NPS 26 Eccentric Butterfly Control Valve for an EDC Plant Expansion: Class 150, a 0.01 bar Drop

A control valve asked to drop one hundredth of a bar across 26 inches of line has, by the usual reckoning, nothing to do. That is the whole engineering problem. At 0.01 bar the valve is not there to burn off pressure — the line does that — it is there to move flow in small, repeatable increments, hold it there for years, and swing shut by itself the moment the instrument air disappears. This page describes the NPS 26 Class 150 pneumatic eccentric butterfly control valve supplied for the EDC plant expansion of PT. Sadang Indoyl Monomer in Indonesia: CF8 body, 304 disc and seat, 17-4PH stem, 24,000 rated Cv, and a fail-close piston actuator.

1. The duty, in six lines

Service: EDC (ethylene dichloride) gas, flow regulation to the reactor feed system — 97 °C, 0.51 barg inlet, 0.01 bar across the valve

Valve: NPS 26 eccentric butterfly control valve, Class 150, RF flanged per ASME B16.5

Body and trim: ASTM A351 CF8 cast body; 304 disc and seat; 17-4PH stem; PTFE packing

Characteristic: approximately equal percentage, 90 degree stroke, rated Cv 24,000 against a calculated 12,075 at maximum flow

Actuation: piston actuator, single-acting spring return, fail close; smart positioner on 4-20 mA, ExdIICT6

Testing: shell per BS EN 1349; seat leakage ANSI/FCI 70-2 Class V

Everything else — the flange drilling, the gland, the air set — follows from these six lines.

2. Why a butterfly, and not a globe at 26 inches

At NPS 26 the case for a globe valve collapses on weight and on length before it collapses on price. A 26-inch globe with a cage trim is a fabrication rather than a casting: a body long enough to turn the flow twice, a bonnet large enough to withdraw the plug, and a face-to-face that the pipe stress engineer then has to design the whole run around.

A butterfly does the same regulation duty inside the pipe's own envelope. The disc sits in the bore, the actuator hangs off a bracket, and the flow path stays straight. At a 0.01 bar differential, the permanent loss a globe trim would add — even a good one — is larger than the drop the whole valve is being asked to produce.

The usual objections to butterflies in control service do not apply here either. A concentric butterfly is a poor throttling valve because the seat rubs through the arc, and any butterfly in flashing or cavitating liquid is a poor idea because the disc edge and the seat take the collapse. Neither is a factor in dry EDC gas at 97 °C with almost no pressure drop. What is left is the one thing a butterfly does better than any other design at this size: a large, smoothly variable flow area inside a short, light, castable body.

3. Where it sits in an EDC plant

EDC sits in the middle of the chlor-alkali to PVC chain. Ethylene is chlorinated — directly or by oxychlorination — to ethylene dichloride; EDC is cracked in a furnace to vinyl chloride monomer; VCM is polymerised to PVC. The valve on this page regulates EDC gas into the reactor feed system, upstream of the cracking furnaces, in a circuit where flow stability reappears downstream as coil temperature and, eventually, as polymer quality.

That position explains the material list far more than the pressure class does. The line is a chlorinated hydrocarbon at 97 °C, carrying the moisture and trace acidity that come with the chemistry. Carbon steel is not on the table: chloride stress corrosion cracking and iron pick-up into a polymerisation-grade product are both disqualifying. Every wetted surface has to be austenitic stainless, and it has to stay that way through excursions, not only at the design point.

It also explains why a full-bore control valve was bought instead of a reduced-bore one with reducers. In a 26-inch line, reducers cost money, add two welds and four flanges, and add a permanent pressure loss the loop can do without. NPS 26 valve, NPS 26 line, nothing in between.

4. What the datasheet fixed

Four decisions were locked in before any metal was cut:

NPS 26, Class 150, RF flanged to ASME B16.5 — line size, line class, line flange. No reducers, no transition spools, no compromise at the tie-in.

Rated Cv 24,000 against a calculated 12,075 at maximum flow — roughly double, and deliberately so. At maximum duty the disc sits at 77.1 % travel rather than at 95 %, which keeps the top of the stroke available for the day the plant debottlenecks or the feed composition shifts.

An approximately equal percentage characteristic — in a circuit where the valve's own share of the total pressure drop falls as flow rises, equal percentage is the characteristic that holds loop gain roughly constant. That is why the normal operating point lands at 72.4 % travel with usable resolution on both sides of it.

Fail close, spring return, single acting — EDC is toxic and flammable. On loss of instrument air the valve has to go shut without waiting for a signal, a solenoid or an operator to notice.

5. Configuration matrix

6. Materials: what touches the EDC

The gas decides the list, item by item:

CF8 body. The cast equivalent of 304, which gives an austenitic pressure boundary in a casting at a size where forging a 26-inch body makes no sense. It also keeps carbon steel out of the chloride path entirely — there is no weak link for the service to find.

304 disc and seat. Not 316, and the reason is worth stating out loud: in dry EDC gas at 97 °C the corrosion mechanism is general rather than pitting, so the molybdenum in 316 buys very little. The money is better spent on the stem and the actuator, where the loads actually sit.

17-4PH stem. A 26-inch disc is a large sail on a small shaft. 17-4PH in the H1025 or H1075 condition carries several times the yield strength of annealed 304, so the stem can be sized for torque and deflection instead of for material, and a smaller stem means less obstruction in the bore and less torsional wind-up between actuator and disc — which is dead band the loop never gets back in tuning.

