is to make the branch as designed, often with a short forged or pulled tee, and not to lengthen it with an extra length of tube to make the fit-up easier.

Distortion and expansion of stainless

Stainless steel expands about half as much again as carbon steel when heated and conducts heat poorly, so heat stays near the weld and the shrinkage is larger and more concentrated. Thin stainless tube distorts readily, and long lines that heat during cleaning grow more than carbon steel lines. The supports must allow the growth, with guides that let the tube slide and expansion loops or bends where long runs heat. Welding Distortion by the Numbers covers the shrinkage of welds in general.

Clamps, ferrules and flanges

Where hygienic lines must come apart, at pumps, valves, instruments and spool breaks for cleaning inspection, they use hygienic clamp fittings: a ferrule welded to each tube end, a gasket, and a clamp. The ferrule is butt welded to the tube with the same care as any other joint, and its inside must align with the tube’s bore. A ferrule welded slightly out of line leaves a step that traps product, so ferrules are welded with an alignment jig or the orbital head, never by tacking them by eye.

Records and traceability

Hygienic clients usually require a full weld record for product-contact lines: each weld’s number on an isometric drawing, the welder or orbital operator, the programme, the coupon for that shift, the tube heat numbers, the borescope image or video of the inside, and the colour assessment. The record proves that every weld meets the specification and lets any weld be traced if a problem appears in service. The contractor’s orbital power sources log each weld’s parameters, which are downloaded to the weld log at the end of each shift. Chapter 10 deals with the weld log in general.

The running case: the product lines

This is an illustration. The dairy’s product and cleaning lines had about 300 welds in 38.1 to 76.2 mm tube. The contractor used two orbital welding sets with heads for the sizes in use, one specialist who developed and qualified the programmes, and four trained operators. Each operator made a coupon at the start of each shift on each size to be welded that day; the coupons, borescope recordings and logged parameters went into the weld log. Supports were set to a 1 in 100 fall with a laser before welding. Manual tungsten arc welding was kept for a dozen joints at equipment where the head would not fit, made by the most experienced tube welder and inspected the same way.

Costing orbital welding

Orbital equipment is costly to buy or hire, and the programmes and coupons take time, so its economics depend on the number of joints. In this illustration, if an orbital weld saves about 4 minutes of arc time and about 10 minutes of total time per joint against a manual weld, and coupons and set-up cost about 45 minutes per operator per shift, the 300 joints save about 50 hours of welding but spend perhaps 30 hours on coupons and set-up over the job. The real gain is in consistency: the manual welds on similar earlier jobs had needed more rework after borescope inspection, and each reworked joint on a hygienic line means cutting out the weld, refacing, refitting, repurging and rewelding.

A failure: the drifted programme

This is an illustration. On another contractor’s job, an orbital operator changed the weld programme slightly to suit a batch of tube that seemed to weld differently, and then forgot to change it back. No coupon was made after the change. For two days, every weld on one tube size had a narrow inside bead with incomplete penetration over part of the circumference, invisible from outside.

The borescope inspection of the line found the problem, and about forty welds were cut out and remade. The programme lock and the coupon after any change, both in the contractor’s checklist above, exist to prevent exactly that.

Exercise: slopes and supports

A cleaning return line is 42 m long, laid to a slope of 1 in 100, with supports every 2.5 m. What is the total fall, and how many supports are there if the first is at the high end and the last at the drain? If a manual weld near the middle kinks the tube by 1.5°, roughly how much height change does that cause over the next 2.5 m span, and what is the risk?

The fall is 42,000 × 0.01 = 420 mm. Supports at 0, 2.5, 5.0 m and so on reach 40 m with 17 supports, and 42 ÷ 2.5 = 16.8 spans, so an 18th support is needed at the drain end, giving 18 supports, with the last span shortened to 2.0 m.

A 1.5° kink changes the tube’s direction by about tan 1.5° ≈ 0.026, so over a 2.5 m span it changes the height by about 2,500 × 0.026 ≈ 65 mm. The designed fall over that span is only 25 mm, so a kink that way could turn the tube uphill for a span, creating a low point where product and cleaning solution pool. The kink must be corrected by cutting and rewelding, or the supports reset, before the line is commissioned.