Pressure Vessels & Storage Tanks: The Complete Engineering Guide — ASME Section VIII Div. 1 Design and Fabrication, from the UG-27 Shell Equation to the UG-37 Reinforcement Ledger, PWHT and Hydrotesting, API 650 Tank Sizing, PESO Licensing in India, and What Vessel Work Costs in 2026

Pressure vessel design and fabrication with worked numbers: ASME VIII-1 shell, heads, nozzle reinforcement, radiography economics, PWHT, hydrotest, API 650 tanks and India costs.

·

Pressure Vessels & Storage Tanks: The Complete Engineering Guide

A pressure vessel is a bomb that works for a living. Our worked example in this guide — a 2 m inside-diameter, 6 m long separator at 12 bar — contains about 21 m³ of gas whose expansion energy, by the simplest engineering measure E \approx P \cdot V, is roughly 25 MJ; counted through a full expansion to atmosphere it is closer to 60 MJ. Either way, that is the energy content of kilograms, not grams, of high explosive. Now place that object beside a road, fill and empty it on a schedule for thirty years, and attach pipes that push and pull on it — and you have the entire engineering problem of the pressure vessel: contain an explosion, permanently, with an inspection trail that proves you did.

Pressure vessels and storage tanks. The distinction is jurisdictional as much as technical: a pressure vessel under ASME Section VIII is designed for more than 15 psi (103 kPa) of internal or external pressure; a storage tank under API 650 is a thin-plate, atmospheric structure whose enemy is liquid head, wind, and settlement; and a boiler is an unfired steam generator that belongs to Section I. The steam boiler guide covered the fired side; this guide covers everything that holds pressure without a flame under it — the separator on a process unit, the air receiver in a compressor room, the LPG bullet in a gas yard, the cryogenic tank behind a hospital's oxygen supply, and the cascade that fills a CNG bus at 250 bar.

It works the whole chain the way a designer, fabricator, or project buyer needs it: the stress basis, the shell equation, the head menu, the joint-efficiency and radiography economics, nozzle reinforcement area accounting, low-temperature rules, supports, shop practice, heat treatment, hydrotesting, atmospheric tank sizing — and finishes with what it all costs in India in 2026, worked on a real vessel weight.

1. The Rulebook, the Stamps, and Who Builds What

ASME Section VIII has three divisions, and their boundaries are economic:

Division · Rules · Typical realm

Div. 1 · Design-by-formula: prescriptive equations, generous margins, well-trodden · ~90 % of the world's vessels — separators, receivers, drums, bullets, up to ~100 bar and moderately thick walls

Div. 2 · Design-by-analysis: mandatory stress analysis, fatigue evaluation, tighter margins, stress tables about 1.4× higher (Rm/2.4 vs Rm/3.5 on tensile) · large, thick, cyclic, or novel vessels where saving 30 % of the wall pays for the analysis

Div. 3 · Impulse and very high pressure, above roughly 70 MPa (10 ksi) · LDPE reactors, waterjet intensifiers, high-pressure hydrogen

The stamp on the nameplate is the whole point. A "U" stamp means the shop holds an ASME Certificate of Authorization, follows a written quality system with an Authorized Inspector (AI) holding hold-points, and ships a U-1 Manufacturer's Data Report that traces every plate heat number to every weld and every test. In India, the same physical vessel goes down one of two regulatory paths: for projects and export, ASME U with National Board registration; for compressed gas storage, PESO under the SMPV(U) Rules, 2016, which accept design and construction to IS 2825, ASME VIII-1/2, PD 5500, EN 13458/13530 or AD 2000 — but add their own licensing, workshop approval, and stage-wise inspection regime on top [1][4].

That dual path matters commercially. A generic MS tank from a regional shop might leave the yard at a fraction of the cost of the same geometry with a code stamp and a document package. The difference buys predictability: a paper trail that a refinery inspector, a PESO officer, or a German insurer can each audit against their own rulebook.

2. Where the Allowable Stress Comes From

Everything starts with one number. The maximum allowable stress S in ASME VIII-1 is the lowest of these criteria (Section II, Part D, Appendix 1) [2]:

where R_m is specified minimum tensile strength and R_e is specified minimum yield strength — both evaluated at temperature below the creep range. For the workhorse material of this entire industry, SA-516 Gr. 70 plate (R_m min 485 MPa, R_e min 260 MPa [9]):

Historical footnote that explains old textbooks: the factor on tensile strength was 4.0 until the 1998 code cases and the 1999 Addenda relaxed it to 3.5 — a 14 % free stress increase that, for carbon steels, was mostly trapped by the \frac{2}{3}R_e ceiling anyway [2]. The relaxed factor is why a modern SA-516-70 vessel is a touch lighter than its 1995 ancestor, and why a thirty-year-old company standard may still say "use 125 MPa".