PTFE packing. Low friction and low hysteresis at 97 °C, which matters more than usual when the loop is trying to resolve a fraction of a percent of travel. PTFE is comfortably inside its range at this temperature; graphite would only add stem friction the positioner has to push through.

ExdIICT6 positioner, epoxy-coated actuator. The unit sits in a classified area, so the positioner enclosure carries the certification instead of a purge system or a local panel, and the RAL 3020 coating is there for the washdown and the marine air, not for the look of it.

7. Where this valve is NOT the answer

Liquid service with a real pressure drop. Flashing and cavitation attack the disc edge and the seat of any butterfly. A cavitating liquid duty belongs to a globe or cage valve with anti-cavitation trim, or to an angle body.

Small-bore control duty. Below about NPS 4 a globe valve is cheaper, tighter and has better rangeability. The eccentric butterfly earns its place by being the only sensible control valve at large bore, not by being a good one at small bore.

Bubble-tight isolation. Class V is a control-valve leakage class, not an isolation claim. A line that needs verified shutoff gets a separate block valve — on a plant like this one, that is what the knife gates and ball valves are for.

Class VI leakage. If Class VI is written into the datasheet, a metal-seated eccentric butterfly at NPS 26 is the wrong starting point. That is a soft-seated high-performance butterfly or a ball valve, and it should be specified that way from the beginning.

Abrasive slurry throttling. The disc edge is a thin section sitting in the flow stream. Solids at velocity will erode it, and eccentricity does nothing about that.

Very high differential. This valve was bought for 0.01 bar. Put the same body across 15 bar and the actuator, the stem and the seat are all the wrong size — a butterfly's usable differential falls away as the bore grows.

8. Verification and document package

Each valve was shell tested to BS EN 1349 at 1.5 times the Class 150 ambient rating, and seat tested to ANSI/FCI 70-2 Class V with the actuator mounted. The leakage figure that means anything is the one measured with the packing, the lever and the linkage in place — not the number from a bare body on a bench. Travel was verified over the full 90 degrees against a 4-20 mA input in both directions, with hysteresis recorded rather than assumed.

The fail-safe test is the one the plant actually cares about: air supply removed with the valve under differential, and the spring-driven stroke timed all the way to the seat. The handwheel was then checked for manual override without disturbing the automatic action — a handwheel that drags the linkage when it is not engaged is a fail-safe claim waiting to be withdrawn.

Documents ship per valve, not per batch: material certificates for body, disc and stem; the shell and seat test record for that serial number; the positioner calibration sheet; and the paint inspection for the RAL 3020 epoxy on the actuator.

9. FAQ

Why not a globe valve at NPS 26?

A 26-inch globe is a special fabrication — heavy, long, expensive — and it needs a trim large enough to pass 70,800 kg/h without adding loss the loop cannot afford. The butterfly's known weaknesses, seat wear in throttling and poor behaviour in cavitating liquid, are respectively solved by the eccentric geometry and irrelevant in dry gas at 0.01 bar. A globe wins at high differential and at small bore. Neither applies here.

What does a 0.01 bar differential actually mean for control?

That the valve has very little authority over the loop. Almost all of the pressure is spent elsewhere, so flow is set mainly by the rest of the circuit and the valve's job is to trim around that point without adding instability. Two consequences follow. The characteristic has to be right — equal percentage here, so that gain stays flat as flow moves. And the mechanical dead band has to be small: a stiff 17-4PH stem, PTFE packing, a positioner with fine resolution. A soft stem and rubber packing at this differential will hunt, and no amount of tuning will fix it.

Why 17-4PH on a stainless valve — is that not a mismatch?

It is a deliberate one. 17-4PH is a precipitation-hardening stainless: it keeps corrosion resistance close to the 300 series while carrying several times the yield strength. On a 26-inch disc the stem sees bending from the flow and torque from the seat, and the limit on how small it can be is deflection, not corrosion. A 17-4PH stem stays straight and stays slim, which is also what keeps the bore clear.

Can a metal-seated butterfly really hold Class V at NPS 26?

Class V is a demanding call at this size, and it is met by geometry and finish rather than by soft goods: an eccentric disc that cams clear of the seat, a lapped 304 seat, and a preload that survives the stroke. It is verified valve by valve on the test stand, which is exactly why the test record is part of the document package. If a plant needs bubble-tight rather than Class V, that is a different valve and it should be specified as one instead of negotiated at the bench.

10. Send us these eight lines and we will quote

Fluid, phase and chemistry — this, not the pressure class, is what decides the trim and body material

Normal and maximum flow, and the turndown you actually need

Inlet and outlet pressure, and the drop you intend the valve to take

Operating and design temperature

Line size and pressure class, and whether a reduced-bore valve with reducers is acceptable

Fail-safe position on air or power loss: close, open, or hold last position

Leakage class, and the standard the plant tests to

Hazardous area classification and control signal (4-20 mA, fieldbus, or discrete)

Valve technical data on this page: customer valve datasheet and selection calculation for the EDC plant expansion project — NPS 26 Class 150 eccentric butterfly control valve, model TD643H-150Lb(P)-26". Dimensions and materials as specified at order stage; confirm against the approved drawing for any repeat order.