A reliability engineer specifying the monitoring layer downstream of an Arrangement 2 or 3 dual seal sees two adjacent plan numbers in API 682 4th Edition. Plan 65 sits in the atmospheric drain; Plan 75 sits in the containment cavity between two seals. Vessel sizes, sensor mix, and alarm setpoints differ, and each instrument output maps to a specific failure mode with a specific action timescale.
How Plan 65 Monitors Atmospheric Leakage
Plan 65 is the atmospheric-drain plan. Leakage that crosses the seal collects in a 3-litre minimum reservoir (per API 682 4th Edition) at the 6 o’clock gland drain, and a level transmitter alarms when the reservoir fills. The 4th Edition splits the plan into 65A and 65B by one hardware swap — orifice vs valve — that controls whether the alarm fires on rate or on cumulative volume.
| Variant | Restrictor | Detection mode | Action timescale |
|---|---|---|---|
| Plan 65A | 5 mm orifice in vertical drain leg | Excessive leakage flow rate | Minutes to hours — isolate or switch to spare |
| Plan 65B | Normally-closed valve in place of orifice | Cumulative leakage volume | Days to weeks — scheduled maintenance |
On 65A, normal weepage passes through the orifice; abnormal flow exceeds what the orifice can pass and the alarm trips. A level rise over minutes means primary seal face failure — fracture, blister-out, abrupt loss of fluid film.
On 65B the closed downstream valve traps every drop (upstream valve locked open, downstream valve closed during operation). A reservoir that fills over weeks means slow weepage — face wear, secondary O-ring relaxation, slow groove erosion. The two opposing valve states are a frequent commissioning failure point: flip them and the plan does the opposite of what it’s designed for.
The 65A-vs-65B choice is operational, not hardware-preference. Sudden failures need 65A’s minutes channel; gradual ones need 65B, because the orifice would let slow weepage drain unnoticed.
How Plan 75 Monitors Containment Leakage
Plan 75 is the condensable-leakage collection plan for the cavity between two seals. Liquid leaking past the inner seal accumulates in a 7-litre minimum collector (per API 682 4th Edition). The vessel carries two channels for the same root cause at different timescales — a level transmitter and a pressure switch.
| Channel | Trigger | What it surfaces | Action timescale |
|---|---|---|---|
| Pressure switch | 0.7 bar gauge | Inner-seal degradation — leakage drives cavity pressure up against the outlet orifice | Hours — controlled shutdown |
| Level switch | Reservoir high level | Same root cause, slow channel — vapor condenses to liquid over hours to days | Days to weeks — scheduled maintenance |
| Plan 72 pairing | Level rise with buffer gas in cavity | Inner seal AND buffer-gas regulation both failed | Check buffer-gas supply first |
AESSEAL, Swagelok, Chesterton, and EagleBurgmann all pin the pressure-switch setpoint at 0.7 barg — the figure any 4th Edition-compliant install will use. A level trip without a prior pressure trip means leakage is well below the pressure-channel threshold: low flow, persistent.
Plan 75 also pairs with either Plan 71 (no buffer) or Plan 72 (buffer gas in the cavity). With Plan 72, process fluid leaking past the inner seal must overcome buffer-gas pressure to reach the collector, so a level rise indicates two failures, not one. The operator’s first check shifts to buffer-gas supply pressure before scheduling inner-seal inspection — a diagnostic step specific to API 682 dual-seal containment installations using Plan 72.
Why the Monitoring Layer Itself Can Hide a Failure
What the spec sheet doesn’t tell you is that every instrument above can suppress its own alarm — silently, usually right when the primary seal is degrading. Four field patterns recur:
- Clogged 5 mm orifice (Plan 65A). Polymerized condensate or coking residue accumulates in the orifice; it becomes more restrictive than spec and alarms hair-trigger on weepage that previously passed. The inverse failure is worse — complete clog, no flow, and the transmitter never sees rising fluid even when the seal fails catastrophically. Run a quarterly orifice clarity check.
- Frozen drain piping (outdoor 65A/B installations). Condensable leakage freezes in the vertical drain leg below the 6 o’clock connection. Flow stops; reservoir stays empty; the signal looks identical to a working seal. Heat-trace outdoor drains where ambient drops below the leakage fluid’s freezing point.
- Fouled level switch (Plan 65B / Plan 75). Polymerized condensate coats the float or capacitance probe. The switch reads low even as the reservoir fills, or stuck-high after a single weepage event. Inspect the probe at every drain cycle and bench-test the switch.
- Alarm fatigue from over-tight setpoint. If buffer-gas supply has 0.5 bar normal variability, a 0.7-barg pressure switch trips eight to fifteen times per shift on process noise. The 4th Edition’s switches-to-transmitters shift addresses this exactly: transmitters allow trending, which catches drift before threshold breach.
If you are still picking the primary flush plan upstream, the step-by-step flush plan selection guide walks that choice.
Where Most Specifications Go Wrong
The most common failure I see on Plan 65 and Plan 75 installs is treating each instrument output as a single binary — alarm or no alarm. Each output is one channel of information about one specific failure mode with one specific action timescale. Read each as a sentence with a verb and a deadline; an operator who treats any as “something is wrong, fix it later” is running a different plan than the one written on the P&ID.