What the margin buys: a simple pressurized cylinder's failure strength is close to the material's tensile strength, so an R_m/3.5 basis means an ideal vessel bursts at ≥ 3.5 × design pressure. A hydrotest at 1.3 × MAWP leaves roughly a 2.5–3 × margin to burst in the tested condition. Those are the two numbers to remember when a customer asks "what happens if it goes to twice the pressure?" — the answer is "the 3.5, not the 1.3, is what keeps you alive."

3. The Shell: UG-27, Worked

The cylindrical shell under internal pressure is governed by UG-27(c). Hoop stress splits the wall lengthwise, so it governs; the longitudinal-stress case is checked as the minority load [1]:

with P internal pressure, R inside radius, S the allowable stress above, and E the joint efficiency for the seam being checked. Both hold while t \le R/2, equivalently P \le 0.385\,SE — for SA-516-70 that ceiling is 0.385 \times 138 = 53 MPa, i.e. 531 bar. Above it, the thin-wall simplification dies and you move to the thick-wall (Lamé) equations of Mandatory Appendix 1-2.

Worked example: our 2 m ID, 12 bar separator

Design conditions: P = 1.2 MPa (12 bar), R = 1000 mm, SA-516-70 at S = 138 MPa, full radiography so E = 1.00.

Circumferential (governing):

Longitudinal (the minority case):

The 2:1 ratio between the two is the whole reason cylindrical shells crack along their length when they fail — and why the longitudinal seam earns the tighter inspection. Take 8.74 mm, add a 3 mm corrosion allowance for wet service: 11.74 mm required, and the plate order rounds to 12 mm.

The back-calculation a vessel welds its nameplate with: with 12 mm plate and 3 mm allowance consumed, the effective wall is 9 mm, and MAWP is recovered from the same equation —

The vessel is stamped 12.3 bar at the design temperature, every other check (heads, nozzles, test) hangs off that number, and the 0.3 bar of headroom over the 12.0 bar design is deliberate: passivation of margins is how the arithmetic survives procurement reality.

Two economies hide in this section. First, the sphere: run the hemispherical-shell equation t = \frac{PR}{2SE - 0.2P} and the same 2 m radius needs only 4.35 mm — half the wall for the same pressure, because every direction on a sphere is a circumferential direction: the membrane stress is a uniform PR/2t in every fibre, with no hoop-versus-longitudinal split to double it. Spheres pay it all back in forming cost and only win at large volumes — LPG spheres are big for exactly this reason, and small ones are always bullets. Second, the longitudinal check reminds you that a vessel wall is a compromise: designers spend steel only where the governing stress demands it, which is why the corrosion allowance (and not pressure) often sets the schedule, exactly as it did in the piping guide.

4. Heads: Three Ways to Close a Cylinder

The head is where pressure stops being easy. UG-32 gives the formed-head menu, and the required thicknesses for our same vessel (D = 2000 mm, P = 1.2 MPa, S = 138 MPa, E = 1.0) separate the options by a factor of 3.5 [1]:

Head type · Equation · Required t (no CA) · Depth for 2 m ID · Character

Hemispherical · t = \dfrac{PR}{2SE - 0.2P} · 4.35 mm · 1000 mm · half a sphere: thinnest, tallest, most forming labour

2:1 Ellipsoidal · t = \dfrac{PD}{2SE - 0.2P} · 8.70 mm · 500 mm · the industry default: balanced

Torispherical (F&D) · t = \dfrac{0.885PL}{SE - 0.1P} · 15.40 mm · ~340 mm · shallowest, easiest to form, thickest wall

Conical · t = \dfrac{PD}{2\cos\alpha\,(SE - 0.6P)} · — · — · transitions, hopper bottoms; often needs a knuckle

Read the table as a three-way trade in depth, thickness, and forming labour:

Two fabrication realities the equations don't show. First, *formed heads are ordered to a minimum thickness after forming — the steel thins as it stretches over the knuckle, so a head that must finish at 12 mm is typically spun or pressed from plate one or two sizes heavier, and the dished-end supplier, not the vessel shop, owns that guarantee. Second, the knuckle radius* on a torispherical head is a stress concentrator (historically the origin of many head failures, back when knuckles were too tight), which is why codes set minimum knuckle radii and why 2:1 ellipsoidal became the default as soon as pressing technology allowed it.

For our separator, both heads are 2:1 ellipsoidal: t = 8.70 mm + 3 mm CA ≈ 11.7 → 12 mm finished, formed from heavier stock. Cylinder plus two heads plus nozzles is the entire pressure envelope, and its weight — shell ≈ 3.6 t, heads ≈ 0.8 t — is the first line of every quotation.

