Gas-Hydraulic Actuated Fully Welded Ball Valve, 2-60 Inch, Class 150-2500

A gas-hydraulic actuated fully welded ball valve is a trunnion mounted ball valve whose body is closed by circumferential welds instead of flanges, and whose operator runs on the gas it is isolating. Pipeline gas at 0.5 to 10 MPa is admitted to a small cylinder, pushed against a piston, and that piston displaces hydraulic oil into a second, larger cylinder that turns the valve. The oil is the transmission, the gas is the fuel, and no compressor, instrument air header or electrical supply is needed at the site. The valve is used where the line is buried, the station is unmanned, or the distance to the nearest power is measured in tens of kilometres.

J-VALVES builds these valves from 2 inch to 60 inch in ASME Class 150 to 2500 at its works in Wenzhou, Zhejiang, with API 6D design, API 6DSS on request for subsea and ASME B16.34 wall thickness. The body is welded from forged pipe pups and body forgings, radiographed in full, and coated for direct burial. This page sets out the construction, the gas-over-oil power chain, the control logic that actually protects a pipeline, and the points that a datasheet usually leaves out.

What the Two Words in the Name Decide

The name carries two independent design decisions and buyers often conflate them. "Fully welded" is a decision about the pressure boundary. "Gas-hydraulic" is a decision about the operator. Either can be changed without touching the other, and a quotation that mixes them up is the usual source of a valve that arrives unable to be buried, or unable to be operated.

The combination is at its best on a gas transmission line of 12 inch and above, buried, with a block valve station every 20 to 30 km. It is at its worst on a short plant line where the gas is not always present, and where a maintenance team would rather unbolt a bonnet than cut a pipe.

Where the Welds Are, and Why It Matters

A fully welded ball valve is not a welded valve in the sense of a fabrication. The body is built from a centre section and two end closures, all forged, joined by two circumferential girth welds. On larger sizes the centre section itself is a forged ring with the ball cavity machined in. There is no longitudinal seam in the pressure-containing shell, and that is the point: a longitudinal seam runs along the highest stress direction, while a girth weld sits in a section that is in membrane tension only.

The welds are made by submerged arc welding for the root and fill on carbon and low alloy bodies, and by gas tungsten arc welding for the root on stainless and duplex. Both processes are mechanised, because a hand-held pass at 60 inch is not reproducible enough for a joint that cannot be inspected from the inside after closure. The weld procedure is qualified to ASME IX, and where the client asks for pipeline practice, to API 1104 or ISO 13847.

Inspection follows the same logic as the design. A welded body has no second barrier, so the weld is examined rather than sampled. J-VALVES performs 100 percent radiographic testing on both girth welds, with ultrasonic testing as a cross-check on thicknesses above about 25 mm where radiography loses sensitivity to planar defects. Dye penetrant is used on the machined seat pockets before assembly, because a crack there is invisible once the seat ring is in.

Three details decide whether the finished weld performs in the ground. The first is fit-up: a mismatch of more than 1 mm at the root concentrates strain and shows up as a toe crack after a few pressure cycles. The second is interpass temperature on stainless, which must stay below 150 degrees Celsius to avoid sensitisation of the heat affected zone. The third is the bore transition. A weld that leaves a internal bead in the flow path becomes a pigging obstruction and a corrosion trap, so the bore is ground flush and the radius blended.

The Gas-Over-Oil Power Chain, Step by Step

A gas-over-oil actuator is a two-stage hydraulic intensifier that is driven by process gas. Understanding the chain explains most of the field problems, because every stage has a failure mode that looks like a valve fault and is not.

1. Take-off. Gas is drawn from the upstream side of the valve through an isolation valve, a filter and a regulator. The regulator drops line pressure to the working pressure of the actuator, typically 4 to 10 MPa. A check valve prevents backflow into the line.

2. Gas piston. The reduced-pressure gas enters the gas cylinder and pushes a piston. The piston area is large, so the force is high even at modest pressure.

3. Oil displacement. The piston drives hydraulic oil out of the oil reservoir through a manifold. The oil side has a much smaller area ratio, so pressure is multiplied by the area ratio, typically 8:1 to 20:1. This is where the torque comes from.

4. Hydraulic cylinder. The pressurised oil acts on the actuator cylinder that drives the valve stem through a scotch yoke or a rack and pinion. The scotch yoke gives high torque at the ends of travel, which is exactly where a ball valve needs it: seating and unseating.

5. Speed control. Throttle valves in the oil return line set the closing and opening speed independently. Restricting the return is what limits speed; restricting the supply only reduces force.

