Look at a large storage tank on a working terminal and the interesting engineering is mostly out of sight. It is in the ring beam under the shell, in the sand pad the bottom plates are lying on, and in a settlement survey somebody has been quietly repeating every six months since commissioning.
This is the part of tankage that catches buyers who have previously only bought pressure vessels. A vessel is an object. You order it, a shop builds it, a truck brings it, you set it on its saddles. A large tank is not an object. It is a construction project that happens to be made of steel, and almost every decision that matters is about what can be done on that patch of ground.
The line between shop built and field erected
Somewhere between a few metres of diameter and a few tens of metres, tankage crosses a line, and everything about how it is produced changes.
Below the line, a tank is fabricated complete in a shop, tested there, and shipped whole. Every weld is made under cover by a welder standing comfortably, with the work rotated to suit. Inspection happens where the equipment lives. This is how we build bullets for liquefied petroleum gas, ammonia vessels and the smaller cylindrical tanks, in diameters up to 8,000 mm and beyond and in lengths past 60 m.
Above the line, transport decides. No road will accept a 40 m diameter tank, so what ships is plate, rolled to radius, marked, and accompanied by a fabrication drawing. The tank is then assembled where it will stand. Shell courses go up one at a time, vertical seams and then horizontal seams, working upward or by jacking. The finished item never existed until it existed on that foundation.
The practical consequence is that the quality of a field erected tank depends heavily on the discipline of site welding, on weather, on access and on supervision. That is why a manufacturer's field erection record deserves at least as much attention as their shop, and why plate marking and fit up control in the shop matter more than they appear to. A shell course that is a few millimetres out of round in the shop becomes a fit up fight on the ring at height.
API 650 and API 620 answer different questions
Two standards govern most of this work and they are not interchangeable.
API 650 covers welded tanks for oil storage at essentially atmospheric pressure. It is the workhorse standard behind the great majority of crude, product, chemical and water tankage in service today. It is well proven, it is economical, and if your service is ambient and near atmospheric it is very likely the right answer.
API 620 covers large welded low pressure storage tanks, and it exists for the cases API 650 cannot reach: modest internal pressures above atmospheric, and low temperature service. Refrigerated ammonia and liquefied gas storage sit here. The design is more demanding, the material requirements are stricter, and the cost is higher, all for good reason.
API 653 is the third of the family and the one owners come to care about most, because it governs inspection, repair, alteration and reconstruction of tanks already in service. It is worth reading before you buy, not after, because a tank designed with 653 in mind is a tank that can be inspected and repaired economically for decades.
For refrigerated and cryogenic tankage there is one further point worth flagging: the minimum design metal temperature drives plate grade selection, and low temperature carbon steels have lead times that do not compress. That decision belongs at the start of the project.
The foundation is not somebody else's problem
A tank and its foundation are a single mechanical system, even when they appear on two different contracts, which they usually do.
Fill a large tank with water for the first time and you apply a load that the ground beneath has never seen. It settles. Uniform settlement is manageable and expected. Differential settlement, where one part of the ring drops relative to another, distorts the shell, strains the bottom to shell joint and can jam a floating roof in its seal. On a floating roof tank, a roof that binds is not an inconvenience. It is a serious operational and safety problem.
Which is why the hydrotest is more than a leak check. Filling the tank with water is the controlled proof load, done deliberately, with settlement measured at points around the circumference as the level rises. It is the moment the foundation demonstrates whether it was right.
One thing to settle early: where the test water comes from and where it goes afterwards. On an inland site with restricted supply, or on a site where discharge is regulated, the hydrotest can become the longest single item in the erection programme. It is a routine question that is routinely asked too late.
The roof decides more than it looks like it does
Roof selection is usually presented as an emissions and product loss question. It is also a maintenance and inspection question for the life of the asset.