5. Joint Efficiency: The Radiography Tax You Can Choose to Pay

E in every equation above is not a property of the steel. It is the credit the code gives a weld for having been inspected — and it is the single most powerful lever on a vessel's weight [1].

For a Type 1 double-welded butt joint (the standard configuration for shell seams and head closures):

Examination of the joint · Joint efficiency E · Our shell t (±3 mm CA) · Plate chosen

Full radiography — every inch of every seam filmed · 1.00 · 8.74 + 3 = 11.74 mm · 12 mm

Spot radiography — a sample length per UW-52 · 0.85 · 10.29 + 3 = 13.29 mm · 14 mm

No radiography — visual only · 0.70 · 12.52 + 3 = 15.52 mm · 16 mm

Follow the consequences. Skipping radiography entirely doesn't just remove a cost line — it moves the vessel from 12 mm to 16 mm plate: one-third more shell steel, heavier heads (the head equations carry the same E), thicker welds with more passes, heavier lifting, and a stiffer, heavier vessel — all to avoid an inspection whose full scope, at Indian rates, is likely cheaper than the steel delta alone. Run the numbers on our vessel: the shell plate area is about 38 m², so the 12 → 16 mm step adds roughly 1.2 tonnes of plate (~₹95,000 at ₹80/kg with cutting allowance), before counting the extra welding hours on every seam. Full radiography of the main seams — perhaps 80–120 films on this vessel — is routinely quoted below that delta.

This is why the economics of NDT, worked in the non-destructive testing guide, are really the economics of steel: E is a dial that trades inspection for plate, and the efficient design point is almost always "shoot the welds, keep the thin wall." The exception is the small, low-pressure vessel where 16 mm is already what the minimum-thickness rules demand — there, skipping RT costs nothing structurally, and only the owner's confidence is the question.

Two rules organise the whole inspection plan. First, the code and the owner both push upward with severity: lethally toxic service, hydrogen service, thick walls, and cyclic duty carry full radiography on the main seams as near-default specification, and spot radiography is defined and counted precisely (UW-52) so that "0.85 credit" is not negotiable after the fact. Second, radiography discovers fabrication, not design: the films are a census of the welder's discipline — which is why progressive regimes and reputation pricing exist — and on a U-stamped vessel the records join the U-1 package where they outlive everyone who made them.

6. Materials: The Weather, the Chemistry, and the Cold

Plate selection for vessels follows the same logic as pipe selection: chemistry first, temperature second, then economics. The workhorses:

Material · Standard · S (ambient) · Serves · Watch

SA-516 Gr. 70 · plate · 138 MPa · the default pressure-vessel carbon steel: separators, receivers, tanks of every kind · needs corrosion allowance; brittle below its UCS-66 curve

SA-516 Gr. 70N (normalized) · plate · 138 MPa · same, but impact-tough: Curve D, exempt to −48 °C at ≤ 1 in wall [10] · premium over as-rolled

SA-537 Cl. 1/2 · plate · ~152–165 MPa · larger/thicker vessels where the R_m/3.5 basis pays · tighter welding controls

SA-240 304 / 316L · plate · ~138 / ~138 MPa · hygienic, chemical, cryogenic, pharma service · chloride pitting — see the corrosion guide; price

SA-285 C / IS 2062 · plate · ~136 MPa · water, air, low-hazard tanks · no code impact envelope; check service

SS-clad carbon plate · roll-bond/overlay · composite · big vessels where solid stainless would double the cost · bond integrity testing; weld overlays at seams

Around the edges live the specialists: 9 % nickel or austenitic stainless for LNG and cryo, 2¼Cr-1Mo for hot hydrogen, NACE HIC-resistant plate for sour service (SA-516 in the HIC version carries a 15–25 % price premium for ultra-low sulphur and mandatory TM0284 testing [11]), and duplex for chloride-rich waters.

The cold rules deserve their own paragraph, because they close the door on innocent designs. Carbon steel loses toughness as it cools and can shatter rather than yield. ASME's answer is the UCS-66 impact-test exemption curves A through D [10]. Given the material's curve and the governing thickness, the curves give the lowest temperature at which the steel may be used at full design pressure without Charpy testing. SA-516-70 (normalized) sits on Curve D — exempt to about −48 °C at one-inch wall; SA-106-B pipe sits on Curve B — only to about −29 °C. Two practical consequences:

7. Openings: The UG-37 Reinforcement Ledger

Every hole in a pressure boundary removes load-bearing metal. The code's answer (UG-37) is the area-replacement method: whatever cross-sectional area of shell was cut away in the opening must be replaced, within a strict zone around the opening, by spare metal from the shell, the nozzle neck, the attachment welds, and — if those fall short — a reinforcement pad [1][12].