6. Return. On a double-acting actuator the gas is vented and the oil returns to the reservoir, driven by the second gas stroke. On a spring-return actuator the spring stores the closing energy and the gas only opens.

7. Manual override. A hand pump injects oil directly into the cylinder, bypassing the gas circuit, so the valve can be cycled with no gas available at all.

The area ratio is the number that decides whether the actuator works at the bottom of the pressure range. A valve that needs 8,000 Nm break-to-open, driven through a scotch yoke with a 0.35 m effective arm, needs about 23 kN of thrust at the yoke. If the hydraulic cylinder is 120 mm bore, that is 4.5 MPa of oil pressure, and at an 8:1 ratio it is 0.55 MPa of gas. That is why a gas-hydraulic actuator can still operate at 0.5 MPa line pressure, and why the same actuator on a 60 inch valve at the same pressure cannot.

The Take-Off Point Is the Part Nobody Specifies

Most gas-hydraulic actuator failures in the first two years are gas quality failures, not actuator failures. The valve is fine; the instrument gas is wet, dirty, or has dropped a slug of condensate into the gas cylinder.

Four properties matter. Water dewpoint must be below the lowest ambient temperature the station will see, or hydrate and ice form at the regulator where the gas expands and cools. Hydrogen sulphide attacks the regulator body and the seals, and on sour service the gas circuit needs NACE compliant materials to ISO 15156 with hardness held below 22 HRC. Condensate carries liquid hydrocarbons that swell elastomer seals and clog the filter. Particulate matter scores the regulator seat and the piston bore, and a 40 micron filter is the practical minimum, with 10 micron preferred on unmanned stations.

Where the gas is not fit for the actuator, there are three honest answers. Take the gas from a cleaner point on the line, install a conditioning skid, or specify a different actuator. What does not work is adding a filter to a wet gas line and hoping: the water stays in the gas phase until it does not.

Closing Speed, Surge and Water Hammer

A gas-hydraulic actuator can be set to close in anything from 15 to 120 seconds, and on large valves longer. That range is not a convenience feature; it is the difference between a controlled shutdown and a pressure spike that splits a line.

The physics is the same as in liquid service, and it is often dismissed on gas lines on the grounds that gas is compressible. Gas is compressible, but a fast-closing valve on a long line still converts momentum into a pressure wave, and the wave travels at the speed of sound in the gas, several hundred metres per second. The Joukowsky relation gives the head rise as the wave speed times the change in velocity divided by gravitational acceleration; on a 500 mm line at 6 m/s the rise can be several bar in the first second. Because the wave reflects off the closed valve and off the upstream station, the peak usually arrives after the valve has already stopped.

Two settings matter. The closing time should be long enough that the deceleration is spread over at least two wave travel times, which on a 30 km block section means a closing time measured in tens of seconds rather than a few. And the last 15 percent of travel should be slower than the first 85 percent, because that is where the flow is throttled and the velocity change is concentrated. A scotch yoke actuator with a flow control on the exhaust, plus a separate cushion at the end of travel, gives that profile without electronics.

Line Break Control Is Not Under-Pressure Protection

Buyers routinely write "low pressure shut-off" on a datasheet when they mean line break control. The two protect against different events, and a valve set up for one will usually fail to protect against the other.

Under-pressure protection trips when line pressure falls below a fixed setpoint for longer than a set delay. It catches a slow loss of containment, a downstream blowdown, or a compressor trip. Its weakness is that it cannot tell a genuine rupture from a normal operating transient, so the setpoint has to sit low enough to avoid nuisance trips, which means it responds only after the pressure has already collapsed.

Rate-of-pressure-drop control compares the fall in pressure over a short window against a threshold in bar per minute. A rupture produces a fall far steeper than any operating transient, so the trip can be set tight without nuisance trips. Most modern line break control logic combines the two: rate of drop for the fast rupture, under-pressure with a time delay for the slow leak, and both feeding an ESD signal.

Whichever logic is used, it has to be built into the actuator control panel and proven at commissioning by an actual trip test, not by a signal injection into the PLC. A valve that closes in 90 seconds on a command but takes 4 minutes on a real line break because the gas supply is being starved by the same pressure collapse is a common and dangerous finding.

Torque, and Why the Margin Is Not Optional

Torque is the number that sizes everything downstream of the ball. It has four components and only one of them is published in a catalogue.

