Isolation is not throttling

Valve datasheets are reviewed for size, class, material, end connection and operator. Those fields are necessary and they are not sufficient, because none of them records what the valve is for. Four failures follow from that gap with enough regularity to be worth naming.

1. Throttling with a gate valve

A gate valve works by lifting a wedge entirely clear of the bore. Fully open, it presents almost no obstruction and almost no pressure drop, which is exactly why it is the default isolation valve.

Held part-open, the same geometry works against it. Flow is forced through a narrow annulus between the wedge and the seat at high local velocity, and the seating surfaces erode — wire-drawing. The damage accumulates on precisely the surfaces the valve depends on to shut off. The failure is therefore delayed and quiet: the valve keeps working as a throttle for a while, and then fails to do the one job it was bought for.

Operations teams throttle gate valves because a hand wheel turns and the flow changes. Preventing it is a design and labelling problem, not a discipline problem — if a line needs regulation, put a valve there that regulates.

2. Treating a globe valve's pressure drop as a defect

A globe valve moves a disc linearly against a stationary seat, and the flow has to turn up through that seat and back down. The path is deliberately tortuous. That resistance is the feature: it is what gives a controllable relationship between stem position and flow.

Substituting a lower-loss valve to "save pressure drop" on a line that needs regulation removes the control along with the loss. If the pressure budget genuinely cannot carry a globe valve, that is a hydraulic problem to solve in the line sizing, not a valve type to swap.

3. Check valve slam

A swing check valve has no operator. Line pressure holds the disc open, and reverse flow closes it. The failure mode is what happens between those two states.

When a pump trips, flow decelerates, reverses, and accelerates backwards. If the disc is still travelling when significant reverse velocity has developed, it arrives at the seat against a moving column of liquid. The result is a pressure transient that can be violent enough to damage the valve, the supports, and the pump it was installed to protect.

Slam is a function of how fast the system decelerates and how fast the disc closes — not of the valve's size or class. A correctly rated check valve, correctly installed, will still slam if the system's deceleration outruns it. Where the deceleration is severe, the selection has to change: a nozzle or axial check with a spring-assisted, short-travel disc closes before reverse flow establishes. That is a system analysis, and it belongs at design stage, not after the first trip.

4. Butterfly valves outside their envelope

A butterfly valve's disc stays in the flow at all times, so there is always some loss. In exchange, face-to-face is short and weight is a fraction of a gate valve of the same size — which is why above roughly 8 in on water and low-pressure service it is usually the right answer, and often the only affordable one.

The envelope matters. High differential pressure across a partly open disc produces significant dynamic torque, and an actuator sized on seating torque alone will stall part-way. Cavitation across the disc on high-drop service will erode both disc and seat. Neither shows up if the datasheet records only size, class and material.

What to put on the datasheet

The common thread is that all four failures come from specifying construction without specifying function. A datasheet that prevents them records, alongside the usual fields:

• Function — isolation, regulation, non-return, or on/off automated shutdown. Explicitly, as a field.
• Operating and design differential pressure, not just the class rating.
• Expected duty cycle — how often the valve is operated, and whether it sits in one position for years.
• For check valves, the system deceleration case: what happens on pump trip.
• For automated valves, the required stroke time and fail position, from which dynamic torque and actuator sizing follow.

None of that is unusual information. It is generally known to the process engineer and generally absent from the document the valve is actually bought against — and closing that gap is far cheaper than discovering it during commissioning.