Industrial valve maintenance is the scheduled inspection, cycling, packing adjustment, and repair of isolation valves so they seal tight, open on demand, and close fully when a line has to come down. It targets the three things that actually fail: stem packing, seat sealing, and the actuation that moves the valve.
Most plants have thousands of manual and automated isolation valves, and most of them get zero attention until one of two things happens: a valve weeps around the stem and starts dripping product on the floor, or a valve that should isolate a line for a repair will not seal, so the job that needed one valve shut now needs half the header drained. This guide covers what to inspect, how often to cycle valves, how to read packing and seat leakage, and how to build a valve program that keeps isolation valves ready for the day you need them.
What does industrial valve maintenance actually cover?
It covers keeping three subsystems working: the stem seal that keeps process fluid from escaping to atmosphere, the seat that stops flow when the valve is closed, and the actuation, handwheel, gearbox, or powered actuator, that makes the valve move. Almost every valve failure a maintenance team sees traces back to one of those three.
A valve is a simple machine with a hard job: it has to move against a flowing fluid, seal against pressure, and stay leak-tight while it sits idle for months. The parts that carry that load are few, and knowing them by name is the whole starting point for inspection.
How is this different from control valve maintenance?
Isolation valves are built to be fully open or fully closed and to seal; control valves are built to sit at 40% and modulate flow all day. That difference changes what wears and what you inspect. An isolation valve fails from sitting still and then being asked to move and seal. A control valve fails from constant travel, trim erosion, positioner drift, and cavitation.
If your problem valve is throttling a process variable against a controller signal, you are looking at a control valve, and it needs positioner calibration, trim inspection, and stroke testing that isolation valves do not. That regimen is covered separately in control valve maintenance. This guide is about the gate, globe, ball, butterfly, and plug valves whose job is simply to isolate, the ones your lockout/tagout program depends on to make a line safe.
What should a valve inspection check?
A valve inspection checks the three failure zones plus the body itself for corrosion and external leakage. You do not need instruments for most of it, a walk-down with a flashlight, a rag, and a torque wrench catches the majority of developing problems.
Work through each valve in this order:
| Zone | What to look for | What it means |
|---|---|---|
| Stem packing / gland | Weeping, crusting, or spray around the stem; gland follower bottomed out | Packing is worn or was over-compressed; adjust or repack before it becomes a fugitive-emission source |
| Seat sealing | Passing flow, temperature, or noise downstream of a closed valve | Seat is eroded, scored, or fouled; the valve no longer isolates |
| Actuation | Hard operation, backlash in the handwheel, actuator not reaching end stops | Stem thread galling, gearbox wear, or actuator/air-supply fault |
| Body & bonnet | External corrosion, bonnet-joint leakage, coating failure | Wall loss or gasket failure; see corrosion-driven wall thinning |
| Position & access | Missing handwheel, buried valve, no position indication | The valve exists on paper but cannot be operated when needed |
That last row matters more than people expect. A valve you cannot reach or cannot tell the position of is a valve you cannot trust in an emergency. External corrosion of industrial equipment is the slow killer here, it seizes stems and thins bodies long before anyone schedules a repair.
Why does packing leak, and what do you do about it?
Packing leaks because the braided or die-formed rings around the stem lose their seal, from age, from the wrong compression torque, from thermal cycling, or from a scored stem cutting the rings every stroke. Packing leakage is the single most common in-service valve problem, and left alone it wastes product, drives up fugitive emissions, and eventually cuts the stem.
The first move is almost never a full repack. It is a measured re-torque of the gland follower: snug the packing enough to stop the weep without clamping the stem so hard the valve will not move. Over-tightening the gland is itself a leading cause of hard operation and stem wear, so it is a balance, not a wrench-it-down job. If a re-torque does not hold, the packing is spent and the valve gets repacked, ideally with the line depressurized, though live-loading and injectable packing exist for valves that cannot be taken out of service.
For valves in hydrocarbon or hazardous service, packing is not just a housekeeping issue, it is a regulated emission point. Modern low-emission valves are type-tested to API 624 (rising-stem) or API 641 (quarter-turn) with an acceptance limit of measured stem leakage below 100 ppmv, which is the number to design toward when you specify replacement packing.
How do you know a closed valve is still sealing?