Break-to-open torque is what is needed to unseat the ball after a period closed, and on a soft seated valve it rises with the time closed as the seat takes a compression set. Running torque is the friction of the ball on the seats and of the stem in the bearings, and it peaks at part-open, not at either end. Seating torque is what is needed to push the ball into the seat at the end of travel, and it is applied last. Bearing and packing friction add a constant that grows with pressure and with time in service.

The margin exists because all four grow in service. Seat swelling in hydrocarbon service, grease aging, slight body distortion after burial, and differential temperature across a 60 inch body all push torque up. The usual allowance is 1.5 on the calculated figure for a clean valve, and 2.0 where the duty is dirty, where the valve will sit closed for long periods, or where the actuator cannot be upsized later. For a 12 inch Class 600 trunnion valve in clean gas, break-to-open falls in the region of 5,000 to 9,000 Nm at full differential, and a metal seated valve at the same size can be two to three times that. A figure an order of magnitude above these is worth questioning before it is written into a purchase order, because it will size a cylinder that cannot be shipped on a standard skid.

Body, Ball, Seat and Stem at Large Size

Body material follows the fluid and the temperature more than the pressure. A105 forged carbon steel is the default for dry natural gas up to about 425 degrees Celsius and is the cheapest route to a 60 inch body. A350 LF2 takes the same duty down to about minus 46 degrees Celsius with Charpy records. F304 and F316 are used where the gas is wet, where CO2 partial pressure is high, or where the client's line pipe is stainless and a dissimilar weld is unwelcome. F51 duplex is chosen for chloride bearing produced water and for high strength at moderate temperature.

Seat design matters more at 60 inch than at 6 inch, and the choice between single and double piston effect is a decision about the cavity. A single piston effect seat relieves trapped cavity pressure into the lower pressure side of the line automatically. That is what you want on a liquid line, where thermal expansion of trapped liquid will otherwise over-pressure the cavity. A double piston effect seat seals in both directions and does not self-relieve, so the cavity needs an external relief valve. On a buried gas valve, DPE is normally preferred because it gives a double barrier: if the upstream seat is damaged, the downstream seat still holds. The relief is then a small spring loaded valve piped to a vent, and it has to be accessible above ground, which is a civil decision as much as a valve decision.

The stem is independent of the ball on a trunnion design and carries no side load from line pressure, which is why a 60 inch valve can still be turned by a modest actuator. Two O-rings and a graphite gasket give a triple barrier, the stem is anti-blowout to API 6D, and a sealant injection port allows the upper stem seal to be refreshed without depressurising the line.

Burying It: Coating, Cathodic Protection and Access

A fully welded valve is usually bought to be buried, and burial changes the specification in ways that are easy to miss because they are not valve standards.

External coating is normally a three layer polyolefin or fusion bonded epoxy system, applied after the weld inspection and holiday tested at 5 to 10 kV. The coating must be compatible with the line coating so that the field joint can be completed with the same material and the same testing. On a buried stainless body, coating plus cathodic protection is a pair that can go wrong: a coating holiday on stainless under cathodic protection concentrates current and can cause hydrogen embrittlement or severe local attack, so stainless buried valves are usually isolated with a monolithic joint and coated fully rather than protected.

The extended stem is the interface between the buried valve and the surface. It carries the stem seals up into the gland plate above ground, so that a seal failure is visible and repairable. The extension is protected by a sleeve, and the sleeve must be sealed at the bottom, or groundwater rises inside it and the seals sit in water. The operator, whether it is a handwheel or a gas-hydraulic actuator, sits on a concrete pad above the flood level, and the pad has to be sized for the actuator's manual override pump and for a technician to stand at it.

Two more items belong on the drawing. A vent and drain on the body cavity, extended to the surface, is required for DBB testing and for drying after the hydrostatic test. And the valve should be installed with the stem vertical or within 30 degrees of vertical: a horizontal stem on a buried valve puts the extension in bending and makes the packing drain badly.

Testing: Shell, Seat, Function and the Loop

Four tests matter and they are done in a fixed order, because each one assumes the previous has passed.

1. Shell test at 1.5 times the 38 degree Celsius rating, with water treated for chlorides on stainless bodies, held for the duration API 6D requires, with no leakage and no permanent deformation.

2. Seat test at 1.1 times the rating, on each seat separately, with the leak rate measured against ISO 5208. On a DPE seat, both seats are tested from both directions, which doubles the test time and is the reason some suppliers quietly test only one.

3. Functional test of the actuator: three full cycles at the specified gas supply pressure, with closing and opening times recorded, the manual override exercised, and the ESD trip proven.

4. Loop test after installation, where the valve is tripped from the actual control system with the actual gas supply and the actual back pressure. This is the only test that finds a starved gas supply or a mis-set throttle.

