Valve Troubleshooting: Leaks, Rising Torque, Vibration and Stuck Actuators

The most common mistake in the field is jumping straight from symptom to part: the valve drips, so the gland gets tightened; the torque feels heavy, so a bigger actuator gets ordered. In practice one symptom usually has three or four plausible causes, and the wrong fix wastes both money and downtime. The order that works is symptom, measurement, likely cause, and only then action.

Diagnosis is only measurable if every critical valve has a reference record. Break torque or motor current logged at commissioning, full stroke time, working pressure and body surface temperature form the baseline that every later reading is compared against. Without that baseline, the phrase 'the torque has gone up' carries no numerical meaning at all.

Symptoms rarely arrive alone. Rising torque usually travels with an over-tightened gland, and vibration often accompanies seat leakage, so a diagnostic sheet should be a small table of readings rather than a single tick box. The inspection routine in our guide to valve maintenance intervals sets how often those readings need to be collected.

A drip around the stem is the most frequent valve fault in industry and the one handled worst. Graphite or PTFE-based packing consolidates over time, loses volume and opens a path between the stem and the gland follower. The first action is to tighten the gland nuts crosswise in roughly one-sixth turn increments, cycling the valve once fully open and closed after each step to see what the leak actually does.

Tightening is an adjustment that buys life, not a repair, and it has a hard limit. Once the gland follower runs out of travel, once leakage returns after two or three adjustments, or once break torque climbs 25 percent above baseline purely because of gland load, the packing set has to come out. Over-tightening scores the stem and shortens the life of every set fitted afterwards.

When leakage keeps coming back at short intervals, the culprit is usually the stem rather than the packing. Scoring you can catch with a fingernail, ovality or a corrosion pit will defeat a new set within weeks. On volatile or hazardous service, stem sealing performance is classified under the ISO 15848 series, and the durable answer there is a live-loaded packing set or a bellows-sealed stem.

A weep at the body-to-bonnet joint usually comes from bolt preload lost through thermal cycling rather than from a failed gasket. When a line swings between 20 °C and 180 °C the bolts change length and preload falls away. The first action is to depressurise, let the valve cool, and re-torque the bolts crosswise to the manufacturer's figure; torquing under pressure is never acceptable.

If the joint still weeps after one proper re-torque, the gasket has to be replaced. Leakage through the body itself, however, is a different category: a crack in a weld, a casting void or a corrosion perforation is outside the scope of any field repair. The judgement is made by measurement: once the ultrasonic wall reading falls below the minimum thickness set by the applicable piping code or the manufacturer, the body no longer carries its pressure class safely and the valve is replaced.

Leakage past a closed valve should be reported as a number, not as the word 'leaking'. Close the valve, drain the downstream side, and collect what comes through the downstream drain in millilitres per minute or as a bubble count. Record the differential pressure alongside it: the same valve can hold tight at 2 bar and show measurable leakage at 10 bar.

A measured figure means nothing until it is compared with a class. For on/off valves, EN 12266-1 Rate A means no visible leakage for the duration of the test; for control valves, ANSI/FCI 70-2 Class IV allows up to 0.01 percent of rated capacity and Class II up to 0.5 percent. Our article on valve leakage classes sets out the test conditions behind those numbers.

Where the leak comes from depends on the valve type. On soft-seated ball and butterfly valves the first suspect is weld slag or a grain of sand trapped on the seating face; if two or three full cycles clear the leak, debris was the cause. If the leak returns after every cycle, the soft seat is worn, or on a metal-seated design the seating face has eroded.

Torque is an earlier warning than leakage and a far less monitored one. On manual valves, measure break torque with a torque wrench instead of estimating it from lever length and hand effort. On actuated valves, use an indirect indicator: motor current on an electric actuator, full stroke time on a pneumatic one.

Actuators are normally selected by adding a safety margin of 25 to 50 percent on top of valve break torque. A reading 25 percent above baseline can therefore wipe out the whole margin on a tightly sized unit, and a reading 50 percent above baseline pushes any actuator to its limit. At the second threshold the valve should be opened up at the next planned shutdown rather than run for another cycle.

Order matters when hunting the cause. Slacken the gland first to separate how much of the torque comes from packing load, then check the stem and bearings, and only then inspect the seat and the internal body surface. Dried-out soft seats, scale and sediment build-up, crystallising fluid and corroded stem bearings are among the most common causes. On isolation valves that operate only a few times a year, a partial stroke test every three months prevents the build-up in the first place.

Valve noise is not random; its character points at the cause. A steady metallic hum usually means high velocity, an irregular sound like gravel travelling through the pipe means cavitation, and a rhythmic slam means a check valve disc fluttering on its seat. Once readings at the measuring point pass 85 dB(A) you are past the occupational threshold where hearing protection is required; trim and body erosion, by contrast, is normally discussed at far higher levels, around 110 dB(A).

Separating cavitation from flashing changes the whole remedy. If downstream pressure recovers above the vapour pressure of the fluid, the bubbles collapse and you have cavitation; if it does not recover, the fluid stays two-phase and you have flashing. Cavitation is solved with multi-stage trim or by splitting the pressure drop across two valves, while flashing leaves only erosion-resistant materials and a rethink of the flow path as an answer.

Part of the vibration you hear is simply the valve running at the wrong point. A butterfly disc flutters in the flow path below roughly 15 to 20 degrees of opening, a control valve becomes unstable below 10 percent opening, and a check valve cannot hold its disc fully open below about half its rated flow. In all three cases the fault is the sizing, not the valve.

The fastest way to split the problem is to uncouple the actuator and turn the valve by hand. If the stem turns at normal effort, the fault sits on the actuator side; if it does not move, or has become noticeably stiffer, there is a mechanical obstruction inside the valve. That ten-minute test eliminates most of the unnecessary actuator replacements that plants pay for.

On the pneumatic side, supply pressure is the first thing to check. Output torque on a double-acting unit varies almost linearly with supply, so an actuated valve assembly fed at 4.5 bar instead of 6 bar gives away roughly a quarter of its rated torque. Work through the regulator setting, a clogged filter in the air preparation set and undersized tubing in that order.

On electric actuators, failure to reach position most often comes from the torque switch tripping before the limit switch. Travel stops before the valve is fully seated and a fault signal goes to the control system. A stroke time 30 percent longer than baseline is the early warning of the same family of problems. Always rule out a genuine torque rise in the valve before changing any switch setting.

The decision is arithmetic rather than sentiment. A packing and gasket kit costs a small percentage of a new valve on most designs, while a seat and trim kit is markedly more expensive. Once parts, labour and downtime together pass half the price of a new valve, replacement wins almost every time, and the process-side cost of that downtime belongs in the same sum.

Size shifts the answer as well. Below DN50, removal and refitting labour on a soft-seated ball valve competes with the price of the valve, so field repair rarely pays. Above DN200, where body and actuator costs dominate, reseating and repacking are the first option. A valve with compromised body integrity, a downgraded pressure class or no spare-parts availability is not a repair candidate at all.

Once replacement is decided, the real opportunity is to close out the root cause instead of fitting the same valve again: a control valve where an on/off valve was being throttled permanently, a metal seat on a crystallising service, a bellows stem where packing keeps failing. Within our valve and valve control systems scope, Vanera reviews your fault history alongside the process data to identify the equivalent or improved option.