A ball valve earns its keep by doing one job extremely well. It gives a clear, round bore when open and a tight shutoff when closed, and it moves between those two states in a quarter turn. That combination is why the design has pushed gate valves out of so much isolation duty over the last forty years, and why a plant walkdown today finds ball valves on utility headers, hydrocarbon lines, sample points, drains and pig launchers that would once have carried a wedge gate.
The sales argument is usually about speed and tightness. Both are real, but the reasons a ball valve outperforms other designs come down to how the closure member is loaded and how the seats are fed by line pressure.
What the quarter turn really buys you
In a soft seated ball valve the seats are rings of polymer held in pockets on either side of the ball. Line pressure acts on the back of the upstream seat and pushes it against the sphere. The harder the line pushes, the harder the seat seals. That is a self energising arrangement, and it is fundamentally different from a wedge gate, where sealing depends on a mechanical wedging action and on two machined metal faces staying flat and clean.
The practical result is that a soft seated ball valve closes bubble tight at pressures where a gate valve of the same nominal rating is weeping past the seat faces. Shell and closure testing to API 598 allows a measured leakage rate for metal seated valves scaled to size, but requires no visible leakage from resilient seated valves. A ball valve routinely passes that with margin, and it keeps passing it after thousands of cycles because the sphere wipes the seat clean every time it rotates.
The full bore version also gives you something a globe or butterfly valve cannot. With the bore matched to the pipe internal diameter, head loss across the open valve is close to that of an equivalent length of pipe, and a scraper pig or an inspection tool passes straight through. On a transfer line where pumping cost over a twenty year life dwarfs the valve price, that matters more than the purchase order ever shows.
Cycle time is the third advantage. Ninety degrees of rotation takes under a second on a small valve with a spring return actuator, and for emergency shutdown duty no multi turn design competes.
Floating ball against trunnion mounted
Two mechanical arrangements dominate, and choosing the wrong one is the commonest specification error I see.
In a floating ball valve the sphere is held only by the two seats and the stem drive. Line pressure pushes the whole ball downstream onto the downstream seat, which does the sealing. It is simple, it is cheap and it seals very well. The penalty is operating torque, which rises with the square of the bore and linearly with differential pressure, because the full unbalanced thrust on the ball is reacted as friction at the seat face. Above roughly DN 150 in Class 300, or at any size in Class 600 and above, the torque becomes awkward and the seat contact stress starts to cold flow the polymer.
A trunnion mounted ball is supported on upper and lower bearings so the sphere does not move. Both seats are spring loaded and pressure energised against a sphere that stays put, and the bearings carry the thrust. Torque is far lower and far more predictable, and the double block arrangement allows a valve to be tested from the body cavity. Trunnion designs also allow sealant injection fittings into the seat and stem area, which on a pipeline valve is the difference between a repair and a shutdown.
The rule of thumb that works is simple. Small and moderate pressure, use floating. Large bore or high class, use trunnion. Pipeline service to API 6D, almost always trunnion.
Seats and seals decide the temperature limit
The ball and body can be a 9 per cent chrome casting good for 550 degrees Celsius, and the valve will still be limited by a polymer ring. Virgin PTFE is the default seat material and is comfortable to about 200 degrees Celsius, but it creeps under sustained load and it has a high coefficient of thermal expansion. Glass filled or carbon filled PTFE raises the load capacity at some cost in sealing finesse. PEEK extends the range further and handles steam and higher contact stress, but needs more actuator torque and will not seal against a scratched ball.
Above those limits you move to a metal seated ball valve with a hard overlay, typically chrome carbide or tungsten carbide applied by high velocity oxy fuel spraying, lapped as a matched pair. These are excellent in abrasive and high temperature service and are common on catalyst, coke and slurry lines. They are not bubble tight in the way a PTFE seat is, they cost several times more, and they require far higher actuator torque. Specify them because the service demands it, not as a general upgrade.
One failure mode deserves naming because it catches people out. A soft seated ball valve closed on a liquid traps that liquid in the body cavity. Heat the valve, whether by process, by sunlight or by a fire, and the trapped liquid expands against a closed volume. Body cavity pressures far above the rating are easily reached. A valve in that service needs a cavity relief feature, usually a vent hole through the upstream side of the ball, and that choice makes the valve unidirectional. Mark the flow direction and respect it.
Where a ball valve is the wrong answer
Throttling is the obvious one. A partly open ball exposes the seat edge to a high velocity jet and the seat is destroyed in weeks. Characterised V port balls exist and work, but they are control valves with a ball trim, not standard isolation valves pressed into control duty. Using a standard ball valve to trim a flow is a mistake, and it shows up as a leaking isolation point when you next need the line dead.
Slurries and solids laden fluids attack the floating design because particles collect in the body cavity behind the ball and in the seat pockets, then grind the seat face on every stroke. Metal seated or a different valve type entirely is the honest answer.
Fire exposure is survivable but has to be designed in. A valve that has to hold after its polymer seats have burned away needs a secondary metal to metal seat and a graphite stem packing arrangement, tested to API 607 or API 6FA. Do not assume a catalogue ball valve has it.
A ball valve left in one position for years in a dirty or polymerising service will seize, and the operator who discovers that is usually the one trying to isolate a leak. Exercise them.
What to pin down on the datasheet
- Bore type, full or reduced, and the actual bore diameter in millimetres rather than the words alone.
- Floating or trunnion, and the calculated breakaway torque at maximum differential pressure.
- Seat material with its temperature limit, and whether the service can see steam out or a thermal excursion.
- Cavity relief and the resulting flow direction, if any.
- Stem design. Specify a blowout proof stem, which is standard in most modern ball valve designs but not in every cheap import.
- Fugitive emission requirement, if the fluid is a volatile organic compound, with reference to ISO 15848 or API 624.
- Antistatic continuity between ball, stem and body for any hydrocarbon or gas duty.
Get those seven right and the valve will usually outlive the actuator fitted on top of it. Get the seat material and the bore wrong and you will be replacing it inside two years.