You know by testing it, because a valve that closes is not the same as a valve that seals. Seat leakage, flow passing through a fully closed valve, is invisible from the outside and is the failure that ruins isolation for a repair job. The standard way to grade it is a seat test to API 598 or ISO 5208, which define allowable leakage rates by valve class, from bubble-tight (Class VI-style, no visible leakage) down to rates that pass a set number of milliliters or bubbles per minute.
In the field, you rarely run a formal API 598 bench test on an installed valve. Instead you look for the practical symptoms of a passing seat: a downstream line that stays warm after isolation, audible flow through a closed valve, a bleed that will not depressurize, or a pressure that creeps back up on the isolated side. Any of those means the seat is compromised and a single valve is no longer a safe isolation point, which is why critical isolations use two valves and a bleed (double block and bleed) rather than trusting one seat.
How often should isolation valves be cycled and inspected?
Cycle critical isolation valves at least once or twice a year so they do not seize in place, and set inspection intervals by how much the valve matters and how nasty its service is. A valve that never moves is the one that will not move when a line is on fire. Exercising valves on a schedule is the cheapest insurance in the whole program.
Use a simple risk-based framework rather than one blanket interval for the plant:
- Rank valves by consequence. Which valves does a safe shutdown, a fire response, or a critical isolation depend on? Those are your program. Pull them from your asset criticality ranking matrix the same ranking that drives the rest of your reliability work.
- Set an exercise interval. Critical isolation and emergency valves get partial or full stroke on a fixed cycle (commonly quarterly to annually) so they stay free. Log the effort it took to move them, rising torque is an early warning.
- Set an inspection interval by service severity. Clean, mild service can go a year or more between walk-downs; abrasive, corrosive, or high-cycle service needs more frequent packing and seat checks.
- Trigger condition-based checks. Tie valve inspection into your condition-based maintenance triggers, a passing seat found on a process upset, or packing weeping spotted on an operator round, jumps the queue.
- Repack and reseat on findings, not on the calendar. Do not blanket-repack good valves. Let inspection and testing drive the intrusive work, and feed every finding back into the interval.
Feeding those findings into a schedule is exactly what a preventive maintenance schedule is for, and tracking whether the exercise routes actually get done is where most programs quietly fail.
The numbers worth knowing
Valve maintenance rarely gets its own budget line, so the case for it leans on general maintenance-strategy data plus the valve standards that define acceptable leakage:
- The U.S. Department of Energy's Federal Energy Management Program O&M Best Practices guidance reports that running equipment to failure (reactive maintenance) typically costs several times more than planned maintenance, and that a functioning predictive program saves roughly 8–12% over preventive maintenance alone. Valves that fail closed on a running line are a textbook reactive cost.
- Low-emission valve packing is type-tested to API 624 and API 641 with an acceptance limit of under 100 ppmv measured stem leakage, the target to specify when you replace packing in hazardous service.
- Seat leakage is graded against API 598 and ISO 5208 allowable-rate classes, from no-visible-leakage up through defined bubble or milliliter rates. Knowing your valve's required class tells you whether a passing seat is a defect or within spec.
Those standards are worth citing in your own procedures because they turn a subjective "it seems to leak a bit" into a pass/fail number a technician can act on.
How do you keep a valve program from going stale?
You keep it alive the same way you keep any route-based program alive: make completion visible and route every finding to a work order. Valve exercise routes are among the first things dropped when a plant gets busy, and because a seized valve is silent until the day you need it, nobody notices the gap until it costs a shutdown.
Three habits hold the line. First, track exercise-route completion as a percentage on the same board as your other maintenance KPIs a route that only happens half the time is not a program. Second, capture the effort and findings on every valve, so rising torque or a new weep becomes a trend, not a one-off note that dies on a clipboard. Third, close the loop: a passing seat found during an isolation gets a work order the same day, not a mental note. Plants that digitize those valve rounds instead of running them on paper, the way Harmony turns floor checklists into searchable, trendable records (see how that works), can finally see which valves are drifting and prove the exercise actually happened. It is the same shift that let the team in our CLS case study replace paper logs with searchable plant knowledge, and it ties valve health back into overall equipment reliability and your predictive maintenance strategy instead of leaving it as an afterthought.