The required area:

where d is the finished opening diameter (nozzle bore, in the corroded condition), t_r the required shell thickness from UG-27 (calculated at E = 1; no corrosion allowance in t_r itself — the method assumes seamless metal at the opening), and F a factor that is 1.0 for the normal radial nozzle. Available area comes from four sources, all clipped to the "limits of reinforcement" of UG-40 (each side: the larger of d or R_n + t_n + t; height: \min(2.5t, 2.5t_n)):

Worked ledger: a 6-inch nozzle on the 12 mm shell

Inputs: d = NPS 6 bore = 168.3 - 2 \times 7.11 = 154.1 mm (Sch 40 neck), t = 12 mm, t_r = 8.74 mm, same material everywhere (f_r = 1). Checking as-new (the mandatory final check runs in the corroded condition, where t shrinks by the CA and d grows — same arithmetic, meaner inputs):

Total available ≈ 812 mm² against 1,347 mm² required — the opening is not reinforced. Now watch the designer's three moves, each of which re-runs the same column:

Neck wall (NPS 6) · t_n (mm) · A_2 (mm²) · Available (with A_1 + welds) · Outcome

Sch 40 · 7.11 · 229 · 812 · deficit 535 mm² → repad needed

Sch 80 · 10.97 · 565 · 1,148 · deficit 199 mm² → small pad, or

Sch 160 · 12.70 · 722 · 1,305 · deficit 42 mm² → weld credit absorbs it; self-reinforced in practice

The repad for the Sch 40 case: A_5 needed 535 mm² at t_p = 12 mm gives a ring 45 mm wide — a pad about 215 mm OD welded around the neck. And the alternative that skips pads entirely: thicken the shell from 12 to 16 mm and A_1 alone (154.1 \times 7.26 = 1{,}119) plus neck and welds clears the ledger — at the price of one-third more shell steel everywhere, forever.

That is the real design conversation behind every nozzle on every vessel: a ₹3,000 pad with an extra weld and an extra NDT joint, versus hundreds of kilograms of shell plate to make the pad unnecessary. The pad usually wins; self-reinforced designs win when the nozzles are many and small, or when a spec simply bans repads to avoid crevice corrosion (a pharma or food-service preference — see the hygienic notes in the piping guide). Two guards on top: closely spaced openings share and compete for reinforcement (UG-42 — their zones may overlap and must be checked jointly), and large openings above the UG-36 size limits move out of the simple method entirely.

Nozzle loads are the sleeper issue. The internal-pressure ledger above treats the nozzle as a witness to pressure alone; real piping hangs on it — weight, thermal movement, occasional loads. VIII-1 allows external loads by analysis (or by the convenient WRC-107/297 bulletin methods), and a vessel with an overhung riser or a long, hot line will have nozzle reinforcement governed by bending rather than pressure. This is the handshake where pressure-vessel and piping engineering must actually talk, and it is worth real money at the fabrication stage — the alternative is discovering it during a site walkdown with the vessel already set.

8. Supports, Wind, and the Things That Hold It Up

A vessel is a pressure boundary and a structure. Horizontal vessels sit on two saddles; the classic analysis is Zick's method (1951), still essentially the industry's language: saddle reaction Q at each end, the shell bending as a beam between supports, and the famous "saddle horn" — the stress peak where the saddle's edge meets the shell, driven by the rigid saddle restraining the shell's natural bulge. Consequence: saddles are set in a little from the tangent lines (spacing roughly 0.4–0.5 L is the rule-of-thumb neighbourhood), the saddle contact is widened (or wear plates added) on thin shells, and one end is anchored to the foundation while the other slides — because a 6 m carbon-steel vessel grows several millimetres on its way from 20 °C to 120 °C, and if both ends are bolted down, the foundation loses a negotiation it never wanted.

Vertical vessels are towers: skirts (a cylindrical shell that carries the load in compression and gets burned — thickest — where wind bending and seismic overturning concentrate, near the base), legs (four I-beams with base plates for small vessels), or lugs (the smallest). The humbling fact of tower design: for most tall columns, the skirt and anchor bolts are sized by wind or earthquake, not by pressure. In India the loads come from IS 875 (wind) and IS 1893 (seismic); internationally ASCE 7 supplies the base shears, and the vessel code's own provisions get you to the stresses. A column designed only for its pressure will be elegant, correct, and flat on its side in a cyclone.

Lifting is its own mini-code: lifting lugs and trunnions designed for the empty weight times a dynamic factor, with shackle geometry and weld sizing to match — because the last thing any fabricator wants is to discover the lift analysis at 7 a.m. on the day the crane arrives.