Documentation is issued by serial number and normally includes the EN 10204 3.1 material certificates with impact results, the weld procedure and welder qualifications, the radiography reports with film or digital images, the shell and seat test records, the torque trace, the actuator functional record, the coating holiday test, and the dimensional report. On a welded valve, the radiography report is not a formality: it is the only record of the pressure boundary, and a buyer who does not receive it has bought an unverified vessel.

Writing the Datasheet Line

Most quotation errors on this product come from a datasheet that omits one of six fields. A complete line reads: size and bore, ASME class, design temperature range, body and trim material, end connection, actuator type with gas supply conditions, and control functions required.

A worked example: 20 inch full bore, ASME Class 600, minus 29 to 80 degrees Celsius, A105 body with F316 ball and RPTFE seat, butt weld ends to ASME B16.25 with 500 mm pups, gas-hydraulic double acting actuator with scotch yoke, gas supply 4 to 8 MPa filtered to 40 micron, closing time 45 seconds adjustable, line break control by rate of pressure drop plus under-pressure trip, manual hydraulic override, SCADA interface 4 to 20 mA and Modbus RTU, external coating three layer PE, extended stem 1,200 mm, DPE seats with external cavity relief, API 6D with 100 percent RT.

Two fields are almost always missing and both change the price materially. The first is the stem extension length, which is set by the burial depth plus the pad height and cannot be guessed. The second is whether the actuator gas is taken upstream or downstream of the valve, because that decides whether the valve can reopen after a trip on a depressurised section.

Frequently Asked Questions

What is the difference between a gas-hydraulic and a gas-over-oil actuator?

They describe the same machine. Gas-over-oil emphasises that line gas pressurises oil directly; gas-hydraulic emphasises that the output is hydraulic. Both use line gas to drive a piston that displaces oil into a cylinder, multiplying force by the area ratio. Some suppliers use gas-over-oil for the simple non-intensified version where gas acts on oil in a single vessel, and gas-hydraulic for the two-cylinder intensified version. The distinction is not standardised, so the datasheet should state the area ratio and the resulting torque rather than the name.

Can a fully welded ball valve be repaired if a seat leaks?

Partially. Emergency sealant injection ports let a technician restore sealing temporarily by injecting grease into the seat pocket, and this is standard on 6 inch and larger. The upper stem seal can be replaced with the valve in line and the cavity vented. What cannot be done without cutting the line is replacing a seat ring or the ball. That is why the seat material and the coating choice are worth getting right at the order stage: on a buried valve there is no second chance.

How long does the actuator take to close, and who sets it?

The closing time is set by throttle valves on the hydraulic return and is adjustable on site between roughly 15 and 120 seconds, longer on very large valves. It should be set at commissioning against the surge calculation for the specific line, not left at the factory setting. The last part of travel should be slower than the first so that the velocity change is spread out. A closing time that is too short is the most common cause of a pressure spike after a trip.

What happens if the pipeline pressure is too low to operate the actuator?

The manual hydraulic hand pump is the answer, and it is why the pump must be reachable from the pad rather than buried under the actuator frame. Below about 0.5 MPa the gas circuit cannot develop enough force on most designs, and the area ratio has to be increased at the order stage if low pressure operation is expected. A valve specified for 4 MPa minimum gas and installed on a line that runs at 1.5 MPa will not operate, and no amount of site adjustment will fix it.

Is a fully welded valve suitable for hydrogen or CO2 service?

The body is suitable; the actuator usually is not. A gas-hydraulic actuator needs gas to burn through its cylinders, and on a pure hydrogen or dense phase CO2 line that gas is either too valuable to vent or, in the case of CO2, the pressure and phase behaviour make a gas take-off impractical. On those duties the welded body is still the right pressure boundary, but the operator is normally electric or electro-hydraulic with an external supply.

What certification should come with the valve?

As a minimum: EN 10204 3.1 material certificates with Charpy results for the body and closure forgings, weld procedure and welder qualification records to ASME IX or API 1104, 100 percent radiography reports, shell and seat test records to API 6D and API 598, torque and functional test records for the actuator, coating holiday test, and a dimensional report. API 6D monogram, API 607 fire test, ISO 15848-1 fugitive emission and NACE MR0175 are available on request and should be stated in the enquiry, because adding them after the body is welded is not possible.

For a quotation, send the line class, the burial depth, the gas analysis and the control philosophy. Drawings are released for approval before the girth welds are made, and any non-standard stem extension or actuator mounting is confirmed on the drawing rather than on the order acknowledgement.