A fixed roof, cone or dome, is simple, cheap and appropriate for products that do not evaporate away or create a hazardous vapour space. A floating roof, internal or external, sits on the liquid and largely removes the vapour space, cutting evaporative loss on volatile products and reducing the fire risk that comes with a large volume of vapour above a flammable liquid. It is also a moving mechanical assembly, with seals, legs, drains and guides, all of which need inspection and eventual replacement.
Neither is better in the abstract. The right question is what the product does at the site's summer temperature, and what the operating team is realistically resourced to maintain.
The bottom is where tanks actually die
Shell corrosion is visible and gets attention. Most tanks that are taken out of service are taken out because of the floor.
The underside of the bottom plate sits against the foundation, invisible, often damp, and frequently in contact with material that holds moisture against the steel. Water and sediment collect on the inside. Between the two, the bottom loses metal from both faces while everybody is looking at the shell.
The countermeasures are unglamorous and effective: a properly specified and drained foundation, an appropriate release or barrier layer under the plate, cathodic protection where the soil conditions justify it, a bottom slope and sump arrangement that actually lets water be drawn off, and an internal lining where the product warrants it. Adding a millimetre or two of corrosion allowance to the bottom plate at the design stage is trivially cheap compared with a floor replacement that requires the tank to be emptied, cleaned, gas freed and taken out of service.
Coatings outlive arguments about steel grade
On tankage, protective coatings and linings usually do more for asset life than the choice of steel, and they are treated with a fraction of the attention.
The external system is fighting weather and, on coastal or plant sites, an atmosphere carrying chlorides or process fumes. The internal lining is fighting the product, which may be benign, aggressive, or benign until it carries a little water at the bottom. Where the product is stored hot, the lining has to hold at temperature, and where the tank is cycled, it has to tolerate expansion without disbonding.
Almost every coating failure traces back to surface preparation rather than to the coating. Blast standard, surface profile, cleanliness and the humidity and dew point at the moment of application decide adhesion. A high specification coating applied over a poorly prepared surface will fail faster than a modest coating applied properly, and it will fail by disbonding in sheets rather than by wearing thin, which is the worse outcome because it takes product with it.
Two practical decisions follow. First, decide whether coating happens in the shop or at site, because for a field erected tank a large share of the internal surface has to be done after erection, in a confined space, with ventilation and access to plan. Second, specify how it will be inspected, because dry film thickness readings and holiday detection are the only ways to know a lining is continuous before it is put into service.
Cathodic protection deserves a mention alongside the coating rather than instead of it. The two work together: the coating does the bulk of the work and the protection system handles the defects, which is why an impressed current or sacrificial anode system is sized on the assumption that the coating is imperfect.
What we build and how it arrives
Our tankage work spans API 650 and API 620 designs, fixed cone and dome roof tanks, floating roof tanks, liquefied petroleum gas and liquefied natural gas bullets, ammonia storage and cryogenic vessels, alongside the jacketed and agitated tanks that sit closer to process service than to storage.
For shipped items, dispatch is treated as part of the engineering rather than an afterthought at the gate. That means export grade packaging, seaworthy preservation for equipment that will spend weeks in salt air, international documentation prepared for the destination rather than adapted on arrival, and coordination of heavy lift logistics. Preservation is the item most often skimped on and it is a false economy. Surface damage sustained in transit is repaired at site, in the open, at the worst possible moment in a programme.
Judge the site organisation, not only the shop
If you are choosing between suppliers for field erected tankage, the shop visit tells you less than you think. Every reasonable fabricator has a tidy shop and a wall of certificates.
Ask instead how the site work is resourced. Who supervises the erection crew, and do they work for the manufacturer or a labour subcontractor. How site welders are qualified and how that qualification is verified on the ground rather than on paper. What the plan is for radiography at height, and for weather. How settlement will be monitored during hydrotest, at how many points, and what the acceptance criteria are.
A tank is not really delivered when the last plate is welded. It is delivered when it has been filled, measured, emptied and accepted. Everything before that is preparation.