9. From Plate to Nameplate: The Shop Sequence

A pressure vessel is built in an order that is itself code, because each stage inspects the previous one. Walking the shop:

  1. Cutting and edge preparation. Plates cut by CNC plasma or oxy-fuel (waterjet for thin stainless where oxidation matters), edges beveled to the joint geometry; the weld preparation itself is often a first inspection point, because a bad bevel is a bad root waiting to happen.
  2. Rolling. Shell courses formed on three-roll pyramid rolls (cheap; needs pre-bending the ends on a press because pyramid rolls leave flat lead-ins) or four-roll rolls (ends bent in the same pass, the modern default). The plate is rolled with the weld seam oriented as designed, the longitudinal seam fitted and tacked — strongbacks, wedges and dogs holding the geometry.
  3. Heads and fit-up. Dished heads arrive formed and trimmed (dished-end shops are their own industry); the head-to-shell girth seams are fit with careful alignment control (UW-33): for ordinary vessel thicknesses up to 38 mm the max mismatch is 6.4 mm, tightening to 3 mm for heavy 38–51 mm walls [1]; beyond that it's re-fit or formal analysis, which is never cheaper.
  4. Welding. The root runs are typically GTAW or SMAW (low-hydrogen electrodes on carbon steel), fill and cap by submerged arc (SAW) on longitudinal and girth seams — the workhorse of vessel shops: an arc buried under granular flux, running at several hundred amps, leaving smooth, dense weld. Root side back-gouged or ground to sound metal, magnetic-particle or penetrant checks on roots where specified. The welding is engineered to the same rules as everything else: WPS/PQR to ASME IX, welder qualification, and the heat input discipline the welding guide works through.
  5. Nozzles, manways, internals. Shell openings cut, nozzles set (set-through the shell more often than not, for the reinforcement arithmetic of Section 7), reinforcement pads welded where the ledger demanded them, internals and wear plates fitted last.
  6. The three gates, in order: radiography → post-weld heat treatment (if required) → hydrostatic test. This sequence is not arbitrary: RT before PWHT catches fabrication defects while they're cheap to cut out; RT after PWHT (the specification's choice for stress-relieved vessels) verifies that the thermal cycle didn't crack anything; and the hydrotest is the final integrated proof of everything.

10. The Three Gates: RT, PWHT, Hydro

Radiography is the census described in Section 5 — full, spot, or none, with E following. What matters operationally is that the plan is written before the first arc: which seams, which percentages, which acceptance standard the films will be judged against, and what happens on a reject (repair, re-shoot, and quite possibly a second round of the same on the neighbouring welds, depending on regime). The NDT methods themselves — RT, UT, phased array, whether the defect is planar or volumetric — are the subject of their own guide.

Post-weld heat treatment is where carbon steel pays its debt to the welding arc. Around every weld, the steel is left hot, constrained, and hard in the heat-affected zone — residual stresses at yield, hydrogen possibly dissolved, toughness locally degraded. PWHT is a controlled furnace cycle that relaxes the stresses and tempers the metallurgy. The triggers, for the main carbon steel family [1][3][5]:

Rule · Value

PWHT required (UCS-56) for P-No. 1 carbon steel (SA-516-70, SA-106-B) · when nominal weld thickness exceeds 38 mm (1.5 in)

Soak temperature · 595–650 °C

Hold time · 1 hour per 25 mm of thickness (minimum 1 hour for all thicknesses)

Heating / cooling rates · controlled: ≈ 220 °C/h per 25 mm heating above 315 °C; ≈ 275 °C/h per 25 mm cooling, all thermocouple-logged

Why else PWHT gets specified · sour service (hardness control, e.g. NACE MR0103 environments), service temperatures that embrittle, codes/specs that simply require it

Our 12 mm separator dodges the furnace entirely — one of the quiet economics of moderate-wall carbon steel. Cross 38 mm, or move to an alloy steel (P-No. 4 Cr-Mo steels are PWHT from ~16 mm up; some are "all thicknesses"), and the project acquires a furnace cycle: the vessel must be fully welded (attachments included — fillets welded after PWHT are a code violation and a punch-list item), thermocouples are placed by the code's map, and the treatment chart joins the document package. The metallurgy of what that cycle actually does to the steel is worked in the heat treatment guide; the project point here is that PWHT is a schedule event, not a cost line — a vessel that needs the furnace can't ship on the day it needs to.

The hydrostatic test (UG-99) is the final proof. The test pressure [1][6][8]:

For our vessel, testing at ambient where S_{test} = S_{design}, the ratio is 1 and:

The rules around the number are worth as much as the number:

Then the last gate is documentary: the nameplate (U symbol, MAWP, MDMT, design temperatures and the year), the U-1 data report signed by manufacturer and AI, and the traceability chain — plate MTRs (EN 10204 3.1 minimum for good work), weld maps with filler-lot numbers, NDT reports, PWHT chart, hydro chart. On a code vessel, that package is the product as much as the steel is; it is what lets a plant forty years from now prove the vessel was born honest.

11. Storage Tanks: The Wider, Thinner World of API 650

Above ground, at atmospheric pressure, the containment problem changes character. An API 650 tank is not a thin pressure vessel — it is a stiffened liquid container whose wall resists hydrostatic head, whose roof carries rain and vacuum, and whose enemies are wind buckling, seismic sloshing, foundation settlement, and corrosion rather than internal pressure. But the design method is elegant in the same way, and it uses the same arithmetic instincts.

API 650's one-foot method (valid up to 61 m diameter; larger tanks use the variable-design-point method) sizes each shell course from the liquid head at a design point 0.3 m above the bottom of that course [3]:

with D diameter (m), H design liquid level to the course bottom (m), G specific gravity, S_d and S_t allowable stresses for design and hydrotest conditions, and CA corrosion allowance. The allowable stresses come from the tank steel tables: S_d is the lesser of \frac{2}{3}R_e and \frac{2}{5}R_m; S_t the lesser of \frac{3}{4}R_e and \frac{3}{7}R_m — for a common tank steel like A283-C (205/380 MPa), S_d = 136.7 MPa, S_t = 153.8 MPa [3].

Worked fleet: a 30 m tank, 12 m of product at SG 0.85, CA 2 mm

Course-by-course, with the governing design thickness (hydrotest numbers stay lower here — the water test is easier on the steel than the product because the design allowables are tighter than the test allowables, and G < 1):

Course (bottom → top) · H at course bottom (m) · t_d (mm) · t_t (mm) · Ordered (mm)

1 · 12.0 · 12.70 · 9.51 · 13

2 · 10.0 · 10.87 · 7.88 · 11

3 · 8.0 · 9.04 · 6.26 · 10

4 · 6.0 · 7.21 · 4.63 · 8

5 · 4.0 · 5.38 · 3.01 · 6

6 · 2.0 · 3.55 · 1.38 · 6 (minimum)

Aggregate shell thickness: 54 mm across six 2 m courses, against 78 mm if every course were built as thick as the bottom one — the step-taper saves roughly 35 tonnes of steel on this tank, which at ~₹60/kg is about ₹21 lakh of plate. That is why API 650 shells are always coursed with heavier plate at the bottom, and why the "minimum thickness" table [3] exists as a floor: 5 mm under 15 m diameter, 6 mm from 15–36 m, 8 mm from 36–60 m, and 10 mm at 60 m and above (all exclusive of corrosion allowance). Our top two courses sit on the 6 mm floor — the design equation wants 3.55 mm at the top, and the code's handling-and-buckling minimum overrules it.

Around the shell, the rest of the tank's anatomy: a bottom of lap-welded plates (with annular plates — thicker rings under the shell — where settlement and shell stresses demand), a roof (self-supporting cone, or the floating roof that rides the product for volatile stocks), shell stiffening rings against wind, and a foundation that is half the project — ring wall, sand pad, and the water test: fill it, hold it, measure settlement for days, before anyone trusts it with product. Cryogenic and low-pressure cousins branch off: API 620 for low-pressure refrigerated service, EN 14620/BS 7777 family for full-containment LNG, where the tank stops being a steel structure and becomes a steel box inside a concrete one.

The commercial context is India's, and it is loud. Refinery and petrochemical capacity keeps expanding; LPG bullet farms answer to PESO; the CBG programme is fabricating 250-bar cylinder cascades for hundreds of plants (Section 13); and the cryogenic end — where Indian shops build everything from dewars to 690 KL vacuum-insulated storage tanks for the world's first commercial liquid-air energy store — has made "vacuum-insulated" a normal line item in Indian procurement [7]. Tanks and vessels are the most fabricated high-value objects on any Indian industrial site, and the shop that wins is the one whose quality system, not its price list, is the differentiator.

12. India 2026: Demand, Licensing, and the Rupee Walk

The demand drivers are public and specific. Refining and petrochemical maintenance/replacement rolling stock consumes the classic vessel tonnage; the compressed biogas build-out adds a new, high-pressure fleet — as of August 2026 India had 217 commissioned CBG plants against the 5,000-plant SATAT ambition, with the GOBARdhan relaunch (₹23,731 crore, August 2026 Cabinet approval) and a blending mandate ramping from 1 % in FY26 to 5 % by FY29 at an administered price near ₹2,110 per MMBtu [13][14]. Every CBG plant needs cascade storage at 250 bar (up from the older 200-bar standard — 25 % more gas mass per cylinder volume, five-stage compression, relief devices at 1.1 × working pressure, everything rated to match) [15], and every cascade is fabricated pressure hardware with its own PESO file. Cryogenics exports: Inox India alone booked ₹1,354 crore of revenue in FY25 from tanks, dewars, and transport systems at −196 to −253 °C [7]. And hydrogen is arriving — storage and transport vessels are where the national mission's money meets the same arithmetic in this guide, at higher pressure.

PESO/SMPV licensing, as it actually works [4]:

What it costs — the anchors. Indicative 2026 levels, ex-GST [16][17]:

Item · Price

MS plate, IS 2062 (non-pressure) · ₹53–70/kg

SA-516 Gr. 70 boiler-quality plate · ₹75–95/kg (listings to ~₹120 for small/cut lots)

SS 304 plate · ₹265–290/kg

Vessel fabrication conversion (MS, rolled/welded/tested, shop scope) · ₹110–180/kg (services listings)

Stainless fabrication conversion · ₹115–400/kg depending on scope

Fully code-documented vessel (design + material + fab + NDT + stamp + docs) · ₹180–280/kg for carbon steel

The rupee walk on our 6-tonne separator

Shell 3.6 t + heads 0.8 t + nozzles/manway/saddles/handling ≈ 6 t all-in:

The same vessel in 316L lands at roughly 2–2.5 × that number, because stainless plate alone is 3.5 × the price and weld efficiency and cleanliness rules add hours on every joint. And the U-stamp premium — the AI hold points, the quality system, the paper — is commonly quoted at 10–20 % over identical un-stamped fabrication; on this vessel that's ₹1.3–3 lakh of confidence, and no refinery will accept the discount version.

13. Specifying a Vessel Job

If you are sending vessel work to a shop — on FabFlow's manufacturer network or anywhere else — the following nine items separate a real quote from a wish:

  1. Design conditions sheet — MAWP, design pressure and temperature, MDMT, corrosion allowance, fluid and service (including lethal/sour flags), snow/wind/seismic site data.
  2. Governing code and stamp — ASME VIII-1 U-stamp, VIII-2, IS 2825 + PESO, or commercial; state it once and it prices the whole documents package.
  3. Material specification with certification — grade, standard, and certificate level (EN 10204 3.1 minimum); impact-test requirements if the MDMT demands them.
  4. Radiography decision in writing — full/spot/none on which seams, and the E values the design assumed. The drawing's thickness and the shop's NDT scope must tell the same story.
  5. PWHT requirements — triggered or not, and by which clause; if triggered, the cycle takes the schedule.
  6. Nozzle schedule with loads — sizes, ratings, projections, and the external loads from the connected piping; reinforcement pads or heavy-bar necks where the ledger says so.
  7. Inspection & test plan (ITP) — hold points at material identification, fit-up, root, final, PWHT, hydro; who signs what.
  8. Surface preparation and painting system — blast profile, primer, DFT per coat, and the stainless contamination rules (no carbon-steel tooling) where they apply.
  9. Document package definition — U-1/design dossier, weld map with heat numbers, NDT and PWHT records, hydro chart, dimensional reports, and packing/marking for transport.

The unglamorous truth of this discipline: every crisis in vessel fabrication is a documentation crisis in disguise. The steel doesn't lie; the paperwork does. A shop whose weld map, MTRs and NDT reports agree with each other on delivery day has told you everything you need to know about the vessel that walked out with them.

The Discipline in One Page

Pressure vessel engineering is one equation and ten disciplines built around it. The one equation is t = \frac{PR}{SE - 0.6P} — hoop stress, material credit, inspection credit — and everything else is the negotiation of its three variables across the rest of the vessel's life. The ten disciplines: the stress basis that sets S and the 3.5× village of safety it represents; the head menu trading depth against thickness; joint efficiency as the checkbox that costs steel if you skip it; the reinforcement ledger that turns openings into area arithmetic; the cold rules that decide whether the material needs to prove toughness; supports that answer wind and earthquake before pressure; the shop sequence whose order is the inspection regime; the three gates of RT, PWHT and hydrotest, each a yes/no on the schedule; the storage tank world where head, not pressure, writes the courses; and the paperwork — the U-1, the weld map, the charts — which is, in the end, the only part of the vessel that speaks for it when everyone who built it has moved on.

Get those in the right order and a vessel is a quiet, boring, thirty-year asset — filled, emptied, inspected, ignored. Skip one, and the 25 megajoules you put in the first paragraph will find it, in the language of a torn seam. Containment is the oldest promise engineering makes; the code exists because the steel keeps a ledger of every time we broke it.

[1] ASME BPVC Section VIII, Division 1 — U-1 (scope), UG-16 (minimum thicknesses), UG-27 (shells under internal pressure), UG-32 (formed heads), UG-37/UG-40/UG-42 (openings and reinforcement), UG-99 (hydrostatic test), UG-100 (pneumatic), UG-84 (impact testing), UW-11/UW-12/UW-33 (radiography and joint efficiency), UW-50 (examination during hydro), UCS-56 (PWHT), Appendix 1-2/1-3 (thick-wall design). [2] ASME Section II, Part D, Appendix 1 — criteria for establishing allowable stress values; the 1998 code cases and 1999 Addenda change of the tensile factor from 4.0 to 3.5. [3] API Standard 650 (13th ed.) — shell design 5.6 (one-foot and variable-design-point methods; design point 0.3 m above course bottom), Table 5.6.1.1 minimum shell thicknesses, allowable stress basis Sd = min(2/3 Re, 2/5 Rm), St = min(3/4 Re, 3/7 Rm). [4] PESO, SMPV(U) Rules 2016 (G.S.R. 1109(E) and amendments) and the PESO SOP for approval/licensing — licence Forms LS-1A/1B, LS-2 series; workshop approval; type approval per drawing; Rule 13 stage-wise inspection; Rule 18/19 relief valve and periodic testing; Rules 33/43 safety certificates; accepted codes IS 2825/ASME VIII-1,2/PD 5500/EN 13458,13530/AD 2000; National Single Window portal. [5] UCS-56 PWHT thresholds — P-No. 1 carbon steel > 38 mm nominal weld thickness, soak 595–650 °C, 1 h per 25 mm, controlled heating/cooling rates; P-No. 4/5 families triggered at lower thicknesses (industry PWHT compliance summaries, 2025–26). [6] UG-99 stress check — test stress limited to 90 % of yield at test temperature; UG-99(b) 1.3 × MAWP × lowest stress ratio (≥ 1.0). [7] INOX India investor materials FY25/FY26 (revenue ₹1,354 cr FY25; Savli capacity 20,000 t/yr stainless cryogenic equipment; 5 × 690 KL tanks for the Carrington LAES facility; liquid hydrogen at −253 °C) — ICICI Securities coverage and company disclosures. [8] Code history on hydrotest pressure — 1.5 × MAWP replaced by 1.3 × MAWP × LSR in the 1998 revision (ASME discussion records; VIII-1 UG-99). [9] SA-516 Gr. 70 plate mechanical data — Rm 485–620 MPa, Re(min) 260 MPa, elongation 21 % (2 in), ASME SA-20/ASTM A20 conformance — mill and stockholder datasheets (Brown McFarlane, 2026). [10] UCS-66 impact-test exemption curves A–D; SA-516-70 normalized curve D (~ −48 °C at 1 in), SA-106-B curve B (~ −29 °C), UG-20(f) limits (P-No. 1 Gr 1/2 only; ½ in min for curve A, 1 in for B/C/D) — ASME VIII-1 and industry guidance. [11] SA-516 Gr. 70N NACE/HIC plate — ultra-low sulphur, mandatory NACE TM0284 testing, 15–25 % price premium over standard plate (stockholder specification data, 2026). [12] UG-37 area-replacement worked methodology — required area A = d·tr·F; shell/nozzle/weld/pad contributions within UG-40 limits; sol-alpha and industry worked examples (2025–26). [13] SATAT programme data — 217 commissioned CBG plants as of August 2026 vs 5,000-plant target; ₹23,731 crore GOBARdhan (National Circular Bioenergy Scheme) Cabinet approval, 6 August 2026; blending obligation ramp 1 %→5 % and administered price revision (MoPNG/SATAT reporting; Bharat Nama policy brief, Aug 2026; CRA Energy analysis). [14] CBG blending mandate schedule FY26–FY29 and administered pricing (SATAT notifications; IMARC CBG project guide 2026). [15] 250-bar CBG/CNG storage — IS 15663, five-stage compression, 25 % mass density gain over 200 bar, relief at 1.1 × working pressure; IS 7285/IS 15490 cylinder standards (industry technical references, 2026). [16] Indian plate pricing — IndiaMART/trade listings August 2026 (MS plate ₹53–70/kg; SA-516 Gr. 70 boiler quality ₹75–120/kg; fabrication services ₹110–400/kg depending on scope and material). [17] Vessel fabrication quotations and service listings — Indian fabricator price ranges for MS/SS pressure vessels (trade marketplaces and fabrication-services directories, 2026).

More FabFlow blog posts